SOPHISMS IN M E D I E V A L L O G I C A N D G R A M M A R

Nijhoff International Philosophy Series

VOLUME 48

General Editor: J A N T. J. S R Z E D N I C K I
Editor for volumes on Applying Philosophy: R O B E R T O POLI
Editor for volumes on Logic and Applying Logic: S T A N I S L A W J. S U R M A
Editor for volumes on Contributions to Philosophy: J A N T. J. S R Z E D N I C K I
Assistant to the General Editor: D A V I D W O O D

Editorial Advisory Board:

L . Broughton {Lincoln University)', R . M . Chisholm {Brown University, Rhode

Island)', Mats Furberg {Göteborg University)', D.A.T. Gasking {University of
Melbourne)', H . L . A . Hart {University College, Oxford)', S. Körner {University of
Bristol and Yale University)', H J . McCloskey {La Trobe University, Bundoora,
Melbourne)', J. Passmore {Australian National University, Canberra)', A . Quinton
(Trinity College, Oxford); Nathan Rotenstreich {The Hebrew University,
Jerusalem)', Franco Spisani {Centro Superiore di Logica e Scienze Comparate,
Bologna)', R. Ziedins {Waikato University, New Zealand)

The titles published in this series are listed at the end of this volume.
 Sophisms in
Medieval Logic
and Grammar
Acts of the Ninth European Symposium
for Medieval Logic and Semantics,
held at St Andrews, June 1990

edited by

Stephen Read
University of St Andrews, Scotland

SPRINGER-SCIENCE+BUSINESS MESIA, B. V .
A C L P . Catalogue record for this book is available from the Library of Congress.

ISBN 978-94-010-4776-0 ISBN 978-94-011-1767-8 (eBook)

DOI 10.1007/978-94-011-1767-8

Printed on acid-free paper

All Rights Reserved

© 1993 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1993
Softcover reprint of the hardcover 1st edition 1993
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
 Sophisms in Medieval Logic and Grammar

Contents

ix Preface

xi Stephen Read, "Introduction"

Part I: Sophisms as a Genre

3 Robert Andrews, "Resoluble, Exponible, and Officiable Tenns in the
Sophistria of Petrus Olai, MS UppsaIa C 599"

17 Appendix 1

24 Appendix 2

31 Mario Bertagna, "Richard Ferrybridge's Logica: a handbook for

solving Sophismata"

45 Sten Ebbesen, Boethius de Dacia et aI. "The sophismata in MSS

Bruges SB 509 and Florence Med.-Laur. S. Croce 12 sin., 3"
56 Appendix

62 List of MSS

64 C. H. Kneepkens, "Orleans 266 and the Sophismata Collection:

Master Joscelin of Soissons and the infinite words in the early
twelfth century"

80 Appendix

86 Roberto Lambertini, "The Sophismata attributed to Marsilius of

Padua"

102 List of MSS

103 Alfonso Maierli, ''The sophism 'Omnis propositio est vera vel faIsa'
by Henry Hopton (Pseudo-Heytesbury's De veritate et falsitate
propositionis )"

115 Appendix
vi CONTENTS

116 Mieczyslaw Markowski, "Die Rolle der Sophismata im Unterricht

der Krakauer Universitat im 15. 1ahrhundert"

128 Fabienne Pironet, "The Sophismata asinina" of William Heytes-

bury"

141 Appendix

144 Paul A. Streveler, "A Comparative Analysis of the Treatment of

Sophisms in MSS Digby 2 and Royal 12 of the Magister
Abstractionum"

154 Appendix 1

168 Appendix 2

185 Andrea Tabarroni, '''Omnis phoenix est': Quantification and

Existence in a new Sophismata-collection (MS Clm 14522)"

200 Appendix

202 Mikko Yrjonsuuri, "Expositio as a method of solving sophisms"

Part II: Grammatical Sophisms

219 Christine Brousseau-Beuermann, "Grammatical sophisms in
collections of logical sophisms: 'Amatus sum' in BN.lat. 16135"

231 Irene Rosier, "La distinction entre actus exercitus et actus significatus
dans les sophismes grammaticaux du MS BN lat. 16618. et autres
textes apparentes"

260 Appendix

262 Mary Sirridge, "Interest mea et imperatoris castam ducere in uxorem:

can 'est' be used impersonally?"

Part III: Logical Sophisms

277 Allan Back, "Who is the worthiest of them all?"

288 10151 Biard, "Albert de Saxe et les Sophismes de l'Infini"

304 Alessandro D. Conti, "II Sofisma de Paolo Veneto: 'Sortes in

quantum homo est animal'''

313 Appendix
 CONTENTS vii

319 Jeffrey S. Coombs, "The Soul of the Antichrist necessarily will be

a being: A modal sophism in 16th century logic texts"
333 Gyu1a Klima, '''Debeo tibi equum': a reconstruction of the theoretical
framework of Buridan's treatment of the sophisma"

348 Simo Knuuttila, "Trinitarian Sophisms in Robert Holcot's Theology"

357 Christopher J. Martin, "Obligations and Liars"

379 Appendix

382 Angel d'Ors, "Hominis asinus/Asinus hominis"

398 Claude Panaccio, "Solving the insolubles: hints from Ockham and
Burley"

413 Index ofManuscripts

418 Index of Names

 Preface

The first Symposium consisted of three people in a cafe in Warsaw in

1973. Since then, meetings have grown in size and have been held in
Leyden, Copenhagen, Nijmegen, Rome, Oxford, Poitiers and Freiburg-
am-Breisgau. The ninth Symposium was held in St Andrews in June
1990, with 57 participants who listened to addresses by 28 speakers. It
was very fitting that Scotland's oldest university, founded in the heyday of
medievalleaming in 1411, should have been given the chance to bring
together scholars from all over Europe and beyond to present their
researches on the glorious past of scholastic rational thought. The topic of
the Symposium was "Sophisms in Medieval Logic and Grammar".

The present volume consists, for the most part, of the papers
presented at the Symposium. In fact, however, it proved impossible to
include five of the contributions. Two of the papers included here were
intended for the Symposium but in the event not delivered, because of the
unavoidable absence of the speakers.

The Symposium received very helpful financial support from one of

the major philosophical associations in Britain, the Mind Association, from
the Philosophical Quarterly, a journal published at St Andrews, from the
University of St Andrews, from the British Academy, and from Low and
Bonarplc.
In organising the programme for the conference and in preparing the
papers for publication I received invaluable help from: Professor E.J.
Ashworth of the University of Waterloo, Canada; Professor Henk
Braakhuis of the Katholieke Universiteit, Nijmegen; Professor Klaus
Jacobi of the Albert-Ludwigs Universitiit, Freiburg; Professor Alfonso
Maierii of the Universita "La Sapienza", Rome; Professor D.P. Henry,
now retired from the University of Manchester; and especially from Dr
Sten Ebbesen of the Institute for Medieval Greek and Latin Philology,
Copenhagen.

The text was prepared on an Apple Macintosh, using Microsoft Word

4.0.

University of St Andrews, July 1992

ix
 Introduction

by Stephen Read

Increasing insight into the medieval genre of sophisms has been

acquired in recent years. Nonetheless, as more is discovered, more
puzzlement arises, and yet more questions are prompted. It was in this
context that three years ago, having offered to host the Ninth European
Symposium for Medieval Logic and Semantics at St Andrews, I proposed
the topic of Sophisms in Medieval Logic and Grammar as the theme for the
conference. It satisfied two criteria: first, it provided a stimulus to the
scholars involved in this regular symposium to work on a topic of much
current interest during the run-up to the meeting, and to present papers
which, particularly as an ensemble, would together shed more light on the
problem. Secondly, it provided a focus, one narrow enough to give the
conference, and the conference volume, a unity, a clear expression of
scholars working together towards a common end; and at the same time a
variety within that common theme, ranging from close textual and
historical study of the actual genre of sophismatic treatises through to the
forever fascinating content of the sophismatic puzzles themselves.

Such is the origin, and the rationale, of the present volume. I believe
the project was successful. Reading through the contributions, one comes
away with a much clearer picture than before of the contribution sophisms
made to the richness of medieval logic and thought. "What is left in logic
which is untouched by British sophisms?", wrote Leonardo Bruni of
Arezzo in his first Disputation dedicated to Pier Paolo Vergerio in 1401. 1
Sophisms played a crucial role in medieval logical and grammatical theory
in providing the spur for investigation, insight and invention.

John Marenbon questions whether "exponents of what might be called

the 'modem analytical' approach of thought in the Middle Ages" are right
to "concentrate on philosophical problems which they believe they share
with medieval scholars".2 The answer is that despite the medievals'
engagement with issues of very particular concern to them, there is indeed
a remarkable similarity both in the problems which they tackled, and in
their methodological approach to them. The roots of our thought lie in
Greek and Roman antiquity, and that heritage is mediated by medieval
philosophy, logic and grammatical theory. The influence of medieval
thinking on our own is too little appreciated, but it is considerable.
Examination of this background well repays the effort, and has moreover a
fascination all its own. The medievals had their own particular problems,
and their own strange methods; but they share many of their central

1L. Bruni, Ad Petrum Paulum Histrum dialogus I, in Prosatori Latini del Quattrocento,
ed. E. Garin, Milan-Naples: Riccardo Ricciardi Editore 1952, pp. 58-60: "Quid est,
inquam, in dialectica quod non britannicis sophismatibus conturbatum sit?"
2J. Marenbon, Later Medieval Philosophy (1150-1350): An Introduction, London:
Routledge and Kegan Paul 1987, p. 86.

xi
xii STEPHEN READ

problems with us, and their overriding method, in logic and grammar, of
analysis, puts them methodologically in touch with modem philosophy of
language. Here are some of history's deeper thinkers - Ockbam, Buridan,
Abelard, St Thomas - tackling our problems - what exists, can processes
proceed infinitely, is self-reference acceptable, what follows from a
contradiction? All these questions are found in sophismatic treatises.

There are three classic treatments of the history of sophismatic works,

starting with Martin Grabmann's monograph published in 1940.3 At the
same time he focused attention on that important Danish thinker at the
University of Paris, Boethius of Dacia. This work was followed up by De
Rijk's monumental three-volume study, Logica Modernorum, in 1962 and
1967, in which he mapped out the origins of terminist logic in the study of
fallacies in the twelfth century.4 Finally, one must note HenkBraakhuis'
dissertation on Syncategoremata, published in Dutch in 1979.5 There.is a
close and intimate connection between sophistic fallacies, the study of
syncategorematic words and the development of the theories of properties
of terms.

But the real growth in studies of sophisms has come only in the last
decade or so. We are still awaiting the publication of Alfonso Maieru's
study of Methods of Teaching Logic during the Period of the
Universities. 6 But there have been important studies by Ebbesen, de
Libera, the Kretzmanns, Rosier and Tabarroni, to name but a few. This
volume, it must be acknowledged, takes that study, and our understanding
of the role of sophisms in medieval thinking, yet further.

An initial question is whether to render' sophisma' as 'sophism', or to

retain the Latin term. Different authors in the present volume adopt each
course. The problem about the word 'sophism' is that, for the English
reader, it threatens to import implications which are unwarranted and not
present for medievals in Latin usage. De Rijk spoke of "ambiguous
propositions") Their role was to cause puzzlement, to invite theoretical
development in the search for clarification. But they did not have the
implication of casuistry, or even necessarily of fallacy. The problem was,
rather, that there seemed to be ways of showing both that they were true
and that they were false. Insight was needed to see which argument was
correct, or to show that, by making a suitable distinction, both were
correct, in appropriately different senses. For example, in Mario

3M. Orabmann, Die Sophismatalitteratur des 12. und 13. lahrhunderts mit Textausgabe
eines Sophisma des Boetius von Dacien, Beitriige zur Geschichte der Philosophie und
Theologie des Mittelalters 36.1, MUnster: Aschendorff, 1940.
4L.M. De Rijk, Logica Modernorum. A Contribution to the History of Early Terminist
Logic, Vol. I: On the Twelfth Century Theories of Fallacy; Vol. II, 1: The Origin and
Early Development of the Theory of Supposition; Vol. II, 2: Texts and Indices, Assen:
Van Oorcum 1962-7.
5H.A.O. Braakhuis, De I3de Eeuwse Tractaten over Syncategorematische Termen.
Inleidende studie en uitgave van Nicolaas van Parijs' Sincategoreumata, Deelll, Ph.D.
Leiden, Meppel: Krips Repro. 1979.
6See below, p. 104 n. 12.
70p .cit., vol. II part I, p. 595.
 INTRODUCTION xiii

Bertagna's paper below, it would appear to follow from Ferrybridge's

theory of truth that the proposition 'Every man is' is equivalent to a
conjunction of propositions of the form 'x is', for each man x existing at
the time of utterance. But each conjunct is only contingently true, whereas
the original proposition, 'Every man is' seems to be necessarily true, for
necessarily every man who is, is. Ferrybridge resolves the difficulty by
distinguishing between the meaning (significatio) and the reference
(significatum) of the proposition. Although the reference of the proposition
changes with time, its meaning, and so its necessity, do not vary.

Another very common distinction that was appealed to was that

between composite and divided senses. For example, in Paul Streveler's
paper we find that, to solve the sophism 'Every man is every man', the
Magister Abstractionum distinguishes the composite sense, in which the
aggregate or collective is referred to, and the proposition is true, from the
divided sense, in which each individual man is separately said to be every
man, and so is false. Again, in Jeff Coombs' article, a similar distinction is
drawn concerning the scope of 'necessarily' in 'The soul of the Antichrist
necessarily will be a being', such that if 'necessarily' has narrower scope
than 'will be', the proposition is true (for when the Antichrist comes to be,
he necessarily will be), while if it has wider scope, the proposition is false
(for there is no necessity that the Antichrist ever will be).

To render the Latin 'sophisma' I shall, for my own part, use the
English term 'sophism'. We should also note that the Latin term was used
both for the "ambiguous proposition" on which attention was focused, and
also for the whole discussion - the proof, disproof and resolution which
we shall look at below. Provided we accept that there is more to a medieval
sophism than a mere quibble, and that many examples are not in the least
bizarre, we shall not be misled.

It should also be noted that three rather different medieval treatments

can all be described as the study of sophisms, and different authors in the
present volume treat of different aspects. First, there is the use of sophisms
at certain points in a work of a different nature: for example, sophisms may
come into a work on supposition theory, to point up a particular problem.
In his paper for the volume, Allan Back looks at discussions of the
sophismatic proposition 'Man is the worthiest of creatures' as a test of an
author's theory of supposition. William of Sherwood thought 'man' here
must have simple supposition, so that it stood for the species - but by
Ockham's lights the proposition would then be false. The only alternative
for Ockham seemed to be to accord 'man' personal supposition, to stand
for individual men - but then which man is the worthiest? Ockham had to
produce a paraphrase: that every man is worthier than every other type of
creature. The hard-nosed rejection, the paraphrase and the reduction
exhibited here are all part of the nominalists' armoury.

Rather different from such use of sophisms is the treatment by a single

author of a succession of sophisms, as for example in John Buridan's or
Albert of Saxony's Sophismata. An example of one such sophism treated
by several authors appears in the collection of sophisms by William
Heytesbury examined below by Fabienne Pironet. It is now known as
Curry's paradox, having been rediscovered in 1940 by Haskell B. Curry.
xiv STEPHEN READ

Heytesbury's treatise has an attractive stylistic unity, in focusing

successively on different problematic arguments all engineered to deliver
the absurd conclusion, 'You are an ass'. The argument in question takes
the form 'This argument is valid, so you are an ass'. Suppose it were
valid; then its conclusion would follow, by Modus Ponens, so by
Conditional Proof, the argument really is valid. So you are an ass, by
Modus Ponens again. The sophism was recognised to be an insoluble, and
is here solved by a restrictive rule, that an insoluble cannot refer to itself.
Claude Panaccio, in the final paper in the volume, extends the restrictive
rules found in Burley and (implicitly) in Ockham, to propose a general
solution to such sophisms.

Finally among these different medieval treatments, there is the text

which consists of a collection of sophisms by various authors, often
showing evidence of oral debate, but often also brought together and
unified by the intervention and comments of the editor. Andrea Tabarroni
examines such a treatise. The author of one of the sophisms (number 9) in
the collection claims that propositions such as 'A man is an animal' will be
false when their subject terms fail of reference - if there are no men. The
author of another (number 6) says that such propositions are true even if
there are no men, for animal is part of the essence of man. This and other
indications show that we are dealing with a collection of multiple
authorship.

For sophisms had a significant role in the medieval curriculum. After

attending lectures on grammar, logic and rhetoric, perhaps for two years, a
student would be required to serve first as opponent, then for a year as
respondent, in a series of disputations. These took place under the
direction of a master, who had in many cases already lectured in the
morning, and then oversaw discussions in the afternoon. A sophism
would be set out, with a hypothesis (casus) - "let's suppose ... " - and
proof and disproof (probatio and improbatio). This would create the
problem, for the opponent first to attack, and for the respondent to try to
resolve. In between, there might be various subsidiary questions; by the
end of the thirteenth century, and particularly in treatises which show
literary and editorial intervention and addition to the oral debate, these
questions could come to dominate the discussion. In the simpler cases, the
treatise may end with the respondent's resolution, and perhaps replies to
the opposing arguments. In other cases, the master may intervene. In such
cases, the debate may be opened by the master who at some point will
dismiss the student's resolution, closing day 1 of the debate. On day 2, the
master returns, and "determines" the sophism at length.

Thus we can set out the general form of a sophismatic disputation as

follows:

Hypothesis
Proof(s)
Disproof(s)
(Questions)
Resolution
(Replies to opposing arguments)
(Determination)
 INTRODUCTION xv

The optional components are given in parentheses. In one of the most

complex examples of disputation, the sophism on the nature of logic by
Bartholomew of Bruges, written at Paris in the first decade of the
fourteenth century, the questions cover 74 of the 76 printed pages. 8 In
fact, if one is surveying the whole range of sophismatic treatises, there is
an indefinite variety of ways in which the elements can be organised. There
may be a whole series of problems posed, discussion of them and eventual
rejoinder. Sometimes the responses will be placed after each question or
problem, though more commonly they are collected at the end.

After a year or so as respondent, the student could proceed to the

degree of bachelor, in which he could himself lecture under the guidance of
his master. Some three years of lecturing, opposing and responding would
then lead to his being allowed to incept as a master himself.

Sophisms were mainly used for teaching logic. There are also
grammatical sophisms and sophismatic treatises, physical sophisms and
theological sophisms. Even the physical sophisms, however, are very
close to logical ones, treating problems of, say, infinity or continuity
essentially as logical and conceptual problems.9 For example, when Albert
of Saxony treats the sophism, 'The infinite are finite', he clarifies the
problem by appeal to the logical distinction between taking 'infinite'
categorematically and syncategorematically. Joel Biard presents the
solution in his paper: Albert identifies three senses for the sophism, in two
of which it is true, namely, when 'infinite' is taken, categorematically,
with respect to number and 'finite' with respect to magnitude - for each of
the infinitely many is finite in size; and when 'infinite'is taken
syncategorematically - for however many finite things one has, there are
yet more. Indeed, whether the possibilities discussed were physically
impossible was largely ignored, the issue being whether they were
logically possible.

The heyday of these treatises recording real disputations is the

thirteenth century. Indeed, on the whole those which survive have come
from Paris. What survive from Oxford are literary works showing less
evidence of original live disputations. Nonetheless, what evolved out of
the discussions of Aristotle's De Sophisticis Elenchis (On Sophistical
Refutations) in the twelfth century was a rich literature of sophismatic
treatises, firmly grounded in the university curriculum. Right through to
the sixteenth century there were little booklets (libelli sophistarum)
designed to help the student in the art of disputation and resolution of

8S. Ebbesen and J. Pinborg, "Bartholomew of Bruges and his sophisma on the nature of
logic", Cahiers de l'Institut du Moyen-Age Grec et Latin 39,1981, pp. iii-xxvi, 1-76.
The structure is set out there on pp. xxiv-xxv.
9See, e.g., N. & B.E. Kretzmann, The Sophismata of Richard Kilvington: Text Edition,
Oxford: Oxford University Press for the British Academy 1990; idem, The Sophismata
of Richard Kilvington: Introduction. Translation and Commentary, Cambridge:
Cambridge University Press 1990.
xvi STEPHEN READ

sophisms.1O To do this, a whole armoury of distinctions and theoretical

concepts drawn from the "properties of terms" was needed.

Some indeed of these treatises are actually entitled Distinctiones

Sophismatum, with a special title in the thirteenth century of Sophisteria or
Sophistria (in the fourteenth century this title came to be used simply as an
alternative to Sophismata, for a single sophismatic treatise). So there are
connections between the various treatises called Distinctiones,
Abstractiones, Syncategoremata, Sophistria, Sophismata, De
Obligationibus, De modo opponendi et respondendi and De consequentiis.

Nonetheless, the treatises we have, and in particular, the treatises

which have been analysed and examined, are only a scattered record of the
work that went on from the late twelfth century to at least the late
fourteenth century - and to some extent later. So it is only by more
research that we can begin to build up an accurate picture of the treatment
of sophisms. In 1989, Andrea Tabarroni gave a counterexample to the
claim that the early treatises were only models, and that actual records of
debate dated only from the time of Boethius of Dacia in the 1270s. 11 He
exhibited a record of debate from the 1250s. Again, Alain de Libera has
only recently corrected the belief that terminist logic was overtaken by
modistic work at Paris in the 1270s by looking closely at two complex
manuscripts. 12 They show that there was a terminist and non-modistic
tradition at Paris from 1250 up to 1270-5, which preserved the teaching of
the masters of the early thirteenth century such as Jean Ie Page.

Another problem thrown up by these sophismatic treatises as we try to

reconstruct the history of logical doctrines in the middle ages, is the very
notion of "the real text".1 3 It is clear from many of the treatises that there
was a real debate which lies beneath it. But that may itself have been a
succession of different debates under different masters, and the literary
rendition may also have been carried out more than once, giving rise to
different successions of related manuscripts, from which no single Urtext
can be recovered. The problem is similar to, but even trickier for the editor
than, the differing reportationes of the lectures of a master.

I have separated the papers collected here into three main categories.
First, in Part I there are those papers which deal largely with the
sophismatic treatises themselves, their historical origins and their structure
and development. Of course, even here the sophisms themselves also play

!OSee EJ. Ashworth, "The 'Libelli Sophistarum' and the use of medieval logic texts at
Oxford and Cambridge in the early sixteenth century", Vivarium 17,1979, pp. 134-58.
11 A. Tabarroni, '''Incipit' and 'desinit' in a thirteenth-century sophismata-collection",
Cahiers de l'lnstitut du Moyen-Age Grec et Latin 59,1989, pp. 61-111.
12A. de Libera, "La litterature des Sophismata dans la tradition terministe parisienne de
la seconde moitie du XIIIe siecle", in The Editing 0/ Theological and Philosophical
Texts/rom the Middle Ages, ed. M. Asztalos, Stockholm: Almqvist and Wiksell 1986,
pp.213-44.
13See e.g., M. Asztalos, "Introduction" in idem (ed.), op.cit., p. 8, and S. Ebbesen,
"Corpus Philosophorum Danicorum Medii Aevi, Archbishop Andrew (t1228) and
twelfth-century techniques of argumentation", ibid., pp. 267-9.
 INTRODUCTION xvii

an important role, but I have brought these papers together since their
dominant theme is the genre of sophisms itself, the use of sophisms in
treatises as their central theme. I hope it will be helpful for the reader to
consider these papers together, separated from those whose focus is on a
particular sophism and its logical or grammatical character.

These other papers I have separated into a small group in Part II which
treat of grammatical sophisms, propositions whose rationale was the
pointing up of a grammatical moral, and a larger group in Part illdealing
with logical sophisms. These papers are less concerned with the historical
origins and form of the treatises themselves, and more with the puzzles
and, primarily, the theories and solutions proposed.

A recurrent theme throughout the conference was the phoenix: the

mythical bird whose existence is forever singular, which must die in the
ashes before its successor is born. The first and last papers at the
conference (Andrews' and Tabarroni's) focused on the ontological
problems to which the existence of the phoenix gives rise. Tabarroni
shows how successive authors took 'Every phoenix is' first to be
grammatically incorrect, on the grounds that 'every' may only be used with
terms actually referring to many; later to be false, for even if used with
such terms as 'phoenix' or 'sun', at least three things must be present for
'every' truly to apply; and finally to be true, for 'phoenix' is different from
'sun' and 'world' in that there are successively different phoenixes, and
'every' only requires that there be in principle more than one. The phoenix
sophism rose again and again from the ashes of medieval theories - it
appears also in the collections studied by Sten Ebbesen and Paul Streveler.
The phoenix can, I think, serve also as a trope, a metaphor for the
recurrent interest which sophisms spark in the logician's and grammarian's
mind. Speaking of the three genres of Distinctiones, Syncategoremata and
Sophisteria, De Rijk wrote: "All three ... comprise tracts which afford a
rich collection of logico-semantical materials which are of paramount
importance for every student of medieval logic and semantics."14 The
authors collected here succeed, I believe, both separately and collectively,
in pushing forward our understanding of these issues of recurrent
fascination.

University of St Andrews

14L.M. De Rijk, Some Earlier Parisian Tracts on Distinctiones Sophismatum,

Nijmegen: Ingenium 1988, p. xi.
 Part I

Sophisms as a Genre
Resoluble, Exponible, and Officiable Terms in the Sophistria
of Petrus Olai, MS Uppsala C 599

by Robert Andrews

As the only surviving collection of sophismata in Sweden, the

Sophistria of Petrus Olai is of significance for the cultural and intellectual
history of northern Europe in the late Middle Ages. For the theme of the
Ninth European Symposium, its chief interest lies in the interconnection
between its sophisms and its interspersed questions drawn from various
logical genres. Petrus Olai reveals his interests in the sophismata by his
choice of questions. Similarly, a historian of medieval philosophy often
chooses a subject because of its interest for current philosophy, as I have
done by selecting the sophism 'Omnis phoenix est', and the question
following it on resoluble, exponible, and officiable terms, with the intent
of indicating parallels to contemporary analysis of language.
MS Uppsala C 599

The manuscript which I have studied, Uppsala C 599, has had the best
imaginable preparation, by Professor Anders Piltz in his Studium
Upsalense. He has provided a manuscript description, a table of contents
with excerpts, a study of the historical background of the texts, and even a
paleographic dictionary of abbreviations distinctive to the manuscript. For
the final work of the manuscript, Petrus Olai' s Sophistria, he has prepared
a list of the sophismata and questions it comprises, and has edited its first
question, on whether a science of sophismata is possible)

The occasion of Professor Piltz's treatment of the manuscript was the

SOOth anniversary of the University of Uppsala, since it is among the
earliest sets of lecture notes from that medieval University. Its content
reflects the intellectual environment of the university, which was much like
other northern European universities in the late 15th century; one of the
German universities was probably the university's model. Doctrinally, the
course of instruction in philosophy was predominantly realist.
Petrus Olai and his works

The author, Petrus Olai, was a somewhat anomalous figure at the

University, because of his interest in the thought of the moderni and the
terminist logicians. All that can be reconstructed about Petrus Olai's life is
presented in Piltz. Petrus, whose vernacular name might have been Per
Olafsson, possibly studied at the University at Rostock, as did many
Swedish students of the time. The lecture notes we possess show him to
be a master at the University of Uppsala sometime between its inception in

I A. Piltz, Studium Upsalense. Specimens of the oldest lecture notes taken in the
medieval University of Uppsala, Uppsala: Acta Universitatis Upsaliensis 1977, p. 297:
"Utrum noticia sophistica, sub arte sylogistica que vere continetur, scientia vocari
possitque appellari, an ei denegetur."

3
4 ROBERT ANDREWS

1477, and 1486, the last dated entry in our manuscripts. 2 Scattered
allusions in his works indicate his conversance with Swedish locations,
including mentions of Uppsala,3 and of Stockholm in the sophisma
sentence •totus thesaurus H olmensis est in cista mea' ,ambiguous between
"the whole of Stockholm's treasure is in my coffer" and "all the treasure I
have obtained from Stockholm is in my coffer."4

Otherwise all that is known of Petrus Olai is his works. Attributed to

Petrus are commentaries on De interpretatione, Sophistici elenchi, and the
Physics; tracts within the parva logicalia on supposition, ampliation,
restriction, appellation, and consequences; and finally the Sophistria.5

His works are markedly derivative. His Physics commentary (MS C

601 ff. 225ra-295rb) explicitly states its dependence on the commentary by
John of Jandun. Its list of questions is almost the same, and many of the
arguments are the same, although sometimes with greater sophistication.6
The commentary on the Sophistici elenchi attributed to Petrus Olai (MS C
599 ff. 214ra-266vb), although not acknowledging the fact, is a verbatim
version of the Expositio supra libros Elenchorum of Giles of Rome,
omitting only the interspersed dubia.7

The treatises on the parva logicalia (in MS C 599) explicitly indicate

their sources of inspiration. The De suppositionibus (ff. 2ra-25va) is based
upon Thomas Maulevelt. The De ampliationibus (ff. 25vb-28va), based
mainly on Maulevelt, also credits the same work by Marsilius of Inghen.
The latter of two questions De restrictione (f. 29ra-va) is based upon the
corresponding section in Albert of Saxony's Perutilis logica. The former
question (ff. 28va-29ra) is of interest because it is said to be based on
Maulevelt, whose work De restrictione has not been recovered. s The De
appellationibus (ff. 29va-32va) discusses the different definitions of
appellation from John Buridan, Albert of Saxony, and Marsilius of
Inghen. The De consequentiis (ff. 32va-36va) takes its starting point from
Marsilius of Inghen.

The Sophistria, with which I shall be primarily concerned, is a curious

amalgam of works. As noted by De Rijk in Some Earlier Parisian Tracts on
Distinctiones Sophismatum, 9 the title Sophistaria or Sophistria here does
not imply a concentration on sophismatic distinctiones as a starting point,

2piltz, Studium Upsa/ense, p. 12.

3/n libros De interpretatione, MS Uppsala C 600 ff. 58rh-68rb; Piltz, Srudium
Upsalense, p. 47: 'Ego fui Vpsalie'; Sophistria, MS Uppsala C 599 f.299vb: 'tu es
Rome, ergo non es in Vpsalia', Piltz, Studium Upsa/ense, p. 33.
4Sophistria, MS Uppsala C 599, f. 290va; Piltz, Studium Upsalense, p. 33.
5Cf. Piltz, Studium Upsalense, pp. 41-9.
6Piltz, Studium Upsalense, pp. 31-2.
7Piltz, Studium Upsalense, p. 45.
SThe work De restrictione is attributed to Maulevelt in L. Thorndike, University
Records and Life in the Middle Ages, New York: Columbia U.P. 1949, pp. 296-7;
cited by Piltz, Studium Upsalense, p. 44 n.78.
9L.M. De Rijk, Some Earlier Parisian Tracts on Distinctiones Sophismatum.
Nijmegen: Ingenium 1988, p. x.
 THE SOPHISTRIA OF PETRUS OLAI 5

but indicates a series of exercitia, meaning questions, on sophismata.I o

The work consists of alternating sophismata and questions. The sophisma
sentences throughout are based on Albert of Saxony's Sophismata. The
first set of questions is based on Richard Billingham's Speculum
puerorum. The next set is based on the second part of Marsilius of
Inghen's Consequentiae ("tractatum posteriorum Marsilii de Ingen 'De
consequentiis'''), and thus serves as a complement to the De consequentiis
earlier in the manuscript, which was based on the first part ("in tractatum
priorum Marsilii de Ingen 'De consequentiis"', ff. 32va-36va). At the end
of the work are questions derived from the Obligationes and Insolubilia of
Marsilius of Inghen and Albert of Saxony.

The sophismata and questions alternate. In order to examine their

relative connections, I have edited one sophisma, 'omnis phoenix est', and
the question following it on resoluble, exponible, and officiable terms,
which are provided as Appendices (pp. 17-30 below).
History of the sophism 'Omnis phoenix est'

Treatment of the sentence' omnis phoenix est' dates back to the time of
early terminist logic, and in Professor De Rijk's Logica modernorum it
appears several times. The interest in the sentence was from the beginning
in the application of the term 'omnis' to a term with a single referent.
Grammatically 'uterque' should be expected to apply to dual referents, and
'omnis' to three or more. Aristotle in the De caeto was taken as the
authority that 'omnis' requires at least three referents'! I The origin of the
sophisma seems to be in considering the peculiar instance of the phoenix
which, according to mythology, has only one exemplar alive at any time.1 2
Aristotle's position seems to militate against the admissibility of the
sentence' omnis phoenix est'.

Among the earliest authors who discussed the scope of 'omnis',

William of Sherwood,13 the Magister Abstractionum,14 and Roger
Bacon l5 considered that at least three supposita should be required for the
term it governs. They therefore considered the sentence 'omnis phoenix
est' to be false and unacceptable. The author of the Tractatus Anagnini
went so far as to declare that 'omnis' taken in the plural requires at least six
suppositaP6 On the opposite side was Peter of Spain, who denied that

IOCf. Piltz, Studium Upsalense, pp. 31-2.

II Aristotle, De caelo I, I 268aI7-9.
12Myths about the phoenix were available from Cicero, and for example appear in the
De animalibus of Albert the Oreat. In the literature of the time, similar questions were
asked about other unique objects such as the sun and the world (mundus).
130uilIelmus de Shireswode, Syncategoremata, ed. J.R. O'Donnell, "The
Syncategoremata of William of Sherwood", Mediaeval Studies 3, 1941, pp. 46-93, p.
49.
14Magister Abstractionum, Abstractiones, MS Digby 24, f. 62ra; cf. MS Digby 2, f.
113ra.
15Roger Bacon, Summa de sophismatibus et distinctionibus, in Opera hactenus inedita
Rogeri Baconi, fasc. 14, ed. R. Steele. Oxford: Clarendon Press 1937, pp. 144-5.
16Tractatus Anagnini, in L.M. De Rijk, Logica Modernorum. A Contribution to the
History of Early Terminist Logic, Vol. I: On the Twelfth Century Theories of Fallacy;
Vol. II, 1: The Origin and Early Development of the Theory of Supposition; Vol. II,
6 ROBERT ANDREWS

'omnis' requires three supposita, for the reason that a universal term is
proportionate to its individuals. 17

Peter's lead was followed by most authors.1 8 Some other writers

who, in the context of discussing •omnis phoenix est', allow that •omnis'
requires only one suppositum include William of Ockham 19 and Albert of
Saxony,20 who is the explicitly credited source for Petrus Olai. There are
other writers on this sophisma whom I have not been able to examine,
such as Peter of Auvergne2 1 and the author of the anonymous Sophismata
associated with the Logica Cantabrigiensis. 22 I would like to note that there
is another sort of treatment of the sophisma •omnis phoenix est' which is
not concerned with there being a single phoenix at anyone time; the
tradition instead seems to concentrate on the historical sequence of
phoenixes, and so concerns generally the application of 'omnis' to things
in the past. Treatments of this type are to be found in Walter Burley23 and
in William Heytesbury and his commentator Gaetanus de Thienis.24
Petrus Olai on 'omnis phoenix est'
The immediate source for Petrus Olai' s treatment of the sophisma was
Albert of Saxony, but other traditions enter in. By identifying sources-

2: Texts and Indices, Assen: Van Gorcum 1962-7; vol. 11,2, p. 300: "Et notandum
quod, sicut hec dictio 'omnis' in singulari exigit tria adminus habere appellata, ita et in
pluraIi sex adminus exigit habere appellata."
17Petrus Hispanus, Tractatus, Called Afterwards Summule logicaies, ed. L.M. De Rijk,
Assen: Van Gorcum 1972, XII, 7-9, pp. 212-6; cf. p. 214: "nullum universale excedit
sua individua neque exceditur ab eis."
18This is so if it was Peter whose opinion had such influence. The anonymous De
solutionibus sophismatum, in De Rijk, "Some Earlier Parisian Tracts", p. 75, if it
indeed can be dated to around 1200, would be by far the earliest known proponent of
this opinion. Its dating is by H.A.G. Braakhuis, De 13de eeuwse tractaten over
syncategorematische termen. Meppel: Krips Repro. 1979, pp. 37-42; cited by De Rijk,
Some Earlier Parisian Tracts, p. xxiv.
19In William of Ockham, Summa Logicae, ed. P Boehner, G. Gat and S. F. Brown,
Opera philosophica, vol. I, St. Bonaventure, N.Y.: The Franciscan Institute 1974, p.
261.
20Albert of Saxony, Sophismata, Paris 1502, f. [3ra].
21Petrus de Alvernia, Sophisma 'Omnis phoenix est', MSS Vat. lat. 14718 (olim in
Frag. MSS Archiv. Vat.), ff.7r-8v; Firenze B. Med.-Laur., S1. Croce, plut. 12 sin., 3,
ff. 67v-69r; Brugge Stadsbibliotheek MS 506, ff.99v-l04r; cf. S. Ebbesen and J.
Pinborg, "Studies in the Logical Writings Attributed to Boethius de Dacia", Cahiers de
l'Institut du Moyen-Age Grec et Latin 3, 1970, pp. 1-54, pp. 9-10; cf. Cahiers de
l'Institut du Moyen-Age Grec et Latin 57,1988, p. 171; cf. Siger de Brabant Ecrits de
logique, de morale et de physique, ed. B. Bazan, Philosophes Medievaux xiv, Louvain:
Publications Universitaires 1974, p. 8.
22Anonymous, Sophismata, MS Cambridge Gonville et Caius 182/215, f.161: 'Omnis
fenix est'; cf. L. M. De Rijk, "The Place of Billingham's 'Speculum puerorum' in
14th and 15th Century Logical Tradition, with the Edition of Some Alternative
Tracts", Studia Mediewistyczne 16, 1975, pp. 99-153, p. 303.
23Gualterus Burlaeus, Quaestiones in librum Perihermeneias, ed. S. F. Brown,
Franciscan Studies 34, 1974, pp. 200--95; Quaestiones quarta et quinta: 'Omnis
phoenix est', pp. 260--95.
24Gulielmus Hentisberus, Regulae eiusdem <Hentisberi> cum sophismatibus. in
Tractatus de sensu composito et diviso cum aliis opusculis logicalibus. Venice 1494,
Sophisma vigesimum sextum 'Omnis fenix est'. ff. 146rb--7ra; Gaetanus de Thienis
super sophisma idem, ff.147ra-b.
 THE SOPHISTRIA OF PETRUS OLAI 7

which is not to say that I have identified all of Petrus Olai' s sources - I can
show the traditions which Petrus uses as a framework to fill in with
original details.

One of Albert's preliminary arguments and its solution concern

contradictories - the contradictory of 'omnis phoenix est' is 'aliqua
phoenix non est', which he takes to be false since a subject demands an
existent referent. A complicated discussion of the contradictory of 'omnis
phoenix est' was present already in the Tractatus de univocatione
Monacensis and the Ars Emmerana,25 but Albert was certainly the source
of Petrus's view, which is present as his argument in oppositione
(Appendix I, lines 13-18) and his determination (109-14).

Albert was also the immediate source for a digression in Petrus Olai in
which he discusses the opinion of the antiqui - that is, of Sherwood,
Bacon, and the Magister Abstractionum discussed above - that the
sophisma sentence is false because' omnis' requires at least three supposita
(24-40).

When Petrus comes to explain why Albert thinks the sophisma

sentence to be true (41-47), he takes from Albert a solution which comes to
be expanded into Petrus's considered opinion. What Albert says is that the
word 'omnis' does not strictly require that the term it governs have several
supposita, but that the term be so imposed that it may have several
supposita. 26 Petrus says that certain common terms, insofar as their
imposition is concerned, are predicable of several things (68-70), although
in fact each is predicated of only one thing, such as 'sun' or 'world'. A
common term such as 'phoenix', "by reason of the form of <its>
imposition, that is, by reason of <its> modus significandi, is designed to
signify several things" (aptus est dici de pluribus) (75-6), although only
successively. In his answer to different objections, Petrus repeats this
solution that some terms have an ability and appropriateness for signifying
several things, although they in fact do not; they have "aptitudinem
supponendi pro pluribus, si plura signata essent in natura" (212-13) and
''fertilitatem termini et aptitudinem quam habet ad significandum plura,
quamvis actu non significet" (224-5).

I suggest that it is consideration of 'phoenix' and similar terms which

affects the defmition of common term in the tradition stemming from Peter
of Spain. Peter of Spain defines 'common term' as what is designed to be
predicated of several things (aptus natus de pluribus praedicari).27 When
John Major in the late Middle Ages comes to define 'common term', he
does so with the careful qualification that it is as regards its imposition that

251n Logica Modernorum, ed. De Rijk, vol. II, 2, p. 339, p. 156.

26Albert of Saxony, Sophismata ii [3rbj. The solution is present embryonically in the
Dialectica Monacensis, in De Rijk, Logica Modernorum, vol. 11,2, p. 469.22-5:
''Terminus communis est qui in una significatione aptitudinem habet plura denotare, ut
isti termini 'homo', 'fenix', 'sol'. Licet enim non possint plures esse soles
communiter - sicut sub hoc nomine 'homo' plures homines - tamen ipsum
aptitudinem habet." (De Rijk's index reference under 'fenix' on p. 871 reading 49~3
should be corrected to 46923.)
27Petrus Hispanus, Tractatus, Called Afterwards Summule logicales, I, 8 p. 4.12-3.
8 ROBERT ANDREWS

a common tenn is designed to signify several things, without prejudice to

the existence of those things. Major in this connection makes explicit
mention of the words 'sol', 'mundus', and 'phoenix'.28

Petrus Olai also derives material from sources other than Albert of
Saxony. For instance, Petrus expands his discussion to consider the
function of 'est' in the sophisma sentence, that is, the role of 'est secundo
adiacente'. The foundation of the analysis is De interpretatione 10 19b15-
22. Petrus gives as his own opinion that of Walter Burley (152-62), that
'est' standing alone indicates that the subject exists. Another opinion, that
the concrete' ens' becomes the predicate (139-47), dates back at least to the
time of Robert Kilwardby.29 I have not been able to identify the third
opinion, that when 'est' stands alone the subject is implicitly repeated
(148-51).

Petrus also appends a discussion of the nature of the phoenix and its
reproduction (163-87), with an unverified reference to Albert (185). Once
again, this is a discussion for which I have not found a citation, although
some authoritative source is to be expected.
The connection between the sophisma and probationes
terminorum

When I began my investigation of Petrus Olai's Sophistaria I was

intrigued by the alternation of sophismata and questions in the text, and I
wondered if there was a deliberate connection among them. Since I have
been interested in resoluble, exponible, and officiable tenns, I decided to
take the question concerning them as a test case. Initially, I was
disappointed to find that Petrus Olai made no explicit reference to the
probationes terminorum in his analysis of the sophisma, for it intuitively
seemed to me that this sophisma, and many others, could be perspicuously
analyzed by the proof of tenns. But as I grew more familiar with the text, I
realized that the preliminary disproof of the sophisma sentence (4-12) and
the response to it (227-33) - the sections beginning and concluding the
sophisma - were concerned with the proper way of resolving the sophisma
sentence, and that Petrus indeed solves the sophisma by means of
resolution.

The preliminary disproof of the sophisma sentence argues that 'omnis

phoenix est' should be divided thus: 'This phoenix is, and this phoenix is,

28John Major, Liber Terminorum, ed. in the Appendix, n. 28, of A. Broadie, George
Lokert, Late-Scholastic Logician. Edinburgh: Edinburgh University Press 1983, p.
210: "Tenninus communis est tenninus cui non repugnat accipi pro pluribus quantum
est ex impositione tennini secundum significationem secundum quam capitur; quia
Iicet sit defectus ex parte rei significate, ut patet in istis tenninis 'sol', 'mundus',
'phenix', tamen non est defectus ex parte modi significandi tennini significantis. Quia
si per possibile vel impossibile pure phisice loquendo producatur unus a1ius sol, iste
tenninus sine nova impositione <ilium> significaret." (my punctuation)
29p. O. Lewry. Robert Kilwardby's Writings on the Logica vetus. studied with regard to
their teaching and method, unpublished D.Phil. thesis, University of Oxford 1978, p.
128; cited in S. Ebbesen, "The Chimera's Diary", in The Logic of Being, ed. S.
Knuuttila and J. Hintikka, Dordrecht: Reidel 1986. p. 124.
 THE SOPHISTRIA OF PETRUS OLAI 9

etc.; therefore every phoenix is'. Since only one ostensive pointing can be
successful, the sophisma sentence must be false, it is argued.

Petrus's response is to say that the correct way to divide the sophisma
sentence is this: 'This is a phoenix; and there are no other things of this
sort; therefore, every phoenix is' (,ista phoenix est; et non sunt plures
tales; ergo omnis phoenix est'). Although Petrus does not label it as such,
his analysis is an example of resolution. Therefore there is an unstated
connection between the Albertistic sophisma and the discussion of
Billingham which follows it.
Probationes terminorum

Billingham's work Speculum puerorum is probably the best known of

the genre of works called probationes terminorum. Billingham's version
was sufficiently influential that its structure is reflected in most other
treatises. It begins with a distinction between mediate and immediate terms,
before moving to the heart of the theory, resoluble, exponible, and
offici able terms. In some writers sections on describable propositions
(descriptiones) and 'on the truth and falsity of propositions' (de veritate et
Jalsitate propositionis) are added. Billingham's work concludes with the
examination of various sentences and sentence-types in light of the
appropriate analysis.

The title 'proof of terms' disguises the content of the literature rather
than reveals it. Although it may be said that the presence of a resoluble,
exponible, or offici able term in a sentence allows that sentence to be
"proven", this is in a special sense of 'proof'. Likewise a term can be said
to be "proven" when a sentence containing it is analyzed. Thus it is clear
that we are dealing with a technical sense of 'proof' that is far different
from its ordinary modem usage.

The intent of each form of analysis is to clarify a sentence in terms of

another sentence or string of sentences. Accordingly, there is always a
proof scheme or form which accompanies each kind of analysis. In the
literature, the method of constructing a proof scheme is indicated often
only by example. While the general outline of a procedure might have been
agreed upon, the details of its accomplishment were subject to varying
interpretation.

Since I am faced with the difficulty of doing justice to many different

accounts of the theory, I shall generalize just enough to give some idea of
how each analysis functioned. I shall treat resoluble and exponible terms
briefly in order to concentrate on officiable terms.
Resoluble terms
Resoluble terms functioning in a sentence invite one to discover
simpler or more evident sentences into which the sentence can be
10 ROBERT ANDREWS

decomposed,30 Such sentences may be simpler in a number of respects,

such as immediacy or specificity: in Paul of Venice's examples, 'movetur'
is resolvable into 'currit', 'alicubi' into 'ibi',31 And resolute simplicity can
be manifested in ostensive pointing, such as when 'Socrates' is replaced
by 'hoc'. Wherever these sorts of simplification - being more immediate,
more specific, or by ostensively pointing - can take place, resolution can
take place.

The proof scheme which accompanies resolution takes the form of a

syllogism in which the conclusion is the sentence to be resolved, and the
premises are propositions which are more evident, in one of the ways
mentioned above - being more immediate, more specific, or by providing
an ostensive definition.

An example taken from Billingham, 'A man runs' ('homo currit'), is

resolved into the syllogism: 'This runs; and this is a man; therefore a man
runs.' Here the resolution or explanatory force is provided by giving an
ostensive reference to the common term 'man', by pointing at this man. In
other cases of resolution, a resoluble term may be replaced by one that is
more evident, but the syllogistic form always contains an ostensive
pointing, usually by the word 'this'. For example, an adverb appearing in
a sentence may allow the sentence to be resolved, such as the sentence' A
man is running at some time' may be resolved: 'This is a man; and this
man is running now; therefore a man is running at some time. '

To repeat the example of our sophisma sentence, 'Every phoenix is'

can be resolved thus: 'This is a phoenix; and there are no other things of
this sort; therefore, every phoenix is.'

Exponible terms

Exponible terms are not easily defined. It seems the more that a notion
has been discussed, the less easy it is to characterize it; and the exponible
was the most discussed of the provable terms among the later medievals.
The exponible term32 and its subsets33 were sometimes accorded their own
treatises, and therefore have been considered most often by modern
writers,34 By confining ourselves to the types listed by Billingham, we get

30Such sentences contain "implicitly molecular propositions", according to J. Pinborg,

"Walter Burley on Exclusives", in idem, English Logic and Semalltics, Nijmegen:
Ingenium 1981, p. 305.
31Paul of Venice, Logica Magna, Part I, fasc. I: Tractatus de terminis, ed. & tr. N.
Kretzmann, Oxford: Oxford University Press 1979, p. 224.
32See the bibliography in E. J. Ashworth, "The Doctrine of Exponibilia in the
Fifteenth and Sixteenth Centuries", Vivarium II, 1973, pp. 137-67, pp. 166-7.
33For a treatment of exclusives, see Pinborg, "Walter Burley on Exclusives"; for
incipitldesinit see N. Kretzmann, "Incipit/Desinit", in Motion and Time, Space and
Matter, ed. P. Machamer and R. Turnbull, Columbus: Ohio State University Press
1976; for maxima et minima see C. Wilson, William Heytesbury: Medieval Logic and
the Rise of Mathematical Physics, Madison: University of Wisconsin Press 1956; for
reduplicatives, see A. Blick, Reduplicatives, Munich: Philosophia, forthcoming.
34Cf. especially N. Kretzmann, "Syncategoremata, Exponibilia, Sophismata," in The
Cambridge History of Later Medieval Philosophy, ed. N. Kretzmann, A.Kenny, and J.
 THE SOPHISTRIA OF PETRUS OLAI 11

some sort of picture of exponible terms, and begin to see the interesting
issues they encompass. Billingham's list includes exclusive terms (such as
'tantum', 'solus'), exceptive terms ('praeter'), universal affirmatives (such
as 'totus'), comparatives and superlatives and such theory-laden terms as
'incipit', 'desinit', 'primum', 'ultimum', 'maximum', and 'minimum'.
Other medieval writers counted reduplicative expressions among the
exponibilia.

I will attempt a generalization of exponible terms, with the

qualification that this account is supported by some authors and not by
others. Exponible terms allow the sentence in which they appear to be
replaced by a string of at least two, and possibly more, sentences which
clarify the original sentence; furthermore, an expounded sentence and its
exponents are convertible. This characterization is confirmed by Petrus
Olai, who points out that to expound is to express more explicitly and
familiarly (in a hypothetical proposition) the sense implicit in a proposition
in virtue of the exponible term it contains.35

It is impossible to give a single proof scheme which accounts for all

the varied terms treated by exposition. Some attempt was made among later
authors to generalize about the proof schemes of various types of
exponibles. For instance, an exclusive sentence is expounded by the
conjunction of affirmative and negative sentences: 'Only man runs' is
expounded 'Man runs; and nothing other than man runs; therefore only
man runs'. An exceptive sentence is expounded by a negative followed by
an affirmative sentence: 'Every man apart from (praeter) Socrates runs' is
expounded 'Socrates does not run; and every man other than Socrates
runs; therefore every man apart from Socrates runs'. A reduplicative
sentence is expounded by however many sentences are needed to capture
the sense: 'Man insofar as he is an animal is sensitive' requires four
exponents: 'Man is an animal; and man is sensitive; and every animal is
sensitive; and if there is an animal, it is sensitive; therefore man insofar as
he is an animal is sensitive'.36 These examples come from Albert of
Saxony. However, Albert says regarding other exponibles (privative,
negative, and infinite terms) that "general rules cannot be given regarding
these; or, supposing that there could be, they would be very long and very
tedious. "37

Many other writers made use of the notion of obscurity in the

definition of an exponible term. An exponible term is an obscure term in a
sentence which can be expounded by an equivalent, clearer sentence. The

Pinborg, Cambridge: Cambridge University Press 1982; and Ashworth, "The Doctrine
of Exponibilia ... ".
35Sophistria, quo 6, 108-11: "Unde notandum 'exponere' est sensum inclusum in a1iqua
propositione, ratione alicuius termini in ea positi, magis explicite per notiora
hypothetice exprimere et convertibiliter."
36These views are Albert of Saxony's from his Logica, cited by A. Maieru,
Termin%gia logica della tarda scolastica, Roma: Edizioni dell' Ateneo Roma: Series
title: Lessico intellettuale Europeo 8, 1972, pp. 422-3.
37 Albert of Saxony, Logica, cited by Maieru, Termin%gia /ogica, p. 425 n. 131: "de
tali bus non possunt poni regulae generales vel, supposito quod possent poni, nimis
longum esset et nimis tediosum."
12 ROBERT ANDREWS

discussion of obscurity in this context, along with the later literature of

exponibles, can be found in E. J. Ashworth's article, "The Doctrine of
Exponibilia in the Fifteenth and Sixteenth Centuries."38
Officiable terms
The first curiosity about officiable terms (officiales or officiabiles) is
their name. Professor Alfonso Maieril has provided the service of outlining
a history of the title from its origin in the word 'officio', meaning "office"
or "official function"}9 In the proof of terms 'officiabile' harkens back to
this root sense - and thus is translated appropriately as "functionalizable"
by Professor Kretzmann40 - since such terms exercise a function over the
clause which follows.

The proof scheme which accompanies officiation consists of a

sentence in which everything within the scope of a term of propositional
attitude is treated as a name of a sentence which is said to signify that
sentence precisely, and is said to be governed by the term of propositional
attitude. This is perhaps easier to see with an example, 'I know that
Socrates is mortal', which can be thus officiated: 'This sentence "Socrates
is mortal", which signifies precisely that Socrates is mortal, is known by
me; therefore I know that Socrates is mortal' .

I shall proceed in my examination of the medieval treatment of

officiable terms with a survey of what modem writers have to say about it.
Not much has been written about officiable terms, and to date they have
been written about only in passing. Most who treat them give workmanlike
summaries of the doctrine, or straightforward translations of what the
medievals have to say. Boh says that "An offici able proposition is one
whose dictum or the accusative-with-infinitive part is determined by a
modal sign or by a verb which signifies an act of the mind in such a way
that the proposition has a composite sense."41 Boh's account is repeated
approvingly by De Rijk.42 Ebbesen and Pinborg say: "Officiation is the
elimination of opacity by means of the simple operation of taking
everything within the scope of the officiable term (modal, alethic, epistemic
terms and the like) to mean 'a proposition with such-and-such

38Vivarium 11, 1973, pp. 137-67.

39Maierii, Terminologia logica, p. 451-4. Maierii notes that, in some of the earliest
appearances of the term in a linguistic context, as in the Summe Metenses and William
of Sherwood, the term is associated with the word 'omnis' - as it also is in Paul of
Pergula. Maierii, Terminologia logica, p. 453 nn. 245-6; p. 254 n.249.
40ln Paul of Venice, Logica Magna, Part I, fasc. I: Tractatus de terminis, pp. 225ff.
411. Boh, "Paul of Pergula on Suppositions and Consequences", Franciscan Studies 25,
1965, pp. 30-89, p. 85. Cf. Paulus Pergulensis Logica and Tractatus de sensu
composito et diviso, ed. Sister M.A. Brown, St. Bonaventure: The Franciscan Institute
1961, p. 74: "Propositio officiabilis est cuius dictum sive oratio infinitiva determinatur
per terminum modalem vel concementem actum mentis in sensu composito, hoc
modo."
42De Rijk, Some Fourteenth Century Tracts, p. *5*.
 THE SOPHlsrRIA OF PETRUS OLAI 13

meaning'."43 Perreiah makes the perceptive observation that officiation

gives a proper analysis of indirect discourse.44

Professor Maierii has two discussions of officiable terms. In the

longer of these he spells out the terms Billingham includes, such as mental
terms (scire, credere, etc.), modal terms (possibile, necessarium), est
regarded impersonally (as est te esse), and finally non and any other
universal negative (that is, also taken impersonally).45 Maierii in his
recapitulation leaves out appetitive and promissory terms which, it is true,
Billingham describes rather obscurely as "whichever are in respect to a
complex, or whatever can be in respect of some universal" (et universaliter
quaecumque sunt respectu complexi et universaliter quae possunt esse
respectu alicuius universalis)46 but which Billingham's example clearly
reveals: 'I can promise you an apple'.47 A variant to the text makes even
more explicit the terms meant: "I promise, I desire, I want, lowe, I am
obliged."48

Most modem writers are subject to the objection that they concentrate
on a single version of officiation while neglecting dissenting elements in
other writers. The fact is that the medieval accounts differ in the elements
they include, so that it is difficult to generalize about officiation. I venture
to say that with officiation the medievals were trying to characterize any
term that governed a whole sentence, or that treated a whole sentence as
modifiable as a unit.

A noteworthy feature, in terms of current linguistic analysis, is that

officiation accomplishes what it sets out to do - it provides a correct
analysis of sentences containing terms of propositional attitudes. Analysis
by officiation clearly reveals that such sentences are composed of two
sentences. This is no small accomplishment, considering contemporary
philosophy's difficulties in explaining the referential opacity of such
sentences. That is, there is the notorious problem of which this is an
example: if Mary knows that Socrates is a philosopher, it does not follow
that Mary knows that Xanthippe's husband is a philosopher, even though
Socrates is Xanthippe's husband. The analysis provided by officiation
shows why the logical form of 'these sentences does not allow for
substitution.

43S. Ebbesen and J. Pinborg, "Thott 581 40, or De ente rationis, De definitione
accidentis, De probatione tenninorum", in English Logic in Italy in the 14th and I 5th
Centuries, ed. Alfonso Maierii, Napoli: Bibliopolis 1982, p. 118.
441n Paulus Venetus Logica Parva, tr. of the 1472 edition with introduction and notes
by A. R. Perreiah, Munich: Philosophia 1984, p. 74.
45Maierii, Terminologia logica, p. 455.
46Maierii, Termin%gia /ogica, p. 345. The peculiar chracterization of these terms at
this spot might interest those interested in intentional/promisory/appetitive terms, and
this locus for discussion can be added to the list of those treated in E.J. Ashworth, '" I
Promise You a Horse': A Second Problem of Meaning and Reference in Late Fifteenth
and Early Sixteenth Century Logic", Vivarium 14, 1976, pp. 62-80.
47Maierii, Terminologia logica, pp. 345-6: 'possum promittere tibi pomum'.
48/bid., p. 345 n. 6: Out prometto appeto desidero debeo teneor'. This list can perhaps
help to make sense of a variant in De Rijk's edition of the recensio italica of
Billingham: De Rijk, Some Fourteenth Century Tracts, p.138, §§ 21-5 n.lO: pomum]
P vel ut promitto assem de Iibero ahictemior (17) add. P.
14 ROBERT ANDREWS

Another criticism which can be directed at modern accounts of

officiation is that they take the medievals' characterization of officiation at
face value. One thought-provoking exemption to this criticism is Sten
Ebbesen's account of how the theory of probationes terminorum could
have been used. 49 He correctly points out that officiation eliminates
referential opacity. Exposition, then, further disambiguates by eliminating
quantifiers, etc., so that a string of propositions which are only singular,
indefinite, or particular remains. Finally, resolution breaks down
propositions into statements which are ostensively definable. Thus, the
probationes terminorum show that a proposition is acceptable because its
components are predicable of (or refer to) chunks of the world.5o

Ebbesen's account reveals many hidden features of the theory, and it

is a coherent account of how the theory could have been used. However, it
is not the way the medievals used it. For one thing, the kinds of terms are
always listed in the reverse order: resoluble,exponible, and then officiable
terms. For another thing, a successive analysis such as Ebbesen presents is
never given in the originals.

I think it is important that the medievals did not use the theory in this
way. I think that each of the analyses functioned independently. Resolution
took care of the cases where ostension provided simplicity, but also the
cases where more simple sentences could be reached by other means.
Exposition allows a sentence, in virtue of an exponible term, to be broken
into a string of sentences which do away with that term; late medieval
followers of the theory found exposition the most interesting and fruitful of
the analyses. Officiation analyzed terms of propositional attitudes. The
medievals had little further concern with officiation, but in fact it provides a
powerful tool of analysis which is free of semantical commitments - it
does not need ostension - in order to establish the truth of a sentence.
Two streams of medieval logic
In order to explain what I take to be the significance of the probationes
terminorum, I shall attempt to characterize what I see as two streams of
linguistic philosophy in the Middle Ages, and in order to do so briefly I
will have to paint with broad brush-strokes. It seems to me that Aristotelian
logic can be seen as developing an ascending order of complexity,
beginning with words, moving to sentences, then to syllogisms, and
culminating in a science. The Aristotelian logical corpus begins with the
Categories, dealing with individual words. It is important that a word not
be equivocal, and we must know under what category a word connects
with the world. Out of these words are constituted sentences, as treated in
De Interpretatione. Out of sentences can be constructed syllogisms, as dealt
with in the Prior Analytics. Lastly, in the Posterior Analytics we learn that
demonstration by universal syllogism is the only basis for a science. It is

49Ebbesen and Pinborg, "Thott 58140 ....., pp. 117-20. .

50Ebbesen and Pinborg, "Thott 58140 ....., p. 119: "Showing that a proposition is OK
is, in the last resort, nothing but showing that its terms are OK, being predicable of
chunks of the world."
 THE SOPHISTRIA OF PETRUS OLAI 15

only through a science that we are secure in our knowledge. This

progression, from words to sentences to syllogisms to science has a naive
appeal. To cite Augustine, "Individual words in language name objects -
sentences are combinations of such names."51 This view of words and the
world has not been accepted at least since the. time of Wittgenstein, who
pointed out that in a language consisting only of the cries 'Slab!' and
'Block!', the word 'Slab!' functions as a sentence.52 Only in the context of
a sentence does a word have a meaning, and only in the context of an entire
language does a sentence have a meaning.

In contrast to Aristotelian logic, terminist logic in the Middle Ages

implicitly acknowledged the preeminence of the sentence. Terminist logic
regarded the sentence as the unit of sense, and analyzed words as they
function in sentences. The most developed tool of terminist logic is
supposition theory. which shows how the noun has different uses in
sentences. Whereas Aristotelian logic tried to construct knowledge founded
upon the word connecting to the world, terminist logic concentrates on the
sentence and tries to disambiguate the words used within it.

I would like to suggest that officiation is a culmination of terminist

logic's attempt to disambiguate language. Officiation requires a
disambiguated sentence in order to operate. I direct attention to the phrase
'signifies precisely' which occursin most accounts of the proof scheme of
officiation. A sentence's' must signify precisely that s in order for
officiation to go through. Officiation relies upon the mechanisms of
terminist logic to have disambiguated a sentence, and then exhibits that
sentence as modifiable by a term of propositional attitude.

Another striking point concerning officiation is that it is neutral as

regards the connection between language and the world. I make the
comparison with Tarski's theory of truth, in which sentences of the form
'''Snow is white" is true iff snow is white' constitute a definition of truth
for a language.53 What is remarkable about Tarski's theory is that it is
extensive - that is, it points from the sentence 'snow is white' to a feature
in the world, namely that snow is white - but it does not make any further
commitment as to how language is linked to the world. In developments of
Tarski's theory, such as in Quine54 and in Davidson's article "Reality
without Reference",55 the suggestion is made that language links to the
world as a totality. There is no provable bond between a word and what it
signifies, but any connection claimed must be based on the whole
language.

51 Augustine, Confessions, I. 8, quoted in L. Wittgenstein, Philosophical

Investigations, tr. O.E.M. Anscombe, New York: Macmillan 1958, p. 2e.
52D. Davidson, "Reality without Reference", Dialectica 31, 1977, pp. 247-53 (repr. in
D. Davidson, Inquiries into Truth and Interpretation, Oxford: Clarendon Press, 1984),
pp. 219-20, calls such an account the "building-block theory", and passes the
judgement that "it is hopeless".
53A. Tarski, "The Semantic Conception of Truth", Philosophy and Phenomenological
Research 4, repro in Semantics and the Philosophy of Language, ed. L. Linsky,
Urbana: University of Illinois Press 1944, pp. 15-16.
54W.V. Quine, Word and Object, Cambridge: MIT Press 1960, Chap. 2.
55Davidson, "Reality without Reference", esp. pp. 220-5.
16 ROBERT ANDREWS

Thus what the medievals have in embryo in the theory of officiation is

an extensional theory of language without a commitment to ostension. It is
extensional in that the phrase 'precisely signifies' which links a sentence
with a name of a sentence has to be satisfied; that is, you have to satisfy
yourself in some way that a sentence properly signifies. But this does not
have to be by ostension, as is the case with resolution. The medieval
theory of officiation, I suggest, has the strength of Tarski's theory. It
suggests that a sentence can be disambiguated, and thereby claims that a
sentence has a unique connection to the world. But it does not make any
further claim about that connection. It is neutral as regards reference.56
Conclusion

In this paper I sought a connection between a sophisma and a question

concerning the probationes terminorum in Petrus Dlai. I found a
connection only implicitly, namely, in Petrus' response to the preliminary
argument, in which he analyses the sophisma sentence 'omnis phoenix est'
by resolution. There remain many more unedited sophismata with
associated questions in Petrus' text. I would like to encourage further
exploration of this document for the interest in late-medieval treatments of
Albert of Saxony's sophismata and their interaction with works of
Billingham and Marsilius of Inghen.

My other interest in this paper was to draw attention to the medieval

theory of officiation. I have suggested that officiation is of interest because
it provides a cogent analysis of the function of terms of propositional
attitude. Furthermore, since it deals with disambiguated sentences, it
provides a culmination for the efforts of terminist logic. Its similarity of
analysis with Tarskian T-sentences makes it of interest for contemporary
philosophers of language.57
The Franciscan Institute, St Bonaventure University

56There has been an interest in considering medieval logic in light of Tarski's theory
ever since the discussion in Ernest A. Moody, Truth and Consequence in Mediaeval
Logic, Amsterdam: North-Holland 1953, p. 101.
57 1 should like to thank those who helped in the preparation of the texts from Petrus
Olai, foremost the members of the Franciscan Institute, Father Gedeon Gal, Father
Romuald Green, Gerard Etzkorn, and Rega Wood; and as well Professors Sten Ebbesen
and Norman Kretzmann, and the participants of the Symposium, for their suggestions
after the presentation of this paper.
 THE SOPHISTRIA OF PETRUS aLAI 17

Appendix 1

Petrus Olai, Sophistria, MS Uppsala C 599 ff.268ra-304vb

<Sophisma secundum> /272va/

OMNIS PHOENIX EST.

Ad sophisma supponitur quod tan tum sit una phoenix pro

tempore, sicut definitur. 58

Tunc improbatur sic. Dividentes sophismatis sunt falsae, ergo et

5 sophisma. Tenet consequentia, quoniam est fonnalis consequentia a
dividentibus ad divisum.

Antecedens probatur quoniam dividitur sic sub sophismate: 'ergo

ista phoenix est, et ista phoenix est, et sic de aliis; ergo omnis
phoenix est'. Et <quod> istae dividentes sunt falsae patet, quia ad
10 veritatem copulativae requiritur ambas partes esse veras; sed una pars
illius copulativae est faIsa, quia subiectum eius pro nullo supponit; et
per consequens sophisma falsum <est>.

Sed probatur sic: Cuius propositionis contradictoria est vera, ipsa

est faIsa, et econverso; sed contradictoria sophismatis est faIsa; ergo
15 ipsum verum. Minor patet quoniam haec est contradictoria
sophismatis: 'aliqua phoenix non est'. Haec est falsa. Patet ex
supposito, nam non est nisi una phoenix. Et sic sophisma relinquitur
probatum et improbatum.

Pro intellectu ergo sophismatis notandum quod illud sophisma

20 disputat Albertus, ut videatur, an syncategorema distributivum possit
addi alicui tennino absque hoc quod habeat tria supposita ad minus
vel plura.59 Et, ut videatur, quid sit praedicatum in propositionibus
de 'est' secundo adiacente.

De primo erat opinio antiquorum Iogicorum et sophistarum

25 dicentium: signum 'omnis' nulli tennino convenienter posse addi nisi
in actu habeat tria supposita.60 Et movebantur ex eo quia signum

58Cf. Albertus de Saxonia, Sophismata ii [2vb]: Sophisma secundum est istud: Omnis
fenix est, supposito quod ilIud quod communiter dicit<ur> sit verum, quod non est nisi
una fenix, Iicet sint plures successive.
59Cf. Albertus de Saxonia, Sophismata ii [2vb]: Tunc improbat sophisma sic: quia ad
hoc quod propositio affirmativa sit vera cuius subiectum distribuitur per hoc signum
'omnis' requiritur quod eius subiectum ad minus habeat tria supposita actu; sed hic non
est.
60GuilIelmus de Shireswode, Syncategoremata (ed. O'Donnell), p. 49: Item regula: hoc
signum 'omnis' vult habere tria appellata ad minus.
Cf. Petrus Hispanus, Tractatus: XII, 7, pp. 212-3; 8, p. 215.
 18 ROBERT ANDREWS

universale additum termino denotat ipsum capi pro omnibus suis

suppositis copulative; sed terminus habens unicum /272vb/
suppositum non potest accipi pro omnibus, ex quo est tantum unum.
30 Inconvenienter ergo additur ei signum distributivum.

Secundo movebantur ex hoc quia 'dici de omni' est in omni

propositione universali affirmativa; sed 'dici de omni' non potest
esse in propositione habente unum suppositum ex quo ibi est dici de
uno tantum. Et addit opinio: "quandocumque syncategorema additur
35 alicui termino non habenti tria supposita ad minus, talis terminus
transit in non ens,"61 id est, supponit pro non ente. Et secundum
illam viam, haec propositio 'omnis phoenix est' valet 'tantum aliqua
phoenix est', et 'aliae duae phoenices, quae non sunt, sunt'; quae
secunda propositio est falsa; quare tota falsa. Et sic dicit sophisma
40 esse falsum.

Albertus tamen aliter sentit de sophismate, volens quod signum

universale convenienter addi potest termino habenti unicum
suppositum.62 Et ratio est quia signum non requirit in termino cui
additur plura supposita actu, sed ei sufficit aptitudo termini
45 communis, ut scilicet aptus natus sit ex forma suae impositionis
supponere pro pluribus, vel quod ei [pro pluribus] non repugnet
supponere pro pluribus.

Ad hoc ergo quod signum universale addatur termino communi,

sufficit quod ipsi ex forma impositionis non repugnare<t>
50 <supponere> pro pluribus. Sic iam est in proposito. Iste enim
terminus 'phoenix' ex forma suae impositionis aptus natus est
supponere simul pro pluribus. Si enim simul et semel essent infinitae
phoenices, omnes illas terminus 'phoenix' significaret et pro eis
supponeret.

55 Secundo probatur idem, duabus suppositionibus praemissis.

Prima suppositio, quod signum universale est dispositio universalis
inquantum tale. Secunda, quod universale inquantum tale non
requirit actu pluralitatem suppositorum, quod praetendit Aristoteles
ostendere cum sic definit ipsum: "universale est quod aptum natum
60 est esse in pluribus."63 Definit enim per aptitudinem et non per
actum. Stantibus suppositionibus, formetur sic ratio: signum
universale est dispositio universalis inquantum est tale; sed
universale inquantum tale non requirit actu muItitudinem

Cf. Albertus de Saxonia, Sophismata ii [3ra]: dixerunt antiqui sophistae quod hoc
sincathegoreuma 'omnis' exigit habere actu tria supposita.
61Cf. Petrus Hispanus, Tractatus: XII, 8 p. 215: Quidam tamen dicunt quod 'omnis'
semper vult habere tria appellata ad minus et dant talem regulam: "quotiescumque
signum universale affirmativum additur termino communi non habenti sufficientiam
appellatorum. recurrit ad non ens."
62Cf. Albertus de Saxonia. Sophismata ii [3ra]: hoc signum 'omnis' non exigit tria
appellata; id est. ad hoc quod propositio sit vera, cuius subiecto additur hoc signum
universale 'omnis', non oportet eius subiectum ad minus habere tria supposita in actu.
63Aristoteles. De illt. 7 17a39-40; Hamesse 305 (10): Universale est quod aptum natum
est praedicari de pluribus, et singulare quod non.
 THE SOPHISTRIA OF PETRUS OLAI: APPENDIX 1 19

suppositorum. Maior est suppositio prima; minor est secunda; quare

65 <etc.>.

Notandum consequenter quod termini communes in duplici sunt

differentia. (Termini communes sunt duplices adn. in mg.) Quidam
enim habent actu tantum unicum suppositum; quantum tamen est ex
parte formae suae impositionis dicibilis est et praedicabilis de
70 pluribus, ut 'sol', 'mundus', 'deus'.

Alius est terminus communis (Alius terminus communis adn. in

mg.) qui tam ex parte materiaequam ex parte fonnae dicibilis est de
pluribus, licet non simul tamen successive, id est, qui tam ratione
suppositi, quod est materiale in termino, quam etiam ratione formae
75 impositionis, id est, ratione modi significandi, aptus est dici de
pluribus, licet non simul et semel, ut 'phoenix'. Similiter et
superlativus tentus affirmative illo modo terminus communis dicitur,
quoniam de uno tantum verificabilis est simul, licet de pluribus
successive. Patet hoc quoniam quidquid per superabundantiam
80 dicitur, uni tantum attribuitur;64 sed superlativus per
superabundantiam dicitur; quare etc.65

Superadditur et tertius terminus communis, /273ral qui tam ex

parte materiae quam ex parte formae verificabilis est de pluribus
suppositis actualiter, ut 'animal', 'homo', et sic de aliis.

85 Unde ergo est (s.l.) notandum quod signum universale

distributivum additur termino communi tertio modo, et facit ipsum
stare pro pluribus et omnibus suis suppositis actualiter. Sed si fuerit
additum (additur a.c.) termino communi sumpto primo et secundo
modis, facit ipsum stare non pro multis actualiter, sed denotat
90 aptitudinem et non repugnantiam supponendi pro pluribus. Sic est in
sophismate; quare <etc.>.

Sed dices, si illud quod minus videtur inesse inest, tunc et illud
quod magis; sed minus videtur quod 'uterque' in actu plura
supposita requirat; ergo et 'omnis' de quo magis videtur. Maior est
95 topica. 66 Minor patet quia 'uterque' est distributivum pro duobus;
'omnis' autem distribuit ad minus pro tribus; quare etc.

Dicendum non est simile quoniam 'uterque' est signum partitivum

partitionem importans; oportet ergo quod habeat duo supposita in
actu, quoniam partitio est inter duos. Sic autem non est de Iy
100 •omnis', quoniam hoc non est partitivum; quare <etc.>.

64Aristoteles, Topica V, 5 134b23-4; Hamesse, 327 (78): Quod per superabundantiam

dicitur, uni soli convenit.
65Cf. Albertus de Saxonia, Quaestiones in Artem Veterem, ed. A. Muftoz Garcia,
Maracaibo: Universidad del Zulia 1988, pp. 304, 322: dico quod illud quod per
superabundantiam dicitur, seu quod est superlativus gradus, polest exponi dupliciter,
scilicet affinnative et negative.
66Boethius, De differentiis topicis II (Patrologia Latina 64) 1191A: Maxima propositio:
si id quod minus videtur inesse inest, id quod magis videbitur inesse, inerit.
 20 ROBERT ANDREWS

Contra: si 'omnis' posset addi termino habenti tantum unicum

suppositum (supposito MS), iam posset addi termino singulari.
Consequens est falsum. Patet sequela quoniam non est ratio
diversitatis, cum uterque terminus habet unicum suppositum.67

105 Respondetur quod ratio diversitatis est haec, quoniam termino

singulari ex forma suae impositionis repugnat pro pluribus
supponere, quod non repugnat termino communi habenti unum
suppositum.68

Tunc dicatur ad sophisma quod ipsum sit verum. (quoniam si h

110 add. et del.) Patet quoniam si hoc esset falsum, hoc esset ideo quia
signum universale additur termino unicum suppositum habenti; sed
hoc non patet, quoniam si ex eo esset falsum, iam duo contradictoria
simul starent in falsitate. Consequens est falsum ex lege et natura
eorundem.

115 Sequela probatur, accipiendo hanc propositionem 'omnis sol

lucet'. Si haec esset falsa, et eius contradictoria esset falsa. lam
ambae essent falsae. Patet quoniam eius contradictoria est 'aliquis sol
non lucet', quod est falsum, ex eo quod non sint plures praeter unum
in firmamento. Sic ergo patet quod ex eo sophisma non est falsum.

120 Secundo probatur: In materia naturali duae contrariae non possunt

simul esse falsae; sed contraria sophismatis est falsa. Si ergo et
ipsum esset falsum, iam duae contrariae essent falsae simul in
materia naturali. Minor probatur quoniam eius contraria est 'nulla
phoenix est', quae est falsa.

125 Tertio probatur. Eius convertens est vera, ergo et sophisma. Tenet
consequentia quoniam a convertente ad conversam est bonum
argumentum. Antecedens probatur quia eius convertens est (aliquid
est add. et del.) 'quod est phoenix', et haec est vera. Quod haec sit
convertens patet, quia universalis affirmativa convertitur in
130 particularem /273rb/ affrrmativam.69

Pro secundo principali notandum, scilicet quid sit praedicatum in

propositionibus de 'est' secundo adiacente.

Considerandum quod propositiones sunt duplices.1 °

(Propositiones sunt duplices adn. in mg.) Quaedam sunt de 'est'

67Cf. Albertus de Saxonia, Sophismata ii [3rb]: Secundo: si 'omnis' non exigeret tria
appellata, tunc convenienter adderetur terminis singularibus, quod est falsum.
68Cf. Albertus de Saxonia, Sophismata ii [3va]: Ad secundum dico quod non est simile
de termino singulari et de termino communi, quia termino singulari quantum ad
modum suae impositionis repugnat supponere pro pluribus; et ideo non est
inconveniens signum universale non posse addi termino singulari, et posse addi
termino communi non obstante quod non habeat nisi unum suppositum actu.
69Cf. Aristoteles, All. priora 1,3 25a7-1O; Hamesse 308 (6): Universalis affirmativa
convertitur in particularem affirmativam.
70Cf. Aristoteles, De imerpretatione 10 19b15-22.
 THE SOPHISTRIA OF PETRUS OLAl: APPENDIX 1 21

135 secundo adiacente in quibus nihil ponitur post copulam, ut 'omnis

phoenix est', 'omnis deus est'. Aliae sunt de 'est' tertio adiacente in
quibus exprimitur praedicatum propositionis, ut 'omnis homo est
animal', 'omnis color est qualitas', et sic de aliis.

Logici ergo diversificantur in hoc quid pro praedicato ponatur in

140 propositionibus de 'est' secundo (adiacente add. et del.) adiacente.
Unde quidam71 dicunt quod praedicatum in eis sit significatum
materiale huius verbi 'est' in eo inclusum; ut in hac propositione
'omnis phoenix est', ens sit praedicatum. Sed hoc non videtur
probabile, quoniam materiale concretum significatum non significat,
145 ergo pro eo non significat in propositione. Verbum autem
substantivum concretum est; ergo pro materiali non supponit in
propositione, sed ipsum tantum ex consequenti dat intelligere.

Dicunt etiam alii 72 quod in hiis propositionibus subiectum

resumptum debeat esse praedicatum; hoc iterum non videtur.
150 Quoniam sic essent propositiones identicae (idemptice MS) (quoniam
add. et del.) cum idem de seipso praedicaretur. Dicendum ergo aliter.

Pro quo sciendum verbum substantivum dupliciter capitur.

(Verbum substantivum capitur dupliciter adn. in mg.) Uno modo pro
actu compositionis, et sic est copula. Alio modo pro formali
155 significato, ut pro mero esse, et pro tali significato formali est
praedicatum in propositionibus de 'est' secundo adiacente. Probatur
sic: illud in propositionibus est praedicatum quod dicitur de subiecto;
sed merom esse in propositionibus illis dicitur de subiecto; ergo tale
est praedicatum ibi; quare <etc.>. Cum ergo convertuntur hae
160 propositiones, illud merom esse ponitur pro subiecto, et exprimitur
per hoc complexum 'quod est', ut sic: 'omnis phoenix est' sic
convertitur: 'ergo quod est phoenix'.

Contra primum arguitur. Primo sic. Impossibile est esse tantum

unam phoenicem. Probatur quoniam species in uno individuo salvari
165 non potest. Probatur quoniam unum individuum, cum sit
corruptibile, non potest aetemaliter permanere; species autem rerum
sunt aetemae; quare species in uno individuo non potest salvari, sed
requirit ad minus duo, ut cum unum fuerit corruptum, in reliquo
maneat.

170 Item duo requiruntur propter hoc ut possint generationem

perpetuare.

Respondetur: speciem salvari in uno individuo intelligitur

dupliciter. (Speciem salvari in uno individuo intelligitur dupliciter
adn. in mg.) Uno modo quoad essentiam; alio modo quoad

71Cf. Gualterus Burlaeus, De puritate artis [ogicae. tractatus [ongior, p. 54: Quando
enim hoc verbum 'est' praedicatur secundum adiacens, tunc est categorema, quia tunc
est praedicatum vel includens in se praedicatum et dicit determinatam naturam, scilicet
esse existere.
72Mihi iglloti sum.
 22 ROBERT ANDREWS

175 perpetuitatem. Quoad primum potest salvari essentia speciei in uno

individuo, sed non quoad secundum, scilicet quoad perpetuitatem.
(specie add. et del.) Unde ex quo individua sunt corruptibilia, tunc
dicatur consequenter de phoenice quod, statim corrupta phoenice, in
potentia propinqua sunt cineres eius ad receptionem novae formae
180 alterius phoenicis. Et non sunt ibi duae phoenices propter
generationem, quoniam est animal genitum non per propagationem
seminis sicut cetera animalia genita, sed virtute caelesti tantum
potentia ilIa propinqua reducitur in actum. Quidquid tamen sit,
verisimile est quod sint plures phoenices in rei veritate, sicut etiam
185 dicit Albertus,73 quamvis /273va/ in uno climate sit (visa add. s.l.)
sola una phoenix; in aliis climatibus etiam inhabitabilibus possunt
esse plures.

Stante tamen casu, sic potest responderi sicut dictum est.

Secundo74 arguitur: contradictorium sophismatis est verum, ergo

190 ipsum omnino falsum. Tenet consequentia ut prius. Antecedens
probatur. Haec est vera: 'aliqua phoenix non est'. Probatur: sequitur
bene 'aliqua phoenix est corrupta; ergo aliqua phoenix non est'.
Tenet consequentia quia corruptum non est. Antecedens probatur
quoniam phoenix, cum sit individuum materiale, aetemaliter
195 permanare non potest. Sic (sit MS) ergo aliqua phoenix corrupta est.

Respondetur 75 negando consequentiam, quia arguitur ab

ampliatio ad non ampliatum. Quia in (maiore add. et del.)
antecedente 'phoenix' supponit pro illo quod est vel fuit, ratione
participii praeteriti temporis; in consequente autem non ampliatur;
200 quare non sequitur.

Arguitur tertio. Syncategorema (pura(?) add. s.l.) non debet addi

termino habenti unum suppositum. Probatur quoniam terminus ad
minus debet habere tria supposita cui addi debeat. Probatur
(Aristoteles add. et del.) quia Aristoteles habet primo Caeli76 quod
205 duos vocamus ambos, et non vocamus duo omnia, sed hoc numero
utimur primo circa tria, dicentes tria (esse exp.) omnia; ubi vult quod

73Locum non inveni.

74Cf. Albertus de Saxonia, Sophismata ii [3ra]: Secundo arguitur. Haec est vera: 'Aliqua
fenix non est'; ergo sophisma est falsum. Tenet consequentia eo quod est
contradictorium sophismatis. Antecedens probatur: aliqua fenix est corrupta; ergo aliqua
fenix non est. Probatur consequentia: nullum corruptum est; aliqua fenix est corrupta;
ergo aliqua fenix non est.
75Cf. Albertus de Saxonia, Sophismata ii [3va]: Ad secundam negatur antecedens et
conceditur consequentia. Et quando dicitur "Aliqua fenix est corrupta; ergo aliqua fenix
non est", conceditur antecedens et negatur consequentia. Et ratio est quia in antecedente
subiectum supponit pro ilIo quod est vel fuit, propter hoc praedicatum 'corrupta', ex
quo habet sic ampliare subiectum, quia est participium praeteriti temporis. Sic autem
non ampliatur subiectum consequentis propter defectum termini ampliativi. Et quando
ulterius arguitur: "nullum corruptum" etc., dico quod non debet inferri ilia conclusio
'aliqua fenix non est'; sed bene ista: 'aliqua fenix quae est fel fuit non est'. Et ratio est
quia subiectum minoris sic supponit.
76Aristoteles, De caelo 1,1 268aI7-9: Quae enim duo, ambo dicimus, et duos ambos;
omnes autem non dicimus, sed de tribus hanc appellationem dicimus primum.
 THE SOPHISTRIA OF PETRUS aLAI: APPENDIX 1 23

signum 'omnia' primo de tribus praedicetur. (Omne et totum super

tria ponimus, primo Caeli adn. in mg.)77

Respondetur quod hoc signum 'omnis' positum circa terminum

210 actu habentem plura supposita denotat ipsum capi pro pluribus
copulative; sed positum circa terminum non habentem plura
supposita, denotat eius aptitudinem supponendi pro pluribus, si
plura signata essent in natura. Aristoteles autem vult quod 'omnia' ea
primo vocamus qui temarium habent, et sic primo utimur hoc
215 syncategoremate circa numerum temarium. Per hoc autem non vult
negare quod (quoniam MS) adhuc alicui possit addi <termino>
habenti unum suppositum actu, aptitudine vero plura.

Arguitur quarto. Si syncategorema addi posset termino habenti

unum suppositum, fieret aliquid frustra in arte. Consequens est
220 inconveniens. Sequela probatur quoniam frustra adderetur signum
distributivum termino communi, cum pro aequivalenti capiatur sine
signo.

Respondetur ad hanc: non fiunt aliqua frustra in arte, quia signum

illud denotat fertilitatem termini et aptitudinem quam habet ad
225 significandum plura, quamvis actu non significet; quare non est
frustrum.

Tunc solvatur ratio in principio adducta (cf. 11. 4-6), qua dicebatur
de dividentibus. Dicendum quod sic dividi non debet, sed sic:
'omnis phoenix est, ergo ista phoenix est; et non sunt plures tales;
230 ergo omnis phoenix est'. Cuius ratio est quia dividere est arguere ad
inferiora. Cum ergo tantum unum habeat inferius, ad unum tantum
est arguendum. Et sic patet quod indirecte dividitur in quarto(?)
argumento; quare [in argumento quare] <etc.>.

77Aristoteles, De caelo 1,1 268all; Hamesse 160 (4): Dmne totum et perfectum super
tria ponamus.
 24 ROBERT ANDREWS

Appendix 2

<Quaestio sexta> /273va/

UTRUM DEFINITIONES TERMINI RESOLUBILlS,

EXPONIBILIS ET OFFICIALIS SINT BONAE
CONSIDERANDUM.

Definitio termini (data add. et del.) resolubilis data per magistrum

5 est haec: "Terminus resolubilis est quilibet terminus communis, sive
nomen sive pronomen sive participium sive adverbium, habens
. aliquem terminum inferiorem vel superiorem se secundum
praedicationem," ratione cuius tanquam naturaliter notioris propositio
in qua ponitur possit probari resolubiliter.1 8 /273vb/

10 Pro quorum intellectu notandum quod ly 'quilibet' in definitione

positum loco generis non distribuit pro singularis generum, hoc est,
pro suis individuis generum, sed pro generibus singulorum, id est,
pro generibus individuorum.1 9 Patet primum, quia non quilibet
terminus communis, sive (1) nomen sive (2) pronomen sive (3)
15 participium sive (4) adverbium, est terminus resolubilis.

(1) Patet quia instantiam patitur de nomine proprio quod non est
terminus resolubilis sed tantum terminus communis et nomen
appellativum.

Etiam (debet add. et del.) de nomine adiectivo claret idem,

20 quoniam tale de se non est terminus resolubilis, cum ex se non
posset esse extremum propositionis.

(2) Patet idem tertio de pronominibus, quae omnia non sunt

termini resolubiles; immo quaedam sunt termini immediati, sicut sunt
pronomina demonstrativa primitiva demonstrative tenta - licet ipsa
25 relative accepta possunt dici termini resolubiles, sicut patet in
propositione hac, 'homo legit, et ille disputat', quae sic probatur
reso<l>ubiliter: 'hoc legit, et hoc est ille homo qui disputat; ergo
homo legit et ille disputat'.

78Richardus Billingham, De probationibus terminorum (prior recensio), ed. L. M. De

Rijk, Some Fourteenth Century Tracts, p. 50: Terminus resolubilis est quilibet
terminus communis, sive nomen sive pronomen aut verbum aut participium, qui habet
aliquem terminum inferiorem se secundum praedicationem. Cf. Maieri'!, Terminoiogia
Logica, pp. 340-1.
79Cf. Albertus de Saxonia, Sophismata iii [4raJ: Ex hoc sequitur quod antiqui sophistae
(Guillelmus de Shireswode, ed. O'Donnell, p. 49) male dixerunt in hoc quod hoc
syncategorema 'omnis' distribuit aJiquando terminum communem pro singuJis
generum, aJiquando autem pro speciebus singulorum.
 THE SOPHISTRIA OF PETRUS OLAI: APPENDIX 2 25

Pronomina etiam gentilia termini dicuntur resolubiIes, ut haec

30 propositio ratione pronominis gentilis sic resolvitur, 'nostras venit':
'hoc venit, et hoc est nostras; ergo nostras venit'.

(3) Neque etiam participia quantum est ex se possunt dici termini

resolubiles, ex eo quod non possunt ex se integrare propositionem;
quantum tamen est de bonitate intellectus, cui +existenti+ dant
35 intelligere <verbum> substantivum, possunt dici termini resolubiles.
Patet quia propositio ratione ipsorum potest probari resolubiliter.

(4) Patet idem de adverbiis. Quae<dam> non possunt dici termini

resolubiles. Patet quia tan tum adverbia prout eis correspondet
conceptus nominalis de bonitate intellectus. (intelligentis MS) Sed
40 adverbia inquantum talia, cum sint syncategoremata pura,80 eis non
correspondet conceptus nominalis; ergo ut sic non possunt dici
termini resolubiles - sed inquantum eis conceptus nominalis
correspondet per bonitatem intellectus; ut huic adverbio 'alicubi'
correspondere potest apud intellectum 'in aliquo loco'; 'aliquando',
45 hoc est, 'in aliquo tempore'.

Claret ergo in definitione ly 'quilibet' non distribuere pro singulis

generum, sed pro generibus singulorum, ut sic stet definitio:
(Terminus resolubilis est adn. in mg.) Terminus resolubilis est
quilibet terminus communis,sive nominalis sive pronominalis sive
50 participialis sive adverbialis sive verbalis; et aliquis terminus
communis de omni genere horum nominatorum, hoc est, aliquis
(aliis pro add. et del.) nominalis, aliquis pronominalis, et aliquis
adverbialis. Alias definitio non valeret ex quo excederet definitum,
quod fieri non debet. Et hoc est quod alii 81 aliis verbis nituntur
55 dicere, quod ly 'quilibet' in definitione non capitur universali sed
particulari tantum.

Unde notandum quod terminus resolubilis capi potest stricte et

large. Stricte accipitur pro termino aliquo habente inferiorem /274ra/
(adn. in mg.: notandum est quod regulatum est ab agente infallibili,
60 duodecimo Metaphysicae82 ) se secundum praedicationem, ratione
cuius propositio tanquam ratione naturaliter (s.l.) notioris possit
probari. Et ut sic distinguitur contra terminum componibiIem,
(Terminus componibilis est adn. in mg.) qui sic definitur: est
terminus habens aliquod superius notius se, ratione cuius propositio
65 possit probari. Sed in definitione data debet accipi communiter
terminus resolubilis, prout non exc1udit terminum componibilem, et
ut sic accipit eum magister in littera. Terminus enim resolubilis in
proposito omnis ille dicitur qui habet aliquod notius se, sive illud sit

80ef. Albertus de Saxonia, Quaestiones in Artem Veterem (ed. Munoz Delgado), p.

492: dupJicia sunt syncathegoremata: quaedam enim dicuntur syncathegoremata pura,
non includentia cathegoremata, sicut Iy 'omnis', Iy 'nullus', etc. Alia dicuntur
syncathegoremata non pura, aliqua cathegoremata includentia.
81Auctores mihi ignoti.
82Locum non inveni.
 26 ROBERT ANDREWS

inferius sive superius. Proprie tamen utendo, terminus ille qui habet
70 inferiorem se notiorem dicitur resolubilis ex eo quod resolutio illa
transit ad posteriora faciens innotescere ea quae in toto fuerunt. Alius
autem qui habet superiorem se, ratione cuius propositio probatur
tanquam per notius, dicitur componibilis ex eo quod priora per
compositionem noscuntur, ex quo superiorum et priorum est
75 componere. In propos ito tamen communiter accipi debet terminus
resolubilis, sicut dictum est.

Item quia additur in definitione 'tanquam naturaliter notioris',

considerandum quod notioritas qua aliquis terminus notior est alia
proven it ex duobus. (Dupliciter aliquid dicitur notius adn. in mg.)
80 Primo enim aliquid (dicitur(?) notius alio add. et del.) dicitur notius
alio eo quod singularius. Singularia enim magis nota dicuntur nobis
eo quod magis cadunt sub sensu, unde nostra cognitio incipit. Et ut
sic terminus resolubilis habens inferius per ea probatur resolubiliter.

Secundo notioritas provenit ex eo quod universaliora sunt notiora,

85 sicut habet Aristoteles primo Physicorum83 hoc sic intelligendo quod
singulare magis universalis prius nobis innotescit quam singulare
minus universalis, ut prius cognoscimus aliquem de longe venientem
sub ratione qua ens, deinde sub ratione qua substantia, tum sub
ratione qua animatus, tandem sub ratione qua homo, ultimo sub
90 ratione qua Sortes vel Plato. Et tali notioritate probatur terminus
resolubilis habens superius se notius, qui alia nomine dicitur
componibilis.

Pro secunda definitione notandum (Terminus exponibilis adn. in

mg.) quod "terminus exponibilis est terminus duas vel plures habens
95 exponentes cum quibus convertitur."84 Et debet definitio intelligi sic:
"Terminus exponibilis est terminus", supple mediatus, "habens",
supple actu vel aptitudine, "duas vel plures exponentes cum quibus
convertitur", hoc est, ratione cuius propositio in qua ponitur
terminus qui probabilis est (s.l.) exponitur per duas vel per plures
100 propositiones convertibiles cum propositione exposita.

Exempli gratia, ubi propositio /274rb/ probatur ratione alicuius

termini per duas exponentes, ut haec propositio 'tantum homo currit'
sic exponitur: 'homo currit, et nihil aliud ab homine currit; ergo
tantum homo currit'. Ubi vero propositio per plures exponentes
105 exponitur, exemplum ut hic 'homo differt ab asino' sic exponitur:
'homo est, et asinus est; et homo non est asinus; ergo homo differt
ab asino'.

Unde notandum 'exponere' est (Exponere est adn. in mg.)

sensum inclusum in aliqua propositi one, ratione alicuius termini in ea

83 Aristoteles, Phys. 1,5 189a5: universa1e enim secundum rationem no tum est,
singulare vero secundum sensum; cf. Hamesse 142 (27).
84Billingham, De probationibus terminorum (prior recensio), p. 51-2: Terminus
exponibilis est qui habet duas exponentes, vel p1ures, cum qui bus convertitur. Cf.
Maieru, Terminoiogia Logica, p. 344.
 THE SOPHISTRIA OF PETRUS OLAI: APPENDIX 2 27

110 positi, (posita a.c.) magis explicite per notiora hypothetice exprimere
et convertibiliter.85 Et hoc contingit dupliciter, quia quandoque per
tres propositiones, quandoque vero per duas tantum, ut dicit
definitio. Unde termini exponibiles sunt (Termini exponibiles sunt
adn. in mg.) omnia exclusiva, signa universalia affirmativa, et
115 exceptiva, 'differt', 'aliud', 'non idem', omnis comparativus et
superlativus, et haec verba 'incipit' et 'desinit', et universaliter omnia
illa ratione quorum propositio potest exponi per quasdam
propositiones aequivalentes propositioni in qua ponitur.86

Considerandum <est quod> magister in textu 87 ponit unam

120 diversitatem et differentiam inter terminos exponibiles et resolubiles,
(Termini resolubiles et exponibiles differunt adn. in mg.) dicens
quod propositio exponibilis convertibilis est cum suis exponentibus,
sic quod ipsa potest inferre suas exponentes, et econverso inferri ex
eisdem; sed sic non est de propositione resolubili, quia quamvis
125 resolventes propositionis resolubilis inferant ipsam formaliter,
econverso tamen ipsa non potest inferre suas resolventes.

Exempli gratia, sequitur bene 'omnis homo currit; ergo homo

currit et nihil est homo quin currat; ergo omnis homo <currit>'. Sed
non sequitur formaliter in propositione resoluta, ut non sequitur,
130 'homo currit; ergo hoc currit,et hoc est homo'; licet bene sequatur
econverso a resolventibus ad resolutam, ut sic: 'hoc currit, et hoc est
homo; ergo homo currit'.

Tertia definitio, scilicet termini officialis, est: "Terminus officialis

est terminus officium mentis vel animae circa propositionem
135 importans, vel est determinatio alicuius complexi vel alicuius

85Cf. Hieronymus de Hangest, Problemata exponibilium, cit. in E.]. Ashworth, "The

Doctrine of Exponibilia ... ", p. 139: signum veTO exponibile est signum ypothetice
denotans implicite quidem etiam mentaliter non merito suppositionis quam termino
tribuat. Et ilIud signum dicitur ypothetice denotare iIIo modo cuius denotatio venit
mediate vel immediate per ypotheticam explicanda.
86Billingham, De probationibus terminorum (prior recensio), p. 52: Et sunt termini
exponibiles: omnis dictio exclusiva, exceptiva, omne signum universale affirmativum,
'differt', 'aliud', 'incipi!', 'desinit', et omnis comparativus gradus et superJativus, et
universaliter omne quod habet duas exponentes, vel plures, cum quibus convertitur; ut
isti termini 'primum', 'ultimum', 'maximum quod sic', 'minimum quod sic',
'maximum quod non', 'minimum quod non'; et sic de aliis. Cf. Maierii, Terminologia
Logica, pp. 344-5.
87Richardus Billingham, Terminus est in quem sive Speculum puerorum (recensio
italica) (ed. De Rijk, op. cit.) p. 121: Et in hoc differt <terminus exponibilis> a
termino resolubili quia terminus exponibilis convertitur cum suis exponentibus et
econverso, scilicet exponens convertitur cum iIIo, sed terminus resolubilis non
convertitur cum suis inferioribus ...
Cf. Billingham, De probationibus termillorum (prior recensio), p. 52: Et in hoc differt
<terminus exponibilis> a resolubili quia licet sequatur formaliter: 'hoc currit; et hoc est
homo; ergo homo curri!" tamen non econverso: sed in exponentibus termini
exponibilis sequitur sic et etiam econverso; ut 'tantum homo currit; igitur homocurrit
et nihil aliud ab homine curri!'; et econverso. Cf. Maierii, Terminologia Logica, p.
344.
 28 ROBERT ANDREWS

universalis tantum sic quod non singularis."88 In hac definitione tres

ponuntur particulae diversae, triplices terminos denotantes officiales.

Primo cum dicitur "est terminus", sic est intelligendum mediatus,

"circa propositionem importans officium mentis vel animae." Et
140 additur 'mediatus' propter pronomina demonstrativa. Illa enim,
quamvis importent actum animae circa propositionem eo quod
singularizent, tamen quia non sunt termini mediati sed immediati,
ergo excluduntur. Omnis ergo terminus mediatus cui correspondet
aliquis actus animae circa totam propositionem potest dici terminus
145 officialis, ut sunt signa universalia, quibus correspondet
<propositio> universalis apud animam. Tota enim propositio dicitur
'universalis' ratione signi. Signum ergo circa propositionem importat
officium /274va/ animae.

Etiam coniunctiones, tam copulativae quam disiunctivae,

150 temporales, (totales MS) et causales89 dicuntur termini officiales
quoniam officium animae important circa propositiones in quibus
ponuntur. Quando copulativa coniunctio, hoc denotat circa
propositionem quod sensus coniuncti per earn nunquam possunt
constituere aliquem sensum totalem. Ipsa enim coniunctio non
155 coniungit idem inter diversa, et ipsa copulative accepta coniungit
propositi ones categoric as pro constitutione hypothetice. Duae
quidem propositiones categoricae unum sensum non constituunt nisi
ambae sint verae, quia "verum vero consonat".90

Et debet addi ad definitionem: 'Terminus officialis est terminus',

160 supple mediatus, ut excludatur pronomen demonstrativum - quod
licet importat officium animae circa propositionem ut singularizaret,
tamen quia est terminus immediatus et non mediatus, ideo per illam
particulam sit exclusum.

Secunda pars definitionis dicit: "vel est determinatio alicuius

165 complexi". Unde terminus dupliciter potest aliquod complexum
determinare. (Determinatio alicuius complexi intelligitur dupliciter
adn. in mg.) Uno modo complexum ex terminis in quantum termini
sunt, et sic non debet intelligi definitio. Alio modo terminus
determinat complex urn, id est, compositionem subiecti cum
170 praedicato, sic quod sit determinativus unius et compositivus
extremorum. Et sic intelligitur definitio. Et ut sic modi constituentes

88Billingham, Speculum puerorum (recensio italica), p. 122: Terminus officiabilis

potest dici quilibet terminus qui importat in se aliquod officium positivum vel
privativum circa propositionem.... Et universaliter: quecumque sunt respectu alieuius
complexi ... ; et universaliter: quecumque sunt respectu alicuius universalis, vel possunt
esse ...
89Cf. Boethius, De syllogismo hypothetico (Patrologio Latina 64), p. 835. Cf.
Gualterus Burlaeus, De puritate artis logicae; tractatus iongior, p. 107: Intelligendum
est, quod communiter assignantur quinque species propositionis hypothetieae, scilicet
conditionalis, causalis, temporalis, copulativa, disiunctiva.
90Aristoteles, Ethica ad Nicomachum I, 8 1098blO-1; cf. Hamesse 233 (15): Omnia
vera vero consonant.
 THE SOPHISTRIA OF PETRUS aLAI: APPENDIX 2 29

propositiones modales dicuntur termini officiales ex eo quod

compositionem determinant in propositione.

Tertia pars definition is est "vel est determinativus alicuius

175 universalis tantum sic quod non singularis". Ubi considerandum
quod haec pars dupliciter potest intelligi. Uno modo sic: terminus
officialis est qui determinativus est universalis sic quod non
singularis, id est, qui terminus additus facit ipsum in universali
supponere pro pluribus suppositis copulative. Et sic non debet
180 intelligi, quoniam sie supervacue adderetur haec particula; quia hoc
satis expressum est per primam partem definitionis, eo quod signa
talia important officium animae circa propositionem. Sed alio modo
intelligitur aliquem terminum esse determinativum universalis sie
quod non singularis, hoc est, quod faciat terminum se sequentem
185 accipi non pro certis suppositis sed indifferenter pro omnibus in
communi, et non pro certo supposito, sie quod non sit verifieabilis
de aliquo certo supposito. Et sic stare debet definitio, quia ut sic
verba importantia actum animae appetitivum dicuntur termini
officiales ex eo quod terminum sequentem non faciunt stare pro
190 omnibus suppositis copulative, sed pro universali tantum, sic quod
non est necesse ipsum verificari de a1iquo supposito.

Sed dices, videtur quod membra praedictae divisionis idem

dicant, quia idem terminus dicitur secundum utrumque modum
"determinativus universalis".

195 Respondetur quod diversitas est illa: quia terminus per primum
membrum signatum denotat term inurn communem stare pro
suppositis suis copulative, et de quo1ibet eius supposito verificabilis
est. Sed sie /274vb/ non est de termino signato per secundum
membrum divisionis, quia ta1is non denotat terminum communem
200 accipi pro pluribus copulative, sed denotat terminum universalem et
signatum adaequatum ipsius termini, sic quod de omnibus
indifferenter sit verificabilis, et non de aliquo certo. Exempli gratia:
ut cum dico 'appeto vinum', terminus communis 'vinum' non
denotat accipi pro omnibus suis suppositis copulative, sic quod
205 verum sit dicere 'appeto hoc vinum, et (s.l.) hoc vinum, et hoc
vinum, et sic de aliis'; sed capitur ibi terminus pro quolibet vino
indifferenter, et non pro aliquo certo. Sed sie non est hie cum dico
'omnis homo currii'. Hie enim cursus denotat omnibus suppositis
simul inexistere, et cuilibet seorsum. Et ergo patet diversitas inter
210 illos terminos.

Notandum quod idem terminus communis et mediatus potest dici

resolubilis, exponibilis, et official is diversis rationibus,91 quod non
est inconveniens, quia membra divisionis possunt coincidere in re,
non tamen in ratione. Sie est hie, quoniam alia est ratio qua dieatur
215 exponibilis, ut quia eius propositio habet alias propositiones sibi

91 Billingham, De probationibus terminorum (prior recensio), p. 52: sunt aliqui termini

qui dicuntur resolubiles exponibiles officiabiles ... Cf. Maierii, Terminoiogia Logica,
p.346.
 30 ROBERT ANDREWS

aequivalentes; et qua resolubilis, ut quia habet terminum inferiorem

et superiorem se notiorem ratione cuius possit propositio probari; et
etiam officialis dici potest ex eo quod officium animae importat circa
propositionem.

220 His praemissis, sit conclusio responsalis: definitiones terminorum

resolubilis, exponibilis, et officialis sunt bonae. Patet conclusio
auctoritate magistri eas ponentis; et ratione communi sumpta ex
conditionibus definitionis cuiuscumque. Ipsae enim definitiones
sufficienter definita notificant et exprimiunt earum circumstantias,
225 facientes eas distingui ab aliis sub definitis non contentis, ut patet in
corpore quaestionis diligenter intuenti. Arguitur primo: definitio
termini resolubilis non valet; ergo <etc.>. Antecedens probatur: nulla
bona definitio debet dari de definito per disiunctionem; definitio
termini resolubilis datur per disiunctionem; quare <etc.>. Maior
230 probatur quia definitio dicit quiditatem definiti, quae sine
disiunctione debet exprimi. Minor patet quia dicitur in definitione
"est terminus habens superiorem vel inferiorem se".

Respondetur: definitio non debet dari per disiunctionem si fuerit

expressiva quiditatis definiti. Si vero fuerit quaedam notificatio
235 exprimens conditiones definiti, quae sibi sub disiunctione insunt,
tunc non est inconveniens.

Arguitur sec undo. Definitio secunda non valet; ergo <etc.>.

Antecedens probatur quia omnia in definitione posita debent esse
vere praedicabilia de definitio; sed sic non est hic; igitur <etc.>.
240 Maior probatur quia in definitione praedicatur genus et differentia,
/275ra/ quae ambo de specie sunt praedicabilia. Minor patet quia ibi
ponitur haec particula "habens duas propositi ones vel plures
aequivalentes", et haec non est vere praedicabilis de termino.
Probatur quia est falsa praedicatio: 'terminus est habens duas vel
245 (prop add. et del.) plures exponentes'. Probatur quoniam (quia a.c.)
exponentes sunt ipsius propositionis, et non termini.

Respondetur quod definitio bene intellecta est bona. Debet enim

sic intelligi: "Terminus exponibilis est terminus habens" etc., id est,
ratione cuius propositio in qua ponitur habet (ex MS) "duas vel
250 plures exponentes"; et non debet ita rude intelligi sicut stat, eo quod
solius termini non est exponi, sed totius propositionis; quare <etc.>.

Arguitur tertio. Tertia definitio etiam est minus bene posita; quare
<etc.>. Antecedens probatur. Nullus terminus importat officium
animae circa totam propositionem. Probatur quia si aliquis, maxime
255 signum universale; sed hoc non. Probatur quia tale importat tantum
actum distribuendi. Actus autem distribuendi non est totius
proposition is, sed solius subiecti, ut patet per regulam logicorum.
Respondetur quod adhuc importat officium animae circa totam
propositionem, quoniam tota propositio eius ratione dicitur
260 universal is. Etiam importat officium animae circa propositionem quia
circa partem eius, ut quia subiectum distribuit; per consequens totam
propositionem universalizat; quare etc.
 Ferrybridge's Logica: a Handbook for solving Sophismata
by Mario Bertagna

The reason why an analysis of Richard Ferrybridge's Logical is

interesting with regard to sophismata is that it contains a general theory of
truth, on which every attempt to "solve" sophisms (i.e. to determine their
truth value) should be founded. Moreover, a connection between his work
and sophisms is established by the author himself: for, quoting a truth
definition for propositions later on disproved, he explicitly describes its
aim as "to prove sophisms and every other proposition". Moreover, as we
shall see, the term sophisma occurs, along with or as a substitute for the
term propositio, in some truth definitions considered by Ferrybridge.

In respect to its content, Ferrybridge's Logica may be divided into

three main sections, dealing in tum with truth conditions for <i> assertoric
propositions, <ii> modal propositions and <iii> hypothetical (molecular)
propositions. In this paper I shall try to illustrate the basic features of this
theory and at the same time to point out, within the limits of our present
knowledge, its main polemical and doctrinal sources. Therefore, I shall
mainly deal with the first part of the Logica, referring to the content of the
remaining parts only when necessary.

When he introduces his own opinion about the question of truth

conditions for assertoric propositions, Ferrybridge sums up the results of
the foregoing sections (especially of the first chapter of the fIrst part) by
claiming to have shown that

1The Logica by Richard Ferrybridge (an English logician from the middle of the XIVth
century, who also wrote a tract on consequences) is preserved in six manuscripts. To
those cited by F. del Punta, "La Logica di Richard Feribrigge nella tradizione
manoscritta itaIiana", in English Logic in Italy in the 14th and 15th Centuries, Acts of
the 5th European Symposium on Medieval Logic and Semantics, ed. A. Maieril,
Napoli: Bibliopolis 1982, pp. 53-85 - namely <i> Padua, Biblioteca Universitaria
1123, ff. 79v-93r; <ii> Vatican City, Biblioteca Vaticana Vat. Lat. 2136, ff. 35r-50r;
<iii> Vatican City, Biblioteca Vaticana Vat. Lat. 2189, ff. 81r-105v; <iv> Rome,
Biblioteca Casanatense MS 85 (D.IV.3), ff. 66r-ll1v; and <V> Seville, Biblioteca
Capitular Colombina 5-1-14, ff. 85r-13Or - we have in fact to add the MS <vi>
Worcester, Cathedral F. 116, ff. 51-61 (incomplete), signalled by L.M. de Rijk, Some
14th Century Tracts on the Probatio Terminorum, Nijmegen, 1982, p. *34*.
Some excerpts from the Logica were edited by F. del Punta in Paul of Venice, Logica
Magna, Part II, Fasc. 6, Tractatus de Veritate et Falsitate Propositionum, Tractatus de
Significato Propositionis, ed. F. del Punta, tr. M. McCord Adams, Oxford: Oxford
University Press 1978, Appendix I, pp. 215-35 (cf. also F. del Punta, "La Logica di R.
Feribrigge...", pp. 63-85). My quotations are from the above-mentioned edition (based
on the Padua MS) or from the Padua MS itself, whose transcription was kindly put at
my disposal by Professor del Punta.

31
32 MARIO BERTAGNA

"<a> It is not the case that if it is in reality as it signifies, a

proposition is true, <b> nor is it the case that if it is not in reality as it
signifies, a proposition is false."2

Clauses <a> and <b> mean respectively that

<c> being such in reality (ita esse ex parte rei) as a proposition

signifies

is neither a sufficient nor a necessary condition for a proposition to be true.

In fact, in the first chapter of the first part Ferrybridge takes into
consideration three consequence-shaped definitions of truth:

<Cl> this sophism signifies exactly as it is, therefore this

sophism is true

<C2> exactly as it is this sophism or this proposition signifies,

therefore this proposition is true;

<C3> things are wholly such as this proposition signifies,

therefore this proposition is true.

Moreover, the converse of <Cl> is examined too, viz.

<CI '> this proposition is true, therefore this proposition signifies

exactly as it is.3

Notice that, while <CI '> expresses a necessary condition for a proposition
to be true, <CI>-<C3> express sufficient conditions. Ferrybridge rejects
all earlier definitions on the ground that, according to their advocates, the
significatum of a proposition is something (aliquid) or a mode of a thing
(modus rei) which actually exists in the physical world (ex parte rei).4

2Cf. Paul of Venice, Logica Magna, Appendix J, p. 229, II. 4-6.

3Cf. Paul of Venice, Logica Magna, Appendix I, p. 215, II. 11-21.
4Consequences <Cl>, <Cl'> and <C2> are often connected by Ferrybridge with a
theory according to which the significatum of a proposition consists of a mode of a
thing (modus rei or aliqualiter esse). This theory is attributable to Richard Billingham,
who holds it in the quaestiones "Utrum idem Sortes et Sortem esse, et sic de singulis.
et ita de re et de termino" and "Utrum propositio affirmativa vel negativa vera sit quae
praecise significat sicut est" (both works were edited by M.J. Fitzgerald in Richard
Brinkley's Theory of Sentential Reference, "De Significato Propositionis" from Part V
of his Summa nova de Logica. Leiden: Brill 1987. pp. 123-42 and pp. 143-53); the
theory of truth contained in the second quaestio, on the other hand. seems strictly
connected to the consequentiae <Cl>, <Cl '> and <C2>. The first and the third
consequences correspond to the secunda and tertia via on the truth of a proposition
quoted by Paul of Venice. Logica Magna, p. 14 and p. 16.
Consequentia <C3>, corresponding to the quarta via in Paul of Venice (Cf. Logica
Magna, p.36), is also quoted by Henry Hopton, De Veritate et Falsitate
Propositionum, Venetiis, 1494, 185ra 14 ff. (the editor ascribes this work to William
Heytesbury; for the ascription to Hopton see L.M. de Rijk, Some 14th Century Tracts,
p. *39*).
 FERRYBRlDGE'S LOGICA 33

On this assumption we can in fact show that <Cl> - and consequently

<c> - is not a necessary condition for a proposition to be true, by
remarking that there are true propositions (such as negative assertoric
propositions, affirmative past- or future-tense propositions, affirmative
present-tense propositions in which an ampliative or a distracting term does
not occur) to whose significatum nothing corresponds ex parte rei. For
instance, the propositions 'Caesar was' and 'No chimera is' are true, yet
nevertheless they do not signify any actually existing object.

On the other hand, the criticism of <Cl>-<C3>, aimed at the

sufficiency of condition <c>, is grounded on an argument we can
reconstruct as follows: <i> every proposition has a significatum; <ii> the
significatum of a proposition is something existing ex parte rei; <iii>
therefore, every proposition signifies something existing ex parte rei; <iv>
so, if <c> were a sufficient condition for a proposition to be true, every
proposition would be true, because every proposition would signify the
way it is (or as things are).

According to this argument, it is not possible to ground the notion of

truth on the notion of significatum and at the same time to maintain that the
significatum of a proposition is an actually existing thing, either a res or a
modus rei.

As we shall see, being unwilling to give up the opportunity for

explaining the concept of truth by that of significatum, Ferrybridge is
progressively led to rework it. Such a task is articulated in two main steps:
<i> firstly, he distinguishes between the meaning and reference of a
proposition and identifies the significatum with the reference of a
proposition; <ii> secondly, he identifies the reference of a proposition with
an object of thought (obiectum intelligibile).

In the second chapter of the first part of his Logica Ferrybridge

determines what the significatum of a true proposition is. To this end he
first sets out to reject two rival theories. According to the first, held by
Richard BilIingham,5 the significatum of a true proposition is a mode of a
thing (modus rei). Understanding such a modus rei, possibly beyond
Billingham's own intention, as an extramental entity distinct from
substances and their properties, Ferrybridge criticizes this theory mainly
because it seems to him a useless and consequently undue amplification of
the Aristotelean ontology. For it is possible to furnish an adequate
description of the significatum of a true proposition without going beyond
the domain of res.

The second theory taken into consideration, ascribed by Richard

Brinkley to William Bermingham,6 states that the significatum of a true
proposition is a compositio mentis, namely the mental act by which our

5Cf. note (4). The theory at issue is quoted and criticized also by Henry Hopton, De
Veritate, I84r bil - 184v al (cf. also Paul of Venice, Logica Magna, pp. 80-4).
6Cf. M.J. Fitzgerald, Richard Brinkley's Theory, p. 3. The theory rejected by Henry
Hopton, De Veritate, 184va 2-50, even if it is clearly similar to that ascribed to
Benningham, nonetheless does not seem to coincide with it.
34 MARIO BERTAGNA

mind compounds (in the case of an affirmative proposition) or divides (in

the case of a negative proposition) two concepts from each other. Such a
theory, looking upon the signiftcatum of a true proposition as an entity that
exists only inside our mind, is criticized by Ferrybridge because, among
other things, it compels us to consider as false propositions commonly
accepted as true. For instance, the proposition 'God is' should be false
according to this theory if there did not exist any mind composing it.

Unlike the above-mentioned theories, Ferrybridge's assigns different

signiftcata to different kinds of assertoric propositions. In fact:

"<i> the significatum of every true affirmative present-tense

proposition whose principal verb or predicate is not an ampliative
term is the thing signified by its terms (...) If the different terms of a
proposition signify different things, as they signify different things
so the signiftcatum of the proposition will consist of more things.
For instance, the significatum ofthe proposition 'God is' (Deus est)
is that God is (Deum esse), and that God is is nothing but God.
Likewise, the proposition 'The sun warms the house' signifies that
the sun warms the house, which is nothing but the sun, and signifies
that the house is warmed by the sun, which is nothing but the house.
Moreover, this opinion holds that <ii> the significatum of no
negative true proposition is something (aJiquid) or a mode of being
(aliqualiter) or some things (aliqua); <iii> nor of a true affirmative
future-tense proposition to whose subject no actually existing thing
corresponds; <iv> nor of a true affirmative past-tense proposition for
which the thing mainly signified by its subject does not exist; <v>
nor of a true present-tense proposition whose predicate is a
distracting term, e.g. 'A man is dead', or an ampliative term and
where the thing signified by its subject does not exist, e.g. 'A
chimaera is thinkable'."7

Therefore, according to Ferrybridge's opinion, the significatum of a

true affirmative present-tense proposition that does not contain any
ampliative or distracting terms is the significatum of its subject and its
predicate term, that is, the thing(s) which they actually supposit for. Yet,
this statement should not be understood in a narrow sense. For
Ferrybridge does not mean that the significatum of such a proposition
exactly coincides with the signiftcatum of its categorematic terms, as if its
syncategorematic terms would by no means contribute to building up its
signiftcatum. The wording of the thesis possibly suffers from the need to
exclude the possibility that the significatum of a proposition depends on
something else apart from the things which its categorematic terms refer to.

7Cf. Paul of Venice, Logica Magna, Appendix I, p. 223, II. 12-28. An ampliative term
is a tenn that enlarges the supposition of the subject term of the proposition in which
it occurs as principal verb or predicate beyond the present things (e.g. the verb 'will be'
in the proposition 'The Antichrist will be' enlarges the supposition of the term
'Antichrist' to those men that will exist). A distracting term is a term that confines the
supposition of the subject term of the proposition in which it occurs as principal verb
or predicate to different things from those that actually exist (e.g. the term dead in the
proposition 'Some man is dead' confines the supposition of the term man only to
those men who were).
 FERRYBRIDGE' S LOGICA 35

However, the signijicatum of these propositions must be identified with a

thing in a certain state (or things in a certain state). For instance, the
signijicatum of the proposition 'the sun wanns the house' does not simply
consist in the sun and the house, but in the sun in so far as it wanns the
house, and in the house in so far as it is wanned by the sun.8

That this is the right interpretation of Ferrybridge' s theory comes out

also from his replies to some objections to it. One of them, for instance,
asserts that according to the opinion held by our author it follows that a
term (e.g. 'God') is convertible into any true proposition in which it occurs
as subject term (e.g. 'God is'), since they possess the same signijicatum.9
Ferrybridge replies to this argument as follows:

" .. .1 agree that the subject of a proposition signifies the same

thing as the proposition in which it is the subject. However the
proposition at issue [sc. 'God is'] signifies something more than its
own subject does, since it signifies everything that its subject or its
predicate signifies. For its predicate, namely the verb 'is' or the term
'being' [since 'God is' is equivalent to 'God is a being'] signifies
being. But what is signified adequately by the proposition is only
what is signified by its terms. I mean by 'adequate signijicatum' that
by which the composition of the predicate with the subject (actually,
that God is - Deum esse) is made. And from that it does not follow
that a proposition and its subject are interchangeable. For, to be
interchangeable we need something more than the signijicatum. For
the proposition 'God is' signifies finitely and determinately that God
is, and by means of this signijicatum our mind is turned to truth or to
falsehood. On the contrary the term 'God' does not signify that God
is. For as the different mode of signifying distinguishes the terms of
a proposition from each other, so it particularly is the cause of the
distinction between a proposition and its terms."10

8This is the common interpretation of the theory held by Ferrybridge, known as the res
theory, that had a wide currency in the Middle Ages and was held by, among others,
William Chatton and Andree de Neufchatel. (Cf. o. Nuchelmans, Ancient and Medieval
Theories of the Propositions, Amsterdam-New York, 1973, pp. 209-25 and pp. 254-9.)
9"ltem ex hoc sequitur quod idem significat subiectum propositionis et tota ilia
propositio, quia cum Deum esse sit Deus, idem praecise significat iste terminus 'Deus'
sicut haec propositio 'Deus est'. Et ex hoc sequitur quod subiectum propositionis et
ipsa propositio convertuntur. Sequitur enim: quidquid significat A significat B, et e
contra; ergo A et B convertuntur" (MS Padua, 83va).
lO"Ad aliud, quando arguitur quod subiectum propositionis idem significat sicut tota
propositio cuius est subiectum, hoc conceditur. Tamen ilia propositio de qua ibi fit
locutio plus significat quam subiectum eiusdem propositionis, quia haec propositio
'Deus est' significat quidquid significat subiectum vel praedicatum eiusdem. Nam
praedicatum eiusdem, sive sit hoc verbum 'est' sive iste terminus 'ens', significat ens
totum. Tamen quod per totam propositionem significatur adaequate non est aliud quam
quod significatur per partes eiusdem. Voco autem adaequatum significatum iIIud pro
quo fit compositio praedicati cum subiecto eiusdem, et iIIud in proposito est Deum
esse. Et ex hoc non sequitur quod propositio et eius subiectum convertuntur. Plus enim
requiritur quam iIIud significatum ad hoc quod convertantur: quia propositio ilia
significat finite et determinate Deum esse, qua significatione intellectus determinatur ad
verum vel ad falsum; et non sic significat terminus <'Deus'> Deum esse, quia sicut
modus significandi distinguit partes orationis ab invicem, ita modus significandi
particulariter est causa distinctionis orationis a parte eiusdem" (MS Padua, 84rb).
36 MARIO BERTAGNA

Ferrybridge opens his reply by saying that a true affirmative present-tense

proposition that does not contain any ampliative or distracting term
signifies ex parte rei nothing but what is signified by its subject term
(namely, the things which it stands for). However, the identity between the
res signified by the proposition and the res signified by its subject cannot
warrant the identity of their significata. For, unlike the term 'God', the
proposition 'God is' does not simply signify God, but it signifies that God
is, namely, God in so far as He is an actually existing being. The mode in
which a proposition signifies a res is in fact essentially different from the
mode in which it is signified by its subject taken in isolation; for every
categorematic or syncategorematic term of a proposition contributes to its
mode of signifying.

With respect to the other kinds of assertoric propositions,

Ferrybridge's theory holds that <i> if their subject term signifies an
actually existing res, then their significatum will consist of this res, <ii>
otherwise their significatum will not be a thing nor a mode of a thing. This
last clause does not mean that the significatum of such propositions is
absolutely nothing, but rather that it is nothing that exists ex parte rei. A
positive definition of the notion of significatum for these kinds of
propositions will be formulated by Ferrybridge in terms of the notion of
intelligible object only in his discussion of the truth conditions for
assertoric propositions.

As he did for the notion of significatum, Ferrybridge does not

establish a general definition of truth for assertoric propositions but gives
singular definitions for each:

"<Dfl> What makes a proposition a true affirmative or negative

present-tense proposition is the fact that it exactly signifies that a
thing that is, is or exactly signifies that a thing that is not, is not;

<Df2> What makes a proposition a true affirmative or negative

future-tense proposition is the fact that it exactly signifies that a thing
that will be, will be or exactly signifies that a thing that will not be,
will not be;

<Df3> What makes a proposition a true affirmative or negative

past-tense proposition is the fact that it exactly signifies that a thing
that was, was or exactly signifies that a thing that was not, was
not."l1

An important specification follows these defmitions:

"I take here 'thing' (res) more generally than 'something'

(a/iquid). For this term means here whatever can be thought or is
thinkable."12

IICf. Paul of Venice, Logica Magna, Appendix I, p. 229, 11. 9-19.

12Cf. Paul of Venice, Logica Magna, Appendix I, p. 229, 11. 20-21.
 FERRYBRlDGE'S LOGICA 37

To understand the foregoing definitions in the right way it is necessary

to point out first that Ferrybridge, when speaking of the significatum of a
proposition (or of the res signified by it), actually means its reference and
not its meaning. On the other hand, the distinction between meaning and
reference of a proposition is actually present in our author's work in the
form of the distinction between the significatio and the significatum of a
proposition, even if it is sometimes formulated in a rather puzzling way
and above all is not used systematicaIly.l3

The distinction between significatio and significatum is made clearly

explicit by Ferrybridge in his discussion of some objections to his
definition of necessary proposition, or rather to a consequence of it,
according to which a proposition is necessary if its significatum is so, that
is, if it signifies a necessary res. One of the objections considered says that
the proposition

(1) Every man is

will be contingent and not necessary according to the theory held by

Ferrybridge. A proof can be given on the ground of the equivalence
between (I) and

(2) Socrates is and Plato is and Cicero is.

For provided that Socrates, Plato and Cicero are the only actually existing
men, according to Ferrybridge's theory (1) and (2) have the same
significatum and so they are equivalent. However (2) is a contingent
proposition; therefore so is (1).14

Ferrybridge replies to this argument by denying that (1) and (2) have
the same significatum. For

"although the universal proposition [sc. (1)] now signifies that

Socrates is and Plato is <and Cicero is>, at another time it will

13The unsystematic use of this distinction leads to two unfortunate consequences: <i>
the verb significare is employed both for 'to mean' and for 'to refer to', <ii> the
infinitive expressions (such as Deum esse) are used both to denote the meaning and the
reference of a proposition. These ambiguities very often give rise to difficulties and
misunderstandings. The modern distinction between meaning and reference relies upon
that between sentence-type and sentence-token (Cf. Y. Bar-Hillel, "Indexical
Expressions", Mind 63,1954, pp. 359-79). Even this last distinction, as we shall see
later on, is implicitly contained in our work. Notice that the aim of the resort to these
concepts is not to reconstruct in a systematic way a medieval theory in the light of
modern concepts, but rather to make clear its implicit assumptions and to understand
its internal difficulties.
14"Item supposito quod de facto nisi tres homines sint - Socrates, Plato et Cicero -
tunc haec propositio 'omnis homo est', significando omnem hominem esse, significat
Socratem esse, Platonem esse et Ciceronem esse. Et non aliud, quia non significat
omnem hominem esse qui non est, eo quod tunc esset falsa (... ) Sequitur ergo quod ilia
propositio 'omnis homo est' significat Socratem esse, Platonem esse et Ciceronem
esse. Ex quo ultra arguitur sic: haec copulativa est contingens, ergo ilia universalis est
contingens" (MS Padua, 87vb-88ra). Ferrybridge does not explain why 'Every man is'
should be a necessary proposition, but this doesn't matter here.
38 MARIO BERTAGNA

signify that Mark or Tully is, and not that Socrates is and Plato is and
Cicero is. For whenever such a proposition will exist, it will signify
only that every man existing at that time is."15

This response is based on an implicit distinction between sentence-

types and sentence-tokens: even if two contemporary tokens of (1) and (2)
may have the same reference, as in the posited case, this does not imply
that the types (1) and (2) have the same meaning, as the author's
counterexample shows. Notice that in this passage Ferrybridge uses the
Latin verb signijicare ambiguously both for 'to have a meaning' and for 'to
have a reference". These two concepts, however, are roughly
distinguished in the response to a succeeding objection to the necessity of
(1), that we can sum up as follows.

Let us assume that (1) is true at time t and that Socrates exists at t: it
follows that (1) at t (also) signifies that Socrates is. Let us now suppose
that at time t', following t, Socrates has ceased to exist. Then, since (1)
will signify at t' whatever it signifies at t, (1) will signify at t' that Socrates
is. But since Socrates does not exist at t', (1) will be false at t' and
consequently it will not be necessary.l6

This reasoning obviously misunderstands the meaning and reference

of a proposition. Ferrybridge points it out in his response:

" ... this does not follow: however a proposition now signifies, it
will then signify; therefore, whatever it now signifies, it will then
signify. For although the meaning (significatio) of this proposition
does not change in time, it will refer (significabit) to different things
at different times, since the references of its meaning will always be
the men who will exist at the time when the proposition will be
uttered with such a meaning."17

As can be seen, in this passage Ferrybridge distinguishes the unchangeable

meaning of a sentence-type, i.e. its signijicatio, from what is referred to by
a sentence-token when uttered, i.e. its signijicatum.

In the light of that, we can conclude that Ferrybridge's theory of

significatum previously outlined is actually a theory of reference.
Consequently his theory of truth must be considered as a theory of truth-
conditions for sentence-tokens grounded on the notion of reference. For

15".. .Iicet haec universalis nunc significat Socratem esse, Platonem esse <et Ciceronem
esse>, alias significabit Marcum vel Tullium esse, et nee Socratem esse nee Platonem
esse nee Ciceronem esse, quia quandocumque haec universalis erit, solum significabit
omnem hominem esse qui erit in eodem instanti" (MS Padua, 89ra).
16"Item sic: corrupto Socrate et manente ilia universali praecise sic significante sicut
ipsa nunc significat, erit haec universalis falsa. Sed nulla talis est necessaria. Ergo etc.
Assumptum probatur: ilia universalis, corrupto Socrate, quodlibet significabit quod
ipsa nunc significat; et ipsa nunc significat Socratem esse; ergo tunc significabit
Socratem esse" (MS Padua, 88ra).
17" ... non sequitur: qualitercumque ilia nunc significat tunc significabit; ergo quidquid
nunc significat tunc significabit. Licet enim non varietur sua significatio, aliud tamen
significabit, cum suae significationi semper subicientur illi homines qui tunc erunt
quando ipsa erit significans sic" (MS Padua, 89ra).
 FERRYBRIDGE'S LOGICA 39

the notion of meaning is useless in order to determine the truth conditions

of a proposition.

In the first chapter of the first part, in response to Ferrybridge's

criticisms of <Cl'> the following reply is made:

" ... when a proposition signifying that the Antichrist will be

actually exists, then it is the case that the Antichrist will be; that is,
what is signified by this proposition is nothing but being in sign
(esse in signo). But this sort of being is not being in an absolute
sense (simpliciter), but in a qualified one (secundum quid). For
according to Aristotle, Posterior Analytics B <c.7, 92b 6-8>, there
are a lot of things that are signified and do not exist, and so they exist
only in their signs. Therefore, from the fact that such being exists in
a proposition as in its sign it does not follow that it absolutely is."18

The respondent seems to conceive the significatum of a proposition as

its meaning and to confer on it an intermediate existence between the level
of language and that of the physical world, depending on the existence of
the proposition that conveys it. However, Ferrybridge argues, it is not
possible to ground the notion of truth on that of meaning. For if a
proposition is true when it is as it signifies, and if what is signified by the
proposition is identified with its meaning, even false propositions would
be true, since they possess a meaning (quaelibet propositio aequaliter
significat suum significatum sicut alia). For in Ferrybridge's terminology
what is signified by a false proposition also exists in signo in the
proposition of which it is the meaning. 19

This objection is not without a response:

" ... there is a difference between the significatum of a true

proposition and the significatum of a false one. For in the case of the
proposition 'No chimera is' (and in the case of any similar
proposition), the cause of such being in the sign or in the proposition
signifying it, is not only that it is the case that no chimera is, but also
that no chimera is and that the proposition 'No chimera is' signifies
this way. This is not the case, however, for the proposition 'A man
is a donkey'."2o

The aim of this argument is to determine additional conditions apart

from the existence of the proposition signifying it to state that the meaning
of a proposition exists (in a qualified sense), in order to reach a distinction
between the meaning of a true proposition and the meaning of a false one

18Cf. Paul of Venice, Logica Magna, Appendix I, p. 218, II. 25-32.

19Cf. Paul of Venice, Logica Magna, Appendix I, p. 218, II. 45 - p.219, I. 3. This
argument should not be mistaken for that against the sufficiency of condition <c>,
although similar to it. For by the former Ferrybridge wants to demonstrate that the
theory of truth cannot be based on the notion of reference conceived in any way as an
entity that exists ex parte rei, while by the latter he shows that it is not possible to
ground the theory of truth on the notion of the meaning of a proposition.
20Cf. Paul of Venice, Logica Magna, Appendix I, p. 219, II. 4-9.
40 MARIO BERTAGNA

and consequently to avoid the preceding difficulty. Let us consider for

instance the following proposition

(3) No chimera is.

Now, the causes of the qualified being of its meaning, namely that no
chimera is, are <i> it is the case that no chimera is and <ii> proposition (3)
signifies that no chimera is. Likewise, the causes of the being of the
meaning (and so of the truth) of the proposition

(4) A man is a donkey

will be <iii> it is the case that a man is a donkey and <iv> proposition (4)
signifies that a man is a donkey. Now, since condition <i> is met while
condition <iii> is not, it follows that the meaning of (3) is (namely, what is
signified by (3) is), while the meaning of (4) is not (namely, what is
signified by (4) is not). Therefore, (3) is true and (4) is false.

The distinction between the significatum of a true proposition and the

significatum of a false one, on which this argument is based, is actually a
distinction between propositions whose reference subsists and
propositions whose reference does not subsist. And this overcomes
Ferrybridge's objection only if the significatum of a proposition is
conceived as its reference, as shown by the fact that the existence of the
significatum is made to depend not only on the properties of signifying
which a proposition posesses but also on the subsistence of a determinate
state of things. Let (P,t) be a token of Pat t, p the meaning of P and (p,t)
the reference of P at t (or of (P ,t»; it follows that according to this
conception a proposition (P ,t) is true iff it is the case that (p,t), that is iff
<a> P signifies that p and <b> it is the case that (p,t) (provided that P is
uttered at t). Now, it is clear that, even if it is possible that every false
proposition meets condition <a>, it is impossible that it meets <b>.

Ferrybridge considers this theory "commendable" (commendabilis)

because it is near the truth. However he criticizes it on the ground that it
contains a petitio principii: for to assume that it is the case that (p,t) (i.e.
that no chimera is) to demonstrate that it is the case that (p,t) (i.e. it is the
case that no chimera is) is to assume what has to be proved. In other
words, to define the truth of a proposition it is not necessary to postulate
that its reference is an entity having a qualified existence caused by its
meaning and by objective reality (in a broad sense), but it is sufficient to
understand it for a member of this reality. So only conditions <a> and <b>
are sufficient to define the truth of an assertoric proposition. 21

However, we have to notice that according to Ferrybridge only

condition <b> always means the cause ex parte rei of the truth of a
proposition, while this holds for <a> only for affirmative present-tense
propositions not containing an ampliative or distracting term. For, when he
explicates the above-mentioned truth definitions, our author clearly says
that

21Cf. Paul of Venice, Logica Magna, Appendix I, p. 219, 11.10-18.

 FERRYBRIDGE'S LOGICA 41

"the cause of the truth of affinnative propositions whose principal

verbs or predicates are ampliative tenns is the meaning (significatio)
by which they signify their reference (significatum) ... <For
instance> the cause a parte rei of the truth of the proposition 'A
chimera is significable' is the meaning by which it signifies that a
chimera is significable."22

The same also holds for (affinnative and negative) future- and past-
tense propositions. The expression causa ex parte rei and the tenn causa
are used by Ferrybridge to point out something existing in the physical
world. The significatio of a proposition is conceived as something real in
the specified way possibly because it constitutes a property of a
proposition and a (mental, vocal or written) proposition is always
considered by medieval logicians as a physically existing entity (for its
existence is a necessary condition for it to be true). The reference
(significatum) of a proposition, on the other hand, does not necessarily
have to belong to the physical world and consequently it cannot be
considered a causa of the truth of the proposition signifying it.

This statement is nothing other than a logical consequence of

Ferrybridge's theory of reference. For let us consider the following two
affinnative present-tense propositions:

(5) God is

(6) You do not run.

According to conditions <a> and <b> the truth of (5) and (6) can be
inferred respectively from

(5a) It is the case that God is

(5b) 'God is' signifies that God is

and from

(6a) It is the case that you do not run

(6b) 'You do not run' signifies that you do not run.

However, conditions (5a) and (6a) do not have the same existential import.
For (5a) means that the reference of (5), namely God, exists ex parte rei
and then expresses a cause of the truth of (5) (ad veritatem propositionis
affirmativae de praesenti tam requiritur significatio ... quam etiam esse rei
sign(ficatae). On the contrary, (6a) does not mean that the reference of (6)
is an actually existing thing and so it does not represent a cause of its truth.
For (6) is trivially true if the tenn 'you' has no reference. On the other
hand, if (6) meant a cause of the truth of (6), the reference of (6) should be

22Cf. Paul of Venice, Logica Magna, Appendix I, p. 230, I. 48 - p. 231, I. 11.

42 MARIO BERTAGNA

an element of the physical world, contrary to Ferrybridge's theory of

reference. 23

But then, which thing is the reference of (6), namely you in so far as
you run, if it is not an element of such world?

Along with the particular truth-definitions <Df1>-<Df3> Ferrybridge

formulates two regulae that express general truth and falsehood conditions
for assertoric propositions:

"<Df4> Every proposition that finitely or determinately exactly

signifies the truth is true;

<DfS> every proposition that finitely or determinately exactly

signifies a falsehood is false. The terms 'true' or 'false' mean here a
true or a false signijicatum."24

With the proviso that the significatum of a proposition is identical with its
reference, according to the foregoing definitions (and disregarding the
clause finite vel determinate), we have that <i> a proposition is true if it has
a true reference and <ii> a proposition is false if it has a false reference.
Thus the notion of truth of a proposition is reduced to the notion of truth of
its reference. This compels Ferrybridge to determine the nature of the
reference of a proposition and its "truth conditions".

Both of these problems are faced and solved by the author immediately
after, when he is about to establish quid sit ratio veri et quid falsi.
According to Ferrybridge,

" .. .' true' and 'false' are terms that can be predicated of the terms
denoting the reference of a proposition and they are ampliative terms
(...) Therefore the true, i.e a true reference, is a conceived being (ens
intellectum) by whose conception (intellectionem) the mind
conceiving it is verified. The false is a conceived being by whose
conception the mind conceiving it is falsified.''25

To assert that the terms 'true' and 'false' are predicated of the terms
signifying the reference of a proposition (such as infinitive expressions like
Caesaremfuisse, which signifies the reference of the proposition Caesar
fuit and of which the term verum is predicable in the proposition Caesarem
fuisse est verum), means to confirm that they supposit for references.
Besides, by saying that these terms are ampliative ones Ferrybridge means
to point out that the reference of a proposition is not necessarily something
actually existing. For generally speaking the reference of a proposition is a

23Cf. Paul of Venice, Logica Magna, Appendix I, p. 229, I. 24 - p. 230, I. 14.

24Cf. Paul of Venice, Logica Magna, Appendix I, p. 231, II. 38-43. The aim of the
clause "detenninately or finitely" (finite vel determinate) is to exclude infinitive
expressions standing for the significatum of a proposition, since they are not
propositions at all ("omnis oralio infinitiva in quantum talis est indelerminata et
infinila. et omnis propositio de se est determinata et finita; ergo nulla huiusmodi oratio
est propositio", MS Padua, 82rb).
25Cf. Paul of Venice, Logica Magna, Appendix I, p. 231, 1. 46 - p. 232, 1. 16.
 FERRYBRlDGE'S LOGICA 43

conceived being (ens intellectum). With this expression Ferrybridge seems

to identify the reference of a proposition with an object of thought in so far
as it is actually thought, as well as a proposition in so far as it is actually
uttered. However, conforming to a common point of view, he then
concedes that the reference of a proposition may be an esse intelligibile,
that is a possible object of thought. He brings this conception back to
Aristotle, who in Metaphysics II says that what is true (ens verum)

"does not have a real existence (esse rationale extra animam), but
is only a conceived or conceivable being (esse intellectum vel
intelligibile)."26

Since it is a conceivable being, the reference of a proposition may consist

not only of a res which actually exists in the external world (an aliquid),
but also of things that were or will be or simply can be, or even of things
that cannot be (such as the chimera). For

"as every sensation is the cognition of a sensible object, so every

concept (intellectio) of a conceivable thing (rei intelligibilis) is the
cognition of a conceivable object (obiecti intelligibilis). However, as
is clear, one may conceive many things that are not and consequently
know a lot of things that are not (... ) For there are things that can be
known and cannot cause a cognition, for instance, non-beings such
as that the chimera is or that Caesar was and so on, which are known
only by means of a cognition produced by other things."27

However we should make sure we do not confuse the conceivable

object which Ferrybridge identifies with the reference of a proposition, and
has an extramental objective existence even if it may not belong to the
physical world, with the concept it produces in one's mind, which has a
merely subjective mental existence.

Finally, as regards the "truth conditions" of a reference, Ferrybridge

confines himself to the rather obscure statement quoted before, according
to which a reference is true if one's mind is verified by its conception (i.e.
achieves a right representation of reality), while it is false otherwise.

To sum up, the analysis of the theory of truth for assertoric

propositions worked out by Ferrybridge as the resolution of a merely
logical problem leads our author to the elaboration of a real ontology,
whose aim is both not to go beyond the traditional domain of the res in
favour of onto logically suspicious entities such as modi rerum, and to
amplify this domain to include every possible object of thought.

In fact, he realizes that it is impossible to ground the theory of truth on

the notion of reference, understood in any way for something actually
existing, since this conception leads to the absurd consequence that every
proposition is true. On the other hand he excludes the possibility that the
reference of a proposition consists of a mental entity, since such a

26Cf. Paul of Venice, Logica Magna, Appendix I, p. 231,11.18-24.

27Cf. Paul of Venice, Logica Magna, Appendix I, p. 234,11. 31-33.
44 MARIO BERTAGNA

psycho logistic conception of truth may lead, when some logically possible
assumptions are made, to the equally unacceptable consequence that every
proposition is false. However, the res which a proposition refers to,
though it cannot always be considered as something existing in the
physical world nor as a subjective mental entity, must necessarily have an
objective existence in order to ground the notion of truth on it. Such
objective reality is conferred on the reference of a proposition by
understanding it for a conceivable object, belonging to the domain of what
is thinkable, including but wider than that of physical objects.

University of Pisa
 Boethius de Dacia et al.
The sophismata in MSS Bruges SB 509
and Florence Med.-Laur. S. Croce 12 sin., 3
by Sten Ebbesen

The manuscripts B (= Brugge, Stedelijke Openbare Bibliotheek, 509),

and F (= Firenze, Biblioteca Medicea Laurenziana, S. Croce 12 sin., 3)
were executed about the same time (in the early fourteenth century, it
would seem), but probably in widely different places, B perhaps in
Belgium or Britain, F in Italy (judging by styles of writing). The two
manuscripts contain partly identical collections of sophismata from the late
thirteenth century. The collections were first described by Grabmann in
1940 (F only)l then in 1962 by Roos,2 who discovered B, and most
recently by Pinborg and myself in 1970.3 A corrected and more detailed
catalogue of the sophismata is found in an appendix to the present paper.
An edition is being prepared for Corpus Philosophorum Danicorum Medii
Aevi IX.4

The Florence MS contains thirteen, the Bruges MS some seventeen

sophismata (some incomplete, which makes counting tricky). Eight
sophismata are shared by the two manuscripts, but in four cases there is
only partial identity of problems (I, V, VI, VIII). Three items (III, V, VI)
are also each transmitted in one other collection of sophismata: III in C =
Cambridge, Gonville & Caius 611/341;5 V in V = Vat. lat. 14718; VI in P
= Paris BN lat. 16089.
With two exceptions (V, X) the sophismata of the Florence manuscript
are ascribed to particular masters, while the Brugesmanuscript leaves them
all anonymous. One of the two exceptions, N° V, is attributed to Peter of
Auvergne in MS V. The order of the shared items is different in the two
manuscripts. My numbers I-XIII reflect the order in the Florence MS. Of
these only I-VIII occur in Bruges, whereas the numbers XIV-XXII
designate items not found in Florence. The Bruges series then runs:

1M. Grabmann, Die Sophismatalitteratur des 12. und 13. Jahrhunderts mit Textausgabe
eines Sophisma des Boetius von Dacien, Beitrage zur Geschichte der Philosophie und
Theologie des Mittelalters 36.1, Miinster: Aschendorff 1940.
2H. Roos, "Das Sophisma des Boetius von Dacien 'Omnis homo de necessitate est
animal' in dobbelter Redaktion", Classica et Mediaevalia 23, 1962, pp. 178-97. Idem,
"Ein unbekanntes Sophisma des Boetius de Dacia", Scholastik 38, 1963, pp. 378-91.
3S. Ebbesen and J. Pinborg, "Studies in the Logical Writings attributed to Boethius de
Dacia", Cahiers de l'Institut du Moyen-Age Grec et Latin 3, 1970.
41. Rosier is working on the three grammatical sophismata, and I myself on the rest.
5Notice that the Cambridge MS was copied in England. Ail its sophismata (ff. 47vB-
6OvB) are anonymous, but the occurrence among them of III, which is by the Parisian
master Peter of Auvergne, may mean that the whole collection consists of Parisian
works. I have earlier argued for a Parisian origin of the Perihermeneias questions on ff.
I-IVv, but an English one for the Elenchi questions on ff. 1-24v. See S. Ebbesen, "The
Dead Man is Alive", Synthese 40, 1979, pp. 43-70.

45
46 STEN EBBESEN

B's order My numbers Authors

Shared w. F B only

1 N Petrus Alv.
2 VIII Petrus Alv.
3-5 XIV -XVI
6 II Petrus Alv.
7 I Boethius
8 VII Petrus Alv.
9 XVII
10 VI Petrus Alv.
11 III Petrus Alv.
12 V Petrus Alv.
13-17 XVIII - XXII

There is no obvious rationale in the order of either manuscript, except

perhaps for one detail: whereas the majority of the sophismata are logical,
three are grammatical, and they come as one block (IX, X, XI) in the one
MS (F) that has them.

The authors, as given by F, V and P, are: Boethius de Dacia (1 ~

sophismata), Nicholas of Normandy (1), and Peter of Auvergne (9b. The
names of Boethius and Peter point towards the 1270s, but that of Nicholas
is of little help, though it is just possible that he was Nicholas of Pressoir,
who became master of theology in 1273 or a little earlier. 6 If so, his name
points to the 1260s or very early 1270s - provided, that is, we can assume
that he was in the Arts faculty when the sophisma was composed. There is
disquieting evidence, however, that not even this much can be taken for
granted. In MS Paris BN nal 1374: 94rB, the sophisma 'Omnis homo de

6B. Haureau, "Sennonnaires", in: Histoire litteraire de la France xxvi, Paris: Finnin
Didot 1873, pp. 438-9, speaking about authors of sermons in Paris BN lat. 14947,
which "se compose de sennons preches it Paris par divers docteurs, durant les annees
1281, 1282 et 1283", says: "Dans Ie meme volume, sous ne nO 100, est un sennon
prononce la meme annee," i.e. 1281, "dans la meme ville, Ie premier dimanche de
I' Avent, par un maitre en theologie nomme Nicolas Ie Normand. Nous supposons que
ce sennonnaire est Nicolas de Freauville, dont nous parlerons dans un des volumes
suivants." Apparently this is the master that P. Glorieux, Repertoire des maitres ell
tMologie de Paris au XlIle sit!cle, 2 vols. Paris: Vrin, 1933, vol. 1 pp. 382-4 registers
as NO 189 "Nicolas de Pressoir. Originaire sans doute du Pressoir, commune de
Boutigny (S.-et-O.)," for he ascribes to him a sermon "In Dom. I Adventus (30
novembre 1281)" from MSS Paris. lat. 15005: f. 95 and 14947: f. 191. According to
Glorieux he is attested as regent master of theology from 1273 and at intervals till
1293. He died in 1302. Glorieux says he was master of arts and probably became
master of theology 1273. The infonnation is repeated in P. Glorieux, La faculte des
arts et ses maitres. Etudes de philosophie medievale LlX. Paris: Vrin, 1971, p. 411.
On p. 263 of the same 1971 book Glorieux gives the authorship of our sophisma
'Albus musicus est' to Nicholas of Paris, the mid-13th century master of arts; this
seems to rest on a misreading of M. Grabmann, Mittelalterliches Geisteslebell I,
Miinchen: Max Hiiber 1926, p. 223, where Nicholas of Paris and Nicholas of
Nonnandy are both mentioned, but only to be distinguished from one another.
 BOEfHIUS DE DACIA Ef AL. 47

necessitate est risibilis' (of unknown date, but no later than early 14th c.)
ends with the words

Haec de hoc sophismate sufficiant determinata a magistro Petro de

Arvemia magistro in theologia.

Now this could mean "determined by Peter, <who is now> master of

theology <when he was a> master <of arts>." But I am more inclined to
believe that it simply means that he was master of theology when he did the
determination. One may think in this connection of a document from 1317
in which <theological?> masters of Sorbonne are allowed to continue a
practice of resuming the teaching of arts.?

The sophismata are not doctrinally homogeneous, but they do seem to

come from the same milieu, which was probably Paris about the 1270s.
They are of varying complexity. In no more than seven cases does one find
the initial probatio, improbatio and solutio (henceforward collectively
corpus sophismatis8 ) and the MSS do not even agree on including or
omitting this section. There are from one to ten problemata in each
sophisma, some 77 problems in all, four of which are just expositions of
some master's opinion - i.e. questions omissis rationibus so that only the
body of the determination is left. Of the remaining problems some 53 are
simple questions with rationes and determination, including solution of
rationes; these problems are structurally indistinguishable from
contemporary questions on Aristotle.

The borders between the genres of sophisma and quaestio can be

fuzzy; as some sophismata contain regular questions, so question
commentaries on Aristotle often contain questions inspired by sophismatic
technique, i.e. asking about the truth of a certain sentence under given
circumstances. "De veritate huius 'Caesar est homo' Caesare non
existente", for instance. 9 'Utrum dictio exclusiva addita uni correlativo
excludat reliquum' is a good question on Physics I, where Aristotle deals
with the thesis 'tantum unum est' and whether 'tantum principium est'
implies 'principiatum non est'. But it is also a good problem in a sophisma
with the title 'Tantum unum est'. Radulphus Brito's Physics commentary
exists in at least three redactions. In two of them there is a question 'Utrum
dictio exclusiva .. .',10 but in the third we read: ll

7p. Glorieux, AIU" origines de fa Sorbonne. I. Robert de Sorbon. Etudes de Philosophie

Medievale 53, Paris: Vrin 1966, p. 210.
8Term borrowed from an English sophisma. See S. Ebbesen, "Three 13th-century
Sophismata about Beginning and Ceasing", Cahiers de l'lnstitut du Moyen·Age Grec et
Latin 59, 1989, pp. 121-80, at p. 141.
9Example from Incertorum Auctorum Quaestiones super Sophisticos Elenchos, Corpus
Philosophorum Danicorum Medii Aevi (abbreviated CPhD) VII, ed. S. Ebbesen,
Copenhagen: DSLlGad 1977, p. 209.
10MSS Firenze BNC Cony. Soppr. E.I.252: 5vB- 6rA; Vat. lat. 3061: 68rA-B. The
text is widely different in the two MSS.
II MS Paris BN lat. 16160: 6vB.
48 srEN EBBESEN

Consequenter posset seu debet quaeri utrum dictio exclusiva addita

uni correlativorum excludat reliquum. Sed hoc quaeratur in
sophismatibus.

And then the text continues with another question. I propose that what
happened was that Brito first composed a question on the subject for his
Physics commentary, but excised it from a later edition because in the
meantime he had published a better version of the question as a problem in
a sophisma (unfortunately this particular sophisma of Brito's remains to be
identified, if it still exists).

However, fuzzy as the borders between question and problem may be,
questions on Aristotle never contain what we find in twenty of the
sophismata in BF, viz. a respondent's preliminary determination before the
final one by the master. Sometimes there is a single set of objections to the
response, but only in eleven cases is there a longer debate with several
interventions by respondent and others.

Mentions of the respondent makes us feel we are close to the live

disputation. Thus: 12

VI.4 quamvis hoc negaret respondens (F 70vA; P 81rB; problem not

extant in B)

In one case we can see that it was he who delivered the solution in the
initial corpus sophismatis:

V.3 Ad illud dicitur [+ quandoque V], sicut respondebat respondens,

quod prima est duplex ex eo quod hoc signum 'omnis' potest
distribuere ilIum terminum 'phoenix' pro phoenicibus actu vel
potentia (F 69rA; V 7rA; problem not extant in B)

This is said at the beginning of the magisterial determination, which in this

problem follows immediately on the rationes with no intermediate
intervention by the respondent. The reference is not to anything the
respondent said within the problem, but nicely fits the end of the corpus
sophismatis:

Solutio. Dicendum quod haec est duplex 'omnis phoenix' eo quod

hoc signum 'omnis' potest distribuere istum terminum 'phoenix' pro
phoenicibus in actu vel pro phoenicibus in potentia. (V 7rA, corpus
sophismatis omitted by B and F).

In at least one case the respondent was a bachelor. In sophisma I we

find the phrases

1.1 Concessit bachelarius [b. om. B] (F 63rA, B 89rA)

12When nothing else is indicated, I quote in the form given by F, and do not record
insignificant variants in other MSS. ·V.I' means 'Sophisma V, problem I'.
 BOETHlUS DE DACIA ET AL. 49

I.4 Ad rationes quae sunt contra rationem bachalarii per praedicta

patet solutio. (F 64rB, om. B).

Sophisma I is by Boethius of Dacia, and Danes belonged to the English

Nation, in whose 1252 statutes one of the requirements to be fulfilled by
the bachelor who wants to become a master is:

det fidem quod per duos annos diligenter disputationes magistrorum

in studio solempni frequentaverit et per idem tempus de
sophismatibus in scolis requisitus responderit. 13

So, at the very least, the sophisma provides evidence that the 1252 statute
was not a totally dead letter to members of the English Nation some twenty
years later. However, there is no reason to think customs were
significantly different in the other nations. Probably, then, all the
respondents mentioned in our collections were bachelors.

On several occasions it is noted that the respondent failed to reply to

the first set of objections against his position and solutions. Thus

I.2 Ad hoc non fuit responsum [f.r.: respondebatur B] (F 63v A, B

9OvA)

1.4 Ad haec non respondebatur (F 64rA, om. B)

V.l Quaedam etiam arguta fuerunt ad quae tamen nihil fuit

responsum, et ideo non est conandum [curandum B] (F 67vB; B
lOOrA; om. V.)

X.l Ad hoc non fuit responsum (F 76rA; sophisma not in B)

XVII.l Contra illud nihil dicebatur (B 94vB; sophisma not in F)

XX.l Ad haec non respondebatur (B 106rB; sophisma not in F)

On another occasion he managed to deal with some of the arguments but

failed to answer them all:

IX.8 Ad ista respondebatur per ordinem et primo ad rationes quae

erant contra positionem quaestionis. Ad primum [...]. Ad secundum
[...] Ad rationes sequentes non respondebatur. (F 74vB)

In IX we see two masters - Peter of Auvergne and Boethius of

Denmark - collaborating, though it is not quite clear what role each man
had. 14 Perhaps Peter was in charge of the first seven and Boethius of the
last three questions. At the end of problem 7 we read:

13H.Denifie & E. Chatelain, Chartuiarium Universitatis Parisiensis I, Paris: Delalain

18.89; see document NO 201, p. 228.
14Por an attempt to interpret the data, see Roos, "Ein unbekanntes Sophisma des
Boetius de Dacia" (see footnote 2).
50 STEN EBBESEN

IX.7 Haec sufficiant de ista oratione 'Syllogizantem ponendum est

terminos' disputata a magistro de Alvernia in qua proponebantur tria
problemata principal iter inquirenda, et de primo alia tria fuerunt
quaesita.

But at the end of the very last problem we find:

IX.IO: Et hoc de tertio problem ate et per con sequens de toto

sophismate sufficiant. Magister Bonus [error for Boethius] Dacus
composuit et determinavit.

It is unclear whether any significance can be attached to the fact that Peter
is said to have "disputed" and Boethius to have "determined". The
formulas used to indicate authorship are these:

I, II, III. Explicit sophisma Boethii de Dacia / magistri Petri de

Alvernia. (F)

V. A magistro Petro de Alvemia. (V)

VII. Per magistrum Petrum de Alvemia. (F)

IV. Determinatum est istud sophisma a magistro Petro de Alvernia.

(F).
XII. Hoc determinatum est a magistro Petro de Alvemia. (F)
XI. Magister Nicholaus de Normandia determinavit. (F)

VIII. Et hoc sophisma [s.: sufficienter F] magister Petrus de Alvemia

disputavit. (F)
XIII. Hoc est sophisma disputatum a magistro Petro. (F)

So, in IV, XI, and XII Peter is said to have determined, in VIII and XIII to
have disputed. But in each of these cases we have only one manuscript to
provide the information. In sophisma VI there are two. And they disagree!

VI. Explicit sophisma diputatum a magistro Petro de Alvemia. (F)

VI. Explicit sophisma determinatum a magistro Petro de Alvernia.

(P)
In N° V we learn that only some of the questions raised at the
beginning of the session were actually discussed:

Circa istam orationem multa fuerunt quaesita, sed solum prosecuta

fuerunt quae circa veritatem quaerebantur, circa quam quaerebantur
tria [Circa illud sophisma plura possunt quaeri, et ea quae circa
veritatem possunt proponi sunt executa V] (F 67vB; B 99vB; V 7rA)

We may also be informed that some part of the oral material has been left
out in the written version. Thus

VI.I Omissis illis quae hinc et inde solvendo [syllogizando F] et

replicando [et r. om. P] dicta fuerunt [+ causa brevitatis P],
consideranda sunt quattuor (F 69rA; B 95vA; P 79rA)
 BOETHIUS DE DACIA ET AL. 51

and

VI.2 Quae autem dicta fuerunt respondendo ad ista et replicationes

[replicando P] ad ista dimittantur [obmittantur P] propter brevitatem
et quorundam inutilitatem [inut.: in universali F] (F 69vB; P 80rA;
the problem does not occur in B.)

Similarly

XVlI.2 Ad hoc dicebatur sicut ad primum in secunda responsione, et

modo simili contra arguitur. (B 95rA)

One strange formulation leaves it unclear whether editorial intervention or

considerations of time during the oral debate occasioned the omission of
the respondent's position and further debate:

IX.9 Ad ista non facit responsum, quia oportet multum repetere, et

ideo statim ad quaestionem dicendum .... (F 75rB; perhaps one
should read 'non fuit responsum, quia oport<er>et').

The editor, however, does not just omit or revise materials derived
from the oral disputation. He introduces quite new items. Thus, in one
case he expressly says that a question was not debated, yet proceeds to
state his opinion on the matter:

IX.7 De ultimo non fuit disputatum, sed est intelligendum, quantum

mihi videtur, quod ... (F 74rB)

Other editorial matter includes a couple of cross-references:

ILl ut dictum fuit in illo sophismate ALBUM POTEST ESSE

NIGRUM. (F 64vA; B 86rB).

V.2: ut visum est [declaratum fuit B] in praecedente sophismate,

scilicet ALBUM POTEST ESSE NIGRUM. [scilicet ... nigrum om.
B] (F68rB,B 101vA)

V.2 ut prius fuit declaratum in illo [+ praecedente B) sophismate

[om. B] ALBUM POTEST ESSE NIGRUM. (F 68vB, B l02vA)

XL2 de nomine adiectivo, de cuius significato visum est in illo

sophismate ALBUM POTEST ESSE NIGRUM. (F 73vB; sophisma
not in B)

In a couple of cases an argument ad principa/e is verbatim identical to

one known from elsewhere. IS In some instances this may merely mean

ISThus XXII: "Praeterea, ista propositio universalis est vera contra quam non contingit
instare; haec est huiusrnodi; ergo etc." = MS elm 14522: 2SvA: "Item, universalis est
vera contra quam non est dare instantiam, sed haec est huiusmodi, ergo etc." An even
more significant case is the third ratio quod non in VIII.3, which is fairly long, and
identical to one in MSS clm 14522: 40rB and Paris BN lat. 16135: 39rA. Also within
52 SJEN EBBESEN

that the person charged with this part of the debate had prepared himself at
home, gathering arguments from published sources. But in sophisma
XVIII - the anonymous 'Si tantum pater est, non tantum pater est' -
something stranger is going on. In the detennination there are solutions of
three arguments in oppositum, but only the first of them occurs in the
initial recital of pros and cons. It could be simply loss of text, but
comparison with the homonymous sophisma in MS W = Worcester Cath.
Q.13, ff. 53r-v, suggests otherwise. The W version starts by announcing
five problems, but only N° 1 and 3 receive any treatment at all. In both
cases we get the initial disputation only, not the magisterial detennination,
which was presumably to follow after the disputation of all the
problems. 16 W's first problem is roughly the same as XVIIIA, but there is
no close relationship between them. By contrast, W's problem 3 is closely
related to XVIII. 1. The following table gives the essentials:

Worcester Bruges XVIlI.l

Disputation: Disputation:
Quod sic
Quod non 1 Quod non 1
Determination:
A. Statement of problem
B. Master's opinion
C. Ad rationes:
Against Quod non 1 Against Quod non 1
Against Quod non 2:
Quod non 2 Summary
Against Quod non 2 Refutation
Against Quod non 3:
Quod non 3 Summary
Refutation

The left and the right column items in the same row are textually
identical save for little details. The order of priority is uncertain, but I
suspect that W's arrangement is the original one, whereas XVIII. 1 is an
incomplete adaptation of the (English ?) sophisma to another fonnat in
which the detennination intervenes between arguments quod non and their
refutation. The adaptation has only been completed for argument 1; in the
cases of 2-3 the arguments are encased in the refutation, ostensibly as a
summary of what was said before the determination, in reality just
preserving the original arrangement. W's omission of the refutation of 3 is
probably an accident. The sophisma is the last of the collection, and

our collection we find recycling of materials. With some verbal changes, the same
arguments are used in VI.2 and XII.2, both by Peter of Auvergne; but in VI.2 they
both occur as rationes quod sic, in XII.2 the first one is a ratio quod sic, the other one
is used in the determination. By some inexplicable mistake a ratio quod non that
belongs in VIII.3 is also inserted, with some minor variants, in VIII.2, with which it
has nothing to do.
161 suppose the intended structure was much like that of 'Quod incipit esse desinit non
esse', which I have published in Cahiers de l'lnstitut du Moyen-Age Grec et Latin 59,
1989, pp. 148-56.
 BOETHlUS DE DACIA ET AL. 53

obviously mutilated. Perhaps W was copied from an exemplar which was

physically mutilated, having lost the leaf or leaves containing the rest of 'Si
tantum pater est .. .' and possibly more sophismata as well.

If XVrn.l is the younger version of the problem, it surely is no report

of an oral debate in which different persons proffered the principal
arguments and refuted them. If XVIII. 1 is the older version of the
question, W is hardly a report of an oral debate, though it is just possible
to imagine that one person did the initial disputation, reciting (reading
aloud?) arguments both pro and con, all of them stolen from elsewhere,
while the determination was left for another (the master). In any event,
there is a high degree of literary dependence in at least one of the two
versions.

Sophisma XVIII is one of those that occur only in the Bruges

manuscript. I have a strong suspicion that in some cases of major
deviations between Band F in the shared sophismata, the Florence text
represents the original form. This squares beautifully with the theory that
in XVIII. 1 the Bruges text is a revision of a text known from another
written source. However, life is not that simple, for in problem 3 of the
shared sophisma VIII (by Peter of Auvergne), B's form of a long
argument known also from another is closer to that source, which is
probably earlier than VIII. So the Florence collection may also have been
subjected to an editor's intervention at some date later than the last common
ancestor of Band F.

The "other source", from which VIII.3 seems to have lifted an

argument, is 'Tantum unum est' in MSS Munich clm 14522: 37vA-43rB
and Paris BN 16135: 37vA-42rB.1 7 This sophisma is precious by
referring to an early master by name, a most unusual phenomenon. The
text reads a little differently in the two MSS. In the Munich version (f.
42rB) it runs:

Quidam autem multum famosus magister de Sicca Villa dixit quod

<the proposition 'tantum unum est'> erat falsa per se, vera autem per
accidens.

The Paris version (f. 40rB-v A) is more ornate:

Quidam autem maximus et famosior tempore nostro, videlicet

magister de Arida Patria, dixit quod erat falsa per se, vera autem per
accidens.

This may help date the sophisma. De Libera in his first description of
the sophismata of MS Paris 16135 suggested a date in the 1270s for the
collectio secunda of that MS, the one in which the above text is found. IS

171 owe my awareness that elm 14522 contains sophismata to Andrea Tabarroni. After
acquiring a microfilm 1 discovered that its collection of sophismata overlaps with Paris
16135.
18A. De Libera, "La litterature des Sophismata dans la tradition terministe parisienne de
la seconde moitie du XIIIe siecle", in The Editing of Theological and Philosophical
54 STEN EBBESEN

However, the master referred to must be John Dry ton, alias John of
Secheville, an Englishman who seems to have started as master of arts in
Paris in the 1240s, was rector there in 1256, and probably moved to
England in the late 50s.1 9 The way he is spoken of suggests a date fairly
soon after his departure. So at least for this sophisma the 1260s seem
indicated rather than the 1270s. But we have no guarantee the collection is
homogeneous. After all, it overlaps with that of clm 14522 without sharing
all items with it. This is a quite general phenomenon: I know of no two
MSS with exactly the same selection of the big 13th-century logical
sophismata-cum-problematibus texts, but we have several cases of
overlap.

Moreover, Dry ton's opinion is debated, though ascribed only to

'quidam', in the 'Tantum unum est' of the collectio prima of the same Paris
MS (f. 20vA). De Libera dated the first collection before, but close to
1250;20 the author's acquaintance with Dry ton's view favours a date after
1250 or very shortly before. 21

Dry ton's thesis continued to be debated into the 1270s when Peter of
Auvergne composed his sophisma VIII, but Peter also just calls its author
'quidam', though the verbal coincidences mentioned earlier make it
probable that Peter knew the 'Tantum unum est' of Paris 16135's collectio
secunda. The Dry ton thesis that 'Tantum unum est' is false per se and true
per accidens may be the ultimate reason why Peter of Auvergne in XIL2
asks whether some proposition may be true per se and false per accidens or
vice versa (in fact, one of the arguments quod sic is derived from
Dryton 22 ), and in VI.2 whether what is per se incompatible with
something may per accidens belong to it. These questions were important,
for as I have shown elsewhere,23 there was a fairly strong tendency in
about the 1270s to introduce two independent sorts of truth and falsity for
propositions, per se and per accidens.

The sophismata of the Florence and Bruges MSS do not evince a very
great interest in syncategoremata on their authors' part; the semantics of
categorematic terms obviously interest them much more. But this means

Texts from the Middle Ages, ed. M. Asztalos, Acta Universitatis Stockholmiensis,
Studia Latina Stockholmiensia XXX, Stockholm: Almqvist & Wiksel\ 1986, p. 217.
19See P.O. Lewry, "The Liber Sex Principiorum", in Gilbert de PoWers et ses
contemporains, ed. J. Jolivet and A. de Libera, History of Logic 5, Napoli: Bibliopolis
1987, pp. 250-78, at pp. 252 ff.
20De Libera, loc.cit.
21 Also of relevance for the date of collectio prima is the relationship to that in elm
14522; thus the latter's 'Si tantum pater est non tantum pater est' (ff. IOvB-18vB)
appears to be an enlarged edition of the former's (16135: 25vB- 28rB), but this requires
further study.
22XII.2: "Praeterea, Commentator sexto Physicorum dicit quod ilia propositio 'omne
quod movetur, contingit velocius et tardius moveri' vera est per se et falsa per accidens;
quare aJiqua propositio est vera per se et falsa per accidens." Cf. elm 14522: 42rB, Paris
BN lat. 16135: 40vA. Cf. also Paris BN lat. 16135: 20vA (collectio prima).
23S. Ebbesen, "Concrete Accidental Terms: Late Thirteenth-Century Debates about
Problems Relating to such Terms as 'album"', in Meaning and Inference in Medieval
Philosophy, ed. N.Kretzmann, Dordrecht: Kluwer 1989, pp. 107-74.
 BOITH/US DE DACIA IT AL. 55

that these texts contain plenty of material for the understanding of a wide
array of philosophical problems discussed in the late 13th century.
Unfortunately these sophismata make unusually great demands of their
editors, for the text often rests on only one bad MS, and when there are
more, they represent different redactions. I doubt whether it will ever be
possible to explain the transmission in detail, but it seems as if patient
study and comparison with other collections of sophismata may teach us a
few new things about how medieval scholars used their predecessors'
work.

In my experience, late 13th-century question commentaries tend on

closer scrutiny to reveal themselves as literary products, heavily dependent
on earlier written sources. 'Sed tu dices' means 'One might object' and
does not refer to a disputant at a definite time and place. Whole questions
may be lifted out of their original context to be re-used in a new collection.
Sometimes the borrowed materials are taken over lock stock and barrel,
sometimes they are revised. It starts to look as if the situation may be rather
similar with the sophismata. I cling to the belief that such phrases as •Ad
hoc nihil dicebatur' refer to what did or did not happen during an oral
debate, but we must reckon with the possibility that some problems or
even whole sophismata were first conceived in the written medium. We
must also reckon with complex developments of the type: 1. oral debate ->
2. strongly revised written version, by determining master or other witness
to the oral debate -> 3. revised edition, by other person. This is what I
suspect happened to Boethius de Dacia's 'Omnis homo de necessitate est
animal' (Sophisma I), in which B's version represents stage 3. But what
was the purpose of producing the B version of this sophisma, and of
gathering that particular collection of sophismata? Are we witnessing a
master who produces his own text-book by revising older works? Did he
even intend to read it aloud in class as a substitute for live discussion?

We still have a lot to learn about the interaction between the oral and
the literary in the faculty of Arts of the 13th century.

University o/Copenhagen
56 srEN EBBESEN

Appendix
Summary descriptions of mss Band F, and
survey of their sophismata

B = Brugge, Stedelijke Openbare Bibliotheek, 509

Provenance: Abbatia Dunensis
Early 14th c., parchment, mm. 272x192, ff. 2 frgm. + I + 107, two
columns a page.
Contents:
Frgm. I Anon., Grammatica
Frgm. 2 Anon., Schemata ad philosophiam pertinentia
1-30v Boethius de Dacia, Quaest. Arist. Top. (ed. in CPhD vi.1)
31-58 Anon., Quaest. Arist. APr.
59-75 Anon., Quaest. Arist. APo.
76-107 <Boethius de Dacia, Petrus de Alvernia et al.>, Sophismata
F =Firenze, Medicea Laurenziana, S. Croce 12 sin., 3
Earlier possessors: Magister lacopinus Cagnelli (78v); S.Croce [.shelf-
mark 302 (Ir)]
Early 14th c., parchment, mm. 325x220, ff. 78, two columns per page.
Contents:
IrA-3vB Petrus de Alvernia, Quaest. in Porph. Intr.
3vB-7vA Petrus de Alvernia + Anon., Quaest. in Arist. Cat. (ed. in
CIMAGL 55-5624 )
7vA-8vA Petrus de Alvernia, Quaestiones de universalibus
8vA-IlrB Anon. + Petrus de Alvernia, Quaest. in Arist. Int.
IlvA-14vA Anon., Quaest. VI Principia
15r-27vA < Boethius de Dacia >, Quaest. in Arist. Top. (ed. in CPhD
vi.})
27v A-37v A 'Petrus de Hibernia', Quaest. in Arist. APo.
39rA-49vB 'Petrus de Hibernia', Quaest. in Arist. SE (ed. in CPhD vii)
50rA-61rB Anonymus, Quaest. in Arist. APr.
61 rB-61 vB Anonymus, Tractatus de generatione syllogismorum
63rA-78rB Boethii, Petri de Alvernia, Nicolai de Normandia, Sophismata
78v Anon., note on logic
78v Radulphus Brito, Quaest. Porph.lntr. 3 (or Quaest APo. 1.4; the two
are identical). Incomplete.
Sophisma I : Omnis homo de necessitate est animal
Author: Boethius de Dacia
MSS: FB.
Edition: Grabmann 194025 (F 1-4), Roos 196226 (B 1-3)

24Cahiers de l'institut du Moyen-Age Cree et Latin.

25M. Grabmann, Die Sophismatalitteratur (see n. I to main text). Text on pp. 77-95.
26H. Roos, "Das Sophisma des Boetius von Dacien" (see n. 2 to main text), pp. 178-
97.
 BOEIHlUS DE DACIA EI AL. APPENDIX 57

Translation: Kretzmann & Stump 198827 (F 1-4)

Problemata:
B I. U. necessitas compositionis possit esse sine necessitate terminorum
(Not in F)
B2. U. sequatur 'omnis homo currit, ergo albus homo currit' (Not in F)
B3. U. sequatur 'omnis homo de necessitate est animal, album est homo,
ergo album de necessitate est animal', i.e.
U. ex maiori de necessario et minori de inesse ut nunc sequatur
conclusio de necessario (Not in F)

Fl = B4. U. haec sit vera 'omnis homo de necessitate est animal' nullo
homine existente.
F2 = B5. U. rebus corruptis oporteat corrumpi scientiam habitam de illis
rebus.
F3 = B6. U. rebus corruptis oporteat terminum cadere a suo significato.
F 4 = B7. U. natura generis existens in specie sit aliquid in actu praeter
ultimam differentiam speciei.
SophismalI: Homo est species
Author: Petrus de Alvemia
MSS:FB.
Problemata:
1. U. illud quod significatur nomine hominis sit quiditas sola vel habens
quiditatem.
2. U. illud quod significatur nomine speciei sit in intellectu sicut in
subiecto (sc. ratio intelligendi) aut aliquid existens in eo de quo dicimus
quod est species.
3. U. haec sit vera 'homo est species'.
4. U. haec sit vera 'aliquis homo est species'.
Sophism a Ill: Album potest esse nigrum
Author: Petrus de Alvemia.
MSS: FB + C (quu. 1-2 and part of 3 only).
Edition: Ebbesen 1988 (only quo 1)28
Problemata:
1. U. iste terminus 'album' significet tantum formam vel aggregatum ex
subiecto et forma
2. U. distinctio bona sit
3. U. sequatur 'quod potest esse album potest esse nigrum, ergo album
potest esse nigrum'
4. U. haec sit vera 'album potest esse nigrum'
Sophisma IV: Animal est omnis homo
Author: Petrus de Alvemia
MSS:FB.
Problemata:

27N. Kretzmann and E. Stump, Logic and the Philosophy of Language. The Cambridge
Translations of Medieval Philosophical Texts I, Cambridge: Cambridge University
Press, 1988.
28S. Ebbesen, "Concrete Accidental Tenus" (see n. 23 to main text). The text edition is
on pp. 182-4.
58 srEN EBBESEN

1. U. hoc signum 'omnis' potest teneri collective vel distributive

2. U. istae propositiones 'non animal est omnis homo' et 'nullum animal
est omnis homo' sint una negatio vel duae negationes et u. una
aequipollet alteri.
3. U. sequatur 'animal est omnis homo, ergo aliquod animal est omnis
homo'
4. U. haec sit vera 'animal est omnis homo'
Sophisma V: Omnis phoenix est
Author: Petrus de Alvemia (1-3, author of 4 unknown)
MSS:FB+V.
Problemata:
1. U. determinatio immediate adiuncta suo determinabili restringit ipsum
2. U. terminus supponens alicui verbo cuiuscumque temporis restringatur
ad supponendum secundum exigentiam eius vel amplietur
3. U. haec sit vera 'omnis phoenix est' (om. B)
4. U. signum adveniens termino alicui substantiali distribuat ipsum
indifferenter pro omnibus tam essentialibus quam accidentalibus (om.
FV)
Sophisma VI: Nullus homo de necessitate est animal
Author: Petrus de Alvemia
MSS:FB+P.
Problemata:
1. U. maiore existente vera de necessario et minori de contingenti sequitur
conclusio de inesse
2. (2.1) U. illud quod repugnat alicui per se possit inesse per accidens
eidem (om. B)
3. (2.2) U. haec propositio 'materia potest esse homo' sit vera de virtute
sermonis (om. B.)
4. (2.3) U. homo possit esse asinus, et u. haec sit vera 'nullus homo de
necessitate est animal' (om. B.)
Sophisma VII: Socrates desinit esse non desinendo esse
Author: Petrus de Alvemia
MSS: FB.
Edition: Ebbesen 1989.29
Problemata:
1. (1.1.1) Utrum sit dare ultimum instans vitae Socratis
2. (1.1.2) U. sit dare paenultimum instans vitae Socratis
3. (1.2) De significato istorum verborum 'incipit', 'desinit' (u. significent
motum)
4. (1.3) U. ista verba 'incipit' et 'desinit' cuicumque adiungantur habent
eandem expositionem.
Sophisma VIII: Tantum unum est
Author: Petrus de Alvemia (F version)
MSS: FB.
Problemata:

29S.Ebbesen, "Three 13th-century Sophismata" (see n. 8 to main text), pp. 121-80.

 BOEJ'HIUS DE DACIA EJ' AL. APPENDIX 59

FI. Dato quod haec dictio 'tantum' excludit (includit F) diversum ab

incluso in forma substantiali et accidentali, u. hoc sit per rationem unam
vel per diversas. (F only)
F2. U. sequatur 'tantum unum est, ergo non multa sunt'
F3. U. haec sit vera 'tantum unum est' (= B5)

Bl (1.1). U. terminus concretus significet formam vel totum aggregatum

ex forma et subiecto vel materia
B2 (1.2). U.li 'tantum' adveniens huic quod dico 'unum' possit excludere
diversum ab ipso ratione formae vel ratione subiecti
B3 (2.1). U. 'unum' addit aliquid supra 'ens'
B4. U. dictio exclusiva addita 'uni' excludat multa (= cf. F 2, but not
identical)
B5. U. haec sit vera 'tantum unum est' (= F 3)
Sophisma IX: Syllogizantem ponendum est terminos
Author: Petrus de Alvemia (1-7) & Boethius de Dacia (8-10)
MS:F.
Problemata:
1 (1.1). U. hoc verbum 'est' possit impersonari
2 (1.2). U. hoc verbum 'est' impersonatum retentum impersonaliter
construatur cum obliquo a parte ante et intransitive
3 (1.3). U. hoc verbum 'est' convenienter construatur cum hoc quod dico
'syllogizantem' a parte post
4 (2.1). U. participium significet per modum substantiae
5 (2.2). U. adiectivum possit habere rationem supponendi
6 (3.1). U. gerundia sint nomina vel verba
7 (3.2). De constructione gerundiorum a parte ante cum accusativo (NB
'non fuit disputatum ')
8 (4.1). U. verbum primae personae vel secundae dat intelligere aliquem
nominativum
9 (4.2). U. nominativus quem dat intelligere sit nominativus nominis vel
pronominis
10 (4.3). U. verbum personae primae sicut 'lego' magis possit facere
orationem perfectam quam verbum tertiae personae.
Sophisma X: Albus pedem est animalia
Author: Anon.
MS:F.
Problemata:
1. De ordinatione huius quod est 'albus' ad hoc quod est 'est'
2. De ordinatione huius quod est 'albus' ad hoc quod est 'pedem'
3. De ordinatine huius appositi quod est 'animalia' u. cum supposito in
singulari numero (u. haec sit congrua 'homo est animalia')

Sophisma Xl: Albus musicus est

Author: Nicolaus de Normandia
MS:F.
60 srEN EBBESEN

Edition: Ebbesen 1988.30

Problemata:
1. Utrum essent eadem principia congruitatis et perfectionis
2. De ordinatione huius quod est 'albus' ad hoc quod est 'musicus'
3. Utrum modus permanentis sufficiat ad hoc quod aliquid habeat rntionem
supponendi respectu verbi
Sophisma XII: Omnis homo est omnis homo
Author: Petrus de Alvemia
MS:F.
Problemata:
1. U. haec sit vern 'ornnis homo est omnis homo'
2. U. aliqua propositio potest esse vera per se et falsa per accidens vel
econverso
Sophisma XIII: Omnis homo est
Author: Petrus <de Alvemia>
MS:F.
Problema: U. ille terminus 'homo' restringatur ad praesens quando
supponit verba de praesenti
Sophisma XIV: Homo non est, animal fuit in archa Noae,
album currit pro albo praeterito et nunc non existente
Author: Anonyrnous
MS:B.
Problema: De veritate huius 'homo non est'

Sophisma XV: ? <sophisma-sentence unstated>

Author: Anonyrnous
MS:B.
Problema: U. significatum individui addat aliquid reale supra significatum
termini specifici
Sophisma XVI: Animal est species specialissima
Author: Anonymous
MS:B.
Problemata:
1. U. animal sit nomen unius rei vel plurium
2. U. quinque praedicabilia sint in rebus vel in anima tamquam in subiecto.
3. U. ista 'animal est species' possit verificari per accidens, sc. per istam
'homo est species'
Sophisma XVII: Homo est animal album
Author: Anonyrnous
MS:B.
Problemata:
1. U. ad veritatem propositionis affmnativae oporteat praedicatum in recto
esse de significato subiecti

30S. Ebbesen, "A Grammatical Sophisma by Nicholas of Normandy ALBUS

MUSICUS EST", Cahiers de l'Institut du Moyen·Age Grec et Latin 56, 1988, pp.
103-16.
 BOEI'HIUS DE DACIA EI' AL. APPENDIX 61

2. U. terminus substantiaIis concretus una significatione et impositione

significat materiam et formam vel formam tantum
3. U. animal possit de homine praedicari et universaliter genus de specie
Sophisma XVIII: Si lantum pater est non tantum pater est
i\uthor:~onytnous
MS:B.
Problemata:
1. U. haec dictio 'tantum' possit esse determinatio praedicati et subiecti
indifferenter.
2. De distinctione quae solet hic assignari sc. quod Ii 'tantum' potest facere
exclusionem ab hoc quod dico 'pater' gratia formae substantialis vel
accidentalis
3. U. sequitur 'tantum pater, ergo pater'
4. U. sequitur 'tantum pater est, non ergo filius est' ita quod dictio
exclusiva addita uni relatorum excludit reliquum
Sophisma XIX: Socrates non est homo
i\uthor: ~onymous
MS:B.
Problemata:
1. De distinctione assignata (de natura negationis)
2. De veritate huius 'Socrates non est homo' supponendo Socratem esse,
utrum aliquo modo possit verificari per hoc quod non esset aliquis
homo sicut Plato.
Sophisma XX: Omnis homo de necessitate est animal
i\uthor:~onytnous
MS:B.
Problemata:
1. U. terminus communis distribuatur pro suppositis essentialibus tantum
vel pro accidentalibus vel pro ipsis indifferenter (announced but not
actually found, see answer to quo 3, however)
2. U. res de se sit universalis
3. U. ex maiori de necessario et minori de inesse ut nunc sequatur
conclusio de necessario
Sophisma XXI: Sola necessaria necessario sunt vera
i\uthor:~onytnous
MS:B.
Problema: De distinctione, utrum sit bona. (i\ fragment).
Sophisma XXII: Omnis homo est unus solus homo
i\uthor: i\nonymous
MS:B.
Problemata:
1. U. terminus commmunis ex sua natura sive per se sine adiunctione
alterius sit determinatus ad aIiquod suppositum vel utrum sit
indifferens.
2. U. per se sit subiectum distributionis
3. U. haec sit vera 'omnis homo est unus solus homo'
62 STEN EBBESEN

List of manuscripts

Brugge
Stedelijke Openbare Bibliotheek, 509
Description: Ebbesen in this volume.

Firenze
Medicea Laurenziana, S. Croce 12 sin., 3
Description: Ebbesen in this volume.

Firenze
Biblioteca Nazionale Centrale
Cony. Soppr. E.I.252
Description: G. Pomaro in: CataLogo di manoscritti filosofici neUe
biblioteche italiane 3, Firenze: Olschki 1982, pp. 49-51.

Miinchen
Bayerische Staatsbibliothek
elm 14522
Description: B. Faes de Mottoni, Aegidii Romani Opera 1.1/5*, Firenze:
Olschki 1990, pp. 145-50.

Paris
Bibliotheque Nationale
lat. 16089
C. Lafleur, Quatre introductions a La philosophie au XIlIe siecle,
Montreal/paris : Institut d'Etudes MedievalesNrin 1988, pp. 17-39.

Paris
Bibliotheque Nationale
lat. 16160
Description: H.V. Shooner, Codices manuscripti operum Thomae de
Aquino 3, Montreal/paris: Presses de l'Universite de MontrealNrin 1985,
pp.314-5.

Paris
Bibliotheque Nationale
lat. 16135
Descriptions: (a) A. De Libera, "La litterature des Sophismata dans la
tradition terministe parisienne de la seconde moitie du xm e sieele", in The
Editing o/Theological and Philosophical Texts/rom the Middle Ages, ed.
M. Asztalos, Acta Universitatis Stockholmiensis, Studia Latina
Stockholmiensia XXX, Stockholm: Almqvist & Wiksell, 1986. (b) c.
Lafleur, Quatre introductions a la philosophie au XIIle siecle,
Montreal/Paris : Institut d'Etudes MedievalesNrin 1988, pp. 72-4.
 BOETHIUS DE DACIA ET AL. 63

Paris
Bibliotheque Nationale
N.aJ. 1374
Description: W. Senko, Repertorium Commentariorum Medii Aevi in
Aristotelem Latinorum quae in Bibliothecis Publicis Parisiis asservantur =
Opera Philosophorum Medii Aevi, Textus et Studia 5/2, Warszawa:
Akademia Teologii Katolickiej 1982, pp. 103-6.

Vaticano
Bibliotheca Apostolica Vaticana
Vat. lat. 3061

Vaticano
Bibliotheca Apostolica Vaticana
Vat. lat. 14718

Worcester
Cathedral Library
Q.13
Description: C. Lohr, "Aristotelica Britannica", Theologie und Philosophie
53, 1978, pp. 97-9.
 Orleans 266 and the Sophismata Collection: Master Joscelin
of Soissons and the infinite words in the early twelfth century
by C. H. Kneepkens

Introduction

In this article, I intend to discuss the opinions on term negation

(infinitatio) which are dealt with in an entry of a collection of early twelfth
century logical sophisms and in texts to which these sophisms are closely
related.

On folios 281 a - 289 b , the manuscript Orleans, Bibliotheque

municipale, 266 has preserved a collection of eighty sophismata and logical
problematic propositions. As Minio-Paluello pointed out in his introduction
to the edition of the Sententie of master Peter, the manuscript itself was
written by several hands in the middle or second half of the twelfth
century.! It was kept at Fleury, before it entered the library of Orleans, but
there is reasonable doubt about how long it was there. Luscombe suggests
a northern French origin for this collection. 2 Apart from the set of
sophismata, the Sententie of master Peter and a small collection of three
sophismata which is appended, all the texts of the codex, comprising
twenty-eight items, directly concern treatises on the logica vetus.3 They
are commentaries on or additions to commentaries on Aristotle's De
interpretatione, Porphyry's Isagoge, and Boethius' logical monographies.

One of the more characteristic features of the texts of this manuscript -

which rightly received the qualification "of high importance in the history
of medieval logic" ,4 but nevertheless remained unedited for the greater part
- is that in many of them there is frequent reference to opinions held by
other masters, who were mentioned by name. Unfortunately, they are for
the greater part denoted by means of their initials only. Helped by the
information already supplied by Geyer,5 I can give the following still
incomplete list: We meet a Robertus, magister Willelmus, magister G.,

1L. Minio-PaluelIo, Twelfth Century Logic. Texts and Studies, II. Abaelardiana inedita,
Roma: Edizioni di Storia et letteratura 1958, p. xlii.
2D. E. Luscombe, The School of Peter Abelard. The Influence of Abelard's Thought in
the Early Scholastic Period, London: Cambridge University Press 1970, p. 73.
3Up to the first decades of the twelfth century, the Latin West had at its disposal from
Aristotle's logical works, the so-called Organon, only the Categories and De
interpretatione. These works, Porphyry's Isagoge in the translation of Boethius,
Boethius' commentaries on these three works and his logical monographs were the
works of what the Mediaevals used to call the logica vetus; cf. L. M. de Rijk, Logica
Modernorum. A Contribution to the History of Early Terminist Logic, Vol. I. On the
Twelfth Century Theories of Fallacy, Assen: Van Gorcum 1962, pp. 14-15.
4Cf. Luscombe, The School of Peter Abelard, p. 72.
5B. Geyer, Peter Abaelards philosophische Schriften, II. Die Logica "Nostrorum
petitioni sociorum". Die Clossen zu Porphyrius, 2., durchgesehene und veranderte
Auflage, Munster i. W.: Aschendorff 1933/1973, p. 595.

64
 MASTER JOSCEUN OF SOISSONS AND INFINITE WORDS 65

who cannot always be identified with master W., magister Goslenus (163 a;
255 b), magister W. (several times), magister Gui., magister Henricus,
magister Ros., in all probability Roscellinus, a dialogue between a master
W. (p. 201: On Boethius' Topica) and a master Galer<annus>, a master
Gauterus, a master Guido, the Camoti, who say that a clause (oratio)
signifies a single (simplex) concept, which can be founded on the soul, a
master W. who is identified as Walterus some lines farther on, a master
Anselmus and the magister noster.

Although it is often difficult and not without risk to try to identify the
masters mentioned in this way, it will be possible, this time, to assign the
references to a master W. in the collection of notes on Boethius' De
dijferentiis topicis to William of Champeaux, since to this master W. the
opinion is attributed that a proposition has a double sense, namely a
grammatical and a dialectical one. So 'Socrates est albus' analyzed
according to the grammarian means: 'Socrates est alba res", whereas the
dialectical sense is: 'Albedo inheret Socrati' (p. 213 b ).6 This master
William and master Goslenus, who can be identified as master Joscelin,
the later bishop of Soissons,7 are the protagonists in an early twelfth-

6MS, p. 213 b: "M. tamen W. dicit unamquamque propositionem et questionem habere

duos sensus, unum gramaticum et alium dialecticum. Verbi gratia. 'Socrates est albus'
habet hunc gramaticum: Socrates est alba res, et hunc dialecticum: albedo inheret
Socrati. Et iterum hec questio 'utrum Socrates est homo uel non est homo' habet ilium
gramaticum, quem proprie generat, et hunc dialecticum utrum predicatum inhereat
subiecto; quem hic dicit Boetius esse communem omnibus predicatiuis questionibus."
The glossing master objects in the following way: "Cui sententie sic opponimus. Si
omnis propositio habet gramaticum et dialecticum sensum, similiter dicam quod omnis
propositio habet quatuor. Habet enim et hos duos et ethicum et physicum. Ethica enim
et phisica similiter per se sunt facultates sicut dialectica et gramatica. Item. Si omnis
propositio habet dialecticum et gramaticum sensum, tunc ista 'Socrates est albedo'
habet gramaticum istum: Socrates est ilIe color albedo, dialecticum uero hunc: albedo
inheret Socrati. Qui sensus cum uerus sit, quare propter ilium non est uera propositio ?
Ad quod dicunt nullam propositionem esse ueram nisi uerum habeat gramaticum
sensum. Item. Si omnis propositio generat duos sensus, tunc omnis propositio est
multiplex. Ad quod dicunt quod multiplex propositio non dicitur nisi propter multos
dialecticos tantum uel multos gramaticos sensus. Sed secundum hoc 'Socrates est
albus' est multiplex propositio, quia generat multos sensus gramaticos. Generat enim
istum: Socrates est res alba, et istum: albedo inheret Socrati, qui est gramaticus, quia
est gramaticus istius propositionis 'albedo inheret Socrati', cuius dialecticus est iste:
inherentia inest albedini respectu Socratis."
Master William's position is finally rejected: "Nos uero dicimus nullam
propositionem nullamque questionem preter multiplices duos proprie sensus habere."
For the evidence that this view is peculiar to William of Champeaux and his followers,
see L. M. de Rijk, Logica Modernorum. A Contribution to the History of Early
Terminist Logic, Vol. II, I, The Origin and Early Development of the Theory of
Supposition, 2, Texts, Assen: Van Gorcum 1967, I, pp. 183 ff.
7For Joscelin of Bourges or of Soissons, see Luscombe, The School of Peter Abelard,
pp. 73-4, P.O. King, Peter Abailard and the Problem of Universals. A Dissertation
Presented to the Faculty of Princeton University in Candidacy for the Degree of Doctor
of Philosophy, 2 vols., 1982, pp. 187-8 and M. M. Tweedale, Logic(i): From the Late
Eleventh Century to the Time of Abelard, in A History of Twelfth-Century Western
Philosophy, ed. P. Dronke, Cambridge: Cambridge University Press 1988, pp. 196-
226, p. 222.
66 C.H.KNEEPKENS

century dispute about the semantics of the infinite terms, the subject of this
paper.

The collection itself, which is titled in the manuscript as

SOPHISMATA, looks like a casual and unstructured congeries of
sophisms. The leading motive, however, of putting these sophismata
together must have been their direct relationship with the other texts
preserved in the manuscript, for a glance at the commentaries shows that in
several of them these and similar sophisms have been used. For instance,
in the Suppietiones to the commentary on the Perihermeneias 8 in the
discussion on the futura contingentia we find the observation that at this
point a sophism of the type '''Socrates amabit" non est propositio' ought to
be made,9 whereas of course Boethius' De topicis differentiis is
responsible for the sophisms on Diogenes and his horns.1O
1. Term negation
(I) Term negation, that is, the negation of a term in a proposition,
which is itself affirmative, is in contradistinction to three
other types of negation:

(2) predicate denial, which negates the predicate term of the

subject and thus creates a negative proposition out of an
affIrmative one;

(3) sentential negation: the denial of the whole proposition;

(4) privation, which is expressed by means of a term with a

negative affix, e.g. 'irrational') I

The propositions 'Socrates is non-rational' (1) with a hyphen between the

negative particle 'non' and the adjective noun 'rational', 'Socrates is not
rational' (2), that is - in the Aristotelian system - the contradictory
proposition of 'Socrates is rational', or - in the Stoic tradition - 'not:
Socrates is wise' (3), and 'Socrates is irrational' (4) which has as its
contrary proposition' Socrates is rational', are instances of these types.

8MS, p. 257 b - 263 b.

9MS, p. 260 a: "Notandum est quod idem est propositionem esse necessariam et ueram
determinate. Hic solet fieri tale sophisma: 'Socrates amabit' non est propositio. Si
neque de contingenti neque de presenti neque futuro, tunc ilIud est. De contingenti non
est; de quo magis uidetur. Si est determinate uera, tunc ilIud est. Sed hoc est, si est
presentialiter uera. Tunc ilIud est, quia omnia presentia determinata sunt. [po 260b] Sed
hoc est. IlIa consequentia est determinanda, que dixit: si est determinate uera, non est de
contingenti, sic: si est uera determinate, idest presentiaIiter, tunc non de contingenti,
non sequitur. Sed si est uera determinate, idest ita quod non potest uerti in aIteram
partem, scilicet non est possibile Socratem non amaturum, bene sequitur. Item.
Notandum est quod hec propositio 'Socrates amabit' est uera: uera est determinate, quia
presentialiter habet ueritatem, quam non potest non habere dum habet."
IOBoethius, Anicii Manlii Severini Boetii De topicis dijferentiis Libri IV, ed. J.P.
Migne. Paris 1847 (repr. Tumhout 1979). vol. 64. 1173C-1216D, I, 118IA.
II For this distinction, see L. R. Hom, A Natural History of Negation, Chicago and
London: Chicago University Press 1989, pp. 17, 25-6; see also G. Englebretsen,
Logical Negation, Assen: Van Gorcum 1981, p. 4.
 MASTER JOSCEUN OF SOISSONS AND INFINITE WORDS 67

Two different opinions on the semantics of infinite terms are discussed

in a rather comprehensive nota in the sophismata collection at issue.
Besides, we find term negation frequently dealt with in several of the texts
of this manuscript, which is natural, since in the works of the logica vetus,
term negation is a topic frequently discussed. These discussions were
provoked by inter alia Aristotle's observations in the Perihermeneias that
compound terms like 'non-homo' are not "ordinary" nouns, but "infinite
nouns" and their verbal counterparts like 'non-currit' "infmite verbs",12 in
the discussion of the quantity of the proposition in the same treatise,13
Boethius' work on the categorical syllogisms14 and De divisione,15 and
his two commentaries on Aristotle's Perihermeneias.1 6
2. Boethius' theory oj term negation
Boethius' contribution is important for the mediaeval discussion since
he broached the question about the grammatical category and the semantic
status of the infinite noun and verb, and thus settled more or less the points
of departure of the mediaeval discussion. In his minor commentary on the
Perihermeneias,11 he first tried to establish the respective positions of the
infinite terms among the linguistic entities. The infinite noun is said not to
be a sentence (oratio), since a sentence ought to consist of verbs and/or
nouns, which is not the case with infinite words, since an infinite noun
consists of a negative particle and one noun or one verb. A negative
proposition (a negatio) is also out of the question, since every negative
proposition must have a truth-value, whereas an infinite noun does not
signify truth nor falsity. On the other hand, semantic considerations also
exclude it from being a noun, for every noun has as its meaning function to
signify something definite. So when we say 'man', we do not mean any
substance, but a rational and mortal substance. One, however, who says
'non-man', removes 'man', but does not determine or say definitely what
he wants to show by means of that "meaning". For that which is not a
man, can be a horse, a dog, a stone etc. Boethius concluded that an infinite
noun signifies as many things as are removed from the definition of man:
the complementary set of the set labelled 'man'.

12AristotIe, De interpretatione, ch. 2 & 3, 16a30-32 and 16b12-16, tr. Boethii,

Aristoteles Latinus, II, 1-2: De interpretatione vel Periermenias. Translatio Boethii
Specimina Translationum Recentiorum, ed. L. Minio-Paluello, Bruges-Paris: Desclee -
De Brouwer 1965, pp. 6-7.
I3AristotIe, De interpretatione, 10, 19b8 ff. (tr. Boethii, ed. Minio-Paluello, p. 18).
14Boethius, Anicii Manlii Severini Boetii De syllogismis categoricis, ed. J. P. Migne,
Paris 1847 (herdr. Turnhou! 1979), vol. 64, 794C-830, I, PL LXIV, 795C ff.: "Non
homo autem quod definitum est perimit" (795C).
15Boethius, De divisione, traduzione, introduzione commento L. Pozzi, Padova: Liviana
1969, chs. 24 and 32.
16Boethius, Commentarii Anicii Manlii Severini Boetii Commentarii in Librum
Aristotelis PERI HERMEENEIAS. Recensuit Carolus Meiser, 2 vol., Leipzig 1877-
80. Aristole's Prior Analytics, in which text (book I, ch. 46) the infinite words are also
dealt with, were not yet rediscovered at the time that this discussion runs; for a
discussion of this place, see A. N. Prior, The Doctrine of Propositions and Terms, ed.
by P. T. Geach and A. J. P. Kenny, London: Duckworth 1976, pp. 44 ff.
17Ed. Meiser, p. 51, 20 ff.: "hoc enim quod dicimus non homo quid sit ambigitur."
68 C.H.KNEEPKENS

In his second and larger commentary Boethius added the observation

that an infinite noun can be employed for existing and not-existing things
equally well. It may, for instance, be used as a predicate term in a
proposition which has the Scylla as its subject. When someone says of the
Scylla that it is a "non-man", this term, viz. 'non-man', signifies, as
Boethius argued, something which does not exist in reality. On the other
hand, when it is said of an existing stone or tree, it will signify a thing, and
the meaning of such a word will always be contrary to that which is meant
by the finite noun.

The question which grammatical category the infinite verb is part of, is
discussed along the same lines. But as soon as Boethius touched upon its
semantics, he there too made the observation that an infinite verb may be
used in combination with words indicating non-existing things, and
nevertheless cause a true proposition. Virtually, we are confronted in both
cases, namely, those of the infinite noun and the infinite verb, with the
same situation: both cause an affirmative proposition, the truth-value of
which is not hampered by the non-existence of the referent of the subject
term, although in the Middle Ages not every logician would follow
Boethius in this respect. Especially as to the infinite verb, many mediaevals
held differing views, provoked by Boethius' observations about this
subject farther on in his Perihermeneias-commentary, whereas, as we shall
see, the greater part of the mediaeval logicians claimed a referent for the
subject term of the propositions with an infinite noun as nominal part of the
predicate.

To sum up, Boethius emphasized the infinitude of the number of

denotata of an infinite word,IS its in-de-finite way of signifyingI9 and the
exclusion of the denotata of the corresponding finite word. 2o Furthermore,
in the extended version of his Perihermeneias-commentary, he pointed out
that an infinite noun can be used as predicate term in a true proposition,
even in the case where the subject term lacks a referent in reality.21

I sEd. prim .• rec. Meiser, p. 52, 19-22: "quare quoniam id quod definite significare potest
aufert in eo negativa particula, quid vero significare debeat definite non dicit, sed multa
atque infinita unusquisque auditor intelligit." Ed. sec., rec. Meiser, p. 63, 11-14: "sed
sit infinitum nomen, non simpliciter nomen, quoniam nulla circumscriptione designat,
sed infinitum nomen, quoniam plura et ea infinita significat." De cat. syll., PL LXIV,
7950: "quoniam ea significare potest infinita sunt, infinitum nomen vocatur."
19Ed. prim., rec. Meiser. p. 52, 16 ff.: "quid autem ilia significatione velit ostendere,
non definit. .... quid vero significare debeat definite non dicit."
20Ed. prim., rec. Meiser, p. 52, 15 ff.: "qui vero dicit non homo, hominem quidem
tollit." Ed. sec., rec. Meiser, p. 62, 3-5: "cum vero dico non homo, significo quidem
quiddam, id quod homo non est, sed hoc infinitum" ... (p.62, 14 ft).: "sublato enim
homine quidquid praeter hominem est hoc significat non homo." De cat. syll., PL
LXIV, 795C: "non homo autem quod definitum est perimit".
2lSee. ed., rec. Meiser, p. 62, 3-10: "cum vero dico non homo, significo quidem
quiddam, id quod homo non est, sed hoc infinitum. potest enim et canis significari et
equus etlapis et quicumque homo non fuerit. et aequaliter dicitur vel in eo quod est vel
in eo quod non est. Si quis enim de Scylla quod non est dicat non homo, significat
quiddam quod in substantia atque in rerum natura non permanet."
 MASTER JOSCELIN OF SOISSONS AND INFINITE WORDS 69

We have to bear in mind, however, that in his comment on the

sections on the noun and the verb of Aristotle's Perihermeneias. Boethius'
principal intention was to expound the name which Aristotle had given to
these two categories, viz. 'infinite'. This "etymological" approach resulted
in only supelficial attention being paid to their semantics: the infmite words
only signify the things they name in an indefinite way, and furthermore
they are able to signify everything except that which is signified by its
finite counterpart. It was Peter Abailard who reproached Boethius for
having paid attention only to the reason why these words are called
'infinite' and for having neglected their "causa impositionis", as Abailard
put it.22
3. The infinite words and grammar

The emphasis on the infinitude of the denotata as the semantic

characteristic of the infinites urged the mediaeval logicians to cross the
boundaries of their own discipline, for it was not only in logic that the
12th-century student of the liberal arts came across the notion of a noun
with infinitude as part of its meaning function. Priscian called those nouns
which signified a "general substance", "quantity" or "quality", such as
'quis', 'quantus' or 'qualis' infinite nouns, since they subsumed in a
general way all the nouns of the underlying species.23 Neither the so-called
Glosule tradition, a series of gloss commentaries on Priscian, nor William
of Conches paid any attention to the logical infinites. On the other hand the
commentary on the Perihermeneias which has been preserved in pages
237a - 252b of our manuscript explicitly refers to the distinction between
the grammatical and the logical infinites. The anonymous master argues
that Aristotle's claim that only a noun and not a case of a noun or an
infinite noun joined to 'est' creates a true or false proposition also holds for
the greater part of the grammatical infinite nouns except for 'aliquis'.24

22Abailard, Logica "Ingrediemibus": Peter Abaelards Philosophische Schriften. ed. B.

Geyer. 3. Die Clossen zu Perihermeneias. MUnster i. W.: Ascheiidorff 1933 (1927).
pp. 307-503. p. 355. 26-31: "At vero Boethius magis ad causam transitionis huius
modi quod est infinitum. respexit quam ad vim significationis eius et ad proprietatem
ex qua ipsum convenit vocibus. Non enim secundum hoc datum est vocibus quod
infinita significant. sed quod infinite. id est remotive. ut diximus."
23Prisciani grammatici Caesariensis Institutionum grammaticarum libri XVIII. rec. M.
Hertz. 2 vol.. Leipzig 1855-57 (Grammatici Latini. II-III) II. 18. ed. Hertz, vol. I. p.
55. 14-20: "ergo 'quis' et 'qui' et 'qualis' et 'talis' et 'quantus' et 'tantus' et similia.
quae sunt 'infinita' sive 'interrogativa' vel 'relativa' vel ·redditiva·. magis nomina sunt
appellanda quam pronomina: neque enim loco propriorum nominum ponuntur neque
certas significant personas. sed etiam substantiam. quamvis infinitam. et qualitatem.
quamvis generalem. quod est suum nominis. habent: nomina sunt igitur dicenda.
quamvis declinationem pronominum habeant quaedam ex eis." Idem. Inst. gram .• II.
30. ed. Hertz. vol. I. p. 61. 7-8: "Infinitum est interrogativorum contrarium. ut ·quis·.
·qualis·. ·quantus·. 'quot', ·quotus· ... Idem. Inst. gram •• XVII. 37. ed. Hertz. vol. II. p.
131.3-6: "itaque sunt nomiua generalem significantia vel substantiam vel qualitatem
vel quantitatem vel numerum. quae necessario et infinita sunt. quippe cum omnia
suarum generaliter specierum comprehendant in se nomina."
24MS. p. 242a; see below. Appendix IV: "Vel aliter. Omne nomen iunctum ......
70 C.H.KNEEPKENS

In this article I will not discuss the grammatical infinites nor the
morpho-syntactic and syntactic aspects of the logical infinites, but only
mention the fact that by some of the 12th-century grammarians, as for
instance Robert of Paris,25 and also in logical treatises such as the
Tractatus Anagnini,26 these aspects received ample treatment, in all
probability stimulated by the extensive discussions in which these terms
played a crucial role.

For the present, I will concentrate on some aspects of 12th-century

theories of the semantics of the category of the logical infmites.
4. 12th-century theories of the semantics of the logical
infinites

We have to bear mind that logicians paid attention to this subject

during the whole of the Middle Ages. In the later Middle Ages Ockham in
his Summa Logicae argues contrary to Boethius that taken literally the
proposition 'Chimaera est non-homo' is false. since one of its exponents.
namely 'Chimaera est aliquid' is false. 27 a position which we also find
maintained in one of the texts of our manuscript. 28 whereas about a
century earlier the author of the Dialectica monacensis29 and Nicholas of
Paris in his Syncategoreumata took pains to distinguish between the
infinite term 'non-homo', with a hyphen and the negative term 'non
homo' .30 in order to save both the current semantics of the infinite noun

25Robert of Paris. Summa "Breve sit": C. H. Kneepkens. Robert of Paris Het Iudicium
Constructionis. Het leerstuk van de constructio in de 2de helft van de I2de eeuw. Deel
II: Een kritische uitgave van Robertus van Parijs, Summa 'Breve sit', Nijmegen 1987.
p. 47, 20 - 51.18.
26Tractatus Anagnini. ed. L. M. de Rijk. in De Rijk. Logica Modernorum. II 2. pp.
215-332, V. pp. 312-314.
270ckham. Summa Logicae. II, cap. 12, ed. Boehner, p. 284: "Ex isto patet quod de
virtute sermonis ista est neganda 'chimaera est non-homo', quia habet unam
exponentem falsam. scilicet istam 'chimaera est aliquid·. Similiter si nullus homo sit
albus, haec est neganda de virtute sermonis 'homo albus est non-homo'. quia ista
exponens est falsa 'homo albus est aliquid· ... For the phrase 'de virtute sermonis·. see
also W. J. Courtenay. "Force of Words and Figures of Speech: The Crisis over virtus
sermonis in the Fourteenth Century". Franciscan Studies, 44.1984, pp. 107-28.
.
28MS. p. 242b; see below. Appendix IV: "Infinitum est quod similiter potest enuntiari

29Dialectica monacensis. in: De Rijk. Logica Modernorum. II 2. pp. 453-638, p.

621.25 - 622. 11: '''Cesar est non-homo' ... Solutio. Ad horum intelligentiam
sciendum quod iIIe terminus 'non·homo· dupliciter potest sumi. Uno modo secundum
quod ponit aliquid quod est et tamen non est homo. ut capra vel leo vel aliquid
consimile. Altero modo potest sumi secundum quod nil ponit et est pura privatio. Et
secundum hoc est' non-homo' idem quod non-ens.
30Cf. Nicholas of Paris. Syncategoreumata, in H.A.G. Braakhuis. De I3de Eeuwse
Tractaten over Syncategorematische Termen. Inleidende studie en uitgave van Nicolaas
van Parijs' Sincategoreumata. Deel II. Ph.D. Leiden 1979. p. 44, 12: "Ad obiecta
dicimus quod 'non homo' potest duobus modis intelligi: sicut terminus infinitus vel
negatio. Et inquantum est negatio: nichil ponit. et secundum hoc potest dici de
imperatore qui non est, quia per ipsum distrahitur presentialitas verbi. sicut si diceretur
'Cesar nihil est'; similiter potest dici 'Cesar non homo est', idest: nichil quod sit homo
 MASTER JOSCEUN OF SOISSONS AND INFINITE WORDS 71

and Boethius' observation about the predication of an infinite noun in a

proposition with a non-referring subject term. And even today, the
semantics of the infinites is still a subject of discussion. For Strawson, a
class expression has both an extension, the set of things the class
expression can be said of in a true proposition, and an exclusion, the
complementary class. However, in Strawson's view the exclusion does
not comprise everything of which the class expression at issue cannot be
truly predicated, but is restricted to those things about which it makes
literal sense to say that they are or are not what is indicated by the class
expression}1 Geach made the observation that "it is at the least intuitively
odd to take 'no A' as a way of referring to the things called 'A"',32 and
Armstrong in defence of scientific realism challenges the adherents of class
nominalism to indicate the "common property" of a heterogeneous class of,
for instance, the "not-blue" things}3

One of the topics the logicians of the late eleventh and twelfth centuries
discussed in this respect, was the question whether general nouns, that is
terms which have such a meaning that they can be used in order to signify
or comprise everything as the general terms 'a/iquid', 'ens', 'res' or
'substantia', could be negated or rather "infinitated" to 'non-res' or 'non-
aliquid'. Garlandus Compotista, in the late eleventh century, refused to
accept the term' non-substantia', since if he did, he would be compelled to
accept the proposition 'non-substantia est non-homo' with the ultimate
consequence that he would also have had to accept the proposition 'nichil
est a/iquid' .34 The grammarian Robert of Paris, in the second half of the
12th century,35 and the Anonymus of the Tractatus Anagnini shared this
view. The latter pointed out that if he accepted negation of these terms, he

est, et sic intellexit Boethius, et sic contradicit ei quod est' homo'. Secundum vero
quod est tenninus infinitus: a1iquid ponit, ut dictum est, et non potest dici de Cesare
vel de imperatore qui non est, nec contradicit ei quod est 'homo', sed est privativum
eius."
31p. F. Strawson,lntroduction to Logical Theory, London: Methuen 1952, pp. 112 ff.
32p. T. Oeach, Reference and Generality. An Examination of Some Medieval and
Modern Theories, Emended Edition, Ithaca, N.Y.: Cornell University Press 1968, p.
87.
33D. M. Armstrong, Nominalism and Realism. Universals and Scientific Realism, vol.
I, Cambridge: Cambridge University Press 1978 (repr. 1988), p. 40.
340arlandus Compotista, Dialectica. First Edition of the Manuscript. ed. L. M. de Rijk,
Assen: Van Oorcum 1959, p. 62: '''Non-substantia' sic probatur nichil significare ...
ergo vera est ista propositio: 'non-substantia est non-homo' idest 'non-substantia' ista
vox est non-homo, idest designatur ab hac alia voce que est 'non-homo'. Sunt alii qui
dicunt 'non-substantiam' esse tale nomen quod nichil significat, et monstrant per
simile dicentes quia 'non-homo' significat et ea que sunt et ea que non sunt. Sed nichil
dicunt. Nam eo respectu quo dicunt significare ea que non sunt, non est nomen - 'ea
enim signijicare que non sunt' quid est dicere? nichil! - quia nulla vox potest imponi
a1icui rei nisi sit vel fuerit vel futura sit. Ergo 'non-substantia' nomen non sit, cum
non sit impositum rei preterite neque presenti neque future."
35Summa "Breve sit", ed. Kneepkens, p. 47, 22-27: "Ad quod dicimus quod nomina que
conueniunt omnibus per appellationem, ut ens, unum, aliquid, .•. et nomina infinita
non possunt infinitari. Inde est quod non concedimus locutionem istam 'Socrates est
non-ens'."
72 C.H.KNEEPKENS

would also have to accept such propositions as 'Non-ens est'.36 Abailard

also entered the discussion on this subject, and turned out to be very
reluctant to admit this kind of infinite noun. In the Logica "Ingredientibus"
he claimed that if one accepts the infinite term' non-res', the consequence
is that the finite noun 'res' can only be employed in order to speak about
existing things,37 whereas in the Dialectica he expressly said that 'res' and
'aliquid' are also nouns which can be used in order to speak of non-
existing things,38

In the texts of the Orleans manuscript this aspect is put forward in the
second commentary on the Perihermeneias. The author broaches the
question, but he is not even sure of their status, namely, whether they are
infinite nouns, and suggests that perhaps they do not exist; however, if
they do exist, 'non-ens' should denote non-existents only and 'non-res'
should be used in order to denote neither existing nor non-existing things.
Actually, their practical value turns out to be negligible, if not zero,39

In the same section the anonymous master comments upon the

meaning function of infinite words, and confronts his audience with two
different views. He attributes one of these to a certain master W. and his
followers. This master held the opinion that an infinite noun signifies
everything which is notwhat is signified by the corresponding finite noun.
In his view 'non-homo' signifies all the things which are not a man, and
'non-currit' all the things which are not dealt with by 'curri!'. According to
the adherents of this theory, however, our commentator observes, it is
possible to say of a stone that "it is a 'non-man' and that it is not a 'non-
man"', which appears an absolutely untenable position to him.40

The same view is mentioned in the twentieth entry of the sophismata

collection. Here it is attributed to a master William. This master argued that
'non-homo' signified everything except humanity.41 We also meet this

36Tractatus Allagllilli, ed. De Rijk, p. 312, 5-7: "Natura autem infinitorum est quod
semper volunt aliquid attribuere alicui. Unde termini omnia continentes non possunt
infinitari. Unde nichil est: 'lloll-aliquid est', 'lloll-res est'."
37 Abailard, Logica "/llgredientibus", ed. Geyer, p. 355, 15-23: "Quod ergo ait ista (sc.
infinita) de quolibet quod continent, tam exsistente quam non exsistente, posse dici
vere, quantum in ipsis est, nil aliud ostendit, quam ea infinite significare, id est non
ponendo esse, sed magis removendo. Quod dum istis infinitis adscribit, omnibus
attribuit in istis, remotive scilicet significare, quod etiam habet non-ens et non-res.
<cum> finitum in designatione exsistentium tantum ponatur. Nam si res tam
exsistentia quam non contineat, infinitari posse non videtur, cum eius infinitum non
habeat, quid comprehendat."
38Dialectica, ed. De Rijk, p. 127,25-31: "Nam 'res' quoque et 'aliquid' significativum,
que infinita non sunt, ea quoque que non sunt continere dicuntur, cum negativa
particula careant, qua finiti significationem perimant. Unde in Primo Periermelleias
dicitur: 'hircocervus enim significat aliquid'; hic enim 'aliquid', ut Boetius ostendit,
nomen est rei non-existentis, ex quo etiam innuitur hircocervum quoque significativum
vocari."
39Cf. p. 242b; see below, Appendix IV: "De 'non-ens' et 'non-res' opponi potes!.. .. "
40MS, p. 242a; see below, Appendix IV: "Dicit enim m<agister> W. et sui .... "
41 MS, p. 283a: "Secundum magi strum Guillelmum lIoll-homo significat omnes res
preter humanitatem et lIoll-albus similiter, et lIoll-legit similiter." See also below,
Appendix I.
 MASTER JOSCELlN OF SOISSONS AND INFINITE WORDS 73

opinion in the commentary on Boethius' treatise on categorical syllogisms,

which has been preserved in pages 17 ta-173 b of the Orleans manuscript.
The master expounds Boethius' remark that "the finite is annihilated by the
infinite" in the following way: the infinite noun produces a concept
(' intellectus' is the Latin word used) which is contrary to the concept of the
finite noun.42

Unfortunately, as often happens in texts in which the opinions of

adversaries are ridiculed, the texts do not enter upon the underlying
reasoning which sustains the thesis of this master William to the effect that
it is possible to say of a stone that is and is not a non-man. This implies
that a stone generates an intellectus which is contrary to the intellectus of
man, and that it does not generate such an intellectus. Abailard, in his
Dialectica (ed. De Rijk, p. 566), shows that in the propositions 'lapis est
non-homo' and 'lapis non est non-homo' we are confronted with the
phenomenon of derived equivocation. The term 'homo' is equivocal, since
it is differently defined as a 'homo verus' and as a 'homo pictus', namely
'animal rationale mortale' and 'assimilatio animalis rationalis mortalis'
respectively. For a stone can be said to generate an intellectus which is
contrary to the intellectus of man, namely "quod non est animal rationale
mortale". On the other hand, a stone in the shape of a man generates an
intellectus which is not contrary to the intellectus of man, taken as an
"assimilatio animalis rationalis mortalis".43 However, this cannot be the
answer to our question, for it is explicitly claimed that according to these
logicians the proposition 'lapis non est non-homo' is true, since a stone is
a stone and not a cow or a horse. Nevertheless, the solution must be
looked for in an equivocal use of the term 'non-homo'. The proposition
'lapis est non-homo' says that a stone is something, namely a stone which
is not a man, for a stone generates an intellectus contrary to the intellectus
generated by a man. The proposition 'bos est non-homo' says that a cow is
something, namely a cow which is not a man. However, the intellectus of
a cow is not the intellectus of a stone. So the lack of a positive common
property each time urges the speaker to change the definition of the infinite
noun in accordance with the subject.

This interpretation finds support in the observation made in the

commentary on Boethius' Categorical Syllogisms in the Orleans
manuscript, that according to this theory an infinite noun, such as for
instance 'non-homo', signifies humanity ('humanitas'), namely in the
same way as it signifies all the other nominata, since, of course, the
following proposition is correct and true: 'Humanity is non-man'.44

42MS, p. I73b; see below, Appendix II: " Vel alio modo hec Boetii Iittera
exponitur.... "
43Abailard, Dialectica, ed. De Rijk: "Sed fortasse quedam possunt esse equivoca, non
communia, cum videlicet <per> eamdem rem possit equivocatio fieri cuiusdam
nominis secundum diversas diffinitiones, ut 'non·homo', quod infinitum veri ac pieti
quidem, sicut suum finitum, equivocum oportet esse, de hoc uno Iapide equivoce
predieatur, modo scilicet cum hac diffinitione quod non est animal rationale mortale,
modo etiam cum hac quod est assimilatio animalis rationalis mortalis."
44MS, p. I73b; see below, Appendix II: "Vel secundum aliam sententiam dieamus ...
cum tamen secundum hanc sententiam non plus principaIiter albationem quam reliqua
sua nominata significet."
74 C.H.KNEEPKENS

The opponents of this view came up with a solution which would

clearly also partly meet the criticism of class nominalism voiced in modem
times by Armstrong,45 with whom they shared the basic assumption that a
"negative property" of the status of "not being F" is hardly acceptable as a
"common property". According to the texts their leader was a master
Goslenus. A master of this name is referred to several times in our
manuscript. In pages 149 to 151 a set of notes, mutilated in the end,on
Boethius' De divisione has been preserved, which is attributed to a master
Goslenus, whereas the text which is appended to these Notule is ascribed
by Peter King to a pupil of Goslenus. 46 In this text we also find the
masters William and Goslenus opposed to each other. In all probability this
Goslenus can be identified with master Goslenus, the later bishop of
Soissons. He was a renowned logician in his time, but his activities in this
field have earned him only some casual references in the history of
mediaevallogic.47

According to John of Salisbury, Joscelin held the so-called Collectio

theory in the dispute about universals. He says of Joscelin that he
"attributes universality to the collection of things, and denies it of each of
them".48 A more elaborate account of the tenets of Joscelin can be read in
the Tractatus de generibus et speciebus, wrongly attributed to Abailard by
its editor Cousin in 1836, but as is now commonly accepted the product of
a partisan of Joscelin '·s. The adherents of this theory held that the genus is
the matter for the species and the species in its tum the matter for the
individual. However, each individual has its own particular matter, its own
essence. 49 So Plato's humanity, which serves as matter for the form
Platonitas the combination of which results in the individual Plato, is not
numerically identical with Socrates' humanity. The species is the collection
of all the singular essences, which is called in the case of the species of

45See above, n. 33.

46The text was edited by Cousin in his Ouvrages inedits d' Abelard, Paris 1836, p. 507-
550, and by Peter King in Peter Abailard and the Problem of Universals, pp. 114*-
185* together with a translation, pp. 186*-212*.
470n Joscelin and his philosophical tenets, see King, Peter Abailard and the Problem of
Universals, pp. 187-92, and Tweedale, Logic(i): From the Late Eleventh Century to the
Time of Abelard, p. 222.
48John of Salisbury, Metalogicon: Ioannis Saresberiensis Episcopi Carnotensis
Metalogicon Libri IIII, rec. C.C.I. Webb, Oxford: Oxford University Press 1929, p.
95, 7-9: "Est et alius qui cum Gausleno Suessioni episcopo uniuersalitatem rebus in
unum collectis attribuit et eamdem singulis demit."
49For the sake of convenience, all the references are to King's edition: p. 162*:
"Unumquodque individuum ex materia et fonna compositum est, ut Socrates ex homine
materia et Socratitate forma; sic Plato ex simili materia (scilicet homine) et forma
diversa (scilicet Platonitate) componitur; et sic de singulis hominibus (homines ed.). Et
sicut Socratitas quae fonnaliter constituit Socratem numquam est extra Socratem, sic
illius (illis ed.) hominis essentia quae Socratitatem sustinet in Socrate numquam est
nisi in Socrate.... Et sicut de homine dictum est, scilicet quod illud hominis quod
sustinet Socratitatem illud essentialiter non sustinet Platonitatem, ita de animali. Nam
illud animal quod formam humanitatis quae in me est sustinet illud essentialiter alibi
non est sed illi indifferens est in singulis materiis singulorum individuorum animalis."
 MASTER JOSCELIN OF SOISSONS AND INFINITE WORDS 75

man, humanity,50 and not the collection of all the individuals!51 So that
part of humanity which is the matter of Plato is, as has already been said,
not the same as but at the same time does not differ from that part of
humanity which is the matter of Socrates, nor is it the species of man.
Every individual - or mutatis mutandis every species etc. - is a
materiatum, that is, the result of the joining of its matter and its form.

It turns out that the adherents of the collectio theory had developed
corresponding theories of semantics and grammar. They refused a
distinction of the noun into substantive and adjective based on the fact that
a substantive noun should signify an essence as essence, and an adjective
noun an accidens in adiacentia. 52 In their view, a substantive has as its
meaning function that of naming something according to its matter, which
it principally signifies, or, in the case of proper nouns, its essence
expressed;53 an adjective noun has as its meaning function that of naming
something according to a certain forma, which the adjective noun
principally signifies and which is found in the named entity.54 This
principal signification55 turns out to be Priscian's quality which is one of
the two signification components of the noun: "nomen significat
substantiam cum (or et) qualitate(m)" (Inst. gram. II, 18). Furthermore,
they accepted a naming or denoting function of the noun. A noun names a
"res" under the principal signification of the noun in question: in the case
of homo, humanity, of album whiteness, of rationale rationality, etc. When

50lbid. p. 163*: "Itaque tota ilIa multitudo quae humanitas dicitur materia est Socratis et
singulorum."
51/bid. p. 162*: "Speciem igitur dico esse non ilIam essentiam hominis solum quae est
in Socrate vel quae est in ali quo alio individuorum sed totam iIIam coIIectionem ex
<singulis om. ed.> iIIis materiis factam, id est unum quasi gregem de essentia hominis
quae Socratitatem (Socrates ed.) sustinet, aliis singulis huius naturae coniunctum."
52lbid. p. 165* (par. 102): "Nam quod dici solet - adiectivum esse quod significat
accidens secundum quod adiacet, et substantivum quod significat essentiam ut essentiam
- ridiculum est vel sine inteIIectu."
53lbid., p. 165* (par. 101): "Et nota quod nomina ilIa tantum dicuntur substantiva quae
imponuntur ad nominandum aliquem propter eius materiam. ut homo et caetera
universalia substantiva, vel propter expressam essentiam, ut Socrates - idem enim
nominat et significat, scilicet compositum ex humanitate et Socratitate."
Ibid., p. 165* (par. 98-99): "Eodem modo homo impositum fuit cuilibet materialiter
constituto ex homine ad nominandum propter eorum materiam, scilicet speciem quam
principaliter significaret. (par. 99) Itaque cum dicitur "Socrates est homo" hic est
sensus: Socrates est unus de materialiter constitutis ab homine; vel ut ita dicam
Socrates est unus de humanatis (humanitis ed.)."
54lbid. p. 165* (par. 101): "Adiectiva vero ilIa dicuntur quae imponuntur alicui propter
formam quam principaliter significat, ut rationale et album res ilIas nominant in quibus
inveniuntur rationalitas et albedo". Ibid., p. 165* (par. 99): "Sicut cum dicitur
"Socrates est rationalis" non iste est sensus: res subiecta est res praedicata, sed Socrates
est unus de subiectis huic formae quae est rationalitas."
55lbid., p. 164* (par. 98): "immo sicut rationale et homo, sic et quodlibet aliud
universale substantivum alterius nomen est, per impositionem quidem eius quod
principaliter significat. Verbi gratia: rationale vel album impositum fuit Socrati vel
alicui sensilium ad nominandum propter formas, id est rationalitatem et albedinem,
quas principaliter significant. Eodem modo homo impositum fuit cuilibet materialiter
constituto ex homine ad nominandum propter eorum materiam, scilicet speciem quam
principaliter significaret."
76 C.H.KNEEPKENS

some of the nouns are used in what later will be called "suppositio
simplex", namely when a noun is used to name the nature which it
principally signifies as for instance the noun 'homo' in 'homo est species' ,
we are confronted with the phenomenon of translatio.56

According to the section on infinite words in the Sophismata

collection, master Joscelin proposed a theory of the semantics of this
category, which essentially differs from the one held by master William,
but is closely connected with his own theory of the signification of the
finite noun. 57

As we saw above, the signification function of the finite noun was

split up by Joscelin into two main components, namely, the principal
signification and the denotation, which regarded the materiata or in the case
of the adjective nouns, the subjecta. Both these aspects were analyzed by
the mind and conceived of in a conjunctive way, which resulted in an
intellectus simplex.

The difference between the meaning function of the infinite noun and
the finite noun is twofold. First, the meaning function of the infinite noun
was broken up into three components: 'tria significat'. The first of them is
the principal signification of the corresponding finite noun. Joscelin needs
it in order to avoid the problem of equivocation which occurred to master
William, So 'non-homo' will always have as its principal signification
humanity, and 'non-album' whiteness, a position which according to Prior
was also maintained by Keynes in his Formal Logic,58 who took the view
that the connotation of a negative term is in each case the same as that of
the corresponding positive term,59 although one could doubt whether
Joscelin's notion of principal signification covers Keynes' connotation.

The second significative aspect consists of signifying or rather

denoting the materiata by the matter in question or the subjecta by the form,
respectively, which correspond with the nominata of the finite noun at the
semantic level. The reason why this component, which is at odds with
Geach's intuition, was claimed is that otherwise it would be difficult to
distinguish between the meanings of for instance 'non-homo' and 'non-
humanitas' or 'non-album' and 'non-albedo', When we analyze the
intellectus simplex of 'non-homo' both these aspects are conceived of in a
conjunctive way,

56lbid., p. 165*-166* (par. 104): "Sciendum est ergo quia vocabula quae imposita sunt
rebus propter a1iud significandum, id est principaliter circa ea, quandoque transferuntur
ad agendum de principali significatione. Ut cum rationale impositum sit substantiis ad
nominandum, et album similiter, translative tamen dicitur 'rationale est differentia' et
'album est species coloris' - nihil a1iud intelligo quam rationalitas et albedo. Sic et
homo transfertur ad agendum de natura quam principaliter significat cum dicitur 'homo
est species' ."
57MS, p. 283a; see below, Appendix I.
58J.N. Keynes, Studies and Exercises in Formal Logic, fourth edition, London:
Macmillan 1906, par. 38.
59Prior, The Doctrine of Propositions and Terms, pp. 84-5.
 MASTER JOSCEUN OF SOISSONS AND INFINITE WORDS 77

Up to this point, the meaning functions of the finite and the infinite
noun run parallel to each other. But finally me infinite noun, for instance
'non-man', has also to signify its own nominata, namely all those things
which are its extension, that is the whole set of things of which it would be
true to say that they are, in this case, non-man. However, unlike the way
in which the intellectus joins the principal signification and the materiata, it
conceives of these nominata in such a way that it separates them from the
other two meaning components conceived of in conjunction. We must bear
in mind that according to this theory an infinite noun, for instance, 'non-
homo' can signify humanity twice, namely once as its principal
signification, and once as belonging to the nominata of 'non-homo', since
'humanity is non-homo'. In the theory of master William only the latter
way of signifying humanity is possible for' non-homo' .

The last item dealt with in the section on the infinite nouns of the
Sophismata collection is the meaning function of the negative universal
sign 'nullus', which according to the commenting master and at least
several of his twelfth-century colleagues, is also an infinite word. In their
opinion they followed Priscian, who was supposed to have considered
'nullus' the infinite of 'ullus' .60

Boethius discusses the question, which grammatical category it

belongs to, and makes some casual remarks about its semantics.61 He
emphasizes the fact that this word has a combined meaning based on its
parts, namely negation and signifying 'unus', since according to ancient
grammar 'ullus' was a diminutive of 'unus' - which it is, indeed. Priscian
also deals with the semantics of 'nul/us', and argues that words like
'nullus', 'numquam', 'nusquam' and so on are "abnegativum", denying
everything which can be signified by their positive counterparts.62

An essential point in twelfth-century discussions of the meaning of

'nullus' is that it has somewhere in the description of its signification the
notion of unity ('unitas'), which when said of a thing is expressed by the
words 'unus' or its diminutive 'ullus', and which, in this context, has the
function of a sort of existential quantifier.

According to master W., in all probability our master William, 'nul/us'

signifies all the non-existing things: "omnes res non existentes" such as the
chimaera, the staggoat and so on. Unfortunately, we are confronted with a
lacuna in the text, which ends with "and all the things that are denotated by
'ullus'. We find a text in the second Perihermeneias commentary preserved
in this manuscript, which refers to 'nul/us' as signifying only non-existing
things. Here it is said that only 'nul/us' is said of "non-existing things" and
that it conceives of unity in a disjunctive way about these "non-existing

60Cf. Priscian, Inst. gram. XVII, 45-6: "necesse est enim 'ullus' omnium intellegere et
cunctos posse numeros ad hoc referri ... Et sciendum, quod composita quoque eorum
abnegativa sunt omnium generaliter, quae per ea possunt significari, ut 'nullus
omnium deorum' vel 'hominum' vel 'eorum quae sunt'."
61Ed. sec., rec. Meiser, p. 146.
62Priscian, Inst. gram. XVII, 46 (ed. Hertz, II, 136,6-7): "composita quoque eorum
abnegativa sunt omnium generaliter, quae per ea possunt significari".
78 C.H.KNEEPKENS

things". So the expressions 'nulla chimaera est' and 'nullus homo est'
would mean 'there does not exist any chimaera' or 'man'. In the latter case
it would assign 'man' to the class of non-existing things. However, a
difficulty must have been recognized, when propositions like' nullus homo
currit' were uttered, since these can be perfectly true even when man is not
a non-existing thing. In order to escape from this problem the anonymous
author added that the intellectus compositus of the sentence of which
'nullus' is a part, should not conceive of those non-existing things.63

Master Joscelin also reflected upon 'nullus' and its meaning. He

presents an explanation which is in agreement with his theory of the
semantics of the infinite nouns. First he argues that 'nullus' and 'ullus'
have the same - principal- meaning, namely 'unity'. The particular aspect
of the meaning of 'nullus' is that it removes this 'unity' of the denotata of
'ullus', whereas in its turn 'ullus' attributes unity to the denotata of
, n ull us'. The source of this circular meaning must be looked for in
Priscian. Priscian says that 'nullus' negates in a general way all the things
which are. 64 So it is said of "all the things that are" which, however,
transposed to the interpretation of master Joscelin is to be interpreted as "of
all the things of which unity can be predicated", that is which fall under the
denotation range of 'ullus'. 'Ullus' in its turn attributes unity to all those
things which fall under the denotation range of 'nullus', namely all the
things of which 'nullus' can be said.

This story is told by Joscelin in order to show the essential difference

between the two classes of infinites, namely, those of the non-homo type
and nul/us, which unlike grammatical infinites such as 'quis' and so on, is
also an infinite word in which negation plays a role. Both types have a
principal meaning which is signified in "a removing way". The main
difference, however, must be looked for in the denotative aspect. Non-
homo signifies humanity in a removing way, but not from the denotata of
its finite counterpart, homo, but of its own denotata. 'Nullus' on the other
hand signifies unity in a removing way, but not of its own denotata, but of
the denotata of 'ullus', namely, of those things which can be subject to
unity, i.e., of which 'ullus' can be said.
Final remarks

I hope to have made clear in this article - which might be considered a

first move towards an edition of the whole collection - that the discussion
of infinite words which is part of the collection of sophismata and notes
preserved in the Orleans manuscript, is closely connected with the theories
on this subject dealt with elsewhere in the manuscript. As I remarked in the
introductory part, there are good grounds to assume that the greater part of
the entries do not diverge in this respect.

The doctrine of infinite words shows the usual picture we have of late
eleventh and twelfth century activities on the borderline between grammar

63MS, p. 283a; see below, Appendix I.

64ef. Priscian, Inst. gram. XVII, 46.
 MASTER JOSCELIN OF SOISSONS AND INFINITE WORDS 79

and logic. 65 Encounter and fruitful interaction of both disciplines often

urged the masters, on pain of confusion, to redefine traditional notions or
to make essential distinctions. In the case of infinite words a
reconsideration of the definition of an infinite word was inevitable, since it
was no longer obvious which words could truly be said to belong to this
category and why. A distinction between grammatical and logical infinite
words was one of the results of these reflections. This even resulted in
sophisms or sophismatical questions like 'Utrum nul/um infinitum sit
jinitum', number 42 in the list of sophismata.

Another aspect is that the several schools developed their own

linguistic theories in order to support their philosophical tenets, or to adapt
the traditional grammatical theory to their philosophical claims. It appears
that this phemonenon can also be pointed out in the doctrine of infinite
words. Joscelin of Soissons - and his followers - tried to harmonize the
traditional doctrine of the semantics of the noun and his view on the status
of the universals, namely, Collective Realism, incidentally using inter alia
the recently discovered or introduced distinction between significare and
nominare, and he fits in his theory of the infinite words in this - his own-
newly created semantic theory.

University of Nijmegen

65See for this topic De Rijk, Logiea Modemorum, II, 1, and Tweedale, Logie(i): From
the Late Eleventh Century to the Time of Abelard, pp. 197-8.
80 C.H. KNEEPKENS

Appendix
Orleans, B.m. 266: De injinitis

SOPHISMATA (pp. 281a-290b)

p.283a:

<De infinitis>

Videndum est de infinitis. Secundum magistrum Guillelmum non-

homo significat omnes res preter humanitatem et non-albus similiter, et
non-legit similiter. Sed magister Goslenus dicit aliter dicens etiam quod
non-homo significat principaliter humanitatem et omnia materiata ab
humanitate et omnia alia que non sunt materiata ab humanitate, et non-albus
eodem modo uidelicet albedinem significat principaliter et omnia illa de
quibus agitur per non-albus, non-legit similiter.

Ex alia parte dicit [add. MS£] m. W. quod nullus significat omnes res
non-existentes: Chimeram et hircoceruum et cetera non-existentia et omnia
illa de quibus agitur per ullus.

Magister uero Goslenus dicit idem significare nullus et ullus. Nullus

significat unitatem remouendo <illam> ab omnibus illis de quibus agitur
per ullus. Vllus significat unitatem attribuendo illa<m> omnibus illis de
quibus agitur per nullus. Et ita idem significant, sed diuersis modis: nullus
remouendo unitatem ab omnibus illis de quibus agitur per ullus tali modo;
ullus uero attribuendo unitatem omnibus illis de quibus agitur per nullus
tali modo. Non-homo habet aliter significare, quia non-homo significat
humanitatem principaliter quam remouet ab omnibus suis nominatis; nullus
uero significat unitatem quam remouet ab alterius nominatis; et ita sunt
diuersa infinita.

Sed si aliquis interrogat quid sit infinitum, dicamus illud esse infinitum
quod significat talem rem principaliter quam disiungit a suis nominatis uel
ab alterius nominatis: illa que disiungit suam principalem significationem a
suis nominatis, ut non-homo, illa que disiungit suam principalem
significationem ab alterius nominatis, ut nullus.

p.285a:

Vtrum nullum infinitum sit finitum ? Aliquid infinitum est

finitum, quia quis est infinitum et finitum. Ideo dicitur finitum, quia
significat intellectum qui concipit aliquam materiam circa omnia nominata
illius. Ideo dicitur infinitum [ideo ... infinitum add. MS£], quia significat
intellectum qui non concipit aliquid circa omnia sua nominata.
 MASTER JOSCELIN OF SOISSONS: APPENDIX 81

Aliud infinitum est, ut dicit Aristotiles: Infinitum est quod significat

intellectum qqi concipit suam principalem significationem disiunctim ab
omnibus suis nominatis. Et nullum tale infinitum est finitum.

II

Commentary on Boethius' De categoricis syllogismis (pp. 171a-173b)

p.173b:

Vt autem infiniti uerbi significationem altius exquiramus, prius quid

infinitum nomen significet, inuestigemus.

Dicit autem Boetius quod non-homo quod diffinitum est, perimit.

Quam litteram doctorum nostrorum sententia[m] duobis modis exponit.

Vel non-homo quod definitum est, perimit, idest intellectus huius

uocis nominata sua tali modo. idest disiunctim a nominatis sui finiti.
concipit. Et hoc modo non-homo principaliter significat humanitatem sicut
et suum finitum. Nominat autem omnia preter sui [MS£ sua MS] finiti
nominata. Ipsam ergo humanitatem inter reliqua sua nominata [add. MS£]
nominat, quam etiam, ut dictum est, et principaliter significat.

Vel alio modo hec Boetii littera exponitur. que dicit quod finitum ab
infinito perimatur sic: Non-homo quod [quoddam MS] definitum est.
perimit. idest intellectum sui finiti intellectui contrarium gignit. Cum enim
intellectus hominis solos homines. intellectus autem non-homo omnia que
non sunt homines. uideat. patet profecto quod non-homo contrarium
intellectum [add. MS£] suo finito habeat.

Omnia uero que de infinito nomine diximus. de infinito uerbo fere

dicere possumus. Hoc namque infinitum uerbum, quod est <non->a/bet.
quandam actionem. idest albationem, que a suo finito uerbo principaliter
significatur. principaliter significat. Omnia autem non-albentia, ipsam
quoque albationem, nominat. Et hoc secundum primam sententiam quam
prius secundum magistrum Goslenum de infinito nomine posuimus.

Vel secundum aliam sententiam dicamus quod non-a/bet [albet MS£

habet MS] quod diffinitum est. perimit. idest contrarium intellectum suo
finito uerbo facit. Nam sicut de infinito nomine diximus. cum a/bet tantum
albentia significet, intellectus non-a/bet non-albentia omnia, ipsam quoque
albationem uidet. Et secundum hoc non-a/bet contrarium intellectum suo
finito habet. cum tamen secundum hanc sententiam non plus principal iter
albationem quam reliqua sua nominata significet.

III

Partial commentary on Boeth .• De categoricis syllogismis (pp. 179a-183b)

82 C.H.KNEEPKENS

p.180a:

Cum enim non-homo aliquid significet, tutum est quod dicamus

humanitatem significare. Et sic expone: illud quod significat, non
continetur in ipso nomine, idest non nominatur ab ipso. Namque non-
homo opposita hominis nominat. Sed in uerbi tractatu prolix ius natura
infinitorum nominum et uerborum patebit '"

Item. Quis iunctum cum est non hoc habet. Magister tamen W. dicit
quis hoc habere, quando ponitur pro aliquis.

Nos uero dicimus quod non est nomen, idest non solet appelari
nomen, quia est infinitum.

p.I80b:

Expulso igitur errore non-albet et alia huiusmodi uerba dicantur, sed

propter solum usum, remouit a uerbo sicut modo <in>finitum nomen a
nomine. Si enim ista nullo modo essent nomina uel uerba, non optime
dixisset aduerbia esse nomina [add. MS£] que satis indigniora sunt. Cum
igitur constet esse uerbum non-albet, et opportet ut sit substantiuum uel
adiectiuum. Sed substantiuum non est. Igitur est adiectiuum. Sed queritur
quam actionem uel quam passionem significat. Quod autem omnes
significet, non est tutum. Mirum enim et ridiculum uideretur, cum uellem
dicere'ille sedet', dicerem 'Ule non-albet'. Dicendum igitur quod nichil
diffinitum monstrat, idest nichil affirmatiue [monstrat expo MS] significat.
Nam significat quandam actionem, scilicet albationem, disiungendo earn a
nominatis. Non-albet teste auctoritate, quod infinitum <est>, idem
significat principaliter quod suum finitum perimendo eius significationem,
sicut non-homo infinitum nomen idem principal iter significat quod homo,
sed quasi disiunctiue, idest disiunctim, uidet humanitatem circa sua
nominata, homo uero coniunctim. Vnde dicitur non-homo nichil diffinitum
significat.

p.I82a:

Nullus est nomen rectum. Si inferatur: Si est rectum, tunc iungitur

cum homo intransitiue, non sequitur, quia [add. MS£] fallit in negatiua
[add. MS£] propositione et in falsa. Et sic nullus nomen sit infinitum teste
Prisciano: Nullus est abnegatiuum ullus,66 et est nomen inpersonale non
agens de aliquo nec gramaticum nomen. Significat unitatem remouendo a
nominatis sui finiti uel unus.

Commentary on Aristotle's De interpretatione (pp. 237a-252b)

66Priscian, Illst. gram. XVII, 46, ed. Hertz, vol. II, p. 135 ff.
 MASTER JOSCELIN OF SOISSONS: APPENDIX 83

p.242a-b:

Vel aliter. Omne nomen iunctum etc.67 Nomen dicit quod secundum
morem et usum solet apellari nomen. Ideo quis iunctum cum est et quisquis
etcetera non significant uerum uel falsum, quia [per expo MS£] infinita, ut
non-homo. Non sola enim dialectica infmita, sed etiam gramatica uoluit
non esse, idest non apellari, nomen. Vel si dicamus 'finitum' 'rectum' esse
de diffinitione, per 'finitum' remouet omnia dialectic a infinita, scilicet [add.
MS£] non-homo etcetera, et omnia gramatica, ut quis, quidcumque.

Si opponatur quod a/iquis iunctum cum est uelfuit significat uerum uel
falsum, uerum est. Sed dicimus omnia dialectica infmita remoueri, ut non-
homo, et etiam grammatica, ut quis. Sed si aliquis, quod est infinitum, hoc
habet, non multum male.

Et sit infinitum uerbum non-currit, scilicet et non-Iabo, quoniam

similiter est in quolibet quod est uel quod non est. Hec est causa propter
quam dicatur infinitum uerbum, scilicet ut quidam exponunt, quia significat
res existentes et non existentes, quod etiam [MS£ omne MS] conuenit
infinito nomini et non infmitis ut est res et ens.

Vnde nunc [MS£ nec MS] prius uidenda est significatio infiniti
nominis et uerbi infiniti. Et prius sic exponemus illam causam quod solum
huiusmodi infinitis conueniet. Nominum ergo quedam sunt sustantiua,
quedam adiectiua, ut homo sustantiuum, album adiectiuum, quorum sunt
infinita: non-homo, non-album, de quorum significatione diuerse sunt
sententie.

Dicit enim m<agister> W. et sui quod non-homo significat omnes res

que non sunt homines, ut eque, non-album omnes res non albas; item non-
currit [MS£ erit MS] omnia illa de quibus non agitur per currit; et sic de
uerbis. Sed secundum illos de eadem re uerum est et 'est non-homo' et
'non est non-homo', scilicet de lapide, quia est lapis et non est bos uel
equus, quod multum absurdum uidetur.

Vnde sic aliter de infmitorum significatione sumus [add. MS£] diuersi.

Dicimus enim quod non-homo tria [MS£ ita MS] significat, scilicet
humanitatem (speciem) et materiatum ab ea et illa de quibus agitur per non-
homo, sed ille simplex intellectus [humanitatem ... intellectus add. MS£ in
marg.] humanitatem et sua materiata coniunctim concipit et illa duo a
nominatis huius uocis non-homo disiungit, idest illa duo et nominata a
non-homo disiunctim concipit. Non-album tria significat, scilicet
albedinem et subiectum illius et nominata a non-album et eodem modo quo
diximus de non-homo. Alioquin nisi per non-homo materiata ab
humanitate, et per non-album subiectum albedinis significari diceremus,
non multum uiderentur in significatione differre non-homo et non-
humanitas, non-album et non-albedo.

67Boethius, Introd. ad syll. categ., Patr%gia Latina 64, 765A.

84 C.H.KNEEPKENS

Verba uero infinita et nullus, quod dicit Priscianus infinitum ul/us,

eiusdem sunt inuentionis, <non [scripsi]> quia sic fuerunt inuenta, ut per
ea de nullo ageretur, sed sic quod unumquodque illorum fuit impositum ad
intellectum significandum disiungentem principalem significationem sui
finiti ab eis de quibus agitur per illud [illam MS] finit<um>, ut nul/us
principaliter significat unitatem, idest principaliter intellectum unitatem
concipientem et eam disiunctim concipit. Nec aliter bene potest legi quod
dicit Boethius in Diuisionibus, scilicet negatio quicquid proponit ab eo
quod est esse disiungit. Et cum dico 'homo', quasi sit quiddam, locutus
sum; cum uero 'non-homo', substantiam hominis negatione destruxi circa
illa eadem de quibus agitur [po 242b] per ul/us. Dicit etiam Boethius in
commento: Infinitum significat unum confusis subiectis.68 Et ita nullus est
inpersonalis uox. Nec ideo facit orationem cum aliquo sustantiuo, ut non
albus cum [non expo MS] homo, quia nul/us non est secundum hanc
sententiam oratio. Et sic de nihil, de nemo et de infinitis uerbis. Ideo dixit
Boethius uerbum infinitum non posse poni in oratione, quin separetur in
negationem et uerbum finitum,69 quia per ilIa de nullis agitur.

Hic etiam solet queri an sint infinita sustantiua uel adiectiua. Ad quod
potest dici quod neque sustantiua neque adiectiua.

De non-ens et non-res opponi potest, si sint infinita. Sed potest dici

quod numquam fuerunt inuenta, uel si sint inuenta, per non-ens tantum de
non-existentibus agitur, per non-res neque de existentibus neque de non-
existentibus.

De nul/us alie sunt sententie, una scilicet que notata est in Notulis, alia
hec potest esse: quod per nul/us de non [add. MS~]-existentibus tantum
agatur; et circa ea unitatem disiunctim concipit intellectus huius uocis
nul/us, sed compositus intellectus orationis cuius est pars nullus, ilIa non-
existentia non concipit.

Nunc redeamus ad causam quam ponit Aristotiles, scilicet quare que

sint infinita, dicantur infinita. Et exponamus ita quod solis huiusmodi
infinitis conueniat, scilicet omne nomen quod est infinitum, infinitum
dicitur quoniam similiter etc.,70 idest quoniam uel disiunctim concipit suam
principalem significationem intellectus illius circa sua nominata uel circa
nominata a suo infinito. Infinitum est quod est similiter in quo[d]libet quod
est uel quod non est. Potest etiam exponi, ut conueniat solis uerbis, sic
scilicet infinitum etc.: Infinitum est quod similiter potest enuntiari de eo
quod est, idest de existentibus, et de eo quod non est, idest de non-
existentibus, uere, ut 'non-currit'. Possumus enim uere dicere 'Chimera
non-currit' et 'homo non-currit', quod non conuenit infinito nomini. De eo
enim quod est, possumus enuntiare uere 'non-homo', sed non de eo quod

68Cf. Boethius, Comment. in librum Arist. PERI HERM., ed. prim., ree. Meiser, p.
127,20 ff.
69Boethius, Comment. in librum Arist: PERI HERM., ed. sec., ree. Meiser, p. 261, 9
ff.
70Aristotle, De interpretatione 3, 16b15: tr. Boethii, ed. Minio-Paluello, p. 7,9 ff.
 MASTER JOSCELlN OF SOISSONS: APPENDIX 85

non est, ut uerum est 'equus est non-homo', sed non est uerum 'Chimera
est non-homo'.

Aliter infinitum etc., idest infinitum est quod est nomen eorum que
sunt et eorum que non sunt ret ... sunt add. MS£]. Sed hoc non conuenit
solis infinitis. Homo enim nomen est [add. MS~-] eorum qui sunt et qui
fuerunt et erunt. Et secundum hoc dicimus quia non est diffinitio, sed
quedam noticia infinitorum. Si uero uolumus esse diffinitionem, dicimus:
Infinitum est quod est nomen et eorum que sunt, et eorum que non sunt, et
significat rem sui finiti perimendo, idest infinitum est quod significat
intellectum qui concipit rem finiti nominis disiunctim et ab eo quod est, et
ab eo quod non est, ut non-homo. Homo ergo etsi nomen est eorum que
sunt et non sunt, non significat perimendo. Non-res, non-ens non sunt
nomina non-existentium.
 The 'Sophismata' attributed to Marsilius of Padua
by Roberto Lambertini

Introduction

The fact of being the author of a masterpiece in political thought, such

as the Defensor Pacis was enough to establish Marsilius as one of the most
interesting political philosophers of the later Middle Ages. The heated
discussions the Defensor Pacis provoked also had the effect, however, of
putting other works of his in almost perennial shade. This is the case, for
example, with other political treatises such as Defensor Minor which, in
spite of its intrinsic relevance, survived in only one copy,! The destiny of
his activity in logic and metaphysics was no better. We are informed of his
presence at the Parisian Art Faculty where he enjoyed the title of magister
and was also rector in the first months of 1313. He came back to Paris
after an Italian parenthesis (to be placed approximately between 1315 and
1319) during which he began his activity on behalf of the Italian
Ghibellines. His Defensor Pacis was in fact finished in the French capital,
on 24th June 1324; a few months later he left that city, together with one of
the outstanding figures of the Parisian Arts Faculty, John of Jandun, to
join the Emperor in his struggle against the Papacy.2

In spite of these proved connections with the Arts Faculty, we only

know of three works which could be related to his activity there. All these
suffer from a very poor manuscript tradition. This is the case, for example,
with a fragmentary questions commentary on Aristotle's Metaphysics
(covering only the first six books) attributed to him, contained in only one
Florentine manuscript.3 In search of the philosophical background of
Marsilius' political theory many scholars have focussed their attention on
this work, providing us with editions of some questions and trying to
establish connections with his political theory.4

lCf. Marsile de Padoue. Oeuvres mineurs. Defensor Minor. De translatione imperii, ed.
C. Jeudy and J. Quillet, Paris: Editions du Centre National de la Recherche Scientifique
1979.
2For biographical data, see C. Pincin, Marsilio, Torino: Giappichelli 1967; for further
infonnation see C. Dolcini, Marsilio ed Ockham, in idem, Crisi di poteri e politologia
ill crisi, Bologna: Patron 1988, pp. 291-426.
3Firenze, Biblioteca Mediceo-Laurenziana, Fiesol. 161, ff. Ira-4Iva. For references to
works taking this Commentary into consideration, cf. R. Lambertini and A. Tabarroni,
"Le Quaestiones super Metaphysicam attribuite a Giovanni di Jandun. Osservazioni e
problemi", Medioevo 10, 1984, pp. 41-104.
4Cf. L. Schmugge, Johannes von Jandun, 1285/89-1328. Untersuchullgen zur
Biographie und Sozialtheorie eilles lateillischen Averroistell, Stuttgart: Hiersemann
1969, pp. 95-119; J. Quillet, "Breves remarques sur les Questions super Metaphysice
Iibros I-VI (Codex Fesolano 161, ff. Ira-41va) et leurs relations avec I'aristotelisme
Mterodoxe", in Die Auseillandersetzungen an der Pariser Universitiit im XIII.
Jahrhundert, ed. A. Zimmermann, Berlin-New York: De Gruyter 1976 (Miscellanea
Mediaevalia 10), pp. 361-85.

86
 MARSIUUS DE PADUA 87

Two other works attributed to him, namely two sophismata, have so

far been neglected. 5 In the present paper I will focus my attention on these
sophismata, providing basic information regarding the manuscript tradition
and a first attempt at locating them in the logical and semantical discussions
of that time.
1. The sophism on concrete accidental terms

The first sophism I will deal with, known as Caius est universale, is
contained in Vat. lat. 6768, ff. 221 vb-223va (to be referred to as V2), the
manuscript studied by Anneliese Maier which conserves so many works of
so-called A verroist philosophers.6 It is devoted to the much debated
question of the meaning of concrete accidental terms.

Before starting a presentation of the position held in this sophisma we

are confronted with a preliminary problem arising from the fact that in Vat.
lat. 3061, ff. 29rb-31rb (VI) another sophism is conserved, which
presents striking similarities to our text, but is not attributed to Marsilius.
There it is ascribed to an otherwise unknown Petrus de Colonia} As far
as I know, we do not possess any information about this author: from the
fact that his only work is part of a collection of sophismata where authors
like Radulphus Brito or Petrus de Insulis are mentioned, one could
tentatively suggest, as a date, the period of time of the academic activity of
these two philosophers. Recently, Irene Rosier has indirectly corroborated
this hypothesis, showing that Gauthier d' Ailly, an unknown author whose
sophismata also appear in the same collection, reveals strong connections
with Radulphus Brito. No cogent reason has emerged, therefore, to cast
doubt on the idea of also placing Petrus de Colonia "au toumant des xme-
XIve siecles."8

5The only exception - to which Professor Maierii alerted me only after this paper was
finished - seems to be M. Grignaschi, "Reflexions suggerees par une derniere lecture du
'Defensor Pacis' de Marsile de Padoue", in Papers in Comparative Political Science.
Estudios de ciencia politica comparada. Trabajos en homenaje a Ferran Valls i Taberner,
ed. M.J. Pelaez, vol. XVI (s.l. et a.), pp. 4507-28, where at pp. 4514-17 Professor
Grignaschi - a specialist in Marsilius' political thought - deals briefly with these
sophismata.
6A. Maier, "Die italienischen Averroisten des Codex vat. lat. 6768", in A. Maier,
Ausgehendes Mittelalter, v.I1, Roma: Edizioni di Storia e Letteratura 1967, pp. 351-66.
For a complete description of the MS, cf. B. Faes de Mottoni and C. Luna, Aegidii
Romani Opera Omnia, I, Catalogo dei Manoscritti, 1/1, Citta del Vaticano, Firenze:
Olschki 1987, pp. 217-9.
7Cf. P. Glorieux, La Faculte des Arts et ses maitres au XllIe siecie, Paris: Vrin 1971, p.
280.
SCf. I. Rosier, "Un sophisme grammatical modiste de Maitre Gauthier d' Ailly", Cahiers
de l'lnstitut du Moyen-Age grec et latin 59,1989, pp. 181-232, esp. pp. 181-2. A very
similar opinion was shared by Martin Grabmann, to whom we owe the first evaluation
of this manuscript for the history of sophisms, in M. Grabmann, Die
Sophismataliteratur des 12. und 13. lahrhunderts mit Textausgabe eines Sophisma des
Boetius von Dacien, Miinster: Aschendorff 1940, p. 60. See also J. Pinborg, "A Note
on Vat. Lat. 3061", Bulletin de Philosophie Medievale 18, 1976, p. 78.
88 ROBERTO LAMBERTINI

This opens a difficult question about authorship. The text does not
offer any element that could strengthen one attribution against the other: the
discussion remains in fact at a theoretical level, without any historical or
geographical reference. We are left, therefore, with the ambiguous
evidence of the attributions at the end of both sophismata, and in the index
of this section of y2, which is at least partly independent of the attributions
contained in the single texts. 9 .

Through a collation of the two texts, however, one can collect clues
which seem to exclude at least the possibility that we are confronted with
two different works or even with two redactiones. V2, the one which
attributes the work to Marsilius, is much less accurate. The scribe (or
somebody else at a former stage of the transmission) did not take particular
care in understanding what he was copying; the result is a high number of
misinterpretations, which in various passages rule out the possibility of
reconstructing the argumentation. In some other cases, this habit simply
has weird effects: in a sophism devoted to concrete accidental terms the
copyist writes contractus instead of concretus for at least half of the entire
work. Moreover,.almost a dozen significant omissions per homoioteleuton
can be ascertained in V2. VI offers a text which has been composed more
carefully. While the scribe of V2 seems on many occasions to reproduce
what he thinks he reads, no matter what it can mean, that of VI's aim is to
give a text provided with a plausible sense and an acceptable level of
syntactical accuracy.1O All this results in two texts which do not depend
directly on each other (both include passages which are omitted - mostly
through oversight - by the other) and also present a high number of
variants, but these can probably be explained without resorting to the
hypothesis of different reportationes. A substantial unity is always
preserved: although one can notice especially in the first part certain
differences in phrasing and diverging stylistic preferences, II arguments,
counterarguments and difficultates, both in thedisputatio and in the
determinatio, are presented in exactly the same order. This complete
coincidence in structure is extended also to features and errors which are
unlikely to appear in two different texts independently. This emerges
clearly in the last part of the sophism, the solutio rationum. Here not only
is the order of the answers changed with respect to that of the rationes in
exactly the same way, but we also find in both VI and V2 the same solutio

9See Maier, "Die Averroisten", op. cit., pp. 357 and 362.
IOThis is the reason why I have decided to base the following quotations on vI,
indicating all variants of v2 , except simple inversions. For some examples of V2 ,s
lack of accuracy: " ... non de significato cui fit impositio sed (cui...sed] cum sit ipso
V2 ) de significato a quo fit impositio (imposita V2 )"; "Item substantia (sensus V2 )
precedit accidens tempore; subiectum est prius accidente (antecedente V2 )..."; " ...et
tamen diversa sunt possunt esse in intelligendo (possunt. ..in] sunt post esse V2 ) .••. "
lIE.g. where Vi (f. 29rb) reads: "Ad secundam quando dicebatur quod esset nugatio,
dicebatur quod non oporteret," V2 (f. 221vb) has: "Ad secundam rationem quando
dicebatur quod terminus concretus causaret nugationem, ipse negabat." vI (f. 30ra):
" ... quia nomen significatione una non significat nisi unum;" V2 (f. 222va): " ... nisi
unum nomen plura significaret ut unum." Moreover, V2 usually prefers more concise
expressions.
 MARSILIUS DE PADUA 89

of a ratio which does not exist in any of the two texts. 12 I think, therefore,
that we are entitled to consider VIand V2 - the high number of variants
notwithstanding - as copies of one and the same work.

The question remains open, which attribution should be considered

wrong, although one cannot exclude, in principle, that one of the two
authors has misappropriated the sophism, a fact which was not unusual for
the Faculty of Arts at that time.13
2. The comparison with Fiesol. 161

Because in the Florentine commentary on the Metaphysics attributed to

Marsilius we find a quaestio bearing the title Utrum terminus concretus
accidentalis significet primo formam,14 it seems suitable to try a
comparison between the doctrines held in the two texts. But, as we shall
see, the result will open more problems than it can solve. In fact sophisma
and quaestio disagree on many crucial points of the debate about concrete
accidental terms. 15

First of all, their main tenets are deeply divergent; the Florentine
quaestio clearly states that a concrete accidental term signifies both form
and subject, adding that the form is signified per prius, and the subject
secondarily.1 6 This solution strongly resembles that of Siger of Brabant
and Siger of Courtrai, although it lacks the specification according to
which the significatio of both form and subject happens sub propriis
rationibus)1 The main part of the sophisma is devoted to the confutation
of this theory; 18 the whole sophisma can be considered as a criticism of

12Cf. VI, ff. 29va and 3lrb; v2, ff. 222ra-223va.

13See the eXaIpple of John of Jandun according to the abstracts of C. J. Ermatinger's
contributions to the first and fifth Saint Louis Conferences, Manuscripta 19, 1975, pp.
72-3 and 23, 1979, p. 7.
14MS Firenze, Biblioteca Mediceo-Laurenziana, Fiesol. 161, f. 30ra-b. A description of
this MS in C. Piana, "Nuovo contributo allo studio delle correnti dottrinali
nell'Universita di Bologna nel sec. XIV", Antonianum 23, 1948, pp. 22 I -54.
15For the identification of the different positions in the debate constant reference will be
made to S. Ebbesen, "Concrete Accidental Terms: Late Thirteenth Century Debates
about Problems relating to such Terms as 'album"', in Meaning and Inference in
Medieval Philosophy, ed. N. Kretzmann, Dordrecht: Kluwer Academic Publishers
1988, pp. 107-74.
16Marsilius (?), Quaestiones super Metaphysicam, 1. V, q. 6 (MS Firenze, Biblioteca
Mediceo-Laurenziana, Fiesol. 161[= F], f. 30ra): "Per hoc dicendum ad questionem:
nomen concretum duo significat, id est subiectum et formam vel compositum ex hiis
quod habeo pro eodem." This statement is rather puzzling, if we consider that the two
formulas seem to belong to different traditions. But in the following the author
specifies: ..... terminus concretus duo significat; hoc non est qUQcumque modo, sed
primo significat formam sed, quia forma accidentalis non potest esse sine subiecto, ideo
dat ex consequenti intelligere subiectum."
17Ebbesen, "Concrete Accidental Terms", p. 118.
18The sophisma squarely rejects the thesis maintained in the quaestio. VI, f. 29vb: "Et
ideo (et ideo] Item V 2 ) alii aliter dicunt sic: quod terminus accidentalis concretus
(contractus V2 ) significat utrumque, ita tamen quod (significat...quod om. V2 ) prius
significat formam (prius ...formam] significet formam per prius V2 ) et ex consequenti
subiectum.. ,. Sed ista opinio non valet, nam (quia V 2 ) si terminus concretus
90 ROBERTO LAMBERTINI

this view and as a defense of the thesis according to which concrete

accidental terms primarily signify the aggregate consisting of form and
subject, sub ratione forme; in fact, the discussion of the 'primo formam,
deinde subiectum' makes up the the main part of the disputatio and of the
solution of difficulties. The other two views, namely that held by Modists
like Boethius de Dacia and that attributed to Avicenna playa secondary role
in the actual discussion. 19

This disagreement is reflected also in the handling of certain arguments

which are typical of the debate about concrete accidental terms. So, for
example, the author of the quaestiones on the Metaphysics denies that a
term must necessarily be either denominative or analogous. 20 The very
same "argument from analogy", based on the claim that every term
signifying different entities according to some order is analogous, is used
by the author of the sophisma to demolish the view held by the quaestio.21

Moreover, while the quaestio maintains that it is possible for the same
term to be denominativus and analogus at the same time, this possibility is
rejected with energy in the sophisma. The fact that a denominative term
cannot at the same time be analogous is, in fact, one of the premises of the
arguments produced by the author of the sophisma. 22 As a matter of fact,

accidentalis significaret primo fonnam et ex consequenti subiectum vel hoc (totum add.
V2) esset significatione una vel pluribus (plures V2 ). Nec (non V2) potest dici quod (sit
add. V2 ) significatione una, quia nomen significatione una non significat nisi unum
(quia... unum] nisi unum nomen plura significaret ut unum V2 ). Item (om. V2 ), nec
pluribus significationibus (om. V2 ) quia tunc non esset tenninus (nomen V2) unus sed
equivocus et analogus: ergo non potest dici quod primo significet fonnam et ex
consequenti subiectum." As a matter of fact, the quaestio uses the expression 'dat
intelligere subiectum' and not 'significat'. I would doubt that in this context the
difference can be relevant; at any rate, the sophisma also criticizes a different opinion
which appeals to the 'dare intelligere'; cf. vi, f. 29vb: "Dicunt enim quidam quod
terminus concretus (contractus V 2 ) accidentalis significat tantum fonnam et non
subiectum sed ex consequenti dat intelligere subiectum... Sed ista opinio non valet...."
19Ebbesen, "Concrete Accidental Tenns", pp. 118-29. Vi, ff. 29vb-30ra.
20Marsilius (1), Quaestiones Met., I. V, q. 6, (F, f. 30ra): "Et dubitatio est hic: si vero
denominativum primo significaret subiectum deinde accidens, tunc esset analogum.
Hoc <est> inconveniens quia hoc nullus diceret; consequentia patet quia iIIe tenninus
est analogus qui significat plura per ordinem... .Ideo aliter dico aliquem tenninum esse
analogum et denominativum simul [esse] potest intelligi dupliciter: primo quod
tenninus analogus esset denominativQs ita quod analogia sit /f.30rb/ denominatio; hoc
falsum est. Secundo quod eidem voci accidat analoga [pro; analogia et] denominatio; sic
non est inconveniens. Unde non est inconveniens eandem vocem ilia duo in se
congregare secundum diversas rationes ...."
21 vI, f. 30ra: "Item, quando aliqua dictio (duo v 2 et add.: plura) significat ordine
quodam, scilicet prioris et posterioris, talis est analoga, ut dicitur 40 metaphysice; si
ergo tenninus concretus significaret fonnam et subiectum ordine quodam, sequeretur
quod esset analogus."
22Vl f. 30ra: ..... si (sic V2) ergo tenninus concretus significaret (significat V2) fonnam
et subiectum ordine quodam et per diversas rationes, tunc erit (esset V2 ) tenninus
analogus, quod falsum est, quia Philosophus in Predicamentis dat differentiam inter
tenninum equivocum, sub quo continetur tenninus analogus, et etiam univocum et
tenninum (etiam ... tenninum] inter tenninum univocum et V2 ) denominativum seu
concretum."
 MARSILlUS DE PADUA 91

one could point out that the quaestio on concrete accidental tenns shows a
much deeper affinity to John of Jandun's writings on the same subject,
than to the views of the sophisma. The problems regarding the relationship
between Jandun and the Metaphysics-commentary contained in Fiesol. 161
are too complex to be discussed here. 23 At any rate, Jandun also supports
the "first fonn, then subject" theory with arguments which are similar,
even in wording, to those of the Florentine question. He even shares the
view according to which a denominative tenn can also be analogous and
equivocal at the same time, at least in a certain sense. 24 The Florentine
quaestio therefore shares Jandun's views on the subject, while the sophism
seems to maintain the opposite, at least as far as the main conclusion is
concemed.25
3. Imposition and meaning of concrete accidental terms

In this sophisma the approach to the relationship existing between

reality and language is rather sophisticated. The exposition of the opinion
held by the magister starts by saying that the mirroring of the ontological
order in the language is never complete. In particular, the imposition of
names does not reproduce the order of things in themselves but rather our
way of understanding them, so that one should avoid drawing conclusions
directly from the ontological order to the semantical one. 26 In the case of
imposition, moreover, one should consider that it has at least two
components, a fonnal and a material one. A name is imposed on something
by means of something, or as the sophisma puts it, one has to distinguish
between illud cui imponitur vox and illud a quo imponitur vox.27 The
fonner is the significatum in its proper sense, while the latter can also be
regarded as a significatum, although in an improper sense. Such a
distinction allows the author of our sophisma to interpret all the auctoritates
which seem to support the opposite view as if they were speaking of the
significatum in the second and "improper sense". For example, the much
quoted passage by A verroes on Metaphysics which states that concrete

23Cf. Lambertini and Tabarroni, "Le Quaestiones", pp. 52-64.

24Johannes de Janduno, Quaestiones in XII libros Metaphysicorum, I. V, q. XV,
Venetiis: apud H. Scotum, 1553 (reprint: Frankfurt/M: Minerva, 1966), f. 64C: "Item
aliter dico quod aliquem terminum esse denominativum et analogum simul potest
intelligi dupliciter: primo modo quod terminus analogus est denominativus, ita quod
analogia sit denominatio: hoc falsum est. Secundo modo quod eidem voci accipit (pro:
accidat) analogia et denominatio; sic non est inconveniens, unde non est inconveniens
illam vocem ilIa duo in se congregare secundum diversas rationes, sicut in proposito
terminus denominativus sicut "album" et consimiles sunt. Et dicuntur analogi
inquantum plura significant secundum ordinem quendam, et hoc est de ratione
analogiae, sed dicitur denominativus inquantum significat formam accidentalem in
habitudine ad subiectum quem denominat."
25Cf. Jandun, Quaestiones in XII libros Metaphysicorum, I,V, q. XV, f. 64E:
"Intelligendum tamen, ut dicit Commentator, quod terminus concretus primo significat
formam, et secundario subiectum."
26Vl f. 30rb: "Quarto (ultimo V2 ) intelligendum est quod rebus absolutis (absolute V2)
voces non imponuntur, sed secundum quod nobis magis note sunt, unde impositio
vocis non sequitur ordinem rerum sed magis nostrum modum (nostrum modum] modos
V2) intelligendi."
27Cf. VI, f. 30va.
92 ROBERTO LAMBERTlNl

terms primarily signify the form can be understood as if by 'significatum'

he intended the ratio of signification, not its proper object. 28 Moreover,
when it is argued that forms must be signified primarily because they give
to the aggregatum its esse, the answer can be that the aggregatum is
signified by means of the form because of that ontological fact, but the
form itself is not the significatum primarium. 29

The position held in the sophism was shared by a large number of

authors, among them Simon of Faversham, Guillaume Arnauld (the
Pseudo-Giles of Rome), Radulphus Brito and Peter of Auvergne.3 o The
sophism I am presenting shows similarities to Peter of Auvergne's Album
potest esse nigrum;31 they defend the same solution, sharing many
arguments and, among them, the distinction between two senses of
significatum. But Caius est universale develops further the defence of his
solution against the argument from nugation and diverges from Peter on an
important point, that of the ontological status of accidents. 32

28 Vi, f. 31ra: "UIterius (ultimo V2) dubitaret aliquis quod illud quod nunc (iam V2)
dictum est sit contra intentionem Commentatoris 50 Metaphysice, qui dicit quod per
terminum concretum significatur primo (am. V2) forma; (item add. V2) videtur esse
contra Philosophum (intentionem Philosophi V2) in Antepredicamentis, qui dicit (qui
dicit am. V2) quod denominativa sunt que (sunt que am. V2) solo casu differunt, id est
sola cadencia. Si ergo sola cadencia differunt (Si...differunt am. V2), conveniunt (ergo
add. V2) quantum ad significatum ad minus principale (ad ... principale am. V2) ... Ad
omnia ista non contradicendo (contradico V2) ipsis (istis V2) sed magis pro ista parte
adducendo (adducuntur V2) considerandum est (considerandum est] considerando quod
V2) de significato - ut dictum fuit (primo add. V2) - possumus loqui dupliciter, uno
modo potest dici significatum principale cui vox principaliter imponitur (ad
significandum add. V2), alio modo dicitur (potest dici illud add. V 2 ) significatum
mediante, quo (una add. V2) vox imponitur ad significandum, seu (am. V2) a quo fit
vocis impositio. Si loquatur (Ioquamur V2) de significato principali secundo (isto V2)
modo sic significatum termini concreti accidentalis erit forma et isto modo intellexit
Commentator. ... "
29Cf. Vi, f. 30ra.
30Cf. Ebbesen, "Concrete Accidental Terms", pp. 117-8.
3lEdited in Ebbesen, "Concrete Accidental Terms", pp.162-74. The text of the responsio
ad quaestionem of our sophisma is not easy to reconstruct, due to textual accidents. VI
ff. 30rb-va: "Ad questionem duo dicenda sunt: primo quod per terminum concretum
significatur totum aggregatum, vel quod terminus concretus (non scr. VI sed del.)
significat subiectum et formam ordine quodam, sed primo et per se totum aggregatum
(vel. .. aggregatum] am. V2) ... Sec undo dicendum est quod terminus accidentalis
concretus (non add. V2) significat substantiam (subiectum V2) et accidens ordine
quodam, sed primo et principaliter significat totum aggregatum sub ratione forme ... "
In the second part I would be inclined to prefer the readings of v2. The presence of the
'non' seems in fact to be more consistent with other passages (see above, nn. 17 and
20-21) of the sophism. For this reason, one could suggest restoring the deleted 'non' of
Vi in the first part, which has unfortunately - apparently due to oversight - no parallel
in v 2 . It might seem that VI has preferred an easier reading, which however
"conflates" two positions and, in the end, blurs distinctions that are on the contrary
treated as relevant in the rest of the sophisma. At any rate, only a critical edition will
provide the evidence necessary to reconstruct this passage in the most plausible way.
32vI, f. 30rb: "Sed tertio considerandum est quod inherentia seu dependentia qua accidens
dependet ad substantiam est de essentia accidentis, et non solum inherentia habitudinalis
(aptitudinalis V2 ), sicut quidam dicunt, sed etiam accidentalis (actualis V2 )." For Peter
of Auvergne, who speaks of 'habitudo ad substantiam', but denies that inherence could
 MARSIUUS DE PADUA 93

As far as the argument from nugation is concerned, the author of our

sophisma is well acquainted with attempts also to use it in case of
signification under different rationes (a discussion on this point is part of
the disputatio) and is aware of the fact it cannot be enough to take refuge in
the assertion of the existence of different rationes in order to neutralize the
argument from nugation. One must be sure that the two rationes under
which a subject is understood and expressed do not reciprocally imply
themselves nor stay in any relationship of specification. For this reason
alone, the complex term 'corpus album' cannot be regarded as a nugatio,
as on the contrary happens for other complex expressions, such as homo
animal or rationale animal.33

As far as the ontological status of accidents is concerned, our sophism

explicitly states that they depend on a subject by essence, joining on this
point the view held by Boethius de Dacia. This does not imply, however,
that abstract accidental terms signify the subject in any way: the subject is
understood, but not signified by means of such terms as albedo. 34 It
would be interesting, at this point, to know whether Marsilius held the
same position regarding the inherence of accidents, Unfortunately, the
question concerning this point is usually placed at the beginning of the
commentary on the seventh Book of the Metaphysics, while the Florentine
commentary attributed to Marsilius ends with the sixth.

From our analysis, then, it emerges that the sophisma described here
represents an interesting contribution to the debate about concrete
accidental terms. Its defence of what Sten Ebbesen calls the URA theory
witnesses to a developed stage of the debate, where arguments have
reached a significant degree of sophistication. It deserves, therefore, closer
scrutiny. Unfortunately, it seems impossible to solve, at the present state
of knowledge, the problem about authorship: no hypothesis can be
excluded beyond any reasonable doubt; even the evidence of a deep
theoretical dissent between the sophisma and the quaestio contained in

be an essential feature of accidents, cf. Ebbesen, "Concrete Accidental Tenns", pp.142-

4.
33v l , f. 30vb: "Cum igitur sit genus aptum natum specificari et detenninari per
differentiam et ipsa (e converso V2 ) differentia (sit add. V2) apta nata specificare et
detenninare, ergo (om. V2) licet sint diversarum rationum, ideo si in diffinitione
preponatur differentia que est apta nata specificare et detenninare et postponatur genus,
quod est aptum natum specificari et detenninari, erit nugatio dicendo 'rationale animal',
ut apparet (patet V2) 50 Metaphysice. Si (ergo V2) igitur subieetum sub alia ratione
intelligitur per tenninum concretum et sub alia ratione exprimitur, quarum nulla
includitur in altera, nee altera est apta nata specificare nec detenninare (nec determinare
om. V2) alteram, nee e converso (e converso] specificari ab alia V2), non erit (est V2)
nugatio (ut add. V2) in oratione qua exprimitur subiectum cum termino concreto
accidentali."
34Cf. VI, f. 31 ra: .....licet subiectum non sit de significato tennini abstracti, tamen est
de intellectuipsius (om. V 2 ), quia accidens, quoquomodo (quocumque modo V 2 )
accipiatur, sine dependentia que est de eius essentia intelligi non potest et (de ...et]
essentia eius nec esse potest nee V2 ) per consequens sine subiecto quod est tenninus
dependentie. Unde aliqua (bene add. V2) sunt de intellectu alicuius que tamen non sunt
de eius significato."
94 ROBERTO LAMBERTINI

Fiesol. 161 is far from being conclusive, as reasonable doubts can be

raised about its reliability as a source for Marsilius' thought. 35

In this case too I find it wiser to include this sophism only tentatively
among the works of the Paduan master, at least until new discoveries
provide us with more facts.

4. The sophisma 'Omne factum habet principium' attributed to

Marsilius
The case of the second sophism is far better. To my knowledge, the
sophisma Omne factum habet principium which deals with the problem of
the number of universals, is known in only one copy and is clearly
attributed to Marsilius in the colophon.36 Although it does not facilitate the
restitution of a reasonably sound text, the uniqueness of the copy does not
seem to allow for many doubts about Marsilius' paternity.

In this text the style of the exposition is lively and conserves the
flavour of an animated discussion. Always speaking in the first person,
the author often allows himself observations which lend the text immediacy
and colour. At a certain point in the disputatio he clearly states that he has
omitted some arguments because they were not of much value;37 then, a
couple of times he refers to iuvenes who need special explanations and
exercises. Other expressions directly hint at the work of someone who is
trying to reconstruct a discussion, such as 'nihil dixit quod recolem', 'ad
hoc nihil dicebat' and the usual admissions not to have heard a part of the
debate because of the noise. In one case, he explicitly says that he has
missed the respondens' argument propter clamorem scholarium
venesanorum. The presence of a group of Venetian students that is
apparently big enough to disturb the discussion seems, however, to
contradict the information contained in the colophon, that the sophism was
held in Paris.38

35Cf. Lambenini and Tabarroni, "Le Quaestiones", pp. 52-64.

36MS Mantova, Biblioteca Comunale, 0.111.19 (445), ff.lvb-3ra (for shon, MS M). M,
f.3ra: "Explicit sophisma universalium (jortasse pro: utilissimum) datum (jortasse pro:
determinatum) Parigius a venerabili viro magistro Marsilio paduano." For a description
of this MS cf. H.V. Shooner, Codices manuscripti operum Thomae de Aquino, Roma:
Editori di San Tommaso, 1973, pp. 303-5. See also C. Piana, Nuove ricerche su Ie
Universitii di Bologna e di Parma, Quaracchi - FIorentiae: Typographia Collegii S.
Bonaventurae 1966, pp. 5-6.
37M, f. I vb: "Non... scripsi replicationes propter brevitatem et quia non sunt magni
valoris."
38According to P. Kibre, The Nations in Mediaeval Universities, Cambridge, Mass.:
Mediaeval Academy of America 1948, it is only for some Italian Universities that we
possess evidence that Venetian students were considered as an independent group; a
Venetian natio is recorded in the Statutes (1331) of the Jurists in Padua (p.1l7). As a
matter of fact, the clue is so uncenain that one could be tempted to amend, as Sten
Ebbesen suggested, 'venesanorum' in 'vesanorum'. On the other hand, if the
hypothesis of an Italian provenance could be corroborated, we would have here - as
Alfonso Maierii pointed out - the first example of a sophisma disputed in Italy, at least
until the date of the Italian activity of Jordanus de Tridentia can be ascenained. V2
 MARSlLlUS DE PADUA 95

S. Marsilius' criticism of previous views

At the beginning, through a complicated explanation which is centred

around the concept of principium, Marsilius connects the proposition
'Omne factum habet principium' with a traditional question in
commentaries on Porphyry, namely, whether universals are five and not
more or less. After the presentation of the usual arguments pro and contra,
that are almost ritually presented and then rejected in several questiones de
sufficientia universalium, the respondens presents the solution which
connects the number of the universals to the different ways of predicating
of many: genus and species are related to two different ways of predicating
in quid; differentia derives from predication in Quale essentiale, while
predication in Quale accidentale can be divided in two, depending on
whether it is predicated per se of the principles of its subject or not. The
first alternative generates the proprium, the second the accidens. Deriving
universalia from the modes of predicatio de pluribus was commonly
accepted. Simon of Faversham39 and Gentilis de Cingul040 adopt it.
Radulphus Brito shows himself to be acquainted with several versions of
this distinction and Marsilius himself refers to it as the via omnium
modernorum.41

The answer of the respondens is attacked at once by someone who

suggests considering not five, but six universals: if we divide predication
in quid into two modes, and also predication in Quale accidentale into two
modes, why should we not also introduce two modes of predication in
Quale essentiale? All in all, some differentiae are said of things which
differ by species, while others are predicated of things differring only by
number; in the case of the predicatio in quid, this is the origin of genus and

contains, in fact, at ff. 174ra-176va Jordanus' sophisma Utrum dimensiones sint

aeternae in materia. Cf. also A. Maier, "Die Averroisten", p. 358. For a study of
methods of teaching in Italian Universities see A. Maieru, "L'insegnamento della
logica a Bologna nel secolo XIVe il manoscritto Antoniano 391", in Rapporti tra Ie
Universita di Padova e Bologna, ed. L. Rossetti, Trieste-Padova: Lint 1988, pp.I-24;
A. Maierii, "Gii atti scolastici nelle universita italiane", in Luoghi e metodi di
insegnamento nell'ltalia medioevale (secoli XII-XIV), Napoli: Congedo 1989, pp.
249-87; A. Maieru, "Methods of Teaching Logic during the Period of the Universities",
in Medieval Schools and the Foundation of Literacy, ed. J.J. Murphy, forthcoming.
39This is what one can grasp from the very short text of Simon de Faversham's,
Quaestiones super libro Porphyrii, q. 6, edited in Opera omnia, vol. I, Opera logica, t.
I, ed. P. Mazzarella, Padova: Cedam 1957, pp. 27-8. Martin of Dacia also seems to be
close to this tradition: cf. Martinus de Dacia, Quaestiones Porphyrii, q. 2, in Opera, ed.
H. Roos, Copenhagen: Gad 1961, pp. 123-6.
40Gentilis de Cingulo, Quaestiones Porphyrii, q.5, MS Palermo, Biblioteca Comunale,
2 Qq. D. 142, f. 75va-b. A short extract from Peter of Auvergne's quaestio on the
subject can be read in J. Pinborg, "Petrus de Alvernia on Porphyry", Cahiers de
l'Institut du Moyen Age grec et latin 9, 1973, p. 51.
41Cf. Radu1phus Brito, Quaestiones subtilissimae super Arte Veteri, q. 10, ed. F.
Macerata, Venetiis: per J. et A. Rubeum [circa 1499] (H3990). Cf. also Scotus, who
prefers this way to other possibilities: Johannes Duns Scotus, Super Universalia
Porphyrii quaestiones acutissimae, q. 12, in Opera omnia, I, ed. L. Wadding, Lugduni:
Durand 1639 (reprint: Hildesheim: 01ms 1969), pp.95-6.
96 ROBERTO LAMBERTINI

species. It appears very soon that the respondens is not able properly to
answer such an objection.

He tries to escape by objecting that the difference is not predicated

primarily of the supposita, but only mediante genere vel specie. For this
reason - we can understand - the distinction between differentia generalis
and specialis cannot cause two different kinds of universal.

His attempt is frustrated at once, however, by a double counter

objection. Firstly, starting from the proposition 'Homo est sensibilis' it is
asked why here 'sensibilis' is considered a universal. Four possibilities are
listed:

a) in virtue of its being predicated of many absolute.

b) in virtue of its being predicated of many differing specie.

c) in virtue of its being predicated of many differing numero.

d) in virtue of its being predicated of many differing either specie or

numero.

The first possibility is dismissed because such a condition applies

indifferently to all universals, and it is improper that the same characteristic
individuate at the same time a genus and one of its species. The third and
the fourth are excluded in virtue of an admission made by the respondens
himself at the beginning of the discussion. Having there rejected the
possibility of considering the existence of as many universals as
predicamenta, he had stated that universals must be derived from different
ways of predicating de suppositis. This further condition rules out,
according to his adversary, the third and fourth alternatives, because it
seems reasonable that a differentia generalis does not regard its supposita
directly.

But all this leaves us with only one answer: 'sensibile' is a universal
because it is predicated of many supposita differing by species. Now, if
we tum to a proposition such as 'homo est rationalis' we have without
doubt a differentia which surely cannot be individuated by "being
predicated of many differing specie". It follows, then, that there are two
species of differentia, and the total number of universals amounts to six,
not to five.

Secondly, the respondens is accused of contradicting himself; as we

have seen, in order to counter the argument from the ten predicamenta, he
had added to 'predicari de pluribus' the specification 'sicut de suppositis'.
Then, to refute the argument in favour of the two species of differentia, he
had said that the differentia is predicated first of genus and species and
then, through them, of the supposita. The differentia risks, therefore,
being excluded from the number of universals. The same happens to
proprium.

At this point of the disputatio Marsilius clearly shows his distrust of

the logical capabilities of the respondens. As a matter of fact, either
 MARSILIUS DE PADUA 97

because of the noisy students, or because of bad memory, he is not able to

reconstruct much of the respondens' defence, but is also persuaded that it
could not have been particularly convincing.42

But this may depend not on the personal weakness of the respondens,
but rather on the difficulties intrinsic to the traditional position. Starting his
own determinatio, Marsilius considers the via communis itself to be
disproved during the disputatio. 43 The discussion of the problem must
start therefore from the beginning. Albert the Great is the first to be
challenged. It is not necessary to reconstruct the refutation of his own
explanation of the fact that universals must be five, no less and no more. It
is enough to remember that Marsilius expresses perplexities towards the
central assumption of Albert's distinction, that is, the opportunity to use
the opposition principium/principiatum to justify the existence of five types
of universal. Albert had suggested considering genus and differentia as
principia, species, proprium and accidens as principiata. 44 Marsilius finds
that the characteristic of being principles seems to be common to all
universals, and not limited to some of them, while 'being principiata' is
rather an accidental feature of universals. To sum up, this distinction seems
to contradict the Aristotelian principle according to which such a division
of the universal into its species must follow not accidental but essential
features of the universal itself. If we now define the universal as esse
innatum in pluribus, it is difficult to see how being principium or
principiatum can be considered an essential feature of it; there are in fact
many principles which are not innata in pluribus.45
6. Marsilius' solution: universals are not Jive, but two

Mter omitting a refutation of the Platonist opinion,46 Marsilius has to

present his own solution. This does not consist in a new justification of the
fact that the types of universal must be exactly five, which could substitute
for those proposed by Albert the Great or by the moderni. Nor does he
intend to accept the idea that universals are six instead of five, either. In his
opinion, universals are only two.

42Por the disputatio, see M, ff. 1vb-2ra.

43M, f. 2ra: "<O>e ista questione semper fuit ambiguitas quantum ad modum
distinguendi species universalis; in conclusione tamen omnes concordant usque nunc,
sed diversimode. Quidam enim distinguunt vel dant causam numerationis eorum per
modum quem dedit respondens et est via quasi communis omnibus modemis. Sed ilia
via sufficienter est reprobata in disputatione. Verum est quod etiam ex positione sive
conclusione quam tenebo destruetur una cum omnibus aliis."
44Cf. Albertus Magnus, Liber de praedicabilibus, tr. II, c.lX, in B. Alberti Magni Opera
que hactenus haberi potuerunt, ed. P. Jammy, Lugduni: C. Prest et socii 1651, t.I, p.
25a.
45Cf. M, f. 2ra-b.
46M, f. 2rb: "Opinio platonica fuit quod essent separata. Sed hanc omitto propter
brevitatem, quia sic(?) earn reprobat magister in 7° Metaphysice, quia si etiam
ponerentur universalia separata secundum esse, nihil refert ad positionem meam et si
referat quantum ad positionem aliorum." I have not yet succeeded in discovering to
which magister Marsilius is here referring.
98 ROBERTO LAMBERTINI

The premises of his position are in fact foreshadowed in his critique of

Albert's views, but he prefers to introduce his demonstrationes with four
suppositiones which seem to him obvious for every sound intellect.

First, Marsilius repeats the definition of universal: esse innatum in

pluribus: to this must be added the clause 'sicut in suppositis'.

Secondly, he emphasizes that a genus has as many species as the

essential differences which divide it.

Thirdly, he asserts that everything divisible can be divided into what

differs either by form or by matter, which is the same as saying, into what
differs specie or numero.

According to the fourth and last suppositio, no substance, either first

or second, essentially implies anything accidental in its first intention.
Marsilius exemplifies this claim with the assertion: "Socrates is nothing
else but animality and rationality".

From such premises the exclusion of proprium and accidens from the
number of universals follows immediately; the universal is in fact the result
of an intellectual operation of abstraction from the supposita which possess
a similarity between them. But this is not the case for the supposita of
proprium and accidens which are supposita per accidens, because
particular substances happen to be capable of laughing or to be black: such
features are not included in their essence. 47 Marsilius exemplifies this with
albedo: it cannot be derived from its supposita per accidens such as man,
horse, stone, because man qua man is not necessarily white. 48

It is true that albedo possesses its own supposita per se, which are the
individual whitenesses. In this respect, however, albedo is not an accident,
but a species, just as colour is a genus. 49

We are left with the conclusion that only universals which derive from
the essence of their supposita and can therefore be predicated of them are
universals in the proper sense. How many are they? According to

47 M, f. 2rb: "Hiis presuppositis tamquam manifestis cuilibet menti non perturbate

formo aliquas rationes infallibiliter ostendentes conclusionem premissam, quarum
prima talis est: ... omne universale debet abstrahi et colligi ex suppositishabentibus
similitudinem ad invicem unde talia, ex quibus sumi possit ratio universalis; accidens
et proprium non sunt huiusmodi respectu suppositorum per accidens, ergo et cetera. Et
intelligo per accidens supposita que sunt extra essentiam universalium et e converso."
48M, f. 2rb: ..... illa, homo vel equus, unde talia non plus includunt albedinem quam
nigredinem; aut igitur ambo aut nullum. Non est dicere quod ambo quia tunc contraria
simul essent, ergo nullum; [sic] sic ergo in eis nihil est, ut sic, unde sumatur [ratio]
ratio universalis ... "
49M, f. 2va: " ... respectu suorum suppositorum tamen respiciunt immediate determinatas
pluralitates, puta color albedinem et nigredinem, que specie differunt, et ut sic reponitur
sub specie universalis que dicitur genus; et albedo respicit immediate hoc album vel
illud, que solo numero vel materia differunt, et ut sic reducitur ad speciem universalis
que dicitur species."
 MARSILIUS DE PADUA 99

Marsilius, universale is esse innatum in pluribus, but there are only two
ways of "esse in pluribus", that is "in pluribus dijferentibus specie" and
"in pluribus dijferentibus numero": only species and genus can therefore
be considered as universals. 50 Even dijferentiae can be reduced to these
two, as far as they are predicated essentialiter of their supposita -
dijferentiae essentiales will be genera, dijferentiae specijicae will be
species: in fact, Marsilius remarks, the distinction between quale and quid
is not relevant for universals. 51
7. Marsilius and Jandun

Marsilius presents his position as a revolutionary change which attacks

a doctrine which dates back to Porphyry, and implicitly connects possible
resistance towards it with attachment to the tradition.52 He considers it a
radically new perspective he has opened, or better, to use his own lively
metaphor:Jenestra quam aperui.53

It is not easy to identify the philosophical context of Marsilius'

reduction of universals to what can be abstracted from the essence of the
supposita and can therefore be predicated of them essentialiter. As far as
the argumentation of the sophisma is concerned, it sounds to me rather
more an assumption than a conclusion. That is, the banning of predication
regarding supposita per accidens from what is relevant for the constitution
of a universal seems more likely to be the starting point of the whole
argumentation. 54 Marsilius presents it as self-evident, but I am persuaded

50M, f. 2va: .....ergo relinquuntur solum universalia dicta essentialiter. Ex dictis itaque
sillogizare oportet: tot sunt species universalis quot sunt eius differentie divisive; sed
tantum sunt due, ergo tantum erunt due species universalis. Maior patet et (pro: ex)
secunda et tertia (seil. suppositione), quia differentia universalis est innatum esse in
pluribus, ut patet ex prima suppositione, scilicet de distinctione (pro: diffinitione 1)
universalis. Ulterius ex tertia habetur quod omne plurificabile vel divisibile vel actu
plurificatum et divisum in plura sicut in supposita est solum duobuscmodis, quia vel
secundum speciem sive formam vel secundum individuum vel materiam. Remanet ergo
ex hiis conclusio principalis, scilicet quod universalis species non skn>t nisi due:
quarum prima potest vocari genus, secunda species."
51 M, f. 2va: ..... prima (seil. species differentiae) reducatur ad primam speciem
universalis, quia eadem differentia specifica constituitur, puta per innatum esse in
pluribus forma differentibus; specifica (pro: secunda 1) vero, (supple: constituitur per
innatum esse in pluribus) substantiali materia <differentibus>, et reducitur +causa+ ad
secundam speciem universalis, unde quid et quale nihil diversificant quantum ad
distinctionem universalitatis, sed solum distinctio aptitudinis ad plura secundum
formam vel materiam ut visum est prius."
52M, f. 2vb: "Reliquum vero est difficillimum omnium quia videtur istud esse omnino
impossibile, eo quod eius oppositum videtur non solum pluribus sapientibus, ymmo
omnibus legentibus in hac questione, PUta Alberto, Boethio et Porphyrio: nam hee est
sententia patrum illorum, quod quinque sunt universalia, et omnes moderni sequuntur
eam in hoc."
53M, f. 2vb: .....et istud posset capere et videre quis per fenestram quam aperui, scilicet
quod universalia distinguuntur ex eo quod aliquid est plurificabile in plura aliqua, de
quibus predicantur solum realiter, et quod nullum plurificabile in aliqua secundum
accidens facit speciem universalis distinctam respectu illorum."
54A very promising perspective can be opened up, if - following a suggestion of Sten
Ebbesen's - one connects Marsilius' move to the problems arising in syllogistic on
100 ROBERTO LAMBERTINI

that many of the socii he invites to discuss his opinion would have had
serious objections.55 Nevertheless, I think that the style of argumentation
suggests at least an affinity with a line of thought which had in John of
Jandun one of his most famous supporters. Jan Pinborg has characterised
John's renewal of the theory of plurality of forms as an attempt to solve the
problem of the semantics of universals. This attempt consisted in doing
away with Radulphus Brito's apparentia or with Herveus' entia rationis
existing in the mind only obiective and stating a kind of one-to-one
correspondence between the semanticallevel and the ontological one.56
Substantial forms, constituting the real components of the essence of
beings, are the signijicata of universal terms. Actually, Marsilius' sketchy
description of the origin of universals is more suitable to this view than,
for example, to that of Radulphus, who through the machinery of
apparentia and modi essendi tried to explain the origin of different
universals from the one substantial form. The same could be said, in a
tentative way, also of Marsilius' attempt to obtain a perfect parallelism
between the universals and common natures existing in individuals.57 It is
likely that for this reason what was traditionally considered a universal but
caused trouble in this scheme had to be eliminated. But other, less essential
elements hint in the same direction, such as the preference for Grosseteste
and the interest in Albert the Great's views.

These are only very vague indications, and would remain too vague if
Jandun had not himself confirmed them. Commenting on the first Book of
De anima, he touches on the problem of the relationship between universal
and singular. Starting from the usual distinction between universale pro
intentione and universale pro re, he defines the second as quiditas apta nata
esse in pluribus. This is not particularly original; but he goes on to refer to
the same passages in Aristotle and A verroes used by Marsilius in the third
premise of his determinatio, in order to show that a plurality can be of only
two kinds, either secundum formam or secundum quantitatem. From this
two-fold division - Jandun continues - the two-fold division of the
universal derives, which can be genus or species. The medieval reader,

account of the supposita per accidens. Marsilius' main assumption could be seen as a
rather "radical" way to get rid of such difficulties, even though the Paduan philosopher
does not express such an intention. Cf. Ebbesen, "Concrete Accidental Terms", pp.
152-7, on the relation between the use of concrete accidental terms and syllogistic
theory. For another view of this problem, cf. R. Huelsen, "Concrete Accidental Terms
and the Fallacy of Figure of Speech", in Meaning and Inference, ed. N. Kretzmann, pp.
175-85.
SSM, f. 3ra: " ... rogo socios istud videntes non primo intuitu istud conari ad
destruendum, sed ad examinandum..."
56J. Pinborg, "A Note on Some Theoretical Concepts of Logic and Grammar", in idem,
Medieval Semantics. Selected Studies on Medieval Logic and Grammar, ed. S.
Ebbesen, London: Variorum Reprints 1984, p. 290: "Jandunus especially tries to found
all concepts, all the theoretical concepts of logic and grammar, as solidly as possible in
the essence of objects."
57Pinborg, "Some Theoretical Concepts", p. 290: "According to the other opinion,
which is corroborated by the authority of Grosseteste, the concepts are derived directly
from the essence of objects. Jandun adheres to this opinion, mainly for two reasons: he
cannot see how general and specific concepts can be derived from accidents ..." In the
sophism Omne factum habet principium Marsilius shares the same attitude.
 MARSIUUS DE PADUA 101

acquainted with Porphyry, could probably not help being surprised; for
him Jandun adds: "whether they can be more than two, is an interesting
object of investigation, but as for now I do not discuss the problem."58 I
do not know whether Jandun treated the subject elsewhere; but Marsilius
did and started from the very same premises.
Conclusion
I hope that future research will discover further texts related to
Marsilius de Padua's activity in the field of semantics and metaphysics. As
for now, the reliable basis for reconstructing his philosophical position is
indeed very slim. Thanks to the fact, however, that sophismata compel
their participants to expand on their views and their implications more than
do other genres, we are able to gain basic information about his
philosophical attitude. This particular sophisma, Omne factum habet
principium, provides us with elements which allow us to consider
Marsilius in the context of that trend in thought which, in the first decades
of the fourteenth century, tried to put together the heritage of Averroistic
discussions and the renewal of the doctrine of plurality of forms, exerting a
lasting influence on the universities of the continent. We have seen further
evidence in favour of a hypothesis already formulated by other specialists
in Marsilius' thought: that in metaphysics and semantics, as in political
struggle, Marsilius of Padua and John of Jandun shared the same ideals.59

Liguria 25, 144042 Cento (Fe), Italy

58Johannes de Janduno, Super libros Aristotelis de Anima, I. I, q. VIII, ed. M. Zimara,

Venetiis: apud haeredem H. Scoti, 1587, reprint: FrankfurtIM: Minerva 1966, c. 37:
"Universale vero pro re est quiditas apta nata, vel cui non repugnat esse in multis, ut
habetur 7. Metaphysicae. Et quia multitudo aut est rerum differentium secundum
formam, aut rerum differentium secundum quanti tatem, ut dicit Commentator 2.
Physicorum, commento 6, ita est duplex multitudo solum ad quam omnis multitudo
reducitur, scilicet multitudo secundum formam et speciem et multitudo secundum
materiam, ut est multitudo individuorum eiusdem speciei. Ideo universale, quod est
quiditas reperibilis in multis, est duplex, scilicet, quiditas reperibilis in multis
differentibus secundum formam, et ista est genus, et quiditas reperibilis in multis
secundum quantitatem differentibus, scilicet in multis, solo numero diversis et haec
dicitur species specialissima. Utrum autem sint plura universalia quam ista duo, bona
perscrutatio est, sed ad praesens non discutio, quia solum de istis duobus universalibus
ad propositam quaestionem loqui sufficit."
59Acknowledgements: The present work is the first result of research started during a
stay at Institute for Greek and Latin Medieval Philology at Copenhagen University
made possible thanks to the financial support of the Danish Ministry of Education and
the kindness of all the people working at IGLM. In particular, I am indebted to Sten
Ebbesen for his most valuable advice. I wish to express my gratefulness to Carlo
Dolcini, who initiated me to Marsilius, to Ovidio Capitani and Dino Buzzetti for their
constant encouragement, to Alfonso Maieru for his precious suggestions and, of
course, to my socii, Andrea Tabarroni and Costantino Marmo.
102 ROBERTO LAMBERT1Nl

List of manuscripts

Firenze, Biblioteca Mediceo-Laurenziana, Fiesol.161

C. Piana, "Nuovo contributo allo studio delle correnti dottrinali

nell'Universita di Bologna nel sec. XIV", Antonianum 23, 1948, pp. 221-
54.

Mantova, Biblioteca Comunale, D.Ill.19(445)

H. V. Shooner, Codices manuscripti operum Thomae de Aquino, Roma:

Editori di San Tommaso 1973, pp. 303-5.

Palermo, Biblioteca Comunale, 2 Qq. D.142

Citto del Vaticano, Vat. lat. 6768

B. Faes de Mottoni-C. Luna, Aegidii Romani Opera Omnia, I, Catalogo

dei Manoscritti, 1/1, Citta del Vaticano, Firenze: Olschki 1987, pp. 217-9.

Citto del Vaticano, Vat. lat. 3061

 The Sophism 'Omnis propositio est vera vel falsa' by Henry
Hopton (Pseudo-Heytesbury's 'De veritate et falsitate
proposition is ,)
by Alfonso Maieru

The 1494 Venetian edition of the works of William Heytesbury

includes a text referred to as a sophism,! but which is not part of the well-
known collection of sophisms by this author published in the same
volume. Curtis Wilson2 did not indicate any manuscript or other editions
of this text. In 1966, I myself used this text attributing it to Heytesbury.3
However, the same year, M. Markowski4 made it known that MS 621 in
the Biblioteka Jagiellotiska in Krak6w contained a copy of the text. But it
attributes the work to another writer,5 so the problem of authorship
presented itself. In 1982, this problem was explicitly formulated by L. M.
De Rijk, who resolved it by attributing the work to Henry Hopton: at the
same time, De Rijk6 pointed out that a further copy of the text existed in
MS 1017 in the Biblioteca Angelica, Rome.? We learn from A. B.
Emden,S that Henry Hopton was a fellow of University College, Oxford in
1357, of Queen's College between 1361 and 1367, Master of Arts, and by
1362, scholar of Theology.

The text we are concerned with here cannot be considered "a small
collection of sophisms," to use De Rijk's words,9 even if the Krak6w MS
contains sophisms 1-21 by Kilvington copied on folios 14vb-19ra, right in
the middle of our own text.!O The text attributed to Hopton, on the
contrary, consists of only one sophism, which according to the

ISee William Heytesbury, De veritate et falsitate propositionis. Venetiis: Bonetus

Locatellus 1494, ff. 183va-188rb; see f. 188rb: "Explicit sophisma magistri Guilelmi
hentisberi de veritate et falsitate propositionis vtilissimum."
2William Heytesbury. Medieval Logic and the Rise of Mathematical Physics. Madison:
University of Wisconsin Press 1960, p. 206.
3"U problema della verita nelle opere di Guglielmo Heytesbury", Studi medievali, 3rd.
series, 7, 1966, pp. 40-72.
""Jean Buridan est-ill'auteur des questions sur les 'Seconds Analytiques'?", Mediaevalia
Philosophica Polonorum 12, 1966, pp. 16-30, p. 30.
5Krak6w, Biblioteka Jagiellonska, MS 621, ff. 14ra-22vb; see f. 22vb: "Explicit
determinacio haptonis: Anno domini M. ccc. lxxxx. amen;" the date furnished is not
that of its composition, since the Sophismata of Kilvington that precede our text in
this MS, ff. 3va-13vb, bear the same date. This MS will be subsequently referred to as
K.
6See "Introduction", in Some 14th Century Tracts on the Probationes Terminorum.
o.
Martin of Alnwick F. M., Richard Billingham, Edward Upton and Others, ed. L.M.
De Rijk, Nijmegen: Ingenium Publishers 1982, p. *39*.
7ff. 7vb-14ra, see f. 14ra: "Explicit sophisma disputatum a magistro henrico de
hoptone." This MS will be subsequently referred to as A.
SA Biographical Register of the University of Oxford to A.D. 1500, Vol. II. Oxford:
Clarendon Press 1958, p. 960.
9De Rijk, Some 14th Century Tracts, p. *39*.
lOIn effect, MS K, ff. 3va-13vb, only contain the introduction "Ad utrumque dubitare
potentes" and sophisms 22-49 by Kilvington.

\03
104 ALFONSO MAIERU

subscriptions of the two MSS, II was disputed (i.e. made the subject of a
public faculty debate between a respondens and one or more opponentes
presided over by the holder of the scholastic act) and then determined (i.e.
defined from the doctrinal point of view) by master Hopton. If the
proposed attribution of authorship is correct, then it must be admitted that
sophisms were disputed and determined in the Oxford schools during the
14th century.

Before proceeding to an examination of the text, I must therefore

briefly review what we know today about sophisms at Oxford. I shall only
discuss two points of the text: first, the doctrine of compounded sense and
divided sense, which Heytesbury took up twice elsewhere, in order to
show from this important aspect that the author of the sophism is not
Heytesbury; and, secondly, the discussion of the truth of a proposition.

The B.A. curriculum at Oxford included a period of listening to

lectures on texts in the syllabus (which, according to the regulations of
1268, covered texts on grammar and philosophy as well as logic) and,
after the first two years, the student had to take part in disputations, firstly
as the opponens, and then as the respondens. The disputations were
initially logical sophisms, and later questions or problems.l 2 After
attending the prescribed courses and participating in disputations on
sophisms and questions, the student was admitted to the bachelor degree
(according to J. M. Fletcher,l3 "a part of the requirements" expected from
the student was respondere de quaestione) and he could determine during
the following Lent.

Recently discussions have taken place not only on the nature of the
above-mentioned disputations, but also on the place where the disputation
was held (that on sophisms in parviso and that on questions in scolis
presided over by the master). E. D. Sylla must be credited with taking up
this discussion, claiming that a comparison between the statutary
provisions of 1268 and those of 1409 "leads to the conclusion that the
disputations in parviso were either identical with the disputations de
sophismatibus or else took their place."14 She disagrees with J. A.

II See above, nn. 5 and 7.

121 have dealt with this terminology in "Methods of Teaching Logic During the Period
of the Universities", forthcoming.
I3"The Teaching of Arts at Oxford, 1400-1520", Pedagogica Historica 7,1967, pp. 439-
41; "The Faculty of Arts", in The Early Oxford Schools, ed. J. I. Caito, Oxford:
Clarendon Press 1984 (The History of the University of Oxford, I), pp. 380-1; "Some
Problems of Collecting Terms used in Medieval Academic Life as Illustrated by the
Evidence for certain Exercises in the Faculty of Arts at Oxford in the Later Middle
Ages", in Actes du colloque: Terminologie de la vie intellectuelle au moyen dge, ed. O.
Weijers, Tumhout: Brepols 1988, p. 50.
14"The Oxford Calculators," in The Cambridge History of Later Medieval Philosophy,
ed. N. Kretzmann et al., Cambridge: Cambridge University Press 1982, p. 543 n. 10;
see also ''The Fate of the Oxford Calculatory Tradition", in L' homme et son univers au
moyen dge, II, ed. C. Wenin, Louvain-la-Neuve: Editions de I'Institut Superieur de
Philosophie 1986, pp. 692-8.
 'OMNIS PROPOSITIO EST VERA VEL FALSA' 105

Weisheipl,15 in whose opinion the disputation de sophismatibus was a

"disputation of logic," whereas the disputation de quaestione was a
"disputation of natural philosophy", and concludes: "I am inclined to
believe that the responses de quaestione are to be identified not by their
subject matter as much as by their location; in particular, 1 think that the
responses de quaestione at issue may be the ones that occurred in the
schools."16 On the other hand, Sylla is convinced that the disputations de
sophismatibus and the disputations in parviso cannot be claimed to be
absolutely identical: "There were certainly disputations de sophismatibus
not in parviso - for instance in the 'determinations' of new bachelors in
Lent and probably also in disputations connected with the ordinary lectures
on logic."17

Alain de Libera shares Sylla's distinction between 'disputatio in

parviso' and 'disputatio in scolis', and thinks that these are two
consecutive stages of the student's training in logic.l 8 However, he is
more wary in dealing with the opposition between 'de sophismatibus' and
'de quaestione', pointing out that it poses a problem both from a
geographical and an institutional point of view: 'sophistalquaesuonista' are
definitely qualifications that imply succession at Oxford, but it is not clear
what 'quaestionista' means (I have already mentioned Fletcher's opinion),
while from the geographical point of view, Paris institutionally requires
sophisms to be disputed in scolis.19

As far as 1 am concerned, 1 too am convinced that generally speaking

the distinction between sophism and question does not correspond to that
between such disciplines as logic and natural philosophy; and that it is
likely that these terms refer to consecutive stages of the training in logic20
(the quaestio, naturally, is also used in all three branches of philosophy).
But 1 am not convinced that the Oxford curriculum was so different from
the Paris one.

As regards 'parvisus', 1 recall, as does J. Fletcher,21 that the term may

refer to an actual meeting place, or have a metaphorical meaning, but it
might have something to do with the parve disputationes that a late
statutary provision (1584) relates to the public schools.22 However, 1
should now like to examine the 1268 ruling:

15"Curriculum of the Faculty of Arts at Oxford in the Early Fourteenth Century",

Mediaeval Studies 26,1964, p. 154.
16Sylla, "The Oxford Calculators", p. 543 n. 10; the author believes (p. 545) that the
ordinary disputes "were linked to the lectures," which does not appear to have been the
case.
17Ibid., p. 545 n. 12.
18A. de Libera, "La problematique de l"instant du changement' au XIIIe si~cle:
contribution l'histoire des sophismata physicalia", in Studies in Medieval Natural
Philosophy, ed. S. Caroti, Florence: Leo S. Olschki 1989, pp. 56-7.
19Ibid., pp. 57-9; on the discussion of sophisms in scolis in Paris see also Sylla, "The
Oxford Calculators", p. 545.
20Maieril, "Methods", nn. 82-6.
21"The Teaching", pp. 431-2, and "Some Problems", p. 48.
22See Statuta antiqua Vniversitatis Oxoniensis, ed. S. Gibson, Oxford: Clarendon Press
1931, p. 435. 6-10.
106 ALFONSO MAIERU

"Et sciendum quod si prius (11) responderint in scolis, publice de

sophismatibus per annum (12) integre debent respondisse, ita quod
nulla pars illius anni in (13) quo de questione responderint in dicto
anno integro computetur. (14) De questione debent respondisse ad
minus in estate precedente (15) quadragesimam in qua sunt
determinaturi. Si autem de sophis(16)matibus publice non
responderint, omnes libros predictos iurent (17) se audisse, hoc
adiecto, quod bis audierint librum Posteriorum. (18) Debent eciam in
audiendo maiorem moram fecisse quam si de (19) sophismatibus
publice responderint.''23

After mentioning the lectures on logic, grammar and philosophy that the
bachalarii determinaturi had to attend, the text distinguishes between those
who have responded publice de sophismatibus (line 11) and those who
have not responded publice de sophismatibus (lines 15-16). The former
had to do so for a whole year, so that the time in which they responded de
questione was not included in that year (lines 11-13; they had to respond
de questione the summer before the Lent of their determinatio: lines 14-15);
the latter had to listen to two readings of the Analytica posteriora and spend
more time attending the courses (lines 17-19). Respondere publice de
sophismatibus was therefore optional and could be replaced by other
exercises. The determinatio was also not compulsory and could be
replaced. 24 The fact that a student could avoid these two acts definitely did
not prevent sophisms from flourishing at Oxford, nor the university from
priding itself (1409) in the determination of Bachelors of Arts and
recognizing how this exercise enhanced mira sciencie logicalis subtilitas. 25
Moreover, the practice of disputing sophisms was also widespread in the
Oxford colleges. The statutes of Balliol (1282) prescribe that a sophism in
tum (circulariter) should be disputed every two weeks, ita ut sophiste
opponant et respondeant et qui in scolis determinaverint, determinent. 26

Let us now return to the 1268 text: the 'si prius responderint in scolis'
in lines 10-11 conflicts with 'si autem de sophismatibus publice non
responderint' in lines 15-16. Weisheipl takes lines 10-11 to mean that "if
an undergraduate wished to undertake responsions pro forma thereby
shortening his course of studies, he must answer objections in the logical
disputations for a whole year; and that year must be distinct from the
period of responding de quaestione.''27 If this were the case, however, the
'si prius responderint in scolis' in lines 10-11 should have been contrasted
with 'si prius autem non responderint in scolis'; but that is not the case.
The question does not seem to be avoidable here, and it comes into the text
primarily to establish the length of the respondere publice de sophismatibus
(per annum integre, in dicto anna integro). The editor, S. Gibson, has

23Ibid., p. 26. 10-19.

24Ibid., p. 34. 25-7.
25Ibid., p. 199.22-3.
26Ciled by O. Weijers, Termin%gie des Universites au Xl/Ie siecle, Rome: Edizioni
delI'Ateneo 1987, p. 351.
27See Weisheipl, "Curriculum", p. 155.
 'OMNIS PROPOSITIO EST VERA VEL FALSA' 107

placed a comma in line 11 after 'in scolis' that has no reason to be there,28
and so, in my opinion, lines 11-13 should read: "si prius responderint in
scolis publice de sophismatibus, per annum integre debent respondisse".29
If this reading is correct, the discussion of sophisms with undergraduates
participating took place in scolis. And if 'parvisus' is to be understood as a
"place outside the schools," the distinction between locations might imply a
distinction between the level of difficulty and the official or ceremonial
nature of the exercise, not a distinction between type of disputation.30 I
agree with Sylla in believing that this 1268 provision seems to correspond
to the one of 1409:31 only I would insist that the specification 'parvisum
interimfrequentantes', etc. in the 1409 statutes should not be understood
as the description of everything the arcista generalis had to do, and
therefore I would not exclude his active participation in the disputation in
scolis, which is not explicitly mentioned here.3 2 (Furthermore, Sylla is
generally cautious, as we have already seen.)

If this is the case, the first part of the Oxford curriculum does not
differ a great deal from the Paris one, which includes the disputation of
sophisms in the schools, distinguishing, as at Oxford, between three
distinct and consecutive stages: responding on sophisms, responding to
questions, and determining. As far as the existing literature is concerned, it
is known that the Paris tradition of sophisms disputed by masters in scolis
is well-documented. At Oxford, there is this determinatio attributed to a
master; it was an act that took place in the schools, and even if there are no
obvious indications of a live debate, I do not feel I can reject the evidence
provided by the two manuscripts that are so far known to exist. The
absence of explicit references to the debate and its participants
(opponentes, respondens) may be explained by the author's editing of the
text. But to my knowledge another determinatio of a sophism is to be
found in MS 92, 15th century, of Magdalen College, Oxford, ff. lr-6v,
and its structure resembles that of our text: Sophisma dei gratia et huius
venerabilis auditor;; benivol<enti>a determinandum est hoc: Nulla sunt
equivoca; after which it reads: Quod sophisma sit falsum satis patet. Quod

28The MS has no sign of any punctuation at that point. My thanks to Nigel Thompson
(Oxford) who examined the MS for me.
29/n scolis publice is also in Gibson, Statuta, p. 200. 35-36.
30See also Fletcher, "The Faculty of Arts", p. 379: "It can hardly be that two different
types of exercise are here [Le. in scolis/in parviso] indicated;" he intends to collocate
the exercises of the undergraduates in the schools.
31See the 1409 statutes in Gibson, Statuta, p. 200.9 and 13-17: "Presentati uero ad
determinandum ... iurabunt ... quod ante responsionem suam ad questionem ad minus
per annum arciste fuerant generales, paruisum interim frequentantes, et se ibidem
disputando, arguendo, et respondendo doctrinaliter exercentes." Fletcher, "The
Teaching", p. 434, compares arciste generales in this text to generalis sophista in a
15th century text (Gibson, Statuta, p. 580.34), while in "The Faculty of Arts" p. 379,
he points out a marginal note to the 1268 statutes (p. 26.15): "Nota quod possit esse
!eneralis sophista per annum."
3 Fletcher, "The Teaching", p. 434, recalls a 1607 provision intended "to restore the
Parvisus exercises to their alleged former glory," in Gibson, Statuta, p. 485: it lays
down that the student, before presenting himself for the Bachelor of Arts, must be
promoted "ad gradum generalis" for four terms, in each of which he has to "semel ad
minus in scholis pubJicis opponere et disputationibus in parvisis diIigenter interesse"
(lines 30-1). This provision, though late for our purposes, is evidence that the student's
exercises at Oxford also took place both in scholis publicis and in parvisis.
108 ALFONSO MAIERU

sit verum arguitur triplici medio. This is followed by three arguments, as in

Hopton's sophism)3 Then the sophism is developed in three principalia
puncta, each of which corresponds to one of the previous arguments
(medium), and is introduced by the discussion of opiniones famose. The
MS bears on f. 1r, at the top in red, the name 'Thomas de Moston' (?),
which is not recorded in Emden's Oxford and Cambridge Biographical
Registers. 34 This text should be examined more closely than I have been
able to do, to see whether it may be traced back to Oxford. But other
determinations of sophisms should be identified, so that a definite solution
may be found for the problem of the disputation of sophisms in the Oxford
schools. In the meantime, I would not give up this possibility simply
because of a comma placed by Gibson. The sophism attributed to Hopton,
in the light of the opinions discussed, may be traced to the Oxford of 1350-
1365.

II

The discussion of the doctrine of compounded and divided senses

(sensus compositus, sensus divisus) of propositions that include a modal
term occurs in the third part of the second article, namely, in the discussion
de veritate propositionis de possibili. By propositio de possibili the author
means any in which the indication of possibility (the term 'possibile')
precedes, is placed in the middle of the other elements of the proposition,
or follows them all, and one wonders what meaning its position
conveys.35

Hopton discusses two opinions. The first (opinio vulgaris) maintains

that a proposition is to be taken in the compounded sense when the modal
term precedes or follows the rest. 36 Thus the proposition 'album esse
nigrum est possibile' would have the same meaning as 'possibile est album
esse nigrum'; but an objection made to this opinion points out that the
former is a possible proposition, whereas the latter is impossible)? The
other opinion maintains that when the modal term follows the dictum, then
the proposition is to be taken in the divided sense. I am not able to identify
the supporters of this second opinion on any firm grounds. 38

33See the Appendix for the outline of the Determinatio of Hopton's sophism.
34Emden, A Biographical Register, II, p. 1324, has only Moston, John, but there are
Morton, Thomas, p. 1321, and Mordon, Thomas, p. 1301. Emden, A Biographical
Register of the Ulliversity of Cambridge to 1500, Cambridge: University Press 1963,
only has a Mordon, Thomas, p. 409, and a Morton, Thomas, p. 414.
35Hopton, Determillatio, 2.3. Here and subsequently I refer to the edition I prepared on
the basis of the two MSS A and K and the 1494 Venetian edition, and I indicate the
division in paragraphs I adopted and not the folios of the Venetian edition or MSS.
36Hopton, Determillatio, 2.3.1.
37/bid., 2.3,l.cl.
38 1 have found the closest traces in the Italian commentators of Heytesbury's De sensu
composito et diviso, such as Sermoneta (MaieriI, Termillologia logica della tarda
scolastica, Rome: Edizioni dell' Ateneo 1972, p. 577) and Landucci (ibid., p. 587), who
claim that propositions with epistemic and volitional verbs have a compounded sense
when the verb precedes a single term ('cognosco Sortem') or an incomplex term that
signifies a complex term ('scio a propositionem'), and a divided sense when it follows
such a term or incomplex term. Peter of Mantua may be mentioned here (ibid., p. 556
n. 197), who claims that propositions with the same verbs may have a compounded
 'OMNIS PROPOSITIO EST VERA VELFALSA' 109

However, the first OpInIOn can much more easily be traced to

Oxford39 and also concerns Heytesbury. The latter, as we know, is the
author of a treatise on compounded and divided senses that had little
influence in England, but was circulated and annotated in Italy.40 He also
dealt with this subject in the De scire et dubitare, which is the second
chapter of the Regule solvendi sophismata, where he maintains that there is
a compounded sense when an epistemic verb precedit totaliter dictum
alicuius propositionis seu sequitur ftnaliter (and where, in actual fact, he
also asks himself if such a verb can be placed at the end of the sentence in a
way that is grammatically acceptable).41 This doctrine was an enormous
success, also thanks to the treatise entitled Termini qui faciunt - taken
from De scire et dubitare. 42 The same position is found in the anonymous
treatise Termini cum quibus43 and in the Opus artis logice attributed to
Bradwardine.44 It seems to me that this opinio vulgaris is nothing more
than the communis fama of which Billingham speaks in his De sensu
composito et diviso: est communis fama quae dicit quod semper est sensus
compositus quando totaliter praecedit vel ftnaliter subsequitur terminus
officialis significans actum mentis.45 Billingham rejects this opinion,
which has no authoritative or rational arguments to support it,46 and, for
his part, believes that there is a compounded sense when an officialis term
(such as possibile) precedes a common term, and a divided sense when the
common term precedes the officialis term; and since the probatio
propositionis occurs starting from the first mediate term, the compounded
sense requires the probatio to start from the officiable term and the divided
sense requires the probatio to start from the common term.47

But what is Hopton's position? He holds that there is a compounded

sense when the term 'possibile' precedes totaliter the dictum of the
proposition, and a divided sense when it mediat; on the contrary, when the
modal term totaliter postponitur vel subsequitur the dictum, then there may

sense when the verbs "totaliter precedunt dictum," and a divided sense "cum inter partes
dicti mediant aut totaliter sequuntur." But, as we can see, these are later positions and
do not coincide with that discussed by Hopton: "Alia opinio dicit quod universaliter
quando iste terminus possibile sequitur et ponitur post dictum propositionis, tunc facit
sensum divisum," and the example given and discussed leaves no room for doubt:
'album esse nigrum est possibile' is to be taken in the divided sense (Determinatio,
2.3.2).
390n the continental masters see Maieru, Terminologia, pp. 499-622; but Parisian
masters deserve reconsideration: Albert of Saxony, Perutilis logica, Venetiis:
Octavianus Scotus 1522, f. 40va, claims that a proposition has a compounded sense
when the modal precedes or follows the dictum, and a divided sense when it ponitur in
medio dicti.
40 See A. Maieru, "1\ 'Tractatus de sensu composito et diviso' di Guglielmo
Heytesbury", Rivista critica di storia dellafilosofia 21, 1966, pp. 243-63.
4lQuoted in Maieru, Terminologia, p. 603.
42Ibid., p. 553, and pp. 601-6, on p. 602.
43/bid., p. 553, and pp. 607-22, on p. 610.5-15.
44See J. Pinborg, "Opus Artis Logicae", Cahiers de l'Institut du moyen-age grec et latin
42, 1982, p. 162.19-24.
45Partially edited by A. Maieru, "Lo 'Speculum puerorum sive Terminus est in quem'
di Riccardo Billingham", Studi Medievali, 3rd series, 10, 1969, 3, p. 389.98-100.
46/bid., p. 390.101-3.
47/bid., p. 387.3-19 and p. 388.28-35.
110 ALFONSO MAlERU

be either the one sense or the other indifferenter. 48 This position is also
found in Ralph Strode's Logica,49 and in the Tractatus aureus where the
same adverb 'indifferenter' is used. 50 Hopton's position on this doctrine,
as we can see, is not that of Heytesbury and the tradition he inspired, with
which our author conflicts. On the other hand, whereas Strode and the
Tractatus aureus are in line with Billingham and adopt a different probatio
for the proposition in the compounded sense and for that in the divided
sense, and though Hopton knows the probatio propositionis,51 he does not
speak of it in this context, and confines himself to explaining how to
interpret the proposition in which the modal term comes at the end. In his
view, this proposition is to be understood either in the one sense or the
other, depending on whether, in uttering it, there is a pause between the
parts of the dictum. In fact the compounded sense occurs quando non cadit
discontinuatio prolationis inter dictum et modum: in this case the
proposition 'album esse nigrum est possibile' is uttered without a pause
and is equivalent to 'possibile est album esse nigrum'. On the other hand,
if there is a pause (interruptio vel discontinuatio) between the parts of the
dictum, so that one part of the dictum is uttered by itself and the other is
uttered with the modal (e.g. 'album. esse nigrum est possibile'), then the
proposition is understood in the divided sense and is equivalent to 'album
potest esse nigrum' .52 All this is not found in Strode and the Tractatus
aureus, but it is in line with the English tradition, and discussions of this
type at the end of the 13th century have been documented by S.
Ebbesen.53

III

But now let us examine some of the other subjects discussed in the
sophism. The sophism-proposition that triggers the discussion is the
principle of bivalence: 'Every proposition (or sentence) is true or false' (I
always use the term 'proposition' for the Latin propositio). Three
arguments are immediately put forward in the text to support the falsity of
the sophism-proposition, and an article is introduced for each of them. The
first article discusses the question numquid deum esse sit deus vel
aliqualiter a deo distinctum. The second article discusses numquid terminus
communis supponens respectu verbi ampliativi supponit indifferenter pro
hiis que sunt vel pro hiis que possunt esse. The third article discusses
numquid inferius significet superius, et e contra superius inferius. 54
Having already examined the part of the second article on compounded
sense and divided sense, I shall confine myself here to referring to the
author's positions on the subjects dealt with in the first and third articles.

48Hopton, Determinatio, R.ad2. 3. 1-2.

49See Maieru, Terminologia, p. 551 and n. 183.
SO/bid., p. 552 and n. 185, and De Rijk, Some 14th Century Tracts, p. 230.
5IOn expositio and ojficiatio, see below, nn. 63 and 72.
52Hopton, Determinatio, R.ad2. 3. 1-2.
53"Suprasegmental Phonemes in Ancient and Medieval Logic", in English Logic and
Semantics from the End of the Twelfth Century to the Time of Ockham and Burleigh,
ed. H. A. G. Braakhuis et aI., Nijmegen: Ingenium Publishers 1981, pp. 331-59.
54See Appendix for the outline of Hopton's Determinatio.
 'OMNISPROPOSITIO EST VERA VELFALSA' 111

The first article opens with the introduction and criticism of three
famous opinions on the significate of a proposition. The first, according to
which deum esse est deus I Sortem esse estSortes, may be traced back to
the position that Ferribrigge calls ceteris veriorem. 55 The second,
according to which deum esse is nothing, but is aliqualiter esse et modus
rei is Billingham's position.56 The third, according to which deum esse est
oratio infinitiva, in Paul of Venice's view derives from the opinion that the
significate of a proposition is compositio mentis;57 this last opinion is
attributed to William ofBe~ngham.58

Hopton advances arguments against the three opinions59 and then

proceeds to discuss the truth of a proposition and to distinguish two
aspects of the question: a) what is necessary and sufficient to make a
proposition true,60 and b) what is the truth of a proposition.61 Both points
lead to an examination of various opinions. In the first case, the author
examines the opinion of the schola, which usually maintains that, in order
for a proposition to be true, it is necessary and sufficient that ipsa precise
signijicat sicut est,62 and discusses various proposals of probatio by
means of expositio of this proposition.63 Further on, under the second
point, he also examines the opinion according to which for a proposition to
be true it is necessary and sufficient that ita est totaliter sicut ilia
signijicat.64 In the second point, Hopton discusses the opinion according
to which the truth of a proposition is an accidens respectivum subiective in
intellectu vel in propositione que est vera jormaliter, an opinion, in the
author's view, deriving from the theory of truth as adaequatio, that he

55The three opinions in Hopton, Determinatio, 1. 1-3. On the first opinion see
Ferribrigge, Logica, in Paul of Venice Logica Magna, II, 6, ed. F. Del Punta and M.
McCord Adams, Oxford: Oxford University Press 1978, p. 223.12, and Richard
Brinkley's Theory of Sentential Reference, ed. M. J. Fitzgerald, Leiden: E. J. Brill
1987, p. 82.4-8.
56Billingham, Utrum idem Sortes et Sortem-esse, in Fitzgerald, Richard Brinkley, p.
135, Rl, and Billingham, De significatio propositionis, ibid., pp. 149-50; see also
Brinkley, ibid., p. 52.6-12.
57Del Punta & Adams, Paul of Venice, p. 90. 29; see also p. 84. 34-35 ('compositio
mentis').
58The attribution is in the margin of the MS of Brinkley's logic: see Fitzgerald, Richard
Brinkley, p. 74.1-7; the Brinkley text ignores the derived opinion discussed by Hopton:
see Fitzgerald's introduction, ibid., p. 32.
59Hopton, Determinatio, I.l.cl-8 ('contra prima opinionem'), 1.2.cl-6 ('contra
secundam opinionem') & 1.3.cl-4 ('contra tertiam opinionem').
6OIbid., 1.4.
611bid., 1.5.
621bid., 1.4.1.
63E.g. "hec propositio significat sicut est et non significat aliter quam est," ibid.,
1.4.1.2.1; "hec propositio significat sicut est et non significat sicut non est," ibid.,
1.4.1.2.2.
64lbid., 1.5.2.
112 ALFONSO MAIERU

attributes to Boethius;65 however, at least it calls to mind Ockham's earlier

discussion. 66

I shall not insist on examining all the various opinions here. I merely
wish to observe that in responding to the sophism and the first article the
author appropriates the objections made to these opinions and accepts and
explains the grounds for them, whilst refusing to grant the significate of
the proposition any form of distinct and independent entity.

The author responds that the sophism-proposition is true, and denies

the antecedents of the three arguments put forward to prove its falsity.67
Hence he rejects the famous opinions (but here the most explicit reference
is to what is upheld by the first two), stating that deum esse is not God nor
is it an aliqualiter distinct from God, and maintains that universaliter non
est ita sicut aliqua propositio signijicat, loquendo de signijicato primario et
adequato, since no proposition signifies sicut est. Naturally this is objected
to by quoting Aristotle, Cat. 5, 4b8-9: ab eo quod res vel non est oratio
dicitur vera vel falsa. However, Hopton does not think this is the universal
definition of a true proposition, since one does not give such a definition.
In fact, we do not possess the terms or concepts that would permit us to
give such a definition; all our concepts lead us to a definition includentem
aliqualiter esse, and there is not an aliqualiter esse as signified by a
proposition. The answer to the question: why is this proposition true or
false? must be that there is no cause of the truth of it in reality. If there
were, the cause of the truth of homo est would be this: quia precise
significat quod homo est, et homo est. But that it precise significat quod
homo est is neither aliquid nor aliqualiter. 'Cause' therefore is not to be
understood as something a parte rei, but as premises capable of inferring a
conclusion, and in this sense of being the cause of it: hec propositio 'tu
sedes' precise significat quod tu sedes et tu sedes. igitur hec est vera 'tu
sedes' .68

IfP = tu sedes and P = 'tu sedes' (the name of p), we have:

if P precise signifies that p and p, then P is true.

As we can see, this formula (that I have taken from P. V. Spade)69 calls to
mind the equivalence made famous by Tarski ('''It is snowing' is a true
sentence if and only if it is snowing"»)O But, as E. J. Ashworth has

65Ibid.• 1.5.1; on the origin of the adaequatio theory see L. Minio-Paluello, Opuscula.
The Latin Aristotle, Amsterdam: Adolf M. Hakkert Publisher 1972. pp. 533-4.
66The closest formulas are in Quodl. V. q. 24: see Guillelmi de Ockham Quodlibeta
septem. ed. J.C. Wey, St. Bonaventure, N.Y.: St. Bonaventure University 1980, pp.
575.27-30 & 576.46-57.
67Hopton, Determinatio, R.ad!.
68/bid., and see Billingham, De signijicato, pp. 149-50.
69"lnsolubilia and Bradwardine's Theory of Signification", Medioevo 7,1981, pp. 115-
34, on p. 127.
70A. Tarski, "The Concept of Truth in Formalized Languages", in A. Tarski, Logic,
Semantics. Metamathematics: Papers from 1923 to 1938, Oxford: Clarendon Press
1956. pp. 155-6.
 'OMNISPROPOSITIO EST VERA VELFALSA' 113

recently emphasised,7! medieval logicians knew very well that it is not

sufficient that p for P to be true (P might not exist) and that it is not
sufficient for p that P be true (P might change signification). The passage
from language to metalanguage and vice versa is not permitted if other
premises are not introduced. The clause 'P precise signifies thatp' seems
to fulfil this function (as may be seen, this is the context of probatio
propositionis, close to the officiatio). Nonetheless, Hopton accepts the
clause only gratia argumenti, to serve in the discussion of the following
articles.72 In fact, he knows that, language being radically conventional, a
true proposition may also signify sicut non est (hec proposino 'homo est'
est vera, et tamen sicut non est signijicat, quia signijicat quod homo est
asinus);73 but he does not dwell on this subject. Having granted the clause,
he states that, for a proposition to be true, it is sufficient that it precise
signijicat sicut est, but it is not necessary. Indeed, neither negative
propositions, nor affirmative propositions with a verb in the past or the
future, nor affirmative propositions with an ampliative verb in the present
signify precise primarie sicut est;74 And therefore he gives a definitive,
articulate response to the question of what is necessary for a proposition to
be true, by distinguishing the non-ampliative affIrmative proposition in the
present tense from all the others. In the former (the non-ampliative
affirmative proposition in the present), for it to be true sufficit et requiritur
quod ipsa precise signijicat sicut est assertive et directe; in the latter (all
other propositions), on the contrary, nichil est in re sUfficiens et requisitum
ad hoc quod ilia sit vera.75
The author responds to the other question: what is the truth of a
proposition, by saying that it is none other than a true proposition: in the
pairs of terms 'true'/,truth', 'false' /'falsity', the abstract form of the noun
has the same semantic value as the concrete form (eadem res precise
importatur per abstractum et concretum, et e converso).16
But let us return to the clause that includes the term 'precise'. We have
found various expressions: 'precise primarie', 'precise assertive et directe'
and also' loquendo de significato primario et adequato'. We must therefore
ask ourselves what these expressions signify, and, more generally, what
follows from a given proposition. Hopton discusses this in the third
article, and in the response to it he concedes that the inferior term signifies
its superior term: e.g., homo signijicat animal.17 But the logical relation
between the inferior term and the superior term is not sufficient, in the
author's opinion, to conclude that the proposition 'Every man runs'
signifies that every animal runs. In fact, the proposition 'homo currit'
signifies the proposition 'animal currit' secondarily, not primarily (he
states that he is using 'primarily' and 'secondarily' in the sense used in the

7! "La semantique du XIIe siecle vue a travers cinque traites oxoniens sur les
Obligationes", Cahiers d' epistemoiogie 8915, 1989, p. 13.
72Hopton, Determinatio, R.adl.4.l.
73Ibid., R.adl.5.2.
74Ibid., R.adl.4.l.
75Ibid., R.adl.5.2.
76Ibid., R.adl.5.l.
77lbid., R.ad3.l.
114 ALFONSO MAIERU

scola communis).78 Moreover, since negation denies what affirmation

primarily affirms, it does not follow that the proposition 'nullus homo
currit' denies that some animals run.7 9 In addition, the superior term
signifies the inferior term;80 therefore the author grants that 'homo'
signifies Socrates and that it signifies every man, but does not grant that
the proposition' homo currit' signifies that every man runs: to obtain that,
you must add in the antecedent (Le. to 'homo currit') that the term 'homo',
as well as signifying every man, has a universal and copulative supposition
for all men, which is false, since the subject in 'homo currit' remains
disiunctive et particulariter.81

The author responds to the further question of whether the negative

proposition signifies its affirmative (Le., whether 'nullus homo est asinus'
signifies that man is an ass), by granting it, on the grounds of Soph. EI.
31, 181b20-30, since the affirmation is understood in the negation. 82 But
he responds to the objection triggered by the axiom: everything that is
signified by a proposition follows from it, because everything that is
included or understood in something follows from it,83 by distinguishing a
double meaning of the proposition. It may signify directe vel indirecte,
assertive uel desertive. The proposition 'No man is an ass' signifies that
man is an ass indirecte et non assertive. That axiom and other similar ones
are false if they are taken without limiting clauses; they must be understood
only in relation to the direct and assertive meaning of a proposition. 84

What does assertive mean here? According to De Rijk,85 Johannes

Venator contrasts 'signifying primarily' with 'signifying assertively': "So
'homo currit' assertively signifies that an animal runs," and this is
supported by a similar use of the term in the 16th century. Here, however,
Hopton seems to place the terms side by side, and not to contrast them.
Thus we have 'to signify primarily' and 'to signify directly and assertively'
on the one hand, and 'to signify secondarily' and 'to signify indirectly and
non-assertively' on the other, since the limitations set by Hopton to define
the area of what it is permissible to infer from a given proposition are far
stricter than those set by Johannes Venator.

Universita "La Sapienza" , Rome

78Ibid., R.ad3.l.cl.
79Ibid., R.ad3.l.cl.a2.
8oIbid., R.ad3.2.
81Ibid., R.ad3.2.cl.
82Ibid., R.ad3.3.
83Ibid., R.ad3.3.al; see Spade, "Insolubilia", p. 120 n. 17.
84Hopton, Determinatio, R.ad3.3.al.cl.
85"Semantics in Richard Billingham and Johannes Venator", in English Logic in Italy
in the 14th and 15th Centuries, ed. A. Maieril, Naples: Bibliopolis 1982, p. 176.
 'OMNIS PROPOSrrlO EST VERA VELFALSA' 115

Appendix
Henry Hopton, Determinatio
Outline

Omnis propositio est vera vel falsa.

Quod sophisma sit falsum arguitur sic:
(1) Hec propositio 'deus est' nec est vera nec falsa; igitur etc.
(2) Secundo sic: Nulla propositio affirmativa ampliativa, ubi verbum
principale est ampliativum, est vera vel falsa; igitur etc.
(3) Tertio sic: Nulla propositio affll1l1ativa est vera, ubi predicatur superius
de suo inferiori, vel e contra; igitur etc.
Ad oppositum et pro veritate sophismatis arguitur sic: Tantum verum vel
falsum est propositio; igitur etc.
Penes materiam primi argumenti est iste articulus pertractandus: Numquid
deum esse sit deus vel aliqualiter a deo distinctum.
Penes materiam secundi argumenti est iste articulus pertractandus:
Numquid terminus communis supponens respectu verbi ampliativi
supponat indifferenter pro his que sunt vel pro his que possunt esse.
Penes materiam tertii argumenti est iste articulus pertractandus: Numquid
inferius significet superius, et e contra superius inferius.
Primus articulus:
(1. 1-3) tres opiniones famose;
(1. 4) de veritate propositionis: quid requiritur et sufficit ad hoc quod
aliqua propositio sit vera;
(1. 5) quid est veritas propositionis.
(R. adl) Responsio ad sophisma et ad primum articulum.
Secundus articulus:
(2.1) contra secundum articulum arguitur;
(2.2) de possibilitate propositionis: quid requiritur et sufficit ad hoc quod
aliqua propositio sit possibilis;
(2.3) de veritate propositionis de possibili.
Tertius articulus:
(3.1) contra tertium articulum arguitur;
(3.2) de modo significandi et de significato terminorum significantium
omnes res imaginabiles (e.g. 'chimera');
(3.3) numquid hec propositio 'nullus homo est asinus', et huiusmodi,
significet affIrmative quod homo est asinus;
(3.4) de veritate illius propositionis ubi predicatur superius de suo
inferiori.
(R.ad2) Responsio ad secundum articulum.
(R.ad3) Responsio ad tertium articulum.
 Die Rolle der Sophismata im Unterricht der Krakauer
Universitiit im 15. lahrhundert

von Mieczyslaw Markowski

Die im 15. Jahrhundert an der Krakauer Universitlit unterriehtete

Logik soUte, ahnlich wie an anderen mittelalterlichen Hochschulen, die
Rolle eines Instruments wissenschaftlichen Denkens spielen. Sie wurde
betrachtet als Einftihmng nieht nur zum Studium der Philosophie und der
freien Ktinste, sondern auch als solche zur Medizin, zur
Rechtswissenschaft und zur Theologie. Aus dies em Grunde wurde die
Logik schon am Anfang des Universitatsstudiums unterrichtet. Die 1404-
1406 verfaBten Statuten regelten gesetzlich, wie lange und wann der
Unterricht in den einzelnen logischen Disziplinen erteilt werden soUte. Bis
auf die Topica wurden das ganze Organon des Aristoteles, die Summulae
logicales des Petrus Hispanus und andere mittelalterliche Traktate tiber
semantische Grundbegrriffe im unteren Kurs erIautert, d.h. bevor der
Student den Bakkalaureusgrad erworben hatte. An der Krakauer
Universitlit soUte damals der Kandidat ca. 18 Monate lang die Vorlesungen
tiber Logik Mren. l Keiner anderen Disziplin so viel Zeit wurde auf dieser
Unterrichtsstufe der Philosophie im Rahmen der freien Ktinste gewidmet.
Man konnte sogar sagen, daB der untere Kurs als "Logikschule" gedacht
war. Den erwahnten Statuten gemaB soUte auch Aristoteles' De sophisticis
elenchis innerhalb von 14 Wochen erlautert werden. 2 In der Regel geschah
dies mit Hilfe der bereits vorhandenen mittelalterlichen Kommentare zu
diesem Werk. In der Handschriftensammlung der Bibliothek der Krakauer
Universitat ist so gar ein Exemplar des handschriftlichen Kommentars
Johannes Buridans zu De sophisticis elenchis erhalten geblieben.3 Bekannt
war auch sein Compendium totius logicae.4 Von den logischen Schriften
seiner Pariser SchUler seien auch die Sophismata des Albert von Sachsen5
und der in mehreren Handschriften ilberlieferte Kommentar des Marsilius
von Inghen zu De sophisticis elenchis genannt. 6

Die ersten Logiker der emeuerten Krakauer Universitat interessierten

sieh auch filr die Oxforder Logik. Besonders beachtenswert sind neben der
Summa totius logicae des Wilhelm von Ockham,7 die Sophismata des
Richard Kilvington8 und des Heinrich Hopton. 9 In Krakau waren auch die

lStatuta nee non Liber promotionum philosophorum ordinis in Universitate studiorum

Jagellonica ab anno 1402 ad annum 1849. ed. J. Muezkowski. Craeoviae 1849. S. XII-
XIII.
2Statuta .... S. XIII.
3Krak6w. BJ (= Biblioteka Jagiellonska). ems 736. f. 51 va-80rb.
4/bid.• ems 662. f. lra-156rb. ems 703. f. 2ra-181 va; Biblioteka KSiezy Misjonarzy. ems
827. f. 3ra-218va.
5Krak6w. BJ. ems 2330. f .lr-113v.
6/bid.• ems 712. f. 54rb-7Orb. ems 711. f. 23va-45vb. ems 713. f. 60ra-154va.
7/bid.• ems 719. f.lra-114rb.
8/bid .• ems 750. f. 99ra-llOv. Vgl. A. Maieru. Terminologia logiea della tarda
seolastica. Rome: Edizioni dell' Ateneo 1972.

116
 DIE ROUE DER SOPHISMATA 1M UNTERRICHT 117

Sophismata longa des Hollanders Johannes von Monickedam bekannt. 1O

Die erwahnten Werke auslandischer Logiker haben sicherlich die Rezeption
der in Krakau gelesenen aristotelischen Sophistik beeinfluBt, bis in den
zwanziger Jahren des 15. Jahrhunderts in Krakau drei eigenstiindige
Schulkommentare zu De sophisticis elenchis entstanden. II 1m letzten
Viertel des 15. Jahrhunderts wurde in Krakau immer hiiufiger eine
unmittelbare Exegese der Sophistici elenchi des Aristoteles vorgetragen, u.
a. von Walentin von Krakau.1 2 Es enstanden auch umfangreiche
Kommentare zu dem genannten Werk des Aristoteles. Johann von Glogau
ist Verfasser des im Geiste des Albertismus und Thomismus des 15.
Jahrhunderts geschriebenen Exercitium "Elenchorum" ,13 welches noch zu
Anfang des 16. Jahrhunderts im Druck herausgegeben wurde.1 4 Unter
dem EinfluB des spiitmittelalterlichen Scotismus schrieb Michael von
Biestrzyk6w einen umfangreichen Kommentar zu De sophisticis
elenchis.1 5 1m Vergleieh mit den demonstrativen und dialektischen
Schliissen wurde den eristischen Schliissen an der Krakauer Universitiit
nieht viel Aufmerksamkeit gewidmet. Die eristischen Schliisse, die damals
sophistische genannt wurden, unterscheiden sich von den vorherigen
sowohl in der Wahrheit der Priimissen als auch in der Giiltigkeit des
Schlusses. Die Priimissen des eristischen Schlusses sind nur scheinbar
wahr und besitzen lediglich den Anschein der Glaubwiirdigkeit. Dieser
scheinbar schliissige Gedankengang fiihrt zu einem TrugschluB, der aus
den Priimissen nicht folgt. I6 Trotzdem muBten die Krakauer Kandidaten
fiir den Grad eines aIle drei SchluBverfahren beherrschen. In den
Vorlesungen erwarben sie die theoretischen Kenntnisse, und wiihrend der
Ubungen und Disputationen lemten sie in der Praxis, Einwande und die
Antworten darauf zu formulieren. Diese Kenntnisse wurden in vielfaIliger
Weise angewandt und dienten als causa docendi, exercitandi, cognoscendi,
dubitandi oder placendi. I7 Schon bei der Bakkalaureuspriifung sollte der
Student seine Leistungen im Argumentieren pro et contra zeigen, und in
diesem Fall nannte man es causa temptandi. Trotz mancher gemeinsamer
Ziige machte sich hier auch das Lokalkolorit bemerbar, das seinen
Ursprung in den ortlichen Statuten und Gebriiuchen hatte. Und das war der
AnlaB dafiir. daB ich mich mit der Rolle der Sophismata im Unterricht der
Krakauer Universitat im 15. Jahrhundert naher beschiiftigt habe.

Da die ersten Professoren auf das wissenschaftliche Gepriige der

Krakauer Universitat einen groBen EinfluB ausgeiibt haben. und zwar nicht

9Ibid.• ems 621. f. 23ra-47va, ems 704. f. lra-24ra; Warszawa. BN (= Biblioteka

Narodowa). ems ake. 1819. f. 117ra-136vb.
IOKrak6w. BJ. ems 2660. f.50ra-125rb.
II/bid.• ems 642. f. 1 Ir-47v. ems 642. f. 48r-49r. 51r-v.
IZ/bid.• ems 1893. f. 198r-252r.
13/bid.• ems 25. f. 155v-175va.
I40ieses Exercitium wurde 1511 in Krakau herausgegeben. vgl. Anm. 72. Es soli noch
eine Edition aus dem Jahre 1506 gegeben haben.
I5Dieser Kommentar ersehien im Druck 1502 in Krakau. Er sollte noch im Jahre 1507
gedruckt sein. aber bis jetzt hat man diese Ausgabe nicht gefunden.
I6Cf. M. Markowski. Dzieje jilozojii sredniowiecznej w Polsce. T. 1: Logika. Wroclaw
1975. S. 86.
17U. Gerber. Disputatio, in Theologische Realenzyklopedie. Bd. 9. Berlin - New York
1982. S. 14.
118 MIECZYSMW MARKOWSKI

nur Philosophen, sondern aueh Juristen und Theologen, die in groBer Zahl
an der Prager Universitat studiert haben, miiBte man meines Eraehtens
erwagen, ob nieht die Prager Sitten an die Krakauer Universitat iibertragen
worden sind. GernaB den Statuten der Prager Universitat muBte der
Kandidat fUr das Bakkalaureat, der wahrend der Priifung zunaehst mit
anderen Studenten zusarnmensaB, auf ein Sophisma antworten. Wie die
Antwort aueh ausfiel, der Magister replizierte nieht. Naeh der Antwort
durfte der Student die Robe anlegen und unter den Bakkalaurei Platz
nehmen. Erst jetzt sehlug ihm der Magister eine quaestio zur Losung vor.
Dann legte der Kandidat fiir das Bakkalaureat seinen Eid abo Der
promovierende Professor hielt dem Kandidaten zu Ehren eine Rede und
erteilte ihm danaeh den Grad eines Bakkalaureus.1 8 Es sei aueh bemerkt,
daB die Seholaren vor der Bakkalaureuspriifung dazu verpfliehtet waren,
an den allgemeinen, fUr aIle verbindliehen Streitgespraehen (disputationes
communes) der Magister teilzunehmen und auf die von ihnen aufgestellten
sophistisehen Thesen zu antworten. Auf die Fragen, welche die Magister
stellten, durften nur die Bakkalaurei antworten.1 9 Es muB noeh betont
werden, daB der Leiter der verbindliehen Streitgespraehe nieht mehr als
drei sophistisehe Thesen aufstellen durfte. An der Disputation durften
hoehstens neun Sophisten teilnehmen. In Prag konnten sieh an der
Disputation einer sophistisehen These Seholaren beteiligen: der eine
(concedens) verteidigte die aufgestellte sophistisehe These, der zweite
(negans) verneinte sie und der dritte (dubitans) zweifelte an beiden.20 An
der Prager Universitat disputierte man wahrend der Ubungen in der
Naturphilosophie mindestens ein Sophisma, und manehmal zwei
Sophismata. 21 Am 21. Oktober 1387 wurden dort noeh Ubungen in der
Sophistik (exercitia in sophistria) eingefiihrt, die dureh drei Vierteljahre

18"Quando baccalariandus vult procedere, magister suus debet sibi proponere unum
sophisma, ad quod respondebit sedendo cum aliis scholaribus, et non inter baccalarios,
et in mantelo suo nec magister contra responsionem replicabit. Ista responsione facta
bidellus faciat eum surgere, et habitum induere, et in loco baccalariorum sedere, et
magister proponat sibi questionem, quam debet honeste determinare, quo facto bidellus
faciat eum jurare tria ultima juramenta supra scripta, et alia statuta, et statuenda et tunc
magister faciat collocationem de eo, et dabit sibi gradum baccalariatus." Liber
decanorum Facultatis Philosophicae Universitatis Pragensis ab anno Christi 1367
usque ad annum 1585, pars I, Pragae 1830, S. 52.
19"Item in plena congregatione facultatis statutum fuit, quod scholaribus post examen
hujusmodi admissis deberet injungi, quod toto tempore, quo non processerint,
disputationes magistrorum visitare teneantur, ad sophismata responderi, sic baccalarii ad
questiones, et sub consimili poena per facultatem eis, si negligentes fuerint, indicenda."
Ibid.
20"ltem 24 die mensis Januarii conclusum fuit, quod de cetero disputatio ordinaria
continuari debeat usque ad horam vesperorum, et quod sophismata per presidentem non
distribuantur plura, quam tria, nec habere debet plures sophistas, quam novem ita quod
ad quodlibet sophisma possint esse tres, unus concedendo, alter negando, tertius
dubitando." Ibid., S. 64.
21"Item Sabbato ante Michaelis conclusum erat in facultate, quod quilibet disputans in
philosophia naturali ad minus deberet unum sophisma, vel duo ad maius disputare, et
pro isto tantum debet dari sicut de exercitio totaIiter philosophicali, et debet sophisma,
vel sophismata ante questiones disputare. Item conc\usum fuit, quod nullus
magistrorum uno actu plures, quam tres questiones vel duas questiones cum uno vel
duobus sophismatibus ad majus, ut praescriptum est, debeat disputare." Ibid., S. 87.
 DIE ROUE DER SOPHISMATA 1M UNTERRICHT 119

tiiglich eine Stunde lang dauerten. 22 Zusammenfassend kann gesagt

werden, daB die Prager Universitiit gegen Ende des 14. Jahrhunderts zu
einer "Schule der Sophistik" wurde. Diese Schulung erfuhren auch die
damals diese Universitiit besuchenden polnischen Studenten, die danach
hiiufig die ersten Professoren der emeuerten Krakauer Universitiit wurden.

Obwohl die Prager Universitiitsstatuten die Grundlage fur die

Bearbeitung analoger Gesetze fur die Facultas Artium Liberalium in
Krakau bildeten und obwohl die Dialektik in der ersten Hiilfte des 15.
Jahrhunderts besonders bevorzugt wurde, sprach man im Lehrprogramm
der erneuerten Krakauer Universitiit del- Sophistik sowie auch den
entsprechenden Ubungen und Disputationen keine groBere Bedeutung zu.
Es sei bemerkt, daB es in Krakau keine "beliebigen" Disputationen, d.h.
disputationes de quolibet gab. Die Form der quaestio spielte aber als
Lehrmethode eine groBe Rolle. In Form der quaestiones doctrinales war sie
schon in den Vorlesungen (lectiones) vorhanden. In Form der quaestiones
dialecticae fand diese Methode ihren festen Platz in den Ubungen
(exercitia), die einerseits der besseren Aneignung des Vorlesungsstoffes
uod anderseits der Befahigung zur Disputation dienten. Auf diese Weise
wurden in der ersten Hiilfte des 15. Jahrhunderts zwei bekannte Krakauer
Handbucher mit Physikubungen vorbereitet, die von dem Dekan der
Facultas Artium Liberalium geleitet wurden. Infolge der Hiiufigkeit solcher
quaestiones begann die Dialektik an dieser Fakultiit zu dominieren. Die
Krakauer Statuten empfahlen, daB die vor der Prufung zu einem
wissenschaftlichen Grad stehenden Studenten an den allgemeinen
Disputationen teilnahmen. Besonders galt das fur die Studenten vor der
Bakkalaureusprufung, die einmal als Zweifelnde (dubitantes) auftreten
muBten23 und mindestens ein halbes Jahr lang vor der Prufung an diesen

22"Item ao. Dni. 1387 die 21 mensis Octobris in plena congregatione facultatis matura
prius deliberatione habita in decanatu magistri Joannis Eliae fuit concorditer conclusum
et statutum, quod ter in anno exercitia in sophistria teneantur, videlicet tribus anni
quartalibus, sic quod per unum quartale quodlibet exercitium continuetur, primum ab
ephiphania Domini usque ad festum paschae, omni die per unam horam, hora 19;
secundum a festo paschae usque ad festum s. Jacobi pro 2do quartali anni, et istud omni
die hora 16. continuetur; tertium a festo s. Jacobi usque ad festum s. Galli omni die
hora 17. continuetur; per quartum vero quartaIe anni nullus teneat exercitia in sophistria
propter temporis illius quartalis brevitatem. Adjectum autem fuit, quod disputans in
sophistria de vespere tantum unum exercitium teneat, vel in logica tantum vel in
philosophia tantum, tribus diebus, quibus sibi placuerit disputando, et quod de quolibet
hujusmodi exercitio 12 grossi persolvantur." Ibid., S. 90.
23"Quoniam plerique studencium responsiones publicas, in disputacionibus ordinariis
magistrorum fieri solitas, ad quas, uigore statuti super hoc facti, ante gradum sue
promocionis obligantur, superfugere volentes, ad vnum zophisma multi simul
dubitando respondere, et, presidenti duntaxat respondentes, recedere consueuerunt;
ideoque conclusum fuit et statutum: quod in antea ad vnum zophisma solum vnus
respondere debeat dubitando; poterit tamen secundus, ex licencia decani pru tempore
existentis, ex causis legitimis, quas decanus facultatis cognoscat, ad idem zophisma
respondendo dubitare: in nullo tamen casu liceat plures, quam duos, ad vnum zophisma
esse dubitantes, sub pena non computandi responsiones. Qui quidem dubitantes ad
finem disputacionis, sicut et alii respondentes, si eis eundem actum pro responsione
computari voluerint, debent pennanere." Statuto ... , S. XVI-XVII.
120 MIECZYSMW MARKOWSKI

Disputationen teilnehmen sollten.24 Nach dem Gesetz von 1462 konnte nur
der Dekan die Thesen der Sophismata fUr die Disputationen und fUr die
Prtifung verteilen. 25 Die Krakauer Universitiitsstatuten spreehen wenig
tiber die Sophismata, betonen dagegen die Pflichten der Studenten. Sie
sehweigen aber davon, wie die Disputationen tiber eine sophistisehe These
verlaufen und wie die literarisehe Form des Sophismas aussehen sollte.
Das wurde von den im 15. Jahrhundert an der Krakauer Universitiit
entstandenen Sitten beBtimmt. Diese Sitten sind einigermaBen anhand der
damaligen Promotionsreden und Sophismata wiederherzustellen, die uns in
zeitgenossisehen Handsehriften tiberliefert sind.26 Ich moehte mieh auf
den handsehriftliehen Codex 2205 der Jagiellonisehen Bibliothek
besehriinken,27 wo auf 369 Folien ca. 260 Krakauer Universitiits-

24"Quia promoti ad gradus nedum in scienciis, verum et in usu scienciarum, qui in

modis legendi, arguendi, et respondendi consistere videtur, debent esse exercitati; quos
quidem modos, precipue magistros arguentes et baccalarios respondentes diligenter
audiendo adiscere (sic) et acquirere contingit, ideoque conclusum est: quod quilibet
studencium propter hoc maxime, ut modum arguendi et respondendi adiscat, per
medium annum ad minus, ante ingressum suum pro gradu baccalariatus ad examen,
disputacionem ordinariam magistrorum assidue visitet sic, quod tres disputaciones
ordinarias immediate sequentes, sub pena unius grossi, nullatenus negligat; quam
penam tociens, quociens predicta facere neglegxerit, incurret, per decanum a quolibet,
qui earn incurrerit, tempore, quo examen intrare voluerit, exigendam." Ibid., S. XYII-
XVIII.
25"Quoniam odia licet restringere, fauores vero et concordiam inter maystros congruit
ampliare, ne igitur odia et indignaciones inter maystros locum habeant, que propter
responsiones studentium faciliter contingere possunt, placuit omnibus maystris protunc
in plena convacacione existentibus, vt a modo decanus facultatis arcium plenam
auctoritatem habeat sophismata distribuendi inter hos, qui secum steterint in exercicio
phisicorum, vel qui in proximo sunt promouendi, et quod amplius presidens non
habeat posse distribuendi sophismata nec dubitaciones; sed omnia procedant de manibus
decani: ne sic per defectum responsionum, vt sepenumero contigit, a promocione
studentes impediantur, et eciam vt illa commutacione, ante introitum examinis, in
studio magis exerceantur, et ipsorum ingenia coram maystris publice respondendo
pateant." Ibid., S. XXXI.
26Wroclaw, BU (= Biblioteka Uniwersytecka), cms I Q 380 (XIY/XY); Krak6w, BJ,
ems 2215 (XIY/XY):, f. 49-53; ems 2216 (XY2-3), f. 56, 58; cms 2231 (XY3-5), cms
2459 (ca. 1425), f. 262-286a; cf. M. Kowalczyk, Krakowskie mowy uniwersyteckie z
pierwszej polowy XV wieku, Wroclaw 1970, S. 17-19 und 71. Prager Einfliisse z. B.
cf.: "In actu presenti iuxta laudabilem consuetudinem facultatis arcium huius alme
Universitatis in huiusmodi actibus hucusque observatam quatuor sunt facienda: Primo
ponendum est unum zophisma, ad quod dominus baccalariandus responde bit. Secundo
proponenda est una questio, quam ipse determinabit. Tercio ad honorem ipsius fiet una
recommendacio. Quarto et ultimo referende sunt graciarum acciones hiis, qui dignati
fuerint facultatem arcium in persona mea exili visitando actum presentem honorare."
WrocIaw, BU, cms I Q 380, f. 60v. cf. M. Kowalczyk. Mowy ...• S. 80. 66-7.
27W. Wislocki. Katalog rekopis6w Biblioteki Uniwersytetu Jagiellonskiego, Krak6w
1877-81, S. 529; J. Zathey. "Biblioteka Jagiellonska w latach 1364-1492". in J.
Zathey, A. Lewicka-Kaminska, L. Hajdukiewicz. Historia biblioteki Jagiellonskiej, T.
1: 1364-1775. Krak6w 1966. S. 108; R. Palacz. "Michal Falkener z Wroclawia. Stan
OOdan", Materialy i Studia Zakladu Historii Filozojii Starozytnej i Sredniowiecznej, VI,
1966. S. 84-5; idem, "Wyb6r kwestii filozoficznych dyskutowanych na Wydziale
Artium Uniwersytetu Krakowskiego w drugiej polowie XY wieku", Materialy i Studia
Zakladu Historii Filozojii Starozytnej i Sredniowiecznej. X, 1969, S. 222-40; J.
Domanski. "Krakowski sofizmat 0 rzekomej wYZszo§ci retoryki nad dialektyka z
 DIE ROLLE DER SOPHlSMATA 1M UNTERRICHT 121

sophismata und -quaestiones aus dem 15. Jahrhundert erhalten sind:

"Sequuntur modi arguendi per quos, quid valet verum, potest probari esse
falsum."28 Dieser von vielen Handen geschriebene Codex wurde zum
Gebrauch derjenigen Krakauer Magistri und Bakkalaurei verfaBt, die an
allgemeinen Universitatsdisputationen (disputationes communes)
teilnahmen. Spater war er auch fiir die Dekane sehr niitzlich, die
Sophismata und Quaestiones fiir die Studenten brauchten, die sich urn
wissenschaftliche Grade bewarben. Bekannt ist auch der letzte Besitzer des
Codex 2205. Es war der beriihmte Krakauer Logiker und Philosoph
Michael Falkener aus Breslau,29 der an der Facultas Artium Liberalium
von 1488 bis 1512 las und in den den Wintersemestem 1499 und 1505 ihr
Dekan war.

An der Krakauer Universitlit wurden die sophistischen Thesen

vorwiegend von den Studenten vor dem Bakkalaureat gelOst, und dies
nicht nur wahrend der allgemeinen Disputationen, sondem auch wahrend
der Bakkalaureuspriifung. Ein solcher von dem promovierenden Professor
oder yom Dekan eroffneter Akt bestand aus drei Teilen: "In presenti actu
tria sunt facienda per ordinem."30 Zuerst wurde die Bakkalaureus-
disputation iiber zwei Sophismata gefiihrt. Ihre Teilnehmer nannte man
Sophisten: "Primo et principaliter disputanda sunt zophismata, ad que
respondebunt domini zophiste."31 Dann disputierte man iiber zwei
quaestiones und die Disputationsteilnehmer hieBen Respondenten:
"Secundo due questiones [disputandae sunt] ad quas respondebunt domini
respondentes."32 Zum SchluB dankte der neue Bakkalaureus all denen, die
sich an diesem Akt beteiligt hatten und zwar nicht nur den Magistem,
sondem auch den Scholaren: "Tercio et ultimo referende erunt graciarum
acciones hiis, qui dignati sunt visitare presentem actum ob reverenciam
huius alme universitatis et me in mea exili persona honorando."33 Ein
solcher Verfahren galt an der Krakauer Universitat bei der
Bakkalaureuspriifung schon im zweiten Viertel,34 urn die Mitte 35 und am

rekopisu Biblioteki JagielIOIlskiej 2005", Materialy do Historii FilozoJii

Sredniowiecznej w Polsce, II, XIII, 1970, S. 9-31.
28Krak6w, BJ, cms 2205, f. 14ra.
29"Questiones scripte, quondam in usu habite a Michaele Wratislaviensi, arcium et sacre
theologie professore, ad disputaciones communes, apprime necessarie (parum propter
dificultatem in triplici professia versant) generalis Studii Cracoviensis et ad
bibliothecam maioris Collegii artistarum spectant." Krak6w, BJ, cms 2205, f. 2r.
30lbid., f. 344r.
3 I Ibid.
32/bid.
33lbid.
34"In nomine Domini, amen. In actu presenti, ut in quolibet sibi simili tria per ordinem
sunt facienda. Primo predisputantur duo zophismata, secundo due questiones, demum et
ultimo graciarum acciones referende sunt hiis omnibus, qui non dedignati fuerint
continuare hunc actum a principio usque ad finem." Ibid., f. 17ra.
"In nomine Domini, amen. In actu presenti sicut et in quolibet actu sibi simili tria per
ordinem sunt facienda: Primo disputabuntur duo zophismata, secundo due questiones,
tercio et ultimo referantur graciarum acciones omnibus, qui dignati fuerint hunc actum
visitare." Ibid., f. 207r.
35"In nomine Domini, amen. In actu presenti tria per ordinem sunt facienda: Primo
disputabuntur duo sophismata, denum due questiones, tercio et ultimo referende sunt
122 MIECZYSkAW MARKOWSKI

Ende des 15. Iahrhunderts. Damals sprach man schon von den
lobenswerten Sitten der Krakauer Facultas Artium Liberalium: "In presenti
actu sieut et in quolibet sibi simili iuxta laudabilem consuetudinem
Facultatis Artium huius inclite Universitatis, in huius modi actibus
hucusque observatam tria per ordinem per me sunt facienda."36 In der
Disputation fiber ein Sophisma genfigte eine Antwort (responsio), doch die
vorgebrachten quaestiones muBten entschieden werden, sie wurden ad
decidendum gegeben.3 7 Die Disputation fiber die sophistischen Thesen
hatte bei den Teilnehmern nieht nur die Befahigung zur Diskussion zum
Ziel, sondern es soUte auch der der SchluBfolgerung des Gegners
entgegengesetzte SchluB bewiesen werden: "Primo duo sophismata
proponuntur, que pro veritatis et falsitatis evidencia probabuntur."38 Das
Hauptziel bei den quaestiones war es, die Offensichtlichkeit der Wahrheit
darzusteUen: "Pro illius questionis veritatis et falsitatis evidencia partes
arguitur ad utrasque. "39

Aus dem Gesetz von 1462 geht hervor,40 daB die Thesen der
Sophismata und die Titel fUr die Fragen, die disputiert werden soUten, yom
Dekan bestimmt wurden, was aber nieht immer beachtet wurde. Vor dem
genannten Statut wurden die Titel der Quaestionen hiiufig von dem
promovierenden Professor und dem Betreuer der Studenten bestimmt. 41
An der Disputation fiber eine sophistische These beteiligten sich mehrere
Personen, die - wie schon gesagt - Sophisten genannt wurden. Der
Defendent verteidigte die aufgesteUte These. Der Opponent verneinte sie.
Ahnlich war es in den Prfifungs-quaestionen. Doch diese wurden von dem
baccalariandus nieht unentschieden gelassen: "Magistri mei reverendi,
necnon domini baccalarii et ceteri domini, audistis quo modo domini
respondentes ad primum zophisma respondendo unus ipsorum concedit et
alter negat et quilibet ad sensum suum. Sed contra quemlibet more solito

graciarum acciones singulis dominis, qui fuerint dignati actum presentem visitare."
Ibid., f. 231r.
"In nomine Domini, amen. In actu presenti tria per ordinem sunt facienda: Primo
disputabuntur duo zophismata, secundo disputabuntur due questiones, tercio et ultimo
referende sunt graciarum acciones singulis magistris, dominis baccalariis et dominis
studentibus, qui dignati fuerint actum presentem visitare." Ibid., f. 234v. "Sophismata
et questiones iste disputate sunt in Universitate Cracoviensi anna Domini millesimo
quadringentesimo quadragesimo nono dominica die in vigilia Concepcionis glosiose
virginis Marie in lectorio theologorum per Johannem Konigsberg de Oppavia
baccalarium Cracoviensem." Ibid., f. 238v.
36/bid., f. 375r.
37"In nomine Domini, amen. In actu presenti tria nobis occurrunt peragenda: Primo
predisputabuntur duo zophismata, ad que domini, qui bus sunt assignata, respondebunt;
secundo proponentur due questiones dominis, qui bus sunt assignate, ad decidendum;
tercio et ultimo agemus inprimis omnium rerum Opifici gracias eisque, qui hunc actum
suis presenciis haud didignati fuerint illustrare." Ibid., f. 276r. Cf. Anm. 4l.
38/bid., f. 263r.
39Ibid., f. 196r.
40Cf. Anm. 25.
41"Questio mihi a reverendo magisitro promotore et preceptore mea mihi semper
observandissimo ad determinandum sub talium sermone verborum est proposita."
Krak6w, BJ, ems 2205, f. 19va; vgl. auch f. 159v.
 DIE ROLLE DER SOPHISMATA 1M UNTERRICHT 123

instabo aliquod mediis. Et primo contra illum, qui concedit. .. [Secundo]

contra negantem .... "42

Die in der erwahnten Handschrift 2205 der Jagiellonischen Bibliothek

erhaltenen Sophismata sind zwar literarisch stilisiert, doch sie spiegeln die
Atrnosphiire der Krakauer Priifungsdisputationen im 15. Jahrhundert
wider. Aus diesem Grunde sollen wir uns mindestens mit einem dieser
Sophismata naher beschaftigen. Als Beispiel wahlen wir ein Sophisma,
welches behauptet, das Universale sei eine Kategorie: "Sophisma
secundum per me disputandum erit hoc: Universale est in predicamento.
Quod zophisma probatur et improbatur."43 Zum Beweis der Wahrheit
dieser These wird die Autoritat des Aristoteles herangezogen, wonach
alles, was durch sich selbst und auf eine spezifische Art in der Kategorie
sich befindet, ein Universale ist. Dies gilt aber nicht fUr das Einzelseiende.
Somit ist das Sophisma wahr: "Probatur primo auctoritate Aristotelis in
libroPredicamentorum, qui ostendit, quod illud, quod per se et proprie est
in predicamento, est universale. Singularia enim, quia non sunt scibilia,
ergo per se non sunt proprie in predicamento. Quare zophisma verum."44
Gegen die Wahrheit dieser These wird bewiesen, daB das Universale ein
transzendenter Begriff ist, was angeblich von allen Logikem zugestanden
wird. Somit ist das Sophisma falsch. Da bereits nachgewiesen wurde, daB
es wahr ist, entsteht ein Zweifel: "In oppositum autem, quod zophisma sit
falsum, patet, quia universale est transcendens. Ergo res in quolibet
predicamento modo transcendens alia non est in predicamento, ut omnes
concedunt loici. Quare zophisma falsum, prius autem probatum verum,
relinquitur ergo dubium."45 Daraufhin werden von dem baccalariandus
fiinf Argumente gegen den Defendenten der erwahnten These vorgebracht:
"Concedenti arguitur sic .... "46 SchlieBlich werden sieben Argumente
gegen den Opponenten vorgetragen: "Neganti arguitur sic ...."47 Ein
charakteristisches Merkmal der sophistischen Argumentation ist es, daB
lediglich die Argumente fiir und gegen die vorgeschlagene These
zusammengestellt werden (arguere in utramque partem), aber nicht
entschieden wird, ob nun der Defendent bzw. die Defendenten oder ob der

42/bid., f. 344v. Vgl. auch: "Diligite et medii, scilicet baccalaurei, sapienciam

acutissime disputandi et respondendi, ut vobis opera sapientis primo Elencorum posita,
scilicet non mentiri et mencientem posse ostendere, qualiter assignentur. Diligite
sapienciam magistrorum vestrorum lecciones et exercicia visitantes, diligenter memorie
comendantes, comendatas aliis sinceriter inparcientes." Anonyma oratio rectoris,
Krak6w, BJ, cms 1587, f. 138r; cf. M. Kowalczyk, Mowy ... , S. 61.
43Krak6w, BJ, ems 2205, f. 17vb.
44lbid.
45lbid.
46"Concedenti arguitur sic: Universale non est in predicamento. Probatur: Nullum
transcendens est in predicamento, sed universale est transcendens, igitur non est in
predicamento. Maior nota, quia iIIud, quod debet poni in predicamento, debet dicere
rectam naturam, modo transcendencia non sunt huiusmodi. Minor probatur, quia
universale reperit principium quo videlicet predicamento et per consequens
transcendens." Ibid.
47"Neganti arguitur sic: Universale est in predicamento. Probatur: Species specialissima
est in genere, sed universale est species specialissima, ergo est in genere. Maior est
Aristotelis in Predicamentis. Minor patet, quia homo est universale et species
specialissima, ut tenet Philosophus." Ibid., f. 18ra.
124 MIECZYSkAW MARKOWSKI

Opponent bzw. die Opponenten recht haben. Es sei hier bemerkt, daB
weder der erste (affirmanti) noch der zweite (neganti) Teil ein integrales
Ganzes bilden, da es keinen Parallelismus der dort vorgebrachten
Argumente gibt. Der Sophist muBte seine Argumentation sowohl den
Disputanten anpassen, welche die im Sophisma enthaltene Feststellung
bestatigten, als auch denjenigen, die sie vemeinten. Dieser Umstand
bewirkte die lose Verbindung der Argumente. Anfanglich wurde das
Begriffspaar concedens-negans verwendet,48 in den spiiteren Jahren des
15. Jahrhunderts treten immer hliufiger die Bezeichnungen affirmans-
negans auf. 49 Selten wurde der Begriff opponens gebraucht. Nur an
wenigen Krakauer Sophismata beteiligten sich auch dubitantes.5o In den
Prager Sophismata des letzten Viertels des 14. Jahrhunderts dagegen traten
die dubitantes viel hiiufiger auf.

Die in der genannten Handschrift 2205 der Jagiellonischen Bibliothek

vorhandenen quaestiones sind auch in formaler Hinsicht nicht einheitlich.
Ahnlich wie in den Sophismata gab es auch in manchen Priifungs-Fragen,
die den Studenten gestellt wurden, Argumentationen fUr beide Seiten: "Ad
cuius question is partes magister meus reverendus arguebat utrasque."51
Erst nachdem die Argumente fUr und wider erwogen und die notwendigen
Bemerkungen und zusatzlichen Erkliirungen gemacht wurden, kam nam
zum Schluss: "Dimissis argumentis pro et contra adductis informacione
salubiori semper salva in huius question is decisione solito progredior
ordine: Primo ponam notabilia, secundo conc1usiones cum corrolariis, ex
quibus patebit responsio ad racionen in oppositum laborantem."52 Eine
solche Vorgehensweise kann man zu den quaestiones temptativae rechnen.
Der promovierende Magister mischte sich bei den sophistischen Thesen
nicht in den Inhalt der genannten Argumente ein, konnte aber den von dem
Studenten gezogenen SchluB in Frage stellen: "Hec sunt dicta, magister
reverende, promotor ac preceptor mihi semper observantissime, vestra
tamen informacione salubiori semper salva. "53 Die Bedenken zu den
disputierten quaestiones konnten selbstverstandlich auch andere Magister
iiuBem, die an der Disputation teilnahmen. 54 An der PrUfungsdisputation

48lbid., f. 17ra-17va, 17vb-18ra, 18v-19r, 63v-65r, 66r, 69r, 97r, 99r, I06r,109r, llOr,
IOOv, 113r, 115r, 116v, 118r, 118v, 119v, 120v, 122r-v, 124r, 126r, 127v, 128r-v,
146r, 148r, 151v, 152v, 207v-208r, 209r-v, 21Or-v, 288r, 298r, 302v.
49lbid., f. 92r, 92v, I02r, 138r-v, 139r, 140v, 150r, 154r, 156r, 168r, 171r, 172r,
180v, 181r, 187v, 189r, 19Ir, 192v, 300v.
50lbid., f. 115v, 209r.
51/bid., f. 159v.
52lbid.
53lbid., f. 16Iv; vgl. auch f. 164v und 21ra.
54"Patres observantissimi eximiique doctores magistrique venerabiles, et quamvis in
presenti et in quacumque alia materia dignitates vestre me informare possent singulari
tamen favore, quo me venerabilis magister Thomas de 01muntcz complectetur pro
habenda informacione et fundamentali presentis questionis decisionem ad eundem me
Iimitto, qui de presentis questionis materia me faciliter informabit dignemini
venerabilis magister eciam questionem per me positam caritative assumere et que vobis
ad presens necessaria videbuntur adducere. Venerabilis magister, Iicet posicio vestra
satis efficaciter posita sit, nihilominus tamen consuetudine laudabili servata contra
ipsam unico instabo medio." Ibid., f. 170r.
 DIE ROUE DER SOPHISMATA 1M UNTERRICHT 125

fiber die gestellten Fragen beteiligten sich schon Bakkalaurei,55 denen die
Rolle der Respondenten zufiel.56

Der Aufbau der Bakkalaureus-quaestio bei der Prfifung erinnert etwas

an die Sophismata, was z.B. bei der quaestio tiber die substantielle
Existenz der Form sichtbar ist: "Utrum forma substancialis ante sui
generacionem habeat aliquod esse substanciale in materia distinctum ab
eodem?"57 Bei ihrer Losung wurde zuerst die Wahrheit des bejahenden
Teils bewiesen: "Ad cuius partes arguitur. Et primo pro prima parte
affirmativa ... "58 Dann argumentierte man fUr die Wahrheit des
vemeinenden Teils: "In oppositum pro prima parte negativa, quod non
arguitur... "59 Eine quaestio, ftir die Argumente pro und contra angeffihrt
wurden, betrachtete man als zweifelhaft: "Et sic questio pro utraque parte
probata relinquitur ergo fore dubia etc."60 In dieser Situation muBte man
zuerst denjenigen Disputanten antworten, welche die bejahende
SchluBfolgerung vertraten: "Ponenti conclusionem affirmativam arguitur
sic ... "61 Daraufhin wurde denjenigen geantwortet, welche fUr die
vemeinende Konklusion optierten: "Ponenti conclusionem negativam
arguitur sic .... "62 Die fUr jede der beiden Seiten genannten Argumente
waren weder ihrer Anzahl noch ihrem Inhalt nach parallel. Eine solche
Argurnentationsweise ist also den quaestiones temptativae zuzuordnen. Bei
solchen quaestiones wurde oft am Anfang bemerkt, daB die quaestio durch
bejahende und vemeinende Argumente zugleich bewiesen und widerlegt
wird: "Hec questio probatur et improbatur statim"63 oder "Ad cuiusquidem
questionis evidenciam arguitur partes utrasque."64 So wie in den
sophistischen Thesen erschienen auch in den Prfifungsquaestionen die
Begriffe concedens bzw. affirmans und negans. 65 Es wurden auch
Argumente gegen die bejahende und vemeinende Konklusion

55"Venerabiles magistri, promotor dignissime, dominique baeealarii et si non pretereant

diginitates vestras zophismata et questiones per me iam disputatas, ut eo tamen tenarius
imprimantur...." Ibid., f.213v.
"Aggredior ultimum et primum gracias ago omnipotenti Deo eiusque Genitriei
omnibusque eelieolis, quorum numine hune letum adepti sumus exitum, demum ad
terrena me divenites (7) grates ago venerabili viro magistro lohanni de Osswyancim
promotorique mea dignissimo, magistris et dominis baeealariis neenon studentibus
preeipue respondentibus, qui bus me pie nunc pollieeor in omni obsequio eis pro
futuro." Ibid., f. 214r.
56"Dieant ergo domini respondentes, quibus assignatum est primum zophisma." Ibid., f.
344v.
57lbid., f. 211v.
58lbid.
59lbid.
60lbid., vgl. aueh f. 215r, 22Or, 236r.
61lbid., f.21Iv.
62/bid., f. 212v.
63/bid., f. 234v.
64lbid., f. 2SOr.
65Cf. ibid., f. 2Sr, 3lv, 33r, 35r, 37v, 45r. 46r, 57v, 60v, 66r, 92r, 92v, 97r, 99r, IOv,
IOlv, l09r, 113r, 116v, llSv, 120v, 125r, 126r, 13Sr-v, 140v, 142v, 143r, 144r-v,
146r, 14Sr, 150r, 15lv, 154r, 155r, 156r, 16Sr, 172r, 174r, 19lr, 192v, 210r-v, 235r-
v, 2Slr, 300v, 302v.
126 MIECZYSkAW MARKOWSKI

vorgebracht. 66 Uber den Disputationsverlauf wachte der Dekan der

Artistenfakultat. Aus diesen GrUnden wurde ihm gegenUber am Ende der
PrUfungsdisputation ein besonderer Dank ausgesprochen. 67

Die Strukturen der Sophismata und der PrUfungsquaestionen waren im

15. Jahrhundert iihnlich. Diese Ahnlichkeit bildete sich infolge der an der
Universitat herrschenden Sitten und nicht infolge der Universitatsstatuten
heraus. In redaktioneller Hinsicht bestanden keine groBen Unterschiede
zwischen ihnen und den Schemata der a,n anderen mitteleuropiiischen
Hochschulen verfaBten Sophismata und Quaestionen, da sie aile ihren
Ursprung in den logischen Schriften des Aristoteles hatten. 1m achten
Buch der Topik gibt es Regeln zur Problemlosung, welche in den
scholastischen Disputationen entfaltet wurden. 68 1m Traktat Uber
sophistische FehlschlUsse werden vier Beweisarten unterschieden: eine
streng wissenschaftliche, eine dialektische, eine untersuchende und eine
sophistische.69 Diese Beweisarten haben ihre praktische Anwendung in
den schriftlich niedergelegten wissenschaftlichen Werken und in den
Universitiitdisputationen gefunden. Zwar ist im 15. Jahrhundert an der
Krakauer Universitiit kein Traktat De obligationibus entstanden, welcher
die Aufgaben des Fragenden und des Antwortenden naher bestimmte, doch
in den logischen Werken, die in Krakau verwendet wurden, war von vier
Disputationsarten die Rede: "Omnis disputacio ordinatur in aliquem finem
vel ergo ad generandum scienciam, vel ad acquirendum opinionem, vel ad
habendum experienciam de aliquo, utrum ipse aliquid scit vel non, vel ad
obtinendum victoriam. Si primum, sic est disputacio doctrinalis; si
secundum, sic est disputacio dyalectica; si tercium, sic est disputacio
temptativa; si quartum, sic est disputacio sophistica vellitigiosa."70 Die
BakkalaureusprUfung, die unter Aufsicht des promovierenden Magisters
einer Disputation iihnelte, prUfte nicht nur die in den Vorlesungen
(leet;ones) erworbenen Kenntnisse, die sich auf die als expositio litteralis,
expositio textus oder quaestiones doctrinaLes verfaBten Werke stUtzten und
zum sicheren Wissen fUhrten. Sie prUfte auch die Fahigkeiten zur
Disputation, die wiihrend der Ubungen (exercitia) und der allgemeinen
Disputationen (disputationes communes) erworben waren und an das

66Cf ibid., f. 9Sv, 107r-108v, 1I6v, 167v, 17Sr, 199v, 21Sr, 220v-22Ir, 237r-238r,
289v, 290v, 338v.
67"Quantum ad uItimum ne ingratitudinis accusaremus vieio, habemus et agimus eas,
quas valemus immortales gracias, in primis summo Opifici Deo, quod nos sua
invisibili gracia hunc actum laudabilem feliciter obiire concesserit. Deinde venerabili
viro domino decano arcium ceterisque magistris baccalauriisque reverendis, qui suis
presenciis ob decorem nostri actus huc sese conferre haud dedignati sunt, agimus et
habemus mille gracias. Denique omnibus ingenuis adolescentibus, qui nostras et si
pueriles disputaciones ascultari minime aspernati sunt, qui bus omnibus pollicemur nos
morem genere et obsequi in omnibus lieitis et honestis." Ibid., f. 238r.
68U. Gerber, Disputatio, S. 14.
69Aristoteles, llEpl uo<!>uTTlKW/I l>.£y'W/I, in Aristotelis Opera, ed. Academia Regia
Borusica, Aristoteles Graece ex recognitione Immanuelis Bekkeri, T. I, Berolini 1831,
2, 16Sa 38 - 16Sb 11; cf. M. Markowski, Burydanizm w Polsce w okresie
przedkopernikaflskim. Studium z historii JilozoJii i nauk scislych na Uniwersytecie
Krakowskim w XV wieku, Wrodaw 1971, S. 44-S.
70Ioannes Buridanus, Quaestiones super "De sophisticis elenchis" Aristotelis, Krak6w,
BJ, cms 736, f. S4rb; cf. M. Markowski, Burydanizm ..., S. 4S.
 DIE ROUE DER SOPHISMATA 1M UNTERRICHT 127

Schema der quaestiones dialecticae ankniipften, die nur zu einer

Uberzeugung (opinio) flihrten. In der disputatio temptativa ging der
geprUfte baeealaureandus von allgemein anerkannten Gesetzen aus71 und
konnte Argumente aller Art verwenden.72

1m Hinbliek auf den Inhalt der Sophismata muB bemerkt werden, daB
die Sophismata logischen Inhalts den ersten und die grammatisehen Inhalts
den zweiten Platz einnehmen. Nicht besonders zahlreich sind die
Sophismata aus dem Bereich der Naturphilosophie, zu der im Mittelalter
aueh die Psychologie zahlte. Bei den Priifungsquaestionen dagegen steht
die Naturphilosophie an erster Stelle. Gelegentlich treten auch
metaphysisehe und astronomische Themen auf. Andere
Wissenschaftsgebiete waren sowohl bei den Sophismata als aueh bei den
quaestiones weniger vertreten. Dieses breite Themenspektrum war nieht
ohne Bedeutung fUr die doktrinaren Veriinderungen an der Krakauer
Universitiit im 15. lahrhundert. In derartigen zu schulischen Zwecken
veranstalteten Disputationen konnten die versehiedensten Fragen kUhn
beantwortet werden. Solche Antworten, die von den allgemein anerkannten
Losungen abwichen, bahnten den Weg fUr neue Ansichten. Den
Sehuldisputationen sollte deshalb - so meine ich - mehr Aufmerksarnkeit
gesehenkt werden, als dies bisher der Fall gewesen ist.

Polish Academy of Sciences. Cracow

71"Et disputacio est quadruplex, scilicet doctrinalis, dyalectica, tentativa et sophistica.

Disputacio doctrinalis est, que procedit ex principiis veris et necessariis alicuius
sciencie et hec habetur inter doctorem seu informatorem et discipulum. Et taIis
disputacio fit, ut generetur disciplina in discipulo et instrumentum eius est syllogismus
demonstrativus, de quo syllogismo agitur primo Posteriorum. Disputacio dyalectica
est, que procedit ex principiis probabilibus, et taIis habetur in actibus ordinariis
magistrorum vel baccalariorum et fit, ut generetur opinio in discipulo. Et
instrumentum eius est syllogismus dyalecticus, de quo in libris Thopicorum.
Disputacio tentativa est, que procedit ex principiis communi bus, qui bus aliquis
tentatur, an sit hic, qualem se existimat et habet fieri in examinibus magistrandorum
vel baccalandiorum. Et fit propter experienciam de respondente, an sit sciens vel
ignorans et instrumentum eius quandocumque est argumentum. Sed disputacio
sophistica est, que procedit ex apparentibus seu ex principiis apparenter veris cuius
instrumentum est syllogismus sophisticus, de quo primo Elencorum et habetur
communiter inter volentes apparere scientes et finis eius est vana gloria." Commentum
Cracoviense super tractatum "De consequentiis," Krak:6w, BJ, cms 2591, f. 94r; vgl..
auch cms 2178, f, 6Ov; M. Markowski, Burydanizm ... , S. 45-6.
72Exercitium novae logicae seu librorum "Priorum" et "Elenchorum" magistri Joannis
de Glogovia pro iunioribus recollectum ac noviter emendatum, Cracoviae 1511, f. 69r-
7Or; vgI. M. Markowski, Burydanizm ... , S. 82-3.
 The Sophismata Asinina of William Heytesbury

by Fabienne Pironet (Aspirant FNRS)

Authorship of the Sophismata Asinina

In his well-known book on Hey tesbury, I Wilson says that the

attribution of the Sophismata asinina to him is uncertain. However, several
criteria allow me to claim that the work was actually written by Heytesbury
himself.

1° The manuscripts I have seen contain almost exclusively tracts

written by Oxford masters. We can therefore conclude that the
Sophismata asinina is also due to an Oxford master.
Furthermore the exercise of variation is typically Oxonian.

2° Three manuscripts (P, M and F) explicitly mention the name of

Heytesbury either in the incipit or in the explicit (cf. infra).

3° The way of arguing in the two works (Sophismata and Sophismata

asinina) is the same, that is, a chain of proofs:
- Statement of a sophisma in consequence form:

- First argument:

- proof of the consequence,

- proof of one of the premises in consequence form,

- proof of the first premise of this last consequence,

- proof of the second premise of this last consequence,

- etc.

- Second argument:

Solved in the same way, and so for all the arguments.

4 ° The MS W, which is the most complete, as we shall see, proposes

different ways of solving the same sophism and concludes by
saying "the two answers are satis probabiles, ideo elige."

This kind of conclusion seems to be characteristic of Heytesbury. We can

find parallel passages in his Sophismata:

Ie. Wilson, William Heytesbury. Medieval Logic and the Rise of Mathematical
Physics, Madison: the University of Wisconsin Press 1960.

128
 THE SOPHISMATA ASININA OF WILLIAM HEYTESBURY 129

Soph.1 ff.77va-77vb: 2 Sed nullum est argumentum cogens ad

negandum aliquam ilIarum, unde sicut probabiliter negatur ilIa, ita
probabiliter sustinetur eadem ... , ideo elige.
Soph.l ff.80ra 4-6: Posset enim probabiliter concedi vel negari,
ideo elige. Utraque enim via satis probabiliter teneri potest.
Soph.5 f.95vb 4-6: Prior responsio est certior quam secunda,
utraque tamen quoddam modo probabilis, ideo elige.
Soph.5 f.97ra 38: Ex his omnibus elige probabile.

In support of these arguments, I make the hypothesis that the works

Heytesbury devoted to sophisms can be divided as follows:

1° Regulae solvendi sophismata: intended for the students in their first

year, it mixes problems that are the concern of logic and of natural
philosophy and provides the students with means to solve every
kind of sophism.

2° Sophismata: intended for more advanced pupils, it essentially deals

with problems of natural philosophy. The discussion of the
sophisms is often long and the argumentation very detailed.

3° Sophismata asinina: it would be the matching piece to the

Sophismata for problems of logic. The mode of presentation is
different from that of the Sophismata since it proposes about thirty
variations on the same statement Tu es asinus put successively in
different cases.
The manuscripts

I am preparing a critical edition of the Sophismata asinina based on a

collection of six manuscripts. Other manuscripts3 also contain this text but
I have not seen them yet.

F Firenze BN Centro V 43 ff.45a-46b (XVth C.)

M Miinchen Clm 19672 F.145 ff.306va-314va (XVth C.)
P Pistoia Cattedrale.61 ff.55ra-57va (end XlVth C.)
Pa Padua Bibl. Univ. 1123 ff.18ra-22vb (end XlVth C.)
V Vienna Oest.Nat. VPL 4698 ff.99r-103v (end XIVth C.)
W Worcester F.118 ff.20ra-24vb and4
48rb-51va (end XVth C.)

21 cite here from the Venice edition 1491, f. 77-170.

30xford, Canon. lat. 278 ff. 83ra-87vb;
Padua Bibl. Univ. 1570 ff. 113r-130v;
Rome Cas. 98 ff. 120-134;
Venice San Marco Lat. Z 310 (1577) ff. 122-126.
4The tract is interrupted by the Suppositiones cOllsequentiarum of Robert Alyngton.
130 FABIENNE PIRONET

Incipit:

M: Sophysmata Hesbri asinina.

Pa: Incipiunt argumenta asinina.
F: Incipiunt sophismata asinina edita ab eodem Hentisbero.
Explicit:

W : et sic est finis.

Pa: expliciunt argumenta asinina juveni valde utilia.
P: expliciunt sophismata Reverendi artium doctoris Tysbari.
M : et sic finita sunt sophismata Hesbri asinina. Anno 1381 in festo
sancti Augustini.
V: the text is interrupted in the middle of the nineteenth sophism.
F: the text is interrupted in the middle of the sixth sophism.
Description of the manuscripts:

De Rijk gives a description of P, Pa, V and W.s

M: f.l In quo continentur Petri Hispani tractatus loycalibus et aliae

materiae loycales (Johannes Muntzinger). (This note dating from
1494).
F: f.] r-12r Inc.: Incipit tractatus Magistri Riccardi de Bilingam.
Terminus est in quem resolvitur propositio ...

Expl.: Explicit tractatus a Magistro Riccardo Bilingam compositus qui

speculum iuvenum nuncupatur quasi Bilingam, id est velut
cantator veritatis et falsitatis propositionum. Deo gratias. Amen.

12r-25v Exp1.: Expliciunt obligationes edictae ab Eximio Magistro

Petro Candie teologiae doctore.

26r-32v Inc.: Incipiunt obiectiones consequentiarum. Deo gratias.

Amen. Causae consequentiarum sicut sunt obiectiones sunt
ponendae et solvendae.

Expl.: Expliciunt obiectiones consequentiarum edictae a quodam probo

doctore loycae (Martinus Anglicus).

33r-38r Inc.: Incipiunt consequentiae Magistri Johannis Buridani.

Consequentia est antecedens et con sequens cum nota
consequentiae, vel est aggregatum ex antecedente et consequente
cum nota consequentiae.

sef. LM. De Rijk, Some 14th Century Tracts on the Probationes terminorum,
Nijmegen: Ingenium Publishers 1982: Pa, p. 31; P, p. 35; V, p. 9; and "Logica
oxoniensis. An Attempt to reconstruct a fifteenth century manual of logic", Medioevo
3, 1979: W, p. 124.
 THE SOPHISMATA ASININA OF WIlLIAM HEYTESBURY 131

Expl.: Expliciunt consequentiae Magistri Iohannis Buridani.6

38r-44v Inc.: Incipit tractatus Hentisberi Anglicani de sensu composito

et diviso. Arguendo a sensu composito ad sensum divisum, et e
contra, frequenter fallit argumentum.

Expl.: Explicit tractatus de sensu composito et diviso editus ab

excellentissimo Magistro Hentisbero Anglicano vocato
Guigleloyus.

45r- 46v Inc.: Incipiunt sophismata asinina edita ab eodem Hentisbero.

Tu es asinus. Probatur sic: ille homo est asinus, te demonstrato;
ergo tu es asinus.

Expl.: The text is interrupted in the middle of the argumentation ... quia
illa dictio 'si' habet vim distribuendi.

47r-70v Inc.: Ob rogatum ... aliqua utilia artis obligatorie ex dictis

antiquorum necnon modemorum divina favente gratia yolo
breviter compilare.

Expl.: Expliciunt obligationes Magistri Iohannis Buser.

Stemma
PWMPaVF
................... ···········:wMPaVF
P

W / MP"::vF

MPa/ \V \

\ Pa
1. Proof of the independence of P.

In many cases, P has a mode of argumentation different from the other

manuscripts: it lacks some arguments (example 1), abbreviates others
(example 2) or completely reverses them (example 3).

The two first characteristics are very frequent; the last has been found at
just one place.
Example 1. (Omission of arguments)
Sophism 1: lste homo est asinus, te demonstrato; ergo tu es asinus.

6As Professor Hubien has shown, this text cannot be attributed to Buridan. but rather to
an English master. Cf. Johannis Buridani tractatus de consequentiis. Edition critique,
Louvain-Paris: Publications Universitaires 1976, p. 7.
132 FABIENNE PIRONE[

Fourth argument. This example also shows clearly the strong relation
between the MSS W, M, Pa, F and V.

W: Item sic: quaelibet M: Item quaelibet Pa: Item sic: quaelibet

singularis est vera; singularis est vera; singularis est vera;
ergo quaelibet ergo quaelibet ergo quaelibet
singularis alicujus singularis alicujus singularis alicujus
universalis est vera. universalis est vera. universalis est vera.

V: Item sic: quaelibet F: Item, quaelibet P:om.

singularis est vera, et singularis est vera;
nulla est singularis ergo quaelibet
nisi quae est singularis alicujus
singularis alicujus universalis est vera.
universalis; ergo
quaelibet singularis
alicujus universalis
est vera.

W: Consequentia M: Probatur Pa:Probo

patet: quia sequitur consequentia: sequitur consequentiam:quia
quaelibet singularis quaelibet singularis sequitur quaelibet
est vera, et nulla est est vera, sed nulla est singularis est vera, et
singularis nisi alicujus singularis nisi alicujus nulla est singularis
universalis; ergo universalis; ergo nisi quae est
quaelibet singularis quaelibet singularis singularis alicujus
alicujus universalis alicujus universalis universalis; ergo
est vera. est vera. quaelibet singularis
alicujus universalis
est vera.

V: contains the F: Consequentia P:om.

probatio in the satis patet, ideo et
utterance of the cetera.
argument itself.

At this place, Pa adds an argument that is not clear. It seems to be a mix-up

of both arguments and I propose to omit it: Minorem probo: quia nulla est
singularis nisi quaelibet est singularis alicujus universalis; ergo quaelibet
singularis aIicujus universalis est vera.

W: Consequentia M: Consequentia patet Pa: Minorem probo:

patet et minorem et minor probatur: quia nulla est
probo: quia nulla est quia nulla est singularis nisi alicujus
singularis nisi alicujus singularis nisi alicujus termini communis,
termini communis, termini communis, et sed quilibet terminus
sed quilibet quilibet terminus communis est
<terminus> communis est universalis; ergo nulla
communis est universalis; ergo nulla est singularis nisi
universalis; ergo nulla est singularis nisi alicujus universalis.
est singularis nisi alicujus universalis.
alicujus universalis.
 THE SOPHISMATA ASININA OF WIUIAM HEYTESBURY 133

V: Minorem probo: F: om. P:om.

quia nulla est
singularis nisi alicujus
termini communis, et
quilibet terminus
communis est
universalis; ergo nulla
est singularis nisi
alicujus universalis.
Example 2. (Abbreviation of arguments)

Sophism 4: Omnis asinus est asinus; ergo tu es asinus.

Second argument: Omnis asinus est asinus; ergo tu qui es asinus es asinus.

Pa here represents the family it belongs to.

Pa: Consequentiam probo: quia si P: Tenet argumentum: quia si non

non valet, stet oppositum valet, tunc sequitur oppositum
consequentis cum antecedente, consequentis, et oppositum est 'tu
scilicet 'tu qui es asinus differs ab qui es asinus differs ab asino'.
asino'. Et tunc sic: 'tu qui es
asinus differs ab asino; ergo tu
qui es, asinus es, et asinus est, et
tu qui es asinus non es asinus'; et
tunc ultra: 'tu qui es, asinus es, et
asinus est; ergo tu qui es asinus
es asinus', et ultra 'ergo tu es
asinus'.

Sed forte dicitur quod hoc datum

non est ejus oppositum, sed hoc:
'tu qui es asinus non es asinus' .
Example 3. (Inversion of an argument)

Sophism 31 (W31, Pa31, M31, P5): Tu es asinus vel tu non es

asinus, sed tu non es asinus; ergo tu es asinus.

The argumentation can be summarized as follows:

1 tu es asinus vel tu non es asinus,

0

sed tu non es asinus;

ergo tu es asinus.

This consequence is valid quia arguitur a tota disjunctiva cum opposito

unius partis ad alteram partem.

2 tu es homo vel tu es asinus,

0

sed tu non es homo;

ergo tu es asinus.
134 FABIENNE PIRONE[

This consequence is valid according to the same rule. The divergence now
appears:

W,Pa: P: Si autem negatur illa

Sed forte dicitur quod illa regula <regula> et dicitur quod non
non valet ubi disjunctiva fit ex tenet nisi ubi pars contradicat...
contradictoriis, et sic non valet
consequentia prima (1.).

Ex illa responsione sequitur quod Ex illa responsione sequitur

numquam valet consequentia a
tota disjunctiva et cetera.
Nam non sequitur
tu es homo vel tu es asinus,
sed tu non es homo;
ergo tu es asinus
- quia antecedens fit ex
contradictoriis.
- quia antecedens est verum et
consequens falsum.
quod non valet ista
consequentia
tu es homo vel tu es asinus,
sed tu non es asinus;
ergo tu es homo,
quod est falsum.
Therefore the consequence 2· is
valid: quia consequentia est
bona et antecedens verum; ergo
et consequens.

Quod antecedens (tu es homo vel Quod antecedens (tu es homo

tu es asinus) sit verum probo:
quia est una disjunctiva cujus
prima pars (tu es homo) est vera.
vel tu es asinus, sed tu non es
homo) sit verum probatur: quia
est una disjunctiva cujus una
pars est vera; ergo antecedens
est verum.

Both texts agree to reject the first argument because in the consequence 1·
the rule a tota disjunctiva et cetera is not directly applied. A correct
application of the rule would give the following consequences:

tu es asinus vel tu non es asinus, tu es asinus vel tu non es asinus,

sed tu es asinus; sed tu non es asinus;
ergo tu es asinus. ergo tu non es asinus.

The response to the second argument is again different: Wand Pa say

that the consequence 2· must be doubted in the following way:

- si arguitur a prima nota: tu es homo (vel tu es asinus),

sed tu non es homo;
ergo tu es asinus,

the consequence must be denied quia consequentia non valet ubi disjunctiva
fit ex contradictoriis.

- si arguitur a secunda nota: (tu es homo vel) tu es asinus,

sed tu non es homo;
ergo tu es asinus,

the consequence is valid quia antecedens est impossibile.

 THE SOPHlSMATA ASININA OF WILLIAM HEYTESBURY 135

To the second argument, P answers: negatur antecedens pro prima

parte, scilicet tu non es homo. It adds that the whole antecedent is not a
disjunctive proposition but a copulative one because' sed' conjoins.

For completeness, I should mention the position of the MS M. It sets

out the rule just as P does and proves the consequence 1 by saying tota
0

disjunctivafit ex contradictoriis; ergo per regulam consequentia tenet de se.

The rest of its argumentation is exactly the same as Pa. I have not found
any other trace of errors shared by P and M that could allow me to conclude
to a contamination of the second by the first. On the other hand its
relationship with Wand Pa are so strongly established that I conclude here
the error is its own.
2. Proof of the independence of W.

In many places, the text of W follows its arguments to their end where
other texts do not (ex. 1), offers the reader additional explanations and more
detailed argumentation or reminds him of several rules (ex.2). It also
discusses opinions of quidam and proposes an answer to them. All these
facts make me think that W contains the text as it was actually written by
Heytesbury .

Example I. (Complete argumentation)

Sophism 6 (W6, M5, Pa5, F5, PlO): Pono quod A convertatur cum
isto terminG 'homo', et B cum isto terminG 'equus', et hoc disjunctum 'A
vel B' cum isto terminG 'asinus'. Isto posito, propono 'tu es A; ergo tu es
asinus' .

M,Pa,F: W: Pom.
Item sic probatur: Item sic probatur:
tu es A, tu es A,
et omne A est terminus; et omne A est terminus;
ergo tu es terminus, ergo tu es terminus,
et, per consequens, tunc sequitur
tu es res inanimata, tu es terminus;
cum omnis terminus sit res ergo tu es res inanimata;
inanimata. ergo tu es non homo,
et tu es;
ergo tu es asinus,
quod fuit probandum.
Example 2. (additional rules)

In sophisms 17 (Pono tibi hunc casum: tu es asinus vel ille casus est
impossibi/is) and 18 (lila consequentia est bona; ergo tu es asinus,
demonstrando per ly 'ilia' eandem consequentiam), Heytesbury recalls that
before saying whether the sophism is true or false we have to specify
which casus the relative term 'ille' refers to. If 'il/e' refers to the casus it
belongs to, then the casus must be denied because, in this case, the
136 FABIENNE PIRONET

restrictive rule, applicable to every insoluble proposition,7 would be

transgressed.

I have chosen this example because it mentions a rule that is not found
anywhere in the Regulae solvendi sophismata, chapter one De
insolubilibus. Does this mean that Heytesbury has changed his mind on the
subject? I do not think he did.

We know that in England most authors solved the insolubles by

applying to them a restrictive rule. This was not said explicitly in the
Regulae so/vendi sophismata, but it is clear that what Heytesbury had in
mind was a restrictive rule. Indeed, here is what we can read in the first
chapter devoted to insolubles: propositio insolubilis est de quafit mentio in
casu insolubili quae, si cum eodem casu significet sicut verba illius
communiter praetendunt, ad eam esse veram sequitur eam essefalsam, et e
converso. It is clear that if a sentence precisely signifies as its words
commonly do, a self-reflection results which is not allowed.

But while other authors said that in an insoluble proposition like 'ego
dico falsum' the term 'falsum' has to refer to another proposition from the
one it is a part of or that the tense of the verb has to refer to a previous
time, Heytesbury just says that an insoluble proposition should not be
admitted by the respondent, and then the respondent has in no way to
argue about it. Heytesbury then inaugurated a new way of solving
insolubles which could be called 'the obligational solution'.

In the Regulae, Heytesbury discusses several opinions and proposes

five rules to solve the insolubles. In the Sophismata asinina, we have, in at
least two places, an application of these rules and the solution proposed is
more detailed. In the response to the sixteenth sophism - ponitur tibi ista
'tu es asinus vel ille casus est impossibilis' - we read:
quando ponitur ille casus, ... dicit sophista 'non intelligo
casum', et causa est quia ad hoc quod intelligatur, oportet quod
aliqua res significata per subjectum sit distincta a praedicato, sed
non est hujus modi: quia signijicata tali impositione 'ista propositio
est vera', nul/us intelligit illam magis esse veram quam istam 'bu et
ba'.

I think that this kind of solution is very close to that of the cassantes, who
used to say that whoever says 'ego dico falsum' does not say anything
(nihil dicit).
3. The relationship between M and Pa.

Here again the examples are very frequent. I propose two of the most
characteristic.

1° Sophism 9 (W9, M8, Pa8, P6) Tu non differs nisi ab asino; ergo tu es
asinus.

7per partem non potest demonstrari totum cujus est pars in una de insolubilibus.
 THE SOPHISMATA ASININA OF WILLIAM HEYTESBURY 137

W:
Consequentia patet per
assimile. Nam sequitur
tu vides hominem,
et tu non vides aliud quam
hominem;
ergo tu non vides nisi
hominem.
M,Pa: P:
Consequentia patet per Tenet argumentum: quia
as simile. Nam sequitur sequitur
tu vides lohannem, tu non vides aliud ab
et tu non vides alium homine;
quam lohannem;
ergo tu non vides nisi ergo tu non vides nisi
lohannem. hominem.

2° W omits a sophism that is discussed by M32, Pa32 and P6: Haec est
vera 'tu es asinus' quae praecise significat te esse asinum; ergo tu es
asinus.

In fact, this utterance is the same as that of the sixteenth sophism but
the argumentation is completely different.
4. The independence of M.

As we have seen earlier, MS M contains several additional sophisms.

It also presents particuliar variants, right or wrong.

- Soph. 17: - Antichristus potest esse W, Pa / deus non potest esse M

- ego sum asinus W, Pa / tu es asinus M

- Soph. 18: - antecedens <non> est intelligibilis W, Pa / antecedens

non est intelligibilis M

Etc.
Comparative table of the sophisms

It is interesting to note that the classification of the sophisms

corresponds to the stemma: independence of P, close connection between
W, M and Pa and between M and Pa, independence of M.

W M Pa F V P
1 1 1 1 1 1
2 2 2 2 2 2
3 3 3 3 3 3
4444 44
5 9 9 7
6 5 5 5 11
7 6 6 6 13
8 7 7 21
9 8 8 6
10 10 10 26
11 11 11 27
12 12 12 28
13 13 13 29
138 FABIENNE PIRONET

W M Pa FV P
14 14 14 30
15 15 15 31
16 16 16 32
17 17 17
18 18 18
19 19 19 33
20 20 20 34
21 21 21 35
22 22 22 12
23 23 23 15
24 24 24 10
25 25 25 14
26 26 26 16
27 27 27 17
28 28 28 18
29 29 29 19
30 30 30 20
31 31 31 5
32 32 8
32 33 33 9
33 34 34 22
34 35 35 23
35 36 36 24
36 37 25
37
38
39
40
41
42
43
44
45
46
Four remarks on this table:

1. Even though P is so different from the other manuscripts, it

contains five sequences of sophisms that are the same as the others.

2. In every manuscript, the four first sophisms are the same.

3. The strong relationship between M and Pa based on so many

paleographical reasons is also confirmed by the fact that they both
contain a sophism (number 32) omitted by W.
 THE SOPHISMATA ASININA OF WIUJAM HEYTESBURY 139

4. The very short sophisms proper to M confirm its independence. I

assume that they are additions made by a zealous student.
Hypothesis on the transmission of the text
As we have seen, several arguments allow me to claim that the text of
W is a transcription of a course given by Heytesbury himself. Pa is a
summary of W written in England and brought back to Italy by one of the
pupils who studied in Oxford.

F, which is an Italian manuscript, would be a copy of Pa.

P was also written in Italy. How it became so different from the

original text is unclear to me. It is perhaps a transcription of a disputation
made by students around the proposition 'tu es asinus' or a copy by a very
unscrupulous student. In the style of Heytesbury one could say, "Both
hypotheses are satis probabiles, ideo elige."

V was written in Prague. The manuscripts in this town often came

from Erfurt and we know how strong was the English influence in Erfurt.

I do not know anything about the origin of M nor about the place
where it was written, but it is so close to P a that it seems reasonable
enough to maintain that their archetype was the same.

Anyway, all these hypotheses should be confmned by an exact study

of the history and the transmission of the texts. Comparison must also be
made with the other manuscripts mentioned supra.
A few words about variations
Very little is known about the special type of exercise called variatio.
The only information I have collected is to be found in the work of Fletcher
on The Teaching and Study of Arts at Oxjord: 8 "One of the most common
of the academic exercises that appear in the university Grace Books is the
variation. Its nature has puzzled many scholars." Fletcher then refers to the
Statuta Antiqua Universitatis Oxoniensis edited by Gibson.9 In the statutes
of 1458 we read:
Supplicat ... Ricardus Ferre, scolaris facuItatis arcium, quatinus
tres anni cum uno termino in eadem facultate, duae variaciones in
parviso possint sibi stare pro completa forma quod possit admitti ad
lecturam alicuius libri facuItatis arcium. Haec gracia est simpliciter
concessa et legitime pronunciata.

This is the only text, as far as I know, that mentions the exercise of
variation.

8J.M. Fletcher, The Teaching and Study of Arts at Oxford 1400-1520, thesis presented
for the degree ofD. Phil. in the University of Oxford, 1960. See pp. 120-3.
9S. Gibson, Statuta Antiqua Universitatis Oxoniensis. Oxford: Clarendon Press 1931,
p. xciii.
140 FAB1ENNE PIRONEI'

Based on MS Lambeth 221, Fletcher gives a description of the way

this kind of exercise was organized: "This opens with a short introduction
which sets out the two questions which are going to form the central
feature of the variation. In this case they are first an pluralitas formarum sit
ponenda in eodem propos ito, and secondly numquam taUs pluralitas
repugnet unitati propositi. There follows a short discussion of the manner
in which the questions are to be approached. The first is to be discussed
under three headings of articuli, and the second under two. There (sic) are
also set out. The centre-piece of the exercise, a discussion of these articuli
then follows. At the termination of each section, the varier presents his
conclusions in the form of a rhyming verse. The exercise is concluded by a
general summary of the authorities used during the variation, and an
acknowledgement by the varier of how much he has depended on these
texts."

To summarize, what we know about variations is:

1. they were usually made by scholars who were studying for the
bachelor's degree;

2. they were made in parviso, that is to say that they were made by
advanced students;

3. they began with two questions and concluded with a rhyming verse;
4. the varier concluded his exercise by mentioning the authorities he
referred to and saying how much he was indebted to them.

Does our text correspond to this description?

The two first characterizations fail because our text has been written,
as I have shown, by Heytesbury himself and in a manner that allows me to
claim that he was not a student when he wrote it. The two last also fail
since our text does not contain preliminary questions nor a conclusion in
the form of a rhyming verse. On the contrary, it begins directly with the
first casus: Iste homo est asinus, te demonstrato; ergo tu es asinus and
ends with the response to the last argument of the last casus.

It is therefore clear that, according to the description given in MS

Lambeth 221, our text is not an exercise in variation. Further investigations
will be necessary to determine in which category the Sophismata asinina
could be classified.

To conclude, I would say that the following questions remain open:

are the texts written on the model of "variations on the same sentence"
current in the medieval literature? Would they be a kind of living exercise?
What role did variations play in the medieval cursus?
University of Liege
 THE SOPHISMATA ASININA OF WILLIAM HEYTESBURY 141

Appendix
List of the sophisms according to W

1. ISle homo est asinus, te demonstrato; ergo tu es asinus.

2. Tu es aliquid; ergo tu es asinus.

3. Omnis asinus est asinus; ergo ille asinus est asinus demonstrato teo (W
omits the first 'asinus' of the conclusion).

4. Omnis asinus est as in us; ergo tu es asinus.

5. Tu es non homo; ergo tu es asinus.

6. Pono quod A convertatur cum isto termino 'homo', et B cum isto

termino 'equus', et hoc disjunctum 'A vel B' cum isto termino
'asinus'. Isto posito, propono 'tu es A'; ergo tu es asinus.

7. Si tu es animal, tu es asinus, sed tu es animal; ergo tu es asinus.

8. Aliqua propositio est, et si illa sit vera significando praecise, tu es

asinus; sed aliqua propositio est et illa est vera; ergo tu es asinus.

9. Tu non differs nisi ab asino; ergo tu es asinus.

10. Tu es homo et tu es asinus; ergo tu es asinus.

11. Tantum tu es asinus; ergo tu es asinus.

12. Sit A ista propositio 'tu es asinus' et B ista disjunctiva 'tu es asinus vel
deus est', et isto posito arguitur sic: A est necessarium: quia Best
necessarium et omne A est B; ergo A est necessarium.

13. Impossibile est falsum et tu es asinus; ergo tu es asinus.

14. Tu es filius asini; ergo tu es asinus.

15. Uterque istorum est asinus, et tu es alter istorum; ergo tu es asinus.

16. Ista propositio est vera 'tu es asinus' quae praecise significat te esse
asinum; ergo tu es asinus.
(arg.: omnis propositio est vera cujus contradictorium est fa1sum)

17. Pono tibi hunc casum: tu es asinus vel ille casus est impossibilis.

18. llla consequentia est bona; ergo tu es asinus, demonstrando per ly 'illa'
eandem consequentiam.

19. Omne animal est asinus, tu es animal; ergo tu es asinus.

20. Omnis homo est asinus, tu es homo; ergo tu es asinus.

142 FABIENNE PIRONEI'

21. Tu non differs ab asino, et tu es, et asinus est; ergo tu es asinus.

22. Tu es frater asini; ergo tu es asinus.

23. Iste asinus visus a me non loquitur; ergo tu es asinus.

24. Pono istum casum: tu es asinus est tibi positum, quo admisso
proponitur ista 'tu es asinus' .

25. Tu es aliquis istorum, et qUilibet istorum est asinus; ergo tu es asinus,

demonstrando per ly 'istorum' duos asinos.

26. Tu es hoc, demonstrando asinum; ergo tu es asinus.

27. Tu es hoc; ergo tu es asinus.

28. Tu es asinus vel hoc est falsum, demonstrando totam disjunctivam.

29. Tu es asinus vel tuus asinus currit, sedtuus asinus non currit; ergo tu
es asinus.

30. Si tu es A, tu es asinus, sed tu es A; ergo tu es asinus.

31. Tu es asinus vel tu non es asinus, sed tu non es asinus; ergo tu es

asinus.

32. Tu non differs ab animali quod est asinus, <et> tu es animal quod est
asinus; ergo tu es animal quod est asinus.

33. Tu es asinus vel duo contradictoria sunt simul necessaria, sed nulla
duo contradictoria sunt simul necessaria; ergo tu es asinus.

34. Tu es asinus vel tu respondes ad A aliter quam affirmative; ergo tu es

asinus.

35. Tu potes esse asinum, et non potes esse aliud quam tu es; ergo tu es
asinus.

36. Tu desinis diferre ab asino; ergo tu es asinus.

Sophisms proper to M

M37: Si tu es hoc animal, tu es asinus; igitur tu es asinus.

M38: Tu non differs nisi ab asino; igitur tu es asinus.

M39: Possibile est hominem esse asinum, <et> si nullus homo est animal,
impossibile est te ipse esse asinum; igitur si aliquis homo est animal,
possibile est ipsum esse asinum.

M40: Omnis asinus est hoc; igitur hoc est asinus.

 HEYTESBURY, SOPHISMATA ASINlNA: APPENDIX 143

M41: Omne quod est Socrates differt ab asino, sed solus Socrates est
Socrates; igitur Socrates differt ab asino.

M42: Omne animal est asinus; igitur et cetera.

M43: Pono tibi istam 'tu es asinus vel duo contradictoria sunt simul vera',
sed nulla duo contradictoria sunt vera; igitur et cetera.

M44: Tantum tu es asinus; igitur tu es asinus.

M45: Probatur quod tu es asinus; igitur et cetera.

M46: <Tu es asinus et tu es homo; igitur tu es asinus>.

 A Comparative Analysis of the Treatment of Sophisms in
MSS Digby 2 and Royal 12 of the Magister Abstractionum
by Paul A. Streveler

Of the six known manuscripts of the Abstractiones of Magister

Ricardus Sophista (Oxford, Bodleian Library, Digby 24; Bruges,
Bibliotheque de la Ville, 497; Paris, Bibliotheque Nationale, 14069;
Oxford, Corpus Christi 293B; Oxford, Bodleian Library, Digby 2; and
British Museum, Royal Fxix 12), only Digby 24 and Bruges 497 appear to
be complete, each containing an "explicit" identifying the author as a
certain Ricardus Sophista. Of the four fragmentary manuscripts, Corpus
Christi 293 is the longest with some eight folios (207ra-215vb); Paris with
only seven (26ra-33-ra). By contrast, Digby 24 runs to almost thirty
folios (61ra-90rb) and Bruges to some twenty-one (74ra-95va). MSS
Royal 12 and Digby 2 are even more fragmentary; the former comprising
only four folios (l12va-115rb), the latter less than six (123r-140v). Not
only are these latter two manuscripts very brief, but also the language used
in treating sophisms appears to be quite different, not only when compared
to the more lengthy manuscripts, but also when compared to each other.
In this paper I want to compare the treatment of sophisms in these two
fragmentary manuscripts with a view toward comparing them to the other
four manuscripts, which are presently in a preliminary edition. (I shall
refer to the four edited manuscripts as the Text.) The editors of the
Abstractiones decided not to incorporate Royal 12 and Digby 2 into the
edition primarily because of the apparent serious differences in these two
MSS. They have decided to simply transcribe these two fragmentary MSS
in an appendix to the edition. Part of the purpose of this study is to
ascertain whether Royal 12 and Digby 2 ought to be incorporated into the
edition of all the MSS of the Abstractiones.

Since space does not permit a detailed analysis of every sophism

treated in· these manuscripts, I shall choose, for varied reasons, a small
selection of sophisms which I shall attempt to analyse in some detail. I
have attached to this paper a brief summary of all sophisms treated in the
fragmentary manuscripts (Royal and Digby 2) along with a brief summary
of their treatment in the preliminary edited text. Although these summaries
are often exceedingly brief, I hope they will be of help in seeing some of
the essential similarities and differences among our MSS. Also attached is
a complete list of the sophisms treated or mentioned in our preliminary
edition of the Abstractiones. L. M. de Rijk's original list of sophisms in
Logica Modernorum II, was transcribed from the Digby 24 manuscript)
Our list makes some additions and corrections to this list: as, for example,
number 125 on our list which is transcribed by de Rijk correctly from

lL.M. de Rijk, Logica Modernorum, Assen: Van Gorcum 1967, II 1, pp. 62-71. (It
should be noted that Sten Ebbesen has recently found a brief fragment of the
Abstractiones in the Royal Library, Copenhagen, KB, fragm. 1075. See S. Ebbesen,
"Bits of logic in Bruges, Brussels and Copenhagen manuscripts", Cahiers de l'Instilut
du Moyen-Age Grec et Latin 60, 1990, pp. 129-44, p. 144.)

144
 SOPHISMS OF THE MAGISTER ABSTRACTlONUM 145

Digby 24 as QUANTO MAGIS SCIS TANTO MINUS SCIS.2 The proof

then progresses: "The more you know the more you drink; the more you
drink, the less you know. So the more you know, the less you know."
Reading 'sms' for 'SCIS' as in the Bruges manuscript (f. 82rb) certainly
gives the "correct", albeit less humorous, reading. This is, unfortunately,
one of the very few instances of humor in the text of the Abstractiones!

A comparative reading of the beginning passages of the texts of Digby

2 and Royal 12 reveals immediately what appears to be an interesting point
of difference between these two MSS on the one hand and the four other
MSS of the Abstractiones. While all six MSS begin with a similar
quotation from Aristotle's de 1nterpretatione, 17b 14-16: "Nulla est
affinnatio in qua universale universaliter sumptum praedicatur", Digby 2
and Royal 12 go on to gloss Aristotle's claim as having a "threefold"
possible reference, whereas all other MSS describe the division as
twofold} (It should be noted that Bruges 497 does not label the division
as "twofold", although its analysis of it is twofold.) Consistent with its
general brevity, Royal 12 does not elaborate upon the nature of the
threefold division, whereas Digby 2 continues at some length to indicate
the nature of the threefold reference. A look at the gloss in the five MSS as
well as Aristotle's remarks in de 1nterpretatione 17, would seem to indicate
that the division ought to be understood as twofold rather than threefold.
Aristotle is obviously not opposed to propositions wherein universals are
taken universally, either explicitly when the subject is quantified or
implicitly when it is not actually quantified, but that is the intention.
Aristotle sees a problem with "propositions" wherein both subject and
predicate are universals taken universally and more explicitly, when the
predicate is quantified. ('Every man is every animal' is Aristotle's
example.) Aristotle's sparse remarks here appear to indicate TWO possible
reasons why such propositions should be rejected: (a) either because they
are not "propositions" at all strictly speaking and are "contrary to truth", as
he says, because they are somehow "degenerate" (and are therefore,
strictly neither true nor false); (b) or because such propositions are
"contrary to truth" simply because they are false. The first line of our text
simply asserts "There is no affirmation in which a universal taken
universally is predicated;" it does not assert flatly that there is no TRUE
affinnation in which a universal taken universally is predicated. And in fact
the only gloss that Royal gives of this comment is that generally Aristotle
is to be understood as saying "that there is no TRUE affinnation" of this
sort. There would be little point to this gloss if Aristotle's remarks could
not plausibly be construed as indicating an additional reason why such
affinnations should be rejected, namely, because they are not propositions
at all.

Under the second possibility (Le. that all such propositions are
contrary to truth in the sense that they are false), the Magister
Abstractionum indicates further such instances. Under this second
division, all MSS (except Royal 12, which has no further gloss), exhaust

2No. 119 on de Rijk's list (f. 73va in Digby 24).

3De Rijk uses this more complex gloss in Digby 2 as a reason for thinking it a later
adaptation of the Abstractiones.
146 PAUL A. STREVELER

the possibilities "whether the superior is affinned of the inferior, or the

inferior of the superior; or the convertible of the convertible." It is here that
these manuscripts survey a threefold reference. It is, of course, only this
second rationale for rejecting all such affinnations (Le. because they are
not true in the sense of being false) that the Magister is interested in
pursuing when offering sophisms which appear to be counterexamples to
Aristotle's dictum.

At this point all manuscripts except Royal 12 jump immediately into

consideration of the first sophism, OMNIS HOMO EST OMNIS HOMO.
SAMPLE I

For purposes of continuity, it may be well first to analyse this

beginning sophism OMNIS HOMO EST OMNIS HOMO:

Proof of this sophism is by induction and appeal to the dictum of

Boethius that "no proposition is truer than this one wherein the same is
predicated of the same,"4 and disproof by counter instance (Le 'Socrates is
every man', which is false) are virtually identical in all MSS.
Nevertheless, it should be observed that a consistent feature of Digby 2,
only occasionally seen in the other MSS, is to respond to both proof and
disproof after a general resolution of the sophism is offered. Sometimes,
indeed, no general resolution is given, but only these responses to both
proof and disproof, making it impossible to conclude what Digby 2 might
consider to be the solution to the sophism.

Solutions to this sophism are similar, but not identical: Royal claims
the sophism to be simply false, the induction in the proof containing the
fallacy of the consequent and the dictum of Boethius is to be understood
only regarding sentences wherein subject and predicate "signify the same"
and are "concepts" ("in conceptu", rather than merely "in voce", as Royal
puts it). As Royal notes, such a sentence as 'Chimaera est chimaera' is not
an example of the same predicated of the same, since 'chimaera' is no
concept; so too, 'Currens est currens' is not an example, if there is no one
running, since the tenns have no referent.

Digby 2 does not assert the sophism to be simply false, but claims that
it is ambiguous as to whether the sign 'omnis' is taken dividedly or
compositely. If it is taken compositely, then it is really a singular
proposition and is true, since it has reference to the whole aggregate of
individual men. If it is taken dividedly, then it is a universal and is false,
because it signifies that one man is every man.

Responding then to the disproof, Digby 2 argues that the counter

instance is not a genuine contrary to the sophism considering it as a
universal but only as a singular proposition. Likewise, the counter instance
is not a genuine contradictory to the sophism considering it as singular, but
only considering it as universal.

41n Liberum de Int., ed. Ja; Patr%gia Latina, L. 1. 64 c. 387CD.

 SOPHISMS OF THE MAGISTER ABSTRACTIONUM 147

Responding to the proof, like Royal, Digby 2 accuses the argument of

fallacy of the consequent. Regarding the dictum of Boethius, Digby 2
concedes it, but argues that here it does not apply, since if we understand
the sentence in the divided sense then 'omnis homo' isn't really predicated,
but only 'homo', and these are not identical. An objection is offered to this
last point: "'Every man is moved' is proved: John writes and Robert
walks, Peter reads; so every man is moved." These are singulars of a
universal, yet the predicates do not all agree. Digby 2 responds that there
is agreement in predicates here, since it is understood that each of these are
examples of motion (i.e. the inference holds: 'Ioannes scribit, ergo
movetur' for each example of movement). But in the induction of the proof
when it is said 'Iste homo est iste homo' etc., we are not to understand
these as singulars of the same universal.

In the Text, this sophism is treated along with the solution to OMNIS
HOMO EST TOTUM IN QUANTITATE. The sophism is said to be
ambiguous because of equivocation since 'omnis' can mean 'all the parts
taken individually' or 'the whole taken together'. In the latter sense
(similar to the composite sense), it is true, but is singular; in the former
sense, it is false and is universal. The Text makes the same point as Digby
regarding the appropriateness of Boethius' dictum to this sophism.
Similarly, the Text accuses the argument of the proof of fallacy of the
consequent.

At this point, the Text offers a counter example to the dictum of

Boethius which has its own casus and appears to be offered as a separate
sophism and will be treated later in Digby 2 as a separate sophism: OMNIS
HOMO VIDET ALIQUEM HOMINEM, under the casus that each man
sees only himself. In the Text this sentence is treated as an example of how
active and passive constructions of the same terms do not always render
the same proposition. It is not exactly clear how this figures as a counter
example to Boethius' dictum and it would appear that Digby 2 is correct to
treat this as a separate sophism. In any event, the sentence is given only
the briefest discussion in the Text. None of this discussion occurs in
Roya/.
SAMPLE II

OMNE COLORATUM EST

I choose this sophism because of the reference to it in the Summa

Logicae of William Ockham,5 and because of its unusally lengthy
treatment in Digby 2 as well as its extremely brief treatment in Royal.

The proof of the sophism is virtually identical in all manuscripts:

White exists, black exists, in between exists; so every colored thing exists.

Royal, Digby 2 and the Text offer the disproof: Every colored thing
exists, every white thing is colored; so every white thing exists. To this,

5Summa Logicae, ed., P. Boehner, G. Gal and S. Brown, St. Bonaventure, N.Y.:
Franciscan Institute Publications 1974, p. 367.
148 PAUL A. STREVELER

Digby 2 and the Text add a second disproof: 'Omnis' requires at least three
appel/ata; there aren't three whites, blacks etc. so .... Thus, the
contradictory to the sophism is true, i.e. 'Not every colored thing exists'.

Royal declares the sophism to be TRUE because it converts with the

truth, 'Every colored thing is a being'. (This is the idea of the
convertibility of 'ens' and 'esse'). But there appears to be some corruption
in Royal here, since its only further remark is to concede the consequence
of the disproof along with the antecedent and consequent, which seems
peculiar since it asserts the sophism as TRUE.

The treatment in Digby 2 continues at some length. First it is noted

that the sign 'omnis' can here distribute the term 'coloratum' on the level
of species or individuals. If the former, the sentence is true and means
simply that every individual species of color exists; if on the level of
individuals, it is false, since it is false that every individual colored thing
exists. This same point is made in the Text in the language of remote and
proximate parts.

Regarding the disproof, Digby 2 accuses this of the fallacy of figure

of speech because in the first premiss (Omne coloratum est) 'omne' is
taken as having reference to species (i.e. referring "quale quid"); whereas
in the second (omne album est coloratum) 'omne' refers individually ("hoc
aliquid") Digby 2 appears obviously corrupt here, for it now concludes:
"I say that the sophism is simply FALSE according to the rule:
Praedicatum non contingit omnibus sub subiecto vere dici contentis."
Obviously, if this is understood as a critique of the argument of the
disproof, it would not follow that the sophism is false unless the proof
also made the same error.

Additionally, Digby 2 accuses the disproof of the fallacy of

equivocation because the term 'est' in the first premiss stands for the 'is'
of existence, but in the second it stands for the habitudinal 'is'. This same
point is made in the Text.

Regarding the disproof that appeals to the rule of three for 'omnis'
(thUS denying that 'omne album est' can be true because there aren't three
whites etc., and so concluding that the opposite is true, i.e. 'non omne
album est', and finally concluding that 'non omne coloratum est' is true),
Digby 2 concedes all of the argument up to the last consequence, and
argues that this last consequence commits the fallacy of the consequent
from destruction of the antecedent, since one cannot infer that there is
something which is not white from the statement that not every white
exists any more than one can infer that there is something which is not
colored from the statement that not every colored thing exists. So it does
not follow: 'Something exists which is not white, therefore something
exists which is not colored'. The Text makes a similar point.

It ought to be noted that the Magister Abstractionum appears

ambivalent regarding the so-called rule of three. Here, as elsewhere, he
does not deny the rule, but points out that the mistake in the argument here
is largely irrelevant to the rule itself, even though to make the argument at
 SOPHISMS OF THE MAGISTER ABSTRACTJONUM 149

all one must assume the rule as operative, as, he says, "some people
maintain."

Ockham argues that the Magister errs here in assigning the fallacy of
accident to the argument of the disproof, but it should be noted that
Ockham appears mistaken on several grounds here. Most obvious, of
course, is that the Magister does not assign the fallacy of accident, but the
fallacy of the consequent here and, perhaps more importantly, it is not the
general argument of the disproof which he accuses of fallacy of the
consequent (that argument he rejects because of fallacy of figure of speech
in Digby 2 or equivocation in the Text; in Royal, as we noted above, he
actually ACCEPTS this argument, as does Ockham); rather, it is the
argument of the disproof wherein the rule of three is assumed which the
Magister (in the Text and in Digby 2) rejects as committing the fallacy of
the consequent (because it is only on the assumption of the rule of three
that the conclusion of the argument 'Omne album est' would be
understood as obviously false). Of course, his objection is firstly to the
argument that follows this, viz., if 'omne album est' is not true, then 'non
omne album est' is true, and so from this it follows that 'non omne
coloratum est' is true. It is this latter argument which contains the fallacy
of the consequent by destruction of the antecedent. Secondly, of course,
the original disproof also commits fallacy of the consequent from positing
the consequent. It is unclear to me how to make Ockham' s remarks
consistent with the text of the Magister Abstractionum. It seems that he
was relying upon a faulty recollection, and did not have the text before
him.

SAMPLE III

OMNIS PHOENIX EST

Although this sophism is given only the briefest treatment in all of our
MSS, it contains reference to a doctrine which Bacon identifies as favored
by Richard Rufus, thus being a ground for our attribution of the
authorship of the Abstractiones to Richard Rufus.

Royal prefaces this sophism with a casus not noted in any other MSS,
which may be grounds for thinking it a quite different sophism in Royal,
although it appears clear that the assumption of the casus is operative in
Digby 2 and the Text as well. The casus is: Sometimes phoenix has being
and sometimes it does not and when it does then there is only one phoenix.
(In any event, this is a strange casus, for it may lead one to think that when
phoenix does not have being, then there can be more than one! The idea
here is, of course, that at any given time there is only one phoenix, since
the new one arises out of the ashes of the old one.) Proofs in all three
sources are virtually identical, viz., this is false 'Some phoenix does not
exist', so its contradictory (i.e. the sophism) is true.

Disproofs all appeal to the rule of three: 'Omnis' requires at least three
appel/ata; there aren't at least three phoenixes, so the sophism is false.

Royal responds that the sophism is doubtable as well as its

contradictory. Regarding the disproof, Royal notes that the rule of three
150 PAUL A. STREVELER

does not always hold good, because 'Omnis sol est omnis sol' is true and
there aren't three suns. (It will be recalled that Royal noted this example in
the opening passages of our treatise as an example of a true universal
wherein the same is predicated of the same. There, Royal did not flatly
deny the truth of the sentence, nor really that it was genuinely universal in
character, but only that 'omnis sol' wouldn't do as a counter example,
since we were here dealing with universals with more than one
suppositum.) Here Royal notes that "To the commoner (vulgo) it appears
that such a proposition might be a false hypothetical whose subject has
only one supposit according to <this> exposition of the term 'omnis'; but
in another idiom <of speech> it must be maintained that such propositions
are true." There are several reasons to believe, therefore, at least on the
basis of the Royal MS, that the Magister Abstractionum is at least
ambivalent, as has already been noted, about the alleged "rule of three" for
'omnis'.

Digby 2 declares the sophism to be SIMPLY FALSE by appeal to two

related rules: "Praedicatum non contingit vere dici omnibus contentis sub
subiecto" and "Terminus communis inquantum est ex se est indifferens ad
omnis eius supposita praesentia et futura signum sibi adveniens distribuit
illum pro omnibus illis suppositis quibus non convenit praedicatum."
Digby 2 does not further explain how these rules apply to the sophism
proof. To the proof itself, Digby 2 declares that the contradictory of the
sophism, as noted in the disproof, 'Some phoenix does not exist' is true,
because its exponent is true, i.e., 'Something not a being does not exist' .

The Text likewise declares the sophism false for essentially the same
reason, although in different language, viz., because it affirms being of
that which isn't in act since it says that this is false, 'Some phoenix does
not exist', when in fact it is proper to negate being of that which does not
exist, and this term 'phoenix' supposits equally for non being and for
being.

This is the doctrine so vehemently opposed by Bacon and atrtributed

to Richard Rufus in Bacon's Compendium Studii Theologiae, Chapter
four, paragraph 86. It should be noted, however, that Bacon attributes to
Rufus a quite general thesis about signification of names, viz., that such
signification can remain in the absence of any referent of the name;
whereas the point at issue in this sophism is the signification of 'phoenix' ,
which is a name of a quite special sort, i.e. a name that never had (or can
have?) real reference. I am reminded here of a remark of Peter of Spain's
in his Tractatus Syncategoremata: "But with regard to that which is
excluded, the following rule is given: An exclusive word added to any
term excludes everything which is different from it as far as natural
supposition is concerned. For anything is different from another in two
ways: in one way according to the essence and natural supposition at the
same time, for example, 'man' and 'horse'; in another way according to
natural supposition only, as 'man' and 'chimaera'. For although
'chimaera' may not be different from 'man' from the point of view of
essence, nevertheless it is different from it from the point of view of
natural supposition. For not only beings are denoted by the term but also
 SOPHISMS OF THE MAGISTER ABSTRACTIONUM 151

non-beings."6 This appears to be the Magister's point as well with respect

to the term ·phoenix'. Thus, I do not believe one can attribute the
Abstractiones to Rufus merely on the basis of this remark in this sophism,
a remark not made in Royal or Digby 2, nor is it clear to me than one can
glean from the Abstractiones a general doctrine about the signification of
terms that have lost or changed existential reference.

The next two sophisms present a related problem. For example, in

QUICQUID EST VEL NON EST EST, the disproof goes as follows: If
whatever exists either does not exist or exists, then Caesar exists or does
not exist, so Caesar exists. The conclusion is obviously false, since Caesar
is no longer existing. The Magister's response to the sophism in general
is, of course, to distinguish the composite and divided senses here; but in
his response to the disproof, he notes that the mistake here is really figure
of speech, because we mistakenly think that the supposit for 'Caesar' is
the same throughout, but it really is unclear whether the statement be taken
in the composite or divided sense. Thus the Magister appears to be
admitting that 'Caesar' does not have the same supposition in •Caesar est'
and 'Caesar non est' and since to think otherwise is to commit the fallacy
of figure of speech, it would appear to follow that the Magister could also
grant that 'Caesar' has different significations here as well, if proper
names can be said to have significations at all.
Sample IV

OMNlS HOMO EST ET ALIUS HOMO EST

This sophism has very lengthy treatment in Digby 2 and the Text, but
a relatively brief treatment in Royal. All three sources give essentially the
same inductive proof: This man exists and another man exists and that man
etc., so every man exists and another man exists.

Disproofs in all MSS are similar: The second part of the conjunction
(Digby 2 consistently refers to the sentence as a disjunction) is false
because it appears to imply that there is some man who is other than every
man, i.e., lacks human nature, which is impossible; so the entire
conjunction is false.

Royal responds to the sophism that it is TRUE and denies the claim of
the disproof that the second part of the sentence is false because if it were
not the case that some man were other than every man, it would be true
that some man was the same as every man, which is obviously false. The
Text makes a similar point.

6Tractatus Syncategoremata, in Tractatus Syncategorematum and Selected Anonymous

Treatises, tr. J.P. Mullally and R. Houde, Milwaukee: Marquette University Press
1964, p. 33. I am assuming here, of course, that these early parts of Mullally's
translation (a 15th century adapatation of Peter's treatise) are consistent with the actual
text of the Tractatus Syncategoremata. See also H. A. O. Braakhuis, "English Tracts
on Syncategorematic Terms from Robert Bacon to Walter Burley", in English Logic
and Semantics from the End of the Twelfth Century to the Time of Ockham and
Burley, Proceedings of the Fourth European Symposium on Medieval Logic and
Semantics, Artistarium Supplementa I, ed. J. Pinborg, Nijmegen: Ingenium 1981, pp.
145-9.
152 PAUL A. STREVELER

In its response to this sophism, Digby 2 presents one response that

relies on thinking of the sophism as a disjunction and another reponse
thinking of the sentence as a conjunction. As a disjunction, one must prove
that both parts of the sentence are false. 'Omnis homo est' is false because
it denotes that man, present, past and future exists, which is false. 'Alius
homo est' is false, as noted above, because it implies that some man lacks
human nature. (This latter argument is strange, since obviously 'Alius
homo est' does not in itself imply that there is a man that lacks human
nature; this is implied only when the sentence is coupled with the first part
of the sophism 'omnis homo est'.)

Regarding the proof, Digby 2 asks whether we are to understand the

term 'alius' throughout as standing for same or diverse supposits. If for
the same, then the reference is only to one singular and either the sophism
is false or there is fallacy of the consequent because of insufficient
induction. Iffor diverse supposits, then either there will only be one true
singular or there will be a singular induction, which induction is
malformed or ought to be censured (quaedam inductio est turperanda vel
vituperanda).

Treating the sophism as a conjunction, Digby 2 presents a second

response, viz., that the sophism is simply TRUE and each part is true.
That 'Omnis homo est' is true is proved because the common term
supposits only for presently existing men, so 'omnis homo praesens est' is
equivalent to 'omnis homo est'. That 'alius homo est' is true is proved:
This man exists, and every man exists; but this man isn't every man; so he
is other than every man. If this were not true, this man would be the same
as every man, which is impossible.

Digby 2 and the Text offer further arguments in support of the

sophism according to the rule: "From a negative, a privative does not
follow" just as "from genus, species does not follow" unless'one adds the
difference. So too, one must add to the negative the positing of the
existence of the extremes of the proposition in order to get a valid
inference. Here the extremes are: 'omnis homo' and 'iste homo'. Thus,
this is a valid consequence: 'This man is not every man, therefore this man
is other than every man' .

Along these same lines, the Text adds a rule and an additional
argument in favor of the sophism: "Diversity with respect to the posterior
follows from diversity with respect to the prior." 'Iste homo' is prior;
'omnis homo' is posterior. It follows: If this man is other than this man,
then <this man> is other than every man. Since only the true can follow
from the true, this is true: 'This man is other than every man'. The
Magister concedes this point, but adds that it still appears inconvenient to
say 'Every man exists and another man exists' because it appears that a
part is copulated of the whole. But, as the Magister notes in the following
similar sophism, a part being copulated of the whole isn't the cause of the
falsity of the sophism, even if it may cause some impropriety.

Space does not permit further analyses of sophisms in this very

interesting text of the Abstractiones. We have only examined briefly a few
 SOPHISMS OF THE MAGISTER ABSTRACTIONUM 153

of over three hundred sophisms, some treated in great detail and

complexity. A perusal of the appended summary of all sophisms common
to Digby 2, Royal 12 and the Text will reveal fundamental similarities and
differences. Differences are more a matter of vocabulary in the actual
treatment of sophisms, rather than in doctrines appealed to in their
resolution, although sometimes the latter reveal differences as well.
Vocabulary can at times be so different as to suggest that Royal and Digby
2 are different versions of at least the beginning sophisms contained in the
larger and more complete manuscripts, suggesting alternative ways of
presenting and, sometimes, resolving these sophisms. This appears to be
the best judgment that can be made at this time: Royal and Digby 2 do not
represent strict copies, albeit fragmentary, of the same text, but represent
different versions of the sophisms themselves, even if not every sophism
treated is given a novel treatq1ent in these fragmentary manuscripts.

In conclusion, therefOr~, it would seem most appropriate to present

the transcription of the Royal and Digby 2 manuscripts either in an
appendix to the edition or, perhaps better, in columns parallel to the other
edited manuscripts so that the different versions can be clearly studied. It
does not appear necessary nor possible to edit these fragmentary versions
of the Abstractiones into the larger Text, since this latter task would not
only overload the critical apparatus, but also cause needless confusion in
the understanding of the actual treatment of sophisms in Royal and Digby
2.
West Chester University
154 PAUL A. STREVELER

Appendix 1

A list of Sophisms treated in the Abstractiones of the Magister

Abstractionum (from a preliminary text edition):7

OMNIS

1 OMNIS HOMO EST OMNIS HOMO (1)

2 OMNIS HOMO EST TOTUM IN QUANTITATE (1)

3 OMNIS HOMO EST UNUS SOLUS HOMO (2)

4 OMNIS HOMO EST ALIQUIS HOMO* (2)

5 OMNIS HOMO EST HOC ALIQUID* (2)

6 OMNIS HOMO EST SINGULARE* (2)

7 OMNIS HOMO EST INDIVIDUUM* (2)

8 OMNES APOSTOLI SUNT XII (3)

9 OMNE ANIMAL EST SANUM (3)

10 OMNE ANIMAL FUIT IN ARCA NOAE (4)

11 OMNECOLORATUMEST(4)

12 OMNIS PHOENIX EST (5)

13 OMNE BONUM VEL NON BONUM EST ELIGENDUM* (6)

14 OMNE ANIMAL VEL NON ANIMAL EST SANUM VEL

AEGRUM* (6)

15 OMNIS HOMO VEL ASINUS EST RISIBILIS* (6)

16 QUICQUID EST VEL NON EST EST *(6)

17 OMNE RATIONALE VEL IRRATIONALE EST SANUM (6)

18 OMNIS PROPosmo VEL EIUS CONTRADICTORIA EST VERA

(7)

19 TU ES QUILIBET VEL DIFFERS A QUOLIBET (8)

20 TU ES QUILIBET VEL A QUOLIBET DIFFERENS (8)

21 TU SCIS QUIDLIBET VEL QUIDLIBET IGNORAS (9)

7 * Indicates sophism sentences considered together.

 MAGISTER ABSTRACTIONUM: APPENDIX 1 155

22 TU SCIS QUIDLIBET VEL NIHIL (9)

23 TU SCIS QUlCQUlD SCIS (9)

24 A DE NUMERO ISTORUM QUORUM QUODLIBET DIFFERT AB

EO QUOD EST IPSUM ESSE NON EST ILLUD (to)

25 OMNIS HOMO EST ET ALIUS HOMQ EST (11)

26 OMNIS HOMO ET DUO HOMINES SUNT TRES HOMINES (12)

27 OMNE VERUM ET DEUM ESSE DIFFERUNT *(12)

28 OMNE VERUM ET DEUM ESSE SUNT DUO VERA* (12)

29 OMNE VERUM ET DEUM ESSE PONUNT IN NUMERO* (12)

30 OMNIS HOMO EST ET QUILIBET VIDENS ILLUM EST ASINUS

(12)

31 OMNIS GRAMMATICUS EST ET QUILIBET SCIENS IPSUM

ESSEGRAMMATICUMESTTANTUMTALIS (13)

32 OMNIS HOMO EST ET QUILIBET DIFFERENS AB ILLO EST

NON HOMO (13)

33 OMNIS HOMO QUI EST ALBUS CURRIT (14)

34 OMNE QUOD EST VERUM SCIRI A TE EST VERUM (14)

35 OMNE QUOD EST VERUM SCIRI A TE EST FALSUM (15)

36 OMNE NESCITUM ATE SCIRI A TE EST FALSUM (15)

37 DEUS ERIT IN QUOLIBET INSTANTI NON EXISTENS (16)

38 QUODLIBET ALIQUORUM ANlMALIUM EST NON HOMO

QUORUM QUILIBET EST HOMO (17)

39 OMNE ALIUD QUAM ANIMAL QUOD ET SORTES SUNT DUO

DIFFERT A SORTE (17)

40 TU ES ALIUD QUAM ANIMAL QUOD EST ROMAE (17)

41 OMNE ALIUD QUAM ANIMAL QUOD EST HICINTUS EST

LAPIS (18)

42 QUILIBET QUALELIBET DE SE TALI SCIT SE IPSUM ESSE

TALE QUALE IPSUM EST (18)

43 SORTES DICIT OMNE ENUNTIABILE IN A ET IN B (19)

44 SORTES VIDET OMNEM HOMINEM IN A ET IN B (19)

45 OMNIS HOMO EST ANIMAL ET OMNE ANIMAL EST HOMO (19)

156 PAUL A. STREVELER

46 OMNIS HOMO EST ANIMAL ET ECONVERSO (20)

47 OMNEM HOMINEM VIDENS EST UNUM SOLUM HOMINEM

VIDENS (20)

48 CUIUSLIBET HOMINIS ASINUS CURRlT (21)

49 CUIUSLIBET HOMlNIS OCULUS EST DEXTER (21)

50 OMNIS HOMO MORITUR QUANDO UNUS SOLUS HOMO

MORITUR* (21)

51 OMNIS HOMO MORITUR QUANDO UNUS SOLUS HOMO

MORIETUR* (22)

52 OMNIS HOMO MORIETUR QUANDO UNUS SOLUS HOMO

MORITUR* (22)

53 OMNIS HOMO MORlETUR QUANDO UNUS SOLUS HOMO

MORIETUR* (22)

54 OMNEM HOMINEM MORl EST IMPOSSmlLE (22)

55 QUICQUID AUDITUR A PLATONE PROFERTUR A SORTE (23)

56 ALBUM FUIT DISPUTATURUM (23)

57 DEUS SCIT QUICQUID SOVIT (24)

UTERQUE

58 UTERQUE 1STORUM EST UTERQUE 1STORUM (25)

59 UTERQUE 1STORUM EST TANTUM ALTER 1STORUM (25)

60 1ST! FERUNT LAPIDEM (26)

61 1ST! SCIUNT VII ARTES (26)

62 1ST! PUGNANT UT VINCANT SE (27)

63 HELENA PEPERIT X FlL10S (28)

64 HELENA PEPERIT DECIES X FlLIOS (28)

65 AB UTROQUE ISTORUM ENUNTIATUM EST VERUM (29)

66 UTERQUE 1STORUM MORIETUR (29)

67 ADAM ET NOE FUERUNT (29)

68 UTROQUE 1STORUM CURRENTE NON CURRIT UTERQUE

ISTORUM (30)

69 NEUTRO 1STORUM CURRENTE ALTER 1STORUM CURRIT (30)

 MAGISTER ABSTRACTJONUM: APPENDIX 1 157

70 NEUTRUM OCULUM HABENDO TU POTES VlDERE (31)

71 UTERQUEISTORUM EST HOMO VEL ASINUS (31)

72 UTERQUE ISTORUM VEL RELIQUUS ISTORUM QUORUM

NEUTER DIFFERT AB HOMINE EST ASINUS (32)

73 ALTER ISTORUM EST HOMO VEL RELIQUUS QUI EST HOMO

EST ASINUS (32)

74 UTERQUE ISTORUM EST HICINTUS ET ALIUS EST HICINTUS

(32)

75 UTRUMQUE ISTORUM ET DUO SUNT TRIA (33)

76 BIS DUO SUNT TRIA ET NON PLURA (33)

77 DUO PATRES ET DUO FILII SUNT TRIA ET NON PLURA (33)

78 PLURA PLURIBUS VI SUNT PAUCIORA PAUCIORIBUS VI (34)

TOTUS
79 TOTUS SORTES EST MINOR SORTE (34)

80 TOTUS SORTES EST HOMO VEL NON EST HOMO (34)

81 SORTES ET PLATO SUNT HOMINES (35)

82 SORTES NASCITUR ANTE B (35)

83 ANIMAL EST PARS ANIMALIS (35)

84 QUAELffiET PARS DISIUNCTIVAE EST VERA CUIUS ALTERA

PARS EST VERA (36)

SI
85 SI ALIQUID EST VERUM EST VERUM IN HOC INSTANTI (37)

86 SI ALIQUID EST VERUM IPSUM ESSE EST NECESSARIUM (37)

87 SI ALIQUID EST IPSUM EST DEUS (38)

88 SI TU ES HOMO ET ASINUS TU ES LEO ET CAPRA (39)

89 FALSUM EST VERUM SI ANTICHRISTUS EST (39)

90 OMNE FALSUM DIFFERT AB A SI ANTICHRISTUS EST (39)

91 UTRUMQUE ISTORUM EST VERUM SI ALTERUM ISTORUM

EST FALSUM (40)

92 SI ALIQUIS DICIT TE ESSE ANIMAL DICIT VERUM (40)

158 PAUL A. STREVELER

93 SI DICO TE ESSE ASINUM DICO VERUM (41)

94 SI QUILIBET EST NON HOMO, HOMO EST NON HOMO (41)

95 SI SORTES DE NECESSITATE EST MORTALIS SORTES DE

NECESSITATE NON EST MORTAUS (42)

96 SI TU SCIS TE ESSE LAPIDEMTU NON SCIS TE ESSE

LAPIDEM (42)

97 SI TU ES UBIQUE TU NON ES UBIQUE (43)

98 SI OMNIS PROPOSmO EST VERA NON OMNIS PROPOSmO

EST VERA (44)

99 SI NULLA PROPOSmO EST VERA ALIQUA PROPOSmO EST

VERA (44)

100 SI NULLUM TEMPUS EST ALIQUOD TEMPUS EST (44)

101 SI NULLUS HOMO EST HICINTUS ALIQUIS HOMO EST

HICINTUS (45)

102 SI NIHIL EST ALIQUID EST (45)

103 ANIMAL CURRERE SI HOMO CURRIT EST NECESSARIUM

(47)

104 IMPOSSIBILE EST TE SEDERE SI TU lACES (47)

105 POSSIBILE EST TE TACERE SI TU LOQUERIS (47)

106 DEUM ESSE SI A NON ERIT VERUM IN A (48)

107 A ESSE SI A NON ERIT ERIT FALSUM IN A (48)

108 SI DE EO QUOD EST SORTES VERUM EST IPSUM ESSE VEL

NON ESSE SORTES EST (49)

109 SI AD HOMINEM ESSE QUOD EST VERUM SEQUITUR

ANIMAL ESSE QUOD EST VERUM HOMO EST (49)

110 SORTES DICIT ID QUOD EST VERUM SI SOLUS PLATO

LOQUITUR (50)

111 TU ES BONUS ET TU ES MALUS SI TU ES MALUS (50)

112 SI VERUM EST TE CURRERE ET TE NON CURRERE TU ES

CAPRA (51)

113 SI SORTES EST SI PLATO EST CICERO EST (51)

114 QUICQUID CONTINGAT SI TU ES ASINUS TU ES CAPRA (51)

115 AD QUODLIBET ENUNTIABILE SEQUITUR IPSUM ESSE

VERUM(52)
 MAGISTER ABSTRACTIONUM: APPENDIX 1 159

116 ALIQUID EST IMPOSSmlLE ET IDEM EST NECESSARIO

VERUM(52)

INQUANTUM

117 ALIQUA INQUANTUM SUNT AEQUIVOCA SUNT UNIVOCA

(53)

118 ALIQUA INQUANTUM CONVENIUNT DIFFERUNT (54)

119 ALIQUA INQUANTUM DIFFERUNT CONVENIUNT (54)

120 ALIQUA INQUANTUM SUNT DISSIMILIA SUNT SIMILIA (55)

121 ALIQUA INQUANTUM SUNT SIMILIA SUNT DISSIMILIA (55)

122 HOMO INQUANTUM EST ANIMAL DIFFERT AB ASINO (55)

123 LOGICUS INQUANTUM CLERlCUS DIFFERT A PHYSICO (56)

124 QUANTO ALIQUID MAlUS EST TANTO MINUS VIDETUR (56)

125 QUANTO MAGIS sms TANTO MINUS sms (57)

126 QUANTO MAGIS ES FOEDUS TANTO MINUS ES FOEDUS (57)

127 QUANTO PLUS ADDISCIS TANTO MINUS SCIS (57)

SIVE

128 SIVE HOMO QUI EST ALBUS EST PLATO SIVE TU ES ASINUS
TU ES CAPRA (57)

QUALECUMQUE

129 QUALECUMQUE EST ALIQUID SI IPSUM EST ALBUM TALE

EST ALIQUID SI IPSUM EST NIGRUM (58)

130 UBICUMQUE EXISTENS EST ALIQUID SI IPSUM EST ROMAE

mIDEM EXISTENS EST ALIQUID SI IPSUM EST PARISIUS
(59)

131 QUODCUMQUE EST ALIQUID SI EST HOMO IDEM EST

ALIQUID SI EST ALBUS* (59)

132 QUOTCUMQUE SUNT ALIQUOT SI IPSA SUNT VII TOT SUNT

EADEM SI IPSA SUNT DUO* (59)

133 QUANTUMCUMQUE EST ALIQUID SI IPSUM EST

BICUBITUM TANTUM EST ALIQUID SI IPSUM EST
MONACUBITUM* (59)
160 PAUL A. STREVELER

QUOTIENSCUMQUE

134 QUOTIENSCUMQUE FUISTI SEDENS TOTIENS FUISTI HOMO

(60)

NISI
135 NULLUS HOMO LEGIT PARISIUS NISI IPSE SIT ASINUS (60)

136 NIHIL EST VERUM NISI IPSUM SIT FALSUM (60)

137 SORTES DECIPITUR NISI DECIPIATUR (61)

138 SORTES DICIT FALSUM NISI DICAT FALSUM (61)

139 NIHIL EST VERUM NISI IN HOC INSTANT! (61)

140 SI ALIQUID EST VERUM EST VERUM IN HOC INSTANT! (62)

VEL
141 NECESSARIUM EST TE SEDERE VEL NON SEDERE (62)

142 OMNE ANIMAL EST RATIONALE VEL IRRATIONALE (63)

143 SORTES DICIT C (63)

AN

144 TU SCIS AN DE MENTIENTE SIT FALSUM SORTEM ESSE

IPSUM (65)

145 SORTES SCIT AN PLATO SCIAT AN SORTES SCIAT AN

PLATO SCIAT ALIQUID DE EO (66)

146 TU SCIS AN OMNE ANIMAL SIT RATIONALE AN ILLUD SIT

IRRATIONALE * (66)

147 TU SCIS AN OMNIS HOMO SIT SORTES AN DIFFERAT A

SORTE* (67)

148 TU SCIS OMNE ANIMAL ESSE RATIONALE VEL

IRRATIONALE (67)

149 TU SCIS ALIQUID ESSE QUOD NON SCIS ESSE (67)

NE

150 DEUS VULT NE FACIAS MALUM ET DEUS PROHffiET NE

FACIAS MAUM, ERGO IDEM VULT ET PROHIBET (68)
 MAGISTER ABSTRACTIONUM: APPENDIX 1 161

151 TU VIS NE TIBI CONCLUDATUR, ET CAYES NE TIBI

CONCLUDATUR, ERGO IDEM VIS ET CAYES (68)

152 NULLUS HOMO NULLUM ANIMAL EST (69)

153 NULLUS HOMO NULLUM ANIMAL EST, ERGO NULLUM

ANIMAL NULLUS HOMO EST (70)

154 NULLUS HOMO NULLUS HOMO EST (70)

155 NIHIL NIHIL EST* (70)

156 NIHIL NULLA RES EST* (70)

157 DE NIHILO NIHIL EST YERUM (70)

158 NULLUS HOMO EST OMNIS HOMO 71)

159 NON ALIQUID EST ET TU ES ASINUS (71)

160 NIHIL ET CHIMAERA SUNT FRATRES (71)

161 ALIQUID NEC EST NEC ERIT ET EST ET ERIT (72)

162 NEC HOMO VIDENS SUUM ASINUM NEC SUUS ASINUS EST
CAPRA (72)

163 TU SCIS QUOD NIHIL SCIS QUOD SI SCIS NIHILSCIS (72)

164 SI TU SCIS QUOD NIHIL SCIS NIHIL SCIS (73)

165 SUNT DUO QUAE DUO SUNT ET SUNT DUO QUAE DUO NON
SUNT (73)

166 ALIQUA CAUSA NON ES HOMO (74)

167 NULLO CURRENTE CRESCUNT TIBI CORNUA FRONTE (74)

168 NULLUM HOMINEM SEDERE EST NECESSARIUM (75)

169 AD NULLUM HOMINEM ESSE SEQUITUR OMNEM HOMINEM

ESSE (75)

170 AD OMNEM HOMINEM ESSE NON SEQUITUR OMNEM

HOMINEM ESSE (75)

171 AD ALIQUEM HOMINEM ESSE NON SEQUITUR ALIQUEM

HOMINEM ESSE (76)

172 NULLUM HOMINEM ESSE SEQUITUR AD ALIQUEM

HOMINEM ESSE (76)

173 NULLUS HOMO EST SI ALIQUIS HOMO EST (77)

174 NULLUS HOMO POTEST SCIRE QUOD ALIQUIS HOMO

POTEST SCIRE (77)
162 PAULA. STREVELER

175 NULLUM CAPUT HABENS EST ALIQUOD CAPUT HABENS

(78)

176 NIHIL VIDENS EST ALIQUOD VIDENS *(78)

177 A NULLO ENUNTIATUM A NULLO VERE DICITUR (78)

178 TU NON POTES VERE NEGARE TE NON ESSE ASINUM (79)

179 NEGATUM ESSE LIGNUM EST VERUM (80)

180 TU ES ASINUS (81)

QUAM

181 SORTES EST ALIUD ANIMAL QUAM BURNELLUS (82)

182 SORTES DESINIT SCIRE PLURA QUAM DESINIT SCIRE (82)

183 PLURA SCIUNTUR A SORTE QUAM SCIUNTUR A SOLO

SORTE(83)

184 PLURA SUNT VERA DE SIBI TOTIDEM QUAM SUNT VERA DE

PAUCIORIBUS SE (84)

185 PLURA SUNT PARIA QUAM INPARIA (85)

186 INFINITA SUNT FINITA (85)

INCIPITIDESINIT

187 QUOD INCIPIT ESSE DESINIT NON ESSE (86)

188 SORTES DESINIT ESSE ALTER ISTORUM (87)

189 SORTES INCIPIT ESSE ALTER ISTORUM (87)

190 SORTES DESINIT ESSE SIMILIS PLATONI (87)

191 SORTES INCIPIT ESSE SIMILIS PLATONI (88)

192 SORTES DESINIT ESSE SIMILIS PLATONI (considered twice)

(88)

193 SORTES VULT ESSE TALIS QUALlS EST PLATO (88)

194 SORTES ET PLATO DESINUNT ESSE TALES QUALES IPSI

SUNT (89)

195 SORTES DESINIT VIDERE OMNEM HOMINEM (90)

196 SORTES INCIPIT VIDERE OMNEM HOMINEM (90)

197 SORTES DESINIT SCIRE QUICQUID SCIT (91)

 MAGISTER ABSTRACTIONUM: APPENDIX 1 163

198 SORTES INCIPIT SORE QUICQUID SCIT (91)

199 SORTES DESINIT SCIRE PLURA QUAM PLATO (91)

200 PLATO DESINIT SORE PLURA QUAM SORTES (91)

201 PLATO DESINIT SORE PLURA QUAM SORTES (considered

twice) (92)

202 SORTES DESINIT SCIRE SE NIHIL DESINERE SCIRE (92)

203 OMNIS HOMO INCIPIT ESSE (93)

204 OMNIS HOMO DESINIT ESSE (93)

205 DEUS DESINIT NUNC ESSE* (94)

206 DEUS INCIPIT NUNC ESSE* (94)

207 SORTES INCIPIT ESSE NON INCIPIENDO ESSE* (94)

208 SORTES DESINIT ESSE NON DESINENDO ESSE (94)

209 SORTES INCIPIT ESSE SI EST ET NON FUIT/DESINIT ESSE SI

EST ET NON ERIT (95)

210 SORTES DESINIT ESSE ALBISSlMUS HOMINUM (95)

211 SORTES DESINIT VIDERE OMNEM HOMINEM PRAETER

PLATONEM (96)

212 TU NON CESSAS COMEDERE FERRUM (97)

PRAETER
213 X PRAETER V SUNT V (98)

214 X ANIMALIA PRAETER DUO SCIUNT SE ESSE ALBA (98)

215 OMNIA X PRAETER UNUM SUNT IX (99)

216 OMNIUM DUORUM FRATRUM UTERQUE PRAETER UNUM

EST ALBUS (100)

217 OMNIS NUMERUS PRAETER BINARIUM EXCEDIT

UNITATEM NUMERO (100)

218 QUOTLffiET PRAETER DUO ET TRIA SUNT PLURA DUOBUS

NUMERO (101)

219 OMNlS HOMO VIDET OMNEM HOMINEM PRAETER SORTEM

(101)

220 NULLUS HOMO VIDET ASINUM PRAETER BRUNELLUM

(102)
164 PAUL A. STREVELER

221 SORTES BIS VIDEBIT OMNEM HOMINEM PRAETER

PLATONEM (103)

222 OMNIS HOMO CURRIT NECESSARIO PRAETER SORTEM

(103)

223 UTERQUE ISTORUM PRAETER UTRUMQUE DIFFERT AB

ILLO (104)

224 OMNIS HOMO PRAETER OMNEM HOMINEM DIFFERT AB

ISTO (104)

225 QUIDLIBET EST QUIDLIBET QUOLIBET EXCEPTO PRAETER

QUIDLIBET (104)

226 OMNIS HOMO EXCIPITUR PRAETER SORTEM (105)

227 SI NON ALIQUID CURRIT ALIQUID CURRIT (105)

TANTUM

228 TANTUM OMNIS HOMO CURRIT (106)

229 SI TANTUM OMNIS HOMO CURRIT NON T ANTUM OMNIS

HOMO CURRIT (106)

230 SI T ANTUM PATER EST NON T ANTUM PATER EST (107)

231 SI T ANTUM ALTER ISTORUM EST NON TANTUM ALTER

ISTORUM EST (107)

232 ALTER ISTORUM EST ALTER ISTORUMETTANTUM ALTER

ISTORUM EST (107)

233 T ANTUM UNUS HOMO EST UNUS HOMO (108)

234 T ANTUM UNUM EST (l08)

235 TANTUM ALTER ISTORUM EST HOMO VEL ASINUS (108)

236 TANTUM ALTER ISTORUM VEL RELIQUUS ISTORUM EST

HOMO VEL ASINUS (l09)

237 TANTUMDUOVEL TRIASUNTTRIA(109)

238 TANTUM HOMINEM ESSE HOMINEM EST VERUM (109)

239 TANTUM DEUM ESSE DEUM EST NECESSARIUM (l09)

240 T ANTUM HOMINEM ESSE ASINUM EST IMPOSSIBILE (110)

241 TANTUM HOMINEM ESSE HOMINEM EST POSSIBILE (110)

242 TANTUM VERUM POTEST ESSE VERUM (110)

 MAGISTER ABSTRACTJONUM: APPENDIX 1 165

243 TANTUMVERUMESTVERUM(III)

244 TANTUM VERUM EST IDEM VERO (111)

245 TANTUM VERUM OPPONITUR FALSO (111)

246 TANTUM ISTA QUAE NON SUNT HOMINES SUNT ANIMALIA

(112)

247 SORTES SCIT T ANTUM TRES HOMINES CURRERE (113)

248 POSSmILE EST SORTEM V1DERE TANTUM OMNEM

HOMINEM NON V1DENTEM SE (114)

249 SI NON AUUD CURRIT TANTUM SORTES CURRIT (115)

250 NON TV TANTUM ES ASINUS (116)

251 TU NON ERIS ASINUS DONEC CRAS (116)

SOLUS

252 SOLUS SORTES EST ALBUS QUO PLATO EST ALBIOR <EO>
(117)

253 SOLI IX HOMINES SUNT QUI NON SOLI SUNT (117)

254 SOLA IX ANIMALIA SUNT ALBA ET ILLA NON SUNT SOLA

(117)

255 SOLIUS BINARII PARS EST UNITAS ET NULLUS NUMERUS

(118)

256 SOLA NECESSARIA NECESSARIO SUNT VERA (119)

257 SOLA CONTINGENTIA ESSE VERA EST VERUM

CONTINGENS (119)

258 SORTES ET DUO <HOMINES> SUNT TRES (120)

259 SOLUS SORTES SCIT VII ARTES (120)

260 SORTES SCIT A (121)

261 SOLUS SORTES SCIT QUOD ALIUS A PLATONE IGNORAT

(121)

262 A SOLO SORTE DIFFERT QUICQum NON EST SORTES NEC

PARS SORTIS (122)

263 SOLUS SORTES EST INDIFFERENS SOLI SORTI (123)

264 SOLUS GENITIVUS PRAECEDITUR A SOLO NOMINATIVO

(123)
166 PAUL A. STREVELER

265 SOLA ASSUMPTIO PRAECEDIT SOLAM CONCLUSIONEM*

(124)

266 SOLA TRIA SUNT PLURES SOLIS DUOBUS (124)

267 SOLUS SORTES V1DETUR A SOLO SORTE (124)

268 SOLUS SORTES VlDET SOLUM SORTEM (124)

269 SOLUS SORTES VIDET SE (124)

270 SOLUS SORTES SCIT QUOD A NULLO ALIO SCITUR (125)

271 SOLI SORTI ALIQUID EST IDEM (125)

272 AD SOLUM SORTEM ESSE SEQUITUR SORTEM ESSE (126)

273 AD SORTEM ESSE SEQUITUR SOLUM SORTEM ESSE (126)

274 SOLUM SORTEM ESSE SEQUITUR AD SORTEM ESSE ET AD

ALlUM ESSE (126)

275 SOLUS SORTES EST SI SORTES ET ALIUS HOMO SUNT (127)

IMPOSSIBILE

276 IMPOSSIBILE FUIT POSSIBILE (128)

277 IMPOSSIBILE POTEST ESSE VERUM (128)

278 IMPOSSIBILE EST TE SCIRE PLURA QUAM SCIS (128)

279 IMPOSSIBILE EST DICI QUOD IMPOSSIBILE EST DICI (129)

280 QUOD NULLI EST DUBIUM OMNIBUS EST CERTUM (129)

281 SICUT SE HABET IMPOSSIBILE AD NON IMPOSSIBILE SIC

SE HABET VERUM AD NON VERUM (129)

282 SICUT SE HABET HOMO AD NON HOMO SIC SE HABET

ANIMAL AD NON ANIMAL (130)

283 SICUT SE HABENT PROPINQUITAS ET DISTANTIA MEl AD

ROMAM SIC SE HABENT PROPINQUITAS ET DISTIANTIA
MEl AD PARIETEM (130)

284 SICUT SE HABET NOVENARIUS AD SENARIUM SIC SE

HABET SENARIUS AD QUATERNARIUM (130)

NECESSARIUM
285 TE SEDERE DUM SEDES EST NECESSARIUM (130)

286 OMNE QUOD EST NECESSE EST ESSE QUANDO EST (131)
 MAGISTER ABSTRACTIONUM: APPENDIX 1 167

287 NECESSARIUM PUIT NON NECESSARIUM (131)

288 NECESSARIUM POTEST ESSE FALSUM (132)

289 OMNIS HOMO DE NECESSITATE EST ANIMAL (132)

290 OMNIS ANIMA EST AllQUA ISTARUM DE NECESSITATE

(133)

291 CONTINGENTIA NECESSARIO SUNT VERA (133)

292 ANIMA ANITCHRISTI NECESSARIO ERIT (134)

293 <DE NECESSITATE SORTES EST HOMO> (135)

294 QUICQUID EST DE NECESSITATE VERUM EST

NECESSARIUM (135)

295 QUICQUID DE NECESSITATE EST VERUM VEL FALSUM EST

NECESSARIUM VEL IMPOSSmILE (135)

296 OMNE QUOD NECESSE EST ESSE VEL NON ESSE NON
CONTINGIT ESSE (136)

297 TE SEDERE NECESSARIO EST ESSE VEL NON ESSE (136)

298 NECESSARIUM ET NON NECESSARIUM ET POSSmILE

CONTINGERE DICUNTUR (136)

299 OMNIS HOMO CURRIT CONTINGENTER (137)

300 SORTES VIDET UTRUMQUE ISTORUM CONTINGENTER (138)

POSSIBILE

301 PossmILE EST OMNEM HOMINEM DIFFERRE AB

ANTICHRISTO (138)

302 POSSIBILE EST ANTICHRISTUM ESSE HOMINEM (138)

303 POssmILE EST SORTEM SCIRE QUICQUID PLATO SCIT (140)

304 POSSmILE EST OMNEM HOMINEM CURERE (140)

305 POSsmILE EST OMNE <ANIMAL>ESSE HOMINEM (141)

168 PAUL A. STREVELER

Appendix 2
Summary of Sophisms treated in Abstractiones, MSS Digby 2
and Royal 12

1. OMNIS HOMO EST OMNIS HOMO

Royal:

Proof: By induction and the dictum of Boethius.

Disproof: The false (i.e. 'Socrates is every man') follows.

Solution: Sophism is simply FALSE. Proof contains fallacy of the

consequent.

Digby:

Proof: Ibid.

Disproof: Ibid.

Solution: By composition and division.

Text:

Proof: Ibid.
Disproof: Ibid.

Solution: By equivocation. Proof contains fallacy of the consequent.

2. OMNIS HOMO EST TOTUM IN QUANTITATE

Royal:

Proof: Induction with respect to various quantities.

Disproof: Sophism has a false instance ('Aliquis homo est totum in

quantitate')

Solution: Sophism is TRUE.

Digby:
Proof: Ibid.

Disproof: Ibid.
Solution: By composition and division. Disproof contains fallacy of
figure of speech. Proof contains fallacy of the consequent.
 MAGISTER ABSTRACTIONUM: APPENDIX 2 169

Text:

Sophism is treated with previous sophism.

3. OMNIS HOMO EST UNUS SOLUS HOMO

Royal:

Proof: (1) By induction (2) Convertibility of 'ens homo' and 'unus

homo'.

Disproof: Sophism has a false instance (i.e. referring to 'woman').

Solution: Sophism is simply FALSE.

Digby:
Proof: Ibid., (1).
Disproof: (1) Opposite is predicated of the opposite. (2)
Syllogistically. (3) 'one man only is one man only'.
Solution: Sophism is simply TRUE. Disproofs (2) and (3) must be
distinguished according to composition and division.
Text:

Proof: Ibid., (1) and (2).

Disproof: Ibid., (1), (2), (3).

Solution: By composition and division. <Solution to this sophism is

coupled with nos. 4, 5, 6, and 7.>.

4. OMNES APOSTOL! SUNT DUODECIM

Royal:

Proof: By induction to each apostle.

Disproof: 'Some apostles are not twelve'.

Solution: By composition and division.

Digby:
Proof: Ibid.

Disproof: (1) Ibid., (2) Its contradictory is true.

Solution: Ibid.
170 PAUL A. STREVELER

Text:

Proof: Ibid.
Disproof: (1) Ibid.

Solution: Ibid.

5. OMNE ANIMAL FUIT IN ARCHA NOAE

Royal:

Proof: By induction.
Disproof: Syllogistically: 'Every man is in the arch of Noah' would
be true.

Solution: Ambiguous according to third mode of equivocation.

<NB. There is interpolated here in Royal a note on the distinction

between 'signification' and 'supposition'.>
Digby:

Proof: Ibid.

Disproof: (1) Through the mode of questioning (Le. Every animal

that is or every animal that was... ?) (2) Consequent is false.
Solution: By distinguishing species and individuals.

Proof contains fallacy of figure of speech and fallacy of the consequent.

Text:

Proof: Patet.
Disproof: Ibid., as in Digby.

Solution: By remote and proximate parts according to some; others

accuse the proof of fallacy of the consequent.

6. OMNE COLORATUM EST

Royal:

Proof: Induction.
Disproof: 'Every white exists' would be true.
Solution: Sophism is TRUE.
 MAGISTER ABSTRACTIONUM: APPENDIX 2 171

Digby:

Proof: Ibid.
Disproof: (1) Ibid., (2) 'omnis' requires at least three appellata.

Solution I: By the distinction of species and individuals. Disproof

contains fallacy of the consequent. Nevertheless, Sophism is simply
FALSE.
Solution II: There is fallacy of equivocation upon the term 'est'.
Second disproof contains fallacy of the consequent.
Text:

Proof: Ibid.
Disproof: Ibid., as in Digby.
Solution: By distinguishing remote and proximate parts.
Disregarding this distinction, it must be responded that the sophism is
simply TRUE. .

7. OMNIS PHOENIX EST

Royal: (casus)

Proof: Denial is false, so sophism is true.

Disproof: 'Omnis' requires three appellata.

Solution: Sophism is DOUBTFUL and its contradictory.

Digby: (no casus)

Proof: Ibid.
Disproof: Ibid.

Solution: Sophism is simply FALSE.

Text: (no casus)
Proof:Ibid.

Disproof: Ibid.
Solution: Sophism is FALSE. Proof is rejected by interemption.
172 PAUL A. STREVELER

8. QUICQUID EST VEL NON EST EST

Royal: <Sophism is not treated in Royal but is mentioned later with

OMNIS PROPOSmO VEL EIUS CONTRADICTORlA EST VERA.>

Digby:

Proof: "Whatever is, is; but whatever is, is or is not, so ... "

Disproof: "Whatever is, is or is not; Caesar is or is not, so Caesar

is."

Solution: By composition and division. Proof contains fallacy of the

consequent; disproof contains figure of speech.

Text:

Proof: Ibid.

Disproof: Ibid.

Solution: <Offered under next Sophism; ibid., as in Digby.>

9. OMNE BONUM VEL NON BONUM EST ELIGENDUM

Royal: <Sophism is treated later after OMNE VERUM ET DEUM ESSE
DIFFERUNT>.

Proof: "Every good is chosen, <every good is good or not good>; so

every good or not good is chosen."

Disproof: 'Every bad is chosen' would be true.

Solution: By composition and division.

Digby:

Proof: Ibid.

Disproof: Ibid.

Solution: Ibid: Proof contains fallacy of the consequent. Disproof

contains fallacy of figure of speech.

Text: <As in previous sophism>

10. OMNE RATIONALE VEL IRRATIONALE EST SANUM

Royal: <Sophism is not treated>

Digby:
 MAGISTER ABSTRACTIONUM: APPENDIX 2 173

Proof: "This rational or irrational <animal> is healthy and that etc.,

so ... "

Disproof: "Every rational or irrational <animal> is healthy, but every

animal is rational or irrational; so every animal is healthy."

Solution: By composition and division. Proof contains fallacy of the

consequent, disproof, figure of speech.

Text:
Proof: Ibid. by induction.

Disproof: Ibid.

Solution: Must be distinguished as in previous sophism (composition

and division).

11. OMNIS PROPOSmO VEL EIUS CONTRADICTORIA EST VERA

Royal:

Proof: Induction.

Disproof: "Every proposition or its contradictory is true, but every

proposition is every proposition or its contradictory; so every proposition
is true."

Solution: By composition and division.

Digby:

Proof: Induction.

Disproof: (1) Ibid. as in Royal; (2) Ibid. as (1) but with reference to
false propositions. (3) "Every proposition or its contradictory is true, no
false proposition is true; so no false proposition is a proposition or its
contradictory. "

Solution: By composition and division. Disproofs (1) and (2) contain

fallacy of figure of speech.

Text:
Proof: Ibid. as in Royal

Disproof: (1) as in Digby; (2) "Every proposition or its contradictory

is true; every proposition or its contradictory is false; so whatever is false
is true." (3) Ibid. as (2) in Digby. (4) "Every proposition or its
contradictory is true of necessity; no contingent proposition is true of
necessity; so no contingent proposition is a proposition or its
contradictory. "

Solution: By composition and division.

174 PAUL A. STREVELER

12. TV ES QUILIBET VEL DIFFERS A QUOLIBET

Royal: <Returns to previous order: Sophism is treated after OMNIS

PHOENIX EST.>
Proof: Induction.

Disproof: It would follow that you differed from yourself.

Solution: Sophism is TRUE, because it is a disjunction, each part of

which is true. Disproof contains fallacy of the consequent.

Digby:

,Proof: Ibid.

Disproof: Ibid.

Solution: Ibid. by the rule: "Quicquid est verum de expos ito est
verum de exponente." Disproof contains fallacy of the consequent by the
rule: "Ad differentiarn respectu posterioris sequitur differentiarn prioris."
Text:

Proof: Ibid.

Disproof: (1) Each part of the disjunction is false (2) Ibid. as in

Royal.

Solution: Sophism is TRUE.

13. TV ES QUILIBET VEL A QUOLIBET DIFFERENS

Royal:

Proof: Induction.

Disproof: As in previous sophism.

Solution: Sophism is simply FALSE. Proof contains fallacy of the

consequent.
Digby:

Proof: Ibid.

Disproof: Both parts of this disjunction are false.

Solution: Sophism is simply FALSE by the rule: "Quicquid est

falsum de exposito est falsum de exponente." Proof contains fallacy of the
consequent (as in Royal).
 MAGISTER ABSTRACTIONUM: APPENDIX 2 175

Text: <No specific proof or disproof is given, Sophism is resolved

essentially as in Royal and Digby.>

14. TU SClS QUIDLIBET VEL QUIDLIBET lGNORAS

Royal: <Sophism is mentioned as TU SClS QUAELIBET VEL
QUAELIBET IGNORAS, but is not treated. Sophism treated in Royal here
is TU SClS QUICQUID SClS. Royal also mentions here TU SClS
QUIDLIBET VEL NIHIL.>
Digby:

Proof: By induction.
Disproof: "You know everything whatever or are ignorant of
everything whatever; but you don't know everything whatever; so, you are
ignorant of everything whatever."

Solution: Sophism is simply FALSE. Proof contains fallacy of the

consequent.

Text:
Proof: Ut praecedens.

Disproof: Ut praecedens.
Solution: Ibid. <Solution is briefer in Text, but not philosophically
different from Digby.>

15. TU SCIS QUIDLIBET VEL NllllL

Royal: Eodem modo sicut praecedentia probatur et solvitur.

Digby: <Sophism is not treated in Digby.>

Text:
Proof: By induction.
Disproof: "You know everything whatever or nothing; but you don't
know everything whatever; so you know nothing."
Solution: Sophism is FALSE. <Solution resembles that in Digby for
previous sophism>.
176 PAUL A. STREVELER

16. TU SCIS QUICQUID SCIS

Royal:
Proof: Induction: "You know this and that...etc. so ... "
Disproof: "Either now first you know whatever you know or not
now. If not now first, then now first you know this: 'God exists' . If not,
then before you knew whatever you knew; therefore before you knew this
'You exist in this instant'. Consequent is false. So sophism."
Solution: Sophism is TRUE. Disproof contains fallacy of the
consequent.
Digby:

Proof: Ibid.
Disproof: Ibid.
Solution: Sophism is simply TRUE by the rule: "Quicquid est verum
de exponente est verum de exposito." Disproof contains fallacy of the
consequent from inference from particular to singular. <There follows in
Digby a discussion and examples of other forms of fallacy of the
consequent.>
Text:
Proof: <Proof is missing in Text. Can be supplied from Royal or
Digby.>
Disproof: Ibid. <but with introductory lacuna>
Solution: Sophism is TRUE <Discussion resembles Digby>

17. A DE NUMERO ISTORUM QUORUM QUODLffiET DIFFERT AB

EO QUOD IPSUM ESSE NON EST ILLUD

Royal: <Sophism is not treated in Royal.>

Digby:
Proof: (casus): Let 'a' be Socrates' being; 'b', a's being; 'c', b's
being. Proof is obvious if the term 'illud' refers to any being whatever.

Disproof: Through the mode of questioning.

Solution: Must be distinguished through the mode of equivocation in
so far as 'illud' can distribute 'Socrates' or any of these beings whatever.
If the former, then the sophism is false and against the casus; if the latter,
then it is true.
 MAGISTER ABSTRACTIONUM: APPENDIX 2 177

Text:

Proof: Ibid.

Disproof: Ibid.

Solution: <Equivocation as in Digby not used in Text.> Sophism is

TRUE under the correct interpretation.

18. OMNIS HOMO EST ET AllUS HOMO EST

Royal:

Proof: By induction.

Disproof: Second part of the conjunction is false, because it denotes

that some man is other than every man, lacking human nature.

Solution: Sophism is TRUE. Disproof has a true antecedent and false

consequent.

Digby:

Proof: Ibid.
Disproof: ''This is a disjunction (!) one part of which is false, so ... "

Solution: Disjunction (!) is false, since each part is false. Proof

contains fallacy of the consequent. In another sense, however, the sophism
is TRUE, since both parts can be proved true if the verb is understood only
in the present tense.

Text:
Proof: Ibid.
Disproof: "This is a copulative one part of which is false, so ..."

Solution: Sophism is FALSE. But there is a proof that it is TRUE

<ibid. as in Digby>.

19. OMNIS HOMO ET DUO HOMINES SUNT TRES

Royal:

Proof: Induction.
Disproof: "Some man and two men aren't three, because Socrates
and Plato aren't three, and they themselves are two men; and some two
men are not three, so some man and two men aren't three."
178 PAUL A. STREVELER

Solution: Sophism is TRUE. Disproof does not contradict the

sophism.
Digby:

Proof: Ibid.
Disproof: Part is copulated of the whole, so sophism is false.
Solution: 'Omnis' here is equivocal as to the composite and divided
senses. To the disproof it is said that the copulation is question causes
some impropriety, but is not the cause of any falsity here.
Text:

Proof: Ibid.
Disproof: Ibid. <as in Digby>.
Solution: Ibid. <as in Digby>.

20. OMNE VERUM ET DEUM ESSE DIFFERUNT

Royal:

Proof: By induction.

Disproof: "Every truth and 'to be God' are different; so every truth
differs from 'God exists'; so 'God exists' differs from 'God exists'; so the
same differs from itself."

Solution: Solution is complex: the sophism ought to be conceded if

one simply understands the dictum (Deum esse) as a material expression.
If, however, it is understood as having meaning and reference to a
proposition of which it is true (Deus est), then the sophism is doubtable.
Digby:

Proof: Ibid.

Disproof: <Given after the resolution. Similar to Royal.>

Solution: Sophism is simply FALSE. Proof contains fallacy of the
consequent.
Text:

Proof: (1) Ibid. (2): "This proposition is true: 'a truth and God exists
are different' for all truths other than this truth 'Deum esse'; so a truth and
'Deum esse' are different. Why? Because if this term 'truth' is distributed
only for supposits for which it is true, this will be true: 'every truth and
"Deum esse" are different', since it would be equivalent to this: 'every
truth other than "Deum esse" is different from "Deum esse"'."
 MAGISTER ABSTRACTlONUM: APPENDIX 2 179

Disproof: Ibid.

Solution: Sophism is FALSE as in the preceding.

21. OMNIS HOMO EST ET QUll.JBET VIDENS ILLUM EST ASINUS

Royal: (casus): Every man sees only himself and there are many asses all
of whom see every man.

Proof: Each part of the sophism is true, so the whole is true.

Disproof: "Every man exists and whoever sees him is an ass; but
Socrates sees Socrates; so Socrates is an ass."

Solution: Must be distinguished according to amphiboly, since

'ilIum' can refer to 'every man' or to 'man' only. If the former, the
sophism is true. If the latter, it is false.

Digby:
Proof: (casus). Ibid.
Disproof: Ibid.

Solution: Ibid.

<Digby 2 ends here>

Text:

Proof: Ibid.
Disproof: Ibid. <Without particular reference to Socrates>

Solution: <No reference to amphiboly, but similar discussion

regarding the relative term.>

22. OMNIS HOMO EST ET QUILIBET DIFFERENS AB ILLO EST

NON HOMO

No specific proof or disproof is given in Text. Proof is by induction in

Royal. Disproof: "Every man exists and whoever differs from him is DOD­
man. Socrates differs from Plato, so Socrates is non-man."
Solution in Royal and the Text are virtually identical. Relative term must
be distinguished as in the preceding sophism. Royal notes the distribution
according to amphiboly; no reference to amphiboly in Text.
180 PAUL A. STREVELER

23. OMNIS HOMO QUI EST ALBUS CURRIT

Casus in Royal: Whatever is a white man runs and no one else. Casus in
Text: Every white man runs and no black nor any in between. Proof by
induction is given in Royal, but not in Text. Disproof is same in both.

Royal asserts the sophism as TRUE and denies the consequence in the
disproof. A counter argument is given to the effect that from accepting the
sophism an absurd consequent folIows, viz., that every white is running.
To this it is responded that it fails because it argues from an inferior to a
superior without an exclusive.
Text appears different here. Solution argues that the sophism is
ambiguous according to composition and division. Two further arguments
are given in the Text: (1) The relative clause can cover 'alI men' or just
'men'; if former, it is false; if latter, true. (2) The term 'qui' can be
understood implicatively or non implicatively.

24. OMNE QUOD EST VERUM SCIRI A TE EST VERUM

Proof and disproof are similar in Royal and the Text.

Royal argues the sophism is ambiguous according to amphiboly. Text
argues sophism is ambiguous according to composition and division and in
the composite sense it is additionally ambiguous according to amphiboly.
Royal considers a counter argument to the solution not in the Text
employing the technique of "obligatio".

25. OMNE QUOD EST VERUM SCIRl A TE EST FALSUM

Proof and disproof are similar in Royal and the Text.
Solution: Royal is corrupt here, obviously missing several sentences. It
notes the sophism is ambiguous (understood essentially in the same way as
the preceding sophism). Text notes the ambiguity as one of composition
and division.

26. DEUS ERIT IN QUOLffiET INSTANTI NON EXISTENS

Proof and disproof are similar in Royal and Text. Proof is by induction.
Disproof: 'Quolibet non' and 'nullum' are equivalent; so God exists at no
existing instant.

Solution: Royal argues that the sophism is false. The consequence of the
proof is good, but both antecedent and consequent are false.

Text has a much more complete treatment. Sophism must be

distinguished according to composition and division, but in BOTH senses
it is false.
 MAGISTER ABSTRACTJONUM: APPENDIX 2 181

27. OMNIS GRAMMATICUS EST ET QUILIBET SCIENS ILLUM

ESSE GRAMMATICUM EST TANTUM TALIS

<NB. Sophism considered in Text previously.>

Casus in Royal: There are three grammarians who are only grammarians
and these know concerning every grammarian that he himself is a
grammarian and only these (know this) and there is another who is a
grammarian, rhetorician and musician and he does not know someone to
be a grammarian except him himself.

Casus in Text: Whoever is only a grammarian knows concerning any

grammarian that he himself is the grammarian and he who is a grammarian,
rhetorician and musician does not know someone to be a grammarian
except himself.
Proof is similar in Royal and Text: This is a copulative each part of
which is true.

Disproof is similar, but more complete in Royal.

Solutions are somewhat different: Text simply declares that there is a

fallacy of the consequent here (apparently in the disproofl). Royal argues
the sophism needs to be distinguished according to amphiboly in so far as
the relative can refer to 'every grammarian' or just 'grammarian'.

28.0MNlS HOMO EST ANIMAL ET OMNE ANIMAL EST HOMO

Proof is similar in Royal and Text by induction. Disproof is similar in

Royal and Text: This is a copulative each part of which is false.
Solution: Sophism is FALSE because of insufficient induction.
Argument is same in Royal and Text.

29. OMNlS HOMO EST ANIMAL ET ECONVERSO

Proof is similar, by induction, in Royal and Text. Disproof is similar in

Royal and Text.
Solutions have similar vocabulary in Royal and Text.

30. CUIUSLIBET HOMINIS ASINUS CURRIT

Casus: Each has an ass that runs and everyone has one ass in common
which does not run.
Proof by induction in Royal and Text. Disproof is similar: Of any man,
an ass runs; so an ass of any man runs. Consequent is false.
182 PAUL A. STREVELER

Solutions are similar in Royal and the Text: Sophism is TRUE and
disproof contains fallacy of figure of speech. <Text now considers what
is, in effect, a new casus with the same sophism.>

31. CUIUSLIBET HOMINIS OCULUS EST DEXTER

Text does not offer a proof or disproof for this sophism. These are given
in Royal. Treatment of the sophism is essentially the same in both MSS.

32. OMNIS HOMO MORITUR QUANDO UNUS SOLUS HOMO

MORITUR
Casus: Every man dies successively, such that one does not die when
another dies.
Proof is by induction and similar in Royal and Text. Disproof (i.e.
sophism leads to a contradiction) is similar in Royal and Text.
Solution is same in Royal and Text: Sophism is FALSE, but the analysis
of the sophism and the confusion involved is somewhat different in the
two manuscripts.

33. QUICQUID AUDITUR A PLATONE PROFERTUR A SORTE

Casus: Socrates says: 'No man is an ass' and Plato hears the whole
sentence except the sign. (Ibid. in Royal and Text).

Proof and disproof are similar in Royal and Text.

Solution: Royal and Text assert the sophism as TRUE. Royal argues the
disproof is to be denied because when saying 'No man is an ass', Socrates
says the very proposition which is true and part of which is false. Text has
a more detailed analysis, arguing that the argument fails because it takes
'aliquid' for 'quale' and commits a fallacy of accident.

34. ALBUM FUIT DISPUTATURUM

Casus: Socrates was white and isn't now nor will be white and Socrates
will dispute, isn't now nor did dispute. (Ibid. in Royal and Text).
Proof is similar in Royal and Text: What was white was about to dispute,
so ...

Disproof is similar in Royal and Text: Whatever was about to dispute,

either disputes, disputed, or will dispute; so ...
Solutions are similar, but Text is much more complete. Royal simply
points out that there is equivocation here because 'album' can supposit for
 MAGISTER ABSTRACTJONUM: APPENDIX 2 183

some whole white thing that exists or that existed. Ifin the first way, the
sophism is false; if in the second, it is true. Text notes that the sophism is
TRUE and the disproof commits fallacy of confusing 'quid' and 'quale'.
For the inference: 'Whatever was about to dispute either disputes,
disputed, or will dispute; this (thing) was about to dispute, so .. .', is a true
conclusion with respect to 'this thing will dispute', but not with respect to
'this white thing will dispute'; because it will not be white. Text now
explains that the changing of supposition from past, present, to future is
the source of the error.

35. DEUS SCIT QUICQUID SCIVIT

Proof is similar in Royal and Digby: God knew everything and forgets
nothing, so ...
Disproof is similar in Royal and Text, but Text introduces "obligatio"
technique: God knows whatever he knew; he knew that you don't exist; so
you don't exist.

Solution: Royal argues the sophism is FALSE because from it

something false follows. Also, given the sophism will be true, it would
follow that God will know the false. Both antecedent and consequent in the
proof should be denied, unless you understand by 'sciri' simply 'to
understand', which does not entail truth.
Solution in Text is more detailed and involves a discussion of God's
knowledge being unchanging and a discussion of a sense of 'is' that does
not have temporal reference.
<Transition to UTERQUE in Text>

36. 1ST! FERUNT LAPIDEM

Casus: Pointing to two men, one carries one stone, the other another
stone.
Proof is same in Royal and Text: Each carries a stone; so they carry a
stone.
Disproof is same in Royal and Text: No stone is carried by both, so ...

Royal argues the sophism to be FALSE and the consequence in the proof
should be denied.
Text presents a more detailed analysis: 'Uterque' signifies two things
through the mode of partition and division, thus rendering a verb singular
in number. 'Ambo' signifies through the mode of collection, thus
rendering a verb plural in number. Thus one cannot infer 'ambo' from
'uterque' .
/84 PAUL A. STREVELER

37. ISTI SCIUNT SEPTEM ARTEM

Casus: Pointing to two, one knowing three, the other four.

Proof is similar in Royal and Text.
Disproof is similar in Royal and Text.

Solution: Royal declares the sophism to be TRUE. The disproof should

be denied, because the antecedent is true but the consequent is false.

Text declares the solution to be clear from the analysis of the preceding
sophism. But there might be a counterexample: 'They know that God
exists, so each knows that God exists'. This isn't a counterexample,
because 'God exists' is a simple; whereas 'septem artes' is complex ...

38. OMNIS ANIMA EST IN TE

<Sophism is not treated in Text>.

39. OMNE QUOD EST VERUM EST VERUM IN HOC INSTANTI

<Royal ends here>
 'Omnis phoenix est': Quantification and Existence in a new
sophismata-collection (MS elm 14522)
by Andrea Tabarroni

It is well known that between (approximately) 1250 and 1275

important innovations took place in the Parisian Faculty of Arts. The
introduction of the Aristotelian libri naturales in the syllabus of 1255 marks
the end of a period when the masters' interest focused mainly on the
disciplines of the trivium'! The flow of exegetical works resulting from
courses on logic and grammar now began to be matched by a comparable
quantity of commentaries devoted to natural and moral philosophy.
Moreover, Averroes' interpretation of Aristotle was more carefully
investigated and assimilated. The free attitude of the Arabic thinker
towards speculation was assumed as a model by some masters, a trend
eventually leading to a clash with ecclesiastical authority concerning the
limits of orthodox philosophical teaching.2

Signs of the new philosophical atmosphere are traceable also in the

most peculiar practice of the Arts Faculty, namely, the practice of logical
and grammatical sophismata held by masters and bachelors with the intent
both of testing doctrines in difficult cases and of training dialectical skills.

As a matter of fact, sophisms in this period display a structure far

more developed than in the past. They are still arranged according to the
four traditional parts (the proof, the disproof, the distinction and the truth
or falsity of the sophismatic proposition). But these parts are now rather
intended as general headings under which special questions, or
problemata, of different sorts can be investigated. 3 Thus, under the
heading de solutione of the sophism 'QUANTO ALIQUID MAlUS EST,
TANTO MINUS VIDETUR' one can find interesting remarks on the
fallacy of accident; and it is even possible to meet a metaphysical analysis
of the different kinds of unity in the sophism 'TANTUM UNUM EST.' 4

ISee P.O. Lewry, "Thirteenth-Century Examination Compendia from the Faculty of

Arts", in Les genres litteraires dans les sources tMologiques et philosophiques
medievales. Definition. critique et exploitation. Actes du Colloque International de
Louvain-Lo-Neuve. 25-27 rnai 1981 (Publications de I'Institut d'Etudes medievales, 2e
serle: Textes, Etudes, Congres, vol. 5), Louvain-La-Neuve: Universite Catholique de
Louvain 1982, pp. 101-16.
2See R.A. Gauthier, "Notes sur Ie debuts (1225-1240) du premier 'averroisme"', Revue
des Sciences philosophiques et tMologiques 66,1982, pp. 321-74.
3See S. Ebbesen, "Three 13th-century Sophismata about Beginning and Ceasing",
Cahiers de l'lnstitut du Moyen-Age Grec et Latin 59, 1989, pp. 121-80, esp. p.127,
n.l. For the use of the term in MS Milnchen, Bayerlsche Staatsbibliothek, clm 14522
see below, nn. 4 and 16.
4For a sophismatic treatment of the fallacy of accident, see clm 14522, ff. 37ra-va:
"Iuxta hoc querltur que diuersificatio medii faciat fallaciam accidentis. Solutio. Dico
quod in predicto problemate sophismatis est fallacia accidentis. Notandum ad
intellectum fallacie accidentis quod duplex accidens exigitur ad hoc quod fiat fallacia

185
186 ANDREA TABARRONI

Moreover, the range of the sophisms most commonly disputed now

became narrower. Sophismata-collections from this period bear witness to
some two to three dozen sophisms coming time and again under
discussion: a far lesser number than those contained in Sherwood's or
Nicholas of Paris' Syncategoremata, not to speak of the more than 300
sophisms collected in Richard's Abstractiones. A greater variety is present
also in the collection contained in MS Vat.lat. 7678, probably dating from
the first half of the century.5

These features point to the fact that, by the 1250s, some sophisms had
reached the rank of canonical frameworks for the discussion of some
difficult logical questions. The most well known example is that of
'OMNIS HOMO DE NECESSITATE EST ANIMAL', where we find
discussed the crucial question of the truth of present tense propositions
whose subjects fail to have extra-mental bearers.6

Hence, this period represents a stage in a development leading at the

end of the century to the so called "modistic" form of the sophismatic
dispute, where attention focuses mainly on metalogical or epistemological
questions and sophismatic propositions are considered only as mere
pretexts. But sophisms dating from the third quarter of the 13th century
bear witness also to important developments in the history of various
doctrines, allowing us to retrace the path leading from the terminist
paradigm of the Summulist period to the modistic approach epitomized in
the works of Simon of Faversham and Radulphus Brito.

accidentis ... Notandum igitur pro regula quod quando medium diuersificatur inter
maiorem extremitatem et minorem fit fallacia accidentis. Set duplex est diuersificatio
medii, scilicet substantialis et accidentalis ... " (f. 37ra). The analysis of unity is
contained ibid. ff. 4Iva-42rb: "Aristotiles etiam posuit quod unum conuertitur cum
ente, et est eiusdem intentionis cum eo, cui us sectator et expositor est Auerois.
Videamus ergo de significato unius, et dicamus cum Aristotile quod eadem sunt
principia substantie et accidentis .... " (f. 42vb) For the converse case of a sophismatic
treatment of the problem of unity in a philosophical commentary see P. Delhaye, Siger
de Brabant. Questions sur la Physique d' Aristote, Louvain: Editions de I'Institut
Superieur de Philosophie 1941, pp. 43-44 (I. I, q. 17: utrum tantum unum ens sit).
5The works of William of Sherwood and of Nicholas of Paris are respectively edited in
J.R. O'Donnell, "The Syncategoremata of William of Sherwood", Mediaeval Studies 3,
1941, pp. 46-93 and in H.A.G. Braakhuis, De J3de eeuwse Tractaaten over
syncategorematische termen. Deel II: Vitgave van Nicolaas van Parijs'
SYllcategoreumata, Nijmegen: Krips Repro Meppe11979. Richard's sophisms are listed
by L.M. De Rijk, Logica Modernorum II.l: The Origill and Early Deve[opmellt of the
Theory of Supposition, Assen: Van Gorcum 1967, pp. 62-71, but see now Paul
Streveler's contribution in this volume. The list of the sophisms in MS Citta del
Vaticano, Bib!. Apost. Vat., lat. 7678 reported by M. Grabmann, Die
Sophismataliteratur des 12. und 13. Jahrhunderts mit Textausgabe eines Sophisma der
Boethius von Dadell. Miinster LW.: Aschendorff, Beitriige zur Geschichte der
Philosophie und Theologie des Mittelalters 36.1, 1940, pp. 33-41 is to be corrected
with the additions by Braakhuis, De 13 eeuwse, Deel I: IlIleidellde studie, pp. 33-65 and
420-2. For the mid 13th-century collections see below, n. 7.
6See S. Ebbesen, "Talking about what is no more. Texts by Peter of Cornwall (?),
Richard of Clive, Simon of Faversham and Radulphus Brito", Cahiers de I'lnstitut du
Moyen-Age Grec et Latin 55, 1987, pp. 135-68, with the relevant literature.
 'OMNIS PHOENIX EST'.' QUANTIFICATION AND EXISTENCE 187

However, for all its stimulating interest for the history of medieval
logic, our knowledge of the sophismatic practice of this period is entirely
dependent on a few sophismata-collections.7 My present purpose is then to
enlarge this base of knowledge by presenting a new collection, hitherto
erroneously attributed to Albert of Saxony. First, I will describe the
collection, which proves to date back to the period before 1275. Then I
will focus attention on the sophism 'OMNIS PHOENIX EST', dealing
with a problem that lies at the heart of the later debate over modistic
semantics, namely the problem of restriction.
1. The collection
1.1. The manuscript

According to the remarkably accurate description provided by Barbara

Faes de Mottoni, clm 14522 is a parchment MS of 208 leaves, consisting
of three parts written by different hands. 8 Part 1 (ff. 1-71, forming 8
fascicles) contains a collection of 13 sophismata plus one fragment,

7The following is a list of the known collections approximately dating to the third
quarter of the 13th century, with the relevant literature: I) MS Erfurt,
Wissenschaftliche Bibliothek der Stadt, Ampl. 4° 328, ff. Ira-73vb: see Braakhuis, De
13de eeuwse, Deel I, pp. 83-7 and idem, "Kilwardby versus Bacon?", in Medieval
Semantics and Metaphysics. Studies Dedicated to L.M. de Rijk, ed. E.P. Bos,
Nijmegen: Ingenium 1985, pp. 111-42; 2) MS Paris, Bibliotheque Nationale, lat.
16135, ff. 3ra-37rb (first COllection): see A. de Libera, "La Iitterature des Sophismata
dans la tradition terministe parisienne de la seconde moitie du xm e siecle", in The
Editing of Theological and Philosophical Texts from the Middle Ages, ed. M.
Asztalos, Stockholm: Almqvist & Wiksell 1986, pp. 213-44; A. de Libera, "La
problematique de I'instant du changement au xm e siecle: contribution a I'histoire des
sophismata physicalia", in Studies in Medieval Natural Philosophy, ed. S. Caroti,
Firenze: OIschki 1989, pp. 43-93; and A. de Libera, "Le sophisma anonyme 'Sor
desinit esse non desinendo esse' du Cod. Parisinus 16135", Cahiers de l'lnstitut du
Moyen-Age Grec et Latin 59, 1989, pp. 113-120; 3) MS Paris, Bibliotheque
Nationale, lat. 16135, ff. 38ra-103vb (second COllection): see de Libera, "La Iitterature
des Sophismata" and "Le sophisma anonyme"; I. Rosier, '''0 Magister .. .':
Grammaticalite et intelligibilite selon un sophisme du xm e siecle", Cahiers de
l'lnstitut du Moyen-Age Grec et Latin 56, 1988, pp. 1-102; and C. Brousseau-
Beuermann's contribution in this volume; 4) MS Paris, Bibliotheque Nationale, lat.
16618, ff. 137r-52vb: see de Libera, "La Iitterature des Sophismata", pp. 213-5; 5) MS
Worcester, Cathedral Library, Q. 13, ff. 24vb-53vb: see C. Lohr, "Aristotelica
Britannica", Thtfologie und Philosophie 53, 1978, pp. 97-9; P.O. Lewry, "The Oxford
Condemnations of 1277 in Grammar and Logic", in English Logic and Semantics from
the End of the Twelfth Century to the Time of Ockham and Burleigh, ed. H.A.G.
Braakhuis et aI., Nijmegen: Ingenium 1981, pp. 235-78; P.O. Lewry, "Oxford Logic
1250-1275: Nicholas and Peter of Cornwall on Past and Future Realities", in The Rise
of British Logic, ed. P.O. Lewry, Toronto: Pontifical Institute of Mediaeval Studies
1985, pp. 193-234; Ebbesen, "Talking about" and Ebbesen "Three 13th-century"; 6)
MS Krak6w, Biblioteka Jagiellonska 649, ff. 253ra-271vb: see A. Tabarroni, "lncipit
and desinit in a thirteenth-century sophismata-collection", Cahiers de l'lnstitut du
Moyen-Age Grec et Latin 59, 1989, pp. 61-111.
8See B. Faes de Mottoni, AegidU Romani Opera omnia. I. Catalogo dei manoscritti
(457-505).115*. RepubbUca Federale di Germania (Monaco), Firenze: Olsckhi 1990, n.
483, pp. 145-50.
188 ANDREA TABARRONI

interrupted at ff. 3lrb-36va by a short anonymous compilation collecting

the common subject-matter (communia) of the Prior and Posterior
Analytics. Part 2 (ff. 72-148) contains Giles of Rome's commentary on
the Elenchi, while part 3 (ff. 149-208) reports four anonymous
commentaries on De sensu, De memoria, De somno et vigilia and De morte
et vita followed by Adam of Buckfeld's works on the Meteora and the
pseudo-Aristotelian De piantis. It is worth noticing that the four
anonymous commentaries are most probably the work of Adam of
Whitby, a little known English philosopher whose activity is located by
R.-A. Gauthier around 1265.9

In 1347 our codex was kept in the library of St. Emmeram's

Benedictine Abbey in Regensburg. The description provided by the ancient
catalogue shows that by that year the codex was already in its present
state. 10 But the ancient foliation and the mention of two owners provide
evidence according to which part 2 was originally an independent
manuscript. On the other hand, parts 1 and 3 probably belonged together,
since both show traces of the activity of an ancient editor supplying a list
of sophisms, headings and notes to the texts.!!

Many of the sophisms recorded in part 1 appear to be incomplete, but

this is only due to the fact that two fascicles have been misplaced. As a
matter of fact, only two sophisms (nr. 4, which is a short fragment, and
nr. 14, which is the last one) are left unfinished. The original order, altered
in the course of binding, can be easily restored; hence, I will number the
sophisms according to the list reported in the Appendix.12

9Whitby's authorship of the commentary on De sensu is suggested by Faes de Mottoni,

p. 148. She points to the similar incipit of the attributed copy in MS Paris, Bibl. Nat.
lat. !6149, ff. 62ra-67va. The Super de memoria has a different incipit from its
companion in the Parisian MS (ff. 6Orb-6Irb), but cf. the passage reported by R.-A.
Gauthier in Sancti Thomae de Aquino Opera omnia, t. XIV,2: Sentencia libri de sensu
et sensato, cura et studio FF. Praedicatorum, Roma: Commissio Leonina 1984,
(Pre/ace) p. 125* with the following: "Ex hi is autem uidetur quod ista pars est de
continuatione scientie tradite in libro De anima, sicut pars precedens que est de
mouente. Sicut enim ibi agitur de uirtutibus apprehensiuis in quantum sunt motiue
uno modo, sic hic agitur de eiusdem uirtutibus a1io modo, licet contrarium huius
appareat ex recapitulatione facta in fine !ibri De sensu et sensato et etiam ex quodam
dicto in hac parte." (clm 14522, f. 155va26-32) The two commentaries on De sompno
and De morte et vita appear to be linked by close affinity to the preceding ones. Adam
of Whitby is not mentioned in Emden's register of the Oxford University.
lOB. Bischoff and C.E. Ineichen-Eder, Mittelalterliche Bibliothekskataloge Deutschlands
und der Schweitz, vol. IV/I, Miinchen: Beck 1977, pp. 159-60: "Item sophysmata de
villa Parysiensi. Item Egidius super librum elencorum. Item super de sensu et sensato.
Item liber de sompno et vigilia. Item de morte et vita. Item super metheororum. Item
super de plantis, omnes in uno volumine."
!!See Faes de Mottoni, Aegidii Romani I, 1/5*, n. 483, p. 146 and 148, under the
section "Correzioni e annotazioni."
!2The following would have been the correct order of the eight fascicles forming part 1
of the codex: I-II-IV-VI-VII-VIII-III-V. Due to the misplacing of the two quires, the
sophisms are presently in the following order: 1 - 2 - 3 (beg.) - 11 (end) - 12 - 13
(beg.) - 3 (end) - 4 (fragm.) - Communia - 13 (end) - 14 (mutil.) - 5 - 6 - 7 - 8 - 9 - 10
- II (beg.).
 'OMNIS PHOENIX EST': QUANTIFICATION AND EXISTENCE 189

The sophisms are all written by the same hand, a small cursive
bookhand probably of German origin. The copyist worked with care,
diligently noting the peculiarities of his model and also reproducing several
marginal glosses, which deserve attention if one is to gain some
knowledge about the use of this sophismatic material in the schools. 13
1.2 The unity of the collection

A detailed analysis of the collection is not necessary in order to discard

the ancient attribution to Albert of Saxony.1 4 We are obviously in the
presence of typical 13th century sophismata, with long sections dealing
with the sophismatic distinction, the truth of the puzzling sentence and the
validity of its proof and disproof. Most of them are structured according to
the same general scheme, requiring that each question be solved before
turning to the next one. Only sophisms 6 and 14 do not follow this pattern,
introducing as they do a general solution only after all the questions have
been disputed. This is a pattern followed e.g. by the sophisms contained in
MS Krak6w, Bibl. Jagiellonska 649.1 5

Accurate recording of the oral debates varies in degree from sophism

to sophism, but all of them have been revised and edited by the master
who determined the disputation. Usually the questions are arranged
according to the canonical form, with the two sets of arguments pro and
contra, but occasionally some fragments survive recording exchanges
between the respondens and his opponents.

In two cases the master's initiative in the arrangement of the

disputational material is made explicit. In sophism 6, before turning to the
general solution, the master says "De quarto problema nihil fuit
oppositum, ideo ponatur in questione" (clm 14522, f. 49vb29-30). In
sophism 7 the fourth section, which collects the two headings de modo
probandi et improbandi and de veritate, is introduced as follows: "Circa
quartum sic proceditur. Et quia minus (fort. pro nimis) prolixe esset hic
ponere, <quia> hec disputata est in uno problemate, ideo faciamus duo"
(elm 14522, f. 53vb6-8).

Hence, it turns out that, while editing the sophism, the master is free
to reorganize the discussion in order to bestow a coherent structure on it.
This accounts for the occurrence of some locutions by which the author
refers to other parts of the sophism, such as 'sicut dictum fuit in solvendo'
or 'in opponendo' or 'in respondendo' and also in one occasion 'superius

I3The copyist's attitude is described by Faes de Mottoni, Aegidii Romani I, 1/5*, n.

483, p. 146.
14The attribution dates back at least to the last decades of the 15th century, when
Laurentius Aicher of St. Emmeram wrote the table of contents at the beginning of the
MS describing the collection as "Sophysmata Alberti de Saxonia." On the history of
St. Emmeram's library in the late Middle Ages see, B. Bischoff, Studien zur
Geschichte des Klosters St. Emmeram im Spiitmittelalter (1324-1525), in idem,
Mittelalterliche Studien, II. Stuttgart: Hiersemann 1967, pp. 115-49 and Bischoff and
Ineichen-Eder, Mittelalterliche Bibliothekskataloge, IV.I, pp. 99-138.
15See Tabarroni, "Incipit and desinit", p. 64.
190 ANDREA TABARRONI

in solvendo corpus sophismatis', referring to the solution to the section de

veritate. 16
All the sophisms are without ascription in the manuscript, but some of
them probably belong to the same author. As a matter of fact, the author of
sophism 12 leaves out discussion of two distinctions by referring to the
treatments already inserted, respectively, in sophisms 1 and 14. In the
same way, the author of sophism 8 refers to the analysis of the fallacy of
accident dealt with in the last question of sophism 13.l7 On the other hand,
the sophisms were not all written by the same author. Evidence for this
conclusion is found in sophism 9, where two references occur.

The first one is to a sophism 'OMNIS ANIMA NECESSARIO EST

IUSTA', referred to as following in the collection (inJerius), but which
actually does not occur; the second one is to a sophism 'OMNIS HOMO
DE NECESSITATE EST ANIMAL'.IS Sophism 6 of the collection
actually bears this title, but a closer look at the content of the two works
definitely rules out their common authorship. The master who determined
sophism 9, in fact, is an opponent of the widespread distinction between
two ways of being, esse actuale and esse essentie or esse habitudinale.
Accordingly, dealing incidentally with the much debated question utrum
'homo est animal' sit vera, nullo homine existente, he denies the
possibility for per se-propositions to be true when their subject terms fail
to refer to actually existing things. 19

l6See e.g. clm 14522, f. 6rb42 (sicut dictumfuit in soluendo), f. 29vb31 (ut uisum est
superius in soluendo), f. 19va3-4 (sicut etiam dictumfuit in opponendo), f. 16va20
(sicut dictumfuit superius in respondendo), f. 59ra34 (et hoc planius maniJestabitur in
soluendo corpus sophismatis in secundo problemate), f. 61va31-32 (superius in
soluendo corpus sophismatis).
17Clm 14522, f. 24rall-13: "De prima distinctione nihil dicemus nisi iIIud quod dictum
est in iIIo sophismate TOTUS SORTES etc. De secunda nihil dicemus nisi quod
dictum fuit in iIIo sophismate T ANTUM UNUM EST"; ibid., f. 57vb36-39: "Circa
quartum sufficienter processum est disputando hoc sophisma 'QUANTO ALIQUID
MAlUS EST TANTO MINUS VIDETUR'; ibi discussum est que diuersificatio medii
faciat fallaciam accidentis et que non."
lSIbid., f. 60vb8-14: "Primum quod ponunt est distinctionem ualere ad propositum.
Quod falsum est, ut iam uidebitur in soluendo quandam aliam distinctionem que
communiter ponitur in iIIo sophismate OMNIS ANIMA NECESSARIO EST lUSTA,
scilicet quod hec dictio 'necessario' potest determinare compositionem principalem
predicati ad subiectum aut potest determinare predicatum gratia compositionis in ipso
intellecte. Quod similiter est falsum, ut iam patebit inferius in eodem sophismate suo
loco. ymmo semper determinat principalem compositionem predicati ad subiectum"; f.
63raI5-20: "Ad ea que postea queruntur, quia inprincipalia sunt in hoc sophismate, set
magis principalia sunt in iIIo sophismate OMNIS HOMO DE NECESSITATE EST
ANIMAL, ideo breuiter pertranseundum est de hiis. Et sciendum quod est controuersia
de hoc inter quosdam. Dicunt quidam quod, <nullo> homine existente, hec sunt uere
'homo est homo', 'homo est animal'. Ad quorum controversiam uidendam notandum
quod duplex est esse uniuersale, scilicet esse actuale et essentiale."
19Ibid., f. 63ra20-28: "Et appellatur esse actuale esse in supposito, appellatur esse
essentiale ipsius uniuersalis in comparatione ad suam diffinitionem. Dicitur
communiter quod quantum ad esse actuale omnes sunt false, quantum ad esse essentiale
omnes sunt vere. Dicunt quod omni supposito destructo adhuc remanet habitudo eorum
ad suam diffinitionem uel habitudo unius ad alterum. Ex hoc est quod dicitur
 'OMNIS PHOENIX EST': QUANTIFICATION AND EXISTENCE 191

On the contrary, the author of sophism 6 is a rabid "essentialist",

asserting as he does that a proposition like 'Socrates is a man' is a per se-
proposition and, even more strongly, that it is a case of per identitatem-
predication, since the individual doesn't add anything but matter to the
universal nature. 20 It turns out, then, that such a proposition is true and
necessary even when Socrates is already dead and, a fortiori, that the same
holds for standard per se-propositions such as 'homo est animal' .21
Hence, the master discards as nonsensical the question whether the terms
are robbed of their signification when their referents cease to exist. 22
1.3 The origin of the collection
It turned out to be impossible for me to find any clue pointing to any
hypothetical attribution of any of the sophisms. On the other hand,
something more can be adduced to establish the geographical and
chronological origin of the collection. A first clue is given by the reference
to our collection in the 1347 catalogue as sophysmata de villa Parysiensi.23
This indication receives confirmation by the fact that most of the sophisms
also occur (albeit in a different form) in the undoubtedly Parisian
collections reported in MS Paris, Bibl. Nat. lat. 16135.24

communiter quod de hiis omnibus 'homo est homo', 'homo est animal' possumus
loqui dupliciter: aut quantum ad esse actuale, et sic sunt false, aut quantum ad esse
essentiale, et sic sunt uere. Set reuera neutrum istorum modorum dici\ur bene quod
essentia alicuius uniuersalis non sit nisi eo existente cuius est essentia. Uniuersale
enim non <est> nisi fuerint eius singularia, uel ad minus aliquod singulare, ut iam
pate bit. Ergo manifestum est quod esse essentie uniuersalis non erit, singularibus
destructis eius omnibus."
20lbid., f. 50rb25-31: "Item, nullo homine existente, equali ueritate erit hec uera 'Sorles
est homo' et 'homo est homo', quia nomen indiuidui et nomen speciei ab eadem forma
imponuntur, nec differunt aliquo modo nisi sicut signatum et non signatum. Et nomen
indiuidui nihil reale addit supra nomen speciei, set ponit modum intelligendi et
discretionem puram. Et ideo dixit Bootius quod species est tota essentia indiuidui. Sicut
ergo hec 'homo est homo' est uera ita quod nulla uerior est ilia, quia idem de se
predicatur. ita et hec 'Sortes est homo', quia idem de se predicatur." Cf. also ibid., f.
50va39-43: "QUOd cum hoc sit inconueniens, relinquitur quod aliquam formam
essentialem non addit singulare supra uniuersale, set solum signatum materie uel
discretionem forme."
21/bid•• f. 50vaI4-18: "Ex hiis manifestum est quid intendo de ueritate et improbatione.
Dico enim quod uera est, nec est ibi aliquod sophisma, et etiam necessaria et qui negat
eam negat omne quod demonstratum est in aliqua scientia, ut uisum est." Accordingly,
the distinction between esse actuale and esse essentie holds good for the master; see
ibid. f. 50vb26-27: "Dicendum quod sicut duplex est esse, scilicet esse actuale et esse
habitudine siue essentie, ita duplex est ens correspondens istis duobus esse ..."
22lbid., f. 5 Iva36-4 I: "Ad illud quod queritur utrum rebus corruptis termini cadant a suis
significatis ego nescio <unde> ista questio uenit, nec uidetur mihi <de>terminabilis.
Immo dico simpliciter quod non cadunt a suis significatis, quia sic numquam
significaremus nisi quod actu est. et sic 'chimera' numquam posset significare. Et
constat quod 'chimera' ita complete significat sicut 'homo', et 'tragelaphus' taliter."
23See above, n. 10.
24See Appendix. The close link between sophisms 2, 12 and 14 of the Munich
collection and their Parisian counterparts was kindly pointed out to me by Sten
Ebbesen. Cursory examination of the microfilm of the Parisian MS revealed that all
192 ANDREA TABARRONI

The link with the Parisian collections studied by Alain de Libera points
to the decade preceding 1270 as the probable period of composition of our
collection. 25 This is confirmed by the fact that the multum famosus
magister de Siccauilla is mentioned in one of the common sophisms,
namely, in sophism 'TANTUM UNUM EST' which is item 14 of our
collection.26 John de Seccheville is known to have been rector in Paris in
1256 and to have acted as regent master there again in 1263.27 So the
period between 1255 and 1270 comes under consideration; but I would
rather point to the last years in this period on the evidence of two facts.
The first one is the renown explicitly attributed to Seccheville in the
sophism, a fitting characterization if referred to his second period of
Parisian teaching, when the master's reputation was certainly increased by
his ties with the English crown. The second fact is that in the solution the
author of the sophism puts forth a distinction between essential and
accidental unity in terms which are very close to Seccheville's De principiis
naturae, which is currently dated around 1265.28
Hence, summing up, I would point to the Parisian Faculty of Arts as
the place where the disputations were held and I would propose 1265 as
the approximate date for the composition of the collection contained in clm
14522. 29
2. The sophism 'OMNIS FENIX EST'
2.1 The structure of the sophism

So much for the context. Let us now tum our attention to the fifth
sophism in the collection, dealing with the sentence 'OMNIS FENIX
EST'. Its structure is rather complex: the sophism is divided into four

but four of the Munich sophisms are strictly related to items of the Parisian collection.
A more thorough inquiry is needed in order to establish the exact relationship linking
the two collections. On the origin of the sophisms gathered in MS Paris, Bib!. Nat.
lat. 16135 see A. de Libera, "Les Appellationes de Jean Ie Page", Archives d' histoire
doctrinale et litteraire du Moyen Age 51,1984, pp. 193-225 and de Libera, "La
litterature des Sophismata."
25See de Libera, "La litterature des Sophismata", p. 217.
26Clm 14522, f. 42rb13-17 (P = MS Paris, B.N. lat. 16135, f. 40rb-va): "Set (+quid P)
de ueritate uarie fuerunt opiniones, maxime maximorum. Nam (om. P) quidam (+enim
P) uolunt soluere per (+duas P) predictas distinctiones. Quidam enim multum famosus
magister de sicca uilla (Quidam ... uilla] Quidam autem maximus et famosior tempore
nostro uidelicet magister de arida patria P) dixit quod erat [per] falsa per se, uera autem
per accidens; et ad hoc induxit (inducebat P) quoddam simile per oppositum."
270n Seccheville's life see R.-M. Giguere, Jean de Secheville, De principiis naturae.
Montreal - Paris: Institut d'Etudes Medievales - Libr. Philos, J. Vrin 1956, pp. 9-12;
A.B. Emden, A Biographical Register of the University of Oxford to A.D. 1500,
Oxford: Clarendon Press 1957-59, vo!. III, pp. 1661-2; Lewry, "Oxford Logic", p. 34.
28Cf. clm 14522, ff. 41va-42ra with Giguere, Jean de Secheville, pp. 92-7. For the date
of this work see ibid., pp. 16-17.
29This date would fit well also with the presence of the exegetical works associated with
Adam of Whitby in part 3 of the MS, see above, n. 9. My proposal has an obvious
import also concerning the date of the first Parisian collection, which gathers at least
ten of the Munich sophisms (see Appendix).
 'OMNIS RHOENlX EST': QUANfIFICATION AND EXISTENCE 193

problemata (or quesita) and each problem is in tum subdivided into at least
three questions. Further, question 2.4 de veritate has four nested
subquestions.3o

The traditional settings of the sophism (its probatio and improbatio)

are omitted by the author as sufficiently known to his audience. They can
be supplied by considering parallel treatments of the same sophism, as for
instance the one contained in the collection doubtfully attributed to Robert
Kilwardby, which is reported in MS Erfurt, Wiss. Bibl., Ampl. 40 328.
Here 'Every phoenix is' - the "first proposition", as it is usually called in
the sophismatic literature - is proved by the Law of Non-Contradiction,
since its contradictory 'Some phoenix is not' is false. 31 Obviously, the
proof depends on the assumption that the subject term in a present-tense
proposition refers only to its actually existing bearers. On the other hand,
the first proposition is disproved by Modus Tollens, since from it the false
proposition 'Many phoenixes are' can be inferred.32 Here, as one can see,
the assumption is that a universally quantified term refers to a plurality of
bearers.

Hence, the distinction by which the sophism is traditionally solved

points to the fact that if 'phoenix' is taken to refer only to actually existing
phoenixes, then the first proposition is true, while if it is taken to range
over all possible phoenixes, then the first proposition is false.3 3
Accordingly, the questions raised in the sophism focus on two rules of the
theory of restriction. 34 The first rule says that a general term having less
than three appellata refers to all its possible bearers in a universal sentence

30The four main problems bear the following titles: de distinctione, de veritate, de
probatione and de improbatione.
31MS Erfurt, Wiss. Bib\., Amp\. 4° 328, f. 19raI7-23: "OMNIS FEN IX EST.
Probatio. Cuius contradictoria est falsa, ipsa est uera; et uidetur ibi esse locus a
contradictorie oppositis. Quod autem eius contradictoria sit falsa patet; hec enim est sua
contradictoria 'aliqua fenix non est'. Set quod ista sit falsa patet, quia iste terminus
'fenix' supponit uerbo de presenti, ergo supponit pro presentibus. Et quia aliqua fenix
est presens, ideo ista est falsa 'aliqua fenix non est'. Et ideo eius contradictoria est
uera"
32Ibid., f. 19ra23-25: "Improbatur sic. Omnis fenix est, ergo plures fenices sunt; et
uidetur ibi esse locus ab inferiori ad superius."
33/bid., f. 19ra28-42: "Ad hoc sophisma communiter respondetur quod prima est duplex
ex eo quod iste terminus 'fenix', cum sit terminus communis supponens uerbo de
presenti non habens sufficientiam appellatorum, cum multiplicetur a signo uniuersali,
potest supponere uel teneri pro suppositis existentibus actu siue pro fenice que est actu.
Et tunc est uera ... Si autem supponat pro fenicibus existentibus in potentia, sic prima
est falsa. Tunc enim, cum iste terminus 'fenix' supponat uerbo de presenti, sequeretur
iam quod fenix non existens esset, quod falsum est. Et ideo tunc propositio falsa."
34Generally on the theory of restriction, see A. de Libera, "On some XIIth and XIIIth
century Doctrines of Restriction", Historiographia Linguistica 7/1-2, 1980, pp. 131-
43; "Supposition naturelle et appellation: aspects de la semantique parisienne au XIIIe
siecle", Histoire Epistemo{ogie Langage 3/1, 1981, pp. 63-77; and "The Oxford and
Paris Traditions in Logic" in The Cambridge History of Later Medieval Philosophy,
ed. N. Kretzmann et a\., Cambridge: Cambridge University Press 1982, pp. 174-87;
C.H. Kneepkens, "'Omnis homo resurget': A Note on the Early Restriction Theory and
12th Century Grammar", in Mediaeval Semantics and Metaphysics, ed. E.P. Bos,
Nijmegen: Ingenium 1985, pp. 93-110.
194 ANDREA TABARRONI

(this is the rule of the sufficientia appel/atorum). The second rule says that,
in a present-tense sentence, a general term refers only to its present
bearers. In the 12th century theory of restriction, as for instance, in the
Fallacie Parvipontane, the first rule was considered as an exception to the
second one. 35

Moreover, a special problem is raised by the proof, which depends on

the presupposition that the terms supposit in the same way in affirmative
and negative propositions. But if reference to the present bearers in the
affirmative proposition is caused by the verbal tense of the copula, how
can the same temporal reference be effected in the corresponding negative
sentence, where the verb is denied ? The third section of the sophism,
under the heading de probatione, deals with such a problem, but I won't
go into further details here, although this is a very traditional problem,
occurring in all the discussions of the theory of restriction from John Le
Page to Roger Bacon and Peter of Auvergne,36
2.2 The history of the sophism

The sophism 'OMNIS FENIX EST' has a venerable tradition in

medieval logic, dating back to the time of the logica modernorum.

The mythical animal reviving from its own ashes had made its
entrance into the logical bestiary of the Middle Ages under the patronage of
Boethius. 37 In his commentary on Porphyry he had pointed to the phoenix
- together with the sun, the moon and the world - as an example of
species having only one individual. He had also added that the term
'phoenix' could be predicated of many individuals secundum potentiam. 38
Hence, discussion of the status of such a term had become a standard item

35L.M. de Rijk, Logica Modernorum I: On the Twelfth Century Theories of Fallacies,

Assen: Van Gorcum 1962, p. 563 10- 32.
36See de Libera. "Les Appellationes", pp. 228-35, nn. 10-27 (cf. also pp. 203-4); L.M.
de Rijk, Peter of Spain. Tractatus, called afterwards Summule logicales, Assen: Van
Gorcum 1972, tr. XI. 15, pp. 205-7; John Ie Page (?), Scriptum super Peryarmenias,
MS Padova, Bib!. Univ. 1589, f. 75va-vb; A. de Libera, "Le traite De appellatione de
Lambert de Lagny (Lambert d' Auxerre)", Archives d' histoire doctrinale et litteraire du
Moyell Age 48, 1981, pp. 227-85, esp. 276-9; A. de Libera, "Les Summulae
dialectices de Roger Bacon. I-II. De termino, De enunciatione", Archives d' histoire
doctrillale et litteraire du Moyell Age 53,1986, pp. 139-289, esp. pp. 284-6, nn. 591-
606; Anonymous, Sophisma 'OMNIS FENIX EST', MS Paris, Bib!. Nat. lat. 16618,
f. 146va-vb; Anonymous, Sophisma 'OMNIS FENIX EST', MS Paris, Bib!. Nat. lat.
16135, f. 15vb-16ra; Anonymous, Sophisma 'OMNIS FENIX EST', MS Paris, Bib!.
Nat. lat. 16135, ff. 63vb-64rb and 67rb-va; Anonymous, Sophisma 'OMNIS FENIX
EST', MS Miinchen, Bayer. Staatsbib!., clm 14522, f. 47vb-48ra; Anonymous,
Sophisteria Toletana, MS Erfurt, Wiss. Bib!., Amp!. 4°276, f. 23va-vb; Peter of
Auvergne, Sophisma 'OMNIS FENIX EST', MS Firenze, Bib!. Med. Laur., S. Croce
Plut. 12 sin. 3, f. 68va-vb.
37The legacy of classical mythology concerning the phoenix is conveniently summed
up by A. Rusch, article "Phoinix 5", in Pauly-Wissowa, Realencyciopiidie der
ciassischell Altertumswissenschaft, vol. XX.39, Stuttgart: J.B. Metzler 1941, col. 414-
23.
38S. Brandt, A.M.S. Boethii In Isagogen Porphyrii Editio secunda, Vindobonae -
Lipsiae: F. Tempsky - G. Freytag 1906 (CSEL 48), pp. 214 19 _215 8 and 219 12- 17 .
 'OMNIS PHOENIX EST': QUANfIFICATION AND EXISTENCE 195

in medieval Porphyrius-commentaries, as for instance in Abailard's Logica

"Nostrorum" }9

Starting from the second half of the 12th century, however, we find
two opposed views concerning the occurrence of the term 'phoenix' in a
universal proposition such as 'Every phoenix is'. Some masters say, as in
the Ars Meliduna, in the Compendium logicae Porretanum and in the
Tractatus Anagnini, that such propositions are grammatically incorrect,
since the sign 'omnis' is to be added only to terms referring to an actual
plurality.40 Others accept such propositions, while affIrming that 'Every
phoenix is' is false by virtue of the rule of the sufficientia appellatorum. It
is the case for instance of the Fallacie Parvipontane and of the Tractatus de
univocatione monacensis.41

In the 13th century we find Roger Bacon still adhering to the old view
of the incorrectness of 'Every phoenix is', while William of Sherwood and
the Magister Abstractionum consider the sentence to be false. 42 But
generally, at least in France, both views were abandoned in favour of the
position according to which the sentence is both correct and true. This is
due to the fact that the sufficientia appellatorum was no longer required as
a precondition for true universal quantification. For the rules of restriction
to apply the universally quantified term had only to be potentially
predicable of many. Hence, in 'Omnis fenix est' the present tense copula
causes the subject term to refer only to its unique present instantiation, thus
verifying the proposition. This solution of the sophism is shared by
several logicians, such as John Le Page, Lambert of Lagny, Master
Matthew, who is the author of the Distinctiones "Quoniam ignoratis
communi bus", the anonymous author of the Tractatus florianus de
solutionibus sophismatum edited by De Rijk, the anonymous author of the
so called Sophisteria Toletana and the anonymous authors of the
'phoenix'-sophismata reported in the following MSS: Vat. lat. 7678,

39B. Geyer, Peter Abailards philosophische Schriften. Die Logica "Nostrorum petitioni
sociorum" , 2. durchges. u. veranderte Aufl. Munster i.W.: Aschendorff 1973 (Beitriige
zur Geschichte der Philosophie und Theologie des Mittelalters XXXI,4), p. 54413 ff.
400e Rijk, Logica Modernorum II.1, p. 320 (Ars Meliduna) and 11.2, pp. 261, 299 and
301 (Tractatus Anagnini) and S. Ebbesen, K.M. Fredborg, L.O. Nielsen,
"Compendium Logicae Porretanum ex codice oxoniensi Collegii Corporis Christi 250:
A Manual of Porretan Doctrine by a Pupil of Gilbert's", Cahiers de l'lnstitut du
Moyen·Age Grec et Latin 46, 1983, pp. 18 and 66. On the theory of grammaticality
sketched in these texts, see S. Ebbesen, "The Present King of France Wears
Hypothetical Shoes With Categorical Laces. Twelfth-Century Writers on Well-
Formedness", Medioevo 7, 1981, pp. 91-113, esp. pp. 98-104 on the use of 'omnis'.
41 De Rijk, Logica Modernorum I, p. 563 16-24 (FaUacie Parvipontane) and Logica
Modernorum 11.2, p. 3392. 8 (Tractatus de univocatione monacensis).
42Roger Bacon, Summa de sophismatibus et distinctionibus, ed. R. Steele, in Opera
hactenus inedita Rogeri Baconi, XIV, Oxford: Clarendon Press 1937, pp. 143-49 (and
see below n. 55 for the Summulae dialectices); C. Lohr, P. Kunze and B. MussIer,
"William of Sherwood, Introductiones in logicam. Critical Text", Traditio 39, 1983,
pp. 219-99, esp. pp. 272-3, n. 5.3.4 (the author deals with the sophism 'OMNIS
HOMO EST', but his solution holds good for the case of the phoenix too). For the
Magister Abstractionum, see P. Streveler's contribution in the present volume.
196 ANDREA TABARRONI

Paris, Bib\. Nat. lat. 16618 and Paris, Bib\. Nat.lat. 16135 first collection
and second collection. 43

Accordingly, by the middle of the century the sophism 'OMNIS

FENIX EST' ceased to be mainly devoted to the study of an exception to
the normal functioning of the distributive sign. Instead it became - together
with sophisms like 'OMNE COLORATUM EST', 'OMNIS HOMO EST'
and' ALBUM FUIT DISPUTATURUM' - a canonical framework for the
discussion of the theory of restriction. The sophism under consideration,
nr. 5 of the Munich collection, belongs to this stage of the development. It
deals only cursorily with the problem of 'omnis', and instead focuses
mainly on the function of the verb in relation to the reference of terms.

Let us now turn, as a conclusion, to a brief survey of the different

positions concerning the theory of restriction as they are found expressed
in some "phoenix"-sophismata from the second half of the 13th century.
2.3 The theory of restriction

The doctrine presented by the author of our sophism is in line with the
traditional Parisian theory of restriction described by Alain de Libera on the
basis of the treatises De appellatione by John Le Page and Lambert of
Lagny.44

According to this doctrine, restriction of the subject term in a tensed

proposition is effected by the combined action of the res verbi, i.e the
signification of the verb, and of the tempus verbi, its consignification. In a
present tense proposition the copulative function of the verb is determined
by the verbal tense; then, it narrows accordingly the denotational range of
the subject. This mechanism was commonly expressed by the dictum
'compositio cohartata cohartat extrema', which is present with some
variation also in our sophism.45 The res verbi was thought to be relevant

43De Libera, "Les Appellationes", pp. 238-41, nn. 39-46; de Libera, "Le traite De
appellatione de Lambert de Lagny", pp. 270-6 (Le Page and Lambert deal with 'OMNIS
HOMO EST'); Magister Matheus, Distinctiones "Quoniam ignoratis communibus" ,
MS Barcelona, Arch. de la Corona de Aragon, Ripoll 109, f. 306rb-307vb; Some
Earlier Parisian Tracts on Distinctiones sophismatum, ed. L.M. de Rijk, Nijmegen:
Ingenium 1988, p. 75 (Tract.flor. de solut. soph.); Anonymous, Sophisma 'OMNIS
FENIX EST', MS Vat. lat. 7678, f. 2ra-4va; Anonymous, Sophisteria Toletana, MS
Erfurt, Wissenschaftliche Bibliothek der Stadt, Amp!. 4°276, f. 23rb-24rb;
Anonymous, Sophisma 'OMNIS FENIX EST', MS Paris Bib!. Nat. lat. 16618, f.
145vb-146vb; Anonymous, Sophisma 'OMNIS FENIX EST', MS Paris, Bib!. Nat.
lat. 16135, f. 14vb-16rb; Anonymous, Sophisma 'OMNIS FENIX EST', MS Paris,
Bibl. Nat. lat. 16135, ff.62vb-67vb.
44See de Libera's works cited above n. 34 and also "Le traite De appellatione de Lambert
de Lagny", pp. 241-9.
45Clm 14522, f. 47va21-24: "Aliter etiam respondetur quod omnis coartatio fit per
immediatum; set cum dicitur uIterius quod 'predicatum non est immediatum cum
subiecto", <tempus> non ipsum immediate restringit, set primo restringit
compositionem, compositio uero immediate respiciens extrema restringit ea," The
dictum occurs also e.g. in the anonymous Sophisteria Toletana (MS Erfurt, Wiss.
Bibl., Ampl. 4° 276, f. 24ra-rb); it is criticized in the second "phoenix"-sophisma of
the Parisian collection (MS Paris, Bibl. Nat. lat. 16135, ff. 63ra and 66vb),
 'OMNIS PHOENIX EST : QUANTIFICATION AND EXISTENCE 197

in distinguishing between restrictive and ampliative verbs. Ifthe meaning

of the verb implies the actual existence of the subject, then the verb has a
restrictive function; otherwise, as with modal or cognitive predicates, the
verb is ampliative, that is to say it leaves unchanged the denotational range
of the subject term.46

Hence, according to the common opinion in Paris, restriction is a

function of the verbal meaning as affected by the verbal tense (res verbi
sub tempore considerata). But in our sophism, a third cause of restriction
is considered, namely, the human mind (intellectus) in its role of
interpreting single speech-acts. In a typical present tense proposition. the
author says, the temporal reference points to a generic present (presens
simpliciter or presens conjusum), but when confronted with the
corresponding token proposition the human intellect takes the temporal
determination as referring to the actual present. i.e. to the time in which the
proposition is uttered (presens ut nunc).47

Moreover, the human intellect acts in assuming restrictive verbs as

implying the existence of their subjects, while taking ampliative verbs as
free from this presupposition. In the latter function. however. according to
our anonymous author, the intellect is not properly acting as a cause of
restriction, but rather as a determinant of the acceptio of the subject term.48

Here, if only in passing. the new notion of acceptio turns up, which
will play a pivotal role in modistic semantics, as pointed out in many
works by Jan Pinborg.49 This reminds us of the sad fate experienced by
the theory of restriction in Paris in the following decades. Restriction as a
property of terms still found eager supporters. such as the author of the

460n this point see de Libera, "La litterature des Sophismata", p. 220; the formulations
of our sophism are reported below, n. 49.
47Clm 14522, f. 47va28-35: "Nmil cum presens ut nunc non accidat uerbo secundum
Petrum Heliam, set presens simpliciter; uerbum autem restringit ratione temporis quod
sibi accidit; cum presens simpliciter non sit presens existens ut nunc, patet quod
uerbum ratione temporis non restringit ad presentes qui sunt, set ad presentes
simpliciter. Unde non sequitur uirtute temporis 'homo currit, ergo homo qui est currit',
set sic 'homo currit in presenti', non hoc uel illo; tamen ratione industrie et intellectus
apprehenditur res in presenti pro presenti ut nunc, et non pro presenti confuso."
48/bid., f. 47va-vb: "Quedam est res uerbi ad cuius esse in presenti sequitur suum
subiectum esse in presenti; ideo intellectus rem suppositam respectu talis presentis
accipere potest solum pro existenti. Sunt autem alia uerba que significant res ad quarum
esse in presenti non requiritur de necessitate subiectum esse in presenti, set indifferenter
potest esse tam ens quam non ens; et ideo tales res supponentes talibus uerbis non
determinant acceptionem pro presenti. Talia autem sunt 'potest', 'Iaudatur', 'opinatur'.
Ad esse enim potentie in presenti non requiritur subiectum esse in presenti, et ideo
'potest' non restringit. Ad esse autem cursus, lectionis, disputationis in presenti
requiritur subiectum esse in presenti, et ideo talia uerba restringunt. Per hoc patet causa
coartationis in uerbo, quoniam non solum gratia temporis, set gratia rei. Et non solum
f!tia rei, set quia adhuc operatur intellectus acceptiones sic determinans."
4 See J. Pinborg, "Die Logik der Modistae", Studia Mediewistyczne 15, 1974, pp. 39-
97, esp. 69-70 and "Some Problems of Semantic Representations in Medieval Logic",
in History of Linguistic Thought and Contemporary Linguistics, ed. H. Parret, Berlin:
De Gruyter 1976, pp. 255-78, esp. p. 263.
198 ANDREA TABARRONI

second "phoenix"-sophisma reported in MS Paris, Bibl. Nat. lat. 16135. 50

As a matter of fact, the latter text is, to my knowledge, the most ambitious
attempt to give the theory of restriction a modistic foundation. The author
says that the convenientia respectiva between the terms in a proposition is
the general cause of restriction. Such a "mutual concord" is then defined in
terms of the basic semantic components of words according to modis tic
semantics: the significatum speciale and the modus significandi essentialis
and accidentalis. 51 The author deals with all the possible combinations,
trying to determine what happens when two components agree against the
third one. The result is a wide-ranging but rather intricate mechanism
purporting to cover all the ways in which the predicate can modify the
subject's reference in a proposition.

But, probably as a reaction to such attempts, the theory of restriction

eventually came under attack in Paris in the 1270s. One of the leading
opponents of the traditional doctrine was Peter of Auvergne, who in his
sophism 'OMNlS FENIX EST' tries to demonstrate that one of the terms
of the proposition cannot modify the other term's way of referring. 52
Hence, according to him, a term in a proposition always supposits
indifferently for all its bearers, be they presently existent or not. The
proposition 'Every phoenix is' is then false, according to Peter, as are
false de virtute sermonis all contingent present tense propositions like
'Every man is running'.53 Peter's view became common in Paris in the

50See above, n. 43. De Libera is currently preparing an edition of this sophism, see
Cahiers de l'Jnstitut du Moyen-Age Grec et Latill59, 1989, p. 115.
51MS Paris. Bib!. Nat. lat. 16135, f. 65va37-39: "Dicendum est igitur quod aptitudo
conuenientie termini causatur ex eis que sunt in termino, scilicet res significata
specialis et modus significandi essentialis et modus significandi accidentalis."
52MS Firenze. Bib!. Med. Laur., S. Croce Plu!. 12 sin. 3, f. 68va: "Si igitur terminus
de se non habet quod supponat pro presentibus nec habeat a predicato, uidetur quod
nullo modo restringatur. Causa autem quare a predicato non restringitur uidetur esse
ista, quoniam omne restringens aliud dicitur esse ut unum extremum (extraneum MS)
cum eo et etiam ut unum intellectum aliquo modo cum eo quod restringit. Ea enim que
significantur ut unum extremum (extraneum MS) aliquod modo intelliguntur ut unum.
Nunc autem predicatum et subiectum, quamquam significentur esse unum, non tamen
ut unum extremum (extraneum MS) nec tamen uno intellectu concipiuntur, set nec
diuersum. Et ita unum non determinat aliud uel restringit ita ut faciat ipsum esse
minus quam esset illud."
531bid., f. 68va: "Et ideo dicendum quod terminus supponens uerbo cuicumque
cuiuslibet temporis supponit suum significatum et per indifferentiam ad omnia
supposita ad que significatum se habet per indifferentiam, siue fuerint presentia uel
preterita uel futura. Hoc enim (etiam MS) accidit significato et suppositis termini.
Sicut enim significatum alicuius termini non includit aliquod tempus, ita nee
suppositum quod dicitur suppositum quia participat significatum per se. Magis tamen
est uerum determinate quod terminus aliquis uerbo de presenti supponens, quamquam
supponat omnia supposita et per indifferentiam presentia <preterita> et futura, ei tamen
denotat inesse predicatum in presenti, ut dicendo 'omnis homo currit' omni homini
siue presenti siue preterito siue futuro actribuitur currere in presenti et non in alio
tempore; sicut dicendo 'omnis homo est albus' denotatur omnem hominem esse album
et non nigrum."
 'OMNIS PHOENIX EST': QUANTIFICATION AND EXISTENCE 199

last quarter of the century, and it was held by Simon of Faversham and
Radulphus Brito.54

In England, in the second half of the 13th century, the theory of

restriction apparently never achieved wide acceptance among logicians. It
had already been attacked by Roger Bacon in 1250 in his Summulae
dialectices, although on opposite grounds from those of Peter of
Auvergne, that is, on the ground that a term cannot refer univocally to
past, present and future things. 55 In the "phoenix"-sophisma doubtfully
attributed to Robert Kilwardby and studied by Braakhuis the two opposed
paradigms - the Baconian and the modistic - are confronted on the
question of the univocal signification of being and non being.56 The author
does not finally choose between the two positions, but he concludes that
on both accounts the theory of restriction is to be dismissed. Even Walter
Burleigh, at the beginning of the new century, sided with Bacon in
declaring that terms refer to their present bearers by imposition and so
there is no need of restriction.57 The latter appears to be the common
position at Oxford at the time of Burleigh and the sentence 'Every phoenix
is' is then considered to be true.58

It is worth noticing, however, that in his sophism 'OMNIS PENIX

EST' - which curiously enough occurs in his questions on Aristotle's
Perihermeneias - Walter Burleigh devotes a long discussion to the thorny
problem of the distinction between being and essence, thus bringing to
light the metaphysical background of the semantic issue at hand.59

Thus, following the traces of the phoenix, the ever reviving animal,
one can revive the different stages of an everlasting problem in 13th
century (as well as in today's) semantics, the problem of the relations

54See Pinborg, "Die Logik der Modistae", p. 69 n. 105 and "Some Problems of
Semantic Representations", pp. 272-4.
55De Libera, "Les Summulae dialectices de Roger Bacon", p. 278, nn. 538-41.
56See Braakhuis, "Kilwardby versus Bacon ?". The following is the author's solution de
veritate: "Idcirco dicendum est aliter quod si terminus sit nomen entis solum, ita quod
non sit nomen entis nisi equiuoce, tunc prima propositio uera est simpliciter ... Si
aUlem terminus sit indifferenter nomen entis et non entis, tunc cum terminus non
possit cohartari a tempore consignificato per uerbum nec ratione rei uerbi, ideo
supponit tam pro ente quam pro non ente; et ideo quia pro non ente falsa est, ideo
dicitur esse falsa, et sic falsa est secundum istam positionem." (MS Erfurt, Wiss.
Bib\., Amp\. 4° 328, f. 26ra-rb)
57S. Brown. "Walter Burley's Quaestiones in librum Perihermeneias". Franciscan
Studies 34, 1974, pp. 200-95. esp. 278-95.
58See for instance Ockham, Summa Logicae 11.4. Opera Philosophica I. ed. P. Boehner,
G. Gal and S. Brown. St. Bonaventure N.Y.: The Franciscan Institute 1974. pp. 260-2.
59 See Brown. "Walter Burley's Quaestiones". pp. 260-78; it is apparent from p. 260,
n. 4.01 that both questions an esse existere sit de essentia rei causatae and de veritate
huius 'omnis phoenix est' are sections of the same sophism 'OMNIS PHOENIX
EST'. On the metaphysical presuppositions oCthe theory oCrestriction see J. Pinborg.
"Bezeichnung in der Logik des XIII. Jahrhunderts", Miscellanea Mediaevalia 8.1971,
pp. 238-81, esp. 249. An analysis of the influence of the Avicennian doctrine of
predication - along the lines recently described by A. Blick, "Avicenna on Existence".
Journal of the History of Philosophy 25/3, 1987, pp. 351-67 - would be of particular
importance in this connection.
200 ANDREA TABARRONI

between time and reference in natural language. The sophism I have

presented belongs to a stage of transition in this development. My aim was
to raise some interest in the collection in which it occurs.

University of Bologna, Italy

Appendix

A List of Sophisms contained in MS elm 14522

I give here only the sophisms' "headings", leaving for another

occasion a more complete description of the collection, with extensive
incipit, explicit and titles of the questions dealt with in each sophism. The
reference to the Parisian MS (when it occurs) is not intended as an
indication of another copy of the sophism in question, but only of a text
strictly related to it. Closer analysis is needed in order to determine the
exact relationship between the two collections.

1. TOruS SORTES EST MINOR SORTE, ff. lra-lOvb

(cf. MS Paris, B.N. lat. 16135, ff. 7rb-9rb)

2. SI TANTUM PATER EST, NON TANTUM PATER EST, ff. lOvb-

18vb
(cf. MS Paris, B.N. lat. 16135, ff. 25vb-28rb)

3. DECEM PRETER QUINQUE SUNT QUINQUE, ff. 18vb-20vb +

29ra-31rb
(cf. MS Paris, B.N. lat. 16135, ff. 34rb-35va)

4. DEUS SCIT QUIDQUID SCIVIT, f. 31rb (fragm.)

5. OMNIS FENIX EST, ff. 44ra-48rb

6. OMNIS HOMO DE NECESSITATE EST ANIMAL, ff.48rb-51vb

7. TANTUM VERUM OPPONITURFALSO, ff.51vb-55ra

8. SI NULLUM TEMPUS EST, ALIQUOD TEMPUS EST, ff. 55ra-58ra

(cf. MS Paris, B.N. lat. 16135, ff. 20vb-21vb)

9. SI SORTES NECESSARIO EST MORTALIS, SORTES

NECESSARIO EST IMMORTALlS, ff. 58ra-63va
(cf. MS Paris, B.N. lat. 16135, ff. 9rb-llrb)

10. OMNIS PROPOSITIO VEL EIUS CONTRADICTORIA EST VERA,

ff. 63va-69va
(cf. MS Paris, B.N. lat. 16135, ff. 28rb-29va)
 'OMNIS PHOENIX EST': APPENDIX 201

11. OMNIS HOMO EST ANIMAL ET ECONVERSO, ff. 69va-71 vb +

21ra-23vb
(cf. MS Paris, B.N. lat. 16135, ff. 17vb-20ra)

12. OMNIS HOMO EST UNUS SOLUS HOMO, ff. 23vb-27ra

(cf. MS Paris, B.N.lat. 16135, ff. 16rb-17rb)

13. QUANTO AUQUID MAlUS EST, TANTO MINUS VIDETUR, ff.

27ra-28vb + 37ra-va
(cf. MS Paris, B.N. lat. 16135, ff. 36rb-37rb)

14. TANTUM UNUM EST, ff. 37va-43rb (mutil.)

(cf. MS Paris; B.N. lat. 16135, ff. 38ra-42rb)
 Expositio as a method of solving sophisms
by Mikko Yrjonsuuri

1. Introduction

Saying anything general about sophisms is difficult for at least two

reasons. First, texts classifiable into the genre of sophisms are so diverse
that there seems to be almost nothing common to all of them. Secondly,
modem discussion of these texts is still somewhat scattered. The texts
simply are not yet known well enough for finding unifying perspectives.

It is beyond the scope of this paper to give a general account of what

sophisms are. Let me therefore limit my characterization to the sense I am
here interested in: sophisms can be understood as problems of deciding
whether a proposition (called 'sophisma sentence' in the following) is to
be granted or denied. Because sophisms belonged to the university
curriculum in the fourteenth century, undergraduates faced this problem in
their disputations on sophisms. In these disputations the respondent had to
decide whether the sophisma sentence has to be granted or denied: he had
to decide whether it is true or false, or - perhaps more usually - whether it
follows or does not follow from the given assumptions, called the casus.!

The other more or less technical term in my title is expositio. In

section 2, I explain this term with reference to Richard Billingham's
Speculum Puerorum, which is probably the best introduction to the theory
of exposition from the fourteenth century.2 In this paper I look at this late
medieval theory as one way of tackling the problems formulated as
sophisms. In sections 3 and 4, I tum to an application of the theory found
in the treatise Dec/aratio sophismatum Climitonis edited by Simo Knuuttila
and Aoja Inkeri Lehtinen.3 A careful look at the application will lead into
comparisons with some modem theories of language in the rest of the
paper. Section 5 introduces the concept "logical form" and section 6 the
concept "semantical game".

1For general descriptions of sophisms see e.g. N. Kretzmann, "Syncategoremata,

exponibilia, sophismata", in The Cambridge History of Later Medieval Philosophy, ed.
N. Kretzmann et aI., Cambridge: Cambridge University Press 1982, pp. 211-45; J. E.
Murdoch, "Mathematics and Sophisms in the Late Medieval Natural Philosophy and
Science", in us Genres Litteraires dans les sources TMologiques et Philosophiques
Medievales (Publications de L'institut d'Etudes Medievales, 2e serie. Textes, Etudes,
Congres, vol. 5), Louvain-Ia-neuve 1982; E. D. Sylla, "The Oxford calculators", in
The Cambridge History ... , pp. 540-63.
2L. M. de Rijk's edition, in "Some 14th Century Tracts on the Probationes
terminorum", Artistarium 3, Nijmegen: Ingenium PublisHers, 1982, contains a few
versions of Billingham's treatise together with some other connected texts and an
introduction. In the following I will refer in passage numbers to the version called
"prior recensio" by de Rijk.
3Knuuttila and Lehtinen, "Plato in infinitum remisse incipit esse albus", in Essays in
Honour ofJaakko Hintikka, ed. E. Saarinen et al., Dordrecht: Reidel 1979, pp. 309-29.

202
 EXPosmo AS A METHOD OF SOLVING SOPHISMS 203

2. De probationibus terminorum
At the beginning of the Speculum puerorum propositions are divided
into mediate and immediate ones. hnmediate propositions seem to be quite
rare: they are like 'hoc est' or 'hoc potest esse'. These propositions are
proved only by the senses and the intellect.4 Mediate propositions are
proved mediately, through other propositions. Billingham's aim in the
treatise is to give suitable rules for finding these other propositions which
are necessary for proofs. The idea is to find ways of reducing complicated
mediate propositions into some immediately evident basic from.

The term 'resolutio' was often used in the fourteenth century in a wide
sense of the process of giving some more readily understandable form for
a proposition, which is in some way complicated. In a resolution, in this
sense, some term with many connotations could, for example, be replaced
by its nominal definition. Billinghamian exposition can be seen as one way
of resolving exclusive, exceptive and comparative propositions as well as
other kinds of propositions, paradigmatically including those concerning
beginning and ceasing.

In Billingham's Speculum puerorum the term 'resolutio' is used also

in a stricter sense. Billingham presents three ways of finding a proof for
mediate propositions, appropriate to three mutually exclusive and, as it
seems, jointly exhaustive classes of mediate propositions, namely
resoluble, exponible and officiable propositions.s In the stricter sense
resolution concerns only resoluble propositions, and it amounts to descent
from a common term to a discrete one.6 For example, the particular
proposition 'homo currit' is proved by the conjunction 'hoc currit et hoc
est homo', where 'hoc' refers to some running man, who is pointed out.?

Officiable terms are modal terms or terms signifying an act of mind,

taken in the composite sense. The proof of these terms is best explained by
an example: 'contingens est te esse' is proved by 'hec est contingens "tu
es", que precise significat te esse'. In the first part the modal term is
predicated of a proposition in material supposition, and in the second part
the signification of this proposition is given. s Since I will not discuss
officiable terms in the following, I leave their description at this level.

An exponible proposition can be identified from certain terms, like

'begins', 'ceases', 'only' or 'except', to mention some examples. These

4"Ideo talis propositio 'hoc est' est propositio immediata, quia non potest probari per
aJiquod prius iIIo. sed solum probatur per sensum et intellectum. Similiter ilia est
immediata: 'hoc potest esse'." BiIIingham, op. cit., nr. 6.
SA proposition may belong to different classes, but only in different senses. Thus an
unambiguous proposition belongs to only one class. See Billingham, nr. 21-22.
6"Terminus resolubilis est quilibet terminus communis, ... qui habet aliquem <alium>
terminum inferiorem se secundum predicationem. Secundum quod resolvitur propositio
in qua ponitur quando capitur inferius eo in eius probatione; et componitur quando
capitur superius eo." Billingham, nr. 7.
?Billingham, nr. 8.
8Billingham, nr. 62.
204 MIKKO YRJONSUURI

terms as terms are also called exponible. In proof through exposition two
or more simpler propositions without the exponible term in question,
called the exponents, are found for the exponible proposition. The
conjunction of these exponents must be equivalent to the original
proposition.9 Billingham gives no other general rules for the character of
the exponents. They are determined separately for each exponible term, as
is certainly necessary, since exposition is so closely connected to the
meaning of the term.
3. Sorles est albior quam Plato incipit esse albus
Let me now turn to one example of the way exposition is used in
sophisms. In the Declaratio sophismatum Climitonis, possibly written by
Billingham himself and at least strongly influenced by Billingham,lO we
find very straightforward applications of the rules of exposition discussed
in the Speculum puerorurn. In accordance with its name, the Declaratio is
based on Richard Kilvington's Sophismata. It deals with Kilvington's
sophisms 1-14, 17 and 19-23.

The text proceeds as a disputation. The opponent puts forward

Kilvington's sophisms as problematic propositions to be evaluated by the
respondent. The respondent gives his evaluation, which consists of
denying or granting the sophisma sentence, and giving it an analysis as
reasons for the answer. The analysis amounts in most cases to exposition
according to Billingham's rules. Often the respondent's answers give
grounds for counter-arguments by the opponent, in most cases beginning
with the word 'contra' in the text. The respondent seems to be required to
solve these counter-arguments.

It is interesting to notice that the sophisms are not discussed with

proof and disproof, but rather as complex or problematic sentences based
on specific casus. If the text is taken as reflecting the way sophisms were
actually considered in disputations, the role of respondent was not so
much that of evaluating arguments pro and contra as that of solving
problems of interpretation pointed out by the opponent. In the following I
will concentrate on the material concerning Kilvington's first two
sophisms. The first sophism is put forward and answered as follows: 11
"[A] Sortes est albior quam Plato incipit esse albus [B] stante casu
sophismatis quod [Bl] Sortes sit summe albus et quod [B2] Plato
incipiat dealbari per remotionem de presenti.
[C] Tunc conceditur sophisma et exponitur sic: [Cl] Sortes est
albus et [C2] Plato incipit esse albus et [C3] Plato non incipit esse ita
albus sicut Sortes est albus. Et sic exponitur ratione primi termini
mediati, scilicet ratione ly 'albior' et prima et secunda ex casu patent,
et si negatur tertia, arguitur: da oppositum: [NC3] Plato incipit esse

9"Terminus exponibilis est qui habet duas exponentes, vel plures, cum quibus
convertitur." Billingham, op. cit, nr. 18.
lOFor the authorship, see discussion in Knuuttila and Lehtinen, op. cit., p. 310.
II Knuuttila and Lehtinen, p. 318.
 EXPOSITIO AS A MEI'HOD OF SOLVING SOPHISMS 205

ita albus sicut Sortes est albus, igitur [C4] Plato immediate post hoc
habebit summam albedinem, igitur."

I have marked out certain parts of the text. 'A' stands for the sophisma
sentence and 'B' for the casus with 'BI' and 'B2' as the two parts of it.
'C' stands for the response, 'Cl' to 'C4' for different propositions used in
argumentation for the response. The respondent grants the sophisma
sentence A and gives an exposition for it. In his exposition the respondent
seems to apply Billingham's rules from the Speculum puerorum. The
exposition follows the rule for the first mediate term, which is 'albior'.
According to Billingham, a comparative is expounded by a conjunction of
(1) its positive form, (2) the positive form of the the compared and (3)
universal denial of equality with the compared. 12 Following this rule
Declaratio gives three exponents:
"[CI] Sortes est albus

[C2] Plato incipit esse albus

[C3] Plato non incipit esse ita albus sicut Sortes est albus."

Strictly speaking the denial of the equality C3 is not universal: if the rules
and examples of Billingham are followed literally, it should read
[C3*] Nihil quod est Plato incipit esse ita albus sicut Sortes est
albus

However, since 'Plato' is a singular term, there seems to be only a

minimal difference between these two propositions.

The respondent does not finish on finding these three exponents: he

points out that CI and C2 hold by the casus (ex casu patent), and that C3
can also be argued for from the casus. The respondent seems to argue for
C3 by expounding its denial NC3, although the argumentation is mostly
implicit in the text. The denial of the third exponent of the sophisma
sentence can be found explicitly from the text:
[NC3] Plato incipit esse ita albus sicut Sortes est albus

According to Billingham's rules 13 the exposition ofNC3 according to the

term 'incipir is the conjunction of denial of the present and affIrmation of
the immediate future:
[C5] Plato non est ita albus sicut Sortes est albus
[C6] Plato immediate post hoc est ita albus sicut Sortes est albus

The latter conjunct C6 together with

12"Comparativus gradus exponitur per suum positivum et per positionem sui

comparati, vel suorum comparatorum, et per universalem abnegationem gradus
equalitatis a comparato, vel comparatis, quoad illud accidens secundum quod
comparantur." Billingham, nr. 49.
13See Billingham, nr. 56.
206 MIKKO YRJONSUURI

[B I] Sortes est summe albus

assumed in the casus implies a proposition explicitly stated in the text:

[C4] Plato immediate post hoc habebit summam albedinem,

which is repugnant with Plato's just beginning the process of getting

white, which is assumed in the casus (B2). The term 'dealbari', which is
used, includes the concept of a process taking some time before Plato is
fully white, and consequently it can be seen without additional
assumptions that Plato will not immediately have full whiteness.

This reductio ad absurdum-argument shows that NC3, or the denial of

C3 leads to inconsistencies with the casus. It is thus shown that NC3 must
be denied, and consequently C3 must be granted, which finishes the
proof, since CI and C2 had already been considered.

The argument for granting the sophisma sentence A can be

summarized as following a two-tier exposition: A is expounded as a
conjunction of three propositions CI, C2 and C3. Two of these
propositions (Cl and C2) are immediately evident from Bl and B2 in the
casus and thus do not need any further proof. The third, C3, is argued for
by reductio ad absurdum of its denial NC3. The reductio ad absurdum-
argument is based on exposition of NC3, the exponents being C5 and C6.
C6 together with B I from the casus imply C4, which is considered to be
evidently inconsistent with B2 from the casus.

From the viewpoint of propositional logic it is interesting to notice that

the method of exposition, which here turns a categorical proposition into
an equivalent conjunction, is used in two ways here. The fIrst exposition
gives a conjunction of three propositions (CI & C2 & C3) equivalent to
the proposition (A), which has to be proved. The rest of the proof consists
of showing that each of these three conjuncts has to be granted. The
exponents are treated as conjunctively implying the desired conclusion. In
the next exposition the conjunction of two propositions (C5 & C6) is
equivalent to the denial of the proposition (C3), which must be proved. In
this case one of the conjuncts is refuted. No attention is paid to the other
conjunct. Here the exponents are treated as being implied by the desired
conclusion and modus tollendo tollens is used.

As I mentioned, exposition is a kind of resolution in the wide sense.

'Resolutio' is Latin for the Greek 'analysis': it is interesting to notice that
as a method of proof exposition looks like analysis in at least two senses.
Firstly, exposition clearly is a method of working from the conclusion
sought for towards what is given in the casus. In this sense the method of
exposition conforms to the basic description of the ancient method of
geometrical analysis given by Pappus.l 4 Secondly, exposition seems to

14Pappi Alexandrini Collectionis Quae Supersunt, vols. I-III, ed. Fr. Hultsch, Berlin:
Weidemann 1876-77, see vol. II, p. 634 ff. See also J. Hintikka and U. Remes, The
Anciellt Method of Analysis, Boston Studies in the Philosophy of Science, vol. XXV,
Dordrecht: Reidel 1974.
 EXPosmo AS A MEI'HOD OF SOLVING SOPHISMS 207

amount to analyzing the exponible proposition into simple elements. In this

sense exposition conforms to the idea of analysis as decomposing a
complex structure into its elementary parts. In our example, when A is
expounded into Ct, C2 and C3, the meaning of the mediate term 'albior' is
decomposed into elements. In the second exposition, when NC3 is
expounded into C5 and C6, the meaning of the mediate term 'incipit' is in
similar way decomposed. In both cases exposition leads towards
elementary propositions easily comparable to the propositions of the casus.
The aim seems to be a reduction of the original complicated proposition
into a set of elementary propositions, whose truth-values are immediately
evident from the casus. It seems that in a full exposition (which is not
given here as it was not necessary for solving the sophism), the process
would lead to immediate propositions like 'hoc est'.

It may be noticed that the process would not lead to a large

conjunction of immediate propositions equivalent to the original
proposition. In our example the second exposition is not an exposition of
one of the exponents of A, but an exposition of NC3, which is the denial
of C3, which in tum is an exponent of A. If we were to formalize the
result of the two expositions presented above, the result would be that A is
equivalent to the conjunction of Ct, C2 and either the denial of C5 or the
denial ofC6 (A <=> Ct & C2 & (...,C5 v...,C6)).
4. Problems with infinitely small whiteness

After the arguments discussed above the text continues with two
counterarguments against the given way of analysing the sophisma
sentence A. In general the continuation of the text illuminates nicely how
the given propositional analysis is fitted into a disputational setting. The
problems solved are those of question and answer.

In the first counter-argument presented in the text the opponent points

out that in the stated reading the sophisma sentence is a comparative which
does not presuppose that both Socrates and Plato are white, as it should,
since comparatives always compare two positive grades. Plato's
whiteness is future, not present. The respondent answers this counter-
argument by simply granting that the comparative compares present
whiteness with future whiteness. The respondent does not take this as
problematic. I5

It seems that the opponent in the text is satisfied with this answer,
since he goes on to another and more interesting problem. He shows that if
the sophisma sentence A is granted, the proportion between Socrates' and
Plato's whitenesses must be either finite or infinite. If it is finite, Plato
begins to be white to a certain degree (in aliquo certo gradu), which is
false, since whiteness is taken to be a continuous quality and Plato begins

15"Contra istam expositionem arguitur sic: Sequitur, quod per comparativum posset
comparari futurum tempus ad presens et per consequens non presupponit suum
positivum in utroque extremorum, quod est contra prius dicta, quia ibi solum ponitur
Sortem esse album et non Platonem. Respondet\lr, quod comparativus in utroque
extremorum presupponit suum positivum, prout copula propositionis requirit sibi
unum de presenti et aliud de futuro." Knuuttila and Lehtinen, pp. 318-9.
208 MIKKO YRJONSUURI

to get more white from not being white at all. 16 The opponent concludes
that
[D] Sortes in infinitum est albior quam Plato incipit esse albus.

The respondent answers this argument by granting D, and giving first a

resolution, and then an exposition. This answer again follows the strategy
of evaluation with an analysis. I?

The opponent continues by putting forward (after some other

considerations) two alternatives: 18
"[E] Vel hoc est quod ista proportio est infinita ex eo, quod [EI]
Sortes infinite vel quod [E2] Plato in infinitum remisse incipit esse
albus."

The respondent chooses E2, and gives an exposition for the proposition
which in the text breaks in the middle.1 9 Intuitively the respondent makes
the right choice, but, as it turns out, the wording of E2 is mistaken
according to the position taken in this text. As Knuuttila and Lehtinen have
shown in their discussion of the text, other authors had different views on
the correct wording of the proposition. 20

After the respondent has granted and analysed E2, the discussion
breaks out again. As a fresh start E2 is again put forward, but now as an
independent sophisma sentence, which is denied. A natural way of reading
the text is that in the actual disputation behind this text the respondent was
taken to have made a mistake when granting E2, and thus the disputation
was halted by the opponent or by the master - depending on the roles of
the people participating the disputation. After it was recognized that the
answer was wrong, the disputation continued, perhaps with some other
student as respondent giving a better answer to E2. The whole analysis is
given also in the text as follows: 21
"Sophisma est falsum et resolvitur: Quantalibet albedine data
remissionem Plato incipit habere. Et exponitur sic: Aliquanta albedine
data remissionem Plato incipit habere et non aliquanta albedo est,
quin adhuc, et cetera. Et tunc primo inducitur sic: Ista albedine data
remissionem incipit habere et tunc oportet dari certa albedo, quia ibi

16"Sed contra. Si SOrles est albior, et cetera, vel hoc est finite vel infinite. Si finite,
sequitur, quod Plato inciperet esse albus in aliquo certo gradu, quia si Sortes est,
exempli gratia, albus ut quattuor, tunc sic Plato inciperet esse albus sub quadruplo,
quod est falsum, et igitur dicitur quod infinite Sorles est albior, et sic Sorles in
infinitum est albior quam Plato incipit esse albus." Knuuttila and Lehtinen, p. 319.
1?"Conceditur secundum casum predictum et resolvitur sic: QuantaIibet albedine data
ultra istam Sortes est albior quam Plato incipit esse albus. Et exponitur sic: Ultra
aliquam proportionem SOrleS est albior quam Plato incipit esse albus et nulla proportio
albedinis est, quin ultra istam Sortes est albior quam Plato incipit esse albus."
Knuuttila and Lehtinen, loco cit.
18Knuuttila and Lehtinen, loco cit.
19"Respondetur, quod ex parte secundi, scilicet remissionis. Et exponitur: Quantolibet
~radu et remissius, et cetera." Knuuttila and Lehtinen, loco cit.
2 Knuuttila and Lehtinen, pp. 311-15.
21 Knuuttila and Lehtinen, p. 319.
 EXPOSITIO AS A MEIHOD OF SOLVING SOPHISMS 209

demonstratur aliquid per ly 'istud' et per consequens Plato habet vel

habebit aliquando in infinitum remissam albedinem. Et si dicitur
quod numquam habebit, tunc numquam incipit habere."

For limitations of space, I won't go into details of this analysis.

However, it is interesting to notice that the analysis is taken to lead to
attributing some infinitely small degree of whiteness to Plato in some
future instant, which is taken to be impossible. For this reason the
sophisma sentence is denied. A similar analysis is asserted to apply also to
[F] In infmitum remissam albedinem Sortes incipit habere

which is put forward as the next sophisma sentence.22

E2 and F have in common the feature that 'infinitum' is before

'incipit' in the word order. In the next sophisma sentence this order is
changed:
[0] Sortes incipit habere in infinitum remissam albedinem.

The respondent grants 0 and states that now the analysis does not lead to
any singular proposition, where 'albedinem' stands in determinate
supposition for some whiteness of infinitely small degree.23 Unfortunately
the text does not give a full analysis, and thus we cannot draw any further
conclusions about the viewpoint at issue.

There are nevertheless some interesting ideas present in the discussion

of these three sophisma sentences E2, F and O. Let me point out two.
First, E2, F and 0 seem to form a sequence, whose purpose is to find a
correct wording for an intuitively understood idea. The proper name
'Plato' (or 'Socrates'), the exponible term 'incipit' and the problematic
compound 'in infinitum remisse albus' are combined in different ways
until an acceptable formulation is found. The word order can be seen to be
very important in analyzing sentences of this type.

Secondly it is interesting to notice that the problem in E2 and F seems

to be that the analysis leads in some sense to an immediate 'hoc est' level,
where a certain infinitely small whiteness is pointed out by a demonstrative
pronoun in an affirmative sentence. In the discussion of 0 it is pointed out
that in E2 and F the confused supposition of 'albedinern' is removed by
the analysis. The analysis leads to a sentence with a demonstrative

22Knuuttila and Lehtinen, loco cit. The answer to this sophism begins: "Sic resolvitur
ut prius et exponitur ut prius et iterum est falsum."
23The sophism is answered as follows: "Conceditur et resolvitur et exponitur ratione Iy
'infinitum' et semper Iy 'albedinem' stat confuse tantum. Sed contra. Sicut in
propositionibus precedentibus confusio istius quod est 'albedinem' est evacuata per
deductionem, quod tandem deveniebatur ad aliquod detenninatum, sic etiam hoc potest
evacuari. Respondetur, quod posset evacuari, sed tamen nullum sequitur inconveniens
quod in primis fuit inconveniens, quia in istis Iy 'albedinem' sequitur Iy 'incipit' et sic
semper stat confusio, et si evacuatur, difficultas iIIius quod est incipit. Tunc tamen ista
albedo est danda per remotionem de presenti et positionem de futuro, et sic numquam
potest tam diu descendere quod veniatur ad singularem in quo Iy 'albedinem' stat
detenninate." Knuuttila and Lehtinen, pp. 319-20.
210 MIKKO YRJONSUURI

pronoun, which has determinate supposition. The respondent granting G

admits thnt it is also possible to remove the confused supposition in G.
However, the demonstrative pronoun suppositing determinately for the
whiteness of infinitely small degree is within the scope of 'incipit', which
is also an exponible term. Thus the analysis leads only to negative singular
statements about the whiteness of infinitely small degree. 'Incipit' is
analyzed by denial of the present and affirmation of the immediate future,
and after this instant there is no first instant which would have the
whiteness of infinitely small degree.24

The respondent's solution seems to amount to admitting that the

degree of the incipient whiteness is indeed infinitely small but that there is
no instant at which Sortes would have that degree of whiteness. Thus
nothing infinite is accepted as existing at any time.
5. Exponents and logical form
In the rest of this paper I will try to show how the technique of
exposition used in these sophisms can be seen as a method of
interpretation. At first sight this might seem to be simply against what the
texts say: in the Speculum puerorum Billingham calls the method of
exposition a method of proof, and in modem terminology 'proof' is not a
semantical term. However, Billingham's statement that immediate
propositions are proved through the senses and the intellect shows that he
did not use the term exactly in our technical logical sense. Sensory
experience cannot be put into inferences in the way sentences can. 'Proof'
does not have such a strict meaning in Billingham's treatise. And
furthermore, Billingham's treatise is not very useful in determining such
metatheoretical questions. Careful reading of Billingham's treatise makes it
clear that the treatise is a manual: it tells quite well how the system works,
but has very few comments on the purpose of the system.

I think that the best way to look at the question of the purpose of the
theory of exposition is to look at its practice: that is, to take a disputational
viewpoint. In a disputation on sophisms the respondent is faced with the
problem of evaluating a problematic sentence. As the facts of the matter are
known from the casus, the problem concerns interpretation, as usually
becomes clear to the modem reader at the first reading. Most sophisma
sentences dealing with exposition are rather weird. I will later return to the
working of the disputation; in the next few pages I will look at how
exposition illuminates the meaning of a proposition.

It seems to have been quite a common fourteenth century idea that

exponible terms like 'incipit' cannot be understood without some analysis.
Let me take William of Ockham as an example. According to him words
like 'incipit' are connotative, and unlike absolute terms like 'homo' they
have nominal definitions expressing their primary signification and all their
connotations explicitly. Ockham's idea seems to be analogous to the
modem idea, to cite an example, that 'unmarried man' gives the analysis of
the term 'bachelor'. 'Bachelor' is in some sense simply shorthand for

24Ibid.
 EXPosmo AS A METHOD OF SOLVING SOPHISMS 211

'unmarried man'. A common theme of modem discussions of such

analysis is that analysans and analysandum should be interchangeable. In
respect to this principle it seems that Ockham's account of connotative
terms gives rise to th~ idea that in fully explicit language -like the mental
language posited by Ockham - all occurrences of connotative terms are
replaced by their nominal definitions. This idea leads naturally to the idea
that the interpretation of a complicated sentence amounts to finding the
nominal definitions and putting them in the places of the problematic
terms. 25

It is readily seen that exposition is not such a technique. The rules of

exposition of the term 'incipit' give a conjunction of two propositions
instead of giving a verbal phrase able to serve the same role in a sentence
as 'incipit' itself serves. And, what is more interesting, there seem to be
quite strong reasons for supposing that full explication of the meaning of
the term 'incipit' in a sentential context demands treatment of the whole
proposition, not just treatment of the term itself.

While discussing the theory of supposition Ockham points out that

'albus' does not have anyone of the standard modes of supposition in the
sentence 'Sortes incipit esse albus', as can be seen from the fact that the
same term has two different modes of supposition in the exponents of this
proposition. It seems that the term 'albus' is seen to function in two ways
in the proposition 'Sortes incipit esse albus', and thus in the fully explicit
form of the proposition, or in the mental language form of the proposition,
'albus' is connected to 'Sortes' in two different ways. In other words, in
the final analysis 'Sortes incipit esse albus' is conjunctive.26 In analyzing
complicated sentences nominal definitions of connotative terms form only
the first step: expositional analysis is also needed. Ockham even states that
every proposition with a connotative term is exponible. As an example he
gives 'Sortes est albus', which looks very simple, but owing to the
connotative term' albus', it has the exponents 'Sortes est' and 'Sorti inest
albedo'.27

When discussing truth conditions of propositions, Ockham says that28

"quaelibet categoric a ex qua sequuntur plures propositiones
categoricae tamquam exponentes earn, hoc est exprimentes quid illa
propositio ex forma sua importat, potest dici propositio aequivalens
propositioni hypotheticae."

I think that Ockham's 'exprimentes quid ilIa propositio ex forma sua

importat' is best translated into modem terminology as stating that
exposition makes the logical form of the proposition explicit.29

25Calvin Nonnore has presented this kind of view of Ockham's mental language in his
unpublished paper "Ockham's Mental Language".
260ckham, Summa Logicae, Opera Philosophica I, St. Bonaventure, N.Y.: Franciscan
Institute 1974, pp. 231-32.
27Summa logicae, pp. 279-81.
28Summa logicae, p. 279.
29When discussing truth-conditions of propositions in his Surrunq logicae Ockham uses
the phrase 'propositio aequivalens propositioni hypotheticae' quite frequently. It seems
212 MIKKO YRJONSUURI

When exposition is understood in this way, it reminds the modern

reader of Bertrand Russell' s famous analysis of
[K] The present king of France is bald

into
[L] There is one and only one present king of France, and every
present king of France is bald.

However, it is interesting to notice that while Russell had in mind the

predicate calculus as a language revealing the logical structure explicitly,
fourteenth century philosophers had only Latin, and they could not turn to
any formal language, except perhaps the language of thought posited by
many fourteenth-century thinkers. A language of thought, however,
certainly cannot be spelled out on paper.

When a respondent using exposition as this kind of method of

interpretation faces an exponible proposition in a disputation, he looks at
the sentence as hiding its real structure. An exponible proposition is only
seemingly categorical.

As will soon become clear, this is just the result I need. Thus it is now
time to tum to the working of the disputation on sophisms.
6. Language games

It would not be too difficult to codify the disputation emerging from

the text of the Dec/aratio as a game, where the players take on the specific
roles of opponent and respondent. As a matter of fact, such games have
been considered by modern commentators on late medieval logic. They
have often interpreted the so-called "Obligations" treatises as giving exact
rules for disputational games very similar to (if not the same as)
disputations on sophisms.30 I cannot discuss this interesting subject at

that only singular propositions are not equivalent to hypotheticals in this sense. (See
esp. part II, chapters 2-20, op. cit., pp. 249-317.) Thus it seems that only singular
propositions would reveal their logical form explicitly. However, strictly speaking
exposition does not offer a finite method of reducing all sentences to the singular level.
Either the system must be seen as potentially infinite so that in actual sophisms
exposition is used only as far as it is needed for the solution, or the system must
contain some other techniques. In the Declaratio, as discussed above, we seem to find
both these solutions. The analysis is not completely carried out, and the techniques
allowing us to replace a common term by a demonstrative pronoun are also employed.
Full account of these features of the system is however beyond the scope of this paper.
30See e.g. treatments of obligational disputations in C. Hamblin, Fallacies, London:
Methuen 1970; I. Angelelli, "The techniques of disputation in the History of Logic",
Journal of Philosophy 67, 1970, pp. 800-15. An account more faithful to the deontic
terminology is given in S. Knuuttila and M. Yrjonsuuri, "Norms and Action in
Obligational Disputations", in Die Philosophie im 14. und 15. Jahrhundert, ed. O.
Pluta (Bochumer Studien zur Philosophie 10), Amsterdam: B. R. Gruner 1989. For
discussion of the relation between Aristotle and obligational disputations see M.
Yrjonsuuri, "Aristotle's Topics and Medieval Obligational Disputations", forthcoming
in Synthese.
 ExPosmo AS A METHOD OF SOLVING SOPHISMS 213

length here, but some comments on the medieval theory of disputation

seem to be appropriate.

As the most important source for the logical theory of disputations

Aristotle's Topics must be mentioned. Book VIII of the Topics considers
discussion in a way that is very naturally read as giving the rules of a
game. The purpose of the dialectical game based on the Topics is to
develop and evaluate arguments against some selected thesis. The main
idea is to develop an inference from premises granted by the respondent to
a conclusion which is contradictory to the thesis discussed. This idea of
refuting a thesis by forcing the respondent into admitting contradictories
seems to be typical also of the rules discussed in medieval "Obligations"
treatises even if obligational disputations are not connected to real truth-
seeking in such a straightforward sense. The important difference between
Aristotle's dialectical games and medieval obligational disputations is that
while Aristotle is concerned with reputable opinions about actual facts,
obligational disputations are concerned with the most remote possibilities.

In connecting sophisms and obligations it may seem problematic that

interest in sophisms is semantical rather than inferential. Typically the
problem in a sophism is that of deciding exact truth-conditions for some
problematic proposition. The primary interest is not in whether the
particular proposition is to be granted or denied, but in general principles
of interpretation. This is clearly the case also in the example discussed
above. However, it seems that "Obligations" treatises exactly suit this kind
of truth-seeking, if any do. Conceptual truths looked for in interpretational
problems are connected to remote possibilities as well as to actual facts.

An interesting modem analogue of Aristotle's dialectical games is

provided in Jaakko Hintikka's game-theoretical semantics)1 Hintikka's
idea is to define logical principles in game-theoretical terms; the games are
dialogue-games between two players, called Myself and Nature, and the
principal idea of the semantics based on these games is that a sentence is
true if and only if it is connected to a game where Myself has a winning
strategy. Results for games on atomic sentences are laid down, and logical
connectives are defined through games connected to them. Thus for
example, a sentence having the form of a conjunction is true if and only if
Myself has a winning strategy for the game beginning with the conjunction
and continuing in accordance with the specific rule given for conjunctions.

The rule for conjunctions is the following: when the game has arrived
at a sentence having the form of a conjunction, Nature chooses one of the
conjuncts and the game continues on that conjunct. Intuitively the idea
behind such rule is that if Myself has a winning strategy for the
conjunction, i.e. if the conjunction is true, Myself has to have a winning
strategy for both conjuncts, i.e. both conjuncts have to be true, since
Myself cannot determine which conjunct Nature chooses. In a similar way

31 A good introduction to game theoretical semantics can be found, e. g., from the first
chapter of J. Hintikka in collaboration with J. Kulas, The Game of Language. Studies
in Game-Theoretical Semantics and its applications (Synthese Language Library, vol.
22), Dordrecht: Reidel 1983.
214 MIKKO YRJONSUURI

in a game connected to a disjunction Myself can choose one of the

disjuncts, and the game continues on that disjunct. Thus a disjunction will
tum out true if Myself has a winning strategy for either of the disjuncts. In
a game connected to a negation roles are reversed and the game is
continued on the negated sentence. Thus if Nature playing the role of
Myself has a winning strategy for some sentence, the denial of that
sentence will tum out true.

Quantifiers are also defined through game rules: in a game on a

universal sentence Nature chooses an individual, and in a game on an
existential sentence Myself chooses an individual. In both cases the game
continues on the instantiated sentence. These rules follow naturally from
the idea of understanding the universal quantifier as an infinite conjunction
and the existential quantifier as an infinite disjunction.

These basic rules are for the language of the predicate calculus, but as
Hintikka has shown, the rules can be developed in ways which are able to
give semantics for richer languages than the basic predicate calculus. For
example, branching quantifiers are easily dealt with by means of these
games. However, I think that for my purposes the simple rules discussed
above give a sufficient intuitive idea of the games.

When we look at Hintikka's semantical games from the viewpoint of

fourteenth century sophisms, it is easy to see that Hintikka tries to reduce
complex sentences to the level of individuals and their properties and
relations in a way which is sufficiently complicated to be able to explain
intricate features of language, but which could serve as a simple unified
seman tical model without the ad hoc solutions all too typical for linguistic
theories.

Now, Hintikka's game-theoretical semantics can say very little about a

sentence like 'Plato incipit esse albus' if 'incipit' is seen to be essentially a
simple verb. The game on an atomic sentence is short: Myself or Nature
wins immediately. If, however, we look at the sentence as a hidden
conjunction, as we saw above that Ockham would do, the game will last
longer, and Myself or Nature wins only mediately.

To put my thesis directly: this is what Billingham means by his

distinction between immediate and mediate sentences. Mediate sentences
hide their real logical structure, and they are "proved" through game-like
processes employing expositional techniques. These "proofs" are similar
to Hintikka's semantical games: the idea is to reduce the proposition to the
immediate level of ostensible individuals and their properties.

Billinghamian analysis of our first example, the sophism 'Sortes est

albior quam Plato incipit esse albus' is almost too simple to be interesting
as exemplifying a semantical procedure like Hintikka's games. As the
explicit form of this sentence is a conjunction of three conjuncts, the first
step in a Hintikka-type seman tical game is that nature chooses one
conjunct. From the text cited above (pp. 204-5) we can see that the
respondent defending the proposition thinks this way: he shows that he
has a winning strategy whichever conjunct the opponent chooses. The first
two conjuncts (CI and C2) are immediately evident according to the casus,
 EXPosmo AS A METHOD OF SOLVING SOPHISMS 215

and the strategy for the third one (C3) is then explained. This third one is
implicitly a negated conjunction (C3 <=> -,(CS & C6)) and thus after
roles have been reversed for the negation, the respondent can choose. The
text shows which one he has to choose in order to win.

Hintikka's discussions concentrate mostly on rules for quantified

sentences. Rules for conjunctions and disjunctions are too simple to
deserve detailed discussion. It is also clear that the sophism 'Sortes est
albior quam Plato incipit esse albus' is interesting mainly because it gives a
background for discussing problems concerning the term 'infinitum'. The
semantical idea behind the treatment of 'infinitum' found in the Declaratio
(as in many other late medieval texts) is the idea of merely confused
supposition, which in Hintikka's game-theoretical semantics would
naturally tum into some kind of quantificational construction. The
interesting thing about the analysis given in the Declaratio is that there the
instantiations given in the spirit of Hintikka's game-theoretical semantics
remove the merely confused supposition. In the final form the whiteness
of a small degree is picked out by a demonstrative pronoun with
determinate supposition.

In the Declaratio the sentence

[E2] Plato in infinitum remisse incipit esse albus

is first rewritten as
[M] Quantalibet albedine data remissionem Plato incipit habere.

Mediately the respondent is led from this sentence to the instantiation

[N] Ista albedine data remissionem incipit habere,

which according to the author should point out a whiteness of infinitely

small degree, and no such whiteness exists. Thus the sentence is false.3 2

The idea behind such reasoning is something like the following: 'In
infinitum remisse' is first simply rewritten in a way synonymous with the
first. The universal quantifier present in this formulation is instantiated in
the step from M to N. In the instantiation - to speak in terms of Hintikka's
game-theoretical semantics - the opponent points to some whiteness. The
sentence is now true if the respondent can point out a suitable minor
whiteness which Plato begins to have. Our author seems to think that the
respondent must point out a definite whiteness of infinitely small degree in
order to do this, and since there is no such thing as whiteness of infinitely
small degree, he cannot do so, and thus the sentence is false.

In game-theoretical semantics quantifiers are read as standing for

suitable individuals; the difference between the universal and existential
quantifiers is only in who chooses the individual. The idea in the
Declaratio seems to be similar. The sentence N picks out two individual
whitenesses, and states that Sortes begins to have the smaller of these two.

32See the analysis ofE2 cited above on pp. 208-9.

216 MIKKO YRJONSUURI

The idea seems also to be connected to modern epsilon-delta analysis,

which in this case would run as follows: the degree of Sortes' incipient
whiteness is infinitely small if, given any however small definite degree of
whiteness (called delta), we can show that Sortes will have also a smaller
degree of whiteness (epsilon). However, while modern mathematicians
never believe that they should pick out an infinitely small value for epsilon,
the author of Dec/aratio requires this. Modern mathematicians wait in the
spirit of Hintikka's game-theoretical semantics until someone else - Nature
- has chosen a value for delta before they choose one for epsilon. The
author of Dec/aratio seems to think that the respondent must have chosen
his value in advance: then it must be infinitely small, since if it were some
small definite value, Nature might choose a still smaller one, and the
sentence would turn out false for that reason.

The text of Dec/aratio is too short to make any far-reaching

conclusions about interesting philosophical ideas behind the judgement. It
may simply be that the strange viewpoint adopted is due to a
misunderstanding of the role of individuals in this intricate issue. Some
general remarks seem appropriate at this point, and they will serve as a
conclusion to my paper.

In the text, exposition is seen as a disputational way of reducing

complicated mediate sentences to the level of immediate judgements about
the properties of ostendable individuals. The analysis points to the
individuals and their individual properties which ultimately verify or falsify
the sentence. Ultimately the analyzed sentence is reduced to the level of
propositions like 'hoc est'. The first sophism is an example of analysing
comparatives, and it is relatively simple. The second sophism is connected
to the first one, but brings in many interesting problems with the term
'infinitum' . According to standard late medieval distinction the word may
have a categorematic or syncategorematic meaning. In the categorematic
meaning the term refers to infinite individuals, and in this meaning
Billinghamian analysis must point to some ostendable real individuals. In
the syncategorematic meaning the term in a sense disappears; no
individuals are required in the analysis. Instead a merely confused
supposition of some other terms is effected. Our example, the Dec/aratio
does not show how the phenomenon of merely confused suppositian
could be incorporated into a Billinghamian system of analysis and in his
Speculum puerorum Billingham does not even mention merely confused
supposition. In the Dec/aratio we have seen in the case of 'infinitum' that a
seemingly very acceptable candidate for analysing merely confused
supposition by a kind of delta-epsilon analysis is not approved of. In this
text it seems that cases of merely confused supposition should be seen as
interesting exceptions to the system of semantical analysis.

University of Helsinki
 Part II

Grammatical Sophisms
 Grammatical sophisms in collections of logical sophisms:
'Amatus sum' in BN. lat. 16135

by Christine Brousseau-Beuermann

The known collections of sophismata are generally homogenous:

either grammatical like Robert Kilwardby's, Roger Bacon's, MS Vatican
lat.7678, or logical; however, some collections are mixed, such as Siger of
Courtrai's in Vatican, lat. 2520, and those contained in the following
manuscripts: Worcester Cath. Q 13, Paris BN 3572,16135 and 16618.1

Among grammatical sophisms, 'Amatus sum' is one of the most often

mentioned; as it is discussed also in merely grammatical works, I will
focus on it, to sketch a typology of the questions raised by this sentence,
and I will show that some aspects of it are relevant to a logical sophism, in
the Parisian collection BN 16135: 'albumluit disputaturum'.

The expression 'Amatus sum' is found in Priscian Minor where it is

given as an example of periphrastic conjugation replacing the missing form
of the passive preterite; in the 12th century, it occurs in the commentary on
Priscian by Peter HeIias, while Robert of Paris uses 'Sor est natusl
nasciturus', in the section of the Summa "Breve sit" devoted to the
construction of the substantive verb with the participle.2 But 'Amatus sum'
is found mostly in 13th century collections of sophismata: Roger Bacon's
Summa Grammatica - which J. Pinborg considered as a collection of
sophisms; the grammatical sophismata and the Commentary on Priscian
attributed to Robert Kilwardby;3 and in mixed collections of grammatical
and logical sophisms: Worcester Cath. Q 13 - where the author of 'Amatus
sum' is a certain William Scardeburh; and three Parisian collections: BN
16135, 16618 and 14927.4 Moreover the periphrase is mentioned in
logical works, in Abelard's Logica Ingredientibus (Glosses on the Peri
Hermeneias), with a slightly different form (doctus sum vel lui), and in

ISiger of Courtrai's Sophismata are not considered here, nor the manuscript Paris BN
lat. 3572.
2/nst. gram., XVII, 81-82, T.2, p.154. Peter Helias, ed. J.E. Tolsson, Cahiers de
l'Institut du Moyen-Age Grec et Latin 27, 1978, pp. 100-1. Robert of Paris, Summa
"Breve sit", Het ludicium constructionis, Deel II, ed. C.H. Kneepkens, Nijmegen:
Ingenium 1988, pp.74-9, 315. This example does not occur in the Priscian
Commentary by Jordan of Saxony.
3R.Kilwardby(?), MS BN lat. 16221, f"19r.
4William Scardeburh, whose name is written on the manuscript, on top of amatus sum
is mentioned by Emden as a magister probably at Oxford; his only known work is the
Sophestria included in this manuscript; but one does not know which sophisms, apart
from the one mentioned, would be by him. For a description and partial edition of the
logical works in this manuscript, cf. S. Ebbesen and J.Pinborg, "Thirteenth Century
Notes on William of Sherwood's Treatise on properties of terms", Cahiers de I'Institut
du Moyen-Age Grec et Latin 47,1984, pp. 1-143, and S. Ebbesen, "Three 13th-century
Sophismata about Beginning and Ceasing", Cahiers de l'Institut du Moyen-Age Grec et
Latin 59, 1989, pp. 121-83. MS BN lat. 14927, ff"191-213v: item sum amatus.

219
220 CHRISTINE BROUSSEAU-BEUERMANN

Jean Ie Page's Appellationes, in the section on ampliatio. 5 But this example

or disputed sentence disappears from the modist and post-modist
grammars. 6

Two groups or families of sophisms may be distinguished, according

to the form of the discussion, the questions raised and the solutions given.
As regards the form of those sophisms, two of them lack an introductory
summary of the questions or 'problemata' (the word occurs in MS 16618):
they are the English sophismata by William Scardeburh and Roger Bacon.?

The questions raised bear upon: the periphrastic conjugation

(circumlocutio) used as a substitutive form (suppletio); the analysis of the
verb phrase and of the subsequent word order (Amatus sum or sum
amatus); on the syntactic structure - is it a word (dictio) or a constructed
phrase (oratio) - and, in the latter case, on the possibility for the participle
to be a subject (suppositum); and on the figure of construction (evocatio),
since the third person of the participle is constrlicted with the first person
of the verb.

The question whether the participle can be a subject, is discussed in

Robert Kilwardby and 16135, and mentioned by Roger Bacon. The three
English sophisms give more importance to the discussion of the evocatio.
Robert Kilwardby, William Scardeburh and the Parisian sophism 16135
discuss the suppletio, the Parisian 16618 deals mainly with this question,
while Bacon doesn't;8 it receives also an extended treatment in the
grammatical treatise of a supposedly "Parisian" author, yet unedited,
Gosvin of Marbais, dating approximately from 1260, i.e. contemporary
with the sophism in 16I35.9

As regards the solutions and positions defended, they differ mainly on

the following points: the analysis of the verb phrase, the question of
suppletion (suppletio) and the tense of the periphrase.
1. The analysis of the verbphrase

For the analysis of the verb phrase, which determines the word order
(Amatus sum or Sum amatus), two descriptions are in competition, and we
may take their presentation from Gosvin of Marbais:

5Abelard, Logica "Ingredientibus", ed. B.Geyer, Beitriige fUr Geschichte der Philosophie
des Mittelalters, XXI, 3, 1919-1927, pp. 348-9. Jean Ie Page, ed. A. de Libera,
Archives d' Histoire Doctrinale et Litteraire du Moyen Age 51, 1984, p. 246.
6Neither in Ps. Albertus Magnus' Grammar, nor in the Daces, nor by Radulphus Brito;
however one of the questions raised by 'amatus sum' - the evocatio - is discussed by
Boethius of Dacia, pp. 230-1; Radulphus Brito, pp. 304, 308; Ps.Albert, p. 94.
7However, the Worcester Cath.Q 13 collection contains at least two sophisms - Sortes
desinit esse albissimus hominum and Quod incipit desinit non esse - with the
preliminary set of questions; cf. Ebbesen, op. cit., 1989, pp. 133-56.
8While Kilwardby(?), in the Sophismata, discusses the suppletio, "he" never mentions
that word in the Commentary on Priscian (ad locum), but uses exclusively the term
circumlocutio (as does Roger Bacon) - BN 16221, f"19r.
9Gosvin de Marbais, BN 15135, ff" 72-84v; C. Thurot, Notices et extra its de mss.
latins, XII, 2, 1869, pp. 337-40.
 'AMATUS SUM' IN BN. LAT. 16135 221

"the verb sum, es, est, can be considered in two ways: either in
the function of composition or in the function of the end (terminus)
(of a movement). In the first way, since composition comes ftrst, the
verb must be placed before the participle, so saying: 'sum vel fui
amatus'. If it is considered in the function of the end of movement,
since the movement precedes its end, the participle must precede the
verb, so saying: 'Amatus sum vel fui'. "lO

This double position, which is also accepted by the author of a

revision of the Glosa Admirantes,ll is alluded to by the author of the
Parisian 16618 who rejects it because of the terminus-movement
description:

"Some people say ... that the sum and the amatus can be ordered
in two ways, that is either by considering that one <sum> is the
composition and the other a participle - and so, the sum is prior, by
sense (sensu) and construction - or by considering that one of them
expresses the movement and the other expresses its end (terminus),
and so the amatus must come first; but this answer fails because of
the equivocation on the word 'terminus'. "12

According to this author, the movement/terminus description applies only

to syntax, in the transitive construction ('I see Socrates'), and not in the
verb phrase Amatus sumP

lOEt notandum quod hoc verbum 'sum, es, est' dupliciter potest considerari: uno modo
in ratione comparationis, et alia modo in ratione termini. Si primo modo, cum
compositio sit ante, sic verbum est ante participium ordinandum, dicendo sic 'sum vel
fui amatus'. Si vero consideretur in ratione termini motus, et cum motus sit ante
terminum sic est participium ante verbum ordinandum, dicendo sic 'amatus sum vel
fui', Thurot, op.cit., p. 340.
IIQuaeritur utrum debeamus <dicere> 'sum amatus' vel 'amatus sum'. Solvitur quod
possumus utrumque bene dicere quia 'sum' dicitur prout terminus, aut in ratione
terminantis, id est oration is, et sic debemus dicere 'amatus sum', cum omne quod
determinat sit posterius ad illud quod determinat, vel prout terminus in ratione
componentis, id est quod ista duo efficiunt dictionem et totum(?) hoc verbum 'sum'
debet praecedere, cum omnis compositio debeat praecedere illud quod est componens et
sic dicemus 'sum amatus'. (BN 18528, f'5) Nothing indicates that terminus, in the first
occurrence, means the end of a movement; but it has the meaning of the end of a phrase
(terminus orationis).
12Aliter dicunt quidam, hoc distinguendo, scilicet quod li 'sum' et Ii 'amatus' possunt
ordinari dupliciter, scilicet aut considerando haec duo [sc. cod. J <secundum> quod
unum est compositio et alterum participium et sic li 'sum' prius est sensu et
constructione vel secundum quod unum istorum dicit motum et alterum dicit eius
terminum et sic li 'amatus' debet praecedere, sed haec responsio fallitur per
equivocationem termini. (BN 16618, 11Ova, 27-31)
13Et ad tertiam rationem contra <i. e., per hoc quod est 'amatus' significatur motus qui
est passio et per hoc quod est sum, terminus illius (f'109rb, 7-8», dicendum quod illud
verum est de motu qui est actio et de suo termino finali, ut 'video Sortem'. Ille tamen
terminus si ordinetur et construatur cum passione precedit passionem sive motum
passionis, ut dicendo 'Sor videtur', hoc est quia iam non significatur in ratione termini
sed subjecti. Cum ergo dicitur quod motus praecedit terminum, verum est de termino
motus qui significat terminum in ratione termini per accusativum vel per ablativum.
222 CHRISTINE BROUSSEAU-BEUERMANN

As is known, the analysis of the verb gets more complex in the 13th
century: to the previous significations and functions - action and passion
according to Priscian and Donatus, predication according to the Peri
Hermeneias and 12th century grammars - is added the signification of
movement which is mostly an effect of the influence of the recently
discovered Physics. 14 This "motional" conception of the verb is
particularly emphasized by the Pseudo-Grosseteste and is found in other
English grammars by Kilwardby and Bacon, before being integrated into
the Danish modist grammars.t 5 For instance, in Kilwardby(?)'s
Commentary on Priscian (ad locum: XVII. 81-82), in the analysis of action
and passion considered in themselves or according to their terminus, the
reference to Book VI of the Physics is explicit. 16 In Bacon's sophism, the
adjective verb - i.e. any verb but the substantive verb - is said to mean
both esse suae rei and movement; it is paraphrased (exponitur) by a
participle which expresses its significate (res) and by the substantive verb
which signifies the being of its significate. Since the verb signifies also the
movement, the substantive verb signifies the terminus - either the end or
the beginning - of the movement, which is signified by the participle; 17 in

Sic autem non signijicatur terminus per hoc quod est' sum' cum potius sit copula. (BN
16618, f"110 va, 21-7)
14-rhe interpretation of the verb as the expression of movement is given by Averroes in
his Commentary on Physics, V. 9: posuerunt nomen cuilibet formae quiescenti et
verbum cuilibet formae mabili: cf. Siger of Courtrai, Summa Modorum Signijicandi,
Sophismata, ed. J. Pinborg, Amsterdam Studies in Hist. of Linguistics 14, 1977, p.
XVII.
15Cf. L. G. Kelly: "Among the early modistae, as in Ps. Grosseteste, Roger Bacon and
in the Compendium modorum significandi in MS Laon 465 <=Sponcius' Grammar,
ed. Fierville>, the verb is actually designated as having a modum signijicandi motus
distantis ab altero. For Siger of Courtrai, the verb has a modus signijicandi per modum
f/uxus, fieri seu motus seu esse": introduction to Ps. Albertus-Magnus, Quaestiones
Alberti de modis signijicandi, Amsterdam: Benjamins,1977, p. XXIV. Idem, "La
physique d' Aristote... ", in A. Joly & J. Stefanini, La grammaire gem!rale des
Modistes aux Ideologues, Lille: Publications universitaires 1977, pp. 107-24. Cum
igitur verbum signijicat substantiam et illam in motu ad peifectionem per agens intra
quod causa compositionis est, erunt tria in verba: substantia,forma motum specijicans
et ipsa compositio, Ps. Robert Grosseteste, Tractatus de Grammatica, ed. K. Reichl,
Veroffentlichungen des Grabmann Instil. 28, Paderbom: Schoningh 1976, p. 46.
16Non est signijicare actionem vel passionemfuturam dupliciter, scilicet in se et in suo
termino vel in se tantum, sicut est signijicare eam praeteritam dupliciter, scilicet in se
et in terminG eius, vel in se tantum, propter hoc quod non est * in motu et actione;
principium? tantum intra determinat, sicut dicit Aristoteles in VI Physicorum. (BN
16221, f"19rb, 19-22)
17Et dicendum est quod unumquodque verbum adjectivum signijicat esse suae rei, unde
cum exponitur, debet exponi per participium quod signijicat rem suom et per verbum
substantivum quod signijicat esse suoe rei, et ideo in circumlocutione cujuslibet verbi
adjectivi cadit verbum substantivum . Adhuc verbum tale signijicat motum; terminus
autem motus, cum ipse motus sit fieri et agere, est esse indivisibile, et ideo cum
verbum substantivum signijicat esse, in significatione cujuslibet verbi taUs cadit
verbum substantivum tamquam terminus mutation is significatae per rem verbi et
tamquom illud quod est terminus et initium ad quod stat resolutio omnium verborum.
Roger Bacon, Summa Grammatica, Opera hactena inedita R. Baconis, fasc. 15, Oxford:
Clarendon Press 1940, p. 148.
 'AMATUS SUM' IN BN. LAT. 16135 223

this case the substantive verb signifies a permanent being (esse cum
perman entia) and 'Amatus sum' is the suppletion of a preterite; but the
substantive verb signifies the existential becoming (esse cum successione),
when 'Amatus sum' is a periphrase of the present 'amor'.1 8

The motional conception is mentioned by the author of 16135: "it is

usually said here that ... there is a suppletion of an act which took place,
compared to its terminus;"19 but he denies that position: "sum is not put
here to signify the end of the movement," and he maintains that sum
signifies the composition, and the participle the act. 20 Thus the Parisian
sophisms in BN 16135 and 16618 retain exclusively the analysis of the
verb sum as a copula - as it is called in BN 16618 - or in the function of
composition - as it is called in BN 16135 - and exclude the motional
conception of the verb.
2. The suppletio: a rational or contingent form (ratio or usus)

The question of the suppletio again opposes the English point of view,
of William Scardeburh and Kilwardby, to the two Parisians ofBN 16135
and BN 16118, while Gosvin of Marbais agrees with the "English" side.
Gosvin, William Scardeburh and Kilwardby give a rationalist explanation
- which is entailed by the motional analysis of the verb - while the
Parisians resort to mere linguistic "usage". Indeed, the problems raised by
the periphrastic conjugation pertain both to grammar and logic. The
linguistic question concerns the supp1etion of a non-existing form, and the
logical one, the extension or suppositio of the subject term, when a verb is
a periphrase with two constituent parts, one of which is in the present and
the other in the past.

Suppletion is due to the lack of (delectus) realisation (impositio) of the

passive preterite in language (Latin, French, or Arabic, as noted by
16618),21 where there is a gap in the realisation of the verbal system. The
concept of suppletio or substitute for a non-existing form rests upon the
distinction between two levels of language: the level of what is said and the

ISDistinguendum tamen est quod per hoc verbum 'est'. potest signijicare esse dupliciter:
vel esse cum successione vel esse cum permanentia; esse cum successione est esse
ipsius agere et ipsius fieri. et hujusmodi esse cadit in circumlocutione presentis et
dicitur de re presenti et idem est sic 'amatus sum' quod 'amor'; esse autem cum
permanentia est esse ad quod terminatur fieri. et sic cadit esse in drcumlocutione
preteriti et dicitur de re preterita ut 'amatus sum, es, est'. Ibid., p. 149.
19Solet enim hic did [quod] communiter quod. .. suppletio est actus in facto esse
comparati ad suum terminum, quod per unam dictionem signijicari non poterat. per
plures dictiones circumlocutio. (BN 16135, f"44ra)
20Ergo patet quod sic debet ordinari 'ego sum amatus', nec ponitur ibi 'sum' ad
signijicandum terminum motus . .. ponitur hoc quod est 'sum' ad determinandum et
dividendum praeterita secundum suam diversitatem ad praesens tempus ut dictum est et
non sub intentione termini sicut ponunt et hoc totum 'amatus sum' circumloquitur
unam dictionem quae si reperta esset in hac 'sum amatus' quorum unum cederet in
actum et alterum in compositionem. (BN 16135, f"44rb, 40, 53-56)
21 ... ut dicunt. in arabico non habent hujusmodi suppletiones. Adhuc in gallico
habemu.r suppletionem in activis. in quolibet praeterito, etfit per hoc verbum habere.
quodpatet si in gallico exponas 'amavi' vel'amaveram'. (BN 16618, f"1l3 va)
224 CHRISTINE BROUSSEAU-BEUERMANN

level of what is meant. The connexion between the two levels is

established by what the medievals called the imposition of words - which
corresponds, to some extent, to what modem linguists call the realisation
or actualisation of a verbal form which belongs to the linguistic system.
The importance given to the concept of suppletio in those texts may be
related to the discussions of the figurative construction, in that two
different semantic levels are considered, secundum intellectum and
secundum sensum: the level of what is meant and the level of what is said;
the suppletio is indeed sometimes considered by the medieval grammarians
as a figure. 22

The lack (delectus) of the past passive was, in Priscian, accounted for
by "usage" (usus), an explanation quoted by 16135 - ista solo usu
defficiunt [44Ra] - and endorsed by Peter Helias who is referred to by
William Scardeburh: "Peter Helias says that the cause of the periphrase
(circumlocutio) is more usage than reason".23 (P. Helias uses the term
circumlocutio while it is in fact a suppletio.)

William Scardeburh, supposing, however, that usage does not explain

why the periphrase occurs in the passive rather than in the active voice,
tries to find a reason. His demonstration is the following: the verb signifies
an act coming out of a substance. The participle signifies an act united to
substance. Any act is in movement and in becoming. Hence an act cannot
be signified by a participle; therefore, the active voice cannot be signified
by a periphrase with a participle, while the passive voice can, since it
signifies that the action is terminated.

So much for the voice, but what about the tense: why is there a
suppletio in the preterite? It is because the preterite signifies that the action
is terminated (hence, united to its substance in the participle). But there are
what we call periphrases in the future (amandus est): William Scardeburh
denies it is a periphrase because the terminus of its action will not take

221n BN 16618, the expression amatus sum is said to be complete according to its
intellectus, not to its sensus (f"110 rb, 3). About this distinction, cf. I. Rosier, "0
Magister . .. , Grammaticalite et intelligibilte selon un sophisme du XIIIe siecle",
Cahiers de l'Institut du Moyen-Age Grec et Latin 56, 1986. The question whether the
suppletio is afigura is discussed: plus repugnat oratio cum dictione quam dictio cum
dictione, sec! quando ab auctore ponitur dictio pro dictione sive pars pro parte, tunc est
figura vel a parte constructionis ut 'sublime volat' pro 'sublimiter', vel a parte
sententiae non mutata constructione ut •spero dolorem' pro' timeo'. Ergo multo fortius
eritfigura si ponatur oratio pro dictione. Sed hocfit in suppletione, ergo suppletio est
figura. (BN 16618, f"I09 va, 23-27) The solution is that it is an oratio figurativa with
an evocatio. (f"1l0 rb, 15)
23De circumlocutione, P. H. dicit quod causa circumlocutionis est magis usus quam
ratio (Worc. , Q. 13, f"38ra). <Verbum substantivum> adjungit alia verba in quibus
preteritum et omnia quae ex ipso formantur, solo usu deficiunt, scilicet in passivis
omnibus et communibus et deponentibus, P. HeIias, Ms. Arsenal 711, f"74ra. Sunt
alia verba, quibus desunt diversa tempora, usu defficiente, non ratione signijicationis,
Priscian,lnst. gram., VIII, xi. 59, T. I, p. 418. Kilwardby makes the same quotation
of P. Helias, whose position he shares, in Commentary on Priscian (ad XVII. 81-82):
et puto quod hoc verbum sit sicut dicit P. H . ... quod omnia hujusmodi verba sive
active vocis sive passive, ex solo usu defectum habent. (BN 16221, f"19 rb, 16-17)
 'AMATUS SUM' IN BN. LAT. 16135 225

place in the "actual present" (presens de presenti) but in the "future

present" (presens defuturo).24 What is interesting here is the attempt to
rationalise the irregularity of language in a semantic way by also resorting
to the motional analysis of the verb.

Much more compelling is the logical argument drawn from the new
motional analysis of the verb, which is defended by Kilwardby and
Gosvin of Marbais: "the suppletio was invented because one word cannot
signify both the movement and the term of movement.''25 Gosvin gives a
rational justification, resting upon the principle that a single ''word'' (dictio)
cannot express a contradiction. The author of BN 16135 is opposed to this
rationalisation.26

More traditionally, the author of BN 16135 resorts exclusively to the

linguistic explanation by usage, which he argues for by saying that if the
passive past participle had been invented, it would have been the
disgraceful 'amatutus' or 'amatumtus' (:F44Ra». The aesthetics of
language is considered as an argument indirectly opposed to the rationalist
explanation. Thus, the motional analysis of the verb contributes to a
rationalisation of grammar, but this modern trend finds some resistance in
Paris among grammarians who are more sensitive to the specific laws of
linguistic usage than to the systematic reconstruction of language.

But whatever explanation one gives of the suppletio, it remains a

question which is of direct interest for the logician: does it count as a word
(dictio) or as a constructed-phrase (oratio), and what is the tense: present,
preterite, or present and preterite? In this discussion, the suppletion of a
non-existing form (amatus sum) is associated with other periphrases

24Sed quare circumloquitur passivum, respondeo quod verbum significat actum

egredientem a substantia; suppono etiam quod participium signijicat actum unitum
substantiae; suppono etiam tertia quod omnis actio est in motu etfluxu quodam; quia
ergo act<iv>um signijicat quodammodo per modum egredientis a substantia, ideo non
potest signijicari ac si esset unitum substantiae; et ideo per participium non potest
signijicari nec circumloqui; passivum autem signijicat rem suam ut terminatam et ita
ut unitam suo subjecto; et ita patet quomodo in passiv~ potest esse circumlocutio et
non in activo . . . Sed quare non est circumlocutio in praesenti et futuro sicut in
praeterito, dico quod passio praesens est egrediens a subjecto et ut nondum unita et
completa. Passio futura est ut egredietur a subjecto sed nondum odquisita nec unita.
Passio praeterita est jam adquisita et unita subjecto et ideo in praeterito est
circumlocutio eo quod significat passionem unitam subjecto . . . Sed estne
circumlocutio infuturo? Quod sic videtur quia possum dicere 'amandus est vel erit', ita
quod 'est' signijicet terminum initialempassionisfuturae. Dieo quod non, quia 'est'
non signijicat tei-minum initialem passionis futurae, quia quando futurum terminatur od
praesens, non erit praesens de praesenti sed praesens de futuro. Praeteritum autem potest
terminari ad praesens de praesenti, et etiam ad praeteritum; et ita potest circumloqui et
non futurum. (Wore. Q. 13, f"38ra)
25 ... per unam dictionem non potuit signijicari motus et terminus motus. Kilwardby,
Sophismata (Bamberg Staatsbibliothek cod.lat. HJ V.l, f"83rb). Dicendum quod
suppletiones fiunt ut motus et ejus terminus qui in unica dictione signijicari non
poterant per plures dictiones signijicarentur, Gosvin de Marbais in Thurot, p. 339.
26This critical attitude toward arguments defended by Gosvin of Marbais is common to
the other grammatical sophism of the same collection, '0 Magister. .. ' edited by I.
Rosier, op.cit.
226 CHRISTINE BROUSSEAU·BEUERMANN

(circumlocutio) which may be expository paraphrases (expositio) of

existing single verbs (erit ambulans = ambulabit). This association of
suppletio and circumlocutio may lead to some confusion if one considers
the suppletion as a periphrase, expository of the verb it replaces, without
making the distinction between two kinds of suppletion: by utility or
commodity and by necessity (propter uti/itatem and propter necessitatem),
which is made by Kilwardby and by Gosvin of Marbais; but in the 12th
century, the difference between suppletion and periphrase was not always
recognised, although Robert of Paris incidentally discusses that aspectP
But this confusion occurs in the first logician who has dealt with the verbal
periphrase: Abelard.
3. The suppletion: word or phrase, and its tense

The question whether the periphrase (circumlocutio) counts as one or

two parts of speech, is first raised by Abelard in the Glosses on the Peri
Hermeneias; he is referring to Priscian:

"when the passive verb is lacking in the perfect and pluperfect

preterite, the noun and the substantive verb replace (subeunt loco)
one verb, as in 'doctus sum vel fui', 'doctus eram vel fueram'. "28

Although this is a case of suppletio, Abelard does not use the term but
the equivalent verb: subire loco. For him, the periphrase is to be
considered as a single part of speech, i.e. as a verb (in vi unius partis, i.e.
unius verbi accipitur; p. 349); the tense is given by the substantive verb,
not by the participle; indeed the periphrase - circumlocutio and not
suppletio - 'erit ambulans' is equivalent to the future ambulabit; he thus
avoids the contradiction between the terms (contradictio in terminis) of the
periphrase, where the verb esse and the participle have different tenses:
otherwise the subject-agent would be walking both in the present and in the
future, hence walking and not walking (since tenses are exclusive of each

27Gosvin makes a distinction between the suppletio propter commoditatem and propter
necessitatem. Kilwardby - in the Sophismata but not in the Commentary - makes the
same distinction: propter utilitatem and propter necessitatem: the suppletion made out
of convenience, is an interpretation: for instance 'philosopher' is interpreted and
developed by the suppletion 'wisdom's lover', or 'antropos' is replaced by 'homo'. The
suppletion is made out of necessity when something cannot be expressed otherwise
(Est autem alia suppletio quae non solum est propter utilitatem sed propter
necessitatem. scilicet quod aliter non potest signijicari) (Bamberg, f"83 ra, 46 ff.); it is
for instance the passive preterite, so that "the movement and its end, which could not
be signified by a single word, were paraphrased by several words" (ut motus et eius
terminus qui in unica dictione significari non poterant, per plures dictiones
circumloquerentur), Gosvin of Marbais in Thurot, p. 339. Robert of Paris gives as
examples of suppletions: ter tria instead of novem, magis quam instead of the
comparative, id est replacing a conjunction, op.cit., p. 315.
28ubi passivum verbum deficit ut in praeteritis perfectis et plus quam perfectis subeunt
loco unius verbi nomen et verbum substantivum ut 'doctus sum velfui', 'doctus eram
velfueram'. Logica lngredientibus, p. 348; cf. M. Tweedale, Abailard on Universals,
Amsterdam: North Holland 1976, p. 285.
 'AMATUS SUM' IN BN. LAT. 16135 227

other).29 Abelard resolves the contradiction by considering the periphrase

as a single part of speech, marked by the tense of the verb esse; the tense
of the participle disappears. But Abelard is not so clear when he comes to
the suppletio, with verbphrases where the participle is in the future
(amaturus) and the verb esse in the present or in the future (est, erit): 'est
amaturus', according to the previous rule, should be a present; however
Abelard does not deny there is a difference between the tenses of those
periphrases ('he is to be loved' and 'he will be to be loved': est amaturus,
erit amaturus) and in fact, he seems to reject the solution he gave for 'erit
ambulans': the reduction to one tense which entails the exchange
(commutatio) of the tense of the participle with that of the verb esse: one
thing is 'he is going to be loved', another 'he will be going to be loved';
'he was going to be loved' is something other than 'he will be going to be
loved'; and when one says 'he was going to be loved', 'going to be loved'
changes its signification neither because of 'was', nor 'was' because of
'going to be 10ved'.30 The only answer Abelard gives is that "sometimes
we want to consider participles with a verb as one dictio. "31

The question is raised after him by the grammarian Robert de Paris

and the authors of sophismata considered here; two solutions are given:

• the reduction to a single part: dictio (Abelard's solution) or suppletio

(the author of Albumfuit disputaturum in 16135 and Gosvin of
Marbais);

• a constructed phrase (oratio) with a single tense, an attraction being

exerted by one of the two tenses: this solution is retained by the
Parisian logician Jean Ie Page, and conceded by the author of 16135.
4. 'Amatus sum' and a logical sophism 'Album fuit
disputaturum' in the same collection, BN 16135

The grammatical sophism 'amatus sum' and the logical one 'albumfuit
disputaturum' occur in the same manuscript BN 16135, but in two
different collections ('amatus sum' belongs to the second collection, dated
by de Libera from 1270, while 'albumfuit disputaturum' belongs to the
first collection, from 1250); it is, however, interesting to underline the
parallelism between the two sentences (they are both periphrases and
suppletions, with a participle marked by a tense different from the tense of
the substantive verb ), and to compare the questions raised from a logical
and a grammatical point of view: the first question bears upon the truth or
falsity, in logic, and the grammaticality and the correct word order, in
grammar; the second question raised by 'albumfuit disputaturum' bears

29Cf. Tweedale, op.cit., p. 289; Kneepkens, op.cit., Deel I, p. 194; De Rijk,

"Abailard's Semantic Views... ", in English Logic and Semantics, ed. H.A.O.
Braakhuis, Nijmegen: Ingenium 1981, p. 29.
30Aliud est 'est amaturus', aliud 'erit amaturus' et aliud 'fuit amaturus' quam 'erit
amaturus'. nec cum dicitur 'fuit amaturus', 'amaturus' propter"fuit' signijicationem
commutat nec 'fuit' propter 'amaturus". Logica "Ingredientibus", p. 349.
31Sicuti quandoque participia cum verbis ita nomina cum verbis in vi unius dictionis
accipi volumus. Ibid., p. 350. P. Helias alludes to this thesis which he rejects: aperte
est contra illos qui 'amatus sum' dicunt esse unum verbum, p. 337.
228 CHRISTINE BROUSSEAU-BEUERMANN

upon the rules of restriction (of the suppositio) and is followed by its
application in the sophism: the rules of supposition of accidental terms
(such as album) in propositions with a verb in the past (fuit) or future
(disputaturum). In the grammatical sophism, a question, parallel to that of
logical supposition, deals with the subject term or suppositum: is it
possible for the participle (amatus) to be subject? - however, the
parallelism is disrupted by the fact that 'amatus sum' is in the first person,
and 'albumfuit disputaturum', in the third, that the participles are
respectively in the passive voice (,amatus') and in the active voice
('disputaturum'); finally, both sophisms analyse the complex tense: is any
of the tenses of the periphrase reduced to the other? Those questions are
specific to each science, but the logical sophism presupposes the solution
of the questions of tense and word order, which pertain to grammar; as a
matter of fact, the logical sophism (album fuit disputaturum) adopts the
word order given as correct by the author of 'amatus sum': 'fuit
disputaturum' as well as 'sum amatus' is constructed with the participle as
an appositum (ex parte post). The fact that chronologically the logical
sophism is earlier than the grammatical one, might lead one to think that the
author of 'amatus sum' was adapting his solutions to those of the logical
sophism, but his solution is commonly held in earlier sophisms, through
the rules of evocatio (Bacon, Kilwardby).
S. Jean Ie Page and MS BN 16135

The problem of the tense of the periphrase 'amatus sum' is solved in

two ways (in 16135): first by the attraction (contractio) of the present of
, s um' by the preterite of the participle; the argument rests upon the
distinction between "being absolutely" (illud quod est simpliciter), that is,
the present, and "to be relatively" (illud quod est secundum quid), that is,
the preterite: the preterite is, relatively to the present, diminished
(diminutum); the rule of attraction of the present by the preterite as given
by 16135, is the following:

"what is diminished relatively to the other, attracts the other in its

nature, if a single understanding comes out of both. "32

This rule is also given in the Appellationes of Jean Ie Page:

"since the present is a simple being compared to the preterite and

the future, the preterite and the future are relative compared to the
present, and preterite and future are a diminution of the present."33

Since the participle, as well as a noun, is an adjunct of the verb 'esse' in

the present, "those adjuncts diminish the verb in the present, attracting it
toward the preterite or the future."34

32Quod est secundum quid et diminutum respectu alterius trahit alterum in sui naturam
si ex illis fiat unus intellectus. BN 16135, 'Amatus sum', f045rb.
33Quoniam presens est ens simpliciter respectu preteriti et Juturi, preteritum et futurum
sunt ens secundum quid respectu presentis. erunt preteritum et Juturum diminutio
presentis. Jean Ie Page, op.cit., p. 247.
34£1 dicimus quod hec adjuncta diminuunt verbum de presenti. trahendo ipsum ad
preterita velfutura. Ibid .• p. 247.
 'AMATUS SUM' IN BN. LAT. 16135 229

As a matter of fact, we may notice that the rule of diminution confirms

and completes Abelard's solution of the contradiction in tenses in 'erit
ambulans': since the participle is in the present and the verb in the future,
the attraction is exerted by the verb, whose tense is relative and
diminished, compared to the present of the participle, or, conversely, the
present of the participle does not diminish the future or the preterite of the
substantive verb.35

The sophism 'amatus sum' uses the same example as Jean Ie Page in
order to illustrate the diminution of relative being (in the past or future): "It
is perfect to say that a relative being is diminished, just as it is to say that a
man is not a diminution of a dead man, but the reverse;"36 but since this
example occurs in Peter of Spain's Tract on Fallacies (before 1250), it is
difficult to say whether our grammarian borrows it from Jean Ie Page or
from Peter of Spain.37

The author of 'amatus sum' rejects the solution given by the rule of
diminution, just as does the author of 'album disputaturum fuit'.3 8 Thus,
the author of the logical sophism, in opposition to Jean Ie Page, considers
the periphrase as a suppletio, hence as an "aggregate" added to one
predicate (totum aggregatum est suppletio unius predicati).39 The tense is
therefore a preterite, and the periphrase is considered as equivalent to the
non-existing form it replaces, in the same way as expository periphrases of
existing verbs.

The solution given by the author of 'amatus sum' is double: he prefers

to keep both tenses (preterite and present) whose significations are
associated in the "confused" preterite which is "joined to" the present (and
can be further analysed into diverse preterite, more or less distant from the
present).40 However he finally makes a concession to the rule of
diminution, hence to Jean Ie Page's position:

35Quia tempus presens est ens simpliciter respectu preterit{vel futuri, non est diminutio
preteriti vel futuri; et propter hoc, dictio signijicans intentionem presentis adjuncta
verbo de preterito aut de futuro non diminuet preteritum vel futurum. Ibid., p. 247.
36Perjecta est enim divisio(diminutio, our reading) entis secundum quid, homo enim
nOll est divisio (diminutio) hominis mortui, sed e converso. Ibid.
37Cf. S. Ebbesen, "The dead man is alive", Synthese 40, 1979, pp. 43-70.
38Dicitur autem a quibusdam quod secundum quid trahat ad se simpliciter dictum, ideo
presens contrahitur ad preteritum; sed unum dictum secundum quid non contrahit aliud;
ideo cum tam preteritum quam futurum sint dicta secundum quid, unum non contrahit
aliud, Sophism 'Album fuit disputaturum', BN 16135, f013 ra = Jean Ie Page's
Appellationes, pars tertia, 1 & 5, ibid., p. 203.
39p 13 ra, ibid., p. 203. This rule is also mentioned by Robert of Paris, who rejects it,
by showing its limitations, op.cit., pp. 78-9.
40Ad representandum illud quod est confusum, dicimus 'amatus sum' .. . ad
representandum autem praeteritum quod est juxta praesens, dicimus 'amatus eram', ad
representandum autem illud quod est valde remotum a praesenti, dicimus 'amatus
fueram', ut vero repraesentemus illud quod est conjullctum praesenti, dicimus 'amotus
sum'. (BN 16135, f044rb,48-52)
230 CHRISTINE BROUSSEAU·BEUERMANN

"one should not say, properly speaking, that the present is here
reduced to the preterite, but according to the rule above <Jean Ie
Page's>, one may reasonably say that, if one intends to signify the
being of a past love so that there be one intellect, an attraction is
exerted by the nature of the preterite."41

Thus, from this parallel between the logical and grammatical sophism,
it appears that there is a common source: the rule of diminution used by
Jean Ie Page, whose presence in the logical sophisms of 16135's first
collection (studied by de Libera) is still active in the second collection;
moreover, the section of Ie Page's Appellationes on ampliatio deals for a
large part with suppletion - 'amatus sum' being the example - and the
author of 'amatus sum' alludes to the amp/iatio, which he leaves to the
logician: "let's leave ampliation to the logician."42 The grammatical
sophism maintains the frontier between logic and grammar, but 'amatus
sum' is relevant for logic, since the ampliation or restriction of the
reference (suppositio), depends, for some part, on the solution given to the
problem of the tense.

One should underline the fact that the notion of suppletio, which is not
really distinct from the circumlocutio in the 12th century, is widely
discussed by 1260 by Gosvin of Marbais and the authors of sophismata;
certainly its importance is due to the discussion of the ampliatio in logic.
The problem is of common interest and may justify the presence of
grammar in a collection of logical sophismata. However, this is not the
case with the other mixed collection: Worcester Q 13, where the only
connection one can find between grammar and logic is rather loose, in the
sense that William Scardeburh, with the motional analysis of the verb, is
also influenced by Aristotle's Physics.

In conclusion, it appears from the discussion of the sophism 'Amatus

sum' that two grammatical trends are in opposition by 1260: the English
authors are revising the analysis of the verb by applying the concept of
movement borrowed from the Physics; it contributes to giving a rational
account of an irregularity of language - the gaps in the verbal system and
their suppletions, which were traditionally explained by usage. But some
grammarians, such as the authors in BN 16135 and 16618, resist this new
trend, by opposing 'usus' to 'ratio', while Gosvin of Marbais holds an
intermediary position. Those texts testify to the tensions existing between
the rationalist wave which is at the same time contributing to the modist
grammar, and conservative trends which represent what we could call the
grammar of usage.

Associee au Centre National de la Recherche Scientifique, URA 381

41 Dicendum quod hic principalius est preteritum et si debe ret ibi reductio fieri. fieret
presentis ad preteritum quia quod est secundum quid et diminutum respectu alterius
trahit .. (BN 16135, f"45rb)
42Qllaliter autemfiat ampliatio. debet in logicis quaeri. BN 16135, f"42vb.
 La distinction entre actus exercitus et actus significatus dans
les sophismes grammaticaux du MS BN lat. 16618 et autres
textes apparentes
par Irene Rosier

1. Le manuscrit BN lat. 16618 comporte, des folios 40r a 114ra, un recueil

de sophismes grammaticaux, inacheve et anonyme, que I' on designera,
d'apres son incipit: Sicut dicit Remigius. n est precede par une collection
de sophismes 10giques, qui a ete decrite par Alain de Libera.l Nous
n'avons trouve aucune mention nous permettant d'assigner aux sophismes
grammaticaux une date ou un auteur. 2 Une allusion au roi de France peut
faire penser qu'il s'agit d'un texte parisien (f. 56va), et l'usage frequent
des articles Ie et del en plus du Ii communement utilise par les philosophes,
semble indiquer un auteur dont la langue matemelle est Ie fran~ais. Les
ouvrages cites, d'une richesse et d'une diversite remarquable, attestent
d'un milieu universitaire, mais ne procurent aucune information quant Ii sa
date de composition}

Sur Ie plan doctrinal, comme nous Ie verrons, notre recueil trouve sa

place dans un ensemble de textes representant un courant particulier de
grammaire speculative, dans lequel on peut placer Robert Kilwardby et
Roger Bacon, mais egalement plusieurs traites anonymes, et en particulier
des sophismes, s'echelonnant sur Ie second quart du treizieme siecle.4 La
caracteristique essentielle de la demarche de ces auteurs, est qu'ils
analysent Ie langage, non pas en etudiant exclusivement les sequences
linguistiques prises en elles-memes, mais en tenant compte de la
production concrete de celles-ci. Ils mettent alors au coeur de leur systeme
l'intention de signifier qui a preside a leur production par un locuteur
donne (intentio pro/erentis). De ce point de vue, ils sont conduits Ii

1A. de Libera, "La litterature des Sophismata dans la tradition tenniniste parisienne de la
seconde moitie du XIIIe siecle", dans The editing of Theological and Philosophical
Texts from the Middle Ages, ed. M. Asztalos, Stockholm: Almquist & Wiksell
International 1984, pp. 213-44.
211 n'est pas rare que les grammairiens utilisent leur propre nom dans les exemples
d'evocatio a la premiere personne, tout comme Priscien disait ego Priscianus scribo.
On trouve par exemple ego Robertus dans Ie commentaire de Robert Kilwardby. On
signalera done que dans notre texte, Johannes est Ie nom utilise: ego Johannes curro (C.
42vb). Ce prenom revient plusieurs Cois (C. 43va, 44ra, 45va etc.).
3En plus de Donat, Priscien et Pierre Helie, et des textes logiques et philosophiques
aristoteliciens (De caelo et Mundo, De generatione et corruptione, De Anima, etc.), on
rencontre des auteurs de l'Antiquite (Ciceron, Seneque, Augustin, Porphyre), ou de
I' Antiquite tardive (Isidore) et du haut Moyen-Age (BMe, Remi d' Auxerre). On
mentionnera egalement Ie Thymee de Platon, Ie Liber sex principiorum, et Aigazel.
L'auteur cite I'ethica vetus et I'ethica nova, mais on n'a pas de reference au liber
ethicorum, traduction de Robert Grosseteste datant des annees 1246-47. II mentionne la
Metaphysica vetus, ainsi que Ie livre X, probablement dans la translatio anonyma.
4Cf. I. Rosier, '''0 Magister .. .': Grammaticalite et intelligibilite selon un sophisme du
XIIIe siecle", Cahiers de l'lnstitut du Moyen-Age Grec et Latin 56,1988, pp. 1-102.

231
232 IRENE ROSIER

s'interesser aux enonces par lesquels Ie locuteur effectue un acte de

langage, par opposition it ceux qui ne lui servent qu'it signifier un etat de
chose, lesquels correspondent aux enonces canoniques declaratifs,
composes d'un nom et d'un verbe. C'est dans un tel contexte que doit etre
situee la distinction entre actus exercitus et actus significatus.

2. Avant d'en venir aux questions doctrinales, on peut tenter de degager

quelques caracteristiques fonnelles du Sicut dicit Remigius, que presentent
d'autres collections de sophismes grarnmaticaux de la meme epoque.5 II
presente une longue introduction tMorique (40r-55vb), suivie des
sophismes proprement dits. 6 Cette introduction est une petite summa,
comportant des considerations generales qui seront, comme dit l'auteur,
necessaires pour la detennination des sophismes qui suivra. Elle se tennine
par des questions disputees, portant notamment sur l'interjection et sur la
question generale de la distinction entre peifectio ad sensum et peifectio ad
intellectum.7 L'expose sur l'interjection servira it la resolution des premiers
sophismes du traite. 8 Quant it la discussion des differents types de
completude, elle est essentielle dans les disputes grarnmaticales, puisque
c'est pratiquement toujours la premiere qui est posee it propos d'un enonce
sophismatique. Celui-ci est en effet choisi precisement parce que
l'assignation de sa correction fait difficulte, qu'il s'agisse d'une
construction atypique ou d'un enonce figure. La Summa grammatica de
Roger Bacon presente la meme organisation: une introduction doctrinale
portant sur des questions de completude, suivie de sophismes.
L'introduction des sophismata grammaticalia attribues it Robert Kilwardby
est beaucoup plus breve.9 Un peu differente est la Summa de grammatica
d'un certain Durandus. lO A la suite de l'introduction, Ie traitement de
chaque probleme se divise en un expose de la regIe, correspondant au
traitement in generali, et une discussion sur des sophismes, correspondant
au traitement in speciali. On citera egalement l'importante collection de
sophismes qui suit la Summa in arte grammatica de Johannes Ie Rus.ll II

5Cf. I. Rosier, "Les sophismes grammaticaux au xme siecle", a paraitre dans Medioevo.
6Sicut dicit Remigius: "Prima particula continet quedam communia que valent ad ea que
sequuntur, ne error in principiis multos producat errores in sequentibus, ut vult
Aristoteles in primo Celi et Mundi. Secunda continet orationes que sunt proprie, non
figurative, tertia figurativas." (f. 4Or)
7Les arguments sont caracoorises comme 'sophistiques': Sicut dicit Remigius: "Ad
omnia ista oportet solvere pro et contra quia sophistica sunt." (f. 54vb, dans Ie cours de
questions sur la completude)
8Proch dolor quia magister non disputat (f. 55vb); Ve tibi (f. 57vb); 0 virum
ineffabilem (60va). Plus loin dans Ie recueil se trouve Ie fameux 0 magister (f. 8Ova).
9Les huit manuscrits et la liste des sophismata de cette collection sont donnes dans
Rosier, op. cite (n. 5). Nous avons decide de considerer Ie MS de Zwettl (= Z),
Stiftsbibliothek, cod. 338, ff. 135-161 comme Ie manuscrit de base.
lOSumma quidam de grammatica, MS Barcelone Ripoll 109, f. 158ra-173ra. Incipit:
Quoniam oratio est ordinatio dictionum congruam perfectamque sententiam
demonstrans. Ce traite, qui se trouve dans Ie mllme manuscrit que Ie fameux "Guide de
I'etudiant" (cf. n. 18) est malheureusement d'une ecriture minuscule et pratiquement
illisible dans la copie que nous possedons.
II Le Tractatus de constructione occupe les ff. 89ra-lOl v du MS Vat. lat. 7678. Elle est
precedee par des sophismes logiques, decrits par M. Grabmann, Die
Sophismatalitteratur des 12. und 13. larrhunderts mit Textausgabe eines Sophisma des
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 233

s'agit en fait d'un recueil composite de questions et de sophismes, comme

en temoigne Ie fait que certains d'entre eux s'y retrouvent plusieurs fois
(cf. infra note 69) , et que plusieurs pages ou demi-pages sont laissees
blanches.

Ces sophismes nous semblent relever des disputationes in scolis, qui

faisaient partie de l'enseignement universitaire des Arts. On y trouve
souvent mention du respondens ou de I' opponens. L' appellation de summa
par laquelles ces recueils de sophismes sont parfois designes est a relever.
Par ailleurs, il semble bien qu'ils proviennent d'une redactio globale plutOt
que de reportationes successives des sophismes: en effet, certaines parties
sont abregees ou non rectigees, qu'il s'agisse de reponses ou de questions,
parce qu'elles relevent, soit d'autres sophismes anterieurement disc utes,
soit de I' introduction doctrinale, auxquels elles sont explicitement
renvoyees. Par ailleurs, il est frequent, comme dans notre recueil Sicut
dicit Remigius ou dans la Summa grammatica de Bacon que la discussion
de toutes les questions se posant a propos d'un sophisme apparaisse
d'abord, par series d'arguments pro et contra successives, et que les
solutions et les reponses aux arguments soit regroupees ensuite, ce qui
oppose clairement la disputatio proprement dite, ala determinatio. On
notera egalement que la longueur des sophismes a l'interieur d'un meme
recueil est variable et que certains peuvent etre tres longs, ce qui fait penser
qu'ils ne correspondent pas necessairement a une seule le<;:on: c'est Ie cas
d'un sophisme contenu dans un recueil acephale d'un manuscrit de
Londres, qui court sur huit folios, et comporte sept questions principales,
chacune s'achevant sur sept ou huit petites questions accessoires.l 2

Differents indices nous font penser que ces collections datent plutOt du
milieu du XIIIe siecle que de la periode posterieure. L'on verra en tous cas,
a partir des questions dont nous traiterons ici, que les nombreux echanges

Boethius von Dacien, Beitriige zum Geschichte der Philosophie des Mittelalters
XXXVI, Aschendorff 1940, p. 33 ff. L'on trouve alors la mention: Explicit Summa
magistri lohannis <Ie Rus> in arte grammatica. Incipiunt latina disputata. Suit ensuite
un premier ensemble de sophismes, puis une interruption (f. 124v blanc), et ensuite des
questions et des sophismes, avec des demi-pages ou des pages blanches au milieu. A la
fin de I'ensemble on peut lire, d'une autre main: Explicit Summa grammatices
Magistri Joannis Ie Rus. Selon les catalogues, iI ne semble pas que ces sophismes et
questions suivent la Summa de Joannes Ie Rus dans les autres manuscrits qui la
contiennent, mais nous n'avons pu les consulter (MS London British Lib. Add. 8167,
ff. 136r-154r, et Miinich eLM 7205, ff. 59-67). Wallerand les a decrits brievement,
Les oeuvres de Siger de Courtrai, Les philosophes belges, VIII, Louvain 1913, p. 29
ff.
12I1 est iI noter que les sophismes contenus dans Ie MS London British Library Royal
8AVI, ff. 36ra-46vb suivent Ie commentaire de Robert Kilwardby sur Priscien Mineur.
Le premier sophisme, dont nous n'avons pas Ie debut, comporte une premiere question
sur la perfection, une seconde sur la distinction entre ad sensum et ad intellectum, une
troisieme sur la construction de in avec un nom, et une quatrieme sur la preposition. La
reponse au dernier argument de la troisieme question nous indique que Ie sophisme est
peut-etre In nomine Patris: "Ad secundum patet solutio per predicta. Nam Iicet non sit
ibi actus significatus per verbum, tamen est ibi actus exercitus de deo sic In nomine
P<atris> etc." L'on trouve ensuite, au f. 38ra, Ie long sophisme Modio vini ad
denariumve illi qui non habet.
234 IRENE ROSIER

et renvois d'un texte a l'autre permettent de poser leur appartenance a un

meme univers de doctrine. Tous ces recueils presentent des ressemblances
frappantes, non seulement sur Ie plan formel, on l'a dit, ou sur celui des
sophismes discutes, que 1'0n retrouve souvent d'un traite a l'autre, mais
egalement sur Ie plan du contenu. D'un cote, il ne s'agit jamais de
sophismes "metalogiques", comme on en trouvera ulterieurement avec
certains sophismes modistes, et l' enonce propose a discussion est
effectivement l'objet de la dispute sans qu'il ne constitue un simple
pretexte. 13 De l'autre, la question sur la completude et la correction, qui est
presque toujours la premiere po see, est souvent l'occasion de
developpements plus generaux, lorsqu'ils ne sont pas regroupes dans des
introductions prevues a cet effet. Ces recueils de sophismes grammaticaux
sont des temoins essentiels de ce que Jan Pinborg appelait Ie "pre-
modisme".

3. Gabriel Nuchelmans a recemment, dans un important article, etudie la

distinction entre actus exercitus et actus signijicatus, en s'interessant
particulierement aux traites sur les syncategoremes du xm e siecle. 14 Sans
chercher a resumer ce travail, on retiendra ici deux points, qui nous
serviront pour l'analyse de ces memes notions dans Ie corpus grammatical.
Le premier est la mise en relation entre deux distinctions: actus signijicatus
et actus exercitus, d'une part, et per modum conceptus et per modum
affectus, d'autre part. Le second est l'opposition qui semble se faire jour
entre deux types d'approches. Certains auteurs mettent en avant l'idee que
les syncategoremes sont la marque (nota) d'un affect, et par consequent
l'instrument d'un acte dont l'agent reel est Ie locuteur, les autres
s'interessent davantage aux proprietes des syncategoremes en tant
qU'unites linguistiques, l'acte realise par leur intermediaire etant une simple
actualisation de leurs proprietes intrinseques. Pour cette raison,
I' effectuation de I' acte par Ie locuteur peut etre laissee de cote: alors que,
chez les auteurs du premier groupe, I' acte exerce, la negation, par exemple,
est celui que realise Ie locuteur au moyen du syncategoreme, chez ceux du
second, on finira par parler de l'acte qu'exerce Ie syncategoreme lui-meme
sur les categoremes, en Ie caracterisant comme signifiant une "disposition"
ou une "circonstance" des choses designees par les categoremes, et non
plus comme signifiant "sur Ie mode de l'affect". La demarche centree sur
l'acte effectue par Ie locuteur semble etre propre a des auteurs anglais
(Roger Bacon, Guillaume de Sherwood), alors que celle qui met l'accent
davantage sur les relations formelles, syntactico-semantiques, entre termes,
para!t caracteriser des auteurs parisiens (Jean Ie Page, Nicolas de Paris,
etc.).

i3Cf. I. Rosier, "Un sophisme grammatical modiste de maitre Gauthier d'Ailly",

Cahiers de l'Institut du Moyen-Age Grec et Latin 59, 1990, pp. 181-232, et la
bibliographie citee.
14"The distinction actus exercitusl actus signijicatus in medieval semantics", dans
Meaning and Inference in Medieval Philosophy. ed. N. Kretzmann, Dordrecht: Reidel
1988. pp. 57-90. Cf. aussi du meme auteur: "Ockham on performed and signified
predication", dans Ockham and Ockhamists, ed. E.P. Bos and H.A. Krop, Artistarium
supplementa IV, Nijmegen: Ingenium 1987, pp. 55-69.
 ACTUS EXERCITUS ET ACTUS SIGNIF1CATUS 235

4. Un autre domaine conceme It la fois par la distinction entre actus

exercitus et actus signijicatus, et par celle entre per modum affectus et per
modum conceptus, signale par Gabriel Nuchelmans, est celui de
l'interjection. L'interjection, en effet, signifie per modum affectus, par
opposition It toutes les autres parties du discours, qui signifient per modum
conceptus. Dans sa Summa de Sophismatibus et Distinctionibus, Roger
Bacon, en un passage essentiel, explique de maniere tres claire la proximite
de I'interjection avec les syncategoremes. II invoque d'abord un argument
tout It fait commun, fonde sur I'opposition entre deux modes de
signification: d'un cote, celui du nom distributio ou du verbe distribuere,
qui signifient la distribution sur Ie mode du concept, de I'autre celui du
syncategoreme omnis, qui la signifie sur Ie mode de I'affect. Les
interjections et les syncategoremes sont ensuite rapprocbes, du fait qu'its
signifient sur Ie mode de I'affect. II y a en fait, dit notre auteur, deux types
d'affects: les uns, provoques par des evenements heureux ou tristes,
affectant l'ame, tels la joie ou la douleur, sont signifies sur Ie mode de
l'affect; les seconds sont des actes de la raison, comme I'union,
I'ordonnancement, la distinction, et de ces actes, les syncategoremes sont
les signes. Ainsi, si l'intellect con\;oit deux choses ne convenant pas I'une
avec l'autre, it est affecte par leur incompatibilite, qui, une fois con\;ue,
peut etre signifiee et exprimee en utilisant Ie signe non, et en produisant la
proposition homo non est asinus. L'interjection, comme Ie syncaregoreme,
constitue, pour Ie locuteur, des marques d'affects, irrationnels et exterieurs
dans Ie premier cas, rationnels et inrerieurs dans Ie second cas. Le passage
se conclut sur une declaration tout It fait importante, que I' on trouvera
frequemment chez les grammairiens dont nous parlerons. L 'homme est
I'agent principal de 1'0pcSration de negation ou de celle de distribution, et
les signes non ou omnis n'en sont que I'instrument, de meme que celui qui
frappe est I'agent principal du coup, dont Ie baton n'est que I'instrument.
On notera que ce passage de la Summa de Sophismatibus et Distinctionibus
se retrouve litteralement dans les Syncategoremata attribues It Robert
Bacon,15 ce qui pourrait etre un argument en faveur de I'attribution de ce
texte It Roger et non Robert Bacon, hypothese plusieurs fois discutee. 16

15Roger Bacon, Summa de sophismatibus et distinctionibus, ed. R. Steele, Opera XIV,

pp. 153:19 ff (nous corrigeons Ie texte edite, la ponctuation, notamment, Ie rendant
incomprehensible). Nous indiquons entre crochet les correspondances avec les
Syncategoremata de Robert Bacon (edites partiellement par H. Braakhuis, De 13de
Eeuwse, Tractaten over Syncategorematische Termine, I,lnleidende Studie, Leyden
1979): "Ad aliud <dicendum> quod significare distributionem vel collectionem est
dupliciter, scilicet, per modum quietis, ut hec nomina distributio, collectio, vel per
modum motus et fieri, ut hec verba distribuere, colligere; aut significari possunt per
moduin affectus, et sic hujusmodi singna. Set duplex est affectus, unus qui causatur ex
tristantibus vel delectantibus vel admirabilibus extrinsecis animam afficientibus,
cujusmodi affectus est gaudium, dolor, admiratio et hujusmodi, et ita significatur per
interiectiones, quia sine deliberatione rationis concipitur illud quod afficiebat, et
exprimitur per interiectionem. Sic non significat signa affectus nec per modum
, conceptus [affectus (!) ed. Steele]. Alius autem affectus animi est, scilicet hujusmodi
actiones rationis, distinctio, dissensus, ordo, unio violenta duorum, et hujusmodi que
sic ponunt, cum anima accipit duo incomplexa disconvenientia discoherentia, ut
hominem et asinum, afficitur quadam dissensioni [diffusione(!), ed. Steele], cujusmodi
dissensionis [diffusionis, ed.] intra est hec dictiQ non signum et nota, et ideo hec dictio
236 IRENE ROSIER

5. La question difficile de la signification et de la construction de

I'interjection est abordee dans plusieurs textes, qui se repondent, et dans
lesquels se developpe une problematique commune. Roger Bacon revient
souvent sur cette question, dans sa Summa grammatica et dans sa Summa
de sophismatibus, dans Communia Naturalium, puis dans Ie De signis, et
enfin plus brievement dans Ie Compendium Studii The%giae. Robert
Kilwardby l'aborde dans son commentaire sur Priscien Mineur, ainsi que
dans les Sophismata qui lui sont attribues. Les exposes Ies plus complets
se trouvent, d'une part dans un traite intitule De Interiectione, qui suit Ie
commentaire sur Priscien Mineur de Robert Kilwardby dans Ie manuscrit
du Vatican (Vat. lat. 298) et Ie commentaire du Ps-Kilwardby dans Ie
manuscrit de Cambridge (Univ. Ubr. Kk III 20), d'autre part dans Ie
chapitre sur l'interjection dans ce demier commentaire, et enfin dans Ies
questions afferentes de notre Sicut dicit Remigius.

La discussion sur Ia distinction entre per modum conceptus et per

modum affectus est posee immediatement a partir d'un conflit d'autorites:
si I'interjection est une partie du discours, elle doit signifier, selon
Priscien, un "concept mental" (conceptum mentis); or Donat Ia definit
comme signifiant un affect; enfin, selon Aristote, tout mot doit etre Ie signe
d'une "passion de l'fune".17 II devait s'agir d'une question pro forma, au
programme des examens, des Ies annees 1240, puisqu'elle se trouve posee
dans Ie fameux "Guide de I' etudiant" conservee dans Ie manuscrit de
Barcelone. 18 Selon Ia reponse Ia plus commune, I'interjection est bien une

non significat necessitatem per modum affectus [= Syncat. p. 141-2: Cum anima
accipit duo incomplexa disconvenientia, ut hominem et asinum, afficitur quadam
dissensione, et huic dissensione, que est intra, respondet hec dictio 'lion' in sermone
extra. Unde iIIius dissensiones que afficit animam nota est hec dictio '11011']. Similiter
cum anima concipit duo complexa afficitur et disponitur ordine eorum, cujusmodi
ordinis actualiter afficientis animam hec dictio si est nota, et ideo hec dictio si denotat
ordinem per modum affectus [= SYllcat. p. 153] (... ) iste autem affectiones non
significantur per interjectiones tum quia non causantur a tristibili vel admirabili
extrinseco, tum quia non concipiuntur et sine deliberatione rationis proferuntur, quod
exigitur ad significatum interiectionis omnino, sicut est de distributione et divisione;
cum enim anima accipit subiectum respectu predicati, ut predicatum conveniat
cuiuslibet parti subiecti, afficitur quadam divisione subiecti, in comparatione ad
predicatum tale, cuiusmodi divisiones omllis est nota (... ). Ad aliud, dicendum quod
sicut homo vel anima est principale agens in operatione negandi, hec dictio 11011 est
instrumentum, et percutiens est principale agens in percutiendo, baculus instrumentum;
sic homo vel anima est principale agens in divisione et distributione subiecti, omllis
instrumentum [= SYllcat. p. 141: Differenter tamen homo negat et hec dictio 'lion'
negat, nam homo negat sicut agens, hec dictio '11011' sicut instrumentum, sicut dicitur:
'homo percutit', 'baculus percutit']."
16A. de Libera, "Les Summulae dialectices de Roger Bacon, I. De termillo, II. De
elluntiatiolle", Archives d' Histoire Doctrinale et Litteraire du Moyen Age, Annee 1986
(1987), pp. 154 ff. Sur les questions d'attribution, cf. Braakhuis, op. cite (n. 15).
17 La discussion sur la signification des syncategoremes utilise Ie m8me type
d'arguments, cf. Nicolas de Paris, ed. Braakhuis, op. cite (n. 15), pp. 2-4.
18Barcelone Ripoll 109, f. 142va. Je remercie Claude Lafleur, qui prepare l'edition du
Compelldium, de m'en avoir communique 1a transcription. Cf. P.O. Lewry,
"Thirteenth-Century Examination Compendia from the Faculty of Arts", dans Les
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 237

partie du discours, a condition d'en exclure les exclamations ou les

transcriptions d'onomatopees, comme 'ha ha hal', et elle signifie un
concept sur Ie mode de l'affect (signijicat conceptum per modum affectus).
L'on n'entrera pas ici dans Ie detail de ces longues argumentations, et 1'0n
ne retiendra que les differentes interpretations de la distinction qui nous
occupe. On en relevera trois principales.

a) La these la plus repandue repose sur l' opposition entre deux

modes selon laquelle la chose est con~ue et signifiee. Une chose, en
effet, peut etre dans l'arne soit par sa species, son image, son intention,
sa similitude, soit y etre reellement. Dans Ie premier cas, elle y est sur Ie
mode du concept - et on cite Ie celebre adage emprunte au De Anima:
"ce n'est pas la pierre qui est dans l'arne, mais sa species;"19 dans Ie
second elle s'y trouve sur Ie mode de l'affect. Cette distinction est tres
claire: comme Ie dit Ie Pseudo-Grosseteste, celui qui pense a la douleur
n'en est pas affecte. 20 L'auteur anonyme du Sicut dicit Remigius
illustre 1'0pposition par une anecdote. Un enfant peut bien dire je
pleure, sans pleurer effectivement (actu). II peut egalement proferer des
sons, des gemissements, qui ne sont pas des mots. Lorsqu'il dit je
pleure, nous ne nous attendons pas ace qu'il pleure, et il ne nous emeut
nullement, alors que lorsque nous l'entendons pleurer, il nous touche.
En effet, les sons qu'il emet alors indiquent que la douleur est
reellement en lui, en tant qu'affect, et non pas en tant que douleur
con~ue ou simulee, en tant que concept, presentant seulement une
similitude avec la chose reelle. 21 Le nom dolor ou Ie verbe doleo

genres litteraires dans les sources thtologiques et philosophiques medievales, Louvain

1982, pp. 101-16.
19De Anima III, 43Ib29-432al: cf. J. Hamesse, Les Auctoritates Aristotelis, Louvain
1974, p. 188, nO 163.
20pseudo-Grosseteste, Tractatus de grammatica, ed. K. Reichl, Paderbom: SchDningh
1976: "Interiectio est pars orationis indeclinabilis mentis affectum significans voce non
disposita. Omnes cetere partes orationis significant mentis conceptum, hoc est
significant in quantum res se ponunt apud animam secundum species suas. Sed affectus
sunt qui secundum se ipsos apud animam se ponunt. Cogitans de dolore non
afficitur, sed dolens." (p. 59: 23-26) II est remarquable que cette opinion soit, dans Ie
Interiectione, explicitement attribuee a un quidam philosophus, auquel il est renvoye en
deux autres occasions (comme Magister grammatice, et comme quidam Philosophus
grammatice). Or, a chaque fois, on retrouve aisement une citation, souvent litterale, du
chapitre sur l'interjection du Tractatus de grammatica (cf. infra n. 44). Voici Ie texte,
pour la question qui nous occupe: "Adhuc sicut dicit 'quidam Philosophus' de
interiectione loquens, significare conceptum est significare rem que per speciem suam
sit apud animam et non per se ipsam. Sed significare mentis affectum est significare
rem que secundum se sit in anima et non per speciem [= p. 59:23-26], cuius est dolor,
gaudium, etc. Iste enim res sunt in se et ipsis apud animam et disponunt ipsam et
afficiunL Quare cum omnis interiectio significet rem existentem apud anima secundum
se - quod patet quia cogitans de dolore non afficitur, sed dolens [= p.59:26]. Patet quod
interiectio affectum significat [argument qui sera retenu, a condition que I' on interprete
signijicare affectus, comme signijicare per modum affectus]." Sur la m8me distinction,
cr. aussi Robert Kilwardby, sophisme Proch dolor 0 socii quia socius noster frangitur
crura (nO 17, texte cite infra, n. 75) ou encore les Syncategoremata de Pierre d'Espagne,
(ed. Braakhuis, op. cite (n. 15), p. 266; Spruyt, Peter of Spain on Composition and
Negation, Nijmegen: Ingenium 1989, pp. 34-5; voir Ie commentaire, ibid., pp. 144-7).
238 IRENE ROSIER

signifient Ia douleur en tant que con~ue par I'enonciateur, I'interjection

heu signifie cette meme douleur en tant qu'elle est reellement en Ie
Iocuteur. 22 Le fait que I'affect so it veritablement en lui Ie pousse a
s' exprimer de maniere subite et sans deliberation. 23

b) Se10n une seconde interpretation, Ia distinction entre "sur Ie mode

du concept" et "sur Ie mode de I' affect" peut etre rapportee a des
modalites de Ia signification. Par rapport a un affect exterieur, soit il y a
deliberation, jugement, et signification, et il est signifie sur Ie mode du
concept, soit il y a reaction immediate, violente, et sans deliberation, et
il est signifie sur Ie mode de l'affect. L' auteur du De interiectione admet
ensemble Ies deux premieres interpretations. 24

A partir de cette idee, certains auteurs, comme Robert Kilwardby25 ou,

plus nettement encore, l'auteur anonyme des sophismes londoniens,

21Sicut dicit Remigius: "Aliter patet planius in exemplo. Puer enim potest dicere
'ploro' Iicet non ploret actu. Potest etiam proferre vocem plorantis. que est iIIiterata
per Priscianum, scilicet cum dicit 'ploro', non attendimus quod vere ploret, vel doleat,
nec per hoc acquirit nobis dolorem vel pietatem, sed per secundum, scilicet per vocem
iIIiteratam plorantis acquirit nobis dolorem sive pietatem. IlIa enim vox significat quod
in ipso est dolor, ut dolor, hoc est secundum veritatem doloris, et non solum
cogitationem, et hunc affectum qui est dolor denotat in eo esse ut verum affectum, non
ut simulatum, hoc est non secundum similitudinem tantum existentem in anima, vel
secundum cogitationem, sed secundum rem et actu. Et hoc est significare per modum
affectus. Unde et ista vox plorantis in hoc non differt ab interiectione grammatica,
scilicet in significando per modum affectus, immo in hoc convenit et esset hec vox
iIIiterata plorantis pars orationis que est interiectio, nisi propter hoc quod non est vox
Iitterata nec significans ad placitum, immo naturaliter. Interiectio vero que est pars
orationis, et est vox Iitterata et etiam significat ad placitum." (f. 49rb)
22Sicut dicit Remigius: "Doleo quod est verbum, significat dolorem, et heu significat
dolorem, sed differenter, quia heu significat dolorem prout dolor est vere in proferente
hanc interiectionem heu, et non ut solum est huiusmodi dolor in cogitatione
proferentis. Et hoc habet a sua impositione, et hoc etiam representat modus proferendi
talem vocem(?). Sed doleD significat ips urn dolorem ut cogitatum in proferente vel
conceptum, quod idem est, et non per modum per quem vere est in proferente, sed
solum secundum quod est in eius cogitatione." (f. 49ra)
23Sicut dicit Remigius: "Affectus, quando vere est in proferente facit ipsum subito loqui
et sine deliberatione ... Hee autem proprietates, scilicet significare affectus ut vere est
affectus et subito loqui et absque deliberatione et huiusmodi, sequuntur significatum
interiectionis, et non sequuntur alias partes orationis. Ideo ipsa sola significat per
modum affectus, idest significat affectum ut vere est affectus, quia ut in fieri et in
exercitio in proferente eum." (f. 49rb)
24De interiectione: "Ex hiis igitur patet quod interiectio significat conceptum et
affectum, et quomodo utrumque ex hiis, et igitur patet quid sit significare affectum.
Hoc enim nichil aliud est quam rem que secundum se sit apud animam, sine
deliberatione precedente, per vocem exprimere." (V, f. 88vb = C, f. 224vb)
25Robert Kilwardby, Commentaire sur Priscien Mineur (ad XVII, 21): "Dicendum etiam
ad secundum obiectum quod interiectio uno modo significat mentis conceptum, scilicet
quantum ad audientem; quantum autem ad proferentem mentis affectum, et ideo potest
esse pars orationis. Vel dicendum quod omnis pars, et interiectio et aliae significant
mentis conceptum. sed aliae partes exprimunt ipsum per modum conceptus, sola autem
interiectio per modum affectus, et ideo dicitur affectum significare et aliae partes
conceptum" (MS Vat. lat. 298, f. 9rb); Sophismata grammaticalia, sophisme 0
magister te non legente parisius dicendum est ve scolaribus: "Vel aliter dicendum quod
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 239

proposent que I'interjection soit dite signifier un affect pour Ie Iocuteur,

et un concept pour I'auditeur: 26 ils estiment en effet que Ie Iocuteur a
bien cette reaction non deliberee et immediate a I' affect, qui s' exprime
dans I'emission de I'interjection; cet affect, cependant, est r~u de
maniere tout a fait rationnelle et deliberee par l'auditeur, et, par
consequent, sur Ie mode du concept. Cette position est souvent citee et
critiquee: 27 on en invoque notamment une consequence absurde seion
laquelle l'interjection ne serait pas une partie du discours pour Ie
locuteur, mais seulement pour l'auditeur.28

significare mentis affectum est dupliciter: vel in comparatione ad audientem vel in

comparatione ad proferentem; et licet interiectio significat affectum in comparatione ad
proferentem, tamen significat mentis <conceptum> quantum ad audientem; et ita potest
<esse> pars orationis." (Z, f. 143 va)
26Sophisme Modio vini ad denarium ve ilti qui non habet: "Ad primum argumentum
quod est contra, dicendum quod interiectio dupliciter consideratur, aut per
comparationem ad pacientem et sic dicit affectum, aut per comparationem ad
apprehendentem passionem extra, et sic interiectio notat conceptum. Et sic supponit
argumentum. Ad secundum patet solutio. Dicimus quod passio denotata per
interiectionem comparata ad subiectum in quo est dicit impetum animi et ab ista parte
non construitur. Comparata vero ad andientem apprehenditur sine impetu ab intellectu."
(MS London BL Royal 8 A VI, f. 39va)
27 De interiectione: "Ad obiectum dicendum quod interiectio non solum significat
affectum mentis, sed mentis conceptum, et hoc non solum in audiente, sed in
proferente ..... (V, f. 88va = C, f. 224vb); Sicut dicit Remigius: "Alii aliter respondent,
scilicet quod interiectio affectum signiflcat in proferente, et tamen cum hoc significat
conceptum, sed in auditore, et ita non solum significat affectum, sed etiam conceptum,
et ideo non concludit argumentum. Iterum aliter dicitur ad idem argumentum, quod licet
in principio significet affectum, tamen significat post conceptum, et ideo significat
conceptum. Sed contra huiusmodi responsiones est hoc, scilicet quod cum interiectio
secundum P<etrum> <Helie> sit vox imposita ad significandum, ideo necessario prius
concepta fuit in mente imponentis, et ita conceptum significat in proferente. Hoc etiam
patet per hoc quod diverse sunt apud diversos interiectiones. Oportet ergo quod primo ab
imponente concipiatur, et ita significat conceptum in proferente, deinde natura vel ...
imponatur, et ultimo proferatur. Sic ergo patet quod a parte proferentis significat
interiectio conceptum, et ita significat conceptum in proferente, quod est contra primam
horum duarum responsionem ..... (f. 49rb); Pseudo-Kilwardby, Commentaire sur
Priscien Majeur: "Dicunt tamen 'aliqui' quod interiectio dicitur significare conceptum
in comparatione ad audientem, affectum in comparatione ad proferentem. Tamen istud
non videtur sufficiens, quia in comparatione ad proferentem utrumque convenit ei, quod
constat quia proferens vocem interiectionalem non significat aliquid nisi quod prius
animo concepit, et sic significat conceptum in comparatione ad proferentem." (A, f.
l04rb =C, f. 215rb)
28Pseudo-Johannes Ie Rus, Questio de distinctione partium orationis: "Pars orationis est
dictio mentis conceptum significans. Ex hoc arguitur quod si interiectio non significat
mentis conceptum non erit pars orationis. Prima vera propositio a Prisciano habetur.
Si dicat quod interiectio significat aliquod quod est affectum et postea fit conceptum et
sic erit pars orationis, contra. Si primo loco significat ut affectum, ergo imposita fuit
ad significandum per modum affectus et ergo ex sua impositione non erit pars
orationis. Iterum si diceret quod significat affectum proferentis et significat conceptum
audientis, quia audiens per vocem prolatam concipit affectum proferentis, tunc
sequeretur quod quantum ad proferentem non esset pars orationis, et quantum ad
audientem esset, et hoc est inconveniens." (f. 125ra)
240 IRENE ROSIER

c) Enfin, certains avancent que signifier sur Ie mode de I'affect

a
equivaut signifier sur Ie mode du bon, alors que signifier sur Ie mode
a
du concept revient signifier sur Ie mode du vraL On trouve une
premiere version de cette position dans un traite bien anterieur aux
textes dont nous parlons, Ie Tractatus de Proprietatibus Sermonum, ou
elle apparait d'ailleurs compatible avec la premiere analyse que 1'0n a
citee. Dne chose peut en effet, dit l'auteur anonyme, affecter soit la
a
virtus apprehensiva, se rapportant aux sens, l'imagination ou a
l'intellect, et etre con¥ue sur Ie mode du vrai, soit la virtus motiva, qui
comprend les vertus concupiscible, irascible et rationnelle, et etre
con¥ue sur Ie mode du bon.29 Cette meme solution est retenue dans Ie
commentaire sur Priscien Majeur du Ps-Kilwardby,30 en invoquant
egalement les deux puissances de I'arne, la virtus cognitiva et la virtus
affectiva. 31 L'anonyme cite les deux premieres interpretations que nous
avons exposees, ce qui pourrait faire penser que ce texte est posterieur a
ceux que nous avons cites. 32 Signalons enfin que Ie meme auteur dit de
I'interjection, ce que Roger Bacon enon¥ait des syncat6goremes: ce sont
pour ainsi dire des instruments par l'intermMiaire desquels nous
effectuons des actes (interiectiones enim sunt quasi instrumenta quibus
actus huiusmodi exercentur33 ). L'accent mis sur l'exercice de l'acte.
actus exercitus, dans ce texte, est tres net.

29Ed. de Rijk, Logica modernorum II 2, Assen: Van Gorcum 1967, pp. 708-9.
30Pseudo-Kilwardby, Commentaire sur Priscien Majeur: "Significare per modum
conceptus est significare aliquid apprehensum per modum veri. Significare per modum
affectus est significare aliquid apprehensum per modum boni, vel eius contrarii,
prosperi vel adversi." (A, f. 105va = C, f. 215va)
31 Pseudo-Kilwardby, Commentaire sur Priscien Majeur: "Voces sunt institute ad
significandum ea que sunt anima. Ea vero que sunt in anima dupliciter sunt, secundum
quod duplex est potentia anime, scilicet cognitiva et affectiva. Quare ea que sunt in
anima aut sunt in ea solum per cognitionem et sic perficiunt cognitivam potentiam,
aut sunt in ea per affectionem, et sic perficiunt affectivam potentiam. Que primo modo
sunt in ea sunt tantummodo similitudines et intentiones. Que secundo modo sunt in ea
sunt veritates rerum." (A, f. 105va = C, f. 215vb)
32Pseudo-Kilwardby, Commentaire sur Priscien Majeur: "Dicunt autem 'quidam' quod
dictio dicitur significare per modum conceptus, quando mud quod significat solum est
in anima secundum sui speciem, significare per modum affectus quando mud quod
significat est in anima secundum veritatem. Sola enim veritas rei per suam presentiam
apud animam existens earn movet motu doloris aut tristicie, gaudii vel admirationis,
qui bus anima habet moveri et affici secundum veritatem. Hee enim sunt passibiles
qualitates animi. Alii vero dicunt quod significare per modum conceptus est significare
aJiquid mediante iuditio rationis deliberantis. Cum enim ratio aIiquid apprehendit
mediante sensu, et post apprehensionem deliberat et iudicat de apprehenso, et
consequenter mud indicat vel significat per vocem. [Alii dicunt add. C] tunc ilIa vox
dicitur significare conceptum per modum conceptus, hoc est mediante iuditio rationis
deliberantis. Sed significare conceptum per modum affectus est significare sine tali
iuditio rationis deliberantis, quod accidit cum aIiquis apprehendit prosperum vel
adversum vehemens, ut mortem patris, vel aliquod magnum bonum, statim ex
vehementi motu iIIius prosperi vel adversi [apprehensi add. C] afficitur animus, in
tantum quod absque iuditio et deliberatione rationis prorumpit in vocem exclamationis
significativam vehementis gaudii vel doloris ..." (A, f. 105va =C, f. 215va)
33Pseudo-Kilwardby, Commentaire sur Priscien Majeur: "Dicendum quod interiectio de
se significat intellectum incomplexum. Significat enim circumstantiam doloris aut
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 241

6. Les idees qui se degageaient de l'anecdote de I'enfant qui pleure, relatee

par I'auteur du Sicut dicit Remigius, sont systematisees de maniere plus
explicite par Roger Bacon. Celui-ci explique, Ii diverses reprises, qu'il
existe en fait non pas deux, mais trois manieres de reagir linguistiquement
Ii un affect exterieur. L'interjection occupe une place tout Ii fait particuliere,
sur Ie plan semiotique, en ce qu' elle constitue un intermediaire entre les
signes naturels et les signes conventionnels: les premiers correspondent Ii
I'expression naturelle et sans deliberation des affects, tels les
gemissements, les seconds correspondent Ii l' expression raisonnee et
deliberee des concepts. Lorsqu'on profere une interjection, il y a
deliberation, mais celle-ci est imparfaite. Ainsi, par rapport Ii une douleur
ressentie, l'on peut reagir soit de maniere purement affective en gemissant,
soit apres raisonnement, en disantje souffre, soit encore Ii la suite d'une
deliberation imparfaite, en utilisant une interjection, telle aie!34 On note un
certain flottement dans les expressions utilisees pour caracteriser
l'interjection: parfois on oppose signification du concept Ii signification de
l'affect, c'est ce que l'on trouve dans les textes les plus anciens comme Ie
Tractatus de Proprietatibus Sermonum 35 ou encore dans Ie "Guide de
l'etudiant"36 ou Ie Tractatus de grammatica, du Pseudo-Grosseteste,37

timoris, gaudii vel admirationis animi, in comparatione ad rem verbi et ad actum

exercitum per quem intelligitur actus verbi. Sicut enim per actum vocandi exercitum
intelligitur verbum vocandi, sic per actum doloris vel admirationis exercite
intelliguntur verba dolendi et admirandi. interiectiones enim sunt quasi instrumenta
quibus actus huiusmodi exercentur." (A, f. 106ra =C, f. 216ra)
34Roger Bacon, Communia naturalium: "Conceptus autem mentis est duplex, scilicet,
sine deliberatione omni, et sic est in vocibus significantibus naturaliter tam hominum
quam brutorum, concipiunt enim et apprehendunt ac anima cognoscente capiunt
tristabile vel delectabile, sed non deliberant in hoc conceptu, sed subito et instinctu
naturali afficiuntur ilIis in dolorem vel gaudium vel huiusmodi; et prevalet affectus
super conceptum, unde magis afficiuntur quam concipiant, quia conceptus transit subito
in affectum. Et secundum hoc sunt agentia naturaliter. Sed postquam homo in conceptu
suo deliberat, potest facere hoc dupliciter, scilicet, vel inperfecte, et tunc profert
interiectiones dolendi et gaudendi, et alias, vel perfecte, et tunc profert voces aliarum
partes orationis. Vnde gemitus tripliciter significatur, vel ut infirmus gemit, et tunc
non est ilia vox pars orationis, vel per interiectionem gemendi, aut per hoc nomen
gemitus, vel hoc verbum gemo, vel hoc participium gemens et hiis modis est pars
orationis, sed secundum quod est interiectio habet vocem absconditam et inperfectam et
informem, quia inperfectus est conceptus, et inperfecta deliberatio, et affectus vincit
conceptum ..." (~d. Steele, Opera II, Oxford, pp. 109-11); De Signis, ~d. Fredborg &
al., Traditio 34, 1978, pp. 83-5; Compendium Studii Theoiogiae, M. T. Maloney,
Brill 1968, p. 62, par. 42 (voir la note de Maloney sur ce passage).
35Ed. De Rijk, op.cite (n. 29): "Dividitur autem vox significativa per se in vocem
significantem conceptum et vocem significantem affectum." (p. 708: 18-19) " ... Et
general iter interiectiones significant affectiones cogitantium dol oris vel gaudii vel
admirationis et non ut apprehensum vel cogitatum. Conceptum autem significant
dictiones que significant aliquid ut apprehensum et cogitatum in anima. Vt hec dictio
'homo'." (p. 708:29-709:2)
36MS Barcelone Ripoll 109: "Interiectio dicit vehementem affectum anime, cum ratio
subcumbit." Vne reponse donn~e plus haut dans Ie texte est plus precise: "Interiectio
significat anime affectum in ratione tamen, qua intellectus invenit se circa ilIum
affectum in concipiendo vocem iIIi affectui similem sub eodem significato. Vnde
differenter significatur affectus per adverbium optandi et per interiectionem. Per
242 IRENE ROSIER

avec la distinction entre significare affectum et significare conceptum;

parfois, et cette position developpe la precedente, on ajoute qu'un terrne
impose ad placitum signifie toujours sur Ie mode du concept, et
l'opposition devient alors entre significare affectus per modum conceptus
et significare conceptum per modum conceptus;38 parfois enfin, on affirrne
que la signification implique toujours qu'il y ait passage de l'affect au
concept, mais qu'elle peut se realiser soit sur Ie mode du concept, so it sur
Ie mode de l'affect, d'ou l'altemative entre significare conceptum per
modum affectus, c'est-a-dire de maniere violente et impulsive, et
significare conceptum per modum conceptus.3 9 Les differents passages
que Roger Bacon consacre a cette question temoignent d'une certaine
hesitation. 40 C'est Ie cas egalement pour l'auteur du Sicut dicit Remigius,
bien qu'il semble finalement preferer la demiere solution, selon laquelle
l'interjection et les autres parties du discours s'opposent en fonction des
modes de la signification, per modum affectus ou per modum conceptus,
tout en signifiant toujours un concept. 41

En fait ces flottements terrninologiques sont tout a fait revelateurs de

l'option originale implicite de tous ces auteurs, qui est exposee de maniere
tres claire par Roger Bacon dans son De Signis. En privilegiant l'acte de

adverbium enim significatur ut conceptus solum, per interiectionem vero in ratione qua
afficit. Unde ibi est ut affectus et conceptus." (f. 142 vb)
37Texte cite supra, note 20: I'auteur y reprend litteralement la definition de Donat.
38Cf. Albert Ie Grand, In Perihermeneias I tr.2, c. 1: " ... interiectiones ... affectum
significant non ut affectum sed ut conceptum affectus voce litterata designatam."
39Sicut dicit Remigius: f. SOra, S6r.
40Dans la Summa grammatica (ed. Steele, Opera XV, Oxford), notre auteur conclut que
I'interjection signifie per modum affectus (p. 96: 34-5); cependant, lorsque
I'argumentation porte sur Ie proces qui conduit 11 la proferation effective d'une
interjection, il tend 11 meUre en avant la signification directe de I' affect (cf. par exemple,
p. 99: 10 ff., 11 propos de la construction d'une interjection avec un cas a parte ante: " ...
Origo interiectionis est ut audita aut visa a1iquo delectabili vel tristabili extrinsecus
intus afficiatur, et subito proferatur interiectio; quare affectus subitus primum
principium est sermonis in proferente; cum ergo contra naturam affectus est ut aliquid
precedat istum affectum ex parte ipsius proferentis, et iste affectus designetur per
interiectionem, contra naturam interiectionis erit ut aliquid in sermone proferentis
precedat ipsum, et ita non debet interiectio construi cum aliquo a parte ante, quare nec
cum nominativo" (= argument du quod non, admis dans la solution). Dans la Summa
de Sophismatibus, Communia Naturalium et Ie Compendium, Bacon dit que
I'interjection signifie per modum affectus, alors que dans Ie De signis, elle signifie per
modum conceptus, licet inpeifecti (par. 9).
41Sicut dicit Remigius: "Anima igitur concipiens vel cogitans dolorem, vel aliquam
huiusmodi passionem, aut vult significare passionem hanc conceptam per modum
conceptus, idest ut cogitatam, et non prout secundum rem est in anima, et tunc anima
dicit dolor vel doleo. Et ideo significant huiusmodi partes conceptum per modum
conceptus. Si autem velit anima significare ipsam dolorem, ut vere est dolor, et ilium
affectum, ut vere est affectus, et secundum rem, et prout presentialiter est in anima, et
non simulatorie, hoc est non in sola cogitatione, sic significabit ipsum dolorem per
interiectionem, dicendo heu ! vel huiusmodi. Et ideo interiectio significat conceptum
per modum affectus, et sic est pars orationis. et etiam specialis pars distincta contra
alias." (f. 49rb-49va) "Interiectio autem est a1terius generis quam a1ie partes, quia alium
modum significandi habet. Partes autem penes modos significandi distinguuntur." (f.
SOrb)
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 243

signifier, ils ont du mal it separer la proferation de l'enonce, correspondant

it l'utilisation des signes, de l'institution de ces memes signes. La difficulte
rencontree it propos de I'interjection est precisement it l'intersection de ces
deux lignes d' argumentation. En effet, si I' on se situe du point de vue de la
proferation de l'interjection, on constate qu' elle correspond it la reception
d'un affect qui, des qu'il devient concept, s'exprime de maniere subite et
violente,42 sans deliberation, marquant ainsi - je cite - que "Ia raison
succombe it l'affect" (ratio subcumbens affectionibus43 ). Dans ce cadre,
l'interjection est plus proche des signes naturels d'affect,44 que sont les
gemissements, les rires, les pleurs, les cris des animaux,45 ainsi que des
exclamations ou des transcriptions d'onomatopees, tous ces signes etant
intentionnels, bien que non imposes. 46 On signale d'ailleurs, pour
developper sa proximite avec les signes naturels, que l'interjection est
souvent monosyllabique, et qu'elle presente une certaine universalite,
puisqu'on constate, par exemple, que les memes voyelles se retrouvent,
dans les differentes langues, pour exprimer tel ou tel affect. Cependant, si
1'0n se situe du point de vue de l'imposition des signes, il est clair que
l'interjection, signe conventionnel, ad placitum, s'apparente davantage aux
autres parties du discours: comme elles, et it la difference des signes
naturels,47 elle resulte d'une triple operation de conception, de deliberation

420n utilise souvent la fonnule prorumpit in vocem, pour rendre compte de l'expression
violente de I'affect, ou de l'affect transfonne en concept (cf. citations notes 32, 50, 58).
On peut en rapprocher ce passage des Regulae d' Augustin, dans lequell'auteur refuse
aux interjections Ie statut de parties du discours: "Interiectio non pars orationis est, sed
affectio erumpentis animi in vocem, et significat aut laetitiam, ut evax, aut
amaritudinem ut heu, ... ergo quot sunt perturbati animi motus, tot voces reddunt, et
vocantur interiectiones, quod interrumpant orationem, ...... Grammatici Latini, ed. H.
Keil, Leipzig: Teubner 1855-80 (GLK) V, p. 524:9-12. On developpe une idee voisine
a partir de l'etymologie Isidorienne de interiectio, ainsi dite parce qu'elle interrompt
(interiacet, interrumpat) Ie discours. Augustin utilise litteralement I'expression
prorumpit in vocem dans son De Dialectica, ed. B. Darell Jackson, Dordrecht: Reidel
1975, p. 90, et dans Ie De magistro (5, 13), pour marquer Ie passage de ce qui est
d'abord dans I'lime vers l'exterieur, celui de l'expression des dicibilia en dictiones.
43Sicut dicit Remigius, f. 50rb.
44Pseudo-Grosseteste, Tractatus de grammatica, ed. Reichl, op. et texte cite (n. 20), p.
59:26-28, repris litteralement dans Ie De lnteriectione: "Oppositum videtur per quendam
Philosophum dicentem quod interiectio est pars proxima hiis que naturaliter significant,
in hoc differens quod natura non dat nomen affectui, sicut in naturaliter
significantibus." (V, f. 89rb = C, f. 225va)
45 A ce sujet, voir l'importante etude de U. Eco, R. Lambertini, C. Marmo, A.
Tabarroni, "On animal Language in the Medieval Classification of Signs", Versus
38/39, 1984, pp. 3-38, repris dans Eco & Marmo, On the medieval theory of signs,
Benjamins 1990, ainsi que A. Tabarroni, "On Articulation and Animal Language in
Ancient Linguistic Theory", Versus 50/51, 1988, pp. 103-21.
46n est essentiel de distinguer les signes naturels, qui sont signes de ce qu'ils signifient
par une relation naturelle d'inference, a partir de leur essence meme (Ia fumee est Ie
signe du feu), et les signes produits naturellement, qui, comme les signes
conventionnels, procMent d'une intention (Ies gemissements). lis s'en distinguent par
Ie fail d'etre produits sans deliberation ni choix. Sur ce point, cf. en particulier Ie De
signis, op.cite (n. 34), par. 3 et ff.
47La discussion de !'interjection est singulierement compliquee par Ie fait que Priscien
parle de trois types d'interjections: I'interjection proprement dite (avec toutes les
difficuItes de sa separation d'avec l'adverbe), les exclamations, signes naturels mais
244 IRENE ROSIER

et d'imposition. Dans cette perspective on distingue Ie caractere naturel de

l'affect du mode conventionnel ou ad placitum de sa signification.48 Selon
I'auteur du De Interiectione, c'est Ie fait que ces operations soient faites
simultanement, et non successivement, qui caracterise l'interjection par
rapport aux autres parties du discours.49 L'altemative est c1airement
exprimee par Ie Pseudo-Kilwardby, dans son commentaire sur Priscien
Majeur. II oppose l'utilisation des signes par Ie Iocuteur (prolatio) et leur
imposition (institutio, impositio). Ce n'est qu'en se situant du point de vue
de I'enonciateur, dit-il, que l'on peut dire que Ia "raison succombe a
I'affect".50 L'auteur anonyme du De Interiectione, dans Ie meme ordre

susceptibles d'~tre ecrits, et les voces correspondant It des imitations de sons non
susceptibles d' ~tre ecrites, comme ha ha, etc. (cf. XV, 40). Sur cette classification et Ie
commentaire qu'en donne Pierre Helie, cf. J. Pinborg, "Interjektionen und Naturlaute,
Petrus Heliae und ein Problem der antiken und mittelalterlichen Sprachphilosophie",
Classica et Mediaevalia 22, 1961. La distinction entre vox Iitterata (= pouvant ~tre
ecrite) et vox articulata (= possedant une signification), permet It Priscien d'avoir une
categorie particuliere pour les signes intentionnels non susceptibles d'~tre ecrits: .....
quaedam, quae non possunt scribi, intelleguntur tamen, ut sibili hominum et gemitus:
hae enim voces, quamvis sensum aliquem significent proferentis eas, scribi tamen non
possunt ... " (GLK II, I, I)
48Pseudo-Kilwardby, Commentaire sur Priscien Majeur: "Unde etsi [et non A]
inquantum est pars orationis significet mentis affectum naturalem, non tamen
naturaliter sed ex institutione vel ad placitum." (A, f. l06ra = C, f. 216rb)
49De Interiectione: "Et dicendum quod <interiectio> ab impositione significat. Anima
enim rationalis simul in conceptione ipsi concepto nomen inponit. Unde simul
concipit et deliberat - si qua sit ibi deliberativa - et imponit et profert." (V, f. 89rb, C,
f. 225va) " ... Ad tertium dicendum quod quedam significantium ab impositione
significant per modum affectus et per modum conceptus. De hiis que significant per
modum conceptus est hoc verum quod supponit, scilicet prius rem concipi, et secundo
circa earn fieri deliberationem, et tertia fieri impositionem. De hiis vero que per
modum affectus, non, quia ibi solum simul vel adminus repente et absque
sensibili distinctione concipitur, imponitur et profertur." (V, f. 89rb, C, f. 225vb)
50Pseudo-Kilwardby, Commentaire sur Priscien Majeur: "Nono queritur - supposito
quod interiectio significat affectum mentis et per modum affectus, causatum [causato C]
ab aliquo extrinsecus apprehenso, per se loquendo incomplexum, non eundem omnino
cum affectum verbi quod in ea intelligitur - queritur utrum sit verum quod communiter
dicitur, quod interiectio significat affectum mentis subcumbente ratione et dominante
sensualitate. Et videtur quod sic, quia interiectio significat eo modo quo profertur...
Contra. Interiectio est pars orationis, sed omnis pars orationis significat quicquid
significat ex institutione. Ergo et interiectio. Sed institutio non est nisi per actum
rationis deliberantis et iudicantis post deliberationem. Ergo secundum hoc interiectio
significabit quicquid significabit ratione deliberante et iudicante. Sed quandocumque
ratio iudicat et deliberat, dominatur rationalitas et subcumbit sensualitas. Ergo
secundum hoc interiectio quicquid significat significat ratione dominante et sensualitate
subcumbente. Dicendum quod vox interiectionalis potest considerari dupliciter. Uno
modo prout refertur ad significatum et modum significandi, quem habet ex institutione,
et sic significat ratione dominante et sensualitate subcumbente. Et sic procedit ultima
ratio. Vel potest considerari in comparatione ad utentes voce interiectionali. IIli autem
utuntur ea quandoque ex vehementi motu prosperi vel adversi apprehensi subito,
sensualitate dominante et ratione subcumbente. Aliquando enim ex vehementi motu
prosperi vel adversi ratio que est coniuncta sensualitati quasi profertur intra vel saltern
non cohibet sensualitatem, nec moderatur earn in suo motu, sicut patet in
apprehensione valde tristis ut in morte patris vel alicuis excellentis boni prosperi, et sic
prorumpit homo in voce que revera illud significat ex institutione ad quod
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 245

d'idees, ne peut justifier l'opinion d'Isidore de Seville selon laquelle

l'interjection signifie naturellement, qu' en expliquant qu' elle se rapporte
uniquement a l'ordre de la prolatio. 51 Roger Bacon privilegie clairement la
premiere option, celIe de I'utilisation intentionnelle des signes par Ie
proferens, alors que les Modistes negligeront cet aspect, en se situant dans·
une problematique de l'imposition. L'auteur du De Interiectione et celui du
Sicut dicit Remigius tentent d'articuler les deux points de vue, celui de la
prolatio et celui de l'impositio. De ce fait la deliberation imparfaite qui
caracterise l'interjection s'applique tantot a son enonciation, en reponse a
un affect, tan tot a son imposition comme partie du discours signifiant ad
placitum. 52 La solution canonique, qui s'imposera petit a petit a partir de
Robert Kilwardby, puis dans les textes Modistes, selon laquelle
l'interjection signifie per modum affectus et s'oppose aux autres parties qui
signifent per modum conceptus,53 ne gardera de ces hesitations et des ces
enthousiastes discussions que quelques rares echanges d'arguments. 54 On
relevera que l'exemple favori, dans Ie corpus Modiste, pour illustrer la
difference entre signifie et modes de signifier, et Ie fait que des parties du
discours differentes peuvent avoir Ie meme signifie, c'est-a-dire signifier la
meme chose sur des modes differents, est precisement celui de la serie
formee sur Ie signifie de la douleur: doleo, dolens, dolet, heu. Ceci indique

significandum homo utitur quasi naturaliter et naturalitate determinante let naturalitate

determinante: quia ratione succumbente et sensualitate dominante C], et sic probantur
rationes ad partem primam facte." (A, f. \06va-vb, C, f. 217ra-rb) Voir aussi pp. 50-65
de I'edition Fredborg et aI., Cahiers de l'Institut du Moyen-Age Cree et Latin 15, 1975.
51 De interieetione: " ... ad primum quod dicit Isidorus, significando quod interiectiones
significant naturaliter, respexit ad modum sue prolationis, et ad communionem earum
cum vocibus naturaliter significantibus." (V, f. 89rb, C, f. 225va) Sur cette notion de
prolatio, vide aussi infra n. 90.
52De Interieetione: "Adhuc lsi add. V] in a1iis <i.e. partibus orationis> est deliberata
prolatio et impositio, sed in interiectione non." (V, f. 89ra, C, f. 225rb) Sieut
dicit Remigius: "Ad ultimum dicendum est quod nostre interiectiones que sub parte
orationis reponuntur, ut heu et pape ad placitum significant, non autem alie que
significant affectum causatum ab intra ut planctus infirmorum et risus. Et quod
obicitur, dicendum quod ilia tria, scilicet rem primo concipere, sive, quod idem est,
intellectu suo earn capere, et secundo deliberare et iudicare sub qua voce et sub qua
proprietate significanda sit, et tertio, earn ad significandum imponere, sunt in
impositione interiectionis, licet minus proprie quam in impositione aliarum partium.
Cum enim imponitur interiectio ad significandum, ratio concipit rem suam per
affectum cui imponitur, sive ad quam significandum imponitur. Sed hec ratio non est
perfecta, nec plene iudicare potest, nec etiam deliberare." (f. 49vb)
53J. Pinborg, "Introduction to the edition of the Commentary on 'Priscianus Maior'
ascribed to Robert Kilwardby", Cahiers de l'lnstitut du Moyen-Age Cree et Latin 15,
1975, p. 8 ff. et Die Entwieklung der Spraehtheorie im Mittelalter, Aschendorff 1967,
p. 50 ff.
54Cf. par exemple Martin de Dacie, Modi signijieandi, ed. H. Roos, Corpus
Philosophorum Danieorum Medii Aevi (CPhD) II, Copenhague, pp. 18-19 et 83-5. On
notera cependant la persistance de I' expression subito et sine deliberatione dans certains
textes Modistes; cf. Boece de Dacie, Modi signifieandi, ed. J. Pinborg & H. Roos,
C P hD IV, Copenhague, p. 296: 40-6. Pour une raison que nous ne savons pas
expliquer, les Modistes admettent que I'interjection puisse constituer une reaction a une
cause interieure, ce que I'ensemble de nos textes s'entendent pour refuser. Selon eux en
effet, c' est Ie fait que I' affect soit exterieur, et non interieur, qui caracterise I'interjection
par rapport aux gemissements ou au cris.
246 IRENE ROSIER

que la question de l'interjection a du jouer un certain rOle dans l'evolution

de la doctrine des modi significandi.

L'interjection est interessante precisement parce que la description de

son utilisation par un locuteur se laisse difficilement rectuire a une pure
actualisation de proprietes intrinseques conferees lors de l'imposition
premiere - ce qui correspond au principe de la semantique Modiste. C'est
un point que developpent a loisir nos grammairiens: I'interjection n'a pas
d'accentuation fixe, elle est une vox incondita, comme Ie dit Donat.
L'accentuation de I'interjection semble donc liee immediatement a la
"disposition" particuliere du locuteur, ala maniere dont il reagit a I'affect,
en accentuant une syllabe plutot qu'une autre. 55 Le commentaire sur Ie De
accentibus du Ps-Priscien, attribue a Robert Kilwardby est tres clair sur ce
point.56 II est interessant de noter d'ailleurs que les Modistes, qui rejettent
la problematique centree sur l'enonciation, nient que l'interjection puisse
avoir une accentuation libre.

7. On constate une evolution sensible dans les definitions de I'interjection

que donnent nos grammairiens, liee a leur lecture des textes philosophiques
aristoteliciens. Les discussions sur Ie point de savoir si les differentes
interjections correspondent aux affects des vertus concupiscible, irascible
et rationelle de l'ame, ou plutot aux parties de I'ame, en temoignent. 57
L'auteur du Sicut dicit Remigius dit clairement que l'affect a l'origine des
interjections conceme exclusivement la partie rationnelle de l'ame, en tant
qu'elle succombe aux affects des autres parties de l'ame. Nos auteurs
expliquent souvent que les termes passio, affectus, conceptus, motus
designent differents moments d'une meme reception de l'affeet: celui-ci est

55ef. Priscien, et Ie commentaire de Pierre Helie, dans Pinborg, op. cite (n. 47), p. 122.
56Robert Kilwardby, commentaire sur Ie De accentibus, ed. P.O. Lewry, Medieval
Studies 50, 1988 (Ad VII, 48): "Ad primum, dicendum est quod hoc quod dicit, non
seruant certos accentus, hoc est intelligendum quando proferuntur sine deliberatione,
scilicet quando significant affectus subito prolatos et sine deliberatione animi.
Aliquando autem significant affectus prolatos cum deliberatione, et quando sic
significant et sic proferuntur cum deliberatione, tunc possunt retinere certum accentum
in medio vel in fine ... Vel aliter potest dici quod hoc quod dicit, interiectiones non
habent certos accentus, hoc est quod non habent accentus determinatos circa aliquem
locum, scilicet in fine tantum vel in medio, set nunc in medio, nunc in fine,
secundum diuersas dispositiones ipsius proferentis. Penes hoc enim quod
ipse proferens afficitur, magis faciendo moram in proferendo supra unam sillabam quam
supra alteram, vel magis deprimendo unam quam alteram, sive sit media, sive sit
ultima, penes hoc, scilicet, diversificatur accentus interiectionis ita quod non habet
ipsum certum, nec in certo loco. Sed hoc quod dicit, accentum habent in medio vel in
fine, hoc intelligendum est ex dispositione ipsius afficientis et proferentis
potest enim sic disponi quod elevet sillabam mediam vel ultimam indifferenter vel
deprimat." (p. 184, par. 278 et 279) L'auteur propose en fait, on Ie voit, deux
solutions: la premiere, refuse 11 l'interjection Ie caractere de vox incondita, Ia seconde
reconna]'t ce caractere, qui confere lui une certaine malleabilite, en lui permettant de
marquer I'intention du locuteur.
57Pseudo-Grosseteste, Tractatus de grammatica, pp. 59-60; De Interiectione (V, f. 89rb=
C, f. 225vb); Sicut dicit Remigius (f. 51 ra ff). Sur cette question, et plus
generalement sur les sources philosophiques de la distinction conceptuslaJfectus et des
classifications des interjections, cf. notre article 11 para]'tre dans Histoire, Epistem%gie,
Langage: "Interjections et expression des affects dans la semantique du xm e siecle".
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 247

dit passion du point de vue de sa reception, concept en tant qu'il est

apprehende, affect quand il dispose a l'action, acte ou mouvement Iorsque
I'arne agit effectivement du fait de sa reception. 58 Ces moments sont
importants pour comprendre que, a partir d'une vision passive de
l'apprehension, on puisse en arriver a une conception active de l'activite
intellectuelle et du Iangage.

8. Les particularites semantiques de l'interjection permettent d'expliquer

celles de sa construction. C'est en fonction que l'actus exercitus que
l'interjection heu peut se construire avec Ie datif mihi, ce cas indiquant Ie
terme de l'acte exerce, de meme qu'un autre verbe peut regir un cas oblique
en vertu de l'acte signifie. Dans un tel contexte est introduite Ia distinction
entre deux modes de Ia transitivite, transitio actus exercitus et transitio·
actus significatus, qui correspond au seul usage commun de notre
distinction dans Ie corpus des Modistes.59

On recourt frequemment a une paraphrase qui retablit vocalement Ie

verbe correspondant a I'acte exerce pour rendre compte des constructions.
Ainsi, si l'on dit ecce homo, avec un substantif au nominatif, c'est parce
que homo indique Ie sujet d'un acte exerce, d'ou Ia glose, "voici I'homme

58De interiectione: "Sic intelligendum est, cum aliquid extrinsecum delectabile, aut
tristabile, aut mirabile, agit in aliquem, facit in iIIo passionem, que quidem passio
afficit animam pereipientis, et est ibi affectus eius sicut gaudium, aut tristitia, aut
admiratio. Cum autem sic afficitur anima patiendo ab extrinseco movente simul vel
repente suscipitur a ratione affectus, et in ipso instanti conceptus sine deliberatione
preambula prorumpit in vocem, et ideo [non C] voce incondita sive indisposita, et non
regulariter accentuata profertur interiectio. Et sic fit eius prolatio. Ideo conceptum per
modum affectus dat intelligere. Sic igitur patet quod ex actione extrinseca causatur
passio, que dicitur affectus, et ipse affectus a ratione concipitur, et significatur per
vocem." (C, f. 88vb, V, f. 224vb) Pseudo-Kilwardby, Commentaire sur Priscien
Majeur: "Et notandum quod significatio interiectionis potest diei passio, conceptus,
affectus et motus. Sed dicitur passio secundum quod ab extrinseco [extrinsecus C]
apprehenso prospero vel adverso subito causatur ... Dicitur autem conceptus. in
quantum apprehenditur per modum veri, sed affectus inquantum apprehenditur per
modum boni vel eius contrarii. Dicitur autem motus [modus P] affectus [effective P] in
quantum inclinat animarn ad actum verbi. Sed intellectus dieitur communiter et in se,
et etiam in quantum per vocem representatur. Voces enim sunt signa intellectuum...
Notandum tamen quod ista quatuor ordinem habent naturalem, ut prius sit passio,
secundo conceptus, tertio affectus, quarto motus; ut dicatur passio in comparatione ad
recipientem passionem ... conceptus in comparatione ad potentiam apprehendentem,
affectus id idem in quantum inclinat ad ·motum, sed motus secundum quod est in
exercitio. Prius enim movet delectabile vel triste, ex cuius motu derelinquitur passio.
Secundo apprehenditur ex cuius apprehensione derelinquitur conceptus. Tertium ipsum
conceptum afficit. Quarto movet. Et potest diei quod intellectus essentialiter dictus non
differt a conceptu." (A, f. l04rb, C, f. 215rb)
59Cf. Thomas d'Erfurt, Grammatica speculativa, ~d. O.L. Bursill-Hall, London:
Longman 1971, pp. 190, 252; Radulphus Brito, Quaestiones super Priscianum
Minorem, ~d. J. Pinborg & H. von Enders, Stuttgart: Frommann-Holzboog 1980:
"Actus significatus est qui importatur per verba vel partieipia ut lego legens. Sed actus
exereitus sive exereitatus est qui realiter exercetur per prolationem adverbii vocandi
quasi in eius modum significandi cadens. Illud ergo adverbium significat rem suam per
modum dependentis sub ratione excitantis vel vocantis seu actus exereitati." (p. 355)
248 IRENE ROSIER

qui court" (Ecce homo currit}.60 En fait l'on rencontre de nombreuses

divergences touchant a la maniere d'integrer la notion d'actus exercitus
dans un modele syntaxique. Elles apparaissent notamment dans les
sophismes abordant la construction de l'adverbe de vocation, 0, avec Ie
vocatif.61 Plusieurs difficultes sont soulevees dans ces discussions; II
s'agit en premier lieu de savoir si l'on peut admettre la construction ou la
rection d'un terme par un autre qui n'est pas exprime, et qui est donc
indetermine. Dans Ie cas de l'adverbe demonstratif ecce, il est possible de
restituer Ie verbe sous-entendu, puisque, dans la situation, la deixis ou
demonstratio s'exerce sans equivoque, ce qui permet de retrouver Ie verbe
sous-entendu: on voit bien si Ie sujet court ou s'il assoit. 62 Ce probleme
est particulierement aigu lorsque c'est en vertu de l'actus exercitus que
semble etre regi tel ou tel cas, puisque comme Ie note Roger Bacon: "actus
exercitus non est pars orationis".63 II fait d'ailleurs exactement la meme
remarque, dans ses Summule logicales, a propos de l'operation logique
exercee par un syncategoreme, qui peut etre fonction de l'actus exercitus
d'un terme. 64 Une des regles communement admises par les grammairiens

61lRobert Kilwardby, Sophismata grammaticalia; Sophisme: "Ecce homo nudus pedes

qui venit de ultra parvum pontem factum sibi sotulares ... Ad primum dicendum quod
hec est oratio perfecta, quia hoc adverbium ecce demonstrativum est substantie sub
aJiquo actu qui exercetur ab ilia substantia, ut si currat Sor, vel sedeat, demonstratur
sub cursu vel sub sessione. Unde sensus est 'ecce homo sedet', vel 'currit', et gratia
demonstrationis appropriatur actus ad ilium actum quem exercet substantia. Unde non
construitur hoc nomen homo cum adverbio demonstrandi sed construitur ilia substantia
demonstrata cum actu sub quo demonstratur. Et si ilIe actus exigat nominativum, tuitc
est ilia substantia in nominativo, ut cum dicitur ecce homo legit vel sedet etc. Et etiam
competenter dicitur ecce quem diligo et ordinatur hoc adverbium ecce cum accusativo
non per sui naturam, sed per naturam diversorum actuum sub quibus demonstratur
substantia." (2, f. 153ra) Voir aussi Roger Bacon, Summa grammatica, op.cite (n. 40),
p.186.
61Cf. I. Rosier, op.cite (n. 4), pp. 46 ff; Sophisme 0 magister, ibid., pp. 90-1.
62Robert Kilwardby, Sophismata grammaticalia, sophisme: "Ecce homo nudus pedes
qui venit de ultra parvum pontem factu sibi sotulares; Solutio ... hoc adverbium ecce
demonstrativum est substantie sub aJiquo actu qui exercetur ab ilia substantia, ut si
currat Sor vel sedeat demonstratur sub cursu vel sub sessione. Unde sensus est 'ecce
homo sedet', vel 'currit', et gratia demonstrationis appropriatur actus ad ilium actum
quem exercet substantia. Unde non construitur hoc nomen homo cum adverbio
demonstrandi, sed construitur ilia substantia demonstrata cum actu sub quo
demonstratur. Et si ille actus exigat nominativum, tunc est ilia substantia in
nominativo, ut cum dicitur ecce homo legit vel sedet, etc. Et etiam competenter dicitur
ecce quem diligo et ordinatur hoc adverbium ecce cum accusativo." (2 f. 153va) " ... Ad
aliud dicendum, quod licet hoc adverbium ecce ut est demonstrativum secundum
habitum, indifferenter se habeat ad plures actus, quia est aptum natum substantiam
demonstrare sub diversis accidentibus et sic nullum actum determinat, sed prout est
demonstrativum in actu, gratia demonstrationis presentis appropriat sive determinat
acturn presentem exerciturn a substantia demonstratur." (2 f. 154va-vb)
63Cf. Summa grammatica, op.cite (n. 40), p. 105:13 et l'ensemble de cette question, it
propos de la construction de 0 socii. Cf. aussi Pseudo-Iohannes Ie Rus, f. 130va ff, qui
y consacre une question: "Queritur de constructione vocativi et primo de constructione
ipsius cum verbo, an reddat ei suppositum. Secundo an aJiqua sit constructio de
adverbio vocandi cum actu exercito. Tertio an sit aJiqua <constructio> ipsius cum
vocativo in se."
 ACTUS EXERCITUS ET ACTUS SIGNIFlCATUS 249

veut que !'interjection regisse, en vertu de l'actus exercitus, Ie meme cas

que celui qui serait regi par I' actus significatus correspondant.65 On trouve
la meme idee dans les traites logiques contemporains: un syncategoreme
agit en fonction d'un actus exercitus semblable It celui qu'exprimerait
l'actus significatus correspondant.66 Ainsi Roger Bacon et Guillaume de
Sherwood expliquent-ils que Ie syncategoreme si correspond It l'acte
signifie sequitur. 67 L'on doit, en second lieu, decider s'il existe toujours
une expression vocale possible des "actes exerces", et si, en d'autres
termes, on peut trouver une paraphrase linguistique exacte de ces actes de
langage, sous forme d'actus significatus: la sequence heu me equivaut-elle
par exemple It dolor afficit me et pape It miror? 68

9. L'on a jusqu'ici rencontre plusieurs types d'expressions linguistiques

qui permettent au locuteur d'exercer un acte: I'interjection, qui sert It
appeler,l'adverbe demonstratif, qui permet de designer. Les grammairiens
remarquent de nombreux autres actes de lang age, et il est frappant de
constater It quel point l'interet pour de telles expressions disparait presque
totalement chez les Modistes. Les plus evidents parmi ces actes sont la
deixis (demonstratio) et l'anaphore (relatio). On citera quelques exemples
plus originaux, qui correspondent It des sophismata se rencontrant dans
diverses collections.

a) Un acte interessant est celui de la glose metalinguistique realisee

au moyen de l'expression verbi gratia, comme dans l'exemple Homo

64Roger Bacon, Summulae dialectices, op.cite (n. 16): "Et quamvis actus exercitus non
possit esse pars orationis, quia omnis actus, qui pars est orationis, est significatus,
tamen bene potest alia pars orationis ad talem actum referri, cum actualiter exerceatur
per aliquam partem orationis, sicut hic per condicionem, vel disiunctionem, vel
copulationem." (p. 258, par. 344) Voir aussi p. 247, par. 232.
65Roger Bacon, Summa grammatica, op.cite (n. 40): "Interiectio eundem casum ratione
actus exerciti exigit quem idem actus significatus." (p. 109:16-18)
66Roger Bacon, Summule dialectices, op.cite (n. 16), p. 247, par. 231, p. 241, par.
178. Cf. aussi Ie sophisme 0 magister, op.cite (n. 4): 2-1-1 "Item, sicut hoc quod est 0
est dictio per quam exercetur actus vocandi, sic hec dictio tantum est dictio per quam
exercetur actus excludendi, et similiter omnis distribuendi. Sed huiusmodi alie dictiones
non determinant sibi aliquod casuale, quare similiter nec hoc adverbium 0, quare melius
dicetur 0 magister quam 0 magistrum. Sed constat quod 0 non construitur cum hoc
quod est magistrum. Quare similiter nec cum hoc quod est magister." (p. 84) ... Ad 2-
1-2 "Ad secundum argumentum dicendum quod hoc quod est 0, quia in iIIo casuali erat
de suo modo siguificandi aliquid quod representet significatum huius dictionis, ideo in
quolibet casuali equaliter potuit reperire dictio exclusiva iIIud quod sufficit ad eius
officium, ut misereor tantum Sortis, congaudeo tantum Sorti, etc." (p. 93)
67Roger Bacon, Summulae dialectices, op.cite (n. 16), pp. 241-2, par. 178; Guillaume
de Sherwood, Syncategoremata, ed. O'Donnell, Medieval Studies 3,1941, p. 79.
68Sicut dicit Remigius: "Natura ilia per quam interiectio dat intelligere verbum est
convenientia inter ilia, quod patet inter pape et miror et sic de aliis intelligendo. Pape
enim significat actum anime vel passionem, quod idem est, et iIIe actus est admiratio,
ut ita dicat et idem significat miror. Item pape significat iIIud ut est in actu et in fieri,
sed tamen per modum affectus, ut dictum est. Similiter miror prout in fieri, per modum
conceptus." (f. 49va) On se reportera au passage de la Summa de Pierre Helie, cite par
Pinborg, op. cite (n. 47), p. 119-20, pour apprecier I'evolution de la terminoiogie. Cf.
aussi Sicut dicit Remigius, f. 51ra, et 0 magister, op.cite, p. 91.
250 IRENE ROSIER

currit verbi gratia filius Platonis. 69 L'adverbe scilicet peut avoir la

meme fonction, de meme que Ie verbe dico, dans un enonce comme
Marcus currit, Marcus dico Tullius.1 o L'analyse en termes d'ellipse, ou
plutot de restitution linguistique du constructible correspondant It l'acte
exerce rencontre des difficulres.
b) Un acte souvent cite est celui de la numeration, que l'on realise
par l'intermediaire de la sequence Unus, duo, tres. Alexandre de
Villedieu recourt It la categorie de "nominatif absolu" pour un tel
enonce.7 1 II est clair, explique Ie Pseudo-Johannes Ie Rus, dans Ie
sophisme qui prend cet exemple pour point de depart, que celui qui
compte ne cherche pas It signifier, It designer quelque chose, mais It
agir, It effectuer l'acte de numeration. 72 Dans plusieurs passages
extremement interessants, l'auteur identifie l'intention de signifier du
locuteur avec son affect et distingue deux types d'actes linguistiques.
Celui qui agit par Ie moyen du langage (agens per sermonem) peut en
effet soit agir en lui-meme (apud se ipsum), comme lorsqu'il compte,
soit agir sur autrui. Cette distinction lui permet d'expliquer qu'il est
inutile de chercher It restituer un verbe sous-entendu dans la sequence
Unus duo tres, alors qu'on peut Ie faire dans les actes du second type,
comme lorsqu'on utilise un vocatif, par exemple, 0 magister, en sous-
entendant les verbes de perception audi ou intelligi.73 L'auteur du Sicut

69Pseudo-Johannes Ie Rus (f. 144vb). On trouve disc ute en plusieurs endroits de la

collection (f. 143ra, 144vb, 150vb, etc.), des sophismes comportant l'expression verbi
gratia, ce qui renforce I'impression d'une collection composite.
70Pseudo-Johannes Ie Rus (f. 98va).
71Doctrinale, ed. D. Reichling, Berlin: Hofmann 1893, v. 1129. Cette sequence
correspond aux premiers mots du Timee de Platon, dans la traduction latine de
Chalcidius: "unus, duo tres. quartum e numero, Timae, vestro requiro ... ". Les gloses
sur Ie Doctrinale donnent plusieurs exemples de nominatif absolu, notamment d'origine
scripturaire, qui constitueront des sophismes grammaticaux (ex. In convertendo
Dominus captivitatem Sion, facti sumus sicut consolati). Cf. aussi Roger Bacon,
Summa grammatica,op.cite (n, 40), sophisme Proch dolor: "Non invenitur ab
auctoribus gramatice quod nominativus ponitur absolute nisi tripliciter: aut cum
subsequitur relatio, ut 'virga tua et baculus tuus ipsa me consolata sunt' aut in
numerando, ut dicciones numeralis 'unus duo tres' , aut in epithafiis aut super ymagines
ut 'beatus Nicholaus', 'Sancta Maria' et huiusmodi." (p. 102: 22-9) Differentes
acceptions de la notion d' absolutio sont etudiees dans C.H. Kneepkens, "ABSOLUTIO:
A note on the development of a Grammatical Notion", dans L' heritage des
grammairiens latins, Actes du colloque de Chantilly, ed. I. Rosier, Louvain: Peeters
1988, pp. 155-69.
72Idee semblable dans la glose Admirantes sur Ie Doctrinale It propos du meme exemp1e,
texte dans Thurot, Extraits de divers manuscrits latins, pour servir Ii I' histoire des
doctrines grammaticales au Moyen Age, Paris: Bibliothi:que Imperiale 1869, pp. 268-
70. On rappellera que la glose Admirantes est proche, sur Ie plan doctrinal, de Robert
Kilwardby, et que nous avons trouve plusieurs concordances litterales avec Ie
commentaire sur l'Ars Maior 11/ attribue It Kilwardby, ed. L. Schmiicker, Siidtirol:
Weger Brixen 1984.
73Ps-Johannes Ie Rus, Sophisme Verbi gratia: "Intentio enim proferentis et affectus est
idem. Intentio enim proferentis non exprimitur nisi cum affectus eius exprimitur ... Ad
hoc dicendum quod dicta oratio <Verbi gratia> est imperfecta simpliciter et rationibus
predictis. Sed tamen intelligendum quod con venit agere per sermon em
multipliciter. Intentio enim agentis per sermonem potest esse agere apud se ipsum
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 251

dicit Remigius rapporte, en citant ce m~me exemple au cours d'une

discussion sur l'interjection composee proch dolor, I'opinion de ceux
qui divisent les enonces selon qu'ils signifient un "concept de l'fune"
(conceptum anime), comme les enonces declaratifs, un "affect de l'fune"
(affectum anime), teis heu me etproch dolor, ou encore une "action de
l'fune" (actum anime), par exemple, unus duo tres.?4 On voit ici que
certains appliquaient la distinction per modum conceptus et per modum
affectus non seulement aux mots, pour opposer l'interjection aux autres
parties du discours, mais egalement aux enonces, avec l'idee sous-
jacente que I'assertion n'etait pas un acte de langage. On notera, sans
pouvoir s'y attarder, que la locution proch dolor, fort prisee dans les
recueils de sophismes grammaticaux, est difficile precisement parce
qu'elle semble presenter une redondance (nugatio), puisqu'elle se
compose d'une interjection signifiant la douleur sur Ie mode de l'affect,
proch, et d'un substantif signifiant la douleur sur Ie mode du concept,
dolor. 75 On repond communement que, puisque les modes de

tantum, ut cum aliquis intendit numerare, dicens unus, duo etc. Vel potest intentio
agentis per sermonem agere cum alio. Si vero intentio alicuis sit agere per sermonem
cum se ipso, sufficit in iIIo sermone quod [qui, cod.1 significat istud quod intenditur,
sicut si intendit numerare sufficit uti nomine significante numerum sine verbo. Si sit
intentio alicuius agere mediante sermone cum alio, non potest agere perfecte sine verba
elSi intelligendum de hoc." (f. 143rb) Sophisme Unus duo tres : "Ad primum dicendum
quod cum dicitur unus duo tres, qui sic agit demonstrare non intendit, sed
numerare tan tum. Nomina autem designantia numerum sufficiunt ad numerandum,
non enim intendit aliquid enumerare de numeratis. Propter hoc non est indigentia verbi.
Unde dicendum quod collectio istorum nominum unus duo tres perfecte orationis
unitatem habent quantum ad agentem, non quantum ad demonstrationem actionis." (f.
I 46va)
74Sicut dicit Remigius, sophisme Proch dolor (f. 56rb); sophisme, 0 magister:
"Quidam ad precedentem orationem respondent dicentes quod congrua est et perfecta et
ad sensum, quia ibi ponilur sub propriis signis ea que debet habere talis oratio, scilicet
id quod significat affectum vocantis, et iIIud circa quod est excitatio. Primum est 0,
secundum est magister in vocativo. Et dicunt quod in dicta oratione non est querendum
appositum et suppositum, sed in orationibus que significant conceptum anime, ut
verum et falsum. Unde ponunt talem diffinitionem, orationum quedam
significant anime conceptum, ut Sor currit, quedam affectum, ut heu me,
proch dolor, 0 Willelme, quedam anime aetum ut unus duo tres, que orationes
significant actum numerandi, vel ut est hic, scilicet quod ista oratio que sequitur actum
anime ut Nominativo hic magister, que oratio significat actum declinandi. Dicunt igitur
quod ad perfectionem orationum que significant anime conceptum primo modo queritur
vel requiritur suppositum et appositum cui inest copula, non autem ad perfectionem
aliarum, quia in aliis ponitur constructibilia, sed tanquam instrumenta" (f. 81 va)
75Cf. Sicut dicit Remigius, sophisme Proch dolor quia magister non disputat (f. 55rb
ff.), Pseudo-Johannes Ie Rus, sophisme Proch dolor meus socius optimus scolarium
istius civitatis frangitur crura in veniendo de ultra parvum pontem (f. 138rb); Roger
Bacon, Summa grammatica, op.cite (n. 40), sophisme Proch dolor 0 socii, p. 95; De
interiectione (V f. 9Ovb); Robert Kilwardby, sophisme Proch dolor 0 socii quia socius
noster frantigur crura: "In anima aliquid potesi esse dupliciter, sicut res, et sicut
species. Sicut species lapidis est in anima et non lapis, scientia vero est in anima sicut
res, similiter dolor potest <esse> in anima dupliciter, sicut res et sicut species. Et cum
est sicut species, sic significatur per hoc nomen dolor. Qui enim apprehendit dolorem,
non propter hoc dolet. Sed cum est dolor sicut res, sic significatur per interiectionem.
Sed simpliciter potest esse dolor in anima sicut res et sicut species, scilicet cum aliquis
dolet et apprehendit dolorem. [argument admis, sauf pour sa conclusion, selon laquelle
252 IRENE ROSIER

signification de la douleur different, il n'y a pas redoublement inutile

d'une meme signification, et que, de la juxtaposition de ces deux
modes, resulte un renforcement permettant une expression plus vive de
la douleur. 76

c) Certains grammamens distinguent deux types d'expressions

correspondant a des actes exerces. Les premieres, parmi lesquelles se
trouvent les interjections, les vocatifs, les imperatifs ou les
demonstratifs, sont telles que c'est en vertu de leur signification ou de
leur consignification qu'elles peuvent exercer un acte. II s'agit d'une de
leurs proprietes intrinseques. D'autres expressions, par contre, exercent
un acte, sans que rien ne les y predispose, mais uniquement par Ie fait
que Ie locuteur decide de les utiliser pour effectuer celui-ci. L'auteur
anonyme du Sicut dicit Remigius donne comme exemple, a cote de
l'enonce de la numeration, unus duo tres, dont on a deja parle, celui que
les enfants utisent pour decliner, Nominativo 'hie magister' ,77 que I' on
peut gloser comme: "au nominatif, on dit ou on decline hie magister".78
Un autre cas communement etudie, emprunte a Priscien (GLK III, p.
62:20), est celui de la sequence bene, bene. En elle-meme, elle n'est
constituee que de la repetition d'un adverbe, mais prise dans Ie contexte
d'une enonciation particuliere, elle a une valeur d'acte.79 Par exemple,
si quelqu'un est en train de parler, et que l'on dit bene bene, cela
signifie quelque chose comme, 'tu parIes bien'. A titre d'anecdote, on
notera un certain sadisme chez nos grammairiens, qui choisissent des
contextes oil une personne est en train d'en frapper une autre, bene bene
constituant un encouragement a poursuivre l'action!80 On pensera bien

proch dolor serait une dictio compositaj Solutio... Et notandum quod dolor potes!
significari sicut affectus solum, vel sicut conceptus, vel utroque modo. Si significetur
sicut affectus, sic significatur per hanc interiectionem heu sive per hanc interiectionem
proch. Si significatur ut conceptus, sic significatur per hoc nomen dolor. Primo ergo
modo significatur per hanc orationem heu mihi, secundo modo significatur per hanc
ego doleo, utroque modo per hanc proch dolor. Quare ergo ille qui profert hanc
orationem afficitur dolore et intendit quod dolor concipiatur ab audiente, propter hoc
oportet significare dolorem utroque modo, scilicet sicut affectum et sicut conceptum."
(nO 17, Z, f. 157rb-157va)
76Roger Bacon, Summa grammatica, p. 100: 8-12.
77Cf. Donat, Ars Minor, GLK IV, p. 356-29:356:2.
78Cf. Robert Kilwardby, Sophismata grammaticalia, n° I; Roger Bacon, Summa
grammatica, p. 165; Magister Durandus, Summa quidam de grammatica (f. 168rb). Cet
exemple est egalement etudie dans les traites anterieurs: on y releve en particulier Ie fait
que nominativo a un usage non referentiel (materialiter ponitur), cf. par exemple,
Quaestiol1es Victorinae, ed. De Rijk, op.cite (n. 29), II, 2, p. 739, Robert de Paris,
Summa "breve sit", ed. C.H. Kneepkens, Het judicium Constructionis. Het Leerstuk
van de Constructio in de 2de Helft van de 12de Eeuw, 1987, del. III, pp. 94 ff, 217 ff.
79Sicut dicit Remigius: "Sed bene bene et huiusmodi non dant intelligere actum per
suam consignificationem, sed quia utimur iIIis adverbiis loquendo ad
alium." (f. 47va)
80Cf. Roger Bacon, Summa grammatica, sophisme 0 socii: "Ad aJiud, dicendum quod
actus exercitus potest exerceri, aut per partem actualem orationis in qua ponitur
adverbium, sive per ipsum adverbium, et tunc refertur ad ipsum, quamvis enim non sit
pars orationis actus talis, tamen ad ipsum ordinatur et refertur pars orationis, quia
actualiter habetur et exercetur per aliquid quod est pars oracionis actualis; sic autem est
hic, quia exercetur per ips urn adverbium et vocativum. Si vero non exercetur per
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 253

sur a la ferule avec laquelle est generalement represente Ie maitre de

grammaire.

L'auteur du Sicut dicit Remigius, utilise l'expression d"'enonces

enclitiques" pour ce type de sequences, qui permettent d'effectuer un
acte du fait de leur enonciation, et non pas du fait d'une propriete de
signification. Dans la terminologie grammaticale, les enclitiques sont
des mots comme -que, -cum qui n'ont pas d'accent en eux-m~mes,
mais, qui, joints a d'autres forment une sorte de compose, un du point
de vue accentuel, par exemple dominusque ou nobiscum. On comprend
bien l'analogie, puisque l'interpretation des sequences qui nous
occupent (ex. bene bene) ne provient pas du sens qu'elles ont en elles-
m~mes: elles ne peuvent ~tre comprises que dans leur relation au
contexte particulier dans lequel elles sont proferees. 81 Comme
l'explique Robert Kilwardby dans son commentaire sur Ie De accentu,
les enclitiques constituent une exception aux regles accentuelles
ordinaires, exception qui peut ~tre justifiee linguistiquement,82 tout
comme dans notre texte, les enclitiques constituent une infraction aux
regles syntaxico-semantiques, qui peut ~tre expliquee, comme dans Ie
cas des figures de construction, par la necessite de l'expression d'un
sens particulier. Roger Bacon, qui distingue de maniere analogue les
deux types d'actes exerces, ajoute une particularite de construction des

aliquam partem oracionis in qua ponitur adverbium, set per aliud extra, tunc necessario
debet addi actus significatus vel intelligi, et non refertur nec ordinatur adverbium ad
actum exercitum, sicut cum dicitur verberanti 'bene, bene', qui quidem actus significat
ex discrecione sermonis ad exercentem actum verberandi intelligitur. Vel, si non placet
sic dicere, dicatur quod per actum exercitum intelligatur actus significatus cum
simili ad quem refertur adverbium, et huis simile manifestum est in Logicis, de
negacione condicionalium et disjunctivarum et copulativarum." (p. 106: 17 ff.) Sur la
demiere reference, cf. Summule dialectices, ed. de Libera, op.cite (n. 16), p. 242, 247,
etc. M~me reference ~ I'acte de frapper dans Ie sophisme 0 magister, ed. Rosier, op.cite
(n. 4), p. 79 ou dans Ie sophisme du m~me intitule de Siger de Courtrai, ed. J.
Pinborg, Amsterdam: Benjamins 1977, p. 66.
81Sicut dicit Remigius: "Forte tamen verius est dicere quod orationes [omnes MS]
encietice sunt que ad acturn exerciturn pertinent, nisi iIle actus exercitus sit
consignificatio et per consignificationem ipsius vocis intellectus, ut vocatio
consignificata per vocativa dat intelligere audi vel [ibi cod.] percipe exercita et
consignificata per ecce dat intelligere quod deest. Sed bene bene et huiusmodi non dant
intelligere actum per suam consignificationem sed quia utimur ilIis adverbiis loquendo
ad alium. Similiter nominativo non significat declinationem, licet eo ut instrumento
declinemus. Unde per suam significationem non·dat intelligere declinationem, ideo
encletica est. Et breviter, quandoque ita est quod id quod deest intelligitur per actum
exercitum qui non est consignificatum sermonis actualiter positi in oratione, ut fit hic
unus duo etc., Ii unus enim non significat numerum in exercitio, sicut heu significat
vel consignificat dolorem in afficiendo, et ideo dicendo heu est oratio omnino perfecta
secundum intellectum, et non est encletica, quia determinate dat intelligere doleo, sed
non sic unus duo etc., suum verbum dant intelligere, immo ita potest intelligi est aut
currit aut aliud." (f. 47va)
82Notons que Ie Pseudo-Priscien n'utilise pas Ie terme enclitique en cet endroit; Ie
commentaire porte sur Ie lemme NECESSITAS PRONUNTIATIONIS (GLK III, p.
520:36). II y est question des trois cas possibles d'infraction aux regles de
prononciation, les enclitiques constituant un de ceux-ci ~ c6te de I'ambigiiite, et de la
necessire de distinguer un mot d'un autre: Lewry, op.cite (n. 54), pp. 142 et 149-150.
254 IRENE ROSIER

expressions que Ie Sicut dicit Remigius designe comme enclitiques, it

propos de bene bene: du fait que l'acte exerce ne peut etre compris que
dans un contexte particulier, il est tout it fait possible d'exprimer
vocalement Ie verbe correspondant it l'acte exerce, ce qui correspond
par exemple it 'bene verbera'.83 Sur ce point encore, la proximite du
Sicut dicit Remigius avec les positions de Roger Bacon est frappante.

d) Plusieurs actes linguistiques renvoient it un univers de discours

particulier, celui de la liturgie. Ite missa est, par ex empIe, permet
d'annoncer que la messe est terminee, tout comme par In nomine Patris,
etc. s' exerce la benediction. On citera egalement I' annonce de la lecture,
realisee par l'enonce: Sequentia sancti evangelii secundum Matheum,
dans lequel on peut comprendre: hec verba sunt Sequentia sancti, etc. II
s'agit d'un contexte ou sont prononces frequemment des enonces
incomplets ou elliptiques, qui, grace aux modalites d'un rituel connu de
tous, sont parfaitement interpretables. 84

e) On doit enfin mentionner, en dehors du corpus grammatical cette

fois-ci, un ensemble de textes dans lequel la distinction entre actus
exercitus et actus signijicatus occupe une place essentielle: il s'agit de
l'analyse des formules sacramentaires par les theologiens. Nous y
avons consacre une etude particuliere, et nous ne nous y etendrons pas
ici. 85 L'efficacite, et donc Ie caractere performatif de ces formules est
inscrite dans leur definition-meme: •sacramentum id efficit quod
figurat'. Cette caracteristique d' efficacite est souvent invoquee par les
theologiens pour distinguer les enonces sacramentaires des enonces
ordinaires: les premiers ont une fonction "factive" ou "operative", alors
que les seconds ont une simple fonction "significative". La discussion
sur cette double valeur apparait it propos de I'analyse de la formule de la
consecration Hoc est corpus meum. Pour certains auteurs, Ie
demonstratif hoc realise une demonstratio ut concepta: la formule
correspond it une citation des paroles du Christ, it un discours rapporte,
Ie demonstratif n'ayant pas une valeur referentielle. Pour d'autres, iI
s'agit d'une demonstratio ut exercita, la deixis, ou demonstratio,
s'exer~ant effectivement, puisque Ie pronom permet de designer une
chose - chose dont la determination fait d'ailleurs egalement I'objet de
vives discussions. L'opposition entre demonstratio ut exercita et
demonstratio ut significata se rencontre egalement chez les
grammairiens du milieu du XIIIe siecle.

10. Apres avoir passe en revue quelques types de sequences exprimant un

actus exercitus il nous faut maintenant montrer it quel point elles constituent
une piece essentielle dans un systeme global, et non pas un ensemble de
remarques eparses et marginales. Ce systeme a pour objet l'etablissement
des diverses conditions de correction et de completude des enonces. 11

83Roger Bacon, Summa grammatica, op.cite (n. 40), p. 107:3-14. Cf. la solution
semblable du Pseudo·Johannes Ie Rus (f. 130va).
84Cf. I. Rosier et B. Roy, "Grammaire et liturgie dans les sophismes du XIIIe siecle",
Vivarium, XVIII 2, 1990, pp. 118·35.
851. Rosier, "Signes et sacrements. Thomas d' Aquin et la grammaire speculative",
Revue des Sciences Philosophiques et TMologiques, 1990,392·436.
 ACfUS EXERCITUS ET ACfUS SIGNIFICATUS 255

s'agit de rendre compte de tous les enonces acceptables, et non pas

seulement de ceux qui presentent la structure canonique sujet-predicat.
Parmi les enonces deviants se trouvent principalement ceux qui comportent
une figure de construction, et ceux qui sont defectueux, elliptiques,
auxquels se ramenent une bonne partie des exemples que nous avons cites
precedemment. Les phrases comportant un sujet et un verbe, dont les
modes de signifier s'accordent, sont completes ad sensum: l'auditeur peut
les comprendre immediatement, sans effort particulier. Toutes les autres
sont completes ad intellectum: l'auditeur doit alors, pour pouvoir
interpreter correctement I'enonce, retrouver la sequence canonique
correspondante, et rechercher l'intentio proferentis qui a rendu l'ecart
necessaire. La revendication la plus claire des auteurs dont nous nous
occupons, est celle qu'enonce Robert Kilwardby, et que reprend Roger
Bacon: l'enonce Ie plus correct est celui qui correspond Ie plus
adequatement a l'intention de signifier du locuteur. 86 Telle est la seule
mesure de la correction de l'enonce. La figure turba ruunt, comme l'enonce
enclitique bene bene ne sont pas corrects simpliciter, mais seulement
secundum quid, c'est-a-dire, en fonction du sens vise par Ie locuteur
(intellectus intentus). L'auteur du Sicut dicit Remigius, en un paragraphe
tres proche, en sa lettre, d'un passage tire du commentaire sur Ie
Barbarismus attribue a Robert Kilwardby, nous en donne un exemple qui
nous ramene a la question de l'interjection. La sequence tiree de Virgile,
par I'intermediaire de Priscien, haec secum est elliptique; il manque Ie
verbe [oquebatur, puisqu'elle a Ie sens "il se disait ces choses a part soi".
Ceci s'explique, dit l'auteur, par I'etat d'emotion dans lequel se trouvait Ie
locuteur. 11 en va de meme dans tous les cas OU, affecte par un mouvement
de l'ame violent et non delibere, par une trop grande joie ou une trop
grande douleur, il ne peut composer un enonce complet: ainsi, si sa maison
briile, il ne demandera pas qu'on lui porte de I'eau, mais dira simplement
aqua, aqua ou ad ignem, ad ignem. 87 On se souvient que c'est Ie meme

8&rexte cite dans C.H. Kneepkens, "Roger Bacon on the double intellectus: A note on
the Development of the theory of congruitas and perfectio in the first half of the
thirteenth century", dans The Rise of British Logic, ed. P.O. Lewry, Toronto:
Pontifical Institute of Medieval Studies 1985, pp. 115-43; cf. I. Rosier & A. de Libera,
"Intention de signifier et engendrement du discours chez Roger Bacon", Histoire
Epistemologie Langage VIII, 2, 1986, pp. 63-79, Mary Sirtidge, "Robert Kilwardby:
Figurative constructions and the limits of grammar", dans De Ortu grammaticae, ed.
G.L. BursiJI-Hall & S. Ebbesen, Amsterdam: Benjamins 1990, pp. 321-37.
87Le passage suivant est un de ceux que I'on trouve litteralement identique dans la glose
Admirantes du Doctrinale. Robert Kilwardby, In Donati artem maiorem III, op.cite (n.
74): "Consequenter in eclipsi ponitur hoc exemplum haec secum, in quo deest hoc
verbum 'Ioquebatur' ... Causa-vero, quare oportuit sic fieri, est necessitas sententiam
exprimendi, quod intenditur; potuit enim significare, quod ipse locutus est apud se, sed
cum tanto impetu, ut vix verba informarent, et ad hoc designandum potuit esse ibi
necessitas, quare fit defectus, vel causa potest esse simpliciter motus animi violentus et
non deliberatus, sicut patet aliquando prae nimio luctu vel prae nimia tristitia, et non
potest perfecte et cum deliberatione inteUectus componere orationem perfectam, immo
imperfectam exprimit verbia gratia: in combustione domus non clamat portare aquam,
sed aquam aquam." (p. 97: 566-79) Sicut dicit Remigius: " ... Hec secum ... Et dicit
quod ibi deest dicebat, et ibi intelligitur, ex quo est sermo autenticus. Quia, licet sermo
eque de se bene reciperet unum de pluribus aliis verbis ab isto, non tamen ita bene,
secundum quod huiusmodi sermo est relatus ad actorem, ut statim dictum. Intendebat
256 IRENE ROSIER

type de reaction impulsive et non deliberee qui etait invoquee pour justifier
l'usage de l'interjection. L'auteur du Sicut dicit Remigius nous propose
alors un systeme complexe ordonnant les differents types d'enonces
complets ad intellectum, et permettant de retrouver les constituants
manquants. Dans certains cas, il suffit de considerer les proprietes de la
sequence Iinguistique, comme lorsque l'on retrouve Ie sujet ego dans Ie
verbe de premiere personne Lego; dans d'autres il faut considerer I'acte
exerce: ainsi, l'interjection heu donne it entendre Ie verbe doLeo; et il est
parfois essen tiel de se rapporter au contexte, comme dans les expressions
enclitiques d'acte exerce, du type bene bene. 88 Dans Ie meme ordre
d'idees, Robert Kilwardby explique, dans son sophisme Nominativo hie
magister, que seuls les enonces par Iesquels nous signifions quelque
chose, et donc reaiisons une assertion, ont besoin d'un nom et d'un verbe,
alors que ceux par lesquels nous agissons et effectuons une action, comme
unus duo tres ou Nominativo 'hie magister' n'ont pas cette exigence.89

On notera, sans y insister ici, la grande parente entre les theses

semantiques de Roger Bacon, developpees particulierement dans Ie De
Signis et celles de nos grammairiens. Le langage est d'abord ce qui permet
d'exprimer intentionnellement quelque chose it autrui: dicere est cum
intentione projerre,90 dit l'auteur anonyme du Sicut dicit Remigius. On
reconnait lit I' influence d' Augustin, dont notre auteur cite d' ailleurs, des les
premieres pages, la definition du signe.91 De meme que dans Ie De signis
Roger Bacon expliquait que chaque locuteur, etre raisonnable, peut
reimposer les signes, et que cette reimposition, s' effectuant dans l' instant

enim dicere actor, quod dicebat velloquebatur ad se ipsum, sed non sine rationali causa
subtacuit Ie dieebat. Causa enim quare actor tacuit Ie dieebat, fuit ut significaret actor
quod ipsa tanto dolore et tanto affectu apud se loquebatur, ut nec ipsa
posset totum sermon em suum perficere. Sicut videmus accidere illis qui
clamant: latus latus, aqua aqua, ad ignem ad ignem et similia." (f. 47ra)
88Sur ces questions, cf. I. Rosier, article cite (n. 4).
89Robert Kilwardby, Sophismata grammaticalia, Sophisme Nominativo hie magister:
"Sed quod ista oratio <Nominativo hie magister> non indigeat verbo quo ad sui
perfectionem ostenditur sic. Duplex est oratio. Est enim una per quam aliquid
enuntiamus aIteri, et est alia oratio per quam non enuntiamus aliquid alteri sed per
quam aliquid agimus sive exercemus. Talis oratio non indiget verbo, quo ad sui
perfectionem, alia vero indiget Cum ergo predicta oratio sit talis, per quam exercemus
aliquid et non per quam aliquid enuntiamus alteri, non indigebat verbo, quo ad sui
perfectionem, sicut cum dicitur quod hic non requiritur aliquod verbum, unus duo tres,
etc. Sed si perficiatur hoc erit per actum exercitum et non per actum significatum.
Similiter dicendum est de hac oratione." (n01, Z f. 137ra) On notera que Johannes
Aurifaber utilise une argumentation analogue, pres d'un siecle plus lard, dans son
sophisme de meme intituJe, "Nominativo hie magister', ed. J. Pinborg, Die
entwieklung der Spraehtheorie im Mittelalter, Beitriige zur Geschiehte der Philosophie
und Theologie des Mittelalters, 42:4, Munster 1967, p. 215.
90Sicut dicit Remigius, f. 40ra.
91Dans ce passage, l'auteur expose la definition de I'enonce, et on notera qu'il utilise
l'exemple de buba dans Ie meme sens que Roger Bacon: "Si dicam bubas, hoc quod dico
bubas non est nomen nec oratio proprie, nec supponitur nomen proprie, quia nec
significativum est proprie. Signum enim aliud a se significat, quod patet per
diffinitionem signi et est facta ab Augustino: signum est quod se ipsum demonstrat
sensui et aliquid derelinquit intellectui." (f. 4Ova) On mentionnera cependant que Bacon
critique cette definition dans son De signis, op. cite (n. 34), par. 2, p. 82.
 ACTUS EXERCrrUS ET ACTUS SIGNIFICATUS 257

de I' enonciation, donc dans un contexte donne, et en fonction des regles

precises de la translatio, n'entrave pas la communication, de meme les
grammairiens, parmi lesquels se trouve Ie Roger Bacon de la Summa
grammatica, s'interessent aux phenomenes de deviation par rapport aux
regles communes, en cherchant Ii les rapporter Ii des types de
fonctionnement regles, et Ii justifier par Ia Ie fait qu'ils puissent etre
produits et interpretes. Le parallele entre la translatio, dans Ie domaine
semantique, et lafigura, dans Ie domaine syntaxique, est explicitement
realise par Robert Kilwardby, dans ses sophismata grammaticalia.92

11. Gabriel Nuchelmans en concluant son etude sur la distinction entre

actus exercitus et actus signijicatus, citait Austin, pour dire que plusieurs
philosophes medievaux etaient tout Iifait conscients du fait qu' au moyen de
certains enonces, on faisait queIque chose en disant quelque chose. II
apparait que cette conception "pragmatique" du langage, deja relevee par
Osmund Lewry Ii propos de Robert Kilwardby, est tout a fait dominante
chez les grammairiens du milieu du XllIe siecle, ce qui explique Ie role
importantjoue par la notion d'actus exercitus, qui deviendra tres marginale
dans les traites Modistes. Comme certains logiciens de la meme epoque,
ceux-ci considerent les signes linguistiques comme des instruments nous
permettant de signifier ou d'agir, en nous-memes ou sur autrui. Les
relations formelles entre les termes pris en eux-memes, simpliciter, sont
moins importantes que les relations entre les formes prononcees et, d'une
part, l'intention de signifier qui a preside a leur production, d'autre part,
l'effet qu'elles produisent sur l'auditeur, dans Ie contexte de leur
enonciation. La solution de nombreux problemes semantiques ou
syntaxiques passe par la distinction entre deux manieres de considerer
l'enonce: d'une part per se ou absolute, en tant que sequence de
constituants, d'autre part, en relation avec son enonciation particuliere par
un locuteur, secundum quid, quantum ad intentionem proferentis, quantum
ad utentes ou encore quantum ad nos. 93 Chaque phrase presente en fait
deux niveaux distincts de completude: selon sa peifectio prima, la phrase
est conctue par rapport Ii son etre (esse), comme un tout agence en parties;
selon sa peifectio secunda, elle est envisagee dans son utilisation, comme
"instrument de la raison".94 L'on doit mettre cette position particuliere en

92Texte cite dans Rosier 1988, op.citi (n. 4), pp. 19-21.
93La distinction entre propriete intrinseque, 'naturelle' d'un tenne ou d'une sequence, et
valeur effective resultant de sa prononciation (prolatio) joue un role essentiel dans un
domaine de problemes different, celui de lafallacia compositionis et divisionis; cf. S.
Ebbesen, "Suprasegmental Phonemes in Ancient and Medieval Logic", in English
Logic and Semantics, ed. H.A.O. Braakhuis et al" Nijmegen: Ingenium 1981; A. de
Libera, "De la logique a la grammaire: Remarques sur la theorie de la determinatio chez
Roger Bacon et Lambert d' Auxerre (Lambert de Lagny)", in Bursill-Hall et aI., De Ortu
grammaticae, pp. 209-26. Voir aussi les developpements ulterieurs entre modus
significandi grammaticalis et modus significandi logicalis chez Duns Scot, les modes
logiques dependant du modus proferendi et se trouvant donc ex parte nostra; A. Maieru,
Terminologia logica della tania scolastica, Roma 1972, pp. 531 ff.
94Sicut dicit Remigius: "Sic patet quid sit perfectio orationis secundum duos modos
perfectionis •.. Comparatio autem duarum talis est, scilicet quod secunda supponit
primam, et hoc quia prima refertur ad esse primum et ad orationem, secundum quod
ipsa est totum ex partibus integrantibus, secunda vero refettur quo ad esse secundum et
quo ad esse quod habet oratio secundum quod est instrumentum rationis ... Nunquam
258 IRENE ROSIER

rapport avec la distinction importante, relevee par O. Nuchelmans, faite

dans Ie Tractatus de Proprietatibus sermonum, entre les deux acceptions du
verbe significare, selon qu'il est dit des termes (voces) ou de l'enoneiateur
(utens). L'auteur anonyme utilise comme metaphores les actions de frapper
et de seier, qui peuvent etre dites soit de l'instrument, soit de l'agent, que
l'on trouvera frequemment utilisees dans nos textes. 95

Du fait de importance de l'idee que Ie langage est un instrument, et

corrolairement de la notion d'intentio proferentis, nous proposons
d'appeler "intentionaliste" l'approche particuliere des grammairiens dont
nous avons iei etudie les textes. II est clair que ces theses ont un echo dans
Ie corpus logique, Roger Bacon en est Ie meilleur temoin. Par ailleurs, les
problemes logiques abordes par les grammairiens Ie montrent assez
clairement. Le celebre traitement de l'equivocation propose par Robert
Kilwardby dans son commentaire sur Priseien Mineur en est un
exemple.96 Et on peut eiter une demiere fois notre Sicut dicit Remigius,
dont l'auteur dit, It propos de l'enonce modal Vel/em esse bonus clericus,
qu'il ne peut etre dit vrai ou faux, bien que celui qui Ie profere dise bien
quelque chose qui est vrai ou faux. 97

La tendance qui, dans la tradition logique, vise, en priviligieant les

enonces assertifs, It renvoyer les autres types d'enonces ala rhetorique ou
ala poetique, position clairement affirmee par un Thomas d'Aquin par

enim oratio est instrumentum nisi sit disposita ap recte sententiam significare." (f.
43va) "Duplex est instrumentum, quoddam quod ordinatur ad aliquid significandum ut
oratio. Unde Plato in Thymeo: 'ad hoc datus est nobis sermo ut presto fiant mutue
voluntatis indicia'. Et aliud quod ordinatur ad aliquid operandum, ut securis ad
scindendum. Sed unumquodque instrumentum recipit perfectionem a fine. Ergo oratio
recipi[enJt perfectionem suam ab actu sue significandi. lIIe vero actus est significare
perfectam sententiam, sicut actus dictionis est significare intellectum simplicem." (f.
52va-vb)
95Tractatus de Proprietatibus Sermonum, M. de Rijk, op.cite (n. 29), p. 710:8-20. Pour
ces metaphores, cf. par exemple Ie texte de Roger Bacon cite supra n. 15.
96Kilwardby utilise une opposition analogue a celie dont il servait pour etudier
!'interjection, entre Ie point de vue du locuteur et celui de l'auditeur, cf. P.O. Lewry,
Robert Kilwardby's writings on the logica vetus, unpub. diss., Oxford 1978, p. 70 ff.
97L'auteur discute ici la question importante de savoir si un enonce complet (perfectum)
est necessairement correct. Sicut dicit Remigius: "Mendacium ilIud non est
attribuendum suo sermoni, sed sue intentioni, et est dicentium et non nominum .,.
Vel/em esse bonus clericus, cum non sit enuntiatio, non est vera vel falsa. Tamen
dicens est verus vel falsus. Nec sequitur quod si virtus est in dicente, quod sit in
oratione, nisi oratio sit talis quod possit verum vel falsum significare, et quod [que
cod.] sit inconveniens signum, quod non est nisi sit congrua et indicativi modi. Vel
potest dici quod loquendo a parte rei verum et falsum supponunt congruum, et hoc
intelligendum est quo ad bene esse et non substantialiter. Per incongruum enim posset
verum significari, sed non ita bene, ut cum dicitur turba ruunt. Loquendo vero per
comparationem ad nos, non semper supponunt, quia per industriam seu
discretionem nostram per incongruum possumus apprehendere verum vel falsum." (f.
47rb-48va) Avec la question de savoir si la virtus est dans celui qui parle, ou dans
l'enonce qu'j) protere, on touche a une question d'une importance capitale en
sacramentologie,lorsqu'on cherche a preciser les raisons de l'efficacite de la formule
sacramentelle. Cf. article cite (n. 85).
 ACTUS EXERCITUS ET ACTUS SIGNIFICATUS 259

exemple,98 rejaillit sur Ia grammaire, qui prit pour modele de l'enonce

complet Ia structure sujet-predicat. Dans ses tentatives Ies plus formalistes,
et en particulier chez Ies Modistes, on cherche a expliquer cette structure,
en termes de regles de formation et de proprietes des constituants, en
marginalisant Ies 'autres types d'enonces. Cette option, sur Ie plan
syntaxique, est corrolaire, sur Ie plan semantique, de celle qui privilegie
I'imposition: I'imposition d'un terme lui confere toutes Ies proprietes
grammaticales ou semantiques qui determineront son usage effectif, tout
comme Ies regles syntaxiques donnent de maniere exclusive toutes Ies
donnees necessaires it l'interpretation d'un enonce. Ce qui nous parait
remarquable, a l'inverse, dans l'approche "intentionaliste", c'est que Ies
grammairiens, prenant Ie point de vue de I' enonciation effective, de Ia
pro/atio, tout en maintenant une exigence formelle et en ne se Iimitant pas a
une pure analyse d'une sequence prononcee hic et nunc - en ne reduisant
donc pas celle-ci a un pur token - cherchent aposer des principes generaux
de fonctionnement des sequences linguistiques, integrant non seulement
ceux qui enoncent simplement quelque chose, mais egalement tous Ies
autres, de types fort varies comme on a pu Ie voir, qui sont Ies instruments
par l'intermediaire desquels nous effectuons des actes. On comprend, dans
une telle optique, l'importance de Ia distinction entre actus signiftcatus et
actus exercitus, que ron a vu o¢rer, tant au niveau des termes, avec Ies
exemples privilegies de I'interjection et des syncategoremes, qu'au niveau
des enonces.99

CenJreNaiondde IaRechenileSden1ifique, UA 381, UnivemtiPais V1I

98Thomas d'Aquin, In Perihermeneias, L. I,!. vii, 87.

99Si I'on ne peut encore proposer de datation absolue pour les textes dont nous avons
parle, il est possible d'en proposer une chronologie relative, assez precise, It partir des
questions etudiees. Le commentaire de Robert Kilwardby, qui date des annees 1240-45
nous foumit un terminus a quo. On sait que la Summa grammatica de Bacon en depend.
O'autres textes sont apparentes d'une maniere ou d'une autre It Robert Kilwardby, qu'i!s
apparaissent, dans les manuscrits, It la suite de ses traites (Ie De interiectione), qu'i!s lui
soient attribues, de maniere incertaine ou erronee (sophismata, commentaire sur Ie
Barbarismus, Ie De accentu et sur Priscien Majeur), ou qu'i!s en soient proches
doctrinalement (glose Admirantes). Sur la question de l'interjection,les questions du
Sicut dicit Remigius s'appuient manifestement sur l'expose du De Interiectione, qui
lui-m~me cite, d'un cOte les arguments donnes par Robert Kilwardby dans son
commentaire sur Priscien Mineur, et de I'autre, litteralement, Ie Tractatus de
grammatica du Pseudo-Grosseteste. Enfin, Ie sophisme '0 magister' contenu dans Ie
manuscrit parisien BN lat. 16135, qui presente un modele global de correction des
enonces, assez semblable It celui du Sicut dicit Remigius, semble dater des annees
1270. Tous les premiers textes que nous avons cites indiquent un milieu "anglais",
qu'il se situe d'ailleurs It l'Universite de Paris ou It celie d'Oxford. Cela ne semble pas
Ie cas, par contre, du Sicut dicit Remigius, comme nous l'avons signale en
commen~ant, ni du sophisme 0 Magister.
260 IRENE ROSIER

Appendix

Manuscrits cites

Anonyme, De Interiectione, MS V = Vat. 298, ff. 88va-91ra; MS C =

Cambridge Univ. Libr. Kk III. 20, ff. 224r-228v. L'on suit V en
indiquant lesvariantes de C uniquement 10rsqu'elles sont significatives.
Une edition de ce traite est en preparation.

Anonyme, Sicut dicit Remigius, MS BN lat. 16618

Anonyme, Sophismata, MS London, British Library, 8AVI, ff. 36ra-

46vb.

Durandus, Summa quidam de grammatica, MS Barcelone Ripoll109, f.

158ra-173ra.

Pseudo-Joannes Le Rus, Sophismes et questions suivant la Summa in arte

grammatica de Johannes Ie Rus, MS Vat.lat 7678, ff. 89ra-lOlv.

Pseudo-Kilwardby, Commentaire sur Priscien Majeur, MS A =

Cambridge Univ. Libr. Peterhouse 191, ff. lr-1l1v, et C = Cambridge
Univ. Libr. Kk II1.20, ff. 25r-224r + 228v-229v (j'ai suivi les memes
principes d'edition que ceux indiques dans I'introduction de I'edition des
extraits qui ont ete publies, p. 3+, en faisant de A Ie manuscrit de base).
Fredborg et aI., "The Commentary on 'Priscianus Maior' ascribed to
Robert Kilwardby", Cahiers de l'Institut du Moyen Age Grec et Latin 15,
1975, 1-143.

Robert Kilwardby (?), Sophismata grammaticalia, cf. note 9. Nous

utilisons Z = Zwettl, Stiftsbibliothek, cod. 338, ff. 135-161, comme
manuscrit de base.

Robert Kilwardby, Commentaire sur Priscien Mineur, MS Vat. lat. 298 -

Extraits dans: Kneepkens, C.H., 1983, "Roger Bacon on the double
intellectus: A note on the Development of the theory of congruitas and
perfectio in the first half of the thirteenth century", in The rise of British
Logic, ed. P.O. Lewry, pp. 115-43.

Principales editions citees

Anonyme, Glose Admirantes sur Ie Doctrinaie, extraits dans C. Thurot,

Extrails de divers manuscrits latins, pour servir al' histoire des doctrines
grammaticales au Moyen Age (reprint Minerva, Frankfurt 1964), Paris,
Bibliotheque Imperiale, 1869.

Anonyme, 0 Magister, MS BN lat. 16135, ed: I. Rosier, "0 MAGISTER

.. Grammaticalite et intelligibilite selon un sophisme du Xllle siec1e",
Cahiers de l'Institut du Moyen Age Grec et Latin 56, 1988, pp. 1-102.
 ACTUS EXERCITUS ET ACTUS SIGNIFlCATUS 261

Anonyme, Tractatus de proprietatibus sermonum, in Logica modernorum :

A contribution to the History of Early Terminist Logic, vol II, 2, 00. L.M.
de Rijk, Assen: Van Gorcum 1967.

Pseudo-Grosseteste, Tractatus de grammatica, ed: Tractatus de

grammatica" eine /iilschlich Robert Grosseteste zugeschriebene spekulative
Grammatik, edition und Kommentar, Karl Reichl, MUnchen 1976.

Robert Kilwardby (?), Commentaire sur Ie De accentibus, ed. P.O.

Lewry, 'Thirteenth-Century Teaching on Speech and Accentuation: Robert
Kilwardby's Commentary on De Accentibus of Pseudo-Priscian",
Medieval Studies 50, 1988, pp. 96-185.

Robert Kilwardby (?), Commentaire sur l'Ars Maior III, ed. L.

Schmuecker, Robertus Kilwardby, O.P., In Donati Artem Maiorem III,
SUdtirol, A.Weger Brixen 1984.

Robert (?) Bacon, Syncategoremata, 00. (partielle) H.A.G. Braakhuis, De

13de Eeuwse Tractaten over Syncategorematische Termen. Deel I, Ph.D.
Leiden 1979, Krips Repro Meppel, p. 105 ff.

Roger Bacon, Summa grammatica, ed. R. Steele, Opera hactenus inedita

Rogeri Baconi, fasc.15, Oxford: Clarendon Press 1940.

Roger Bacon, Summule dialectices, 00. A. de Libera, "Les Summulae

dialectices de Roger Bacon, I. De termino, II. De enuntiatione", Archives
d' Histoire Doctrinale et Litteraire du Moyen Age, Annee 1986, pp. 139-
290; et " ... illDe argumentatione", ibid., Annee 1987, pp. 171-278.

Roger Bacon, Summa de sophismatibus et distinctionibus, ed. R. Steele,

Opera hactenus inedita Rogeri Baconi XIV, Oxford.

Roger Bacon, Communia Naturalium, ed. R. Steele, Opera hactenus

inedita Rogeri Baconi II, Oxford.

Roger Bacon, De Signis, ed. K.M. Fredborg, L. Nielsen and J. Pinborg,

"An unedited Part of Roger Bacon's Opus Maius : 'De Signis"', Traditio,
34, 1978, pp. 75-136.

Roger Bacon, Compendium Studii Theologiae, ed. T. S. Maloney, Brill

1988.
 Interest mea et imperatoris castam ducere in uxorem: Can
'est' be used impersonally?
by Mary Sirridge

There is an obvious syntactic similarity between such sentences as: 1

1. It is daytime. (est dies)

2. It is between black and white. (est inter album et nigrum2 )

and:

3. It rueth me. (paenitet me)

4. There is running by me. (curritur a me)

And there seems to be a similarity of meaning as well. for in each case

there is an absence of a specific subject to whom the action of the sentence
is ascribed. But in fact. both (3) and (4) have impersonal verbs and can be
rewritten so as to make it clear that the real subject of the sentence is some
action or state.

3' . Rue affects me. (paenitentia habet me)

4'. A run is being made by me. (cursus fit a me)

Of course (1) and (2) have no such concealed reference to an action. It is

thus very tempting to suggest that they do not have true impersonal verbs
at all. but have a misleading surface grammar. that they could be less
misleadingly rewritten as:

IThe texts in this paper are taken from ms. Zwettl 338, ff. 135-161 (= Z). I have
consulted MS Seville Biblio. Capitular Columbina 5.5.9, ff. 53r-l04v (= S). I have
not noted variants unless they are used to justify divergences from Zwettl in the printed
text. The other MSS of SG are Bamberg Staatsbiblio. Misc. Astr. IH. IVI ff. 65r-
JOOv: Erfurt 80 JO ff. 47-82; Erfurt 40 220, ff. 1-38; Basel Univ. Bib. B.VIIIIA, ff.
49v-76r; Firenze, Bibl. Naz.: Conv. Soppr. 0.11045; St. Florian SB: XI.632, ff.64r-
86r. Regulae: x (S: MS = y): S has the previous text x, as in the printed text, and Z
has y; <x>: neither S nor Z has x, but it is needed; [xl: the manuscript(s) being
followed has x, but it should be deleted; x (x: S: MS am.): S has x, which Z omits; x
(MS = y: Sam.): Z has y for x in the printed text, and the text is missing in S; -x-:
deleted by the scribe in Z. I have normalized orthography in the texts, except where
classical orthography would demand ae in place of e. and where medieval forms are
constant across the MSS, e.g.• intelligere. I have usually given the full text of
passages abbreviated in the paper.
2The usual sentence is est inter canem et lupum, which presents difficulties. The first,
as Roger Bacon notes, has to do with the lack of difference in this case, so that the
meaning is best expressed by indistinctio est inter canem et lupum. Secondly, as Bacon
also notes, the whole sentence is a metaphor for est crepusculum. A slightly different
meaning for (2) is suggested by Bibliotheque National 16618, f. nv, which claims
that est inter album et nigrum is equivalent to distat inter album et nigrum.

262
 CAN 'EST' BE USED IMPERSONAlLY? 263

1'. The light of the sun is in the air.

2'. (3x)(M(x) & R[x, (black, white)])

(1') offers a "scientific" paraphrase of a colloquial sentence with

misleading surface grammar. (2') seems to show that this apparent
impersonal construction is actually existential in form; it says that among
the things there are is an intermediary between white and black. And it
might be hoped that we can find a similar paraphrase for such sentences as
ponendum est terminos. Some grammarians of the 13th century yield to the
temptation to treat such sentences as (1) and (2) in this way.3

The set of grammatical sophismata attributed to Robert Kilwardby

(KS), is an exception in this respect. It proposes several strategies to
preserve the intuition that est in such sentences functions like an active or
passive impersonal verb, while acknowledging the force of considerations
which seem to show that est cannot really function impersonally. The
official discussion of this issue occurs in the course of dealing with the
sophism Interest mea et imperatoris castam ducere in uxorem; but the issue
arises repeatedly in discussions of the other sophisms of KS which raise
problems with respect to person: 4 dicto de genere dicendum est de specie;
sillogizantem ponendum est terminos; amatus sum vel lui quia legitur
Virgilium.

The verb interest is a traditional locus of grammatical interest. This is

partly because Priscian makes special mention of its peculiarities of
construction.5 More importantly, if interest is regarded as a compound in
which est has a genuine occurrence, then discussions of interest lie, so to
speak, at a very rocky theoretical crossroads; at this point, the theory of the
meaning and function of the verb 'to be' (verbum substantivum) and the
theory of impersonal constructions intersect.

Impersonal constructions present a theoretical challenge to the

grammarian. For one thing, impersonal constructions are quite diverse,

3Robert Kilwardby, the commentary on Priscian Minor, hereafter cited as KS (MS

Cambridge Peterhouse 191 ff. 112-229 (= cod. vol. 2, ff. 1-118); Roger Bacon, Summa
Grammatica, ed. R. Steele, fasc. 15, Opera hactenus inedita Rogeri Baconi. Oxford:
Oxford University Press 1940, hereafter cited as SG; Anon. Bibl natI. 16618, 'Sicut
dicit Remigius' hereafter cited as SdR. See M. Sirridge, "Can est be used
impersonally? A Clue to the Understanding of the Verbum Substantivum", Histoire,
Epistemologie, Langage, 12, fusc. 2,1990, pp. 121-38.
4KS as a whole is consciously stnictured according to the syntactic relations among the
various parts of speech. Some time previously, the author, having dealt with
pronoun/noun and pronoun/verb constructions, has turned his attention to noun/verb
constructions, subdivided in tum into those attendant upon number (ex. Turba ruunt);
those which involve person; and those having to do with the requirement of noun in
the nominative case and finite verb for a complete sentence (ex. Urbem quam statuo
vestra est, Ecce homo qui venit, etc.) This architectonic is not found in Bacon's SG,
which takes up many of the same sophismata according to the various figurae, though
like our KS, SG ends with an ill-assorted rush of leftover cases.
5Priscianus, Institutiones Grammaticae, ed. M. Hertz, Leipzig: 1855, Grammatici
Latini, vol. II, III, hereafter cited as /G. Books XVII-XVI of IG are known as Priscian
Minor. Cf. XII.29, XVII.92.
264 MARY SIRRlDGE

and there is considerable disagreement about borderline cases. This

diversity and the fluctuating boundaries of the class make it very difficult to
arrive at some basic functional feature of all instances. The press of theory
requires that this feature be plausibly expressed by saying that the verb has
been "deprived of substance", which becomes a usual way of explaining
how it is that a third person form can be said to have undergone a privatio
personae. But the notion of a privatio substantiae is ontologically
problematic - what can it possibly mean to say that the action referred to by
a verb has been deprived of the substance in which it inheres? This
ontological problem is particularly intensely felt in connection with the verb
'to be'. Here, grammatically anomalous behavior tends to be explained by
pointing to the fact that the verb 'to be', paradigmatically in the form est,
has as its meaning that very act of existence which is common to all
substances by definition and which serves thereby as the foundation of all
combinations of inhering forms. It is by virtue of this meaning, for
example, that est takes both a preceding nominative and a subsequent
nominative form in agreement with the first. The problem, then, is that the
very theories of meaning and function used to explain the syntactic
anomalies associated with est seem to show that it can never function
impersonally. For how can the most semantically unambiguous form of the
verbum substantivum undergo aprivatio substantiae? Not surprisingly, a
fair number of theorists decide that it cannot.

1. A first theory about 'est' taken impersonally

The author of KS discusses Interest mea etc., in his usual manner,

going through the sophism sentence from front to back, raising in
succession the difficulties which might render the sentence
ungrammatica1. 6 Having determined that it is possible for verbs generally
speaking, to function impersonally - the author admits that the discussion
is long overdue - KS presents both an argument that interest in the
sophism sentence must be taken impersonally and a number of arguments
that it cannot. One of these negative arguments runs: the preposition inter
cannot render est impersonal; thus, since est is in itself personal, so is its
compound interest. Thereupon arguments are proposed to establish the
premise that est in itself cannot function impersonally, given its function
and meaning. I paraphrase the three of these which bear most directly on
the analysis of the impersonaI.7

6KS (Z 141 va; S 78va): "Circa istam orationem sex possunt queri: primum est utrum
verbum aliquod possit impersonari vel non; secundum est, supposito quod sic, utrum
hoc verbum interest possit impersonari, et gratia huius queritur utrum hoc verbum est
possit esse impersonale; tertium est utrum personalia passive vocis possint
impersonari quemadmodum (S: MS = quem ad) activa et neutra; quartum est de
constructione huius verbi interest ad hoc pronomen mea et ad hoc quod dico
imperatoris; quintum est utrum infinitivus ducere possit reddere suppositum huic verbo
interest vel cui aJteri; sextum est de hac prepositione in et de eius liZ: 141 vbll
constructione cum suo casuali." I translate impersonari and fieri impersonale "to be
rendered impersonal" or "to be used impersonally," since KS takes the view that the
impersonal is derived from the personal verb fonn by a process or action for a reason.
7KS (S 79rb; Z 147va): "Occasione huius queritur utrum hoc verbum est possit esse
impersonale, et quod non ostenditur sic, til quia significat substantiam uniuscuiusque,
et ita suus actus cum substantia est idem; sed a suo actu non potest privari, quare nee a
 CAN 'EST' BE USED IMPERSONALLY? 265

A I: est signifies the substance of anything whatsoever, and so its act

is identical with substance; since it cannot be deprived of its act, it
cannot be separated from substance and hence cannot function
impersonally.

A2: est is that element in the genus of verbs which is the first,
smallest indivisible element, as the tone is the smallest indivisible
unit of melodies; but all impersonal verbs are divisible by exposition,
i.e., into the specification of the thing designated by verb (res verbi)8
and fit; hence est cannot be used impersonally.

A3: est signifies composition, but this composition has no meaning

without the elements of composition, i.e., suppositum and
appositum. Thus est cannot be deprived of a suppositum; hence est
cannot function impersonally.

Those who hold that est cannot be rendered impersonal accept these
arguments, our author says. He disagrees, however, and must therefore
rebut the arguments. Indeed, his solution is composed wholly of a series
of answers to them.

The point underlying Al is surely that the very thing or res verbi
designated by forms of to be, and pre-eminently by est, is substantial
being, i.e., that very act or realization by which a substance is at all, as
opposed to some mode of action like running or reading, which it might
easily lack while remaining in existence as the substance it is. Our author
answers that what is designated by est can indeed not be wholly separated
from substance, in which it inheres, "for that [substance] is identical with
the thing designated by the verb." But it can be separated from substance in
which it inheres as that in which it inheres. When we say sum faciens

sua substantia. [ii] Item in quolibet genere ut vult Aristoteles in Posterioribus est
ponere unum primum minimum indivisibile, ut in melodiis tonus, in ponderibus
uncia; ergo in genere verborum est ponere unum indivisibile; sed non est nisi hoc
verbum est quod erit indivisibile; sed cuiuslibet verbi impersonalis intellectus est
divisibilis (S: MS = indivisibilis), ut patet per expositionem ipsius, quare hoc verbum
est non erit impersonale. [iii] Item si hoc verbum est possit impersonari, cum sit de
intellectu alterius verbi, et ilIa sunt verba per naturam ipsius, omnia alia possunt
impersonari, quod quidem est impossibile, quare et primum, et sic reddit idem quod
prius, scilicet, quod non possit impersonale fieri. [iv] Item visio est in aliquo
dupliciter: vel sicut in organo, et sic est in oculo, vel sicut in causa et //Z:148vb// sic
est in anima. Similiter hoc verbum est est sicut radix et causa respectu aliorum
verborum, et licet visio possit privari ab organo in quo est, non tamen ab anima, quia
est in ipsa sicut in causa, quare similiter cum persona sit in hoc verbo est sicut in
causa, in aliis vero verbis sicut in organo, et si possit privari ab aliis verbis, non
tamen ab hoc verbo est, et ita ut prius. [v] Item, hoc verbum est, ut dicit Aristoteles
in \ibro Perihermeneias, significat //S:80ra// compositionem quam sine compositis non
est intelligere; sed composita sunt suppositum et appositum, quare non poterit intelligi
sine supposito et apposito hoc verbum est, quare oportet quod non intelligatur in
oratione vel quod non possit poni impersonaliter." Arguments AI, A2, and A3 in the
text correspond to [i], [ii], and [v], respectively.
8The res verbi is whatever single thing is designated by the various inflected forms of a
verb. Cf. C.H. Kneepkens, "'Legere est Agere:' The First Quaestio of the First
Quaestiones-Collection in the MS Oxford, CCC250", Historiographica Linguistica
VII.2, 1980, pp. 109-30.
266 MARY SIRRIDGE

domum, he elaborates, being is connected to a substance in which it

inheres as that in which it inheres and as that from which it issues; by
contrast, in est faciendum it is related to the substance only by efficient
causality.9

The author explicitly refers us back to his prior explanation of why it

can be said that the crucial connection with substance is retained in the
impersonal. There our author has explained that in a sentence like Sortes
currit, the referent of currit is designated as being connected to the
substance in which it inheres as its material and efficient cause, "for
Socrates is the subject of the running and the effector of the running." But
in a sentence like a Sorte curritur the action is designated as removed from
substance as material cause and as receiving another conjunction which
involves efficient causality; the verb is thus rendered impersonal, since it
now designates an action as conjoined to substance "in virtue of efficient
causality only." Similarly, in a sentence like paeniteo, the action of the verb
is signified as having a twofold connection with substance: as the receiver
of the act and as the subject in which it inheres. In paenitet me, the thing
referred to by the verb is not referred to as connected with substance as that
in which it inheres, but as being connected with substance as recipient. tO

This account has a certain intuitive plausibility for a Sorte curritur and
paenitet me. In each case it makes sense to say that the substance in which
an action inheres is not expressed, grammatically speaking; moreover, in
both cases, that substance is specified in the sentence as a whole, i.e. by
pronouns in oblique cases, but in a manner extrinsic to the verb. The
plausibility does not stand up to close scrutiny, however. The view that the
verb must have a connection with substance or a supposit really rests on
the requirement that the syntactic noun/verb structure reflect the
substance/attribute structure of reality. It has been objected that this
requirement does not seem to be met by, say, curritur a me, precisely
because the substance in which the action inheres is not specified as the

9KS (S 80ra; Z 148vb): "Ad prim urn sic: cum dico sum faciens domum, esse
comparatur hie ad substantiam in qua est res verbi sive ad illud in quo est, et sicut a
quo egreditur (S: MS = agreditur). Sed cum dicitur est faciendum, comparatur in
causalitate efficiente tantum (sic!). Et ita per eandem causam per quam dictum est
superius activa et neutra impersonari hoc verbum est fiet impersonale. Et dieendum
quod Iicet essentia non sit separabilis a subiecto penitus, tamen potest privari a
substantia <ut> in qua est. Ab ea enim in qua est non potest privati cum sit eadem
cum re verbi; sed ab ea in qua est et ut in qua est potest privati."
lOKS (Z 148va; S 79vb): "Verbum personale (S: MS = impersonale) a quo descendit
impersonale passive vocis significat actum comparatum ad substantiam in duplici
causalitate, scilicet in causalitate materiali et efficiente, ut patet cum dicitur Sortes
currit; Sortes enim (S: MS = non) est subiectum cursus et efficiens cursum. Sed cum
privatur habitudo que est actus <comparatus> ad substantiam causalitate materiali
recipiendo alteram compositionem que est in causalitate efficiente fit personale [et]
impersonale, et hoc modo significatur actus cum dicitur a Sorte curritur, ut comparatur
ad substantiam in causalitate efficiente solum (fit...efficiente: S: MS am.) ... Similiter
verbum impersonale a quo descendit impersonale active vocis comparatur ad
substantiam dupliciter: ut est recipiens actum, et subiectum (S: MS = sieut verbum),
ut patet; cum dicitur peniteo substantia importata per hoc verbum peniteo est
subiectum penitentiae et recipiens; sed cum (S: MS = causa) dicitur penUet me (me S:
MS am.) comparatur ad substantiam huius pronominis me sieut ad recipiens tantum; et
ita privatur una comparatione recipiendo a1teram."
 CAN 'EST' BE USED IMPERSONALLY? 267

substance in which the action inheres. As an answer to this objection,

KS's claim that a subject of inherence can be referred to in some other way
is irrelevant. Even if it were claimed that curritur a me was derived from
curro by a trivial meaning-preserving transformation, we would still have
no explanation of how its syntax allows it to mean what curro does; and
how, for example, are we to explain the fact that curritur alone is a well-
formed complete sentence? And in fact, the author of KS cannot hold that
curritur a me and curro are meaning-preserving transformational
equivalents, since he explicitly holds that the impersonal is required
because it has a characteristic signification that personal forms lack. It is
exactly this special signification that this account does not explain, for it
carries on as if the oblique complement has exactly the same syntactic
relationship to the impersonal verb as the nominative subject has to the
personal verb. There is good reason, then, to be generally suspicious of
this model of the impersonal.

More to the present purpose, although it seems true that in est

faciendum there is no subject specified as that in which the predicate
inheres, the explanation offerred for the parallel with genuine impersonals
is problematic for a number of reasons,ll most notably because it is
difficult to see any expression of displaced causality in our author's chosen
example, est faciendum. Moreover, though imperatoris of our sophism
sentence might be said to designate a recipient of sorts, it is hard to see any
application of the account to (1) or (2).12 The metaphysical problem with
this first model for impersonal function is that it depends upon efficient
causality playing a role in the meaning of the verb. It is thus inapplicable to
a verb like est, which seems neither to express an action, on the one hand,
nor to have an effect on a recipient, on the other. The grammatical problem
with the solution as a whole is, as we have seen, that it offers us a
semantic solution to a syntactic problem. Put another way, it asserts a
semantic equivalence where we already knew there was one; and it does
nothing to explain how the equivalence is syntactically possible.

A2 again raises the question of the meaning of such verbs as est and
interest by pointing out that est does not fit the canonical model of analysis
for impersonal verbs: Verbi = Ni+ fit. On this model, N is a noun which
refers to the thing which the various inflected forms of the verb designate

I lOur author has, for example claimed that in est faciendum, est is conjoined with
substance as efficient cause only, like active impersonals, and claimed that this is in
accord with the account we have just reviewed. But in that account active impersonals
are conjoined with a substance designated ut est recipiens actum, which is surely more
a matter of material than of efficient causality. In that account, being connected to
substance only by efficient causality is connected explicitly only with the passive
impersonals. Possibly, our author has misspoken, and meant to say that substance is
retained only in material causality in estfaciendum.
l2The suggestion that the intended parallel to me in paenitet me is the accusative
faciendum is not exploited here. In fact, in the discussion of sillogizantem ponendum
est terminos, KS proposes that the accusative gerund indicates the terminus ad quem of
the referent of est. This would perhaps secure a marginal analogy with me in paenitet
me, construed as a receipient of action. But this theory is itself poorly developed, and
in fact is used to support the claim that est iri such contexts is personal, since
paraphrase (i.e., ponendum est =causa est ponendi) shows that it has the causa as an
implicit nominative subject.
268 MARY SIRRIDGE

verbally (res verbi), e.g., the running named by cursus and designated
verbally by curro, -it, etc. Since est by its nature is indivisible, the
argument runs, it cannot be analyzed in this way and is therefore not
impersonal.

Our author's answer to A2 is initially not very promising: The rule that
every impersonal verb is analyzable in this way, he says, applies only to
those verbs which "signify a double motion" and thus signify a "perfect
action."13 Verbs like curro, he says, signify a double action because they
express in themselves a twofold change which can be represented by the
analysis cursumfacio: a transition from act to act, and something which is
brought about over and above the action as pure act. 14 Such verbs
"signify a perfect action." Thus, on our author's view, most verbs can be
analyzed into a neutral core of verbishness represented by some form of
ago or flo, and a noun which imparts quality to that core by its reference to
the res verbi. It is only such verbs whose impersonal forms can be
analyzed into fit and a nominative specifying the nature of the action. A2
points out that interest cannot be analyzed in this way. To respond, as our
author does, that interest is not of this kind, since it cannot be analyzed into
a designation of pure act signified by facio or ago and a nominal
specification of the quality of act effected, seems simply to admit that A2
has detected an important difference between interest and the other
impersonal verbs.

But KS has a further explanation of how interest can still be

considered to be impersonal. It is deprived of a nominative designating the
substance in which the action designated by the verb inheres, and in
addition:

''The preposition which is attached to it [sc. inter] conjoins it with

the action toward which it is directed, and this action it has as a
supposit with respect to its finite mode, but not with respect to its
<own> action."15

13KS (S 80ra; Z 148vb): "Ad aliud dicendum quod cum dicit intellectum cuiuslibet verbi
impersonalis <esse> condivisibilem (MS = condivisibile: Sam.), intelligendum est de
impersonalibus significantibus duplicem motum, et que actum perfectum significant,
ut habitum est prius."
14KS (S 80ra; Z 148vb): "Ad primum dicendum quod quedam verba significant duplicem
actum quia in eis est duplex mutatio, una que est ab actu in actum, alia que est ab actu
vel ab agente supra naturam, quod patet; curro (S: MS = curritur) idem est quod cursum
facio et [ego similiter. Per hoc quod dico ago (S: MS = [ego) vel facio importatur
mutatio que est ab actu in actum, et per hoc quod dico curro vel [ego importatur
mutatio que est ab agente supra naturam ... Et talia sunt que significant actum
~rfectum. Tale autem non est hoc verbum interest, ut prius visum est."
1 KS (S 80ra; Z 148vb): "Ad primum dicendum quod hoc verbum interest est
impersonale sicut pretendebat prima ratio, unde notandum quod impersonale est a
privati one nominativi substantie in qua est res verbi; et hoc est quia prepositio (MS
propositio?) que ei apponitur comparat ipsum ad actum ad quem est, quem actum habet
pro supposito respectu modi finiti, et non respectu actus." Also ibid., "Ad aliud
dicendum quod hec prepositio inter facit compararl hoc verbum est ad actum qui redditur
ei pro supposito, et facit ipsum comparari ad substantiam in causalitate tantum, et
propter hoc fit impersonaIe."
 CAN 'EST' BE USED IMPERSONALLY? 269

Thus interest exhibits the feature of having a res verbi as a supposit,

though not quite in the usual way. In the usual case, curritur, consisting of
a twofold action, is analyzed into cursus fit, where fit designates the bare
doing and cursus designates the res verbi which is the supposit of fit,
thereby specifying what kind of bare doing is occurring. Similarly, interest
has a designation of a res verbi as supposit, i.e., ducere. Ducere cannot
function as a supposit in the sense of a substance in which an action
inheres, and thus cannot be a personal subject. Interest is thus impersonal
because it does not have a personal subject for a supposit, but rather some
res verbi.

Whatever its virtues as an explanation of the impersonal and of

interest, this second pattern of explanation does not explain how est can be
taken impersonally: est cannot be analyzed into a nominal designation of
the res verbi and fit; nor can it in its proper signification be conjoined with
the infinitive, as verbs like interest can. Indeed, our author explicitly links
the capacity of interest to conjoin with an infinitive supposit to the function
ofthe component inter.16 The reason for the anomaly is obvious: est in its
proper signification does not express an action or a relation or an
inclination to act.

We have thus been given two accounts of the impersonal, of which the
latter in terms of the res verbi seems the more inclusive and the more
satisfactory. But neither is very useful in explaining why est in est
faciendum or est dies is different from est in Socrates est, and indeed,
different in a way which strongly suggests the impersonal function of the
verbs in curritur a Sorte and paenitet me.

2. A Second Theory about 'est'

In his discussion of the sophism Syllogizantem ponendum est

terminos, our author takes a different position:

"It is to be said that this verb est cannot be made impersonal like
other verbs. For those verbs are not said to be impersonal only
because they are deprived of the substance 'in which the action
inheres, but in addition it is required that in verbs of this sort there be
understood a twofold action, one of which is like a suppositum, the
other like an appositum, as when we say curritur: cursus fit. But this
verb est does not include within itself the notion of two actions, and
because of this it cannot be said to be impersonal like other verbs.
Nonetheless if it is said to be impersonal, this is by virtue of being
deprived of a definite subject. Or it is possible to answer otherwise
that it is not said to be impersonal on the basis of deprivation of
substance, but is said to be impersonal because it asserts a departure
from inaction into action, and thus there corresponds to it a general

16Cf. n. 15 above.
270 MARY SIRRlDGE

substance, as, for example, when we say est inter canem et lupum,
est dies, etc)7

According to this account, A2 is correct. est may be said to be used

impersonally only improperly and in an extended sense. One interpretation
of this extended sense suggested here, i.e. that est can be said to be
impersonal just because it lacks a definite subject, has already elsewhere
been pronounced by our author an insufficient condition for impersonality
proper. The other, that est here has a general substance, or perhaps
substance in general, as a subject seems to have been suggested to our
author by the examples est dies and est inter canem et lupum, which he
cites here. Something resembling (2') looms here; but obviously our
author is still thinking of a simple subject-predicate paraphrase with, say,
'something-or-other' as its subject term.

In the discussion of the sophism Legitur Virgilium, with its

inconvenient accusative, our author gives a fuller statement of this
approach to the impersonal. He first distinguishes between impersonal
verbs "properly speaking" and others which are impersonal verbs "loosely
speaking". The former are called impersonal verbs because [a] they are
deprived simpliciter of the substance in which the res verbi inheres; and [b]
they have a res verbi for a supposit. Impersonal verbs only loosely
speaking are impersonals only by virtue of privation of a definite subject,
and not privation of a subject simpliciter; they fail condition [a] and a
fortiori condition [b]. Impersonals properly speaking are further divided:

Impersonalia proprie loquendo

I. que habent rem eiusdem verbi II. que habent rem alterius verbi
pro supposito pro supposito
I. privatur a substantia in I. rem alterius verbi significantis
causalitate materiali; remanet in passionem
causalitate efficiente (Example: debet legi a me)
(Example: curritur a me)
2. privatur a substantia in 2.rem alterius verbi significantis
causalitate efficiente; remanet in actionem
causalitate materiali (Example: interest regis ducere)
(Example: paenitet me)
3. privatur causalitate in
causalitate materiali; remanet in
intentione causae finalis
(Example: placet mihz)

17KS (z: 147vb:S 77vb): "Dicendum quod hoc verbum est non poterit fieri impersonale
sicut alia verba. IlIa enim verba non dicuntur (S: MS = dicunt) impersonalia solum
quia privantur a substantia in qua est actus, sed cum hoc exigitur quod in huiusmodi
verbis intelligitur [dupliciter] duplex actus quorum unus est sicut suppositum, reliquum
sicut appositum, ut cum dicitur curritur: cursus fit; sed hoc verbum est non habet in se
intellectum duplicis actibus, et immo si dicatur impersonale, hoc est a privatione
substantie definite. Vel aliter potest dici quod non tenetur impersonaliter per
privationem substantie, sed tenetur impersonaliter quoniam dicit exitum ab otio in
actum, et tunc respondet ei substantia generalis extra, ut cum dicitur est inter canem et
lupum, est dies, etc."
 CAN 'EST' BE USED IMPERSONALLY? 271

Only 1.3 is really new relative to previous treatments. But the division
makes it clear that the essential characteristic of the impersonal for our
author is the taking of some res verbi for a suppositum consequent upon
the privation of the substance in which the res verbi inheres. Obviously
habet dubitationem, legendum est musas, est dies, and legitur Virgilium
have no place in this schema. They are impersonals only large loquendo,
by virtue of a "proportionality" with the genuine impersonals:

"That is, just as in the case of the other impersonals a subject is

provided by the res verbi which ought to be [construed] a parte post,
and not a parte ante, similarly in the case of these verbs when they
are rendered impersonals the suppositum is provided by something
which is construed a parte post. Thus, when we say habet
dubitationem, est dicendum, or the like, the suppositum of this verb
habet is provided by the accusative dubitationem. And similarly
when we say est inter canem et lupum, est legendum musas, and in
the same way when we say legitur Virgilium."18

Even if our author raises additional problems by introducing the matter

of construction a parte post and a parte ante, the point is clear enough.
KS's paradigm of the impersonal, however it is explained, will not support
classifying est in any of the usual examples as an impersonal verb, strictly
speaking. But our attention is called to a syntactic similarity with genuine
impersonals, which explains why est is sometimes said to be used
impersonally. And the author of KS seems to think that est is thereby
shown to be sufficiently impersonal that the standard examples of the
impersonal use of est do not require paraphrase which would reveal a
concealed nominative subject.

It is fairly obvious that KS does not offer a single consistent treatment

of the impersonal generally or of the purported impersonal use of est.
Possibly the author's right hand has simply forgotten what the left has
written. A more radical approach might be to ask whether KS was in fact
intended to put forward a unified and cumulative grammatical doctrine, or
whether the aim was not instead to meet each challenge as it presented
itself, maintaining a generally consistent approach on important points, and
to introduce specific innovations which might or might not ultimately prove
theoretically consistent in the strict sense. The answer to this question is
beyond the scope of this paper, in part because it cannot be answered on
the basis of an examination of a single instance of grammatical doctrine.

18KS (z: 152vb; S: 87rb): "Notandum tamen quod hec non proprie dicuntur
impersonaiia, cum non habeant rem verbi pro supposito, sed dicuntur impersonalia ad
proportionemaliorumimpersonalium.scilicet quod sicut in aliis impersonalibus
redditur suppositum per rem verbi que proprie debeat esse a parte post et non a parte
ante. Similiter in his verbis secundum quod flunt impersonalia datur suum suppositum
per iIIud cum quo construuntur a parte post, unde cum dicitur habet dubitationem, est
dicendum, vel aliquid tale, [habet] suppositum habetur huius verbi habet per naturam
huius accusativi dubitationem. Et similiter cum dicitur est inter canem etlupum, est
legendum musas, et eodem modo cum dicitur legitur Virgilium; in his enim omnibus
constructionibus habent ista verba a1iquid pro supposito per naturam articuli et quod
datur eis per naturam eorum que construuntur cum ipsis a parte post."
272 MARY SIRRIDGE

3. The Remote Sources of the Theory

About the remote sources of KS's doctrine we can be more definite.

Our author's treatment of the impersonal verb is surely indebted principally
to Priscian's meandering discussions of impersonal verbs at Institutiones
XVII1.51-60 and XVII.90-92. At IG XVIII.51, Priscian discusses
constructions acceptable with active and passive impersonals, respectively.
In the course of dealing with active impersonals, he observes that they may
be conjoined with infinitives, e.g., placet discere, or with various oblique
cases, e.g., placet mihi, iuvat me, pudet me illius rei. He digresses from
his project of listing the various oblique constructions to observe that debet
is impersonal when, conjoined with an infinitive of passive force, it
encompasses all three persons, as in debet fieri a me, ate, ab ilIo. He is
prompted to add an important reminder:

"sed si quis et haec et omnia impersonalia velit penitus inspicere,

ad ipsas res verborum referuntur et sunt tertiae personae, etiam si
prima et secunda deficiant." (IG XVIII.53)

He intends, it seems, to point out that the impersonal verb itself is

syntactically and semantically third person by virtue of the verb's own
intrinsic reference to some "thing", independently of other personal
references in the sentence. He certainly seems to be extending this claim to
all impersonal verbs, and not just to those with active endings.

In the subsequent discussion of passive impersonals, Priscian seems

to make a related point. Passive impersonals, Priscian points out, cannot
be conjoined with infinitives, and are intransitive,

"for if I add a nominative, bellatur gens, a transition from one to

another person ensues, and the form develops into a passive,
although Apollonius in Peri syntaxeos III shows that also in
impersonals we can understand the nominative of the action
designated by verbs (ipsius rei verborum) ... Similarly in all the
impersonals which we use frequently, those which we take from
exactly this sort of Greek verbs, that is from melei, dei chrei, ... we
can have an understanding of the nominative of the very thing which
is understood in the verb. For when I say curritur, I understand
cursus, and sedetur, sessio." (IG XVIII.55-56)

Difficulties of interpretation notwithstanding (how exactly does

Priscian go from such verbs as 'it is necessary' to 'curritur' and 'sedetur'?)
this vexed passage appears to make an overall point which is related to
Priscian's previous claim: even ifimpersonals can be made personal by the
addition of a subject, they already as impersonals intrinsically have a
reference, a reference to the thing to which the verb refers by means of an
implicitly contained nominative. Thus curritur contains cursus.

This section of IG closes with a parenthetical remark that impersonals

which are conjoined simultaneously with the genitive and accusative case,
e.g., pudet me tui, signify by the accusative the person in whom the
passion resides (in qua est) and by the genitive the person from whom it
 CAN 'EST' BE USED IMPERSONALLY? 273

arises (ex qua/it). Now, the function of oblique fonns with impersonals
has already been discussed, notably at IG XVII.90, in the context of a
discussion of the function of pronoun subjects. There Priscian points out
that pronoun subjects are redundant with first and second person verbs
except for purposes of contrast and emphasis, whereas they are needed to
limit and define the uncertain or limitless application of third person verbs.
He is thereupon struck by the reflection that impersonal verbs, too, may
have their limitless application delimited and defined by pronouns which
impose person and number, in this case by those in the various oblique
cases:
"Impersonal verbs, too, since in themselves they are infinite, are
delimited with respect to person and number by the addition of
pronouns; and in the case of perfect verbs, this extends through all
the modes." (/G XVII.90)
Priscian's discussions seem to be governed more by the variety of
usage than by conscious architectonic. He lists a general rule, then
prominent exceptions or applications, fairly often digressing into matters
suggested by some incidental feature of these examples, fmally returning
more or less to the matter under discussion. His theoretical objectives often
seem quite limited. At XVIII.50, for example, he makes no attempt to give
a common account of active and passive impersonals, though in each case
he has something to say about a reference to a res verbi, which almost as
an afterthought he generalizes at least to all impersonals. In neither case is
the initial rule about the admissibility or inadmissibility of an associated
infinitive actually related to a doctrine of the res verbi; and nowhere is it
actually asserted that a contained nominative which refers to the res verbi is
the subject of the sentence. Moreover, Priscian makes no attempt to
generalize the special rationale which he offers in XVllI.50 for such
constructions as pudet me tui to all oblique fonns with impersonal verbs;
nor does he connect that causal rationale with the delimiting function of
oblique fonns with impersonal verbs which he has described earlier at
XVII.90.

Priscian's discussion of the impersonal presents us with a striking

example of what we might call "the portentous silence of Priscian." Many
of the features of the discussion of KS are easily explained as a matter of
making connections which Priscian did not make. It is easy to see, for
example, that Priscian's remarks suggest the two quite different models for
detennining the supposit of the impersonal verb whch we find in KS; and
his explicit inclusion of modal verbs suggests to the author of KS that
interest can qualify as an impersonal verb by taking the res verb; of another
verb as its supposit. Moreover, Priscian's chance use of the phrases in qua
est and ex qua fit in the discussion of pudet me ill;us clearly offerred an
immediate point of entry for the machinery of Aristotelian causality which
we see in KS's systematic rationale for the use of oblique fonns with
impersonals.

The use of ex qua est and in qua fit atIG XVill.50 is much more
obviously consonant with Aristotelian causality than the theory that the res
verb; of the impersonal verb is its real supposit, which may explain why
the author of KS attempts to work out this theory.
274 MARY SIRRIDGE

But the reference to De Interpretatione in A3 points to. an even

more significant Aristotelian influence on the discussion. A3 objects
that without suppositum and appositum there can be no complete
sentence involving est; this is a verbatim quotation from Boethius'
translation of De Interpretatione 1:3. Our author's answer is brief: The
relationship of subject to verb expresses material and efficient
causality, and either one of these can be dispensed with, so long as
the other is expressed. 19 This response points up the fact that
specifying some source of a supposit for the impersonal verb is a major
objective of KS's whole discussion - an objective which Priscian does
not share. Both of KS's accounts of the supposit of the impersonal verb
are products of the assumption that the impersonal verb must have a
supposit if the sentence is to be complete as a sentence, and not just
in the fragmentary way in which nouns and verbs in isolation are
complete. This conception of the meaning of nouns and verbs in
relation to sentential meaning and structure can be found in Aristotle's
De Interpretatione I: 3, particularly as elaborated by Boethius:
"Thus, it has been rightly said, 'verbs themselves said alone
are nouns,' because 'he who speaks forms an understanding,
he who hears comes to a resting point.' Or the exposition will
be better if we speak thus: verbs themselves said in isolation
are nouns, for the reason that they signify some thing. If the
verb signifies something which is alwa,ys in another or
predicated of another, it does not thereby sigmfy nothing. Nor, if
It signifies something which cannot be without a subject, does it
thereby signify that which is a subject ... Therefore verbs said
alone mdeed signify something and are the names of things, but
do not yet signify in such a way that something is <thereby>
determined to be or not to be F, i.e., in such a way that they
constitute an affirmation or negation."20
A reader of Boethius' commentary would know, too, that a few
pages earlier, Boethius had ruled that such sentences as paenitet
Socratem are perfect sentences. A subject must therefore be located
for sentences with impersonal verbs. Priscian's remarks about the res
verbi which explains why impersonal verbs have third person forms,
and his claim that oblique forms delimit the application of the verb now
become suggestive. Moreover Priscian's talk of a res verbi can be
assigned a fairly clear technical interpretation on the basis of
Boethius' discussion of "the thing to which the verb refers as inhering
in another." Indeed, it seems clear that KS's theory of subject and
verb as a whole has been inflenced by an Aristotelian conception of the
semantics of sentential structure. The result, as we have seen, is a
theory which presents a number of different ways of retracing the
impersonal to this sentential model and which is irresolute in its
handling of constructions which offer resistance to this model.
Louisiana State University

19K5 (s: SOra; Z: 14Svb): "Ad aliud dicendum quod sine dubio omnis comppsitio habet
composita, et ita appositum et suppositum. Sed duplex est subici habitudo, sicut
dictum est superius. Comparatur enim ad actum in causalitate materiali et efficiente,
et ab una istarum potest privari, sicut visum est."
20Soethius, Commentaries on Aristotle's "De Interpretatione", ed. Leonardo Taran,
Garland: New York and London 19S7. Reprint of ed. Carolus Meiser, Teubner:
Leipzig, IS77-S0, v.], ]:3. pp. 72-5.
 Part III
Logical Sophisms
 Who is the worthiest of them all?
by Allan Back

Here I consider the sophism, 'Man is the worthiest of creatures'. The

main use of this sophism was to test classifications for different types of
supposition. For William of Sherwood, 'man' in that proposition has
simple supposition. Yet the difficulty was that in typical examples of
simple supposition, like 'Man is a species', the subject term was said to
refer to the species, the universal man, and not to the individuals.
However, in 'Man is the worthiest of creatures', the subject term seems
not to refer to the species. And yet it does not refer in the usual way, of
personal supposition, to individuals in a usual definite (e.g. 'this man') or
indefinite (e.g. 'some man') fashion: since there are many human beings,
and the proposition appears to concern each of them equally, each then
would be the worthiest of creatures, which is absurd.

So this sophism came to have interest as something more than a

particular puzzle. It provided a testing ground for the adequacy of
competing theories of supposition. And, more fundamentally and more
darkly, it figured in the debate between nominalists and realists in the
thirteenth century.

First I shall consider early medieval solutions of this sophism. I shall

then discuss the use made of it by Ockham and Burleigh, as a test case for
the success of Ockham's nominalism. I shall end by adjudicating on their
controversy.
1. Early Medieval Treatments of the Sophism
The Fallacie Parvipontane considers the sentence, 'Man is the
worthiest of creatures' to be an instance of univocation. It defines
'univocation' thus: "Univocation is the varied supposition of a name
(nominis) when the signification remains the same; since, although the
supposition varies, still there remains the same supposition."l The
signification of a name remains the same when the same definition may
replace the different occurrences of that name, salva veri tate; the
supposition will remain the same only when the name stands for, or refers
to, the same things in its different occurrences. Now the Fallacie
Parvipontane claims that 'man' in 'Man is the worthiest of creatures'
continues to have the same signification, even though it contains an
ambiguous reading, so as to have two different possible references:

''The second species (of univocation) is when some expression is

taken to deal sometimes with some (part) of things of some manner
(maneriei), and other times with the manner of things as such, as

lLogica Modernorum, ed. L. M. De Rijk, Vol. I, Assen: Van Gorcum 1962, p. 562,
11-12: "Univocatio est manente eadem significatione variata nominis suppositio; quia
etsi variatur suppositio, manet tamen eadem significatio.

277
278 ALLAN BACK

when it is said, 'Man is the worthiest of creatures'. For this

(assertion) can be understood to make some claim about some (part)
of the referents (appellatorum) of this name, 'man'; it can also be
understood to make a claim about the manner of things as such. The
following propositions must be understood in the same way: 'Gold
is the most precious of metals'; 'Pepper is sold here and at Rome'.''2

This text sets out what was to become the main issue of this sophism,
but does not resolve it. It claims that 'Man is the worthiest of creatures' has
two readings. On one reading, what is said to be the worthiest creature is
one or more of the things called or labelled correctly by the term 'man'.
That is, one or more of the individual things that are human would be the
worthiest of species. (The superlative, 'worthiest', seems to rule out there
being more than one such thing, but below I shall suggest a way in which
there may arise a kind of multiplicity here.) Here then the generic
description, 'man', does allow a descent to, and perhaps a reduction to,
singulars.

On the other reading, what is said to be the worthiest creature is the

sort (maneries) man itself. If'maneries' means 'species', then here it is the
species man that is the worthiest of creatures. I am not certain how
plausible this reading is; it is dubious whether a species is a creature. Even
in a Platonist ontology, where a species would be a Form, God need not
create species) Perhaps, then, 'maneries' here indicates only that a descent
to singulars is not justified by this assertion, and so no individual human
being will be the worthiest of creatures. Rather, all human beings, in a
certain mode or manner of being, are the worthiest of creatures. That is,
human beings, in a certain respect, are the worthiest of creatures. This
more intelligible reading agrees also with the account of William of
Sherwood's, discussed below.

So the Fallacie Parvipontane does not judge between these two

readings, or even explain them much. It is likely merely to record current
views on the meaning of 'Man is the worthiest of creatures'. However,
later authors did not scruple to take sides.

2. William of Sherwood

Sherwood discusses 'Man is the worthiest of creatures' as an instance

of simple supposition. In simple supposition, the term, say, 'man', stands
for or refers to the concept itself. The standard example is 'Man is a
species', where 'man' stands for a concept, or a secondary substance if

2Logica Modernorum, p. 562, 20-6: "Secunda species est quando aJiqua dictio
transsumitur modo ad agendum de aliqua rerum alicuius maneriei, modo de tali manerie
rerum, ut cum dicitur: 'homo est dignissima creaturarum'. Potest enim sic intelligi ut
fiat sermo de aliquo appellatorum huius nominis 'homo'; potest enim intelligi ut fiat
sermo de tali manerie rerum. Eodem modo intelligendum est cum dicitur: 'Aurum est
preciossismum metallum'; 'Piper venditur hic et Romae'.
3That is, God need not create the ideas (€LSl]) as existing in the divine intellect, although
He must create the ideas or species as existing in reo If the species is existing in re only
insofar as its exemplars exist, sc. as a secondary substance exists, then the species at
least in one sense is no creature. Cf. Aquinas, Summa Theologiae 1.44.l.r; 1.l5.l.r.
 WHO IS THE WORTHIEST OF THEM AIL? 279

you like, and where no descent to individuals is permitted: no individual

human being is a species. In contrast, in personal supposition, the term
refers to the indviduals that fall under the concept. So, 'every man' in
'Every man is seated', has personal supposition, since a descent to (at least
some) individual human beings is valid.

Sherwood apparently classifies 'Man is the worthiest of creatures' as

an instance of simple supposition since he holds that no descent to
singulars (without qualification) is justified. For the claim seems to
concern humankind in general, and not an individual, unusually saintly
human being. Humans may be considered the worthiest because they have
moral agency and rational, immortal souls. Moreover, as there seems to be
just as much reason for one human being as another to be the worthiest (in
many cases or on the above grounds), so Xanthippe and Lysis would each
be the worthiest of creatures. But then there is a contradiction: a superlative
like 'worthiest' demands uniqueness. So Sherwood has some reasonable
motivations to consider this claim to be an instance of simple supposition.

Yet, on the usual reading, 'Man is the worthiest of craatures' and

'Man is a species' have logical differences: the former seems somehow to
concern individual human beings, as individuals, while the latter does not.
Sherwood recognizes and addresses this difference:

"The second mode (of simple supposition) occurs in this way:

'Man is the worthiest of creatures'. And this (mode of) supposition
is not like the first mode, since there is no predication of the abstract
species itself, but of it insofar as it is in things. Whence, such a
predicate can be said of any thing of the species, insofar as it
participates in the nature of the species. Whence, it can be said, 'This
man insofar as man is the worthiest of creatures' . And there is this
(mode of) supposition whenever things of a species take on some
predication only with reduplication of the species, and not otherwise.
E.g., if! were to say, 'Man is animal', this is personal supposition,
since individuals receive this predication without reduplication of the
species. "4

So Sherwood holds that 'man' in 'Man is the worthiest of creatures'

has simple supposition of a special type, with reduplication of the species.
In this type of simple supposition, a descent to singulars can be made, not
without qualification (simpliciter) as in personal supposition, but with the
qualification (secundum quid) of the reduplication of the species. So, given
that Judas and Peter are human beings, it follows from 'Man is the

4William of Sherwood, Introductiones ad Logicam, ed. M. Grabmann, Sitzungsberichten

der Bayerischen Akademie der Wissenschaften: Phil.-hist. Abteilung, 1937: 10, p. 77,
19-28: "Secundum modus fit sic. Homo est dignissima creaturarum et non est similis
haec suppositio priori, quia non predicator de ipsa specie abstracta, sed in quantum est
in rebus. Unde potest tale predicatum diei de qualibet re speeiei, in quamtum partieipat
naturam speciei. Unde potest dici: 'iste homo, in quantum homo, est dignissima
creaturarum. Et est hec suppositio, ubicumque res speciei suscipiunt aliquam
predicationem solum cum reduplicatione speciei et non aliter, ut si dicam: homo est
animal, haec est suppositio personalis, quia individua recipiunt hanc predicationem sine
reduplicatione speciei." My translation; also see Norman Kretzmann, trans.,
Sherwood's Introduction to Logic, pp. 111-2.
280 ALLAN BACK

worthiest of creatures' that Judas insofar as he is a man is the worthiest of

creatures, and that Peter insofar as he is a man is the worthiest of creatures,
and so forth. However, the reduplicative phrases cannot be dropped
without fallacy. So it does not follow: Judas qua man is the worthiest of
creatures; thus, Judas is the worthiest of creatures. For then we would
have two undesirable consequences: the material falsity of a sinner being
the crown of creation, and the formal absurdity that every individual
human being is the worthiest of creatures. For, likewise Peter, Xanthippe,
et aI., would each be the worthiest of creatures. But, again, for a
predication of a superlative to be true, it cannot be true of more than a
single subject.

In my On Reduplication,5 I have discussed reduplicative propositions

extensively. In my terms, Sherwood is using accidental qua phrases: these
reduplicative phrases cannot be dropped from the sentences. Moreover,
their presence and attachment to the subject terms change the reference of
the subject: it is not "Judas" who is the worthiest of creatures, but "Judas
qua man". For, given that it is true to say of any human being, qua human,
that he or she is the worthiest of creatures, and given that a superlative can
be true of at most one subject, 'Judas qua man', 'Xanthippe qua man',
'Peter qua man', they must all name the same thing, according to
Sherwood's analysis. That thing is the worthiest. But that thing is not an
indvidual human being. What then might that thing be?

In the Aristotelian tradition, human beings are distinguished from the

brutes through being rational. The rational is the divine, moral, free spark
in us. It is the humanity, more precisely, the rationality, in each individual
human being, that gives that individual a superlative worth. It is, then,
rationality, not the abstract property of rationality, but the rationality in the
individual human being, that is the worthiest of creatures. This rationality
constitutes a certain aspect of every individual human being. As this aspect
remains invariant in a multiplicity of human beings, it subsists as a single
thing. Being an aspect of individual things, it is created along with them -
even if there are no Forms (of Rationality) or if Forms are not created. In
Platonic jargon, this worthiest thing is an immanent Form, "the rationality
in us"; in the Aristotelian, it is a subjective part of the individual (sc. an
element in the formula of the essence of a human primary substance). Such
then is the ontology towards which Sherwood appears to have been
pointing.

Aquinas seems to share this interpretation of 'Man is the worthiest of

creatures' with Sherwood:

"But whenever something is attributed to a universal so

considered, which clearly is apprehended by the intellect as one, still
what is attributed to it does not pertain to the act. of the intellect, but
to the being (esse) which a nature apprehended in things external to
the mind has; e.g. if it is said that man is the noblest of creatures. For
this also agrees with human nature in virtue of the fact that it is in
individuals. For any individual man is nobler than all irrational

5Allan Blick, On Reduplication, Munich: Philosophia Verlag, forthcoming.

 WHO IS THE WORTHIEST OF THEM AIL? 281

creatures, but still all individual men are not a single man external to
the mind, but only in the view of the intellect. And in this way the
predicate is attributed to it as to a single thing. "6

For Aquinas too, in 'Man is the worthiest of creatures' we are speaking of

human nature insofar as it is in individuals. He agrees that we are speaking
of a single thing here, even though it is true, for each and every individual
human being, that he or she is worthier than any other individual of any
other species. The superlative demands unity. Thus the subject of the
statement is not any and all individual human beings. Being an Aristotelian
and not a Platonist, Aquinas does insist that this humanity in individuals
does not exist in re as a separate immanent Form, but rather is a unity
existing only in intellectu that depends upon mentally abstracting a
common feature of individual human beings.

Whatever the metaphysical conclusion, it is clear that this analysis of

generic descriptions like 'man' in 'Man is the worthiest of creatures' will
complicate the ontology. In addition to individuals and universals (primary
and secondary substances, if you like), now there are (in the category of
substance) intermediate complex objects which might be styled,
indifferently, the universals insofar as they appear in individuals, or the
individuals insofar as they exemplify a universal. Before making this
move, which will complicate the ontology considerably, it seems adviable
to look sympathetically at Ockham's nominalist program, of reducing talk
of such complexes to individuals.

3. Ockham

As Ockham does not want to admit to the real existence of anything

but individuals, it is understandable that he rejects the claim that in 'Man is
the worthiest of creatures' 'man' has simple supposition. For Sherwood
insists that 'man' in that sentence has simple supposition at the cost of
asserting the existence in re of a complex universal, the rationality in this or
that human being. Rather, Ockham says:.

" ... it must be said that the opinion of those who say that in 'Man
is the worthiest of creatures' the subject has simple supposition is
simply false; rather 'man' has personal supposition in it.

"Nor is their argument valid, but it is against them, for they prove
that if 'man' has personal supposition then that (proposition) would
be false. But that argument is against them, since if 'man' stands
simply in that proposition and not for some singular thing, therefore

6St. Thomas Aquinas, 1n Aristotelis Libros Perihermenias Expositio, p. 126f. Cited in

Ignacio Angelelli, Studies on Gottlob Frege and traditional philosophy, Dordrecht:
Reidel 1967, p. 131: "Quandoque vero attribuitur aliquid universali sic considerato,
quod scilicet apprehenditur ab intellectu ut unum, tamen id quod attribuitur ei non
pertinet ad actum intellectus, sed ad esse, quod habet natura apprehensa in rebus, quae
sunt extra animum, puta si dicatur quod homo est dignissima creaturarum. Hoc enim
con venit naturae humanae etiam secundum quod est in singularibus. Nam quilibet
homo singularis dignior est omnibus creaturis irrationalibus; sed tamen omnes
homines singulares non sunt unus homo extra anima, sed solum in acceptione
intellectu; et per hunc modum attribuitur ei praedicatum, scilicet ut uni rei."
282 ALLAN BACK

it stands for something else, and that is the worthiest of creatures.

But that is false, since something common or a species is never
nobler than a singular thing under it, since in virtue of their mode of
speaking the inferior always includes its superior and more.
Therefore, that common form, since it is part of that man, is not
nobler than that man. And so if the subject in 'Man is the worthiest
of creatures supposits for something other than a single man, it
would be simply false."7

Ockham's main argument is simple: 'man' in 'Man is the worthiest of

creatures' must have either personal or simple supposition; it does not have
simple supposition; therefore it has personal supposition.

His reason why it does not have simple supposition consists in

claiming that a common attribute never has more perfection than do the
individuals under it. For the individual would have the perfection of that
attribute plus the perfections of other attributes. So to assert that a common
attribute, like man, is worthier than the individuals that have that attribute,
the individual human beings, would always be false.

Ockham's reasoning is not compelling, unless you make many

ontological assumptions. Why could not the imperfections of the
individuals, when added to their perfections, result in their having a net
worth less than that of their common attributes? You need not be a
Platonist to suspect that humanity in general might have more worth than
Hitler or Judas.

Moreover, it is not clear that a complex like 'the rationality in Judas' is

a universal, common attribute. Rather it looks to be more like an individual
accident, and even Ockham recognizes these. Now different individuals
have different individual accidents and essences, and yet there are common
characteristics of those esences. Still, because of the superlative,
'worthiest', 'the rationality in Judas' would have to denote a common
attribute, if 'Man is the worthiest of creatures' is to be true. Otherwise,
there would be many individuals, each the worthiest.

So there remain two choices, either to take 'man' to have simple

supposition, and so to stand for a certain sort of complex universal-

7William of Ockham, Summa Logicae, ed. P Boehner, G. Gill and S. F. Brown, Opera
philosophica, vol. I, St. Bonaventure, N.Y.: The Franciscan Institute 1974,1.66,26-
41, pp. 200-1: "Ad primam istorum est dicendum quod opinio dicentium quod in ista
'homo est dignissima creaturarum' subiectum habet suppositionem simplicem, est
simpliciter falsa; immo 'homo' habet tantum suppositionem personalem in ista.Nec
ratio eorum valet, sed est contra eos, nam probant quod si 'homo' haberet
suppositionem personalem quod tunc esset falsa, quia quaelibet singular est falsa. Sed
ista ratio est contra ipsos, quia si 'homo' stat simpliciter in ista et no pro aliquo
singulari, igitur pro aliquo alio, et per consequens ilIud esset dignissima creaturarum.
Sed hoc est falsum, quia tunc esset nobilius omni homine. Quod est manifesta contra
eos, quia numquam commume vel species est nobilius suo singulari, cum secundum
modum eorum loquendi inferius semper includat suum superius et plus. Igitur ilia
forma communis, cum sit pars istius hominis, non est nobilior isto homine. Et ita si
subiectum in ista 'homo est dignissima creaturarum' supponeret pro aliquo alio ab
homine singulari, ipsa esset simpliciter falsa.
 WHO IS THE WORTHIEST OF THEM AU? 283

singular attribute, or to take it to have personal supposition, and so to stand

for one or more individual human beings. Wishing to eliminate universals
existing in re, Ockham chooses personal supposition. Yet, curiously
perhaps, Ockham agrees with his opponents that no individual human
being is the worthiest of creatures, i.e. that replacing 'man' by the name of
any individual man would yield a false proposition. I suppose the rationale
consists in observing that, although some human beings are worthier than
others, still there is no definitive ranking of all human beings relative to
their moral worth.s So no individual human being is the worthiest of
creatures. However, as personal supposition requires a descent to at least
one singular, 'Man is the worthiest of creatures' will then be false.
Ockham admits this:

"Therefore it must be said that 'man' supposits personally, and

the proposition is strictly speaking false, since any singular
proposition under it is false. Nevertheless, in virtue of the intention
of those who maintain the proposition, it is true, since they do not
intend that man is nobler than every creature universally, but that
man is nobler than every creature that is not a man."9

So Ockham claims that 'Man is the worthiest of creatures' is, strictly

speaking false. For it to have personal supposition would require that at
least one individual human being be the worthiest, and that requirement, he
says, is not met. Moreover, even if it had simple supposition, it would be
false, since then a common attribute would be asserted to be the worthiest
and that would be false too.

Yet people take the claim to be true. Ockham gives a paraphrase by

which he attempts to save the phenomenon of ordinary speech: 'Man is the
worthiest of creatures' means that any individual human being is worthier
than any individual of any other species.

In effect, Ockham is following the reading of Aquinas given above.

For Aquinas too offers the parsing that any indvidual man is nobler than all
irrational creatures. Yet, Ockham deletes Aquinas' talk of 'man' in that
proposition standing, in the view of the intellect at least, for a common
human nature insofar as it is in individuals. Ockham might partly agree
with this point. However, he would not take such a realist stance towards
the cornman attributes as Aquinas seems to.1 0 Ockham's reading has the
merit of being simpler than Aquinas'. However, Aquinas' analysis has the
merit of explaining how, or at least acknowledging that, there comes to be

SNote that a similar point is made in modern discussions of the allocation of medical
resources. Although Albert Schweitzer may have a greater claim for treatment than
Adolph Hitler, in most cases it is not clear which individuals should be judged to have
the most merit and hence the greatest claim on the medical resources. Cf. Jonathan
Glover, Causing Death and Saving Lives, Harmondsworth: Penguin 1977, pp. 223-6.
90 ckham, Summa Logicae, I c.66, 42-5, p. 201: "Ideo dicendum est quod 'homo'
supponit personaliter, et est de virtute sermonis falsa, quia quaelibet singularis est
falsa. Tamen secundum intentionem ponentium earn vera est, quia non intendunt quod
homo sit nobilior omni creatura universaliter, sed quod sit nobilior omni creatura quae
non est homo."
IOC£. Summa Logicae I c.65.13-5: "Sed in ista propositi one 'homo est species', quia
'species' significat intentionem animae, ideo potest habere suppositionem simplicem."
284 ALLAN BACK

a single thing that is the subject of the proposition. On Ockham' s reading,

it would be better to say, in the plural, that all human beings have ultimate
worth, or, as Kant puts it, are beyond price.

4. Walter Burleigh

That I have given a reasonably accurate picture of the issues

concerning generic descriptions like 'man' in 'Man is the worthiest of
creatures' can be confinned by looking at Burleigh's discussion of this
example. In effect, Burleigh sides with Sherwood, and gives 'man' simple
supposition. He is willing to admit that 'man' here stands for a complex
attribute, 'the rationality in individual human beings', and not for the
individuals:

"Absolute simple supposition occurs when the common term

supposits for the significatum, as it is in its referents (suppositis).
Comparative simple supposition occurs when the common term
supposits for its significatum in virtue of the fact that it is predicated
of its referents ... Still it can be said that absolute simple supposition
occurs when the tenn supposits for its significatum absolutely, not in
comparison to its referents nor in regard to being in or being said of.
But comparative simple supposition occurs when the common tenn
supposits for its significatum in relation to the referents or for some
of its inferiors that have referents. In the first mode 'Man is the
worthiest creature of creatures' is true'; in the second mode 'Man is a
species' is true."!!

So, Burleigh claims, 'man' in 'Man is the worthiest of creatures', stands

for the significatum, as it is in its referents (suppositis). That is, it refers to
human nature insofar as it is evidenced by individual human beings.
Human nature, thus qualified, constitutes a complex abstract object or
conception. Burleigh calls this sort of simple supposition "absolute" in
order to emphasize that it refers to such an abstract, indissoluble entity. It
is that complex conception, half universal and half individual, that has the
greatest worth. Burleigh thus contends, against Ockham, that such an
abstraction of "man in common" may have a greater worth than an
individual human being:

"And it (,Man is the worthiest of creatures') should be understood

to claim that among corruptible creatures man is the worthiest of
creatures. And when it is said that Socrates is a creature worthier
than man in general, it should be said that this is not true, since

!lWalter Burleigh, De Puritate Artis Logicae Tractatus Longior, ed. P. Boehner, St.
Bonaventure, N.Y.: The Franciscan Institute 1955, 1.1.3, p. 11, 5-18: "Suppositio
simplex absoluta est, quando terminus communis supponit absolute pro suo
significato, ut est in suppositis. Suppositio simplex comparata est, quando tenninus
supponit pro suo significato, secundum quod praedicatur de suppositisooPosset tamen
dici, quod suppositio simplex absoluta est, quando terminus supponit pro suo
significato absolute non in comparatione ad supposita nec quantum ad esse in, nec
quantum ad dici de. Sed suppositio simplex comparata est, quando tenninus communis
supponit pro suo significato in comparatione ad supposita vel pro aliquibus
inferioribus suis habentibus supposita. Primo modo est haec vera: 'Homo est
dignissima creatura creaturarum'; secundo modo est haec vera: 'Homo est species'."
 WHO IS THE WORTHIEST OF THEM AlL? 285

although Socrates includes the perfection of man, still he does not

include it necessarily, but contingently, since when Socrates is
destroyed (corrupto) Socrates is not a man. And so it is clear that the
following inference is not valid: 'Socrates includes the entire
perfection of man, and also some perfection added in addition;
therefore Socrates is more perfect than human nature', but it is
necessary to add that Socrates necessarily includes the perfection of
the species of man, or that lle includes the perfection of the species of
man as a part of himself, and neither of these [additions] is true. "12

Ockham has argued that an individual human being has greater worth than
man in common, since he or she will possess the worth of man in common
along with the worths of his or her other common attributes. Like Aquinas,
Burleigh denies this; the complex abstract object, 'Man insofar as it is in
individuals', has greater worth than any individual human being. 13 I do
not find his reason compelling: he claims that the abstract object has more
worth since it is man permanently and necessarily, whereas an individual
human being is human only contingently and, presumably, temporarily.
But a good Aristotelian, at any rate, should deny that an individual human
being is human only contingently, as well as that the permanence of being
human makes a difference in how perfectly human something is. 14
However, I have given reasons above that might be more persuasive.

So, in effect, Burleigh tries to save the phenomenon - the

phenomenon here being the truth of 'Man is the worthiest of creatures'.
Ockham can have it true only by working a major paraphrase: any man is
worthier than any other creature. But this paraphrase seems to modify the
meaning of the original sentence. To preserve the superlative, Burleigh
must locate a unique object to be the worthiest. He is willing to "bite the
bullet" and admit complex abstract objects and a more complicated
ontology. In modem terms, this would amount to the claim that certain
sorts of generic descriptions (Heyer's contingent personal ones) point to
the need of a more complex domain for a logical language intended to
handle ordinary language: not only predicates and singulars, but also the
complex, a predicate insofar as it is in singulars.

5. Conclusions
What shall we conclude on the debate between Ockham and Burleigh?
At this point, there should remain no doubt on the historical significance

12Walter Burleigh, De Puritate Artis Logicae Tractatus Longior, 1.1.3, pp. 13,35-14,
10: "Et intelligendum sic, quod inter creaturas corruptibiles homo est dignissima
creatura. Et quando dicitur, quod Sortes est dignior creatura quam homo in commune,
solet dici quod iIIud non est verum, quia quamvis Sortes includat perfectionem
hominis, tamen non necessario includit earn, sed contingenter, quia Sorte corrupto
Sortes non est homo. Et ita patet, quod ista consequentia non valet: 'Sortes includit
totam perfectionem hominis, et etiam aJiquam perfectionem superadditum, ergo Sortes
est perfectior natura humana', sed oportet addere, quod Sortes necessario includeret
perfectionem speciei humanae, vel quod includeret perfectionem speciei humanae
tamquam partem sui; et neutrum iIIorum est verum."
13Aquinas, Summa Theologiae 1.14.1I.r.
14Aristotle Categories 3b37-9.
286 ALLAN BACK

and philosophical import of this sophism. There is left for us to judge the
contest. 15

In brief, I hold that Burleigh has the better of Ockham in this contest,
as stated, but that in general the nominalist position wins out over the
realist.

Ockham has had to conclude that, strictly speaking, 'Man is the

worthiest of creatures' is false. He has managed to give an account
according to which it is true only by rewriting the sentence radically. In
contrast, Burleigh has offered an account that saves the phenomenon of the
commonly accepted truth of the sentence. Now it may not be obvious that
this sentence is a phenomenon worth saving. Yet there are many other
sentences of the same logical type that have less obscurity, and abound in
ordinary and in scientific discourse: e.g. 'The whale is the largest
mammal'; 'The atom is not the smallest material particle'; 'The roach has
great survival value'. Such have come to be called generic descriptions in
modern discussion. We seem to reckon some such statements as
meaningful and even as true.

However, Ockham's analysis clearly does not suit such examples.

E.g., take 'The whale is the largest mammal'. According to Ockham, by
analogy with his analysis given above, this claims that every whale is
larger than every other mammal. But that is false: there are large elephants
and baby and dwarf whales. So once more Ockham's approach gives
unsatisfactory results.

Perhaps a more elaborate nominalist paraphrase of these generic

descriptions, that reduces their significance to talk of individuals and thus
that makes their supposition personal will succeed. Indeed, I shall suggest
such a paraphrase below. Still, I claim that in the historical debate between
Ockham and Burleigh the latter, realist approach has won out. At the least,
it has recognized the importance and the distinctive features of generic
descriptions. Ockham appears less sensitive.

Still, the realist approach complicates the ontology. It advocates a new

type of abstract entity that in effect is a mere mirror of the grammar of
sentences that might be offered to explicate generic descriptions.

So I am inclined to look at nominalist analyses further. Let me merely

propose my candidates here (with some simplification). 'The whale' in
'The whale is the largest mammal' refers to that individual whale that is
larger than any other mammal that is not a whale. Why then the generality
of the generic description? Which whale is the largest will change over
time; indeed, may change frequently over time. Consider that the period of
time for which the assertion is supposed to hold is a rather long one, and
during that period the largest mammal has consistently been a whale. This
situation suggests that there is some genetic regularity in the nature of
whales that causes them to be quite large. Moreover, generally, when one

l5What follows is quite brief. I develop a fuller analysis in a longer version of this
paper.
 WHO IS THE WORTHIESF OF THEM AIL? 287

individual of a species has a certain characteristic, like large size, naturally,

other individuals of that species tend to share in that characteristic to a
greater or lesser extent.

Likewise, 'man' in 'Man is the worthiest of creatures' would refer to

some man at every time. It is not that man is the only creature having any
worth; otherwise the original claim would be vacuous. Indeed, it was
commonly held that there are worthy candidates from other species;
compare Anselm's hierarchy of goodness in the Monoiogion. It makes no
sense to say that S is the most P, if S alone can be P. Rather, at any time
when human beings exist, there is a human being having more worth than
any other being in the domain of beings having worth. Who this person
might be could vary from minute to minute. Indeed, such variance would
strengthen the need to talk of human beings in general, and would
emphasize our common humanity.

Like Ockbam's analysis, this analysis does require paraphrasing and

rewriting the original sentence. But consider that utterances in ordinary
language are generally imprecise and fuzzy. Further, I find this approach
preferable to complicating the ontology in an ad hoc manner simply
because a sentence resists a straightforward analysis. So I end up favoring
Ockham's method although not his results.

}(utzto}Vn lJniversi~
 Albert de Saxe et les sophismes de I'in/ini
par Joel Biard

Albert de Saxe traite l'infini comme un probleme de philosophie

naturelle dans ses Questions sur la Physique et dans ses Questions sur Ie
Traite du ciel, mais ill'examine aussi, dans ses Sophismes, comme un
terme facteur d'ambigulte au sein d'une proposition. Ainsi, alors que Jean
Buridan n'accordait aucune place particuliere au terme 'infini' dans son
propre recueil, Albert y consacre douze sophismes.t De ce fait, il examine
principalement l'infini en liaison avec une reftexion sur les
syncategoremes, puisque tel est explicitement Ie champ que, d'une maniere
fort traditionnelle, il assigne aux Sophismata.2

lci convergent donc deux lignes de traitement de l'infini. D'une part,

l'etude de l'infini dans Ie cadre des commentaires sur la physique
aristotelicienne; centree sur Ie continu, elle s'est considerablement
developpee a partir de la seconde moitie du XIne siec1e et Albert de Saxe en
re~oit l'heritage par Jean Buridan. D'autre part, l'analyse de l'infini comme
terme syncategorematique, qui a donne lieu a des reftexions sur l'infini
cou¢es de toute theorie proprement physique. Un traite De solutionibus
sophismatum, date de 1200 environ, examine deja la phrase ambigue
'Infinita suntfinita', et introduit Ie principe que Ie terme 'infini' peut etre
pris comme chose ou bien comme mode ou signe.3 Des Ie traite De
sincategoreumatibus attribue a Robert Bacon (ca. 1200-1210), il est dit que
Ie terme 'infini' peut etre un categoreme - auquel cas il signifie ce en
dehors de quoi il y a toujours quelque chose - ou un syncategoreme -
auquel cas il ne signifie pas une infinite reelle mais l'infinite du sujet en
relation avec Ie predicat, autrement dit il signifie que Ie sujet est dans un
rapport qui n'est jamais termine avec Ie predicat.4 Et c'est grace a cette
distinction que se resout l' ambiguYte de la phrase 'infinita sunt finita'. Par
suite, on retrouve des formulations tres voisines dans Ie traite sur les
syncategoremes de l'anglais Guillaume de Sherwood,5 vers 1230-1240, et

lL'edition la plus accessible est I'edition de Paris, 1502, reproduite par Georg Olms
Verlag, Hildesheim-New York, 1975; cette edition etant souvent fautive, Ie texte sera
cite 11 partir du manuscrit latin 16 134 de la Bibliotheque nationale de Paris. La
premiere reference sera celie du manuscrit, la seconde, entre parentheses, celie de
I'edition incunable.
2ef. Sophismata, BN Lat. 16 134, f. Ira (Paris 1502: sign. a II, ra): "~b rogatum
quorumdam scholarium, deo !avente, quedam conscribam sophismata ex parte
diversorum sincathegoreumatum difficultatem habentia [...J;" par suite, I'ouvrage est
divise selon les syncategoremes concernes.
3ef. Tractatus jiorianus de solutionibus sophismatum, in Some Earlier Parisians Tracts
on distinctiones sophismatum, ed. L.M. De Rijk, Nijmegen: Ingenium 1988, pp. 49-
145: voir pp. 57-8.
4Des extraits de ce texte ont ete publies par H.A.G. Braakhuis, De J3de eeuwse tractaten
over syncategorematische termen, 2 vol., Meppel: Krips repro. 1979, vol. I, pp. 106-
67; voir en part. pp. 126-7.
5ef. Guillaume de Sherwood, Syncategoremata, ed. J. R. O'Donnell, Mediteval Studies,
3, 1941, pp. 46-93, en part. pp. 54-5.

288
 ALBERT DE SAXE ET LES SOPHISMES DE L'lNFlNl 289

cette tradition se prolonge, enrlchie et renouvelee, jusque dans les traites

sur les sophismes des calculateurs anglais du XIVe siecle.

A Paris, Jean Burldan reprend cette distinction entre l'usage

categorematique et I'usage syncategorematique du terme 'infini' dans ses
Questions sur fa Physique. Cet ouvrage propose en meme temps des
definitions auxquelles bon nombre de discussions du XIVe siecle vont se
referer et dont on trouve I'echo dans Ie recueil de sophismes d'Albert. Mais
a la difference de son maitre, Albert de Saxe, sans doute dans Ie
prolongement direct de Guillaume Heytesbury plus que de I'ancienne
tradition parisienne, examine beaucoup plus systematiquement (et en
visant, comme il en est coutumier, une certaine completude) l'infini comme
terme facteur d'ambigui'te et, par consequent, Ie cadre logico-linguistique
de toute discussion possible sur I'infini.

La presente etude concemera les sophismes UI a LXIII de la premiere

partie des Sophismata, qui tous, sauf Ie premier, contiennent explicitement
sous une forme ou sous une autre Ie terme 'infini'.

Quelles definitions de l'infini ?

De nombreux elements definitionnels sont fournis, d'abord a
I'occasion du sophisme LII, 'Quotlibet entia suntfinita' -puisque Ie terme
'infini' a un rapport avec les expressions syncategorematiques 'aussi grand
que I'on veut'et 'aussi nombreux que I'on veut' - puis dans Ie sophisme
LIII, 'Infinita sunt finita'. C'est en premier lieu une serle de cinq 'modes'
selon lesquels se dit l'infini;6 en deuxieme lieu, greffee sur Ie cinquieme de
ces modes, une division quadripartite de I'infini, fondee sur Ie double
crltere du continu et du discret d'une part, de la composition et de la
division d'autre part; enfin une serle de definitions qui sont proches des
formulations buridaniennes, et qui presupposent la distinction - precisee a
cette occasion - du sens categorematique et du sens syncategorematique de
I'infini.

Certains des sens enumeres dans la premiere serle ne presentent qu 'un

faible interet; un tel inventaire traduit surtout, d'emblee, un souci de
completude dans la presentation de I'infini. Les quatre premiers sont relies,
de f~on plus ou moins lache, a la notion aristorelicienne de I'infini comme
"ce qui ne peut etre parcouru."

Le cinquieme mode, I'infini selon la composition et la division,

souleve en revanche des questions rlchement debattues au XIve siecle. Le
premier exemple invoque est celui du continu, par consequent l'infini selon
la division, puis symetrlquement I'infini selon la composition (ou
I'apposition). Dans les deux cas, il ne semble etre ici question que d'infini
en puissance.1 C'est a cette occasion qu'est introduite une nouvelle
division des sens de l'infini, nous rapprochant des reftexions et des debats
du XIve siecle tandis que les sens precedents ne constituaient guere que des
preliminaires.

6Cf. Albert de Saxe, Sophismata, ff. 18vb-19va (sign. d VI, vb - d VII, rb).
7Cf. ibid., f. 19ra (sign. d VII, ra): "[...J isto modo dicitur infinitum in potentiam."
290 JOEL BIARD

Le point de depart est que Ie fini et I'infini concernent la quantite. La

quantite etant so it continue so it discrete, et I'infini pouvant se dire selon la
composition ou la division, on obtient quatre possibilites.

L'infini selon la grandeur par division est Ie continuo A son sujet,

surgissent surtout des problemes de philo sophie naturelle, mais deux
sophismes lui seront consacres ulterieurement. L'infini selon la grandeur et
selon la composition ne peut pas exister dans la nature, conformement a ce
que dit Aristote, pas plus que l'infini selon Ie nombre, qui appelle quelques
precisions puisqu'il peut etre entendu de deux fa<;ons. Ce peut etre une
infinite de choses actuellement distinctes les unes des autres, ne faisant pas
un par elles-memes, et de la sorte il est impossible qu'il y ait un infini selon
Ie nombre. En un autre sens, sont dites infinies selon Ie nombre des choses
qui ne sont pas actuellement distinctes les unes des autres, bien qu'elles
soient separables, et un tel infini est dit infini selon Ie nombre en
puissance. 8

Cette division quadripartite ne fait donc finalement que reformuler

I'opposition de l'acte et de la puissance et Ie refus de I'infini en acte, si bien
que I'infini ne se rencontre (en puissance) que dans Ie continu: soit
directement comme I'infini par division selon la grandeur, soit
indirectement - au prix d'une confusion omnipresente dans les reflexions
de cette epoque - comme Ie nombre infini (en puissance) des parties
proportionnelles de ce meme continu.9

Mais a l'occasion du sophisme LID, Albert de Saxe examine plusieurs

definitions courantes de I'infini, y compris certaines formulations tres
proches de celles qui avaient ete proposees par Jean Buridan ou par
Gregoire de Rimini. C' est la qu' Albert de Saxe introduit la difference entre
Ie sens categorematique et Ie sens syncategorematique de I'infini, la
presentant comme centrale pour la resolution du sophisme.1 0 Cette
distinction, dont nous avons deja dit que les premices remontent au debut
du xme siecie, est presentee comme "communement re<;ue" par Guillaume
Heytesbury,11 et Jean Buridan en a codifie I'emploi pour la physique
parisienne. Chacun de ces sens fait I'objet d'une ou de plusieurs
"expositions".

En ce qui concerne Ie sens syncategorematique, Albert de Saxe

examine quatre definitions, qui sont centrees sur la definition de l'infini
selon Ie nombre plus que selon la grandeur.

La premiere propose I' exposition suivante:

Aliqui dicunt quod debet exponi pro ly non tot quin plura. 12

8Cf. ibid., ff. 18va-19ra (sign. d VII, ra).

9Cf. ibid., f. 19ra (sign. d VII, ra-rb).
IOCf. ibid., f. 19ra (sign. d VII, rb).
IICf. Guillaume Heytesbury, Sophismata, XVIII, in Tractatus de sensu composito et
diviso, Regule eiusdem cum Sophismatibus, Venise 1494, f. l30ra.
12Albert de Saxe, Sophismata, f. 19ra (sign. d VII, rb)
 ALBERT DE SAXE ET LES SOPHlSMES DE L'INFlNl 291

Une telle definition reprend partiellement celIe de Jean Buridan. qui

definissait l'infini selon la grandeur aliquantum .et non tantum quin maius.
et I'infini selon Ie nombre aliquanta et non tot quin plura l3 -encore qU'une
partie de I' expression ne se retrouve pas explicitement. Cette definition
semble devenue courante a Paris. puisqu' elle est aussi mentionnee par
Gregoire de Rimini. qui ne la retient pas.l 4 Par ailleurs. sans en faire une
definition. Guillaume Heytesbury se servait d'une expression voisine au
debut du sophisme XVIII de SOIl recueil. I5

Albert de Saxe va alleguer deux critiques possibles de cette definition.

La premiere repose sur une objection qu'il ne va pas retenir: 'Non tot quin
plura sunt finita' est une proposition negative. tandis que 'infinita sunt
finita' est une proposition affirmative; or on ne peut pas inferer une
affirmative d'une negative. si bien que les deux propositions ne sont pas
convertibles. Albert ne se satisfait pas de ceUe refutation car. selon lui.
'non tot quin plura' n'est pas une expression negative mais. 'selon la fa~on
de dire des anciens. simplement adversative·.l 6 II est vrai neanmoins que
les deux expressions ne sont pas convertibles,11 quoique pour une autre
raison. comme Ie montre un exemple: 'On ne peut pas inferer "l'infini est
deux. donc pas tellement nombreux qu'il ne soit de plus grand nombre
[non tot quin plura] est deux ... • Une telle exposition ne doit donc pas etre
retenue. Ce raisonnement suppose que I'infinite numerique comprenne en
elle chaque nombre. ainsi que d'autres definitions l'expliciteront. alors que
cette premiere definition insistait sur I' iteration au dela de tout nombre
(fini).

La deuxieme exposition ne fait que reprendre la definition la plus

usuelle de l'infinite numerique. entendue comme la suite indefinie des
nombres finis:

Ideo alii dicunt quod ly infinita sumptum sincathegoreumatice

debet sic exponi: infinita, id est duo. tria. quattuor. etc.l s

13Cf. Quaestiones in libros Physicorum Aristotelis, III, quo 18: "Sequitur dicere de
infinito syncathegoreumatice sumpto, de quo notandum est quod diversis modis soli
exponi hoc nomen 'infinitum' syncathegoreumatice sumptum. Uno modo in
magnitudinibus, quia aliquantum et non tantum quin maius, et de multitudine, quia
aliquanta et non tot quin plura ..." (Paris, 1509, f. LXIvb).
14Cf. Gregoire de Rimini, Lectura super primum et secundum Sententiarum, II, dist. 2,
quo 2, art. 1, ed. A. D. Trapp et V. Marcolino, vol. V, Berlin-New York 1979, p. 294.
15Cf. Guillaume Heytesbury, Sophismata, XVIII, f. l3Ova.
16Bien que Ie manuscrit de Paris parle des "anciens logiciens", cette notion est plut6t
grammaticale. Priscien la cite en effet parmi les dix-sept especes de conjonctions;
cf.lnstitutiones grammaticae, XVII, I, 1, ed. M. Hertz, «grammatici latini» II et III,
Lipsiae, 1855-1859, vol. II, p. 93, I. 13-16), et iI 1a definit comme suit: "Adversativae
sunt, quae adversum convenienti significant, ut 'tamen', 'quamquam', 'quamvis', 'etsi',
'etiamsi'. 'saltem'" (ibid., XVII, 11,9, p. 99, I. 12-13).
17Albert de Saxe, Sophismata, f. 19ra (sign. d VII, va).
IS/bid., f. 19 ra-rb (sign. d VII, va). Cf. Henri de Gand, Syncathegoreumata, ed. H.A.G.
Braakhuis, p. 353: "/NF/NITA SUNT F/N/TA. Probatio: duo sunt finita, tria sunt
finita. et sic in infinitum; ergo infinita sunt finita."
292 JOEL BIARD

Cette definition suscite une objection de meme nature que la precedente: les
expressions ne sont pas convertibles. En effet, on ne peut pas inferer:
'L'infini est deux, donc deux sont deux, trois sont deux, etc.' La verite de
I'antecedent, qui n'est pas mise en doute, suppose la possibilite d'une
distribution du c6te du predicat, mais une distribution similaire ne saurait
etre effectuee du c6te du sujet sans dissoudre I'acte meme de designation
de cette sommation indefinie. D'autres illustrations (comme celle d'une
infinite de bateaux Hree par des hommes 19 ) viennent confirmer
!'impossibilite d'une telle fragmentation.

La troisieme exposition ne renvoie pas simplement a la suite des

nombres, elIe pose un exces par rapport a un quelconque nombre fini
donne:
Adhuc alii tertia exponunt ly infinita si'ncathegorematice dicentes
'infinita, id est duobus plura, et tribus plura, et sic de aliis. '20

Une telIe definition insiste sur Ie fait que l'infini est toujours au-dela.
L'idee n'est certes pas nouvelle. ElIe se trouve d'une certaine maniere chez
Guillaume de Sherwood ou chez Pierre d'Espagne; on la retrouve chez
Guillaume Heytesbury.2 1 Ce pourrait etre la simple reprise d'une
caracterisation banale de l'infinite en puissance. Toutefois, cette
formulation, comme celle proposee par Gregoire de Rimini,22 n'en reste
pas a une caracterisation negative de !'infini mais esquisse une position de
l'infini. En regard de celIe de Gregoire, toutefois, la presente formulation
est moins explicite, moins forte.

Albert de Saxe critique encore cette definition a partir d'un exemple

analogue a celui qui avait ete utilise a propos de la precedente exposition:
de ce que l'infini est deux, on ne peut pas inferer que plus que deux est
deux, plus que trois est deux, etc. En verite, il y a la bien des confusions,
puisque poser l'infini comme ce qui est au dela de tout fini devrait interdire
de telles distributions; mais Albert tient que la verite de la proposition
'infinita sunt duo' est patente, puisque Ie predicat serait contenu sous
!'infinite du sujet.

Entin, une quatrieme exposition est indiquee:

Quarto potest exponi infinita, id est de qualibet specie numeri
aliquot entia.23

19yariation sur un vieil exemple: cf. Guillaume de Sherwood, Syncategoremata, p. 54.

20Albert de Saxe, Sophismata, f. 19rb (sign. d VII, va).
21Cf. Guillaume de Sherwood, Syncategoremata, p. 55: "Item sophisma: infinita sunt
finita. Probatio: duobus plura suntfinita et tribus et sic deinceps; ergo infinita etc.;" cf.
aussi Pierre d'Espagne, Tractatus called afterwards Summule logicales, XII, M. L.M.
de Rijk, Assen: Van Gorcum 1972, p. 231, ou Guillaume Heytesbury, Sophismata,
XVIII, f. 130va.
22Cf. Gregoire de Rimini, loc. cit., p. 294: "Sed puto quod magis proprie diceretur
'quantocumque finito maius' vel 'quotcumque finitis plura' ."
23Albert de Saxe, Sophismata, f. 19rb (sign. d VII, va).

\
 ALBERT DE SAXE ET LES SOPHlSMES DE L'INFINI 293

Le developpement qui suit indique que par 'etres d'une espece de nombre',
on peut entendre par exemple une trinire de choses. Autrement dit, il s'agit
la d'une definition qui considere Ie nombre comme defini a partir de
classes d'objets, et l'infini comme defini a partir de l'indetermination de
'n'importe lequel'. lei encore, la meme objection est faite, appuyee sur Ie
meme raisonnement que precedemment.

Mais apres avoir recuse chacune de ces definitions comme n' elant pas
convertibles avec Ie defini, Albert ajoute que 'de n'importe laquelle [de ces
definitions] au terme "infini" la consequence est bonne', et c'est la ce qui
caracterise son analyse. Chacune de ces expressions, par consequent,
designe en un certain sens l'infini, sans qu'aucune ne Ie definisse
completement. Toutefois, les quatre prises ensemble procurent une
definition satisfaisante de l'infini:
[. ..J quattuor dicte expositiones simul sub disiunctione equivalent
Ly infinita, ita quod aliquando alicui attribuatur ei una earum,
aliquando alia, secundum exigentiam predicati propositionis in qua
ponitur £.. .J. Aliquando est verificatio pro una illarum expositionum,
aliquando pro alia £.. .].24

S'agit-il d'un simple eclectisme, ou d'une conception de l'infini cherchant

confusement a en faire la totalisation des differents infinis entrevus en
chacune de ces definitions? Albert, en tout cas, assigne a chacune des
propositions precedentes un sens qui permet de la verifier.

Le sens categorematique de l'infini est traite beaucoup plus

succinctement. En ce sens, en effet, l'infini designe 'des etres qui sont
infinis', soit selon la grandeur, soit selon Ie nombre. Cette acception est
donc beaucoup plus simple, que 1'0n accepte ou non l'idee que quelque
chose puisse correspondre a une telle infinite au sens categorematique.
Visiblement, a la suite de Buridan, c'est l'infini au sens syncaregorema-
tique qui suscite Ie plus de reftexions.

Quelques problemes souleves a propos de I'in/ini

1. Variations sur Ie rapport du jini et de l'injini dans La proposition

Dans Ie sophisme 'lnfinita suntfinita', Guillaume Heytesbury avait

evoque differentes combinaisons de ces deux termes au sein d'une
proposition. Ce faisant, il avait pose des regles limitant les variations
possibles et codifiant la maniere dont chacun de ces termes doit etre
entendu, en fonction de sa place dans la proposition. Apres avoir rappele la
distinction, communement re~ue, entre Ie sens caregorematique et Ie sens
syncaregorematique de l'infini, Guillaume Heytesbury jugeait en effet que
Ie sophisme etait vrai seulement selon l'usage syncategorematique de ce
terme, mais faux selon son usage categorematique. Ce demier reviendrait
en effet a dire que 'alique que sunt infinita sunt finita', ce qui n' est pas

24Ibid.• f. 19rb-va (sign. d VII, vb).

294 JOEL BIARD

possible. Ce n'est donc pas ainsi qu'it faut comprendre l'enonce

sophismatique, ce qui Ie conduisait a fonnuler les regles suivantes:
[... J quando iste terminus 'infinita' ponitur a parte subiecti et
precedit propositionem aliquam, nullo determinabili precedente
ipsum, tunc iste terminus tenetur sincathegoreumatice et non
cathegoreumatice, nisi disputans velit omnino uti sic isto termino.
Sed universaliter, quando iste terminus 'infinita' vel aliquis talis
ponitur a parte predicati alicuius propositionis, tunc iuxta communem
modum loquendi tenetur cathegoreumatice et non
sincathegoreumatice.25

La ou Ie maitre anglais s'en tenait a l'examen d'un seul sophisme, Albert

de Saxe multiplie lesphrases a examiner, ajoutanta '[nfinita sunt finita'
deux autres propositions ambigues: 'Finita sunt infinita' (sophisme LIV),
et '[nfinita sunt infinita' (sophisme LV). Ce faisant, it paralt tenir les
contraintes pour moins rigoureuses que ne Ie faisait Guillaume
Heytesbury.

Deja les reponses donnees par Albert au sophisme Lm contreviennent

partiellement aux exigences logico-linguistiques posees par Guillaume
Heytesbury. En combinant les deux distinctions, du sens categorematique
et du sens syncategorematique d'une part, de la grandeur et de l'etendue
d'autre part, trois cas peuvent etre distingues concernant Ie sophisme
'[nfinita suntfinita'.

Premierement, en prenant l'infini au sens categorematique selon Ie

nombre et Ie fini selon la grandeur,le sens de la phrase 'infinita sunt finita'
est que toutes les parties du continu prises collectivement sont infinies
selon Ie nombre et finies selon la grandeur; de la sorte, Ie sophisme est
vrai. On notera au passage qu' Albert de Saxe accepte l'idee que 'toutes les
parties du continu' soit une expression ayant un sens lorsqu'elle est prise
collectivement, ou, plus preeisement, une expression designant quelque
chose - puisque Ie sujet doit supposer pour quelque chose afin que la
proposition affinnative soit vraie. Cela se confinnera dans les sophismes
consacres plus loin au continu, ainsi que dans Ie sophisme 'Omne totum
est maius sua parte' qui distingue differents sens du tout. 26 Or Jean
Burid~, pour sa part, recusait l'idee que l'expression 'toutes les parties'
puisse etre prise en un sens collectif pour designer toutes les parties du
continu.27 Plus tard, l'acceptation du sens collectif de cette expression
pennettra de conferer une signification reelle a l'infini dans son acception
categorematique. Mais deja Gregoire de Rimini jugeait faux de dire que Ie
mot 'toutes', pris collectivement, ne peut etre ajoute a un tenne qui
suppose pour des choses en nombre infini. lei, Ie fait qu' Albert de Saxe
accepte l'expression 'toutes les parties' comme sujet d'une proposition
vraie, et qu'it admette qu'elles sont en nombre infini semblerait temoigner

250uiIlaume Heytesbury, Sophismata, XVIII, f. 130va.

26 Albert de Saxe, Sophismata, Ia pars, soph. XLIX, f. 16ra-17vb (sign. d III, va - d V,
rb).
27Cf. Jean Buridan, Quaestiones in octo libros Physicorum Aristotelis, III, quo 16, f.
XLIXva.
 ALBERT DE SAXE ET LES SOPHISMES DE L'lNFlNl 295

qu'it reprend implicitement a son compte, au moins partiellement, les

a
argumentations qui avaient conduit Gregoire de Rimini accorder que I' on
peut concevoir l'etre en acte de l'infini. Mais Albert - et sans doute
manque-t-il ici de coherence - ne va pas jusqu 'a tirer cette conclusion.

Deuxiemement, si dans la proposition 'infinita suntfinita', l'infini est

pris de maniere categorematique selon la grandeur, tandis que Ie fini est
pris aussi selon la grandeur, la proposition est a l'evidence fausse
puisqu 'un terme est predique de son oppose.

Troisiemement, si l'infini est pris de maniere syncategorematique, la

proposition est vraie. La justification fait encore appel aux parties
proportionnelles du continu, cette fois d'une f~on fort classique puisque
ces parties proportionnelles, infinies au sens syncategorematique, sont
finies. Dire qu'elles sont infinies au sens syncategorematique, c'est
simplement dire que Ie terme 'infini' modifie la signification des termes
categorematiques de la proposition. Ainsi, les choses finies sont aussi
nombreuses que I'on veut.

Pour revenir a notre propos, en tenant pour vraie la proposition

'infinita sunt finita' lorsque I'infini est entendu au sens categorematique
selon la grandeur, Albert de Saxe contrevient a la premiere regIe enoncee
par Guillaume Heytesbury.

Le sophisme suivant, quant a lui, contrevient a l'autre regIe, puisqu'il

prend l'infini de maniere syncategorematique du cote du predicat. Ce qui
permet, en effet, de tenir pour vraie la proposition 'finita sunt infinita',
c'est l'induction du cote du predicat a partir de "deux demi sont trois tiers,
et deux demi sont quatre quarts, et ainsi a I'infini..." Le predicat est donc
entendu - l'auteur prend so in de Ie preciser28 - en un sens
syncategorematique.

Le sophisme suivant procMe a une induction analogue du cote du

sujet: deux moities sont infinies, en vertu de ce qui a ete demontre
precedemment, trois tiers sont infinis, etc., donc 'infinita sunt infinita' .29
Ici encore Ie predicat doit etre entendu selon I'usage syncategorematique,
c'est pourquoi on peut tenir pour vraies ala fois 'infinita sunt finita' et
'infinita sunt infinita'. 11 resterait toutefois a preciser si Ie sujet est ici a
entendre de maniere categorematique ou de maniere syncategorematique,
ce qu' Albert ne fait pas. Le renvoi au sophisme precedent semblerait
plaider pour une comprehension syncatcSgorematique de I'infini, y compris
du cote du sujet; mais it est difficile de concevoir quel sens aurait une
proposition dont Ie sujet et Ie predicat seraient tous deux entendus selon
I'usage syncategorematique du terme, a plus forte raison comment une telle
proposition pourrait etre vraie, d'apres les principes de la semiologie et de
la theorie de la verite qui sont ceux d'Albert de Saxe. Puisque certains
passages du sophisme LID autorisaient une comprehension de I'infini au
sens categorematique en position de sujet (liee a une apprehension
collective des parties d'un tout), tel semble devoir etre ici Ie cas puisque

28Cf. Albert de Saxe, Sophismata, f. 20ra (sign. d VIII, va).

291bid.
296 JOEL BIARD

1'0n procede a une sorte de passage a la limite de la division

potentiellement infinie en designant les parties comme infinies.

En fin de compte, tout en multipliant les enonces combinant les termes

'finita' et 'infinita', Albert de Saxe reste assez succinct, et est plutot moins
precis que ne l'etait Guillaume Heytesbury. Voulant multiplier les
variations, il fait 6clater Ie cadre, plus rigide mais plus Tigoureux, qu'avait
fixe Ie maitre anglais. Ce faisant, il autorise une comprehension
categorematique de l'infini en situation de sujet logico-linguistique, sans
paraitre toutefois en percevoir toutes les implications concernant la position
d'un etre infini verifiant la proposition. Ce probleme va se retrouver a
propos du nombre et a propos du continuo

2. Infinite, unite et pluralite.

Trois sophismes sont consacres aux relations de l'unite, de la pluralite

ou du nombre, et de l'infinite. Parmi eux, c'est Ie troisieme qui presente Ie
plus d'interet.

Le premier (c'est-a-dire Ie sophisme LVI: 'Infinite unitates sunt in

aliquo numero finito') n'apporte Tien de nouveau. Pour etre vraie selon
l'acception syncategorematique de l'infini, la phrase suppose seulement
que 1'0n puisse continuer indefiniment la serie des nombres, en ayant
toujours affaire a un nombre fini.

Le deuxieme (sophisme LVII) souleve une question plus interessante

en examinant la phrase 'Infinitis infinita sunt plura', soit: par rapport a une
infinite, une infinite est plus nombreuse. Le sophisme etant tenu pour vrai,
la question surgit de savoir si une pluralite d'infinis est ici impliquee,
parmi lesquels certains seraient plus grands que d'autres.

L'argumentation semble montrer que non, puisque l'infini doit etre

entendu de maniere syncategorematique pour que l'enonce soit vrai:
[...J concluditur quod infinitis infinita sunt plura. et etiam infinita
sunt plura in/mitis. et hoc accipiendo ly in/mitis sincathegoreumatice.
Sed utrum accipiendo ipsum cathegoreumatice secundum
multitudinem. dubium est an hec sit concedenda: 'infinita sunt plura
infinitis' [ ...J.30

Le seul argument ad oppositum evoque la possibilite d'une inference

entre 'infinitis infinita sunt plura' et 'infinita sunt plura infinitis', et juge
que 'Ie consequent parrot faux', si bien que l'antecedent ne saurait etre
vrai. Cet argument fait l'objet d'une refutation nuancee: l'inference n'est
pas formellement valide, cependant 'en vertu de la matiere de l'argument'
Ie consequent, a savoir 'infinita sunt plura infinitis', peut etre admis.
L 'auteur Ie demontre en deux temps: d' abord, deux moities sont trois tiers
qui sont plus que deux, deux moities sont quatre quart, qui sont plus que
trois, etc.; ensuite, deux moities sont plus qu'une infinite, trois tiers sont
plus qu'une infinite, etc, donc une infinite est plus qu'une infinite. Ainsi,

30/bid., f. 20va (sign. e I, fa).

 ALBERT DE SAXE ET LES SOPHlSMES DE L'INFINI 297

on pourrait aussi bien admettre 'infinitis inftnita sunt plura' que 'infinita
sunt plura infinitis'. Mais a l'evidence, ceci n'a de sens qu'en prenant
I'infini de maniere syncategorematique.

On resterait sur sa faim concernant I'infinite numerique entendue de

maniere categorematique, si Ie probleme n'etait repris dans Ie sophisme
suivant (sophisme LVIII): 'Inftnita sunt plura infinitis'. Puisque cet enonce
ambigu a deja ete evoque a l'occasion du sophisme precedent, I'examen va
s'en tenir a un aspect limite mais particulierement interessant: il s'agit de
savoir si par rapport a un nombre infini, il pourrait y avoir un nombre plus
grand, 'utrum multitudine infinita possit esse multitudo maior'. 31 Cette
fois, il s'agit bien d'esquisser une reftexion sur la possibilite d'un nombre
infini et sur la comparaison d'infinis. Sans aucun doute, cette seule
hypothese, dfit-elle etre finalement recusee, est marquee par les reftexions
de Gregoire de Rimini.

Albert de Saxe, pour sa part, refuse clairement toute relation

d'inegalite entre deux infinis:
Pro responsione sophismatis, vola probare quod una multitudo
infinita non potest esse maior alia. 32

Albert souleve Ie meme probleme dans ses Questions sur Ie Traite du ciel.
La, pour eviter les paradoxes de l'infini, suscites en particulier par
I'hypothese de l'etemite du monde, il nie que 1'0n puisse appliquer les
relations d'egalite ou d'inegalite a differents infinis.3 3 Examinant ici la
question d'un point de vue logique, il montre que l'id6e selon laquelle un
infini pourrait etre plus grand qu'un autre est absurde. Cela Ie conduit a
tenir Ie sophisme pour faux, des lors que Ie terme 'infini' y est entendu de
maniere categorematique:
Ad sophisma respondeo quod ipsum est falsum, capiendo ly
infinita cathegoreumatice, et similiter ly infinitis. 34

Le premier argument est un lieu classique de la reftexion sur l'infini en

relation avec la cosmologie aristotelicienne et les commentateurs arabes. Si
l'on suppose Ie monde infini, comme c'est Ie cas pour Aristote, il s'est
ecoule une infinite de jours et une infinite d'annees. Ainsi, comme il y a
plus de jours que d'annees, on obtiendrait deux infinites dont l'une serait
plus nombreuse que l'autre.

L'argument ad oppositum juge tout simplement inacceptable l'idee

qu'un infini puisse en exceder un autre. L'argumentation assez detaillee
par laquelle Albert de Saxe justifie sa position aborde la question de la
divisibilite a l'infini du continu, qui sera traitee plus longuement dans les
deux sophismes suivants. Le raisonnement est assez contoume et, me
semble-t-il, peu rigoureux. Albert veut montrer que si I'on acceptait Ie

31/bid.
32lbid.
33Cf. Albert de Saxe, Questiones super quattuor libros Aristotelis de celo et mundo,
Venise 1492, I, quo X, sign. B VI, ra.
34Sophismata, f. 20va (sign. e I, rb).
298 JOEL BIARD

sophisme, l'indivisible serait divisible. II evoque a cet effet un probleme

classique, herite d' Aristote et discute par de nombreux Medievaux en
relation a la question de la divisibilite du continu: il s'agit de deux mobiles
c et d, se deplacrant en un temps egal sur deux lignes a et b, dont l'une est
double de l'autre.35 Selon Albert, si I'on admet qu'une infinite est plus
grande qu'une autre, il faut admettre que Ie point (mobile) b a traverse
deux fois plus de points que Ie point (mobile) a. Mais comme les deux
parcours ont pris Ie meme temps, donc Ie meme nombre d'instants, il faut
en conclure qu'au moment ou Ie point c cO'incidait avec un point (donc un
indivisible) de la Jigne a, Ie point d devait co'incider avec deux points de la
ligne b. L'argument est confirme par Ie fait que tous les instants de
l'intervalle de temps e, pris ensemble, doivent etre egaux a tous les points
de la ligne a, pris ensemble, mais devraient aussi etre egaux a tous les
points composant la ligne b, pris ensemble; il s'ensuivrait que que la
collection des points de la ligne a serait egale a la collection des points de la
a
ligne b, puisqu'egales une meme troisieme. Mais tel n'est pas Ie cas si
une infinite peut etre plus grande qu'une autre.

Le confusion vient ici d'une apprehension peu rigoureuse de

l'indivisible. Albert de Saxe semble raisonner dans l'hypothese d'une
infinite d'indivisibles compos ant Ie continu, selon la these soutenue par
exemple par Henri de Harclay,36 et toute sa demonstration toume autour
de l'idee qu'il ne peut pas y avoir deux infinires de cette sorte dont l'une
serait plus grande que I'autre. Mais cette demonstration se mene en
appJiquant des operations qui n'ont de sens que pour comparer des
grandeurs finies. Or c'est precisement ce qui conduisait Henri de Harclay a
distinguer deux sens de la relation entre Ie tout et les parties, selon une idee
a laquelle Gregoire de Rimini a ajoute l' exigence de distinguer aussi, pour
etre coherent, plusieurs sens de la relation d'inegalite. Albert, face au
meme probleme et partant de formulations voisines, delaisse ces
innovations.

Cette proximite avec les questions traitees par Gregoire de Rimini dans
son Commentaire des Sentences, en meme temps qu'une relative cecite aux
innovations de ce dernier, sont confirmees par un autre argument,
considerant l'exces d'un infini sur un autre. Tout quantum superieur a un
autre quantum est divisible en une partie par laquelle il depasse I' autre, et
une partie qui est egale a ce qu'il excMe. Supposons a partir de la qu'un
infini soit superieur a un autre infini. Nommons a l'infini qui est depasse et
b l'infini qui depasse Ie premier. L'infini b peut alors etre divise en c, qui
est la partie par laquelle il depasse l'autre, et en d, qui est egal a l'infini
ainsi depasse. Alors, du fait que d est depasse par c, il s' ensuit que d est
fini. Or I'on a pose qu'il etait infini. Donc I'infini est fini. Cette
contradiction dans les termes suffit a recuser l'idee qu'un infini puisse etre
plus grand qu'un autre infini.

35Cf. Aristote, Physique, VI, 2, 232 a 23 - 233 a 9.

36Henri de Harclay traite lui-mame ce probleme dans une perspective indivisibiliste: cf.
J. Murdoch, "Henry of Harclay and the Infinite", in Studi sui XlV secolo in memoria di
Anneliese Maier, ed. A. Maieril et A. Paravicini-Bagliani, Roma 1981, pp. 219-60, en
part. pp. 250 sqq.
 ALBERT DE SAXE ET LES SOPHISMES DE L'INFINI 299

La encore, Albert de Saxe part d'un paradoxe classique suscite par des
relations entre infinis pretendument inegaux. Son questionnement et ses
formulations sont tres proches de Gregoire de Rimini. Mais alors que ce
demier distinguait deux sens du tout et de la partie, de telle sorte qu' en un
sens une grandeur infinie puisse etre dite partie d'une autre, Albert s'en
tient a I'apprehension la plus immediate des relations valant dans Ie
domaine du fini, et, constatant comme on l'avait fait de longue date que
leur application a I'infini engendre des contradictions, il refuse toute
comparaison entre des infinis.

3. La divisibilite du continu

La question du continu fut au XIVe siecle un des lieux privilegies de

debats sur I'infini, en particulier dans les reuvres physiques,37 Quelques
considerations sur ce sujet ont deja ete introduites a I'occasion du
sophisme precedent. La question est abordee directement dans les
sophismes LIX, 'In infinitum continuum potest esse divisum', et LX,
'Continuum potest esse divisum in infinitum'. Les developpements sont
assez longs, et je n' en indique que quelques elements.

Pour demontrer la verite du premier sophisme en prenant l'infini au

sens syncategorematique, Albert met en reuvre de nombreux instruments
logiques:

l'idee d'un sens syncategorematique de I'infini, par opposition au

sens categorematique.

la distinction du sens compose et du sens divise, courante a

I' epoque mais surtout systematiquement utili see par Guillaume
Heytesbury dans ses Sophismata et entre autres dans son sophisme
'Infinita suntfinita'.

la distinction entre proposition possibilis et proposition de

possibile. 38

des criteres permettant de determiner la quantite de certaines

propositions, criteres qui, combines avec la distinction precedente,
occupent une bonne partie des pages consacrees a ce sophisme,39

la theorie de la supposition.

En ce qui conceme I'avant-demier point, la phrase 'in istas partes

continuum esse divisum est possihi/e' est une proposition indefinie dont Ie
sujet est la totalite de l'expression 'in istas partes continuum esse divisum'.
Un tel dictum suppose pour lui-meme mais aussi pour plusieurs

37Cf. John Murdoch, "Infinity and Continuity", in The Cambridge History of Later
Medieval Philosophy, ed. N. Kretzmann, A. Kenny et J.Pinborg, Cambridge:
Cambridge University Press 1982, pp. 564-91.
38Cf. Sophismata, f. 20va-vb (sign. e I, vb).
39Cf. ibid., f. 20va-21ra (sign. e I, vb - e II, rb).
300 JOEL BIARD

proposItIOnS similaires, ecrites ou parlees. Comme il suppose

disjonctivement pour plusieurs propositions en meme temps, la
proposition est indefinie. De meme, la proposition 'in omnes partes suas
continuum esse divisum est possibile' est indefinie et non pas universelle.
QueUe serait l'universelle correspondante? 'Omne "in omnes partes suas
continuum esse divisum" est possibile'.

Pour ce qui est de la theorie de la supposition, la SUpposItIOn

seulement confuse est introduite avec une reflex ion sur l'acte et la
puissance, en relation avec Ie verbe 'potest'. L' auteur veut montrer que' in
infinitum continuum potest esse divisum' ne renvoie pas a une puissance
precise mais a une infinite de puissances, de telle sorte que la phrase ne
signifie pas une possibilite deterrninee de divisions infinies. En effet, la
puissance a laquelle renvoie Ie verbe 'potest' dans cette proposition est a
prendre de maniere seulement confuse, en vertu du syncategoreme qui
precede et qui renvoie a un nombre. Et s'il est vrai, conformement au
principe aristotelicien, que 'omnis potentia potest reduci ad actum', ce
n'est pas vrai de toutes les puissances en meme temps, comme Ie montre
l' exemple de puissances opposees. Par consequent, il peut y avoir une
infinite de puissances dans la divisibilite du continu sans que l'actualisation
concerne cette totalite, donc une division actueUement infinie. II s'agit
bien, en vertu du mode de supposition, d'une infinite de divisions
possibles et non d'une division infinie.

Le second de ces sophismes consacres au continu, 'Continuum potest

esse divisum in infinitum', est declare faux par une argumentation fondee
elle aussi sur les modes de supposition. La puissance denotee par Ie verbe
'potest' doit etre tenue pour determinee puisqu'aucun syncategoreme
facteur de confusion ne precede ici 'potest', a la difference de l'enonce
precedent. La phrase signifie donc qu'il y aurait une puissance
(deterrninee) par laquelle Ie continu pourrait etre divise a ]'infini. Or c'est
faux, d'apres ce qui a ete precectemment dit de la puissance. Une fois cette
these po see, les differents arguments sont examines a I' aide de deux
principaux instruments logiques. En premier lieu, Albert fait encore appel a
la distinction entre supposition determinee et supposition simplement
confuse, en relation avec, d'une part, ce qui a ete etabli dans Ie sophisme
precedent concernant la puissance et l'acte, d'autre part la discussion
esquissee dans un sophisme anterieur a propos de l'induction ex parte
praedicati.40 En second lieu, la fonction du syncategoreme varie selon sa
place. C'est ici que joue pleinement la difference entre les deux acceptions
du terrne 'infini' et que prend tout son sens, par consequent, Ie traitement
de ces questions dans Ie cadre des' sophismes'.

En somme, a I' occasion de ce sophisme, nous voyons Albert de Saxe

utiliser de multiples instruments de la logica modernorum, selon une
demarche, deja courante, dont Jean Buridan fournissait un exemple dans
sa Physique avec la codification et la mise en reuvre systematique de la
distinction entre usage categorematique et usage syncategorematique du
terrne 'infini'. A vrai dire, l'essen tiel des pages consacrees a ces deux
sophismes consiste non seulement en precisions de nature logico-

40Cf. ibid., f. 21rb (sign. e II, vb).

 ALBERT DE SAXE ET LES SOPHISMES DE L' lNFlNI 301

linguistique sur les enonces - ce qui est lie a la nature meme du recueil -
mais encore en developpements proprement logiques concernant des
elements dont la mise en reuvre est ici requise. Pour ce qui est, en
revanche, des positions soutenues, on voit Albert s'en tenir ici au sens
syncategorematique et eviter tout ce qui pourrait aller dans Ie sens de la
position d'un etre en acte de l'infini.

4. L'infini et les quaJittfs

L'influence anglaise se manifeste clairement dans les derniers

sophismes consacres a l'infini. Deux sophismes concernent l'infinite
qualitative. Un autre concerne la force, selon un rapprochement qui avait
deja conduit certains auteurs d'Oxford a traiter de problemes du
mouvement a l'aide d'instruments tMoriques elabores pour mesurer les
variations qualitatives.

Le sophisme LXI, 'In infinitum Sortes erit albior Platone' est examine
selon Ie casus suivant: Socrate et Platon ont la meme blancheur, et Socrate
garde ce meme degre de blancheur pendant une heure ou deux, heure
durant laquelle la blancheur de Platon diminue continiiment jusqu'a
atteindre un seul degre de blancheur. Le sophisme LXII, 'Infinitam
albedinem habet Sortes', a pour casus: la blancheur de Socrate augmente
continiiment pendant un certain temps jusqu'a un certain temps (ou,
imaginairement,jusqu'a un instant) qui constitue une limite ext6rieure de
cet intervalle de temps et qui est Ie premier moment du non-etre de Socrate,
etant entendu que si celui-ci survivait a ce moment, il aurait une blancheur
finie.

Les moyens mis en reuvre pour resoudre ces sophismes sont

identiques. II s'agit de la distinction entre Ie sens categorematique et Ie sens
syncategorematique d'une part, de la difference entre supposition
determinee et supposition confuse d'autre part. A partir de la,les reponses
sont similaires. Chacun de ces sophismes est vrai au sens
syncat6gorematique et faux au sens categorematique.

Au sens syncat6gorematique, Ie premier sophisme signifie en effet que

du double, Socrate sera plus blanc que Platon, que du triple il sera plus
blanc, et ainsi de suite a l'infini - ce qui, dans l'hypothese de depart, est
vrai. Au sens categorematique, il signifierait que par un certain degre de
blancheur, Socrate serait infiniment plus blanc que Platon - ce qui n'est
pas vrai. De la phrase 'in infinitum Sortes erit albior Platone', on ne peut
donc pas inferer 'Sortes erit albior Platone in infinitum', car ce serait
passer d'une supposition seulement confuse a une supposition determinee.
Dans la premiere proposition, la supposition est seulement confuse parce
que Ie temps qui est consignifie par Ie verbe 'erit' est confus, en vertu du
syncategoreme qui precede et qui inc1ut une notion de nombre. Dans la
seconde, n'etant pas precede par Ie syncategoreme, il renvoie a un temps
determine, et un degre determine de blancheur est de ce fait signifie.

De la meme far;:on, dans Ie deuxieme de ces sophismes, de 'Infinitam

albedinem habebit Sortes' , on ne peut pas inferer 'Sortes habebit infinitam
albedinem', puisqu'ici encore, dans la premiere, Ie temps consignifie par Ie
verbe est rendu confus du fait du syncategoreme qui precede et qui inc1ut
302 JOEL BIARD

une notion de nombre, tandis que dans la seconde la supposition est

determinee.

Ce sophisme, en depit de sa proximite avec Ie precedent, introduit

implicitement une question suppIementaire, puisque Ie raisonnement
suppose que, quel que soit Ie moment considere dans l'intervalle durant
lequella blancheur de Socrate croit, ce moment n'est pas Ie dernier, si bien
qu' Albert se trouve confronte a un probleme de limite.41 La divisibilite a
l'infini du continu temporel est simplement ici invoquee pour une
paradoxale croissance a l'infini de la blancheur, alors meme que si Socrate
survivait, il n'aurait qu'un degre fini de blancheur.

Des questions similaires sont traditionnellement evoquees dans les

problemes de force et de resistance, qu'aborde Ie sophisme LXIII:
'Infinitum pondus Sortes potest portare'. De tels problemes sont aussi bien
traites par les Anglais que par Jean Buridan dans Ie cadre de sa philosophie
naturelle. A vrai dire, et en depit d'un assez long developpement, Albert de
Saxe n'approfondit pas tous les aspects de la question, renvoyant
explicitement a d'autres ouvrages. Je ne fais que resumer les conclusions
auxquelles il aboutit apres avoir longuement critique une autre position,
soutenant qu'il faudrait assigner un maximum que Socrate puisse porter, si
bien que la proposition sophismatique serait fausse, de quelque maniere
qu'on l'entende. Albert de Saxe, assez proche ici de Jean Buridan, pose
une serie de quatre conclusions concernant la puissance active, a laquelle il
convient d'assigner un minimum quod non, puis une serie de cinq
conclusions concernant la puissance passive, a laquelle it convient
d'assigner un maximum quod non.42

Apres de longs detours au cours desquels il prend ainsi position par

rapport aux debats sur Ie maximum et Ie minimum, Albert de Saxe repond
au sophisme en Ie jugeant vrai si l'on entend 'infini' de maniere
syncategorematique, et faux si on I' entend de maniere categorematique.
Mais auparavant, il renvoie a son commentaire sur Ie Traite du ciel,
estimant que de nombreuses difficultes qui surgissent a propos de tels
enonces ne concernent pas son propos dans Ie present ouvrage "puisque ce
sont des difficultes qui relevent de la philosophie naturelle [difficultates
naturales] et qui ne proviennent pas des syncategoremes."43

Ainsi se marquent Ies limites principielles de ce traitement de l'infini

dans Ie cadre des Sophismata. Neanmoins, on a vu que Ie besoin
d'eclaircir Ies enonces contenant Ie terme 'infini' etait au premier plan, et
dans la mesure ou les passages concernes du Traite du ciel mettent en
reuvre, comme I'avait fait Jean Buridan, Ia distinction (ici decisive) entre Ie
sens categorematique et Ie sens syncategorematique, sans doute Ies
precisions delivrees par ces sophismes conditionnent-elles ce qui est
pensable en physique ou en cosmologie.

41Cf. ibid., f. 2lvb (sign. e III, va).

42Cf. ibid., f. 22va-22vb (sign. e IIII, va - e V, ra).
43/bid., f. 22vb (sign. e V, ra).
 ALBERT DE SAXE ET LES SOPHlSMES DE L'INFINI 303

Le traitement de l'infini propose par Albert laisse de c6te certaines

questions renvoyees a la philosophie naturelle, mais il pretend cependant a
une certaine completude. II illustre une liaison intrinseque entre l'etude de
l'infini et l'analyse d'un terme, moyennant la distinction entre Ie sens
categorematique et Ie sens syncategorematique de l'infini. C'est une
demarche qui se repand au XIYC siecle, en meme temps que se multiplient
les champs d'analyse de l'infini. Elle est ici mise en evidence et accentuee
du fait que 1'0n examine les questions de l'infini dans Ie cadre d'un traite
sur les sophismes. Ainsi sont examinees les conditions logiques et
linguistiques de tout enonce sur l'infini.

Pour ce qui est du contenu, Albert de Saxe reste assez eclectique. On

per-roit des echos de la demarche burldanienne, de Guillaume Heytesbury
et peut-etre d'autres oxoniens, de Gregoire de Rimini... Albert ne semble
pas toujours percevoir ce qu'il y a de plus novateur dans les differentes
doctrines auxquelles il puise. II est souvent rapide et parfois confus. Mais
par cette volonte de prendre en compte l'etat des differentes reflexions sur
l'infini et meme de les exposer systematiquement, Albert de Saxe est un
bon temoin des reflexions sur l'infini qui ont cours a Paris au milieu du
XIye siecle.

Centre National de la Recherche Scientifique - Paris

 II sofisma di Paolo Veneto: Sortes in quantum homo est
animal
Alessandro D. Conti

II sofisma di cui qui ci occuperemo (il 41 0 dei 50 raccolti nei

Sophismata di Paolo Veneto) risulta parzialmente anomalo, in quanta non
consiste, propriamente, in una argomentazione fallace, rna in una
proposizione reduplicativa, 'Sortes in quantum homo est animal' ,I di cui
sembra possibile dimostrare tanto la verita che la falsita - e in cio
consisterebbe, per I' appunto, il suo carattere sofismatico. Infatti, una volta
che la proposizione in oggetto si risolva nelle sue esponenti, queste si
strutturano secondo Ie scansioni tipiche di un sillogismo ipotetico la cui
conclusione, cioe 'Sortes in quantum homo est animal', e garantita nella
sua verita dalla verita delle premesse: 'Sortes est homo', 'Sortes est
animal' e 'si aliquid est homo illud est animal'.2 Dall'altra parte,
assumendo la proposizione reduplicativa in questione come premessa in
quattro diverse argomentazioni, sembra si possa pervenire a conclusioni
palesemente inaccettabili, pur nell' apparente rispetto delle regole formali
del ragionamento.

La strategia operativa messa in atto da Paolo Veneto per risolvere

questa evidente impasse logica mette capo ad una sillogistica generate delle
proposizioni reduplicative di notevole interesse soprattutto per gli stretti
legami che manifesta con la dottrina della predicazione e1aborata dal
Nicoletti stesso,3 nella quale Ie proposizioni reduplicative sono assunte
come esempio di predicazione formale - un nuovo tipo di predicazione
introdotto appunto dal maestro agostiniano. Di tale abbozzo di una teoria
generale delle proposizioni reduplicative, dei suoi risvolti logici e dei suoi
presupposti ontologici mi occupero in questo paper.

II primo dei quattro argomenti di cui Paolo Veneto si serve per provare
la falsita della proposizione reduplicativa in esame si basa sui fatto che da
una normale proposizione affermativa e sempre possibile ricavare una
nuova proposizione nella quale un termine posto "piii in alto" del predicato

lSi definiscono reduplicative quelle proposizioni occultamente composte 0 esponibili iI

cui soggetto viene introdotto con la precisazione 'in quanto', 'in quanto tale', 'per se
stesso', ecc., per restringere iI senso del termine ad un suo valore formale 0 essenziale.
2Qui nei Sophismata, dunque, Paolo Veneto, assume che Ie exponentes di una
proposizione reduplicativa positiva siano tre, non diversamente da quanto fa nella
Logica parva (cap. de propositione reduplicativa). Nella Logica magllQ, invece, Ie riduce
a due soltanto, eliminando la seconda proposizione (quella cio~ in cui iI predicato della
proposizione reduplicativa viene attribuito aI suo soggetto - nel nostro caso 'Sortes est
anima/,), in quanto la giudica ridondante (cf. la Logica magna, ed. Venetiis 1499, f.
41rb). Alia stato attuale delle conoscenze non e ancora possibile avanzare delle ipotesi
suI perche di queste oscillazioni e differenze, e nuovi studi e ricerche sull'argomento
sono percio necessari.
3Sulla nuova teoria della predicazione elaborata da Paolo Veneto cf. Johannes Sharpe,
Quaestio super universalia, Alessandro D. Conti (a cura di), Firenze: Olschki 1990,
parte seconda: "Studio storico-critico", pp. 318-24.

304
 PAOLO VENETO: SORTES IN QUANTUM HOMO EST ANIMAL 305

nella scala di generalita infracategoriale si predica con verita del soggetto;

cosl da 'Socrate e uomo' possiamo inferire 'Socrate e sostanza'. Da
'Sortes in quantum homo est animal' si puo allora ricavare 'Sortes in
quantum substantia est animal'. Quest'ultima proposizione e pero falsa, in
quanto da essa consegue che se qualcosa e sostanza allora e animale - il
che certo non e. Per conseguenza la proposizione di partenza, 'Sortes in
quantum homo est animal', deve essere, a sua volta, falsa.

La seconda argomentazione muove dalla constatazione che il termine

'animal', che compare come predicato nella proposizione in esame, e
mobile, in quanto puo essere sostituito da un termine posto piil in alto nella
scala di generalita infracategoriale senza che it valore di verita
dell'enunciato cambi. Da 'Sortes in quantum homo est animal' possiamo,
infatti, inferire 'Sortes in quantum homo est corpus, et est substantia' ecc.
Se il termine animal e mobile, allora esso autorizza non solo a salire (per
cosl dire) nella scala di generalita infracategoriale, rna anche a scendere; da
·SOl·tes in quantum homo est animal' potremmo cosl ricavare 'Sortes in
quantum homo est iste homo vel iste homo', che e pero falsa, in quanto
tale risulterebbe una delle sue proposizioni esponenti, e cioe la
condizionale 'si aliquid est homo, illud est iste homo vel iste homo', la cui
protasi (antecedens) e una proposizione necessaria e la cui apodosi
(consequens) e una proposizione contingente.

Molto piil interessante risulta il terzo argomento, che solleva l'annoso

problema dello statuto logico del proprium. Poicbe i termini 'uomo' e
'capace di ridere' sono estensionalmente equivalenti (convertuntur), da
'Sortes in quantum homo est animal' dovremmo poter ricavare 'Sortes in
quantum homo est risibilis'. Ora, pero, questa conseguenza pare
inaccettabile (in quanto 'uomo' elogicamente anteriore rispetto a 'capace di
ridere', cioe non 10 presuppone, rna ne e presupposto) e percio e falsa la
proposizione condizionale nella quale la seconda reduplicativa (ovvero
'Sortes in quantum homo est risibilis') viene risolta, e cioe 'si aliquid est
homo illud est risibile'.

II quarto, ed ultimo, argomento propone un ragionamento per

analogia: se 'Sortes in quantum homo est animal' e vera, allora, per i
medesimi motivi, anche 'isosceles in quantum isosceles habet tres angulos
aequales duobus rectis' dovrebbe essere vera, in quanto i tre termini che
compaiono in essa stanno tra loro nella medesima relazione dei termini che
compaiono nella prima reduplicativa. Cosl pero non e, poicbe, in verita,
non in quanto isoscele, rna in quanto triangolo it triangolo isoscele ha i tre
angoli intemi eguali a due retti. Anche 'Sortes in quantum homo est
animal' deve percio considerarsi come una proposizione falsa.

Paolo Veneto ritiene, ovviamente, che l'enunciato in questione sia

vero e corretto; si impegna, per conseguenza, a "smontare" Ie
argomentazioni pro Jalsitate e risolvere cosl il sofisma. Per ottenere questo
risuItato egli premette alcune tesi che, insieme con Ie confutazioni dirette
dei quattro ragionamenti contra, formano una sorta di abbozzo di una teoria
generale delle proposizioni reduplicative.

La prima tesi (conclusio) afferma che, non diversamente dalle

proposizioni esclusive e dalle eccettive, anche Ie reduplicative, considerate
306 ALESSANDRO D. CONTI

in relazione alla loro quantita, si dividono in universali, particolari,

singolari ed indeterminate. Queste ultime sono queUe nelle quali la nota
reduplicationis precede l'intera proposizione, come nel caso di 'in quantum
Sortes est homo SO/'tes est animal'.

La seconda tesi afferma invece che se in un sillogismo una sola delle

due premesse e una proposizione reduplicativa, la conclusione non puo, a
sua volta, essere una proposizione reduplicativa, Da 'omne animal est
album' e 'omnis homo in quantum homo est animal' non segue, infatti,
'omnis homo in quantum homo est albus'. Soltanto nel caso in cui si tratti
di un sillogismo in Barbara la cui premessa maggiore sia una proposizione
reduplicativa, anche la conclusione sara una reduplicativa. Infatti, da
'omne animal in quantum animal est album' e 'omnis homo est animal'
segue 'omnis homo in quantum homo est albus'. Invece nel caso in cui
entrambe Ie premesse siano proposizioni reduplicative, la conclusione e
sempre tale. Da 'omnis homo in quantum homo est sensitivus' e 'omne
risibile in quantum risibile est homo' segue infatti 'omne risibile in
quantum risibile est sensitivum'.

La terza tesi conceme Ia differenza tra Ie proposizioni reduplicative

propriamente dette e Ie reduplicative specificative. Paolo Veneto, in verita,
non presenta tale distinzione come una divisione tra due tipi diversi di
proposizioni reduplicative, rna come i due distinti sensi secondo cui ogni
proposizione reduplicativa puo essere interpretata. Egli afferma, infatti, che
Ia nota reduplicationis assunta con valore reduplicativo propriamente detto
non e logicamente equivalente (non convertitur) a quella assunta con valore
specificativo. Cosl che, ad esempio, l'enunciato 'Sortes in quantum animal
est coloratus' e vero se la nota reduplicationis e interpretata come avente un
valore reduplicativo propriamente detto, rna falso se interpretata come
avente un valore specificativo, poiche in questo caso starebbe a significare
un legame necessario tra l'essere animale e l'essere colorato ("quia per
ipsam denotatur quod Sortes sub ratione animalis est coloratus"), che
invece non c'e.

Nella quarta conclusio, infine, Paolo Veneto nega che Ie proposizioni

reduplicative possano essere soggette a "conversione" - nega, cioe, che
l'enunciato ottenuto scambiando di posto il soggetto e il predicato di una
proposizione reduplicativa vera sia anch'esso un enunciato vero. Da
'Sortes in quantum homo est animal' non si puo ottenere 'ergo animal in
quantum homo est SO/'tes'.

Sulla base di queste premesse il Nicoletti e in grado di ribattere punto

per pun to Ie argomentazioni pro falsitate riportate in precedenza. Data la
loro stretta connessione, Ie risposte al primo e al secondo argomento
possono essere esaminate assieme. Paolo Veneto obietta, dunque, al suo
ideale oppositore che Ie norme relative all' ascensus e al descensus neUe
proposizioni categoriche standard e in quelle reduplicative sono differenti.
La prima e la seconda argomentazione contra, percio, rispettano solo
all'apparenza Ie regole formali del ragionamento, poiche non tengono
conto del diverso valore che tali regole assumono nel caso delle
proposizioni reduplicative. Quando si ha a che fare con enunciati di questo
tipo, infatti, il termine che compare nella reduplicatio e mobile non in
maniera assoluta, rna solo rispetto al descensus, mentre il termine che
 PAOLO VENETO: SORTES IN QUANTUM HOMO EST ANIMAL 307

funge da predicato e mobile solo rispetto all'ascensus. II termine che

compare nella nota reduplicationis pub, quindi, essere sostituito da un altro
che 10 segua nella scala di generalita infracategoriale senza che il valore di
verita della proposizione cambi, rna non da uno che 10 preceda; mentre per
il termine che funge da predicato accade proprio il contrario.

La risposta al terzo argomento consente a Paolo Veneto di operare

alcune precisazioni a proposito del problema dello statuto del proprio. II
Nicoletti ammette che la proposizione 'Sortes in quantum homo est
risibilis' sia vera e che tale risulti anche la condizionale nella Quale si
risolve, cioe 'si Sortes est homo Sortes est risibilis', in quanto ripugna
logicamente assumere che qualcosa appartenga alIa classe degli uomini
senza nel contempo assumere che appartenga alIa classe degli esseri che
sono capaci di ridere. Ma questo non e per lui incompatibile con la
conclamata anteriorita logica di 'uomo' rispetto a 'capace di ridere', in
quanto il fatto che si possa benissimo intendere il significato di 'uomo'
senza fare alcun riferimento al significato di 'capace di ridere' e non
viceversa, non e sufficiente a eliminare la perfetta equivalenza estensionale
tra i due termini - sulla Quale si fonda la verita della proposizione
condizionale 'si aliquid est homo, illud est risibile'.

La risposta al quarto argomento fa riferimento alIa distinzione tra

valore reduplicativo propriamente detto e valore specificativo della nota
reduplicationis. Paolo Veneto assume cosl che la proposizione 'isosceles in
quantum isosceles habet tres angulos aequales duobus rectis' e vera se 'in
quantum' viene interpretato come avente un senso reduplicativo
propriamente detto, rna falsa se 'in quantum' viene interpretato come
avente un senso specificativo, giacche l'avere la somma degli angoli intemi
eguale a due retti euna proprieta connessa necessariamente con I' essere un
triangolo e non con l'essere un triangolo isoscele.

Da un punto di vista puramente formale la teoria delle proposizioni

reduplicative elaborata da Paolo Veneto non si discosta poi molto da quelle
avanzate dai piu importanti, e autorevoli, logici del XIV secolo: Ockham,
Burley, 10 stesso Wyclif (che godeva di larga considerazione da parte del
Nicoletti).4 Sono gli stretti legami con la nuova dottrina della predicazione
e la diversa concezione della struttura del reale che Ie fa da sfondo che,
conferendo alla teoria delle proposizioni reduplicative del maestro
agostiniano un valore del tutto particolare, la differenziano dalle altre - alle
quali pure si ispira. II punto di partenza per una interpretazione corretta ed
adeguata della teoria sulle reduplicative va cercato, dunque, nella
concezione ontologica di Paolo Veneto, e in particolare nella sua teoria
della composizione metafisica dell' ente finito corporeo.

II logico italiano (il cui pensiero sull'argomento e debitore delle

elaborazioni di Scoto e Wyclif) sostiene che l'ente finito corporeo si
compone di sostanza individuale e di una congerie ordinata di forme
accidentali che ad essa ineriscono. La sostanza individuale, per parte sua,

4Cf. in proposito Massimo Mugnai. "La expositio reduplicativarum chez Walter

Burleigh et Paulus Venetus", in English Logic in Italy in the 14th and 15th Centuries,
ed. A. Maieru, Napoli: Bibliopolis 1982, pp. 305-20.
308 ALESSANDRO D. CONTI

risulta dall'unione di materia e forma particolari. Quest'ultima, a sua volta,

deriva dalla congiunzione di una forma specifica con un principio
individuante (haecceitas, 0 ratio suppositalis) che funge come da materia
rispetto alIa specie, in quanto la particolarizza, portandola dal piano
dell'essere indeterminato e possibile a quello dell'esistenza determinata ed
attuale nel singolare. 5 Secondo il Nicoletti, quindi, neIl'ente finito
corporeo vengono ad essere, in qualita di elementi componenti (che non
possono pero inficiarne l'unita numerica, poiche non in grado di esistere
autonomamente): la forma specific a, la ratio suppositalis, la materia e Ie
forme accidentali; rna senza che sia possibile stabilire una scala di priorita
metafisica nei vari ordini di composizione se non tra l'individuo sostanziale
(forma specifica + ratio suppositalis + materia) e Ie sue forme accidentali.

Le conseguenze piii rilevanti di questo approccio alIa questione della

composizione metafisica dell' ente finito corporeo concemono il problema
del rapporto tra universali e singolari, e la connessa teoria della
predicazione. Secondo Paolo Veneto, gli universali formali (0 in re)6
esistono in atto nei loro individui (supposita), coi quali si identificano
realmente (realiter) distinguendosene, pero, formalmente iformaliter)}
Detto altrimenti, egli ritiene che universali e singolari, pur coincidendo
nella loro realm fisico-empirica, che e quella del singolare, abbiano pero
principi formali costitutivi differenti (il che e sufficiente ad assicurare loro
una forma, debole, di distinzione). Gli universali in re, dunque, sono parti
costitutive dell'essenza degli individui, mentre questi sono partes
subiectivae di quelli, giacche nel loro insieme ne rappresentano il reale
sostrato d'esistenza. Cio vuol dire che tutti i generi, Ie specie e gli individui
di un determinato settore categoriale risultano essere realmente identici tra
loro. 8

L' aver ridotto in maniera cosl sensibile la distinzione tra universali e

singolari poneva il problema di determinare con precisione se ed
eventualmente come fosse possibile attribuire tutte Ie proprieta di un
individuo sostanziale aIle sue forme universali e viceversa.
Esemplificando: (i) se il singolare e l'universale sono un'unica e medesima
cosa nella loro realta fisico-empirica, non si dovra, allora, nel caso in cui
Socrate 0 un qualche altro individuo umano sia bianco, attribuire in
qualche modo la proprieta di essere bianco anche all'universale uomo
(homo in communi)? E (ii) piii in generale, non si dovra ammettere che
universale est singulare?

Gia Wyclif, prima di Paolo Veneto, si era posto queste domande,

dando ad entrambe una risposta in senso affermativo. II teologo inglese

5Cf. la Summa philosophiae naturalis, pars vi, c.5, ed. Venetiis 1503, ff. 95vb-96ra; e
la Lectura super librum Metaphysicorum, Iibro III, tr.l, c.l, MS Pavia, Bib!.
Universitaria 324, ff. 83vb-84ra.
6Come tutti gJi autori realisti del suo tempo Paolo Veneto accetta una tripartizione degli
universali in: allle rem 0 ideali, cioe Ie idee divine, esemplari e archetipi delle cose
singolari esistenti; in re 0 formali, cioe Ie nature comuni esistenti nelle cose singolari;
post rem 0 intenzionali, cioe i concetti mentali, corrispondenti agli universali esistenti
in reo Cf. la Summa philosophiae naturalis, pars vi, c.3, f.94rb-va.
7Cf. la Summa philosophiae naturalis, pars vi, c.2, f.94ra.
SCf. la Quaestio de universalibus, MS Fabriano, Bib!. Comunale 34, f. 25v.
 PAOLO VENETO: SORTES IN QUANTUM HOMO EST ANIMAL 309

aveva cosl elaborato una nuova teoria dei tipi di predicazione che
modificava la tradizionale dottrina aristotelica, basata sull' opposizione tra
predicazione essenziale e predicazione accidentale (0 inerenza propriamente
detta). Wyc1if distingueva tra tre principali tipi di predicazione, ciascuno
piu generale dei precedenti: formale, secundum essentiam, e secundum
habitudinem, che gli permettevano di coprire anche quei casi di inerenza
indiretta di una forma accidentale propria di un individuo negli universali
sostanziali che ne costituiscono l'essenza. 9

Paolo Veneto riprende in parte e rielabora criticamente tale teoria. Egli

distingue due principali relazioni predicative, mutuamente esc1udentisi:
l'identica e la formale (praedicatio identica e praedicatio formalis).l0 La
prima si ha quando, affinche la proposizione in esame risulti vera, e
sufficiente che la forma connotata (0 anche direttamente designata) dal
termine che funge da soggetto e quella connotata dal termine che funge da
predicato abbiano almeno uno dei sostrati d' esistenza propri in comune.
Cosl e, ad esempio, nel caso di 'homo est animal' 0 di 'homo in communi
est album'. La seconda, invece, si ha quando per la verita della
proposizione in oggetto e necessario che la forma connotata dal predicato
sia presente nei so strati d'esistenza propri della forma connotata (0
designata) dal soggetto in virtu di un principio formale a sua volta presente
in essi e dichiarato nella proposizione stessa. Questo si verifica ogni volta
che in una normale proposizione assertiva i cui estremi siano termini di
prima intenzione vengono introdotte alcune espressioni particolari, quali
'per se', 'formaliter', 'in quantum', in grado di modificare il senso
dell'enunciato, rendendo appunto necessario per la sua verita che tra
soggetto e predicato valga la particolare relazione intensionale teste
ricordata. 0 anche quando il predicato dell' enunciato in esame sia un
termine di seconda intenzione, quale' species', 0 'genus' - dal momento
che la predicazione formale e quella normalmente comportata dai termini di
seconda intenzione usati in funzione predicativa. Cosl e, ad esempio, nel
caso di enunciati quali 'homo formaliter est animal' 0 'animal est genus'.

Come si puo facilmente vedere, nel caso della predicazione identica si

fa appello esc1usivamente a requisiti di tipo estensionale per i1 controllo
della verita della proposizione, essendo siffatta predicazione basata sulla
capac ita che soggetto e predicato hanno di identificare estensionalmente la
medesima c1asse di oggetti. Nel caso della predicazione formale, si tratta,
invece, di requisiti di tipo intensionale, facendo siffatta predicazione
appello ad un rapporto di convenienza tra soggetto e predicato,
modalmente determinato, in sen so necessario, in funzione del significato
concettuale dei termini stessi - e percio, in ultima analisi, fondato sulla
composizione metafisica delle entita designate da soggetto e predicato.

A seguito di questa ridefinizione dei rapporti predicativi Paolo Veneto

articola diversamente da Wyc1if la sua risposta alle due domande iniziali.
Se, infatti, egli ammette che sia possibile attribuire in qualche modo una

9Cf. John Wyclif. Tractatus de universalibus, c.l. ed. I.J. Mueller.'Oxford: Clarendon
Press 1985, pp. 28-36.
IOCf. la Summa philosophiae naturalis. pars vi, c.2. f. 93va-b e la Quaestio de
universalibus, ff. 27v-29r.
310 ALESSANDRO D. CONTI

proprieta inerente ad un individuo agJi universali che sono costitutivi della

sua essenza (e la predicazione identica serve appunto per coprire anche
questi casi), si rifiuta pero di accettare come vere quelle proposizioni in cui
due tennini di seconda intenzione vengono predicati l'uno dell' altro, come
nell'enunciato 'universale est singulare' - giacche in questa caso si
tratterebbe di una predicazione fonnale, rna nessuna intenzione seconda, in
quanta quella particolare intenzione seconda, e un'altra intenzione seconda.
Le intenzioni seconde, difatti, non si dispongono secondo la linea diretta di
predicazione infracategoriale (linea praedicamentalis) come invece Ie
intenzioni prime; percio nessuna intenzione seconda rientra nell'essenza di
un 'altra come suo principio costitutivo: genus non e specie 0 genere di
species.

Dalla sua analisi della predicazione Paolo Veneto ricava alcune regole
relative alia sillogistica che tengono conto della mutata prospettiva
(intensionalistica vs estensionalistica) che la predicazione fonnale introduce
rispetto a quella identica. Una di esse e, per noi, di grande interesse, dal
momenta che e la trascrizione quasi fedele della seconda conclusio
presentata per la soluzione di questa quarantunesimo sofisma, rna con una
differenza di un certo peso che converra discutere. Eccone il testo: ll

"Quintum notabile <est> quod praemissae mixtae ex praedicatione

fonnali et identica non concludunt in aliqua figura conclusionem
formalem, quia non sequitur 'omnis humanitas est species; sed
Sortes est humanitas; ergo Sortes est species'; sed bene sequitur
'ergo aliquid quod est Sortes est species', quae est praedicatio
identic a propter limitationem tennini relativi, qui habet limitare ad
praedicationem identicam. Tamen bene sequitur 'omnis humana
natura est species; sed Sortes fonnaliter est humana <natura>; ergo
Sortes est species', sed negatur minor, quoniam Sortes per rationem
Sortis, quae est sorteitas, non est homo nec humanitas, sed per
rationem humanitatis, quae est universalis ratio et specifica. Ex isto
notabili sequitur primo quod praemissis exsistentibus identificis non
sequitur conclusio fonnalis, quia non sequitur 'omnis homo est
coloratus; sed omne risibile est homo; ergo omne risibile est
formaliter coloratum.' Secundo sequitur quod praemissis
exsistentibus fonnalibus non sequitur conclusio identica. Patet, quia
non sequitur 'omnis animalitas est fonnaliter genus; nulla humanitas
est fonnaliter genus; ergo nulla humanitas est animalitas,' sed bene
sequitur quod nulla humanitas est fonnaliter animalitas."

Per il maestro italiano, dunque, se in una soltanto delle premesse di un

sillogismo la relazione predicativa edi tipo fonnale, la relazione predicativa
propria della conclusione sara identica. Perche sia fonnale e necessario,
infatti, che entrambe Ie premesse contengano una relazione predicativa di
tipo fonnale.

A differenza di quanta avviene per Ie reduplicative, neppure i

sillogismi in Barbara la cui premessa maggiore contenga una relazione
predicativa di tipo fonnale possono sottrarsi a questa regola generale.

11Quaestio de universalibus, f.28v.

 PAOLO VENETO; SORTES IN QUANTUM HOMO EST ANIMAL 311

Cio significa che l'identificazione tra proposizione reduplicativa e

predicazione formale, che a prima vista potrebbe sembrare scontata, non e
completamente automatica. In realal, soltanto nelle reduplicative
specificative il rapporto predicativo e di tipo formale e, per conseguenza,
puramente intensionale, giaccbe riducibile, in sostanza, ad una relazione di
analiticita semantica; mentre nelle reduplicative propriamente dette la
relazione tra soggetto e predicato resta ancora di tipo estensionale. La
differenza tra Ie due interpretazioni possibili della nota reduplicationis (cui
Paolo Veneto fa riferimento nella terza conclusio) si configura, quindi,
come una differente determinazione del valore della copula in funzione dei
possibili rapporti semantici vigenti tra gli estremi dell'asserto enunciativo.
Se la nota reduplication is viene interpretata come avente un valore
reduplicativo in senso stretto, allora implicitamente si assume che per la
verita della proposizione in questione sia sufficiente che il predicato
identifichi una classe di oggetti che contenga come suo sottoinsieme quella
determinata dal soggetto. La copula puo, quindi, essere letta come il segno
di appartenenza ad una classe nel caso di un enunciato singolare, e come il
segno di inclusione fra classi negli altri casi. Le cose vanno diversamente
se si attribuisce alIa nota reduplicationis un valore specificativo, percbe
allora essa sta ad indicare che tra soggetto e predicato vale una particolare
relazione di tipo intensionale che presuppone un legame necessario ed
immediato tra la forma designata dal termine reduplicativo (ad esempio, nel
nostro caso, la forma dell'humanitas) e quella designata dal predicato (nel
nostro caso, la forma dell'animalitas), che deve essere quella da cui la
prima si origina direttamente per l'aggiungersi di una differenza specifica.

Questo differente valore della copula quando si interpreti la nota

reduplication is come avente valore propriamente reduplicativo 0
specificativo, implica che quanto Paolo Veneto dice a proposito
dell'ascensus e del descensus, nelle prime due risposte alle argomentazioni
pro Jalsitate, sia valido solo nel caso in cui 'Si abbia a che fare con
proposizioni reduplicative la cui nota reduplicationis sia vista come avente
un senso puramente reduplicativo, rna non quando la si legga come avente
un senso specificativo. In realta, tanto il termine reduplicativo che il
predicato di una proposizione reduplicativa specificativa sono immobili.
Ne potrebbe essere diversamente, dato il particolare rapporto ontologico
che lega tra loro Ie forme da essi designate. Una conferma di questo fatto e
data da quel che Paolo Veneto dice in risposta alla quarta argomentazione
proJalistate. Egli infatti ammette che la proposizione 'isosceles in quantum
isosceles habet tres angulos aequales duobus rectis' e vera se la nota
reduplicationis e considerata come avente un valore esclusivamente
reduplicativo, rna falsa in caso contrario - dal momento che la proprieta di
avere la somma degli angoli interni eguale a due retti e connessa
necessariamente con l'essere triangolo e non con l'essere un triangolo
isoscele. Ora, se Ie regole relative all' ascensus e al descensus fissate dal
Nicoletti fossero valide anche per Ie specificative, da 'isosceles in quantum
triangulus habet tres angulos aequales duobus rectis', che e vera,
dovremmo poter ricavare 'isosceles in quantum isosceles habet tres
angulos aequales duobus rectis', che e invece falsa. II che significa che
esse hanno valore solo per Ie reduplicative in senso stretto.
312 ALESSANDRO D. CONTI

Un caso a parte e rappresentato, all'interno di questa teoria, dallo

statuto del proprio. Si e visto, infatti, come, in linea di massima, nelle
reduplicative assunte con valore specificativo debba esserci un legame
necessario tra la forma designata dal predicato e quella designata dal
termine reduplicativo. Normalmente esso e dato dal fatto che la forma
designata dal predicato e il costituente principale della forma designata dal
termine reduplicativo. Nel caso del proprio, pero, questa legame manca,
dal momenta che la forma designata dal proprio non rientra tra i costituenti
della forma specifica cui esso inerisce. Malgrado cio proposizioni
reduplicative quali 'homo in quantum homo est risibilis' sono considerate
vere dal Nicoletti, anche qualora si interpreti la nota reduplicationis come
avente un val ore specificativo. Questo dipende dalla natura particolare del
proprio (i) la cui forma presuppone necessariamente quell a della specie cui
inerisce, e che (ii) identifica estensionalmente la medesima classe di
oggetti. Queste due condizioni, per conseguenza, vengono difatto (se non
di diritto) considerate da Paolo Veneto equivalenti alIa relazione di
analiticita semantica (specificamente interpretabile come omogeneita
intensionale 0 subintensionalita), tipica delle proposizioni reduplicative
specificative. SuI piano puramente ontologico questa peculiarita del proprio
si spiega tenendo conto che per il Nicoletti la forma da esso designata non
ha un suo principio individuante che la faccia venire ad essere in qualche
individuo, rna viene contratta, seppure secundarie, da quello stesso
principio individuante che contrae la specie, cosl che il proprio, a
differenza dell'accidente comune, inerisce in primo luogo alIa specie e solo
in secondo luogo (tramite questa) all'individuo: 12

"Respondeo quod genus contrahitur per differentiam, <et> ex his

fit species, sicut ex specie et ratione suppositali fit individuum et
singulare. Proprium vero non habet propriam rationem suppositalem,
sed contrahitur per eandem ration em suppositalem per quam
contrahitur species; ita quod sorteitas contrahit speciem humanam et
risibilitatem. Non tamen aeque primo, sed per prius speciem,
secundarie autem et per accidens contrahit risibilitatem. Et isto modo
possumus concedere quod eadem ratio suppositalis speciei etiam
contrahit genus et differentiam, non tamen per se, sed ratione speciei.
Ex quo sequitur quod genus et differentia et proprium non habent alia
individua quam individua specierum specialissimarum - quae quidem
individua <sunt proprie individua> specierum, communiter autem et
secundarie sunt individua generis, differentiae et proprii."

Un interessante elemento di riflessione sembra emergere abbastanza

chiaramente da questa breve analisi della teoria delle proposizioni
reduplicative qui abbozzata da Paolo Veneto: quello della stretta
connessione, nelle sue opere, tra visione del mondo e concezioni logico-
linguistiche. Non solo queste ultime appaiono fortemente condizionate da
quella anche nei loro risvolti pill tecnici, rna la conoscenza delle dottrine
pill specificamente metafisiche del Nicoletti risulta indispensabile per una
corretta lettura, definizione e valutazione delle sue teorie logiche. Da questa
simbiosi tra logica ed ontologia e daIIa dipendenza della prima dalla

12Quaestio de universalibus, ff. 40v-41r.

 PAOLO VENITO: SORTES IN QUANTUM HOMO EST ANIMAL 313

seconda deriva - ritengo - quella «eccedenza» tipica di ogni sistema logico

tanto antico che medievale rispetto a qualsiasi tentativo di trascrizione
puramente formale che miri a ricondurli entro un ben definito sistema
logico (e/o linguistico) modemo.

Seuola Nazionale di Studi Medievali, Roma

Appendix
Paulus Nicolettus Venetus: Sophismata

41 m sophisma: 'Sortes in quantum homo est animal'

ed. Venetiis 1493, fol. 46va_47rb

Quadragesimum primum sophisma: 'Sortes in quantum homo est

animal'.

Quod sophisma sit falsum arguitur sic: Sortes in quantum homo est
animal; ergo Sortes in quantum substantia est animal. Tenet consequentia
ab inferiori ad suum superius affirmative absque impedimento, quia si
aliquod est impedimentum, videtur quod nota reduplicationis impediat,
ratione immobilitatis quam inducit in terminum immediate sequentem, sed
contra: sub isto termino 'homo' contingit descendere, ergo non stat
immobiliter. Probatur antecedens, quia sequitur: Sortes in quantum homo
est animal; et isti sunt omnes homines; ergo Sortes in quantum iste homo
vel ille homo est animal, et sic de singulis. Sed illud consequens est
falsum, scilicet 'Sortes in quantum substantia est animal', cum ex isto
sequitur quod si aliquid est substantia illud est animal. Et per consequens
antecedens est falsum - quod est sophisma, scilicet' Sortes in quantum
homo est animal'.

Secundo quaero utrum ly 'animal' stet mobiliter vel immobiliter. Non

immobiliter, quia generaliter ab inferiori ad suum superius a parte
praedicati, dictione reduplicativa praecedente, est bonum argumentum.
Nam bene sequitur: Sortes in quantum homo est animal, ergo Sortes in
quantum homo est corpus, et est substantia, et est ens. Si mobiliter, ergo
contingit descendere. Sed qualitercumque descenderetur, con sequens est
falsum, quia quaelibet talis esset falsa 'Sortes in quantum homo est illud
vel (et ed,) illud animal', et etiam quaelibet talis 'Sortes in quantum homo
est iste homo vel iste homo', et (vel ed.) sic de singulis; sicut haec
condicionalis 'Si aliquid est homo illud est iste homo vel iste homo', et
(vel ed.) sic de singulis, quoniam huiusmodi condicionalis antecedens est
necessarium et consequens contingens.

Tertio: si Sortes in quantum homo est animal, ergo per idem Sortes in
quantum homo est risibilis. Patet consequentia, quia homo et risibile
314 ALESSANDRO D. CONTI

convertuntur, et non homo et animal; sed consequens est falsum, quia si

Sortes in quantum homo est risibilis, ergo si aliquid est homo illud est
risibile. Sed contra: omne prius stat esse sine suo posteriori; sed prius est
homo quam risibile, quia prius generaliter est subiectum sua passione
(possessione ed.); ergo stat hominem esse sine risibili. Et per consequens
oppositum istius consequentis stat cum antecedente; ergo illa consequentia
non valet 'aliquid est homo, ergo illud est risibile'. Quod enim omne prius
stet esse sine suo posteriori patet per Philosophum in Postpraedicamentis,
capitulo de priori, dicentem: 'Prius est illud a quo non convertitur
subsistendi consequentia'.

Quarto: si Sortes in quantum homo est animal, hoc est quia Sortes est
homo, et si aliquid est homo illud est animal. Sed contra: sequitur per idem
quod isosceles in quantum isosceles habet tres angulos aequales duobus
rectis, quia isosceles est isosceles, et si aliquid est isosceles illud habet tres
angulos aequales duobus rectis. Deinde, sicut se habet homo ad (et ed.)
sensibile, ita isosceles ad <habens> tres angulos <aequales> duo bus
rectis, ergo si est verum quod homo in quantum homo est sensibilis, sic
isosceles in quantum isosceles habet tres angulos aequales duobus rectis.
Sed oppositum istius dicit Aristoteles, I Posteriorum, scilicet quod
isosceles non in quantum isosceles, sed in quantum triangulus habet tres
angulos aequales duobus rectis.

Ad oppositum arguitur sic: Sortes est homo; et Sortes est animal; et si

aliquid est homo illud est animal; ergo Sortes in quantum homo est animal.
Tenet consequentia ab exponentibus ad expositum; antecedens est verum;
ergo et consequens - quod est sophisma.

Circa praedicta sit haec prima conc1usio: aliqua reduplicativa est

alicuius quantitatis et aliqua nullius est quantitatis. Prima pars patet de istis:
'quilibet homo in quantum homo est animal', 'aliquis homo in quantum
homo est animal', quarum prima est universalis et secunda particularis -
remoto vero signo reduplicativo est indefinita. Sophisma autem est
propositio singularis. Secunda pars patet de istis: 'in quantum Sortes est
homo Sortes est animal', 'in quantum Sortes est homo Sortes non est
asinus'. Non enim videtur cuius quantitatis sint, ex quo nota
reduplicationis praecederet totam propositionem. Sicut enim exc1usivarum
et exceptivarum aliquae sunt alicuius quantitatis et aliquae nullius, ut patet
in duobus praecedentibus sophismatibus, ita in proposito.

Ex ista conc1usione sequitur quod aliqua reduplicativa est reduplicative

probabilis et aliqua non. Primum patet de reduplicativa nullius quantitatis,
ac etiam de reduplicativa alicuius quantitatis, cuius primus terminus est
nota reduplicationis. Verbi gratia, haec 'in quantum Sortes est homo,
Sortes est animal', sic exponitur 'Sortes est homo, et si Sortes est homo,
Sortes est animal'; haec vero 'Sortes in quantum homo non est asinus', sic
exponitur 'Sortes est homo, et si Sortes est homo, Sortes non est asinus'.

Item, haec singularis 'tu in quantum homo es risibilis' sic exponitur

'tu es homo, et si aliquid est homo est risibile'; haec autem 'tu in quantum
homo non est asinus' sic exponitur 'tu es homo, et si aliquid est homo,
illud non est asinus' .
 SORTES IN QUANTUM HOMO EST ANIMAL: APPENDIX 315

Secunda vero pars correlarii patet de istis 'omnis homo in quantum

homo est risibilis', 'aliquis homo in quantum homo non est asinus',
quarum prima est exponibilis ratione signi universalis et secunda est
resolvenda ratione termini exponentis determinate.

Secunda conc1usio: altera praemissarum exsistente reduplicativa non

sequitur conclusio reduplicativa. Patet, quia non sequitur in prima figura
'omne animal est album; sed omnis homo in quantum homo est animal;
ergo omnis homo in quantum homo est albus'. Etiam in secunda figura
non sequitur 'omne animal est album; nunus homo in quantum homo est
albus; ergo nunus homo in quantum homo est animal'. Item, nec sequitur
in tertia figura: 'omne animal est album; omne animal in quantum animal
est sensitivum; ergo omne sensitivum in quantum sensitivum est album'.

Ubi tamem utraque praemissarum est reduplicativa sequeretur

conc1usio reduplicativa. Verbi gratia: 'omnis homo in quantum homo est
sensitivus; omne risibile in quantum risibile est homo; ergo omne risibile in
quantum risibile est sensitivum'. Immo, in prima figura, si maior est
reduplicativa et non minor, potest inferri conclusio reduplicativa. Nam
bene sequitur 'omne animal in quantum animal est album; omnis homo est
animal; ergo omnis homo in quantum homo est albus' - sed maior est
impossibilis. Sicut etiam bene sequitur 'omnis homo necessario currit; tu
es homo; ergo tu necessario curris', sed non sequitur 'omne corpus
movetur; hoc necessario est corpus; ergo hoc necessario movetur'.

Tertia conc1usio: nota reduplicationis reduplicative tenta non

convertitur cum seipsa specificative tenta. Patet, quia haec propositio
'Sortes in quantum animal est coloratus' est vera reduplicative tenta, quia
Sortes est animal, et si aliquid est animal, illud est coloratum; specificative
autem tenta est falsa, quia per ipsam denotatur quod Sortes sub ratione
animalis est coloratus. Quod est falsum, quia si Sortes sub ratione animalis
est coloratus, ergo in ratione seu definitione animalis inc1uditur color -
quod est falsum. Similiter haec est vera 'ens in quantum ens est subiectum
metaphysicae' tenendo ly 'in quantum' specificative, quia ens sub ratione
entis est subiectum metaphysicae; sed reduplicative est falsa, quoniam ex
ista sequitur quod omne ens est subiectum metaphysicae - quod est
falsum. Interdum vero talis propositio redditur vera, sive ly 'in quantum'
teneatur reduplicative sive specificative; ut 'homo in quantum homo est
risibilis', 'medicus in quantum medicus est sanativus' . Quando ergo ly 'in
quantum' tenetur specificative, tunc est determinatio subiecti, quando vero
tenetur reduplicative, tunc est determinatio praedicati. Et hoc intendebat
Aristoteles, I Priorum, dicens quod reduplicatio semper tenet se a parte
praedicati.

Quarta conc1usio: propositio reduplicativa sive sit alicuius quantitatis

sive nullius nullam omnino recipit conversionem. Patet, quia non sequitur
'Sortes in quantum homo est animal, ergo animal in quantum homo -est
Sortes', nec sequitur 'nunum animal in quantum homo est Sortes, igitur
nunus Sortes in quantum homo est animal'; etiam non sequitur 'in
quantum Sortes est homo est, ergo in quantum homo est Sortes est'.
Cuiuslibet istarum consequentiarum antecedens est verum et consequens
falsum, ut patet assignando suum contradictorium.
316 ALESSANDRO D. CONTI

Ex praedictis sequitur primo quod aliqua in quantum conveniunt

disconveniunt. Probatur: haec in quantum conveniunt disconveniunt,
demonstratis Sorte et Platone; et haec sunt aliqua; ergo etc. Maior
declaratur: nam haec conveniunt, cum sunt eiusdem speciei specialissimae;
et si haec conveniunt disconveniunt; ergo etc. Probatur minor: nam si haec
conveniunt haec sunt; et si haec sunt haec sunt multa; et si sunt multa
disconveniunt; ergo, a primo ad ultimum, haec si conveniunt
disconveniunt. Consimiliter conceditur quod aliqua in quantum sunt eadem
differunt.

Secundo sequitur quod in quantum Sortes est homo, Sortes est

animal, non tamen in quantum Sortes est homo, animal est Sortes. Prima
pars patet, quia Sortes est homo; et si Sortes est homo, Sortes est animal.
Secunda consimiliter, quia suum oppositum est falsum, scilicet 'in
quantum Sortes est homo, animal est Sortes', cum ex ista sequitur quod si
animal est homo, animal est Sortes, tamquam ab exposita ad alteram
suarum exponentium; modo haec est una condicionalis cuius antecedens
est necessarium et consequens contingens.

Tertio sequitur quod Sortes non in quantum homo est Sortes nec in
quantum homo non est Sortes. Primum patet, quia si Sortes in quantum
homo est Sortes, ergo si aliquid est homo illud est Sortes, tamquam ab
exposita ad alteram suarum exponentium - sed consequens est falsum.
Secundum etiam est notum, quia si Sortes in quantum homo non est
Sortes, ergo si aJiquid est homo illud non est Sortes - quod iterum est
falsum. Nec ista duo contradicunt 'Sortes in quantum homo est Sortes',
'Sortes in quantum homo non est Sortes', sicut nec ista 'tantum homo
currit', 'tantum homo non currit', quia negatio debet semper praecedere
notam reduplicationis sicut exclusionis.

Quarto sequitur quod Sortes in quantum homo non est asinus; non
tamen in quantum animal non est asinus. Primum est notum, quia Sortes
est homo, et si aliquid est homo, illud non est asinus. Secundum etiam
patet, quia suum contradictorium est falsum, scilicet 'Sortes in quantum
animal non est asinus', cum (tamen ed.) ex ista sequitur quod si aliquid est
animal, illud non est asinus. Sicut ergo ab inferiori ad suum superius cum
nota condicionis fallit consequentia, ita cum nota reduplicationis fallit
consequentia; igitur etc.

Ad primum nego illam consequentiam 'Sortes in quantum homo est

animal, ergo Sortes in quantum substantia est animal', quia arguitur ab
inferiori ad suum superius confuse tantum immobiliter ratione dictionis
reduplicativae. Et cum dicitur: 'sub isto termino contingit descendere, ergo
stat mobiliter tantum', negatur consequentia. Unde aliquando terminus
confundens confuse tantum immobiliter solum immobilitat ad ascensum,
aliquando solum quo ad descensum, et aliquando utroque modo. Primum
patet de ly 'contingenter', quod immobilitat in ascendendo, quia non
sequitur 'contingenter iste homo est homo, ergo contingenter homo est
homo', sed bene sequitur 'contingenter homo est homo; et isti sunt omnes
homines; ergo contingenter iste homo vel iste homo, et sic de singulis, est
homo'. Secundum vero patet de ly 'necessario', quia immobilitat quo ad
descensum et non quo ad ascensum, quia sequitur 'necessario homo est
 SORTES IN QUANTUM HOMO EST ANIMAL: APPENDIX 317

homo, igitur necessario animal est homo', sed non sequitur 'necessario
homo et homo; et isti sunt omnes homines; ergo necessario iste homo est
homo, vel iste, et sic de sin9ulis, est homo'. Tertium vero est manifestum
de illo verbo 'incipit', quia Isicut non sequitur 'tu incipis videre aliquem
hominem, ergo tu incipis videre aliquod animal', ita non sequitur 'tu
incipis videre aliquem hominem; et isti sunt omnes homines; ergo tu incipis
videre hunc vel hunc, et sic de singulis' .

De dictione ergo reduplicativa est dicendum quod suum casuale

confundit confuse tantum mobiliter non quo ad descensum, sed quo ad
ascensum, sicut etiam dictio exceptiva. Sicut non sequitur 'omne animal
praeter hominem currit, ergo omne animal praeter animal currit', sed bene
sequitur 'omne animal praeter hominem currit; et isti sunt omnes homines;
ergo omne animal praeter hunc et hunc hominem, et sic de singulis, currit'.

Ad secundum dicitur quod ly 'animal' stat mobiliter et immobiliter:

mobiliter quidem quo ad ascensum, immobiliter quo ad descensum. Ita
quod dissimiliter confundit immobiliter suum casuale et praedicatum suum,
quia (quidem ed.) casuale immobilitet quo ad ascensum et non quo ad
descensum, sed (si ed.) praedicatum immobilitat quo ad descensum et non
quo ad ascensum. Supra subiectum vero, quando est propositio alicuius
quantitatis, nullam omnino habet vim mobilitandi vel immobilitandi.

Ad tertium concedo quod Sortes in quantum homo est risibilis, et cum

dicitur istam condicionalem esse falsam: 'si Sortes est homo Sortes est
risibilis', negatur. Immo repugnat aliquid esse hominem et illud non esse
risibile.

Ad Aristotelem vero dicitur quod in capitulo illo posuit multos modos

prioritatis, inter quos unus est ille a quo non convertitur subsistendi
consequentia. Et sic unum est prius duobus, et genus specie. Non tamen
omne quod est altero prius stat esse sine suo posteriori, quia punctus est
prior linea, et linea est prior superficie, et tamen non stat punctum esse sine
linea, nec linea absque superficie; sicut etiam non stat instans esse sine
tempore nec gradum esse absque intensione. Concedo tamen quod stat
intelligere hominem non intelligendo risibile, sicut stat intelligere punctum
non intelligendo lineam.

Contra: si aliquid est homo illud est risibile, et non e contra, quia
Antichristus est risibilis et non est homo.

Dicendum quod ly 'risibile' sumitur uno modo pro termino ampliativo

correspondente illi verbo 'potens (potest? ed.) <ridere>', alio modo pro
isto termino convertibili cum isto termino 'homo' - et sic dicitur passio
eius. Primo modo non dicitur esse posterior iste terminus 'risibile' isto
termino 'homo', sed bene secundo modo - et sic sumo in proposito.
Sumendo ergo pro termino ampliativo istum terminum 'risibile' sicut non
sequitur 'hoc est risibile, ergo hoc est homo', sic non sequitur e contra,
quoniam possibile est aliquem hominem ita complexionatum esse quod
nullo modo posset ridere. Sed sumendo istum terminum 'risibile' ut est
passio istius termini 'homo', sic omnis homo est risibilis et e contra.
318 ALESSANDRO D. CONTI

Ad quartum concedo quod isosceles in quantum isosceles habet tres

angulos aequales duobus rectis.

Et cum inducitur Aristoteles in contrarium, I Posteriorum, dieo quod

ibi sumitur ly 'in quantum' specificative et (etiam ed.) reductive, modo
verum est quod isosceles sub ratione isoscelis non habet tres angulos
aequales duobus rectis, sed sub ratione trianguli, quia proprium primo
inest speciei et per speciem individuis, secundum Porphyrium in
Universalibus. Igitur etc.
 The Soul of the Antichrist Necessarily Will Be a Being: A
Modal Sophism in 16th Century Logic Texts
by Jeffrey S. Coombs

In logic texts of the early 16th century, logicians wondered whether

the proposition 'The soul of the Antichrist necessarily will be a being'
(anima antichristi necessario erit ens) is true or false. Their discussions
primarily revolve around the issue of scope, that is, whether the modal
term 'necessarily' is within the scope of the temporal term 'will be',
although along the way other interesting topics arise, such as multiple
conceptions of necessity, the temporal connotations of the modal terms,
and the truth conditions for divided modal propositions. We will focus on
the works of two logicians, both of whom taught at the University of Paris
within the first two decades of the 16th century: the Spaniard Jeronimo
Pardo, and the Scot Robert Caubraith.

The structure of these discussions betrays their sophismatic roots.

This proposition 'The soul of the Antichrist necessarily will be a being' is
presented first with arguments for its truth and then with arguments for its
falsity, just as in Sophismata tracts of the 14th and 15th centuries. In fact,
many arguments are taken (with citation) from the sophismata tracts of
William of Heytesbury and Paul of Venice.! Pardo and Caubraith,
however, do not discuss this sophisma in an independent Sophismata tract,
as is the case for Paul of Venice. They instead use this sophisma (and
others) as a counterexample to their general method for determining the
truth and falsity of modal propositions. Thus, Jeronimo Pardo discusses
this sophisma in chapter 8 of his Medulla dyalectices, a work devoted to
determining the truth-conditions of many different types of statements.
Chapter 8 of Pardo's text is devoted to presenting rules for determining the
truth and falsity of modal propositions. Robert Caubraith considers this
sophisma in the fourth chapter of his work, the Quadrupertitum in
oppositiones, conversiones, hypotheticas, et modales. 2 Again the
sophisma appears within a dicussion concerning rules for determining the
truth and falsity of the modal propositions.

Pardo's discussion of the sophisma 'The soul of the Antichrist

necessarily will be a being' (anima antichristi necessario erit ens) begins
with the usual presentation of arguments for and against its truth. Pardo
holds that 'The soul of the Antichrist necessarily will be a being' is true,
and in section I we will consider his defence of it. Pardo's defence relies

!See Paul of Venice, Sophismata magistri Pauli Veneti, Venetiis 1491, ff. 38rb-39vb.
I have seen the copy in Italian Books before 1601, microfilm series, Watertown,
Mass.: General Microfilm Company, roll 118.7.
2Jeronimo Pardo, Medulla dyalectices, Parisius 1505. I have used the copy of this text
reproduced in the Vatican Film Library, microfilm series, roll 37.11, cited hereafter as
'MD'. Robert Caubraith, Quadrupertitum in oppositiones, conversiones, hypotheticas,
et modales, Parrhisiis 1509. I have used a copy of this edition which can be found in
the British Library, cited hereafter as 'Quad.'

319
320 JEFFREY S. COOMBS

on the idea that the modal term 'necessarily' is within the scope of the
temporal term 'will be'. In section II, we will see Caubraith's answer to
Pardo. Caubraith will opt for the "common" view that the sophima is false
and he argues against Pardo's defence by claiming that the temporal
expression is within the scope of the modal term.

Pardo states that the argument showing the truth of 'The soul of the
Antichrist necessarily will be a being' runs as follows. Assume that the
soul of the Antichrist will be produced at some time. After it is produced, it
will be eternal and incorruptible. This soul necessarily exists because
'necessarily' is understood in terms of 'physical' necessity. An entity is
physically necessary if and only if it cannot be destroyed by any created
agent.3

The usual (common) opinion, Pardo claims, is that, on the contrary,

'The soul of the Antichrist necessarily will be a being' is false because it
entails that the soul of the Antichrist cannot possibly fail (non poterit non)
to exist. 4 This is false because it is quite possible that the soul of the
Antichrist not come into existence at all, since God might choose not to
produce the Antichrist, or the Antichrist's parents might never meet, etc.

After Pardo outlines attacks on the "common" opinion, he proposes

his own argument for the truth of this sophisma. His main argument is

The soul of the Antichrist necessarily will be [a being] after time t*.

Therefore, the soul of the Antichrist necessarily will be [a being].

Time t* is the moment at which the soul of the Antichrist comes into
existence, and since this soul is immortal, it cannot be destroyed after t*.5
Pardo proceeds to defend the truth of this argument's premise, and then the
argument's validity.

Pardo defends the truth of the premise 'The soul of the Antichrist
necessarily will be a being after t*' by following his general approach to
determining the truth value of divided modal propositions. Divided modal
propositions were modal propositions in which the mode appears in the

3M D. ff. cviii(rb-va). "Dubium [est]: an haec propositio sit concedenda 'anima

antichristi necessario erit ens' posito quod anima anti christi aliquando producetur. Et
cum erit producta, erit eterna et incorruptibilis, et diceretur necessario esse necessitate
phisica, ita quod virtute nullius agentis creati poterit desinere esse." ("The question is
whether this proposition should be conceded: 'The soul of the Antichrist necessarily
will be a being' assuming that the soul of the Antichrist were at some time produced.
And since it will have been produced, it will be eternal and incorruptible, and it would
be said to 'be necessarily' in the sense of physical necessity, such that by the power of
no created agent will it possibly be ended.")
4Jbid .. f. cviii(va). "Ad hoc est responsio communis quod ilia propositio est falsa quia si
anima antichristi necessario erit, non poterit non esse, quod est falsum."
5Jbid. "[I]psa [anima antichristi] necessario erit post a, ergo necessario erit. Antecedens
probatur. quia post a erit aeterna et incorruptibilis, ergo post a necessario erit."
 THE SOUL OF THE ANTICHRIST 321

middle of the proposition and the mode separates the terms of the
proposition.6 The primary rule for determining the truth value of divided
modal propositions was, according to Pardo (and many others in the early
16th century),

"Each true divided modal proposition is reducible, that is, its truth
is made manifest, through one or through many necessary or
contingent, possible or impossible non-modal (de inesse)
[propositions] .... If in fact the proposition contains the mode
'possible', it must be reduced to one or many possible non-modal
(de inesse) [propositions]. That is, its truth must be made manifest
through one or many possible non-modal (de inesse) [propositions].
Similarly if it contains the mode 'necessary', its necessity must be
made manifest through one or many necessary non-modal (de inesse)
propositions. "7

Thus, a modal proposition is changed to a non-modal (de inesse)

proposition, of which the evicted mode is predicated. The mode then
becomes a 'metalinguistic' predicate of the sentence. For example,

Jeff is possibly drinking sherry,

which is a divided modal proposition because 'possibly' appears in the

middle, becomes

'Jeff is drinking sherry' is possible.

The primary aim of this approach is the clarification of the scope of the
modal term. For although determining the non-modal counterpart of modal
propositions is easy when there are no logical operators, quantifiers, or
modal or temporal terms present, when these latter sorts of expressions do
appear in the proposition, additional rules had to be proposed to deal with
the ensuing confusion. For example, the proposition 'The soul of the
Antichrist necessarily will be a being after t*' requires this additional rule
for determining its truth value since it contains a future tensed verb:

6Caubraith, Quad.: f. cxxii(va). "In modali divisa modus mediat inter partes dicti et non
in modali composita. Et ilia est ratio potissima quare una vocatur composita et altera
divisa. Dicitur enim composita... quia partes dicti ex parte eiusdem extremum
componuntur unum extremum totale constituentes. Et divisa quia partes dicti
adinvicem per modum dividuntur ad diversa extrema se applicantes." ("In a divided
modal proposition the mode is in the middle between the parts of the dictum, but [this
does] not [occur] in composed modal propositions. And this is the strongest reason
why one is called composed and the other divided. For [the one] is called composed
because the parts of the dictum are composed into the same extreme [term], [and the
parts of the dictum] make up one total extreme [term]. And [the other is called] divided
because the parts of the dictum are divided from one another by the mode, placing
themsel ves into separate extreme terms.")
7MD, f. cvii(va). "Quaelibet modalis divisa vera est reducibilis, id est, manifestabilis
sua veritas, per unam vel per plures de inesse necessarias vel contingentes, possibiles
vel impossibiles .... Nam si sit de possibili, debet reduci ad unam vel plures de inesse
possibiles, id est, debet sua veritas manifestari per unam vel plures de inesse
possibiles. Similiter si sit de necessario. debet manifestari sua necessitas per unam vel
plures de inesse necessarias, et sic de aliis modis."
322 JEFFREY S. COOMBS

"divided modal propositions, when the modes are joined to verbs

of the past or future tense, first must be reduced to propositions of
the present tense while keeping the modes the same, and after that
they must be reduced to de inesse propositions."8

This statement of Pardo's view may lead one to conclude that we should
simply ignore the tense of verbs in divided modal propositions, which
would be a very strange result. However, Pardo accepts (in a modified
form) the common opinion (ascribed to Buridan in the margin of Pardo's
text) that

"propositions about the past or future are not true unless a

proposition about the present corresponding to them was or will be
true."9

Thus, the tense of the verbs in a proposition about the past are not ignored;
they are pushed into what we would now call the 'metalanguage,' just like
the modes. Thus, 'Socrates ran' is true only if 'Socrates runs' was true.
As Pardo himselftells us:

"a proposition about the past or the future is said to be true or

false in virtue of a proposition about the present. Thus a proposition
about the present is called by logicians 'absolutely first' (simpliciter
prima) because other propositions must be called true or false in
virtue of it."10

Pardo finally states that the proposition 'The soul of the antichrist
necessarily will be a being after t*' is true because

"every proposition about the future, which at some time will have
a true present tense proposition to which it truly reduces, is true. "11

In the particular case under consideration, 'The soul of the Antichrist

necessarily is a being after t*' is true because the following at some time in
the future will be true:

This soul necessarily is a being after t*,

8Ibid •• f. cviii(rb). "Modales divisae quando modi iunguntur verbis praeteriti aut futuri
temporis. primo habent reduci in propositi ones de praesenti talibus modis manentibus.
et postea iIIae habent reduci in propositiones de inesse."
9Jbid .. ff. Ixxiii(va-vb). "Propositiones de praeterito vel futuro non sunt verae nisi
propositio de praesenti eis respondens fuit vel erit vera."
tOlbid•• f. Ixxv(rb). "Propositio de praeterito vel futuro dicitur esse vera vel falsa ratione
propositionis de praesenti. unde propositio de praesenti vocatur a logicis simpliciter
prima quia ratione eius a1iae propositiones habent dici verae vel falsae." This view
probably has its source in Ockham. See W.Ockham, Summa Logica. Opera
Philosophica: vol. I.. ed. P. Boehner et al.. St. Bonaventure. N.Y.: Franciscan
Institute 1974. pp. 269-70.
I IJbid .. f. cviii(va). "Omnis propositio de futuro. quae semel habebit unam de praesenti
vera in qua vere reducitur. est vera."
 THE SOUL OF THE ANTICHRIST 323

speaking, of course, in tenns of physical necessity.12 The assumption here

seems to be that the statement 'This soul necessarily is a being after t*' is
only true after t*. In essence, then, Pardo is suggesting that the tense tenn
has the modal tenn within its scope. The final analysis of 'The soul of the
Antichrist necessarily will be a being after t*', using 'F' for 'it will be the
case' and 'L' for 'it is necessary that', is:

FL(the soul of the Antichrist is a being after some time t*).

Pardo next defends the validity of the implication from 'The soul of
the Antichrist necessarily will be [a being] after time t*' to 'The soul of the
Antichrist necessarily will be [a being]'. 13 Common opinion, Pardo tells
us, argues against this implication because according to it 'necessarily' and
'impossibly' are "universal modes." Universal modes are such that when
they are added to a verb connoting a time they distribute the verb for every
time connoted by it. 14 The modes 'possibly' and 'contingently' are
"particular modes" because they make the verbs to which they are attached
connote for every time usually connoted, except they do so
"disjunctively."15 In the case at hand, the proposition 'The soul of the
Antichrist necessarily will be' expresses that idea that for every time t later
than the present moment m, the soul of the Antichrist necessarily is at t.
The proposition 'The soul of the Antichrist necessarily will be a being after
t*' is a weaker claim, assuming, as Pardo does, that there are moments of
time between the present moment and t*.1 6

12/bid. "[S]ed ista 'anima antichristi erit [ens] post a' est reducibilis in propositionem de
praesenti veram. Probatur, quia ista semel erit vera 'haec anima necessario est ens post
a' et loquor de necessitate phisica." ("But this: 'The soul of the Antichrist will be [a
being] after a' can be reduced to a true [proposition] about the present. This is proven
because this at some time will be true: 'this soul necessarily is a being after a' and I
speak of physical necessity.") A problem arises in this text concerning how much
Pardo is mentioning and how much he is using. Pardo, of course, does not give any
quotation marks. So, the question is, is he mentioning only 'anima antichristi erit
[ens]' or 'anima anitichristi erit [ens] post a'? I think the latter reading is the only one
which will make sense of this passage. The second sentence is the deciding one since
the use of 'semel' and 'post a' together makes no sense.
l3lbid., f. cviii(vb). "Sed an sequatur anima antichristi necessaria erit [ens] post a, ergo
anima antichristi necessaria erit [ens] dubium videtur."
14/bid. "Unde quidam est modus dicendi communis qui solet teneri videlicet quod isti
quattuor modi sic se habent videlicet quod ly 'necesse' et ly 'impossibiJe' sunt modi
universales qui, quando adduntur verbo connotanti tempus, distribuunt ips urn pro
quolibet tempore importato per ipsum."
15lbid. "[S]ed ly 'possibile' et Iy 'contingens' sunt modi particulares et faciunt accipi
pro omni tempore importato disiuntive."
16lbid. " ... qui modus [dicendi], si teneatur, notum est quod ista est falsa 'anima
antichristi necessario erit' sumendo ly 'erit' ut importat tempus, quia ly 'erit'
distribuitur pro quolibet tempore futuro. Ideo non valet consequentia: anima antichristi
necessaria erit post a, ergo anima antichristi necessaria erit, cum arguitur ab inferiori ad
superius cum distributione superioris." (" ...if this manner [of speaking] is held, it is
obvious that this is false: 'The soul of the anti christ necessarily will be' understanding
'will be' insofar as it implies time, because 'will be' is distributed for every future
time. Therefore this argument is not valid: the soul of the Antichrist necessarily will
be after a, therefore the soul of the Antichrist necessarily will be since [in it] one
argues from an inferior to a superior when the superior is distributed.")
324 JEFFREY S. COOMBS

Pardo answers by stating that there are reasons for thinking that
'necessarily' does not distribute the time connoted by the verb. First, such
a view must concede that a term is distributed in two contradictory
propositions, such as

Socrates necessarily is running

and

Socrates possibly is not running.

Pardo seems to think that the view of his adversaries must interpret the first
as:

(\7' t)(Socrates necessarily is running at t)

since 'necessarily' distributes 'is'. His adversaries must think the second
says:

(\7't)(Socrates possibly is not running at t)

since 'not' distributes 'is' .17 So interpreted, of course, they are no longer
contradictory. The real contradictory of the first should be:

(3t)(Socrates possibly is not running at t).

A second problem is that the propositions

Socrates is not necessarily running

and

Socrates is possibly not running

are usually thought to be equivalent. However, 'is' in the first is not

distributed because two distributing terms ('not' and 'necessarily') appear
together. IS For similar reasons, the two distributing terms 'not every' do

17lbid. "Sed [iste modus dicendi] de Iy 'necessario' non videtur verum, quod probatur
quia tunc sequeretur quod in ambabus contradictoris terminus distribueretur ut in istis
duabus 'Socrates necessario est currens', 'Socrates possibiliter non est currens', Iy 'est'
in utraque distribuitur ut patet in prima per Iy 'necessario', in secunda per negationem."
("But [this manner of speaking] about 'necessarily' does not appear to be true, which is
proven because then it would follow that a term would be distributed in two
contradictories, just as in these two: 'Socrates necessarily is running', 'Socrates
possibly is not running', 'is' is distributed in both as is obvious in the first because of
'necessarily', in the second because of the negation.")
ISlbid. "Sequeretur etiam quod in equipollentibus secundum equipollentias quae dantur in
consequentiis modalium non esset consimilis acceptio terminorum. Nam iIlae duae
equipollent 'Socrates non necessario est currens' et 'Socrates possibiliter non est
currens' in quibus tamen Iy 'est' non eodem modo accipitur. Nam in prima non
distribuitur propter duo distributiva." ("It would also follow that in equivalent
[propositions] according to the equivalences which are given in the [section on] the
modal consequences, there would not be a similar meaning of the terms. Indeed, these
 THE SOUL OF THE ANTICHRIST 325

not distribute when together. So, Pardo believes that this view would
interpret the fIrSt as

(3t)(Socrates is not necessarily running at t)

and the second as

('v't)(Socrates is possibly not running at t).

The second is distributed for all times because of the distributing term
'not'. So interpreted, the two propositions are no longer equivalent.

Third, the view requires that two contradictory propositions would be

false at the same time. This is false:

The soul of the Antichrist necessarily will be a being

assuming that a lot of time (multum temporis) will pass before the
Antichrist is created.1 9 Remember, 'necessarily' distributes 'will be' for all
future times. This, too, is false:

The soul of the Antichrist possibly will not be a being

because it implies (given that 'not' distributes 'will be')

The soul of the Antichrist possibly will not be a being at t*,

where t* is a time after the soul of the Antichrist is created. This latter
proposition is false because given the physical necessity of the created soul
of the Antichrist, its contradictory 'The soul of the Antichrist necessarily
will be a being att*' is true. 20

Pardo finally points out in what sense he thinks 'necessary' is a

universal mode. It is called universal because the necessary implies the
possible, but not vice versa, just as universal propositions imply indefinite

two [propositions] are equivalent: 'Socrates is not necessarily running' and 'Socrates is
possibly not running', in which however 'is' does not have the same meaning. In fact,
in the first statement it is not distributed because there are two distributed terms.")
19lbid. "Ideo infertur inconveniens tale quod duae contradictoriae essent simul falsae.
Primo ista est falsa 'anima antichristi necessario erit ens', posito quod multum
temporis pertransibit antequam ipsa producetur." ("Therefore, an unacceptable result
follows such that two contradictories will be false at the same time. First, this is
false: 'The soul of the Antichrist necessarily will be a being', assuming that a lot of
time will pass before this [soull is produced.")
20lbid. "[I]sta etiam est falsa 'anima antichristi possibiliter non erit ens' cum una
descendens sit falsa, scilicet ista 'anima antichristi possibiliter in hoc instanti
(demonstrando aliquod instans quod erit post productionem istius animae) non erit ens',
quia sua contradictoria est vera, scilicet ista 'anima antichristi necessario erit in hoc
instanti'." ("This too is false 'The soul of the Antichrist possibly will not be a being'
since one of its descendants is false, namely this: 'The soul of the Antichrist possibly
at this instant (when referring to some instant which will be after the production of this
soul) will not be a being', because its contradictory is true, that is, this: 'The soul of
the Antichrist necessarily will be at this instant' .")
326 JEFFREY S. COOMBS

(that is, particular) propositions, but not vice versa. 'Necessary' is

therefore not 'universal' in virtue of distribution but in virtue of a rough
analogy between the implications among the modes and those among the
quantifiers. 21 So, 'The soul of the Antichrist necessarily will be a being'
is true because

"its truth is manifested through one [proposition] about the

present which at some time will be true, namely, through this 'The
soul of the Antichrist necessarily is a being. "'22

II

Robert Caubraith considers a slightly different proposition from

Pardo's, namely 'The soul of the Antichrist will be' instead of 'The soul of
the Antichrist will be a being'. Although these logicians were very
concerned with slight variations in expression, the difference between
these two propositions does not seem of importance to them.

Caubraith will ultimately claim that 'The soul of the Antichrist will be'
is false, but first he reveals the sophimatic roots of his discussion by
giving arguments for and against accepting it. Here, as for Pardo, the
main concern is that these arguments count against the usual rule for
understanding divided modal propositions containing the mode
'necessarily':

"Every affirmative divided modal proposition is either reduced to

one non-modal proposition or to many; to one with regard to singular
terms, to many in the case of common terms."23

The proposition 'The soul of the Antichrist necessarily will be'

violates this principle since, Caubraith argues, it is true, but the non-modal
proposition to which it is reduced is not necessary. The non-modal
proposition to which it is reduced is

This soul will be

2 IIbid. "Ideo dico quod non oportet quod ly 'necessario' distribuat copulam. Quod autem
dicitur modus universalis in ordine ad ly 'possibile', hoc est quia ad ly 'necesse'
sequitur ly 'possibile' sed non e contra, quemadmodum ad propositionem universalem
sequitur indiffinita sed non e contra, et ideo non vocatur modus universalis eo quod
habeat aliquam virtutem distribuendi." ("Therefore I say that it is not obligatory that
'necessarily' distribute the copula. That, however, it is called a universal mode in
relation to 'possible', is the case because 'possible' follows from 'necessary' but not
vice versa, just as an indefinite proposition follows from a universal but not vice versa,
and thus it is not called a universal mode from the fact that it has some capacity to
distrubute [the copula].")
22Ibid., ff. cviii(vb)-cix(ra). "Ideo concedo istam propositionem 'anima anti christi
necessario erit ens' loquendo de necessitate praedicta, cuius fundamentum totum est
hoc, quia sua veritas manifestatur per unam de praesenti quae semel erit vera, scilicet
per istam 'anima antichristi necessario est ens'." The necessitas praedicta is physical
necessity.
23Quad., f. cxxiiii(rb), "Omnis modalis affirmativa divisa reducenda est ad unam de
inesse vel ad plures, ad unam ex parte termini singularis, ex parte vero terminorum
communium...ad plures."
 THE SOUL OF THE ANTICHRIST 327

where 'this' refers to the soul of the Antichrist. This latter proposition
Caubraith claims is contingent.

Caubraith presents two arguments for the truth of 'The soul of the
Antichrist necessarily will be'. The first is the familiar argument borrowed
from Paul of Venice and abbreviated by Pardo:

"(Ia) The soul of the Antichrist in this time (referring to the time at
which it will be produced) necessarily will be. Therefore, (Ib) the
soul of the Antichrist at some time necessarily will be. Therefore,
(Ie) the soul of the Antichrist necessarily will be."

Premise Ia is true, Caubraith says, because at the time the soul of the
Antichrist exists it cannot be destroyed by any natural power since every
soul is etemal.24

The second argument supporting 'The soul of the Antichrist

necessarily will be' is:

God sees the future soul of the Antichrist.

Therefore, the soul of the Antichrist cannot fail to be about to be.

This argument holds because if one denies it, then it is possible that God
see the future soul of the Antichrist and that soul possibly fail to be about
to be. If we assume that this possibility is actually the case, then God
would see the future soul of the Antichrist even though that soul will not
exist. Thus, it would be possible that God be deceived.25

Both arguments fail, in Caubraith's opinion. In the first argument,

Caubraith concedes the validity of argument from Ib to Ie but denies the
premise Ib, 'The soul of the Antichrist at some time necessarily will be. '26
His reasons for accepting the validity of this argument are not very strong,
since he seems to agree with an attack on it which can be found in Paul of
Venice. This attack states that the argument is invalid because it argues
from a non-distributed term to one that is distributed.27

24Jbid. "[A]rguitur haec propositio est vera 'anima antichristi necessario erit', et tamen
sua de inesse non est necessaria, ergo regula est nulla. Minor patet, haec est sua de
inesse 'haec anima erit' quam constat esse contingentem, et maior probatur, anima
antichristi in aliquo tempore necessario erit, ergo anima antichristi necessario erit.
Consequentia videtur tenere, et antecedens probatur, anima antichristi in isto tempore
necessario erit, tempus in quo erit producta demonstrando, cum in ilIo tempore ilia
anima non poterit destroi per potentiam saltern naturalem. Omnis enim anima a parte
post est perpetua."
25Jbid. "Item, deus vidit ilIam animam futuram, ergo ilia anima non potest non fore.
Consequentia tenet quia dato opposito consequentis cum antecedente haberem quod deus
vidit iIIam animam futuram et quod ilia anima potest non esse, quo posito inesse
habebitur quod deus vidit illam animam futuram et quod ilia anima non erit, ex quo
clare sequitur deum decipi posse."
26Jbid., f. cxxxvi(va). "Aliter dico concessa consequentia, negando antecedens."
27Jbid., f. cxxxvi(va). "[R]espondeo negando iIIam propositionem, et consequenter
negatur maior argumenti principaiis, et ad eius primam probationem respondet Paulus
Venetus. negando consequentiam quoniam arguitur a non distributo ad distributum...ex
328 JEFFREY S. COOMBS

We earlier saw this point discussed by Pardo. Even though it may be

the case that the soul of the Antichrist will be necessary at some time in the
future, it does not follow that that soul will be necessary at all future times,
that is, for all times after the present moment. Caubraith makes the point
explicit by claiming that arguing that Ib implies Ie is like arguing:

The soul of the Antichrist at some time necessarily will be.

Therefore, the soul of the Antichrist will not at some time possibly
not be. 28

It seems that the point can be reconstructed in this way: the argument
moves from

(3t)(the soul of the Antichrist necessarily will be at t)

to

-(3t)(the soul of the Antichrist possibly will not be at t ),

which surely is a misapplication of the modal rule which states that

'necessarily p = not possibly not p.'

Without citing any reason, Caubraith decides to accept the validity of

both of these arguments. He instead attacks the truth of Ib since he thinks
it could be shown false by arguments similar to those he had already given
against 'The soul of the Antichrist necessarily will be'. Furthermore,
Caubraith rejects la, 'The soul of the Antichrist at this time (referring to a
time after it is produced) necessarily will be'.29 If someone responds to
this rejection that the soul of the Antichrist cannot be corrupted after it is
created, and is thus necessary, Caubraith answers that he rejects the idea
that it will necessarily be, but not because he denies that it could not have
been corrupted after it was created, but rather because it was possible that it
never had been produced in the first place.30

parte de ... tempore ... " ("I answer by denying this proposition, and consequently the
major [premise] of the principal argument is denied, and to its first proof Paul of
Venice answers by denying the implication because it argued from a temporally
undistributed [term] to a temporally distributed.")
281bid. " ...ac si argueretur: anima antichristi in aliquo tempore necessario erit, ergo
anima antichristi non in aliquo tempore possibiliter non eri!."
29lbid. "[S]ed quoniam illud antecedens est falsum, ut probaliter potest probari, sicut de
principali propositione probatum est, [alliter dico concessa consequentia negando
antecedens, et ulterius nego hanc 'anima antichristi in hoc tempore necessario erit,'
tempus in quod erit producta demonstrando."
30lbid. hEt quando arguis in ilia non poterit corrumpi et in ilia erit, ergo tunc necessaria
eriE, nego consequentiam, non enim illam [sc., 'in illo anima antichristi poterit
corrumpi'] concedo quia postquam illa anima est, possit corrumpi, sed [illam concedo]
quia potest nunquam produci." ("And when you argue at this [instant] it will not be
possible to destroy [the soul of the Antichrist], and at this [instant] it will be,
therefore, then it will necessarily be, I deny the implication, for I do not accept this
[statement: 'at this instant the soul of the Antichrist could by corrupted'] because after
this soul is, it could be destroyed, but because this may never be produced.")
 THE SOUL OF THE ANJ'ICHRIST 329

The response to the second argument, Caubraith maintains, is to deny

its validity. Caubraith holds that

"it is possible that both God sees the future soul of the Antichrist
and that soul can fail to be about to be."3!

Since 'God sees the future soul of the Antichrist' is true and 'The soul of
the Antichrist can fail to be about to be' is a necessary truth, then the
conjunction is possible.3 2

Caubraith correctly points out that the original argument made a

mistake. The original claims that the possibility of both God's seeing the
future soul of the Antichrist and that soul's capacity not to exist entails that
this be possible:

"God sees the future soul of the Antichrist and that soul will not
be."33

Caubraith points out, however, that all this possibility entails is that this be
possible:

"God sees the future soul of the Antichrist and that soul can fail to
be in the future."

This latter is possible since "God's foreknowledge does not destroy the
contingency offuture entities."34

Caubraith's answer is acceptable as long as God's knowledge of a

statement p does not entail that p is necessary. Caubraith explicitly denies
this by saying that

God knows that p will be

does imply that

P will be

but it does not imply

p necessarily will be.35

3!Ibid. "Ad secundam probationem eiusdem... [r]espondetur negando consequentiam, et

ulterius do oppositum consequentis cum antecedente, et nego illius copulative
impossibilitatem... "
32Ibid. ..... cum eius una pars sit vera et altera necessaria."
33Ibid. "Et quando arguis illa copu[ativa posita illesse, etc., nego praesuppositum. IlIa
enim copulativa secundum se totam non esse inesse ponibilis."
34lbid. "[S]ed praecise secundum secundam partem, cuius de inesse est possibilis stante
adhuc huius veri tate 'deus vidit illam animam futuram'. Praescientia enim dei non
tollit contingentiam rerum futurarum."
35Ibid., f. cxxxvi(vb). "Ex quo infertur primo quod licet haec consequentia sit bona deus
scit hoc esse futurum, ergo hoc erit tarn respectu dei quam creatura, ex ilia IlOIl sequitur
ergo hoc Ilecessario erit."
330 JEFFREY S. COOMBS

Caubraith supports this claim with no lesser authority than the New
Testament. For when Christ said to Peter, 'You will deny me thrice', he
knew Peter would sin but not that Peter would necessarily sin. For if Peter
had necessarily sinned, he would not really have sinned, because there can
be no sin when the action must take place necessarily.3 6 So too supposing
God knows that the Antichrist will be damned, the Antichrist still is
possibly not damned, although it would never come to be that he is not. It
is still within the power of the Antichrist not to be damned. 37

Caubraith attacks another possible defence of 'The soul of the

Antichrist necessarily will be' by distinguishing "logical" from "physical"
necessity as Pardo did.3 8 If the necessity in question is logical necessity,
then all three of the following propositions must be true:

The soul of the Antichrist necessarily was a being,

The soul of the Antichrist necessarily is a being,

and

The soul of the Antichrist necessarily will be a being,

if 'The soul of the Antichrist necessarily will be' is true.39

Implicit in Caubraith' s remarks here is that if something is logically

necessary, then it is necessary at all times. 4o Since this soul is not now
necessary nor was in the past, then 'The soul of the Antichrist necessarily
will be' is false when interpreting 'necessarily' in this case as a logical
necessity.

Some unnamed logicians (perhaps Pardo?), Caubraith claims, hold

that 'The soul of the Antichrist necessarily will be' is true given the
"physical" notion of necessity.41 Caubraith argues against this solution by
pointing out that 'The soul of the Antichrist necessarily will be' implies

The soul of the Antichrist not possibly will not be,

36Ibid. "Patet per simile: quando christus dixit petro 'ter me negabis', ipse sciebat
petrum peccaturum [esse], attamen propterea petrus non necessario peccavit quoniam
dato opposito, quod necessario peccasset, sequeretur quod non peccasset, quia null us
peccat in hoc quod de necessitate evenit."
37lbid. "Ex quo subsequitur quod licet deus sciat antichristum esse damnandum,
antichristus tamen potest non damnari, licet nunquam ita eveniet. Hoc tamen est in
potestate antichristi."
38Ibid., f. cxxxvi(va). "Ad argumentum respondetur illam maiorem distinguendo velly
'necessario' dicit necessitatem logicam vel praecise phisicam."
39lbid. "Si primum, constat iIlam propositionem esse falsam, ex qua utraque istarum
sequitur 'anima antichristi necessario est', 'anima antichristi necessario fuit'."
40Ibid. "[S]ic enim utendo de ly 'necessario', omnes istae copulae convertuntur
'necessario est', 'necessario fuit', 'necessario eri!'."
41Ibid. "Si secundum (respondet quidam) propter rationem factam iIlam propositionem
concedendo."
 THE SOUL OF THE ANTICHRIST 331

which must be false, he thinks, because its contradictory

The soul of the Antichrist possibly will not be

is true.42 Caubraith proves this latter point by arguing that

The soul of the Antichrist will not be

is possible because

"the generation of the of the Antichrist can be prevented and

therefore so too the production of the soul of the Antichrist."43

A further argument against 'The soul of the Antichrist necessarily will

be' is that if this soul, which does not now exist, necessarily will be, it
follows that everything which will come to be, necessarily will come to be.
This outcome, Caubraith says, conflicts with Aristotle's views. 44 For
these reasons, Caubraith decides, contra Pardo, to deny that 'The soul of
the Antichrist necessarily will be' .45

Caubraith seems to be aware of Pardo's discussion of this sophisma

although he never mentions Pardo by name. He presents Pardo's view in
this way:

"Another argument could be made to show the truth of 'The soul

[of the Antichrist] necessarily wiII be', and thus its non-modal
[proposition] at some time will be true, and since it is of the future, it
follows that the original is true. The assumption is clear: after the
production of the Antichrist, this is true 'this soul necessarily is'
when understanding 'necessarily' always in a qualified way, that is,
as it expresses natural necessity of the future and of the potential."46

Caubraith chooses to respond only "briefly" to such a view by

denying its presupposition that one first reduces such a proposition with
regard to its future tense copula rather than its modal term. 47 Caubraith

42lbid. "Contra quem arguitur: sequitur bene anima antichristi necessario erit, ergo ipsa
non possibiliter non erit. Falsitas consequentis patet per veritatem eius contradictoriae,
quae est haec 'anima antichristi possibiliter non erit'."
43lbid. "[QJuod autem illud antecedens sit possibile patet quia generatio antichristi
potest impediri et per consequens productio animae antichristi."
44lbid. "Item, si anima antichristi, quod nunc non est, necessario erit, sequitur quod
omne quod eveniet de necessitate eveniet, contra Aristotelem primo perihermenias."
Caubraith may be referring to De Int., chap. 9, l8b26-l9a7.
45 1bid. "Quare aliter respondeo negando illam propositionem, et consequenter negatur
maior argumenti principalis."
46lbid. "Posset fieri aliud argumentum ad probandum illam veram 'haec anima
necessario erit', et hoc sic sua de inesse aliquando erit vera, et cum sit de futuro,
sequitur illam esse verarn. [AJssumptum patet post productionem antichristi, haec est
vera 'haec anima necessario est' proportionabiliter semper capiendo ly 'necessario', puta
ut dicat necessitatem naturalem futuritionis et potentialitatis."
47lbid. "Ad primum istorum dicitur breviter negando praesuppositum, ilia enim non
prius debet ex parte copulae de futuro ad de inesse reduci quam ex parte modi." ("To the
first of these I briefly answer by denying the assumption, for this must not be reduced
332 JEFFREY S. COOMBS

seems to be making the interesting point that the modal term is not within
the scope of the temporal expression, but that the temporal expression is
within scope of the modal term.48 Thus, Caubraith would interpret this
sentence as:

LF(the soul of the Antichrist is),

where 'L' is 'it is necessary that' and 'F' is 'it will be the case that'.

In Pardo's defence, we can still point out that if this proposition is

truly ambiguous, then Pardo has made some progress by explaining one
sense in which it is true. However, Pardo would certainly not be justified
in thinking (as he seems to) that his interpretation is the only one possible.
In fact, these logicians now need to give arguments for thinking that one of
these interpretations should take precedence over the other.

Caubraith also apparently had not read Pardo very closely, since he
proposes problems with Pardo's approach for which Pardo had already
provided solutions. Caubraith claims that the copula in 'The soul of the
Antichrist necessarily will be' is modified by a universal term. 49 Caubraith
seems to be referring to the doctrine we have already seen that 'necessarily'
has a universal connotation, and accordingly the proposition 'The soul of
the Antichrist will be' would have to be true at all future times in order for
'The soul of the Antichrist necessarily will be' to be true. Pardo has
already answered this by denying that 'necessarily' has the universal
connotation with regard to time.

To summarize, we have seen that while Pardo has insisted that the
temporal term 'will be' should take precedence over the modal term
'necessarily' in 'The soul of the Antichrist necessarily will be a being',
Caubraith thinks that the temporal expression is within the scope of the
modal term. Neither, however, sufficiently proves that his interpretation is
the stronger - or only - interpretation.50

Our Lady of the Lake University, San Antonio, Texas

to the non-modal proposition with regard to the copula in the future tense before [it is
reduced] with regard to the mode.")
48Caubraith's opinion is not crystal clear, however, since it is not completely clear
whether he accepts the claim that the non-modal counterpart of 'The soul of the
Antichrist necessarily will be a being' is 'This soul will be a being', which would have
to be (but is not) necessary in order for 'The soul of the Antichrist necessarily will be a
being' to be true.
49 Ibid. "Secundo dicitur negando consequentiam; non sufficit hoc ad eius veritatem
postquam copula determinatione universaJi determinatur, sed requiritur quod semper sit
vel erit vera."
5°1 would like to thank the Philosophy Department of the University of Texas at Austin
for travel funds. I would also like to thank the sponsors of the symposium for
supplying room and board during the conference.
 'Debeo tibi equum': A Reconstruction of the Theoretical
Framework of Buridan's Treatment of the Sophisma
by Gyula Klima

1. Introduction

In this paper my main aim is to sketch a formal reconstruction of the

theoretical framework of Buridan's treatment of the famous sophisma. In
this way I hope to show not only that Buridan's treatment is an organic
part of his general semantics and philosophy of language and mind,l but
also that the theoretical framework of his treatment, namely his theory of
appellation, can be given such a rigorous formulation as to make it a
genuine rival to contemporary theories too.2 But in order to have a better
understanding of Buridan's theory, we have first to take a closer look at
what paved the way to it, namely Ockham' s theory of connotation.

2. Denomination, connotation and Ockham's innovation in the

theory of signification
As is well-known, Buridan's theory of appellation owes much to and
has many features in common with Ockham' s theory of connotation.
Ockham's theory, on .the other hand, may best be understood starting from

1In respectful opposition to the judgement of L.M. de Rijk, who wrote: "As a matter of
fact, Buridan himself seems to feel quite uneasy about his 'debeo tibi equum' case (see
Scott, pp. 141-2). It must be considered an intruder, indeed. The last case of appellation
('equus' in 'debeo tibi equum') is not a correct one, since it is just a case of
supposition. Buridan's extending the 'venientem' case to the 'equum' case (i.e. the
adjective noun cases to the substantive noun cases) seems to be rather abortive."
"Buridan's Doctrine of Connotation", in The Logic of John Buridan, ed. J. Pinborg,
Copenhagen: Museum Tusculanum 1976, p.l00. For a more sympathetic evaluation of
Buridan's doctrine see A. Maieru, "Significatio et connotatio chez Buridan", in the
same volume. Cf. also Joel Biard, "Le cheval de Buridan: Logique et philosophie du
langage dans I'analyse d'un verbe intentionnel", in Die Philosophie im 14. und 15.
Jahrhundert, ed. Olaf Pluta, Bochumer Studien zur Philosophie, Amsterdam: Griiner
1989.
2For a critical assessment of Buridan's solution by the lights of contemporary logic see
P.T. Geach, "A Mediaeval Discussion of Intentionality", in his Logic Matters, Oxford:
Blackwell 1972. My claim is that if we go further with our reconstructions and take
Buridan's own semantic ideas seriously, then his theory works well, indeed even better
than modern theories like those of Quine and Montague, taking refuge in "easier"
paraphrases, leaving the semantics of the original problematic sentences simply
unexplained. Indeed, Buridan himself considers and rejects as "superficial" such evasive
attempts. See his Sophismata, ed. T.K. Scott, Stuttgart-Bad Cannstatt: Frommann-
Holzboog 1977, p. 87. For more criticism of and references to the modern theories see
S.L. Read, '''I promise a penny that I do not promise': The RealistINominalist Debate
over Intensional Propositions in Fourteenth-Century British Logic and its
Contemporary Relevance", in The Rise of British Logic, ed. P. Osmund Lewry, O.P.,
Papers in Mediaeval Studies 7, Toronto: Pontifical Institute of Mediaeval Studies,
1985, pp. 335-59, and G. Klima, "Approaching Natural Language via Mediaeval
Logic", in Zeichen, Denken, Praxis, ed. J. Bernard and J. Kelemen, Budapest-Vienna,
1990.

333
334 GYULA KLIMA

considerations connected with the Aristotelian conception of paronymy,

i.e. denomination. Indeed, as is pointed out by Marilyn McCord Adams in
her recent book on Ockham, one of Ockham's main semantic innovations
was his reversal of the traditional priorities of the significates of
denominative terms} As she writes:

"Whether or not Ockham' s criteria of primary and secondary

signification are adequate, his predecessors and contemporaries
thought that Ockham had the priorities exactly reversed. For
example, in De Grammatico, Anselm says that the per se
signification of 'literate' is literacy, while 'literate' appellates, or
signifies per aliud, one who is literate. Since the name 'literate' is
imposed on Socrates through something else - viz., literacy -
Anselm suggests that literacy is the primary and the one who is
literate the secondary signification of 'literate' and not vice versa. 4
Burleigh takes a similar line in De Puritate Artis Logicae."5

Of course, this reversal of priorities, as Burleigh clearly sees in his

criticism of Ockham's views, is closely connected with Ockham's
reductionist ontological program. For by this reversal Ockham could claim
that what a general term primarily signifies is not some common nature that
can be found individualised in the singulars falling under this term, but the
singulars themselves. So what these terms signify are just their possible
supposita, i.e. the things they refer to in several propositions.
Accordingly, it comes as no surprise that it is precisely over this point that
Burleigh feels especially at odds with Ockham.

However, although I agree with Adams' characterisation of Ockham's

innovation as a reversal of priorities, I think that anyone reading Burleigh's
argumentation may get completely false impressions as to exactly what was
exchanged by Ockham for what. For from Burleigh's words it is quite
clear that what he takes to be the primary significate of a general term is a
universal, whatever the ontological status of that universal is, whether it is
something outside the soul, or is a concept existing in the sou1.6 So in
view of Burleigh's line of argument, Ockham's "reversal of priorities"
should amount to his claiming that the secondary significates of
denominative terms are universals, which he definitely identifies with

3That this tradition was far from being uniform can be seen from the extensive
discussion of the earlier· positions concerning denominatives by Sten Ebbesen,
"Concrete Accidental Terms: Late Thirteenth-Century Debates About Problems
Relating to Such Terms as 'Album"', in Meaning and Inference in Mediaeval
Philosophy, ed. N. Kretzmann, Dordrecht: Kluwer Academic Publishers, 1989. See
also the related papers by R. Andrews and R. Huelsen in the same volume.
4Cf. sect.3. of D.P. Henry, The Logic of Saint Anselm, Oxford: Clarendon Press 1967.
(Adams' note.)
5M. McCord Adams: William Ockham, Notre Dame: University of Notre Dame Press
1987. vol. I. p. 325. The reference to Burleigh is Walter Burleigh. De Puritate Artis
Logicae Tractatus Longior with a revised edition of the Tractatus Brevior, ed. P.
Boehner, St. Bonaventure, N.Y.: Franciscan Institute 1955, c.3. pp. 7-10.
6"Sed sive iIIud commune sit res extra animam sive sit conceptus in anima non curo
quantum ad praesens; sed tantum sufficit, quod iIIud quod hoc nomen primo significat
est species." Burley, ibid. p. 8.
 'DEBEO TlB/ EQUUM' : BUR/DAN'S TREATMENT 335

concepts or intentions of the souI.7 Consequently, he would have to claim

that a denominative term, say, 'white' primarily signifies its supposita, the
things falling under it, that is, white things, and secondarily signifies the
concept of whiteness existing in the soul. However it is definitely not this
concept that Ockham takes to be the secondary significate of 'white' as
opposed to its primary significate, namely the white thing.

For he says that "a denominative term in the strictest sense is one to
which there corresponds an abstract term importing an accident that exists
formally inhering in something else,"8 namely in a thing having this
accident as its form, like whiteness in a white thing. But of course in the
case of a white thing it is not anybody's concept of whiteness that exists as
aform of the white thing, for this concept, being a quality of the mind
either distinct from or identical with the mental act itself by which one
thinks of a white thing, is an accident formally inhering in one's mind, and
not in the white thing that is thought of.9 What inheres in the white thing is
this thing's whiteness, by which the thing is denominated 'white'.
Consequently, it is this whiteness which, in Ockham's view, is the
secondary significate of the term 'white'. But, to be sure, this whiteness is
not a universal. For the whiteness of this white thing is as individual as the
thing itself, being numerically distinct from other things as well as from the
whitenesses of other white things.IO

As a matter of fact, if we take a closer look at the theory of

signification of the antiquiores defended by Burleigh against Ockham, we
find explicit reference to such individualised forms as what are signified by
general terms in the individuals. So it seems that also in the framework of
the older theory we have to make a distinction between the form signified
by a general term immediately, in an absolute sense, the "forma in
communi" on the one hand, and the forms signified by the same term
ultimately, in the individuals falling under it, on the other.!1

So such a term signifies these individualised properties in respect of

the individuals in which they inhere, i.e. it signifies such a property for
this individual, and another for that one, etc. But so, signification being a

7See e.g. chs. 12, 14-15 of Ockham, Summa Logicae, Opera Philosophica I, St.
Bonaventure N.Y.: Franciscan Institute 1974, (hereafter SL).
80ckham, Expositio in Librum Praedicamentorum Aristotelis., Opera Philosophica /l,
St. Bonaventure, N.Y.: Franciscan Institute 1978 (hereafter OP-/l) p.147.
9Cf. Adams, op.cit., p. 74 n. 10.
lOSee e.g. Ockham, Expositio in Librum Porphyrii de Praedicabilibus, OP-/l, pp. 24-5.
11 Such an individualised form is what e.g. S1. Thomas Aquinas refers to as "forma in
supposito singulari existens per quod individuatur", ST l.q.13.a.9. This distinction is
made also very clearly by St. Thomas in his De Ente et Essentia, c.4. For a
reconstruction of this distinction in St. Thomas see my '''Socrates est species': Logic,
Metaphysics and Psychology in St. Thomas Aquinas' Treatment of a Paralogism", in
my Ars Artium: Essays in Philosophical Semantics, Mediaeval and Modern, Budapest:
Hungarian Academy of Sciences 1989, also forthcoming in Acts of the 8th European
Symposium of Mediaeval Logic and Semantics,'ed. K. Jacobi. Cf. also William of
Sherwood, introductiolles in Logicam, ed. M. Grabmann, Sitzungsberichte der
Bayerischen Akademie der Wissenscha[ten, 1937, 10, p.78; Peter of Spain: Tractatus,
ed. L.M. de Rijk, Assen: Van Gorcum 1972, pp. 83-8. See also Ebbesen, op. cit., pp.
133-4.
336 GYULAKLlMA

semantic relation, we may represent the signification of a general term as a

semantic function assigning inherent properties to individuals. A significate
of a one-place general term, therefore, can be denoted as the value of the
signification function of this general term for an individual, like this: 12
SGT(P)(u)=v

(Read: a significate of P in respect of u is v, where u and v range over

elements of the universe of discourse W, plus 0, a zero-entity falling
outside the universe of discourse. The case v=O, of course, represents the
situation when P signifies nothing in respect of u.13)

The usefulness of this move will tum out immediately if we consider

how easy it renders the explanation of the nature of Ockham' s innovative
"reversal of priorities" between the signijicata of denominatives.

For with this formulation at hand we can say that both on Ockham's
and on the older view there is something that such a term primarily
signifies, whatever it is, that takes the place of v, and there is something
that is signified secondarily by the same term, whatever it is, that takes the
place of u in the above equation. Ockham's "reversal", then, simply
consists in his claiming that, e.g. in the case of 'white', what takes the
place of v is the white thing, not the thing's whiteness, while what takes
the place of u is, conversely, the thing's whiteness, not the white thing. So
we can bring out the contrast between Ockham's and the older view by the
following instances of the above equation:
Older view: SGT(' album ')(Socrates)=albedo Socratis
Ockham's view: SGT(, album ')(albedo Socratis)=Socrates

But Ockham's innovations, of course, did not only concern

denominative terms,a subclass of the class of connotative terms, but
connotative terms in general, i.e. terms which have both primary and
secondary significata, indeed, in Ockham's view all terms belonging in
categories other than substance and quality.14

120f course, by describing a semantic relation occurring in a mediaeval theory as a

semantic function, we do not ascribe anachronistically the possession of the concept of
a semantic function to the mediaevals, any more than one describing someone else's
argumentation as a case of ignoratio elenchi thereby attributes to him possession of the
concept of ignoratio elenchi.
13Notice here that SGT is not a two-place function, with a term in its first, and a thing
in its second argument-place, but a one-place function, which for a one-place term in
its argument-place yields another one-place function, which for a thing in its argument-
place yields an individualised property (the property signified by this term in this
thing). In this way, applying the same trick, we can give a uniform treatment of the
signification of general terms of any arity, and so, generaJIy speaking, we can denote a
significate of an n-place general term like this: SGT(T")(Ul) ... (Un). For further
technical details and the role of 0 see n. 22 below.
14See e.g. SL, pp. 37-8.
 'DEBEO TIBI EQUUM': BURIDAN'S TREATMENT 337

For example, the relative term 'father' signifies those men who have
sons in relation to their sons: 15
SGT('father')(son)=father

In contrast to this, on the older view, heavily criticised by Ockham,16 the

term 'father' was taken to signify a relation, 'fatherhood', holding between
two persons, the son and his father:
SGT('father')(son)(father)=fatherhood of father

Indeed, this formulation immediately shows why the question whether the
same person's having several sons multiplies his fatherhoods so naturally
arises in this context'!?

On the other hand, it also shows why Ockham did not have to worry
about such questions: for him what such a term signifies are just the things
which it can stand for in propositions, namely its possible supposita,
signified in relation to other things, the term's connotata. This is why he
could define personal supposition in terms of signification, thereby
incurring the charges raised by Burleigh of challenging the opinion of the
antiquiores .IS

Clinging much less to older views, however, Buridan willingly

adopted this conception ofOckham's in his own semantic theory.

3. Buridan's theory of appellation

Indeed, in line with T.K. Scott's characterisation of Buridan's relation

to Ockham, according to which "if Ockham initiated a new way of doing
philosophy, Buridan is already a man of the new way;'19 Buridan defines
the personal supposita of a term as its significata as a matter of course. 20
He finds no difficulty either in defining appellative terms as those that
connote other things beyond the ones which they are apt to supposit for:

"... omnis terminus connotans aliud ab eo pro quo supponit

dicimus quod est appellativus et appellat illud quod connotat per

15Cf. e.g. Guillelmi de Ockham, Quodlibeta Septem, ed. J.C.Wey, Opera Theologica
IX., St. Bonaventure N.Y.: Franciscan Institute 1980. V.25, VI.20-5.
16Cf. e.g. SL, chs. 50-4.
17A theologically more intriguing question of this kind was whether Christ's being the
son of both the Holy Mother and the Heavenly Father multiplies his filiations. See
e.g. St. Thomas Aquinas, ST III, q.35.a.5.
ISCf. SL, pp. 195-6.
19T.K. Scott, John Buridan, Sophisms on Meaning and Truth, New York: Appleton
Century Crofts 1966. n.1. p.13, quoted at greater length also by J. Pinborg in
discussing in more detail this relationship in his "John Buridan, The Summulae,
Tractatus I De Introductionibus", in idem, Mediaeval Semantics, London: Variorum
Reprints 1984.
20"Et vocatur suppositio personalis quando subiectum vel praedicatum propositionis
supponit pro suis ultimatis significatis vel pro suo ultimato significato." Buridan,
"Tractatus de Suppositionibus", ed. M.E. Reina, Rivista Critica di Storia della
Filosofia 12, 1957, pp. 175-208 and 323-52, p.201.
338 GYULA KLIMA

modum adiacentis ei pro quo supponit, ut 'album' appellat albedinem

tanquam adiacentem ei pro quo supponit."21

Buridan explains the notion of "adjacence" used here in the following

manner in his Sophismata:

" ... terminus appellat illud quod appellat per modum adiacentis
aliquo modo vel per modum non adiacentis ad illud pro quo supponit
vel est natus supponere. Dico per modum adiacentis, si appellat illud
positive, et dico per modum non adiacentis, si appellat illud
privative, ut 'album' appellat albedinem positive, tanquam
inhaerentem ei quod est album ... Sed 'caecum' supponens pro oculo
appellat visum privative per modum non adiacentis illo oculo."22

This adjacence or non-adjacence of the appellata of a term has a

prominent role in determining the supposition of appellative terms and,
consequently, in determining the truth-conditions of categorical sentences
formed with them. For an appellative term supposits only for those of its
significata in a proposition to which its appellata are adjacent (or non-
adjacent, as the meaning of the term requires) relative to the time connoted
by the copula of the proposition.23 But since according to Buridan an
indefinite affirmative categorical is true only if its terms supposit for the
same thing or same things, the sentence: 'A man is white' is true only if
both 'man' and 'white' supposit in it for the same thing or things.24

21Ibid., p. 343. Although as far as I know he never says so explicitly, it seems that for
Buridan being appellative or non-appellative are properties of terms independent of
context, but appellation is a property of terms only in propositions. On the other hand,
we find explicit statement of this in the work of his pupil, Marsilius of Inghen. See
Marsilius of Inghen, Treatises on the Properties of Terms, ed. E.P. Bos, Dordrecht:
Reidel 1983, pp. 128-36. So we may say that while supposition is reference in a
proposition to a term's signifcata, appellation is (oblique) reference in a proposition to
an appellative term's connotata.
22Buridan, Sophismata, ed. Scott, p.61. This means that if an appellative term connotes
something positively, say as 'videns' connotes sight, then it signifies what it is apt to
supposit for in relation to its connotatum only if the connotatum belongs to its
significatum, while an appellative term connoting something privatively, as 'caecum'
connotes sight, conversely, signifies what it is apt to supposit for only if its
connotatum does not belong to its significatum. Otherwise these terms signify nothing
in relation to these connotata. Making use of our previously introduced notation we
may reconstruct this difference between positive and negative adjacence in the following
way: SGT('videns')(u)E W, if UE W, and SGT(,videns')(u)=O, if u=O, while
SGT('caecum')(u)=O if UE W, and SGT('caecum')(u)E W, if u=O. Accordingly, a
suppositum of the term 'videns' in a proposition the copula of which connotes some
time t is definable as: SUP('videns')(t) = SGT('videns') (SUP('visus')(t», where,
'visus' being an absolute term, SUP('visus')(t) = SGT('visus'), if SGT('visus') is
actual at t, otherwise SUP('visus')(t)=O. (Here, of course, SGT('visus') is an element
of a subset of the universe of discourse determined by the natural signification of the
concept, or ratio signified immediately by this term, namely, the set of individual
sights that there were, are, will be or can be.) That is to say, the term 'videns' will
supposit for someone in this proposition only if the term 'visus' may successfully
refer to his sight, i.e. if he actually has sight (at t), while the term 'caecus', in
accordance with its negative connotation, would refer to this person only if he did not
have sight.
23Cf. Tractatus de Suppositionibus, pp. 184-5 and pp. 347-51.
24See Sophismata, pA3, Duodecima conclusio.
 'DEBEO TIBI EQUUM',' BURIDAN'S TREATMENT 339

However, a thing can be a suppositum of the term 'white' in this

proposition only if it actually possesses whiteness at the time connoted by
the copula, namely the present time of the formation of the proposition. On
the other hand, in the sentence 'A man will be white' the same term
supposits for only those things that will possess whiteness at some future
time relative to the present time of the formation of the proposition.

As Buridan notes, there is a substantial difference between the

appellation of the subject and of the predicate terms of such a proposition.
For although the predicate term appellates what it connotes strictly for the
time connoted by the verb, or even determined more by some explicit
temporal adverb, the subject term always appellates "under disjunction to
the present":

"Nam praedicatum appellat rem suam pro tempore verbi solum,

quantumcumque tempus verbi fuerit restrictum. Et si pro isto
tempore non correspondet modus adiacentiae rei appellatae ei pro quo
terminus est natus supponere, non supponit pro illo, quamvis bene
correspondeat modus adiacentiae pro tempore praesenti. Verbi gratia
si dico 'Socrates heri fuit albus', iste terminus 'albus' in ista
propositione non supponit pro Sorte nisi heri albedo ei adiacebat,
quamvis etiam modo ei adiaceat et quamvis ante ei adiaceret. Et sic
propositio esset falsa. Sed subiectum appellat rem suam indifferenter
modo disiunctivo pro tempore praesenti et pro tempore verbi, sicut
etiam est de suppositione. Dnde haec est vera 'Album fuit heri
nigrum', si ei quod heri fuit nigrum nunc adiaceat albedo, licet heri
non adiacebat sibi. Propter quod exponitur 'Album fuit heri nigrum',
idest 'Quod est album vel fuit heri album fuit heri nigrum' ."25

As Buridan adds later, this difference between the mode of appellation of

the subject and predicate does not mean that only one of them, namely the
predicate would have appellation in such a proposition. On the contrary,
both terms have appellation, provided they are both appellative, but in a
different manner: the subject under disjunction to the present, while the
predicate only for the time of the verb.

So appellative terms always appellate the things in a proposition that

they connote as being determinant of the things they refer to, and
consequently whether they refer to this thing or that thing in the given
proposition depends on whether what they appellate as the determination of
the thing in question does (or does not, as the meaning of the term
requires) actually belong (in the appropriate manner) to the thing relative to
the time connoted by the verb of the proposition.

The case is different, however, with absolute terms, which do not

connote anything as a determination of the things they signify, but simply
supposit for those things which they signify, like substance-terms or
abstract terms from the category of quality.26 Since the type of appellation
discussed above is a kind of secondary reference in a proposition to what a

25Sophismata, pp. 62-3.

26See Tractatus de Suppositionibus, p. 184.
340 GYULAKLIMA

term connotes also outside a proposition, this kind of appellation can

belong only to appellative terms, which do have connotation, or secondary
signification.27

On the other hand, also absolute terms can have appellation, i.e.
oblique reference to what is not supposited for in a proposition.
Obviously, this is the kind of reference they have as oblique parts of
complex terms. 28 But they also can have a peculiar type of appellation in
the special context created by intentional verbs like the one in our problem-
sentence, the kind of appellation that Buridan and his followers called
appellatio rationis.

4. A reconstruction of Buridan's treatment and his theory of

appellatio ration is
Nevertheless, despite its specific character, Buridan's appel/atio
rationis is justly called appellation, insofar as it is a sort of oblique
reference to something in a proposition which is not supposited for in that
proposition. However, this is not something connoted by a connotative
term, but something signified immediately by every vocal term, namely the
concept to which this vocal term is subordinated, due to which it is a
meaningful term at al1. 29 For according to Buridan a term, or in general,
any meaningful vocal expression of some language has meaning only
insofar as it is associated to some cognitive act of a human mind, a
conceptus, intentio, or ratio, whether it is simple or complex, by which we
conceive things in some way, and it is only through this association, or
subordination as Ockham and Buridan call it, that vocal terms can signify
the things we conceive by the concepts to which they are subordinated.

Now Buridan's treatment of our sophisma rests on his claim that

intentional verbs and other terms deriving from them which signify some
mental act force the terms following them to appellate their rationes
according to which they were imposed to signify the things conceived by
these rationes.

27This is the type of appellation that Marsilius de Inghen calls appellatio formalis
sigllificati, which he sharply distinguishes from the other type of appellation, the
appellatio ratiollis. From this doctrinal point of view it is quite evident that in the
otherwise excellent edition (see n. 21 above) of Marsilius' text all occurrences of the
phrase 'rationem suam' on pp. 150-2. should read 'rem suam', the standard traditional
expression for what he calls 'sigllifieatumformale'.
28"Terminus obliquus substantivus appellat iIIud pro quo rectus suus supponeret per
modum adjacentis ei pro quo rectus regens ips urn supponit." Tractatus de
Suppositiollibus, p. 347.
290n the other hand, we may say that the ollly difference between a case of appellatio
ration is and the appellatio an oblique term has in a complex term is that what the
oblique term appellates is what would be its personal suppositum, while what a term
having appel/atio rationis appellates is what would be its material (or simple, according
to earlier terminology) suppositum, if it itself were a whole term of a proposition. (Cf.
previous note.) Hence Buridan's comparison of the latter case (without, however,
subsumillg it!) to material supposition: "isti accusativi quodammodo videntur
participare suppositionem materialem, quia appellant conceptus suos, Hcet non
supponant pro eis." Traetatus de Suppositiollibus, p. 335. Cf. n.l.
 'DEBEO TlBI EQUUM'.' BURlDAN'S TREATMENT 341

He also gives a short explanation for this peculiarity of these verbs in

contradistinction to other, extensional verbs: namely that the terms
constructed with these verbs have to appellate their concepts because "we
think of things by these concepts, but it is not by these concepts that fire
heats water or a stone hits the ground. "30 So since a mental act concerns its
object only through the concept by which the thing is conceived (since
unless it were conceived by some concept it could not be an object of a
mental act at all), it is no wonder that a verb signifying such a mental act
makes the term signifying the object of this act appellate the concept by
which the term signifies the object; and it is equally no wonder that other
verbs which do not signify such mental acts do not force such an
appellation)!

So Buridan's theory can already be said to have scored a point over

several others in that it is able to give a plausible philosophical explanation
for the peculiar behaviour of these verbs. However, the further question is
whether Buridan's theory also works as a logical theory, in that it is able to
justify the intuitively acceptable, and invalidate the unacceptable logical
relations of our problem-sentence to others, and/or to provide compelling
arguments to accept possible unintuitive results.

But in order to know this first we have to see how this kind of
appellation affects the truth-conditions of sentences in which it occurs.
Buridan writes about this as follows:

"Nam quia eandem rem possum cognoscere secundum multas

divers as rationes, et isti rei secundum diversas rationes diversa
nomina imponere ad significandum earn, ideo talia verba faciunt
terminos cum quibus construuntur appellare rationes secundum quas
imposita sunt nomina ad significandum, et non solum res cognitas ad
extra, sicut faciunt alia verba. Aliter tamen a parte ante et a parte post.
Nam a parte post, illi termini appellant determinate suas rationes
proprias. Sed a parte ante appellant eas indifferenter sub disiunctione
ad alias rationes quibus res significatae possunt significari et intelligi.
Propter quod iste propositio non est vera 'Cognosco venientem',
proprie loquendo, nisi cognoscam earn secundum earn rationem,
secundum quam dicitur veniens, licet cognoscerem bene secundum
alias rationes. Et sic non sequitur, cognosco Sortem, et sit veniens;
ergo cognosco venientem. Quia licet cognoscam illum secundum

30Tractatus de Suppositionibus, p. 335.

31Cf. "oo.sciendum est quod ista verba 'inteIligo' cognosco', 'scio' et huiusmodi de
qui bus post dicemus, et participia et nomina inde descendentia, ut 'intelligens',
'cognoscens', 'intellectio', 'cognitio', etc., faciunt in terminis cum quibus construuntur
quosdam modos speciales appellationum. 00' Postea notandum est quod nomina
imponimus ad significandum mediantibus rationibus quibus res intelligimus. Ideo
etiam istud verbum 'significo' tales facit appellationes, sicut 'inteIligo' vel 'cognosco',
et ita etiam hoc verbum 'apparel', et haec verba 'scio', 'opinor', 'puto', 'credo', etc.
Praeterea etiam quia appetitus nostri fiunt in nobis mediante cognitione, ideo sequitur
quod similes appellationes faciunt nobis ista verba, scilicet 'volo', 'appeto', et
'desidero'. Et adhuc etiam quia sub aliquibus conceptibus facimus nostras promissiones
et obligationes, sequitur quod etiam isla verba 'debeo', 'promitto', 'obligo', etc. et alii
termini ex ipsis descendentes faciant huiusmodi appellationes." Sophismata, pp. 73-4.
342 GYULA KLIMA

illam rationem secundum quam dicitur Sortes, non tamen cognosco

illum secundum illam rationem secundum quam dicitur veniens. Sed
a parte subiecti, bene sequitur, Sortem cognosco, et Sortes est
veniens; ergo venientem cognosco .... Non ergo sequitur, venientem
cognosco, ergo cognosco venientem. Immo est possibile quod
ignoro venientem. Sed bene sequitur, cognosco venientem, ergo
venientem cognosco; sicut in aliis appellationibus non sequebatur,
album erit, ergo erit album, sed bene econverso, erit album, ergo
album erit."32

Since according to Buridan the "canonical form" of such sentences as

'Cognosco venientem' or 'Debeo tibi equum' is obtained by analysing their
verbs into copula and participle, and the result is true if the terms so
obtained supposit for the same, the question of the truth-conditions of such
sentences reduces to the question whether the terms 'cognoscens
venientem' or 'debens tibi equum' supposit for me in the sentences 'Ego
sum cognoscens venientem' or 'Ego sum debens tibi equum'. But in these
terms the intentional participle makes the oblique term following it appellate
its own ratio, whether this term is appellative or not, and so these complex
terms can supposit for me only if their participles signify me in relation to
this ratio and, of course, their other connotata required by their
signification.

So if we return to the notation previously introduced for representing

the signijicata of connotative terms we can spell this out quite simply in the
following manner:

SUP('cognoscens venientem')(t) =

= SGT(' cognoscens ')(RAT(' veniens') )(SUP(' veniens ')(t) )(t)

SUP('debens tibi equum')(t) =

= SGT(,debens ')(SUP('te'»(RAT('equus '»(SUP('equus ')(t'»(t)

where, taking into account the ampliative force of 'debens', t:::; t' (but
henceforth I disregard this complication), RAT is a function assigning to
terms the concepts they are subordinated to and S UP is a function
assigning to terms their supposita. 33

32Sophismata, pp. 73-4.

33Cf. n. 22. To be sure, in view of equivocations there is no such simple functional
connection between words and their rationes, i.e. concepts to which they are
subordinated. But by adding a further argument-place for several impositions,
associating the same words with possibly different rationes like this: RATCD(i), where
i ranges over different acts of imposition, we may approximate better the real situation.
Furthermore, since these rationes are individualised qualities of individual minds, we
might even add another argument-place reserved for individual minds like this:
RAT(DCi)(m), where m ranges over human minds. Nevertheless, as from our present
point of view both equivocations and individual conceptual differences may be
neglected, to simplify the formulation I omit these argument-places. But we should
keep well in mind that these have theoretical significance in Buridan' s discussions of
sigllificatio ad placitum, as e.g. in chs. 1 and 6 of his Sophismata, or qq. 2 and 3 of his
Questiones Longae super Librum Perihermeneias, ed. R. van der Lecq, Nijmegen:
 'DEBEO TIBI EQUUM': BURIDAN'S TREATMENT 343

The question of the truth of 'Cognosco venientem' and 'Debeo tibi

equum', as I said, boils down to the question whether I am one of the
above defined supposita of the terms concerned. But as we can see, these
supposita are functionally dependent on the rationes appellatae of the
names following these verbs. So changing the name may change the ratio
appel/ata, and this may change the supposita of the terms in question,
which changes the truth-conditions of the corresponding propositions. So
this is why the passage from 'Cognosco Socratem' to 'Cognosco
venientem' does not preserve truth-value despite the identity of reference of
'Socrates' and 'veniens', and so the inference from the first to the second
is not valid.

On the other hand, since when the noun in question precedes the verb,
or the participle deriving from the verb, it does not determinately appellate
its own ratio, but appellates it under disjunction to other rationes, we can
determine the supposita of the corresponding complex terms in question in
the following manner:

SUP('venientem cognoscens')(t) =

= SGT('cognoscens ')(RAT(X))(SUP('veniens ')(t))(t)

SUP('equum tibi debens')(t) =

= SGT(,debens ')(SUP('te '))(RAT(X))(SUP('equus ')(t))(t)

where X is some expression that can take the place of 'venientem' or

'equum' in the above complex terms.34

Ingenium 1983. On the other hand, even this simplified formulation well represents the
case when two expressions are synonymous: in this case RAT has the same value for
both expressions. This immediately shows why synonyms are interchangeable salva
veritate even in such contexts.
34Note that here we quantify over expressions, not the rationes themselves. So in this
context we need not worry much about the identity-criteria of rationes. The theory only
says that if two expressions have the same ratio, then their interchange in intentional
contexts a parte post does not affect the truth of the sentence in which they occur, but
the same interchange involving non-synonymous expressions does, while the case is
different a parte ante, where the interchange of any terms referring to the same thing or
things leaves the truth-value of the sentence unaltered. (Cf. note above.) As P.T. Geach
in his paper mentioned above correctly points out, for quantifying over rationes we
would need some criterion of identity for them. However, it is simply not true that "we
have not even one example <from Buridan> of the same ratio differently expressed,
from which we might divine a criterion of identity." (op. cit. pp. 132-3) Of course, all
synonyms are different expressions of the same ratio, and, in particular, all definitiones
exprimentes quid nominis, are different expressions of the ratio expressed by the nouns
they define, as Buridan carefully explains in several places. (E.g. Sophismata, pp. 24-
35.) So from these explanations we may "divine" the following criterion of identity:
different expressions express, or are subordinated to, the same ratio if and only if the
sum total of their significata and connotata, in the case of connotative terms, is the
same, whether these are things outside the mind or concepts existing in the mind (e.g.
in the case of complex expressions containing syncategoremata). So the identity-
conditions of rationes are definable in terms of the sets of significata and connotata of
the terms subordinated to them. But sets are quite well-behaved entities with respect to
344 GYULA KLIMA

This formulation immediately justifies Buridan's claim that with this

type of appellation 'Cognosco venientem' implies 'Venientem cognosco'
and 'Debeo tibi equum' implies 'Equum tibi debeo', but not conversely.
For if I am one among the supposita of 'cognoscens venientem' with the
determinate appellation of the concept of 'veniens', then I also have to be
among the supposita of 'venientem cognoscens' with the appellation of
some concept, one among these possible appel/ata being the concept of
'veniens', but not conversely, just as if I see Socrates, I see someone, but
if I see someone, it does not follow that I see Socrates.

There is, however, a further, at first sight rather unintuitive claim

made by Buridan with respect to the implications of 'Debeo tibi equum'.
For he claims not only that 'Debeo tibi equum' implies 'Equum tibi debeo',
but further that 'Equum tibi debeo' implies 'Omnem equum tibi debeo', on
the basis that my obligation concerns particular horses through the general
concept of 'horse', which, in tum, concerns equally every horse, whence
my obligation, through the general concept of 'horse', concerns equally
every horse. On the other hand, according to Buridan, this does not mean
that lowe you every horse, for you can claim from me only what is
signified by the term under which the obligation was made, but I did not
oblige myself to give you every horse, only a horse. (In 'lowe you every
horse' the ratio appellata would be that corresponding to 'every horse', not
only that corresponding to 'horse'. Hence, since change of appellation
changes truth-conditions, the passage from 'lowe you a horse' to 'lowe
you every horse' does not preserve truth.)

Without going into detailed discussion of the matter, I only wish to

show here very briefly what kind of further semantic considerations may
prompt such a conclusion.

But to this end, first we have to see in general what effect the addition
of a distributive sign to an oblique term in a complex term may have on the
supposition of the complex term. For example, it is clear that the term
'videns asinum' supposits for me if I actually see an ass, but the term
'videns omnem asinum' supposits for me only if I see every ass, i.e. every
suppositum of the term' asinus'. But this means that I am a suppositum of
the term 'videns omnem asinum', only if for any choice of a suppositum of
'asinus' the term 'videns' signifies me in relation to that suppositum, that
is, in general, a suppositum of this complex term is a thing that is signified
by 'videns' for any choice of a suppositum of 'asinus'.

This may be formulated as follows:

SUP(,videns omnem asinum')(t) =

= SGT('videns')(SUP'('asinus')(t))(t), for any SUP',

identity even on modern standards. So along these lines, in the framework of a

thoroughgoing model-theoretical reconstruction of Buridanian semantics we may
eventually appease our modern qualms concerning quantification over rationes.
 'DEBEO T/B/ EQUUM': BUR/DAN'S TREATMENT 345

if there is aVE W such that SGT(,videns')(u)(t)=v for any u, otherwise

SUP('videns omnem asinum')(t)=O

where UE (w: for some SUP', w=SUP'('asinus')(t)}, i.e. u is a

suppositum of 'asinus' at time t, and, adding the clause familiar from
quantification theory, SUP' is the same as SUP except for the value
assigned by it to 'asinus', which makes it clear that SUP functions here as
an ordinary value-assignment, and 'asinus' and the other general terms as
restricted variables in many-sorted quantification.35

But regarding Buridan's claim concerning 'Ornnem equum tibi debeo'

this amounts to saying that if this sentence is true, then for any choice of an
individual suppositum of the term 'equus', the term 'debens' signifies me
with respect to that horse considered under any concept whatever.

Making use of our notation:

SUP('equum tibi debens')(t)

=SUP('ornnem equurn tibi debens')(t)

=SGT('debens ')(SUP('te '»(RAT(X»(SUP'('equus ')(t»(t),

for any SUP'. But this assumption is indeed quite plausible: after all it is
something like this that we mean by owing just any horse in general, no
matter which one in particular. But this plausible semantic assumption does
have as its consequence the above-mentioned unintuitive sounding
implication. However, despite its rather unintuitive sound, on seeing that it
is a consequence of a plausible semantic assumption, one may eventually
accept the conclusion that 'lowe you a horse' implies 'Every horse lowe
you', without, however, implying 'lowe you every horse', of which
Buridan's arguments also try to convince us.

5. Conclusion
In conclusion I wish to give only a brief illustration of the explanatory
power of Buridan's theory and remark on its consistency with his more
general philosophical ideas in the light of the above reconstruction.

In particular, I would like to show that with a minimal modification of

its starting assumptions Buridan's theory is also able to account for the
intuitions backing Ockham's treatment of sentences with intentional
verbs.36 To see this, we only have to notice that Ockham's contrary claim

35Concerning this topic see my earlier reconstructions in my Ars Artium (see n.ll.
above). See also Sophismata pp. 191-6 and 347-8. Concerning SUP('asinus')(t) see
the similar case of SUP('visus')(t) in n. 22 above.
36For Ockham's treatment of the similar case of 'Promitto tibi equum' see SL Part I
ch.n, pp. 219-21. cf. Part II ch. 7. Cf. also Guillelmi de Ockham, Scriptum in librum
primum Sententiarum Ordinatio, St. Bonaventure N.Y.: Franciscan Institute 1967-
1979; d.2.q.4., pp. 145-8.
346 GYULAKLlMA

(namely, that the proposition 'Equum tibi promitto' is not true in the
posited case because the term 'equum', being preposited to the verb,
supposits in it determinately) may be obtained in Buridan's theory from the
assumption that if the term is preposited to (the participle derived from) the
intentional verb, then, instead of having a disjunctive appellation, it does
not have appellatio ration is at all, but simply stands for some of its
supposita. With this assumption, contrary to his own claim, Buridan's
theory predicts that 'Promitto tibi equum' does not imply 'Equum tibi
promitto'.

For with this assumption it may well be the case that although in the
complex term 'promittens tibi equum' the participle 'promittens' signifies
me in relation to the ratio of 'equus', and so the complex term supposits
for me with the appellation of this ratio in 'Ego sum promittens tibi
equum', in which case this sentence is true, still, 'equum tibi promittens'
does not supposit for me without any appellation of the ratio of 'equus', or
of any other term in 'Ego sum equum tibi promittens', in which case this
sentence is false. For if we determine the supposita of 'equum tibi
promittens' in the following manner:

SUP('equum tibi promittens')(t)

=SGT('promittens')(SUP('te'))(SUP('equus')(t))(t),

then I may be one among its supposita only if I am signified by

'promittens' in relation to this horse, or in relation to that horse, and so on,
that is, if I promise you some horse determinately, which is not the case
when only 'Promitto tibi equum' is true.

So Buridan's theory is able to account for the intuitions of those who

feel uneasy about his actual solution: these intuitions dictate that we take
'equus' as standing determinately and without any appel/atio rationis in
'Equus tibi promittitur', in which sense this proposition is, of course,
false, if no determinate horse was promised to you, but only a horse in
general. So those who share their intuitions with Ockham and claim that no
horse is promised to you in such a situation may respect at least Buridan's
theory, even if they dislike his actual solution, insofar as it is this theory
that may account even for their intuitions.

On the other hand, in this case it is their task to account for the other,
epistemologically important cases of 'Cognosco triangulum' and
'Triangulum cognosco' where it would indeed be valde durum, to use
Buridan's phrase, to accept 'Nullum triangulum cognosco'. As we know,
Buridan's solution was also motivated by these epistemological
considerations.37 We also know that his rationes playa prominent role in
his general theory of significatio ad placitum. Finally, we know that since
these rationes are individualised qualities of individual human minds, they
fit in nicely with Buridan's nominalistic metaphysics and philosophy of
mind. So whatever particular misgivings we may have concerning his
actual treatment of the case of 'Debeo tibi equum', we cannot but respect

37See Sophismata, p. 86 and pp. 76-8, and Tractatus de Suppositionibus, pp. 333-5.
 'DEBEO TIBI EQUUM': BURIDAN'S TREATMENT 347

the consistency of Buridan's theory with his more general philosophical

considerations. 38

Institute a/Philosophy a/the Hungarian Academy a/Sciences

381 wrote this paper during my stay in Helsinki as member of Simo Knuuttila's project:
Ockham and the via moderna. lowe thanks to the Finnish Academy for their generous
financial assistance and to all the Finnish friends and colleagues for their help and
encouragement. But my thanks are due above all to Simo for the extremely useful
discussions and his help and advice in all kinds of problems, both practical and
theoretical. The final version of the paper was prepared during my stay in St. Andrews
as Gifford Visiting Fellow at the Department of Logic and Metaphysics. 1 am grateful
to Stephen Read for helpful comments and for correcting the English of the paper.
 Trinitarian Sophisms in Robert Holcot's Theology
by Simo Knuuttila

1. In his Questions on Peter Lombard's Sentences, Robert Holeot

discusses many arguments allegedly showing inconsistencies or
contradictions in the articles of faith or in the received interpretations of
them.' Holcot thought that, in so far as the conclusions of these arguments
are contrary to Catholic doctrine, they could be called sophisms. One must
suppose, he said, that something was wrong either in their form or in their
matter, independently of whether any philosophical disproof is available.
(Sent. I, q. I, a. 6, q 5.) Ever since the works of Konstanty Michalski,
scholars have discussed the question whether Robert Holeot could be
characterized as a representative of scepticism or not. 2 Clearly, Holcot's
theological approach does not have much in common with the Pyrrhonian
scepticism of Sextus Empiricus, but one might try to see it as a fideistic
version of the dogmatic scepticism, represented by the Academics, which
started from the axiom that nothing can be known} Holeot's fideism
assumes that the human intellect is so thoroughly corrupted in the original
fall that it is unable to understand divine matters without grace. God causes
faith in some people, but the articles of faith remain mysteries even to
them. (Sent. I, q. 1, a 5.)

When discussing the question 'Whether God is three distinct persons',

Holeot first lists sixteen arguments for the negative answer. According to
him, the premises of some of the Trinitarian sophisms are true and they are
not formally faulty in the light of Aristotle's logic. Holcot then argued that,
since their conclusions are in spite of this not acceptable for Catholics, they
should think that Aristotle's logic does not hold universally. It is not
formal in the sense of being valid in every domain of discourse. It can be
called formal in the restricted sense that "it does not admit counterexamples
in natural inquiry into things known to us by the senses". (Sent. I, q. 5)
According to Holeot, one can construct from revealed Trinitarian and
Christological truths examples of the cases in which natural logic and a

'In quatuor libros Sententiarum quaestiones, Lugduni 1518, repro Frankfurt am Main:
MinelVa Nachdruck 1967.
2See H. Gelber, Logic and the Trinity: A Clash of Values in Scholastic Thought 1300-
1335, University of Wisconsin Ph. D. dissertation 1974, pp. 268 ff.
3For ancient skepticism, see Doubt and Dogmatism, ed. M. Schofield, M.F. Burnyeat,
and J. Barnes, Oxford: Clarendon Press 1980; J. Annas and J. Barnes, The Modes of
Skepticism: Ancient Texts and Modern Interpretations, Cambridge: Cambridge
University Press 1985; J. Barnes, The Toils of Skepticism, Cambridge: Cambridge
University Press 1990. For slightly different views on Holcot's position, see H.
Oberman, "Facientibus quod in se est Deus non denegat gratiam: Robert Holcot, 0 .P.,
and the Beginnings of Luther's Theology", Harvard Theological Review 55, 1962, pp.
317-42; F. Hoffmann, Die theologische Methode des Oxforder Dominikanerlehrers
Robert Holcot (Beitrlige zur Geschichte der Philo sophie und Theologie des
Mittelalters), N.F. 5, Miinster: Aschendorff 1971; M.H. Shank, "Unless You Believe,
You Shall Not Understand". Logic, University, and Society in Late Medieval Vienna,
Princeton: Princeton University Press 1988, pp. 74-9.

3~8
 TRINITARIAN SOPHISMS IN HOLCOT'S THEOLOGY 349

different logic of faith (rationalis logica fidei) do not overlap. But the
nature of the relationship between natural and supernatural logic in Holcot
remains somewhat unclear, however. Holcot seems to think, as does the
early fourteenth century author of the Centiloquium,4 that certain revealed
propositions imply a supernatural logic which is based on principles very
different from those which are valid in the logic of natural reason; thus, the
principles of the latter do not constitute the universal foundation of
rationality. What makes the seemingly compelling arguments against
articles of faith sophisms is their failure to attend properly to different types
of rationality. Hester Gelber has argued that Holcot changed his position
later, and it seems that he did not in fact develop further the suggestion of a
special logic of faith.5

Historically interesting is that Holcot, in the Questions on the

Sentences, as did some others, denied the validity of traditional logic with
respect to supernatural matters but not with respect to the natural world. A
little earlier, Landulf Caraccioli and his followers had qualified the primum
principium complexum in another way. They defended the principles of
logic while speaking about the Trinity and other non-temporal theological
matters, but they simultaneously asserted that two different instants of
nature can exist with contradictory contents within a single Aristotelian
instant of time. In connection with beginning and ceasing, contradictory
sentences are simultaneously true. Nature, so to say, is not strong enough
to be totally free from contradictions.6

Many of the theological sophisms mentioned by Holcot were

commonly known and widely discussed by his contemporaries. In the
fourteenth century these problem cases were increasingly treated in
connection with the question of the validity of syllogisms, the rules of
consequences, and other logical matters in revelational theology. Standard
examples and their relations to doctrines in logic are listed systematically in
the first article of the third question of Roger Roseth's Commentary on the
Sentences'? It is a short introduction to these sophisms, or paralogisms, as
they usually were called.

Number 1 in the list of Trinitarian paralogisms by Roger Roseth is the

famous case of the expository syllogism:

haec essentia divina est pater

4See P. Boehner, "The Medieval Crisis of Logic and the Author of the Centiloquium
Attributed to Ockham" in idem, Collected Articles on Dckham, ed. E.M. Buytaert, St.
Bonaventure, N.Y.: Franciscan Institute 1958, pp. 351-72. H. Gelber has suggested
that the author is Arnold of Strelley: H. Gelber, Exploring the Boundaries of Reason.
Three Questions on the Nature of God by Robert Holcot, D.P, Toronto: Pontifical
Institute of Mediaeval Studies 1983, pp. 45-6.
5Exploring the Boundaries of Reason, pp. 26-7.
6See S. Knuuttila and A. I. Lehtinen, "Change and Contradiction: A Fourteenth century
Controversy", Synthese 40, 1979, pp. 189-207; S. Knuuttila, "Remarks on the
Background of the Fourteenth Century Limit Decision Controversies" in The Editing
of Theological and Philosophical Texts from the Middle Ages, ed. M. Asztalos,
Stockholm: Almqvist and Wikselll986, pp. 245-66.
7MS Oxford, Oriel Coil. 15, 278(270)rb - 279(271)ra. O. Hallamaa (University of
Helsinki) is working on an edition of Roseth's Commentary on the Sentences.
350 SIMO KNUUTfILA

haec essentia divina est filius

ergo pater est filius.

Then appears the much debated example pertaining to the Barbara mood:

omnis essentia divina est pater

omnis filius est essentia divina
ergo omnis filius est pater.

The examples which follow are alleged violations of the consequences ab

iriferiori ad superius affirmative sine distributione and ab affirmativa de
predicato infinito ad negativam de predicato finito. Examples 5 to 8
concern consequences based on the rules of conversion. He ends by
discussing three alleged violations of the law of contradiction. First, even
though it is said that God's essence is the Son, it also is something else,
and, since this is not the Son, God's essence, consequently, is not the
Son. Secondly, even though it is said that the Father and God's essence
are really the same, the essence is communicated, but the Father is not
communicated. The third is an analogous example based on the notions of
generating and being generated. These three examples were usually
considered as problematic with respect to the primum principium
complexum, including the principles of identity and non-contradiction.
They also are the starting points for Petrus Rogerii, later Pope Clement the
Sixth, when he denied, in a controversy with Franciscus Mayronis, the
Law of Contradiction. It may be that Petrus Rogerii was not as simple
minded as Franciscus Mayronis portrays him; Anneliese Maier has
described his views much from the point of view of Franciscus, without
paying attention, e.g., to Petrus Rogerii's idea of connecting the
contradiction with infinite objects.8

Roger Roseth thought, as did most of his contemporaries, that

standard logic is universally valid. According to him, the Trinitarian
paralogisms are solved by noticing, for one thing, that the terms in
Trinitarian propositions may supposit either for the essence or for the
persons, which may cause confusion if not carefully noted. Furthermore,
one should explicate the exact form of the syllogistic moods when
discussing divine matters. So in an expository syllogism, 'hoc X' means
'omne quod est hoc X' and omne and nul/um in syllogisms should be read
as omne illud quod est and nul/um quod est, respectively. With the help of
this device, Roseth believes, one can solve all paralogisms. The result is
that the nature of the religious terms prevents one from putting them into
certain valid logical forms.

There is nothing very original in Roseth' s short chapter about the

connection between logic and the Trinity, but it is a good example of how
the standard problem cases could be arranged and treated with the help of
logical tools. I shall not discuss the question concerning the extent to
which the doctrine of the Trinity actually had influence on the development

8See A. Maier, Ausgehendes Mittelalter II (Storia e Letteratura 105), Rome: Edizioni di

Storia e Letteratura 1967, pp. 257-85.
 TRINITARIAN SOPHISMS IN HOLCOT'S THEOLOGY 351

of logic. 9 Instead I shall make some remarks concerning the question of

conceivability and intelligibility as they appear in the discussions of
Trinitarian sophisms.

2. John Buridan states in his questions on Aristotle's Prior Analytics that

Aristotle did not see any problems with expository syllogisms, because he
thought that it would be impossible that there could be two things which
are the same as a third while yet being different from each other.1 o
Ockham, Buridan, Roger Roseth, and many others accepted the view that
such entities are correctly believed to be possible, but they were not very
interested in explaining how they are possible. In order to elucidate this, let
us have a look on Ockham' s discussion of the positio impossibilis rules of
obligations. 11

Ockham stresses that in a positio impossibilis disputation only those

propositions must be granted as relevant which follow from the positum
and/or the correctly granted propositions by natural and simple
consequences which are evident and per se notae. The propositions
presented as posita must be propositions which are not logically
impossible, because otherwise all answers could be acceptable. Ockham
seems to think that a typical impossible positum is a natural or doctrinal
impossibility .12 The only examples discussed are Trinitarian propositions
and the point is to show that theological indirect proofs are often as much
matters of faith as the articles themselves. Suppose that the positum is as
follows: the Holy Spirit does not proceed from the Son. It is then
proposed: the Holy Spirit does not differ from the Son. This, Ockham
says, is irrelevant and false and must be denied. However, if the opponent
has earlier put forward the conditional, 'If the Holy Spirit does not proceed
from the Son, the Holy Spirit does not differ from the Son', a Catholic
believer grants this as irrelevant and true. And after this, the respondent
should accept the proposition, 'The Holy Spirit does not differ from the
Son'. Ockham remarks that the game does not go in this way, if the
respondent is an infidelis.

At the beginning of part III of his Sum of Logic, Ockham states that
the articles of faith are not probable and appear false to natural reason.13
We could think that the Trinitarian forms are presented as posita to a
respondent who is using only his or her natural reason. Which kinds of

9See Ph. Boehner, op. cit.; H. Gelber, Logic and the Trinity; A. Maieru, "Logique et
tMologie trinitaire dans Ie moyen-age tardif', in M. Asztalos (ed.), op. cit., pp. 192-6.
lOQuaestiones in duos libros Analyticorum priorum Aristotelis, ed. H. Hubien
(forthcoming), I, q. 6.
11 William Ockham, Summa logicae ed. P. Boehner, G. Gat and S. Brown (Guillelmi de
Ockham, Opera philosophica et theologica, Opera philosophica I), St. Bonaventure,
N.Y.: Franciscan Institute 1974, III-3, ch. 41, pp. 739-41.
12Fourteenth century distinctions between various types of necessities are discussed in
S. Knuuttila, "Nomic Necessities in Late Medieval Thought", in Knowledge and the
Sciences in Medieval Philosophy. Proceedings of the Eighth International Congress of
Medieval Philosophy II, ed. S. Knuuttila, R. Tyorinoja and S. Ebbesen, Helsinki:
Luther-Agricola Society 1990, pp. 222-30.
130p. cit., III-I, ch. 1, p. 360. 29-32.
352 SIMO KNUUTTILA

conclusions should they grant? According to Ockham, the distinctio non-

identitatis formalis between the essence and the persons should be granted
when it evidently follows from the Scriptures or the determinations of the
church. 14 So it seems that if someone has accepted such a positum, the
formal distinction should also be regarded as a valid principle, but this is
only because the naturally impossible positum has been accepted as
possible. Ockham says that it is not any easier to accept the formal
distinction than to accept the notion of the Trinity of persons with the unity
of essence. The formal distinction does not explain anything; it is exercised
only by those who believe that the Trinity is possible. IS

If the respondent grants that the Trinitarian entity is possible, it seems

natural to think that it is an entity, as David Lewis would say, in an alien
world. Is it reasonable to speak about such entities? Perhaps we need such
worlds in order to be able to analyze sentences like, 'If the world were
totally different from what it is, we could not understand it at all'. It is
thinkable that the world could be such that we could not understand it at all
by those conceptual tools by which we think we understand the actual
world. 16 Although Ockham thought that certain articles of faith are
incomprehensible, he did not want to argue for the view that no natural
theology is possible. One could, however, develop his ideas in that
direction by stressing, e.g., that all references to God are references to the
Trinity. If the articles of faith, with respect to natural reason, are like the
posita of positio impossibilis, they could be accepted as truths about a
reality which is totally different from our familiar world. One might then
think that even when the positum and some granted impertinent
propositions about the naturally known world seem to imply a certain new
proposition, it should not be granted but only doubted, because the
meaning of the supernatural propositions is unclear.

This seems to be how Holcot thought, after having given up his earlier
idea of a special logic of faith. According to him, a Catholic must believe
that the articles of faith concerning the Trinity are true. The propositions
must be accepted separately, one by one, and the believer is permitted to
draw, with respect to them, only those consequences which are similarly
included in holy doctrine. In the question Utrum haec est concedenda,

14Scriptum in librum primum Sententiarum. Ordinatio, ed. S. Brown and G. Gal

(Guilleimi de Ockham, Opera philosophica et the%gica, Opera theologica II), St.
Bonaventure N.Y.: Franciscan Institute 1970, d. 2, q. 1, q. 11, pp. 17-18,364-78. See
also H. Gelber, Logic and the Trinity, Ch. 7; M.H. Shank, op. cit., pp. 65-71; A.
Maieru, "Logic and Trinitarian Theology: De Modo Predicandi ac Sylogizandi in
Divinis", in Meaning and Inference in Medieval Philosophy. Studies in Memory of Jan
Pinborg (Synthese Historical Library 32), ed. N. Kretzmann, Dordrecht, Boston,
London: Kluwer Academic Publishers 1988, pp. 253-4.
ISIn the Summa logicae II, ch. 2, p. 254.2-3, Ockham states that "it is impossible to
find among creatures a numerically one thing which is really many things and is each
of those things, as is the case with God." Ockham probably thinks that the invariant
structures of the actual domain of finite beings exclude such things, i.e., they are
naturally impossible.
16Cf. D. Lewis, On the Plurality of Worlds, Oxford: Blackwell 1986, p. 91. See also
M. Kusch, Language as Calculus vs. Language as Universal Medium. A Study in
Husserl. Heidegger and Gadamer (Synthese Library 207), Dordrecht, Boston, London:
Kluwer Academic Publishers 1989, pp. 93-102.
 TRINITARIAN SOPHISMS IN HOLCOT'S THEOLOGY 353

Deus est Pater et Filius et Spiritus Sanetus, edited by Hester Gelber,

Holcot says:

"Catholicus nulla logica debet uti in concedendo et negando sive

propositiones sive consequentias nisi detenninatione ecclesiae."17

Although the views of Ockham and Holcot differ here, both of them
thought that if one starts from the notion of the possibility of strange
entities, then, instead of trying to construct the laws of possible alien
worlds, one should say that we possibly could understand them if we
leamed a new system of symbolic representation. Ockham applied this idea
only to some Trinitarian forms, while Holcot considered those examples as
paradigm cases for theology in general. Holcot's positivist attitude towards
revelation is an exaggerated version of Ockham's theory, according to
which our insufficient knowledge of the properties of the terms of the
Blessed Trinity prevents us from using them as replacements of variables
in logical forms.i8

3. Some analogous questions were treated in fourteenth century

discussions of the interpretation of the rules of obligational disputations.1 9
The discussion originated from the Scotist revision of the old positio rules,
which made intelligible the list of answers that earlier was nonsensical. In
the Scotist model, the list of -answers can be understood as a partial
description of a possible world at the moment of disputation. In the
traditional rules, it was fIrst stated that everything which is positum and put
forward during the disputation must be granted. Propositions were then
divided into pertinent and impertinent ones. A pertinent proposition either
necessarily follows from the positum and/or what has been granted and/or
the opposite(s) of what has been correctly denied or is incompatible with
them. Other propositions are impertinent, and they are granted, denied, or
doubted according to their own qUality. The rule deleted by Duns Scotus
stated that if the positum is a contingent proposition which is false at the
present instant, then it must be denied as repugnant that now is that instant.
As the rules concerning pertinent and impertinent propositions were not
qualifIed in the Scotist approach and the consequences used were supposed
to be logical consequences, one easily slipped into worlds very different
from the actual one. This fact probably led some people, like Richard
Kilvington, to think that impertinent propositions should be given the
responses which would be appropriate if the positum were true and the

17Op. cit., 35. 102-4.

18For a similar interpretation of the Ockhamist view in Luther, see R. Tyl)rinoja,
"Proprietas verbi. Luther's Conception of Philosophical and Theological Language in
the Disputation Verbo carum factum est (Joh. I: 14), 1539", in Faith, Will, and
Grammar. Some Themes of Intensional Logic and Semantics in Medieval and
Reformation Thought, ed. H. Kirjavainen, Helsinki: Luther-Agricola-Society 1986,
gpo 141-75.
1 See, e.g., P.V. Spade, "Three Theories of Obligationes: Burley, Kilvington and
Swyneshed on Counterfactual Reasoning", History and Philosophy of Logic 3,1982,
pp. 1-32; E. Stump, Dialectic and its Place in the Development of Medieval Logic,
Ithaca and London: Cornell U.P. 1989, pp. 177-269; E.J. Ashworth, "The Problems of
Relevance and Order in Obligational Disputations: Some Late Fourteenth Century
Views", Medioevo 7,1981, pp. 175-93.
354 SIMO KNUU1TlLA

imagined world should otherwise differ as little as possible from the actual
world. This was an attempt to qualify obligational rules with the help of
counterfactual conditionals, and it is connected with the discussions of the
distinction between logical and nomic necessities in the fourteenth century.
Another revision of the rules was offered by Roger Swyneshed. It
contained two much debated principles which were formulated by
Swyneshed as follows: One need not grant a conjunction in virtue of
having granted all its conjuncts, and one need not grant any part of a
disjunction in virtue of having granted that disjunction. Without entering
into the details, we may note that, instead of operating with a one matrix
model of answers, Swyneshed made use of two distinct matrices, one for
the positum and relevant propositions, and one for irrelevant propositions.
Propositions are considered relevant or irrelevant solely by their relations
to the positum. Both columns are treated separately, except that an
irrelevant proposition already discussed during the disputation can be
presented again as a second positum and added into the positum matrix.20

In Swyneshed's approach one treats simultaneously two possible

worlds which have nothing in common. Similarly Holeot thinks that a
Catholic has in his mind a column for the answers of natural reason and a
column of faith and that there is no connections between them, except that
something may be added to the column of faith as a new positum. The
philosophical similarity between Holeot's treatment of propositions of faith
and Swineshed's obligational rules is based on the view of unconnected
domains of discourse. In Holcot's time people were acquainted with a
further version of obligations rules which has not been much attended to in
modem discussion and which may shed new light on the habits of thought
reflected in Swyneshed's obligations rules and in Holeot's theological
method, respectively. In his questions on Aristotle's Prior Analytics,
Richard of Campsall sketches rules for games in which there are two
opponents and correspondingly two different posita. As the answers given
to one opponent become pertinent only with respect to the answers given to
that opponent, the Swyneshed type conjunction and disjunction rules are
considered natura1. 21 Holeot may have been familiar with these rules,

20For the rules of Swyneshed, see P.V. Spade, "Roger Swyneshed's Obligationes:
Edition and Comments", Archives d' histoire doctrinale et litteraire du moyen age 44,
1977, pp. 243-85; E. Stump, "Roger Swyneshed's Theory of Obligations", Medioevo
7, 1981, pp. 135-74, repro in idem, Dialectic and its Place ... , pp. 215-49; S.
Knuuttila and M. Yrjonsuuri, "Norms and Action in Obligational Disputations" in Die
Philosophie im 14. und 15. Jahrhundert. In memoriam Konstanty Michalski (1879 -
1947). ed. O. Pluta (Bochuiner Studien zur Philosophie 10), Amsterdam: Gruner 1988,
pp. 199-202.
21 The Works of Richard of Campsall I: Quaestiones super librum Priorum
Analeticorum, ed. E. A. Synan, Toronto: Pontifical Institute of Mediaeval Studies
1968, 15.11-15.18, 15.47, pp. 227-9, 237-8. Campsall writes: "Et quando dicitur quod
tunc ista copulativa, facta ex istis duabus positis ['tu sedes', 'tu non sedes'], esset
concedenda, quia utraque pars iIlius copulative est concedenda, dicendum est quod
copulativa ista non est concedendi quia, si esset concedenda, hoc esset vel in una
disputacione, vel in alia, sed in neutra est concedenda, cum in utraque tantum possibile
est obligatum." Op. cit., p. 237. For this interpretation of Swyneshed's rules by
Stanislaus ofZnaim, who taught in Prague in 1388-1413, see his De vero etfalso, ed.
V. Herold, Prague: Czechoslovak Academy of Sciences 1971, pp. 209.12-211.13.
 TRINITARIAN SOPHISMS.IN HOLCOT'S THEOLOGY 355

although he does not refer to them.22 When formulating the rules of what a
Catholic should concede or deny, Holcot remarks that the answer to the
question of whether something is contradictory or not is different,
depending on whether it is given in the light of natural reason or in
accordance to the Catholic doctrine.23

One traditional theological doctrine was also connected with the theme
of different domains. In his book De genesi ad litteram, Augustine had said
that God is the author of two books, the book of Scripture (tiber
scripturae) and the book of Nature (liber naturae). This idea was
commonly employed in medieval theology in connection with the question
of faith and reason. The divine authorship of the book of nature gave
certain legitimation to natural philosophy and theology which, in the
Augustinian tradition, was thought to be improved by divine revelation,
which included truths available only through faith. In the thirteenth
century, the two books model was sometimes modified to the effect that
the divine intellect itself was called a book containing everything that can
be known. The books of Nature and of Scripture could then be regarded as
two partial editions of the original book.24 In HoIcot's fideistic approach,
reading one book is a quite different activity from reading the other.

When one finds some prima facie contradictory descriptions of divine

entities in the book of faith which one believes consistent, why not try to
make them understandable? Why did Ockham and Holcot think that we
cannot improve our modes of understanding so that the allegedly possible
supernatural states of affairs would become comprehensible to us?
Ockham claims that the categorematic signs of mental language have a
natural signification, which means that semantic relations of mental
language, the language of thinking and understanding, are inaccessible. 25
Human understanding, consisting of natural categorematic terms,
syncategorematic terms, and their combinations, is a fixed system. If
something which cannot be conceptualized in it is claimed to be possible, it
must be treated as an element of a world which is totally different from the
one connected with our mental language. As such it cannot be understood.
Perhaps we cannot understand it in the strict sense that in order to
understand it, we should leave the mental language which is the universal
medium of human understanding. If this was the reason why Ockbam and
Holcot thought that it was impossible to construct new intelligible modes
of understanding, a critic could have said that it is not necessary to give up
all meaning simultaneously. Nor do we have to leave the world we
understand and leap in one bound to another world which we don't

22Ho\cot describes the positio rules in Sent. II, q. 2, a. 9.

23Utrum haec sit concedenda, ed. Gelber, 39.177-187.
24See U. Kopf, Die An/tinge der theologischen Wissenscha/tslehre im 13. Jahrhundert
(Beitriige zur historischen Theologie 49), Tiibingen: J.C.B. Mohr 1974, pp. 233-4,
238,249-50; I. Dalferth, Theology and Philosophy, Oxford: Blackwell 1988, pp. 67-
75. Dalferth states, referring to B. Lohse's book Ratio und Fides, Gottingen:
Vandenhoeck and Ruprecht 1958, that the via moderna "could only avoid making
theological matters utterly incomprehensible by proposing an epistemological dualism
between philosophy and theology, couched in terms of a double logic, a double truth
and a double mode of knowing". As for logic, the characterization is exaggerated.
25Summa logicae I, ch. I, pp. 7-9.
356 SIMO KNUU1TlLA

understand. This would be a total loss of meaning. We could also proceed

step by step, thinking that semantic relations, instead of being inaccessible,
are inexhaustible. 26 It seems that Ockham and Holcot would have not
accepted this idea; it was beyond the purview of their conception of
understanding.

University of Helsinki

26Cf. M. Kusch, op. cit., pp. 2-10; J. Hintikka, "Wittgenstein's Semantical

Kantianism", in Ethics. Foundations, Problems and Applications, ed. E. Leinfellner et
aI., Vienna: HOider-Pichler-Tempsky 1981, pp. 375-90.
 Obligations and Liars
by Christopher J. Martin

The best efforts of some of the best historians of logic have failed to
find a clear source in material inherited from the ancient world for the
earliest mediaeval discussions of the Liar paradox.! The obvious
candidates, and most importantly Aristotle's reference to the puzzles of
oath-breakers and liars in the Sophistical Refutations, seem to have been of
little importance in the first theorising about insolubilia.2 Despite this I
would like to suggest that there was at least a very distant cause in late
antique logic for the appearance of the Liar and its relatives in the twelfth
century. It cannot be emphasized enough, however, that twelfth century
logicians devised the paradoxes for themselves and that their solutions
were all their own work.

Three things are important in the construction of the Liar and

variations upon it. As Saul Kripke3 has emphasized, in the appropriate
circumstances just about any statement involving truth or falsity may tum
out to be paradoxical. Once this is pointed out it is easy to see that there can
be no syntactic or semantic sieve in which to trap all instances of
ungroundedness. Whether or not, for example, utterances of the sentences
'Everything that Socrates says is false' and 'Everything that Plato says is
false' are paradoxical depends upon who says them, what else they say,
when they are said, and the way the world happens to be then. No doubt
circularity sometimes occurs by chance but in general we need to arrange
the circumstances of utterance, inscription, or thought so as to guarantee it.
This stage setting was done in the middle ages by means of a hypothesis or
positio. The theory of such hypotheses was regimented in treatises on the
obligatio of the same name and it is to these, I suggest, that we should look
if we want to understand how the mediaeval concern with insolubilia
originated and the character of the earliest solution.

lAbbreviations: TEF =Tractatus Emmeranus de Falsa Positione; TEl =Tractatus

Emmeranus de Impossibi/e Positione; both in L.M. De Rijk, "Some Thirteenth
Century Tracts on the Game of Obligation", Vivarium 12, 1974, pp. 94-123. OP =
Obligationes Parisiensis. in idem, "Some Thirteenth Century Tracts on the Game of
Obligation", Vivarium 13, 1975, pp. 22-54. 1M = Insolubilia Monacensis, in idem,
"Some Notes on the Mediaeval Tract De Insolubilibus with an Edition of a Tract
Dating from the End of the Twelfth Century", Vivarium 4, 1966, pp. 83-115. Here see
De Rijk, "Some Notes"; P.V. Spade, "The Origins of the Mediaeval Insolublia
Literature", Franciscan Studies 33, 1973, pp. 292-309, reprinted in Lies, Language and
Logic in the late Middle Ages, London: Variorum Reprints 1988; and P.V. Spade,
"Five Early Theories in the Mediaeval/nsolubilia Literature", Vivarium 25, 1987, pp.
24-46.
2See in particular Spade, who, in "The Origins ... ", discusses St. Paul's letter to Titus,
Gellius' Noctes Atticae, Cicero's Academica and Augustine's Contra Academicos. The
earliest use of the distinction between being X simpliciter and being X secundum quid
in the solution of the Liar seems to date from the beginning of the thirteenth century.
There is much disagreement over the appropriate X.
3Saul Kripke. "Outline of a Theory of Truth", The Journal of Philosophy 72, 1975, pp.
690-716.

357
358 CHRISTOPHER J. MARTIN

Secondly, in order to achieve paradoxical circularity or self-reference a

certain amount of semantical and syntactical equipmentis necessary. In
particular 'true' and 'false' must be available as predicates in the language
and must behave in the appropriate way in conjunction with a suitable
referential apparatus. The paradoxes arise because of a conflict between
our intuitions about truth and falsity and our intuitions about the power of
natural languages to deploy the corresponding predicates. Furthermore the
choice between sentences, statements, and propositions as bearers of truth
will lead to different formulations of and different solutions to the Liar. An
inscriptionalist like Buridan can, for example, make moves which would
be entirely unacceptable to someone who maintains a realistic account of
propositional contents.

Thirdly, the acceptance or rejection of particular logical principles may

facilitate or prevent the derivation of disturbing consequences from the Liar
sentence. I have in mind for example the Parvipontanian adoption of strict
implication and the associated paradoxes and, on the other hand,
Abaelard's qualifications of the principles of quotation and disquotation
and the requirement of the Melidunenses and Porretani that for a
conditional to be true it is necessary for the antecedent and the consequent
to be about the same thing. In this paper I will limit myself to a discussion
of the first two points. 4

1. Obligations
The writers of the mediaeval treatises on obligations are irritatingly
vague about both the history and the purposes of their works. There are
some clues, however, to the development of the obligation known as
positio and these suggest an account of its intended application.

1.1 Eudemian hypotheses and procedures

The treatises published by de Rijk as the Tractati Emmerani <TE =
TEF + TEl> and the Obligationes Parisiensis <OP> are certainly the oldest
accounts of obligations discovered so far. Each justifies impossible positio
with an appeal to Aristotle who, it claims, requires that "an impossibility be
posited in order to see what follows from it."5 There seems to be no such
injunction in Aristotle's writings but something rather like it is found in
Boethius' De Hypotheticis Syllogismis. 6 There Eudemus is credited with

4For the beginnings of an account of the different logics invented in the twelfth century
see my "Embarrassing Arguments and Surprising Conclusions in the Development of
the Theory of the Conditional in the Twelfth Century", in Gilbert de Poitiers et ses
Contemporains, ed. I. Iolivet and A. de Libera, Naples: Bibliopolis 1988.
5"lmpossibile est ponendum ut videatur quid inde sequatur" <OP, 52.10-11; cf. TEl
117.28-31>: "Quod impossibile positio habeat sustineri sic probatur. Sicut enim nos
dicimus quod possibile est concedendum ut videatur quid inde sequitur, similiter
habemus ab Aristotile quod impossibile est concedendum ut videatur quid inde accidat";
TEF 103.21-23 gives the justification for positio possibilis.
6"Hypothesis namque, unde hypothetici syllogismi accepere vocabulum, duobus, ut
Eudemo placet, dicetur modis: aut enim tale allquiescitur aliquid per quamdam inter se
consentientum conditione[m] quae fieri nullo modo possit, ut ad suum terminum ratio
perducatur; aut in conditione posita consequentia vi coniunctionis vel disiunctionis
 OBLIGATIONS AND LIARS 359

distinguishing between a hypothesis, or positio, which is the condition of a

consequence, i.e. the antecedent of a conditional, and a hypothesis by
means of which "something which can in no way come about is agreed to
... in order that reason may be pursued to its limit."7 As an example
Boethius cites the separation in thought of the form of a body from its
matter so that "for a little while that is understood to be so which can in no
way come about" in order that we may consider what follows.

Let us call these conditions of agreement Eudemian hypotheses and the

arguments carried out under them Eudemian procedures. Garlandus
Compotista characterizes Eudemian hypotheses as being per consensum
and offers the example of conceding "that a man is a stone, in order to see
how far reason can go."s Nothing like this hypothesis is found in
Garland's authority The Book, which consists of the works of Boethius',
but the use of positio impossibilis in the exploration of the logic of
impossible unions is the main topic of TEl. An account of such unions is
also of great importance to Abaelard in developing his theory of the
conditional. Abaelard characterizes a Eudemian hypothesis as by "consent
and concession ... a proposition which is not taken as true in itself, but is
rather conceded for the sake of argument, so that we may see what can be
extracted from it."9 Notice that he does not insist upon the impossibility of
the hypothesis.

Boethius' example is not particularly helpful but one thing is clear, he

thinks that Eudemus had in mind something quite different from the
method of reduction to an impossibility employed in the theory of the
syllogism. There the aim is to show that a hypothesis is impossible by
showing that an impossibility follows from it and appealing to Aristotle's
rule that the impossible cannot follow from the possible. In a Eudemian
procedure on the other hand we start with something agreed to be
impossible and set out to explore its logical structure. As we will see,
Aristotle's rule construed as requiring actualizability justifies possible

ostenditur. Ac prioris quidem propositionis exemplum est, veluti cum res omnes
corporales materiae formaeque concursu subsistere demonstramus. Tunc enim quod per
rerum naturamfieri non potest, ponimus, id est ?omnemformae naturam? a subiecta
materia, si non re, saltem in cogitatione separamus; et quoniam nihil ex rebus
corporeis reliquum fit demonstratum atque ostensum putamus eisdem convenientibus
corporalium rerum substantiam confici, quibus a se disiunctis ac discedentibus
interimatur.ln hoc igitur exemplo posita consentiendi conditione, ut id paulisper fieri
illtelligatur quod fieri non potest, id est ut formae a materia separentur, quid consequatur
intendimus, perire scilicet corpora, ut eadem ex iisdem consistere comprobemus."
A.M.S. Boethius, De Hypotheticis Syllogism is, ed. L. Obertello, Brescia: Paedeia
1969, I, ii, 5-6.
7The example recalls the "striping away" of Metaphysics Zeta 3, and the procedure
Aristotle's reflections on definition in Zeta 10, II. The distinction between two kinds
of inseparability discussed below is found in a primitive form in De Anima in the
distinction between "spatial" separability and separability in account
S"Quando aliquid impossibile conceditur ut homo est lapis gratia videndi a quemfinem
ratio perveniat." Garlandus Compotista, Dialectica, ed. L. M. de Rijk, Assen: Van
Gorcum 1959, p. 127.
9o'Per consensum et concessionem hypothesis propositio ilia dicitur quae non in se vera
recipitur, sed gratia argumentandi conceditur, ut quid ex ea possit extrahi videatur."
Petrus Abaelardus, Dialectica, L. M. de Rijk, Assen: Van Gorcum 1970, IV, I, p.471.
360 CHRISTOPHER 1. MARTIN

rather than impossible positio. It is indeed cited to this end in some later
accounts of obligations though it does not appear in TE or 0 p.1O

Boethius employs Eudemian procedures on a number of occasions but

most explicitly in his treatment in Quomodo Substantiae of the problem of
how created things can be good in so far as they exist yet not be substantial
goods. If we look at some twelfth century commentaries on Quomodo we
find an interesting development in the descriptions of Boethius' request
there that we do in thought what is impossible in act. Gilbert of Poitiers
does not have a special term for it but Thierry of Chartres refers to the
exercise as positio impossibilis per consensum. Finally Clarembald of
Arras calls Boethius' hypothesis a positio impossibilis and finds in his
argument a positio possibilis."

None of these writers mentions obligatio but it is clear that the kinds of
thought-experiment that Boethius proposes can be well conducted only if
there are some rules about what counts as an acceptable inference.
Philosophers spend much of their time reasoning under hypotheses and in
the middle ages they registered this activity with the words 'ponitur' and
'supponitur'. One of the functions of the positio was certainly to regiment
the varieties of hypothetical reasoning and to provide practice in the rules
of inference applicable in them.

Unfortunately Boethius tells us nothing about the formalities of the

kind of "counterpossible" reasoning employed in Eudemian procedures.
As I have shown elsewhere, however, from various of his works it is
possible to extract two notions of inseparability and corresponding to them
two accounts of consequence. In the weaker sense two things are
inseparable if they cannot actually be separated but they may nevertheless
be conceived apart. I call this strict inseparability. The corresponding strict
conditional is true just in case it is impossible for the antecedent to be true
and the consequent false at the same time. In the stronger sense two things
are inseparable if they cannot even be conceived apart. This I call
conceptual inseparability. The corresponding conditional holds just in case
one cannot conceive the antecedent without conceiving the consequent. 12

Though poorly conducted, the thought-experiment of Quomodo gives

some content to the distinction between strict and conceptual inseparability.
If a situation, albeit an impossible one, can be constructed in which two
things exist in some sense independently of one another, then they are not
conceptually connected and so neither belongs in the definition of the

IOTE in particular seems not to know Aristotle very well. In giving examples of
understandable impossibilities. it cites him as speaking of a fish removed from water in
such a way that nothing occupies its place <TE/IIS.4-6>. Aristotle seems nowhere to
come close to saying this.
II Thierry of Chartres. "Commentum Super Ebdomadas Boetii". p. 421.24. 423.16. in
Commentaries on Boethius by Thierry of Chartres and his School. ed. N. Haring
Toronto: Pontifical Institute of Medieval Studies 1971; Clarembald of Arras.
"Expositio Super Librum Boetii 'De Hebdomadibus· ... in Life and Works of
Clarembald of Arras. ed. N. Haring. Toronto: Pontifical Institute of Medieval Studies
1965, p. 212.4 ff.
l2See "Embarrassing Arguments".
 OBLIGATIONS AND LIARS 361

other. Boethius' authority makes this sort of reasoning respectable and

important but he offers no hint as to how we are to avoid begging the
question of what is conceptually separable from what.

1.2 Impossible Positio

In the twelfth century the most important work on this problem was
done by Abaelard and its results can be seen in the Tractatus Emmeranus
de lmpossibile Positione. Although insolubilia are not introduced in this
kind of positio, I would like to look at the work briefly since I think it can
provide some useful clues to the early history of their treatment and also
help to answer a singularly vexed question in the history of logic.

TEl, in accordance with the principle attributed to Aristotle, will allow

us to posit any impossibility which may be understood. It claims that since
we can understand the claim that that God is man is true and that this is
harder to understand than that that a man is an ass is true we may posit the
latter. The comparison and the example tum out not to be accidental. TEl is
for the most part devoted to a discussion of the various kinds of impossible
union and has a logic to offer for reasoning about them.

In impossible positio the positum is of course impossible but since the

aim is to avoid conceding contradictory proposita it cannot be such that
contradictory opposites may be derived from it by means of the principles
of inference appropriate for this sort of positio. Thus in agreeing to take
part in an impossible positio we must forgo the pleasure of appealing to the
thesis, attributed to the Adamati in TEl, that whatever you like follows
from an impossibility. That is not all we must give up, however, since
inconsistency may be derived from an impossibility independently of the
rule. Suppose, for example, that we posit as TEl suggests that a man is an
ass. We may then appeal to the locus from opposites to prove that a man is
not a man. This move is exactly what Abaelard's theory of the conditional
is devised to avoid. TEl embraces at least part of that theory by insisting
that consequences with affirmative antecedents and negative consequents
must not be accepted in impossible positio and in allowing no appeal to the
locus from opposites.

More interestingly TEl rehearses but confuses Abaelard's arguments

against the truth of 'If it's a man, then it's not an ass'. Abaelard maintains
that conditionals of mixed quality are false since the sense of the antecedent
cannot 'contain the sense of the consequent and so they do not express
natural consequences. TEl, on the other hand, claims that such
conditionals may express natural consequences but that since we can
understand the combination of the forms of, say, a man and an ass in a
single subject, the understanding of the one does not exclude that of the
other. Since it is understanding rather than nature with which impossible
positio is concerned we are not allowed to appeal to such consequences
there.

TEl thus relies on Abaelardian logic in reasoning from an impossibility

and this is also the logic associated with the Nominales. There is,
362 CHRISTOPHER J. MARTIN

however, at least one curious thesis identified with that school which, pace
Calvin Nonnore, Abaelard apparently would not have accepted. 13 Jaques
of Vitry tells us that, in opposition to the Adamati, the Nominales
maintained that nothing grows. 14 It is striking that just this conclusion
follows immediately from the principles of predication set down in TEl.
It's possible, then, that in the Tractati Emmerani we have works connected
with the Nominales and interesting infonnation on their beliefs about
essential and personal predication.

Having mentioned one outrageous positio nominalium let me note that

another, obviously related to it, is found in the Obligationes attributed to
William of Sherwood. In what is known as vicarious positio we are asked
to respond to questions as someone else would. Walter Burley gives the
example of being asked to respond as Zeno would to the proposition that
nothing moves. In the parallel passage in pseudo-Sherwood the proposal
that something moves is put to someone required to answer as a
Nominalis. The proper response is to deny it.

It seems to me that the evidence of this positio on its own outweighs

all the similarities pointed out by Spade and Stump in arguing that Burley
was the author of the work attributed to Sherwood. ls It is surely much
more likely that Burley was updating an old, perhaps standard, text and in
this case providing an example that would make sense to his
contemporaries. The only other argument offered by Spade and Stump for
the late dating of pseudo-Sherwood is the sophistication that they find in its
distinction of different kinds of consequence. They are quite wrong,
however, in supposing that such sophistication is a late development in
mediaeval logic. It is all available and much more in Abaelard's
Dialectica. 16

l3See Calvin Normore, "Medieval Nominalism", in Studies in Medieval Philosophy,

ed. J. F. Wippel, Washington D.C.: Catholic University of America Press 1987.
Abaelard considers the claim that nothing grows and agrees that it is tnJe if you identify
a thing with its integral parts and by growth you mean that it undergoes a
multiplication of these parts. His response is that this is not the proper way to
understand growth. To mark growth we must locate something that remains fixed as
extra parts are added. What remains fixed Abaelard claims is at some level of generality
the kind of composite. Thus if to a heap of three stones I add a fourth, the heap of
stones grows but the heap of three stones does not. If I add a horse to a collection of
three men the size of the collection of animals increases. IfGod adds matter to the flesh
of a child, the child becomes larger. Abaelard's view seems to me to be exactly the
opposite of that constructed for him by Normore. He holds, to use Normore's
terminology, but not I think his own, that things certainly grow but that growth only
occurs where some status remains the same. See Petrus Abaelardus, Logica
Ingredentibus, in B. Geyer, Peter Abaelards philosophische Schriften, Beitriige zur
Geschichte der Philosophie und Theologie des Mittelalters XXI, Munster i. W.:
Aschendorff 1919-33, ad Cat. cap 14, pp. 296-305.
140ie Exempla des Jakob von Vitry, ed. J. Graven, Heidelberg: Carl Winter 1914: cf.
exemplum 105, p. 62.
ISp. V. Spade & E. Stump, "Walter Burley and the Obligationes attributed to William
Sherwood", History and Philosophy of Logic 4, 1983, pp. 9-26.
16See my "Embarrassing Arguments". Spade and Stump (n. 22) also make a curious
mistake in supposing that 'dubio' is the answer to be given when one is in doubt. It is
rather the form of the answer. One responds doubtfully by giving the answer 'proba!'.
What we have here is not, as they suggest, a new and an old terminology but simply
 OBUGATIONS AND UARS 363

Though I think that reflection upon Eudemian procedures and their

regimentation was quite probably an important factor in the early
development of positio it cannot be denied that in the texts that have
survived positio impossibilis appears as very much a poor relative of
positio falsa. This is, I suppose, how it should be. It is not often that we
are required to indulge in counterpossible inference and nowadays, indeed,
the prevailing view seems to be that it can't be done at allP Counterfactual
reasoning where the fact is contingent is, however, something that we do
all the time and something which needs to be codified.

1.3 The Purpose of Possible Positio

That much being obvious, there has been considerable disagreement
over exactly what purpose is served by the rules for positio. Recently the
debate has turned upon whether or not they were intended to provide a
means for evaluating counterfactual conditionals, conditionals, that is, such
as 'If you promised me a penny, I would be very suspicious of your
motives'. Paul Spade has argued that positio was indeed constructed for
the evaluation of such sentences but that it was not very good at its job.l 8
Eleonore Stump on the other hand thinks that this was not its job at all. 19
As far as I know Stump does not address Spade's positive arguments but
rather takes the failure of positio to give the right results as a proof that it
was not supposed to give them. Stump's view is, I am sure, correct.
Whatever motivated the introduction of the so-called nova responsio in
about 1330 there is no suggestion in treatments of positio falsa current
before that date that the device was to be employed in the evaluation or
generation of counterfactuals.20 The discussion has been confused by a
failure to distinguish between 'would' and 'might' conditionals and
between the corresponding forms of counterfactual reasoning. With these
distinctions made we can give a much more plausible account of the
workings and purposes of positio.

Accounts of positio provide two things at once: firstly a procedure for

putting claims of possibility to the test of actualization suggested by
Aristotle's criterion of possibility; secondly, and in consequence, the rules

the standard response and a description of it. See for example OP, p. 46: "Si dicas ad
hanc 'Antichristum est coloratus', 'proba!" cedat tempus. Petitum erat ut non
responderes dubie nisi ad aliquid ad quod directe obligatus esses ad dubitandum."
17But see T. Yagisawa, "Beyond Possible Worlds", Philosophical Studies 53,1988, pp.
171-204.
18p. V. Spade, "Three Theories of Obligationes: Burley, Kilvington and Swyneshed on
Counterfactual Reasoning", History and Philosophy of Logic 3, 1982, pp. 1-32.
19E. Stump, "Roger Swyneshed's Theory Of Obligations", Medioevo 7,1981, pp. 135-
74.
20Curiously, something more like 'would' counterfactual reasoning is found in ors
very brief account of the way in which one responds in rei veritas. If it has been
established in rei veritas that the Antichrist exists and it is then proposed that the
Antichrist is white, the proper response is 'proba!'. The Antichrist certainly would
have a colour but we cannot say what colour. In a positio with the corresponding
positum the appropriate response would be 'falsum est'. Rei veritas thus perhaps
anticipates Kilvington's notorious response to Sophism 47.
364 CHRISTOPHER J. MARTIN

for an exercise in which one may gain practice in recognising the truth of
conditionals and the validity of arguments. The latter is made more
complicated by the cumulative character of positio. At each stage the
respondent has to decide the relationship of the propositum to the set of
sentences all of which he is required by his earlier answers to hold true.
What follows from or is repugnant to this set is more than what follows
from or is repugnant to the positum alone since at any point arbitrary
truths, logically independent of what has gone before, may be added to the
set. This feature of positio seems to serve an essentially pedagogical
function. It complicates the antecedent that the respondent has to consider
in deciding how to reply and can make delightfully difficult the exercise of
detecting consequence and repugnance, the more so since there is no
requirement that the principle of compositionality be observed in proposing
compound proposita. Thus, notoriously, starting from the same positum
one may be required to concede a given propositum in one positio and
deny it in another.

Spade finds accounts of positio reminiscent of treatments of the

semantics of counterfactual conditionals in recent philosophical logic by
means of possible worlds. 21 In such accounts, however, possible worlds
are taken as primitive and a given counterfactual conditional is evaluated by
considering the properties of the worlds in which the antecedent is true. In
the various treatises on positio on the other hand there is absolutely no
suggestion that the problem is to decide whether a given counterfactual is
true or not. Nor is there any hint that at some point in a positio one may
stop and conditionalize to generate a true counterfactual with the positum
for its antecedent and the last conceded propositum for its consequent.
Indeed, at least according twelfth and thirteenth century logic, the result of
such a conditionalization would often be false. In the theory of
consequences there was no place for a contingently true conditional with a
false antecedent. A necessary condition for the truth of a conditional was
always that it be more or less strictly necessary and on the rare occasions
when subjunctive conditionals were noticed it was not in virtue of any
semantical peculiarity but rather because they do not satisfy Boethius'
requirement that conditionals have propositions as their parts.

What one obtains in a positio starting from a possible positum is not a

sequence of propositions which would be true if the positum were true but
rather a sequence of propositions each compossible, or better cotenable,
with the positum. The rules for handling irrelevant proposita guarantee that
the relationship between the positum and any given concessum can be no
stronger than that. Thus the most that can be "read off' a properly
conducted positio falsa is that the positum is cotenable with a given
concessum or the opposite of a given negatum.

Spade is impressed by the relationship which exists between the

positum P and a propositum Q properly conceded in a properly conducted
positio. It is (1) transitive and such that both (2) strengthening the
antecedent and (3) contraposition fail. These are certainly all properties of
'would' counterfactuals but that shouldn't impress us too much since they

21 Spade, "Three Theories ... ", p. 11.

 OBLIGATIONS AND LIARS 365

are so in virtue of the necessity that the antecedent and consequent of a true
counterfactual be compatible, unlike, say, the antecedent and consequent of
a true material implication. The important point is that the relationship
between P and Q has the properties of cotenability which 'would'
counterfactuals do not have. It is (4) symmetric and (5) such that if P and
Q are contingent and logically independent, then both Q and not Q may be
properly conceded in positiones with positum P. This latter because
cotenability follows from logical independence plus contingency.

Far from regarding it as an embarrassing difficulty to be explained

away accounts of positio dignify this last feature as the rule that: "With a
possible falsehood as positum any contingent proposition may be proved
and conceded."

The Obligationes Parisien sis <31.13-27> proves the case for a

contingent falsehood as follows. Let 'P' stand for 'You are in Rome' and
'Q' for 'You are a bishop'.
Rei veritas: -,P,-,Q
Possible: P

Po : P the positum,

Prl: P &-,Q doesn't follow from Po, and it's

false, so it is to be denied,

Pr2: -,(P & -,Q) follows from the denial of Prl so

it is to be conceded,

Pr3: Q follows from Po and Pr2 so it is

to be conceded.

Having stated the rule for possible posita and contingent falsehoods,
however, 0 P goes on to claim that the "rule will not hold for the
consequentia Nominalium" since if one takes for Q an impossibility, the
Nominales would still have to deny Prl. Given OP's account of the tests
of relevance in terms of the truth of appropriately constructed conditionals,
the alternative must have been to believe that the addition of a necessity to
the consequent of a true conditional makes no difference. That is to say the
Nominales rejected Nec(Q) 1= P ~ (P & Q). Note that this is perhaps a
weaker claim than the second Parvipontanian thesis that a necessity follows
from anything and in rejecting it the Nominales are perhaps rejecting
more. 22 Assuming that what they required was containment of sense then
the rejection of this principle seems more than reasonable. It is not clear
quite where OP stands in the debate since in the discussion of depositio,

22Correcting Normore, op. cit., p. 204 on this point. In the example given the
Nominales might perhaps have cited the principle that a conjunction of an affirmative
and a negative is negative and their general rule that a negative does not follow from an
affirmative. 0 P, however, states the objection as a general one and so as holding
presumably for affirmative P and Q with Po = P and Prl = (P & Q). For the
Nominales' account of the quality of compounds see my "Embarrassing Arguments".
366 CHRISTOPHER J. MARTIN

for obvious reasons, it instructs those who uphold the rule that anything
follows from an impossibility not to accept a necessity as a depositum.

Rather than talking of cotenability here we could use the so-called

'might' counterfactual, for example: 'If you were to promise me a horse,
then I might ask to see it'. The only true counterfactual that can be read off
the positio in the example is the relatively uninteresting one: 'If you were in
Rome, you might be a bishop'. If P is false but might be true, then a
respondent accepting it as positum and obeying the rules of positio will
never be led to contradict himself. If at some point he concedes Q, then the
relation between P and Q is such that we can say that if P were so, then Q
might be so. Conversely if this is so, then there is a positio starting from P
in which Q is conceded. That is what we might say but mediaevals would
not have done so since the positio already says it all,23

Though the procedure of positio cannot be construed in terms of the

most recent possible worlds accounts of counterfactual conditionals it does
recall the use of the so-called Ramsey test in earlier modem attempts to deal
with such propositions.24 The "metalogical" accounts of counterfactuals
proposed by Chisholm, Goodman and Mackie, each in their simplest form
propose that a counterfactual 'P > R' is true just in case there are
propositions Qt, ... , Qn which are true of the actual world, compatible
with P, and from which, in conjunction with P, R follows in virtue of
some general principle L. The case of the man who is a bishop when he is
in Rome illustrates why such accounts must fail. If P and R are logically
unrelated and R is contingently true then -,(P & -,R) is cotenable with P
and true. By modus ponens R follows from P and -,(P & -,R). This result
trivializes such "metalogical" accounts of counterfactuals and positio
interpreted as one of them. Since the result is known and not opposed in
any way in the early treatises we surely cannot regard them as attempts at
an account of 'would' counterfactuals but at best as exploring the structure
of cotenability.

I am not claiming that the procedure of positio cannot be used in the

construction of possible worlds. Indeed, my suggestion about the origins
of insolubilia will require just the opposite. A positio in which the
respondent can succeed in avoiding inconsistency despite the best efforts
of a logically omniscient opponent provides a partial valuation of the
sentences of their language. This valuation may be extended by
Lindenbaum's lemma to yield a possible world. The mediaevals did not
know Lindenbaum's lemma but the development of a positio exactly
parallels its proof. The basic intuition is expressed in the rule that an
impossibility cannot follow from a possibility. What positio provides is an
interpretation of that rule which has it that a proposition is possibly true
only if it may consistently be supposed to be true. That is to say there is a
consistent valuation in which it is assigned truth or, equivalently, some

23, should emphasize that' am not claiming that positio was intended to provide a
model of 'might' counterfactual reasoning. The treatment of irrelevant falsehoods
means that it is no more useful for that purpose than for 'would' counterfactual
reasoning.
24See for a start the introduction to lis, ed. W.L. Harper, R. Stalnaker and G. Peatce,
Dordrecht: Reidel 1981.
 OBLIGATIONS AND LIARS 367

possible world in which the state of affairs that it describes is actual.

Positio used in this way is especially common in fourteenth century
discussions of future contingency.

Notice that the arguments here for the "cotenability interpretation"

cannot apply to impossible positio. None of our sources provide rules for
concession and denial in such positiones but it would hardly make sense to
allow the introduction of arbitrary contingent truths in reasoning there.
From a positio impossibilis we may thus read off, that peculiar being, a
true conditional with an impossible antecedent. And this indeed is what
Clarembald takes Boethius to derive from the impossible positio set out in
Quomodo. The resulting counterfactual will not be contingent, however,
but rather true in the strongest sense of all, that the antecedent contains the
consequent.

2. Liars

2.1 Positio and Paradox

The aim of the respondent in false positio is to avoid being forced to

concede a per se impossibility, that is to say, both a proposition and its
contradictory opposite or a proposition like 'A man is a an ass' in which
"the form of the predicate is naturally repugnant to the subject thing."25
The respondent is bound to fail if he starts off by allowing a positum
which is inconsistent. Consider for example a twelfth century version of
Russell's barber. Suppose that as a matter of fact Socrates is not looking at
himself and I posit and it is admitted that Socrates is looking at all and only
men who are not looking at themselves. I next propose that Socrates is not
looking at himself. This does not follow from the positum - since a
categorical does not follow from a conditional - and it is true. It should
therefore be conceded. My respondent is now required to concede both a
proposition and its contradictory opposite. Consequently the positum
should not be accepted.

Propositions and sets of propositions such as this are the topic of the
little treatise De Petitionibus Contrarium. 26 None of the sophisms dealt
with there tum upon the application of seman tical terms. They correspond
rather to what in the earlier part of this century were called the logical
paradoxes. The solution to them is perfectly straightforward. The situation
described by the propositions is impossible and so imponible. The solution
offends no intuition and seems to threaten no theoretical claims.

It is quite otherwise with two other classes of propositions. If, as

Buridan does in Sophism 8 of his Sophismata, I posit that Socrates utters
only the sentence 'What Plato says is false' and that Plato utters only the
sentence 'What Socrates says is false' and I then propose 'What Plato says
is false', it seems that no consistent response is possible. Yet the positum
is perfectly well formed and could easily enough be realized in practice.

25TEF 113.12-21.
26Edited in L.M. De Rijk, "Some Thirteenth Century Tracts On The Game Of
Obligations", Vivarium 14, 1976, pp. 26-49.
368 CHRISTOPHER J. MARTIN

Our first intuition is surely that the propositum must have a determinate
truth-value but a moment's reflection seems to show that if it is either true
or false, then it is both true and false. Something deeper is wrong here and
Buridan and the rest of the fourteenth century theoreticians of insolubilia
try to save the intuition at the expense of naive Aristotelian semantics.

Neither TE nor OP notices insolubilia generated in this way but TE

warns of and OP explores an even more pathological phenomenon. In
order to generate a Liar paradox we need to find a predicated 'P' uniquely
true of the sentence 'If a sentence is P, then it is false'. Today we might
use a demonstrative or the Godel code of the sentence for' P'. The second
alternative obviously was not available in the twelfth century and curiously
the first seems not to have been employed. A suitable P was available,
however, from positio. The sentence 'The positum is false' is perfectly
well formed and true when uttered by a witness to a false positio. TEF,
however, insists that neither it nor anything convertible with it is acceptable
as a positum. The hypothesis that I would like to propose is that the earliest
mediaeval interest in insolubilia developed out of a concern with avoiding
just this sort of positum. The best evidence for this, I think, is the form
taken by the earliest known solution.27

TEF describes the problem created by the enuntiabile that afalsehood

is posited as follows: 28

"Given that the enuntiabile that a falsehood is posited may be

posited, there would follow a contradiction if that a falsehood is
posited were posited, and afterwards it was said 'Let the time cease'
and asked 'Was the positum false or was it true?' If it was true,
therefore that a falsehood is posited was true, therefore a falsehood
was posited. But nothing save that. Therefore that was false.
Therefore the positum was false. If it was false, therefore that a
falsehood is posited was false. Therefore a falsehood was not
posited. But something was posited. Therefore a truth. But nothing
save that. Therefore it was true and it was said that it was false. On
account of this it should be said that this may not be posited since
from its positio there follows a contradiction."

27Eleonore Stump has suggested on a number of occasions that there is a connection

between the sophisms found in texts on obligations and insolubilia. She is obviously
right about TE and OP but in the examples she gives from ps.-Sherwood the reference
to the respondent in the posita is quite accidental. The opponent needs to find a
positum or propositum which will change its truth value while the positio is in
progress. He could watch the clock but it's a lot easier to fix on the utterances of the
respondent. See E. Stump, "William of Sherwood's Treatise On Obligations",
Historiographia Linguistica 7: 1/2,1980, pp. 249-64.
2S"Dato enim quod hoc enuntiabile falsum poni possit poni. inde sequitur contradictio,
si ponatur falsum poni. et postea dicatur cedat tempus. Et quaeratur. Positum aut fuit
falsum aut fuit verum. Si verum ergo verum fuit falsum poni. Ergo falsum ponebatur.
Et nihil nisi hoc. Ergo hoc fuit falsum. Et hoc fuit positum. Ergo positum fuit
falsum. Si falsum. ergo falsum fuit falsum poni. Ergo falsum non ponebatur. Et
ponebatur a/iquid. Ergo verum. Sed nihil nisi hoc. Ergo fuit verum. Et dictum <est>
quod falsum. Propter hoc debet dici quod hoc non potest poni cum ex eius positione
sequatur contradictio." TEF 104.7-15.
 OBLIGATIONS AND LIARS 369

To draw a parallel with modern work on the Liar let us say that an
enuntiabile is ungrounded with respect to positio just in case it cannot be
consistently assigned a truth value as the positum of a positio. I take it that
the problem with admitting that the positum is false as a positum is the
implication that the positum expresses an enuntiabile. Since an enuntiabile
must be either true or false the proof shows that if the sentence is admitted
it cannot express an enuntiabile.

We may also say that a sentence is ungrounded in apositio just in case

it cannot be consistently assigned a truth value in that positio. Buridan,
reinterpreting ungroundedness in a positio as what George Hughes has
called "contextual inconsistency", will argue that such sentences are false.

The first problem for a respondent in positio is thus to decide whether

he is confronted with a claim of possibility which when demodalised yields
an enuntiabile ungrounded with respect to positio. OP and TE assume
without argument that such enuntiabilia are equivalent to that a falsehood is
posited. 29 Some simple examples from TEF are that the positum has the
same truth-value as afalsehood and that the positum has the opposite truth-
value to a truth. Let us say that these enuntiabilia are simply ungrounded
with respect to positio. They cannot be posited by anyone or at any time
without becoming circular. In contrast there are enuntiabilia which are only
accidentally ungrounded with respect to positio. For example that the
positum has the same truth-value as that you are a bishop. This enuntiabile
will be ungrounded with respect to positio just in case you are not a
bishop, which is to say just in case the world happens to be a certain way.
The same holds mutatis mutandis of the disjunction or conjunction of the
enuntiabile that the positum is false with a contingent enuntiabile.

Insisting that the respondent not accept as a positum an enuntiabile

which has 'positum' for its subject term will not save him. It will exclude
perfectly acceptable enuntiabilia while failing to trap all insoluble posita. I
will justify the qualification 'insolubles' in a moment. One example
mentioned by TE is that Socrates posits a falsehood and another that a
falsehood is conceded. The first is accidentally ungrounded with respect to
a positio in which Socrates is the opponent. The second certainly describes
a possibility since it is true of most well conducted positiones. Furthermore
no inconsistency follows if we suppose simply that it is posited. It is
ungrounded, however, with respect to any positio in which it is proposed
since as positum it must not be denied but if it is conceded, it can no longer
be consistently assigned a truth value. Though TEF does not prove it, it is
easy to see that on being conceded, that a falsehood is conceded becomes
equivalent to that the positum isfalse.

TEF offers no instructions to a respondent for dealing with cases like

this but an appropriate answer can I think be recovered from the treatment
of two other enuntiabilia. The first of them is decidedly peculiar. The
enuntiabile, that the positum and the propositum have the same truth-value

29Neither TE nor OP considers the yet more pathological enuntiable that the
propositum is false which can neither be posited nor proposed without circularity.
370 CHRISTOPHER J. MARTIN

can be maintained, according to TEF, if the casus falls out in one way but
not if it falls out in another. Ifthe first propositum is true then the positio is
perfectly safe. This, incidentally, shows us, I take it, how TEF would deal
with the truth-positing enuntiabiLe, that the positum is true. If we start off
by agreeing to maintain its truth, no inconsistency can result. On the other
hand, that the positum and the propositum have the same truth-value
cannot be posited if the first propositum is false. In that case, according to
TEF, the positio is destroyed (interimitur) by the proposal.

But what exactly is it that is destroyed? Throughout TEF the

distinction between sentences and enuntiabilia, or dicta, is carefully
observed with the latter picked by means of accusative plus infinitive or
quod constructions. The aim in positio is to answer as if the posited
enuntiabile were true. Unlike pseudo-Sherwood, TEF allows that the
positum may be proposed in a different form of words to that in which it
was accepted. If the positio is destroyed it is because the positum can no
longer be treated as true and this because it can be neither frue nor false.
But in that case the positum sentence can no longer express an enuntiabiLe
since all there is to being an enuntiabile is being either true or false. Putting
this in more modem terms the problem is that of responding to utterances
of sentences in situations in which they fail to express propositions.

TEF does not say how such proposals should be greeted but I think
that it would require that the respondent say 'nugaris' or 'nil dicis', "you
aren't saying anything". This at any rate is what we are told to say to the
propositum 'mulier albus est' in apositio with positum 'quod hec vox
mulier sit masculinis generis'. The point is that the metalinguistic positum
says nothing about the gender of 'albus' and we should not give a reply to
the propositum which presupposes its congruity.3o

If this is the appropriate response for an insoluble positum according

to TE, then it agrees with the Obligationes Parisiensis. Both works
maintain that 'that the positum is false' and anything convertible with it is
unacceptable as a positum but the sophisms that appear in 0 P are very
much more complicated than those presented in TEF.

For example, prior to Scotus' criticism of the account of time and

modality presupposed by the theory of positio a standard rule was what we
may call "the Present Instant Principle" (PIP): "IfA is the name of the
present instant, then, if it is proposed in a false positio, 'A exists' must be
denied." OP thus sophistically proposes that it is possible that I should
concede that A does not exist. This indeed seems eminently plausible since

30TEF, p.110 is here dealing with the problem of how one should answer a quaestio
disciplinalis such as 'quid?' or 'quare?' in apositio. It describes the respondens under
such circumstances as being in an apparentia - something that seems like the truth but
isn't so. This unusual word is used as an alternative to 'fantasia' by the Summa
Sophisticorum Elenchorum and in the Fallacia Parvipontanae in giving the division of
sophisms according to "Alexander". Apparently alluding to Top., VIII, 156a7 but
giving Aristotle's advice to the respondent rather than the opponent TEF maintains that
in such cases the respondent should refuse to answer the question' quia debet celari rei
veritas in falsa positione'. 'Nugaris' as used by TE does not seem to be connected with
refutation by nugatio as described in Soph. El. 3, 165b15.
 OBUGATIONS AND UARS 371

it is just what I must do in a false positio. Thus, unsuspecting, I allow it to

be posited. Though the dialectic is a little complicated my opponent can
relatively quickly convict me of an improper response no matter how I
reply to his proposals. I attach the details in an Appendix. What has gone
wrong of course is that I have admitted a proposition simply ungrounded
with respect to positio. Here is the proof: "If you should concede that A
does not exist, you should concede a falsehood, and if you should concede
a falsehood, then the positum is false." Conversely: "If the positum is
false, that A exists is to be denied, <and> from this that you should
concede that A does not exist."31

OP offers the following advice: 32

" ... one should consider whether in the time of the positio, the
positum might become convertible with 'that a falsehood is posited'.
And then it should be said that the positum departs <cedit> at the
time at which it becomes convertible with that a falsehood is posited.
Whence if after that time 'Stop the time' should be said, or
something be proposed, one should say 'nugaris', and respond to
the proposita as if there were no positum. For indeed it has then
departed <cessit>. If it does not become convertible with that a
falsehood is posited in the time of the positio, one should consider
whether an insoluble arises from it and if it does then the casus is
terminated by you."

In the exercises which follow 0 P sophistically proposes that it is

possible that you are a man is repugnant to the positum and what has been
conceded. 33 Again it certainly looks possible but if we accept it as our
positum we are by a rather complicated series of proposita led to
embarrassment. The solution is to notice that at a certain point the positum
became convertible with that a falsehood is posited and thereupon departed
<cessit>. At that point we should have replied 'You are talking nonsense'
<'nugaris'>. Again I attach the details in an Appendix.

310p 33.19-23.
321 have made some obviously necessary corrections to the text: "Praeterea, quodam
enuntiabile potest poni et permanere positum et <non> cadere a positione, ut Sortes est
albus; quodam potest poni et permanere positum sed [non] potest cadere a positione, ut
falsum positum <!concedi?> <for de Rijk's cadere> et quadlibet convertibile cum iIIo.
Unde quotiens intendit aliquis ponere, diligenter considerandum est an sit convertibile
cum falsum poni an non. Et si sic non recipiatur. Vel si fiat positione facta, simile vel
convertibile, similiter non recipiatur. Si vero non sit nee fiat, recipiatur et tum
consideretur an in tempore positionis propter propositum aliquod vel aliud fiat
convertibile cum falsum poni. Et si sic, dicendum quod cadit in eodem tempore
positum in quo fiat convertibile cum falsum poni. Quare si post iIlud tempus dicat:
cedat tempus vel proponat aliqua, dicendum <est> nugaris et ad proposita
respondendum est ac si positum non fuerit. lam enim cessit. Si vero non fiat
convertibile cum falsum poni in tempore positionis, considerandum est an ex ea fit
insolubile et si sic, casus tibi terminatur. Solet tamen nul/um tale a quibusdam recipi
1uod non potest cadere, quia fit insolubile ex casu." OP 36.16-32.
3 OP, p. 40, correcting de Rijk's 'respondere' with'repugnare'.
372 CHRISTOPHER J. MARTIN

What has happened, of course, is that the positum sentence has

become ungrounded with respect to positio. The opponent may utter it but
if he does, he succeeds in saying nothing at all.

Exactly the same may happen in depositio, that form of obligatio in

which one has to respond to proposita as if the enuntiabile accepted as the
depositum were false. The skill in depositio is essentially that of applying
modus tollens, of grasping at a glance the antecedents of the depositum.
The hazard is the same as in positio and it will only be avoided if one is
able to detect propositions that are or will become ungrounded with respect
to depositio by becoming convertible with the enuntiabile, that the
depositum is false. OP's description of what happens here is, however,
interestingly different from that which it gives for positio.

In depositio just as in positio the Parvipontanian machine is a threat to

those who take it seriously. They must not accept as the depositum a
disjunction of two contradictory opposites:34

"By those upholding the rule that from an impossibility anything

follows a necessary <enuntiabile> is not to be accepted in depositio.
From this it is clear that a disjunction consisting of contradictory
opposites is not deponible. Thus this has to be disambiguated: it is
deposited that there will be a naval battle or that there will not. Ifit is
intended to deposit the disjoined dictum, then the depositio should be
cassed .... If the depositum is disjoined, the depositio is to be
admitted ... Those which have for one of the contradictory opposites
a necessary <enuntiabile> and for the other an impossible
<enuntiabile> should be accepted in deposition only for the
impossible, unless an insoluble should prevent this. On account of
this that a falsehood is deposited or not deposited is not deponible,
neither for the whole nor for the parts because that a falsehood is not
deposited is per se necessary, that a falsehood is deposited is
impossible and is as well insoluble if it is deposited."

The proof of its insolubility is the standard one. The consequence is that: 35

34"Praeterea, sustenentibus hanc regulam quod ex impossibile sequatur quidUbet,

necessarium non est recipiendum in depositione. Et ex eodem patet quod disiuncta
COllStans ex contradictoria oppositis non est deponibilis. Unde distinguenda est hec:
deponatur navale bellum fore vel non fore. Et si dictum disiuncte intendat deponere,
cassetur depositio ... Si depositum disiungitur, admittenda est depositio ... Que autem
habent unum contradictorie oppositorum necessarium, alterum impossibile, pro
impossibile tantum recipiatur in depositione, nisi impediat insolubile. Et propter hoc
hec disiuncta: falsum deponi vel non deponi, non est deponibilis nec pro toto nec pro
partibus, quia hec per se est necessaria: falsum non deponi, hec autem est impossibilis:
falsum deponi; et cum hoc est insolubile si deponatur." OP 48.17-32.
35"Cassetur ergo eius depositio et omnis convertibilis cum ipsa, vel ut nunc vel
simpliciter. Ut nunc, ut deponatur depositum esse dissimile te sedere et contingit te
sedere ill rei veritate. Simpliciter, ut deponatur depositum esse dissimile Deum esse
<for de Rijk's dictum esse>. Praeterea, aUe circumstantie que causant insolubile in
falsa positione cassande sunt etiam in depositione." OP 49.7-12.
 OBLIGATIONS AND LIARS 373

"Its deposition is cassed as is that of everything convertible with it

either ut nunc or simpliciter. Other circumstances which cause
insolubles in false positio are also to be cassed in depositio."

In Scotland 'cassed' is still available as a translation of cassatur. It is used

in the law in exactly the right sense of cancellation or annulment. The
solution to the problem of propositions ungrounded with respect to positio
or depositio found in OP and TE is thus cassatio, apparently the earliest of
the solutions to the Liar. The word is certainly cassatio rather than
quassatio. 36 What is proposed is cancellation and annulment not shaking
and breaking. The terminology of obligationes indeed reeks of the law and
it would certainly be worthwhile to investigate late twelfth century thinking
about legal disputation as a source for the structure adopted in regimenting
the logical procedure.

It seems to me that it makes much more sense to speak of cancellation

or annulment with respect to a positio than it does apart from that context. I
would thus suggest not only that twelfth century logicians first became
aware of insolubles via obligations but also that they developed their first
solution in terms of them. To see what that solution really amounts to,
however, we must now turn finally to what seems to be the oldest
surviving treatise entirely devoted to insolubility. This is the Insolubilia
Monacensis found in the same manuscript as the Tractati Emmerani.

2.2 Insolubilia, Cassatio and Truth-value Gaps

Like both TE and OP, 1M is careful to maintain a clear distinction

between the propositional content expressed in the assertive utterance of an
indicative sentence, the enuntiabile or dictum, and the sentence itself, the
propositio. Strictly speaking, according to 1M, the enuntiabile, that what I
assert is false is not insoluble. It would only be so if no appropriate answer
could be given to someone proposing it. 1M does not explictly appeal to
positio here but clearly needs it to guarantee that this is the only thing that I
sayY In such a positio, that what I say is false is ungrounded. However,
far from there being no suitable response to someone who proposes only
this, 1M claims that it and any questions that may be asked about it can be
answered. The questions it suggests are "is what I assert false?", "is what I
assert true ?", "do I assert something ?". We are not told the answers but
we will see that they must be "no", "no" and "no".

36As it is called in the treatise on Insolubilia in B.N. Lat 11.412 excerpted by De Rijk
in "Some Notes". 'Cassatio' is a very uncommon term. 'Cassare' is more common and
interestingly is used by Abaelard in just the right sense to describe the way in which a
"formally" good consecution "can in no way be cancelled <cassari>" by a uniform
substitution of terms: Dialectica, III. I, p. 255.31-34.
37The description of insolubilia in the tract in B.N. Lat. 11.412 is much more
explicitly formulated in terms of positio: "Et ut melius pateat quod quaerandum est,
faciamus deductiones ad hanc propositionem: ego dico falsum. Si dicitur quod verum
est. sequitur hoc est verum et ego dico hoc. ergo ego dico verum, ergo ego non dico
falsum, ergo hec est falsa: ego dico falsum; et eoncessisti <quod> verum, ergo male. Si
dicatur quodfalsum. hoc est falsum. et ego dieo hoc, ergo ego dico falsum, ergo hee est
vera: ego dieo falsum, et eoncessisti <quod> falsum ergo male." Op. cit., p. 94.
374 CHRISTOPHER J. MARTIN

Like later treatments of insolubilia 1M appeals to the Aristotelian

distinction between being X without qualification and being X in some
qualified way. The later treatments will have the Liar sentence false
simpliciter but true secundum quid. 1M on the other hand has it insoluble
simpliciter but soluble secundum quid. What is meant, I think, is that
neither 'verum est' nor 'falsum est' is an appropriate response but
nevertheless there is one. This I suppose is arguably evidence that the
solutions which apply to Aristotle's distinction to truth and falsity had not
been developed when 1M was written.

1M goes on to claim that enuntiables insoluble secundum quid are all

characterized by circularity: "What is insoluble is a circular and necessary
deduction to both sides of a contradiction. "38 It follows from its account of
the significant use of sentences that there are three ways in which an
enuntiabile might force itself to go round in circles. 1M distinguishes
between enuntiabilia, the physical and mental acts involved in presenting
enuntiabilia to ourselves and others, and the results of those acts, the token
physical, or mental, embodiment of the enuntiabiles. For example, in an
act of assertion I assert the enuntiabile, that snow is white by uttering a
token of the sentence 'Snow is white' assertively.

That snow is white is of course perfectly straight. For there to be the

possibility of a circular deduction the enuntiabile must in some way refer to
one of the three elements required to make it manifest - the enuntiabile
itself, the act in which it is made manifest, and the sentence which
manifests it. For example in asserting that I assert a falsehood I refer, or
try to refer, to the act of assertion and so long as the only thing that I utter
is the corresponding sentence the circular deduction of both parts of a
contradiction is immediate. 1M proposes in effect that an enuntiabile of the
form that what I X is false is simply insoluble with respect to the act of
Xing. Thus that what I concede is false is simply insoluble with respect to
conceding and that what I think is false with respect to the mental act of
thinking.

According to 1M, the solution secundum quid to the problem of

responding to speech acts of the form 'what I X is false' is to realise that
with such an utterance I cannot X on pain of circularity - with it posited, of
course, that this is the only thing that I say out loud, if X is a physical
speech act, or say to myself, if it is a mental speech act. The appropriate
response is thus to say 'You aren't Xing anything'. In general we must
cass the act of Xing just as we cass the act of positing when the proposed
positum is ungrounded with respect to positio.

While cassing seemed to make good sense in the context of a positio it

is rather less clear what it amounts to here. 1M goes on to offer an

38"/lisolubile est ad utrumque partem contradictionis circularis et necessarie deductio."

1M, p. 105.12-13.
 OBLIGATIONS AND LIARS 375

argument in favour of the objection that in asserting that what I assert is

false I cannot fail to assert something: 39

"To assert something is none other than first to judge and then to
utter. But the judgement is not to be cassed, since from it no
embarrassment follows. Posit that someone judges but does not utter
and one may with impunity maintain that it <i.e. the enuntiabile> is
false. Similarly the utterance is not to be cassed nor is it usual to cass
it. And so since neither the judgement nor the utterance is to be
cas sed and these two are what it is to assert, the assertion is not to be
cassed."

1M's solution parallels the account of the negation of implications found in

the treatise De Implicationibus in the same manuscript. Neither assertion
nor utterance alone is to be cassed but rather the one in respect of the other.
That is to say, I take it, that when I respond to your words with 'nil dicis'
my claim is that you failed to achieve the combination of judgement and
utterance necessary for assertion.

As it stands the solution is entirely schematic and ad hoc - assertion

succeeds except when it cannot. Furthermore the enuntiabile itself still
seems to be insoluble secundum quid, since surely if I fail to assert
anything with an assertive utterance of 'what I assert is false', then the
enuntiabile, that what I assert is false is false etc. Whether 1M has an
answer will tum upon what it supposes an enuntiabile to be. If it can show
that somehow there is no enuntiabile in this case, then perhaps it does have
an answer. Here we need to know more about 1M's theory of meaning.
The question of the nature of enuntiabilia is raised directly but it is hard to
see whether the answer helps with our problem: 40

"The appellation that a man runs stands for an enuntiabile. It is

asked what that is, whether a thing or an expression or an
understanding. In so far as we are dealing with insolubilia, we
maintain that an enuntiabile is an understanding or a conjunction of
things or understandings."

Later, addressing the problem of the insolubility that arises when an

enuntiabile is in some way about the enuntiabile itself, 1M develops this
point a little in a version of Grelling's paradox almost identical with that
presented recently by Tyler Burge.41 Unlike Burge, however, 1M is happy
to solve the paradox by forcing a gap between truth and falsity even here: 42

39"Sed contra. Dicere nil aliud est quam prius asserere et postea proferre. Sed assertio
non est cassanda, cum ex ipsa non sequatur inconveniens. Data enim quod asseratur et
non proferatur, inpune potest sustineri quod sit falsum. Similiter prolatio non est
cassanda nec solet cassari. Et ita cum assertio non sit cassanda nec prolatio et ista duo
sum dicere, dictio non est cassanda." 1M, p. 106.3-8.
40"Et cum tractandum sit de enumiabilibus insolubilibus, potest queri quid sit
enuntiabile. Verbi gratia, hec appellatio hominem currere supponit enuntiabile.
Queritur quid sit id, sive res sive voces, sive intellectus. Secundum quod tractandum est
de insolubilibus, sustineamus quod enuntiabile sit imellectus sive coniunctio rerum
sive intellectuum." 1M, p. 106.18-22.
41Compare: "Suppose I conduct you into a room in which the open sentence type 'It is
not true of itself' is written on the blackboard. Pointing at the expression, I present the
376 CHRISTOPHER 1. MARTIN

" ... the third species of insoluble is that which comes from the
existence of the insoluble. And for this reason the enuntiabile should
be cassed. But in cassing the enuntiabile we do not cass the
substantial but rather the accidental. Granted that an enuntiabile is a
thing or an understanding, nevertheless it is not a thing or an
understanding which is cassed but rather a conjunction of things or
understandings. Since the conjunction of things is an enuntiabile and
that is accidental, the enuntiabile may indeed be cassed."

To cass an enuntiabile it seems that we must say that it is not an

enuntiabile (propter hoc debet dici quod non est enuntiabile).43 Is there a
confusion here between the enuntiabile as an understanding and what is
enuntiabile, or assertable, a combination of things or understandings and
presumably of things and understandings? Perhaps the intended answer is
a development of the familiar notion of semantic congruity. An enuntiabile
is an understanding consisting of a combination of understandings of
things or, presumably, of understandings. A combination which cannot
occur cannot be understood and in consequence cannot constitute an
enuntiabile.

Whatever the details of their epistemology and semantics the cassers

have in effect provided a truth-value gap solution to the problem of
insoluble sentences. Assertive utterances may be either true or false or say
nothing at all. Let me conclude by considering briefly how well they
understood the demands of this kind of solution.

An account the Liar should do its best not to offend too many of our
intuitions about meaningfulness and truth. While the claim that if! say only
'What I say is false', I really say nothing at all perhaps conflicts with a
general intuition about the meaningfulness of well-formed sentences, it
seems to accord well with our intuitions about this particular instance. If
the author of the treatise on insolubles in B.N. Lat. 11.412 is to be

following reasoning: Let us consider it as an argument for its own variable or pronoun.
Suppose it is true of itself. Then since it is the negation of the self-predication of the
notion of being true of, it is not true of itself. Now suppose it is not true of itself.
Then since it is the negation of the self predication of the notion of being true of, it is
true of itself." Tyler Burge, "Semantical Paradox", Journal of Philosophy 76, 1979,
pp. 169-98; with: "Dato enim quod hoc esset enuntiabile: aliquid non es<se> verum
pro se, inde sequitur contradictio sic: hoc aut est verum pro se aut non est verum pro
se: si est verum pro se [non] ergo ei convenit suus predicatus; suus predicatus est non
es<se> verum pro se; ergo ei convenit non esse verum pro se; ergo non est verum pro
se; et dictum est quod verum; si non est verum pro se, sed est verum pro omni eo cui
convenit suus predicatus; sed suus predicatus convenit ei; ergo est verum pro se; et
dictum est quod non est verum pro se. Propter hoc debet dici quod non est enuntiabile."
1M. p. 115.11.17.
42"Sequitur de tertia species de insolubilium que provenit ex essentia insolubilibus. Et
propter hoc debet cassari enuntiabile. Sed cassando enuntiabile non cassatur substantiale
sed accidentale. Licet enim enuntiabile sit res vel intellectus, tamennon cassatur res vel
intellectus sed coniunctio rerum sive intellectuum. Cum enim coniunctio rerum sit
elluntiabile et illa sit accidentalis, enuntiabile bene potest cassari" - continued in n. 41.
431M, p. 115.17.
 OBLIGATIONS AND LIARS 377

believed it also agrees with the intuitions of the rusticus whose opinion is
canvassed there.

Our intuitions about a sentence like 'This sentence is in English' are

quite different and it would be highly offensive to them to propose that in
uttering it I am merely babbling. 1M does not deal with such sentences but
presumably would have to assign this one to the truths. More problematic
is the "truth-teller", the enuntiabile, that what I assert is true, which locates
itself exactly at the point at which our intuitions about the connection
between meaningfulness and truth-value become unsure. It is certainly not
paradoxical but if we assign it a truth-value we are immediately led back to
that assignment in an infinite circle. 1M agrees that the sentence is not
insoluble in the sense that a contradiction follows from it but against
indifference it offers an argument from symmetry. Suppose that of two
people hearing the sentence one remarks that it is false and the other that it
is true. They cannot both be right so they must both be wrong. Some grant
the argument, 1M tells us, and consign the enuntiabile to the insolubilia and
the utterance, presumably, to those with which nothing is said. That is
perhaps the best thing to do with it since there can be no matter of fact to
settle its truth-value. 1M does not agree. Prior to the the utterance of the
corresponding sentence the enuntiabile is surely false. When it is uttered it
either remains false or it becomes true. The argument presumably is that it
certainly doesn't become both true and false and unlike the Liar there is no
reason to suppose that it becomes neither. It seems that the enuntiable is
either true or false but we cannot say, or even know, which.

Though solutions like cassatio can reasonably claim to have

despatched the Liar they can never rest content in their victory. Rather they
must be perpetually on guard against an army of its fortified descendants
threatening to pour through the gap. One such descendant arrives to take
up the challenge in 1M. In an objection to the rule that if an insoluble is
disjunctively combined with a true enuntiabile the disjunction is true but if
with a false enuntiabile the whole is insoluble we are introduced to the Son
of the Liar: 44

"The enuntiabile, that I assert a falsehood or that I assert nothing

is either true or it is false. If it is true but it is not true on account of
the part: that I assert nothing, therefore it is true on account of this
part: that I assert a falsehood; therefore it is true that I assert a
falsehood, therefore I assert a falsehood, and nothing except this,
therefore this is false, and it was said to be true. If it is false,
therefore in virtue of both parts, therefore in virtue of this one: that I

44"Hoc enuntiabile me dicere falsum vel nil dicere aut est verum aut est falsum. Si
verum, sed non est verum ratione istius parte me nil dicere; ergo est verum ratione
istius partis me dicere falsum; ergo verum est me dicere falsum; ergo ego dico falsum;
et nil nisi hoc; ergo hoc est falsum; et dictum est quod verum; si falsum, ergo ratione
Ulriusque partis; ergo ratione istius me dicere falsum, ergo falsum est me dicere falsum,
ergo non dico falsum. et dico aliquid. ergo verum. et nil nisi hoc. ergo hoc est verum.
et dictum est quodfalsum. Et ita cum hoc sit quoddam verum me nil dicere secundum
commune iudicium respectu huius enuntiabilis. verum est me dicere falsum <vel nil
dicere>. Hic verum enuntiabile disiungitur ab insolubile et tamen ex responsione illius
sequitur contradictio." 1M, p. 111.9-19.
378 CHRISTOPHER J. MARTIN

assert a falsehood, therefore it is false that I assert a falsehood,

therefore I do not assert a falsehood, and I assert something,
therefore a truth, and nothing but this, therefore it is true and it was
said to be false."

But what about the possibility that I assert nothing with an utterance of
the sentence? 1M notes that that is just what is usually said but seems to tire
of difficulty here and does not pursue the problem any further. It follows,
it claims, that if I assert nothing, then the disjunction that I assert is true.
But then, of course, if the disjunction is true, it is not true and if it is not
true, it is true. All 1M has to say is that its rule was not intended to cover
disjunctions like this in which the same act appears in both disjuncts.
Though it runs through the proof, it seems not to notice that in agreeing
that with this disjunction I say nothing victory has been handed over to the
Son of the Liar.

What would an appropriate resolution be ? Recent gappy accounts of

the Liar have suggested a number but all seem to be objectionable in one
way or another. In any event the critics of the cassers did not trumpet the
return of the Liar. Rather, confiating assertion and utterance, they
contented themselves with the facile request that we read their lips for a
refutation of cassatio.45 This, of course would not have impressed the
cassers but exactly what the details of their semantical theory were we
cannot yet say. Until that problem is solved we will not be able fully to
evaluate their analysis of the Liar. Let us hope that it can be solved and
soon, since cassatio is arguably the most interesting and certainly the least
contrived of the mediaeval treatments of the test case for semantical
theories.

University qf Auckland

45See Spade "The Origins", pp. 307-8.

 OBUGATIONS AND UARS 379

Appendix
Two Sophisms from the Obligationes Parisiensis

Sophism 1: It is possible that you should concede that A does not exist (see
p.370).

Proof:

1. It is possible that a possible falsehood is posited.

2. If a possible falsehood is posited, then you should deny that A exists.
3. If you should deny that A exists, you should concede that A does not
exist.
4. :. It is possible that you should concede that A does not exist.

LET THE POSSIBILITY BE POSITED

Let Q be 'A exists'.

The Positum, Po, is 'you should concede not Q'.

LET THE TIME OF OBLIGATION CEASE

Po WAS EITHER TRUE OR FALSE.

o. Po WAS TRUE ~ Po WAS FALSE

1. Po was true Hypothesis.

2. Q was true Positio took place at A
3. Q was not repugnant to Po A truth is not repugnant to
a truth.
4. You should not have denied Q 3, rules for positio.
5. You should not have conceded not Q 4, rules for positio.
6. 'You should concede not Q' was false 5, quotation.
7. Po was false 6, substitution.

8. Po WAS FALSE ~ Po WAS TRUE

9. Po was a possible falsehood Hypothesis.

10. If a possible falsehood is posited you should PIP (see p. 370)
deny Q
II. You should have denied Q 9, IO
12. You should have conceded not Q 11, rule for positio.
13. 'You should conceded not Q' was true 12, quotation.
14. Po was true 13, substitution.
380 CHRISTOPHER J. MARTIN

Sophism 10: It is possible that you are a man is repugnant to the positum
and what has been conceded. (See p. 371)

LET IT BE POSITED.

Let Q be 'You are aman'.

Let C at step n be the conjunction of everything conceded before step n.
Let Pri be the ith propositum.

Let R(P,Q) be 'P is repugnant to Q'.

The positum, Po, is 'R(Q,(Po & C))'.

O. Q IS REPUGNANT TO Po & C Po: the positum.

I. Only this, indicating the sentence Prl.

displayed at 0, has been posited.

2. Not Prl Prl: true but denied

3. LET THE TIME OF OBLIGATION CEASE.
4. You have denied something true which is not
repugnant to the positum so you have
responded badly.

5. Prj Pr I: true and conceded

6. No truth has been denied Pr2: true and conceded
7. No falsehood has been conceded Pr3: true and conceded
8. LET THE TIME OF OBLIGATION CEASE.

9. Po WAS EITHER TRUE OR FALSE.

10. Po WAS TRUE --7 Po WAS FALSE Claim.

11. Po was true Hypothesis.

12. C was true 5,6,7.
13. Q is true You are a man.
14. Q was not repugnant to (Po & C) Truth is not repugnant to truth
15. 'Q is repugnant to (Po & C), was false 14, Quotation
16. Po was false. Substitution

17. Po WAS FALSE --7 Po WAS TRUE Claim

18. Po was false Hypothesis

19. (Po & C) --7 not Q (!)
20. Q is repugnant to Po & C Definition of repugnance
21. Po was true Substitution & Quotation

We have to prove (!) : (R(Q,(Po & C)) & C) --7 not Q

 OBUGATIONS AND UARS: APPENDIX 381

o. R(Q,(Po & e»
-7 not Q Claim
I. R(Q,(Po & e» Hypothesis
2. R(Q,(Po & e»is true Quotation
3. Only R(Q,(Po & C» is posited Prl
4. Po is true Substitution
s. Pri is true Prl, Quotation
6. Pr2 is true Pr2, Quotation
7. Pr3 is true Pr3, Quotation
8. Q is repugnant to truths 2,4, S, 6, 7.
9. Q is false Whatever is repugnant to truth
is false
10. notQ 9, Disquotation

SOLUTION:

"The positio is to be accepted and the proposita are to be conceded. But

when he says 'let the time of obligation cease' after the last concession (i.e.
8.) you should say 'NUGARIS' because <the positio> ceases in the time
of the last concession and in 'let the time etc.' <the positum> is convertible
with 'the positum is false. ,,,

PROOF:

o. Po IS FALSE H R(Q,(Po & e» Claim

I. R(Q,(Po & e» -7 Po is FALSE Claim

2. R(Q,(Po&e) Hypothesis
3. R(a truth, (Po & C». Q is true
4. Po is false or e is false. 2, Definition of repugnance,
De Morgan
S. Po is false. e is not false, Disjunctive Syll.

6. Po is FALSE -7 R(Q,(Po & e» Claim

7. Po is false Hypothesis
8. (S -7 not Q) 7, since the conditional is both
true and necessary and a
necessity follows from
anything, where S indicates the
positum and concessa

9. R(Q,S) 8, Definition of repugnance

10. S = (Po & e) As soon as Pr3 is conceded
11. R(Q,(Po & e» 9, 10, Substitution

Since S becomes equal to (Po & C) at the moment that Pr3 is conceded,
"up until then the positum may remain, but in the moment of the last
concession it falls."
 Hominis AsinuslAsinus Hominis

by Angel d'Ors

As is well known, medieval logicians gave a lot of attention to the

analysis of different sophisms in which there are propositions that include
complex terms one of whose parts is a term in an oblique case. The
propositions 'euiuslibet hominis asinus eurrit', 'ab omni homine
enuntiatum est verum I ab utroque istorum enuntiatum est verum', 'omnem
hominem videns currit', 'omnis homo videt omnem hominem', 'utrumque
oeu/um non habens potes videre', 'omne caput habens est unum solum
caput habens', which receive attention in all kinds of treatises l (Instantiae,2
Abstraetiones,3 Distinctiones,4 Sophismata,5 Syncategoremata,6
Summulae,7 in the commentaries to the Analytiea Priora 8 and to the
Sophistici Elenehi,9 etc.), seem to be the most famous representatives of

1Concerning the history and characteristic features of different logical literary genres,
see: H.A.G. Braakhuis, De 13de Eeuwse Tractaten over Syncategorematische Termen, 2
vols .• Meppel: Krips Repro. 1979; N. Kretzmann, "Syncategoremata, exponibilia,
sophismata", in The Cambridge History of Later Medieval Philosophy, ed. N.
Kretzmann, A. Kenny, J. Pinborg, Cambridge: Cambridge University Press 1982, pp.
211-45; A. de Libera, "La Litterature des 'Abstractiones' et la Tradition Logique
d'Oxford", in The Rise of British Logic, ed. P.O. Lewry, Papers in Mediaeval Studies
7, Toronto: Pontifical Institute of Mediaeval Studies 1985, pp. 63-114; L. M. De
Rijk. Some Earlier Parisian Tracts on Distinctiones Sophismatum, Artistarium 7,
Nijmegen: Ingenium 1988.
2y. Iwakuma, "Instantiae. A Study of Twelfth Century Technique of Argumentation
with an Edition of Ms. Paris BN Lat. 6674 f. 1-5", Cahiers de l'lnstitut du Moyen Age
Grec et Latin 38, 1981, pp. 40-1.
3See Alain de Libera, "Les 'Abstractiones' d'Herve Ie Sophiste (Hervaeus Sophista)",
Archives rfHistoire Doctrinale et Litteraire du Moyen Age 52, 1985, pp. 163-230
(especially, pp.168, 176, 187 (sophism 26),188 (sophism 29),190 (sophism 44».
4De Rijk, Some Earlier Parisian Tracts, especially, pp. 61-68 and 197-202.
5Albertus de Saxonia, Sophismata, Paris 1502, repro Georg Dims Verlag, Hildesheim-
New York, 1975, sophisms VI, VII, VIII and XLIII.
6J. R. O'Donnell, "The Syncategoremata of William of Sherwood", Mediaeval Studies
III, 1941, pp. 46-93, especially pp. 51-3.
7In chapters devoted to the "suppositio terminorum", or to the ''fallacia figurae
dictionis", the analysis of this kind of proposition is a commonplace. A. de Libera,
"Les Summulae Dialectices de Roger Bacon. I-II De Termino, De Enuntiatione",
Archives rfHistoire Doctrinale et Litteraire du Moyen Age 53, 1986, pp. 171-289,
especially, p. 267(434): De Suppositione; A. de Libera, "Les Summulae Dialectices de
Roger Bacon. III De argumentatione", Archives rfHistoire Doctrinale et Litteraire du
Moyen Age 54, 1987, pp. 171-272, especially, pp. 256-7 (585 and ff.): De fallacia
figura dictionis); Peter of Spain, Tractatus. called afterwards Summule Logicales, ed.
L.M. De Rijk, Assen: Van Gorcum 1972, especially pp. 222-4; Logica (Summa
Lamberti), ed. F. Alessio, Firenze: La Nuova Italia Editrice 1971, especially pp.I72-3.
8The question 'De syllogismis ex terminis obliquis' is one of the places in which the
analysis of this kind of proposition is ordinarily tackled. See R. Kilwardby (attributed
to Aegidius Romanus), In Libros Priorum Analyticorum Expositio, Venetiis 1516,
repro Frankfurt: Minerva G.M.B.H. 1968, especially f. 46rb; lohannis Buridani
Tractatus De Consequentiis, ed. H. Hubien, Louvain-Paris: Publications Universitaires
1976, especially, pp. 98-104; Albertus de Saxonia, PerutilisLogica, Venetiis, 1522,
repro Hildesheim-New York: George Dims Verlag 1974, especially ff. 29vb-30rb.

382
 HOMINIS ASINUS/ASINUS HOMINIS 383

this important kind of proposition. The study of such sophisms, of the

difficulties to which they give rise, and of the procedures invented by
medieval logicians to solve them are of great interest for contemporary
readers.

This is so for at least two reasons. First, because from the starting
point of such sophisms medieval logicians developed an important part of
what we could call the Medieval Logic of Three-Term Propositions,IO a
logic which paved the way for a General Logic capable of giving an
account of logical properties and of logical relationships between
propositions of any number of terms. Medieval logicians made a great
effort to develop the most elementary formal logical doctrines, of
Aristotelian origin (opposition, conversion, equipollence, syllogistic),
which were constituted in their origins as doctrines relative to Two-Term
Propositions, in order to make them applicable to propositions having any
number of terms. To achieve this, they had to overcome theoretical and
technical difficulties which are of interest.

The second reason is that in relation to these sophisms medieval

logicians also developed an important part of what we could call the
Medieval Logic of Relations. ll This Logic offers important differential
features as compared to the Contemporary Logic of Relations. Whereas the
latter proposes, defends and accustoms us to analyze relations as polyadic
predicates, the former attempted, over many centuries, to analyze them as
monadic predicates, and in that attempt it shaped concepts, developed
methods and encountered technical difficulties which - due to their
undoubted logical interest and unquestionable philosophical relevance - are
worthy of a detailed and systematic study which we can compare to the
concepts and methods of contemporary Mathematical Logic, as well as to
achieve a more adequate measure of the logical and philosophical relevance
of the differences that exist between the two approaches.

Each of these sophisms gives rise to peculiar difficulties by virtue of

the function (subject, predicate or both) which such a complex plays in the
proposition, by virtue of the nature (nominal or verbal) of the nominative
term, by virtue of the nature (singular or general) or of the number
(singular or plural) of the determination in the oblique case, by virtue of the
nature of the syncategoremata (quilibet, omnis, uterque, non, etc.) that rest
on one part or another of such propositions, etc. Due to both the number
and the logical complexity of the questions raised - questions, on the other
hand, which have been debated for over four centuries by many authors,
and therefore questions having a complex historyl2 - it is impossible to

9nte question 'De fallacia figura dictionis' is another of the places in which the analysis
of this kind of proposition is usually tackled.
WOther important parts of this Medieval Logic of Three-Term Propositions are those
referring to propositions that have, as parts, disjunct or conjunct terms, or to
reduplicative propositions.
11 Other important parts of this Medieval Logic of Relations are those referring to the
terms of the category of relation, or to the relative pronoums.
12Such questions, already quite clearly stated in the second half of the twelfth century,
were the object of attention until at least the beginning of the sixteenth century, when,
in my opinion, they attained their maximum development on the ocasion of the
treatment of the supposilio mixla, on the part of the logicians of John Malr's school.
384 ANGELD'ORS

treat them all here. For this reason, I will focus on the two most general
questions which have arisen around the two most elementary sophisms of
this class, the sophisms 'cuiuslibet hominis asinus currit' and 'ab omni
homine enuntiatum est verum I ab utroque istorum enuntiatum est verum':
one which refers to the logical form of such propositions, and the other
which refers to the logical value of the order of the parts (nominative term
and oblique term) of such complex terms. The examination of these
questions will give us the key for the adequate interpretation of an
interesting passage of Sherwood's Syncategoremata, in which this author
examines one of the difficulties which these sophisms give rise to.

1. On the Analysis of Logical Form

The proposition 'omnis homo est iustus' can be considered as a

paradigm case of Two-Term Propositions. Such propositions can be
analysed in two ways, depending on whether they are considered as a
whole, that is, as an enunciation, or as part of a syllogistic complex.
According to the former analysis - the one corresponding to the doctrine of
enunciation - 'homo' is said to be the subject of the enuntiation and
'iustus' is said to be its predicate. According to the latter analysis - the one
corresponding to the syllogistic doctrine - both 'homo' and 'iustus' are
said to be also, according to the particular case, the middle term or an
extreme term of the syllogistic complex. According to the medieval
doctrine of the enunciation (Aristotelian in origin), the subject term (in this
case 'homo'), which has to be in the nominative case, carries out at least
the following three functions: it is the term that determines what we are
talking about, the one that gives the suppositum for the predicate and the
enunciation, and, in virtue both of its nature (singular or common) and of
the nature of the syncategorema that rests on it, the term that gives the
enunciation its quantity. According to the syllogistic doctrine, the subject
term must also be the middle term or one of the extreme terms of the
syllogistic complex. According to this analysis, the proposition 'omnis
homo est iustus' (whose subject, 'homo', is a common term affected by a
universal syncategorema, 'omnis') is a universal proposition which speaks
of human beings and by which there is attributed to human beings - all of
them - the property of being just.

See E. J. Ashworth, "Multiple Quantification and the use of Special Quantifiers in

Early Sixteenth Century Logic", Notre Dame Journal of Formal Logic 19, 1978, pp.
599-613. repro in idem. Studies in Post-Medieval Semantics. London: Variorum
Reprints 1985. Discussions regarding such questions have at times gone beyond the
limits of purely logical discourse and have attained considerable cultural significance.
Thus. on the one hand. some of the theses defended in this regard were the object of the
Oxford condemnations of 1277, and. on the other hand. humanists such as Thomas
More and Vives found in these questions the occasion for their critique of scholastic
logicians and philosophers. See. regarding these matters, V. Munoz Delgado. "La
L6gica en las condenaciones de 1277". Cuadernos Salmantinos de Filosofla 4. 1978.
pp. 17-39; P.O. Lewry. "The Oxford Condemnations of 1277 in Grammar and Logic".
in English Logic and Semantics. From the End of the Twelfth Century to the Time of
Ockham alld Burleigh. ed. J. Pinborg. Artistarium. Supp1ementa I. Nijmegen:
Ingenium 1981. pp. 235-78.
 HOMINIS ASINUS/ASlNUS HOMINIS 385

The doctrine of the enunciation, which has just been presented, with
similar remarks on the predicate term, which is characterized by the
concurrence of multiple requirements and functions on each of the two
terms of the proposition, allows us to give an account of the elementary
logical doctrines (opposition, conversion, equipollence, syllogistic) within
the realm of Two-Term Propositions, but - as I have already pointed out -
it runs into serious difficulties when such doctrines are extended to the
realm of Three-Term Propositions, and, particularly, when extended to the
realm of propositions whose subject is a complex one of whose parts is a
term in an oblique case.

The proposition 'cuiuslibet hominis asinus currit' can be considered as

a paradigm case of such propositions. Prima facie, we could ignore the
peculiarities of this kind of proposition and analyze them as if they were
Two-Term Propositions. We could say, for example, that the whole
complex 'hominis asinus' is the subject of this proposition and that 'currit'
is its predicate. Such an analysis, however, immediately gives rise to
multiple conflicts of a different nature which deserve a detailed
examination.

a) 'Subiectum enuntiationis I propositionis'

In the first place, it gives rise to a conflict between its analysis as an
enunciative whole and its analysis as a part of a syllogistic complex, in so
far as it can happen that iUs not the whole complex term, but only one of
its parts - in particular, the oblique term - which carries out the role of
syllogistic middle term. Thus, if it is claimed that 'hominis asinus' is the
subject of such a proposition, and that' hominis' is the middle term of the
syllogism, then there arises a mismatch between the doctrine of the
enunciation and the syllogistic doctrine.

In order to solve this first conflict and to reestablish the articulation

between both doctrines, medieval logicians attempted at least two different
procedures whose common feature is the break-down of the complex into
its parts. In agreement with the first procedure, which is characterized by
the attempt to keep hold of the dyadic analysis of enunciations into subject
and predicate terms, it is claimed that the subject of the proposition
'cuiuslibet hominis asinus currit' is the oblique term 'hominis' on its own,
and that the other part of the complex, the nominative term, is on the side
of the predicate. The oblique term 'hominis', the middle term in the
syllogism, determines what is being spoken of in the proposition, and
gives to it the suppositum and the quantity. Such a solution, whose roots
were already found in the Ars Meliduna,13 or in the Sophismata Parisius
Determinata,14 finds an explicit expression in the works of Walter Burley,
Albert of Saxony, and Vincent Ferrer. ls

13See nn. 18 and 20.

14See n. 17.
IS"Si vero terminus obliquus praecedat terminum rectum, tunc nihil est subiectum,
loquendo de subiecto quo ad logicum, nisi terminus obliquus, et totum residuum se
tenet ex parte praedicati, ut patet in ista: 'cuiuslibet hominis asinus currit', et in
386 ANGELD'ORS

This solution immediately comes up against an objection of

Aristotelian origin: to affirm that the oblique term' hominis' is on its own
the subject of a proposition, is tantamount to admitting, against the
authority of Aristotle, that the subject of a proposition need not be in the
nominative case. 16 The defenders of such a solution reply to this objection
by making a distinction between enunciation and proposition and -
derivatively - between the 'subiectum enuntiationis', to which the
Aristotelian requirement applies, and the 'subiectum propositionis', to
which it does not. 17

consimilibus; hie nihil est subiectum nisi li 'hominis', et residuum se tenet ex parte
praedicati." W. Burley, De Puritate Artis Logicae. Tractatus Longior, p. 41.
"Conclusio ergo quod sicut in sophismate solum Iy 'hominis' distribuitur ita solum Iy
'homillis' subiicitur et non hoc aggregatum 'homillis asillus'. Sed diceres contra: in
praedicta propositiolle ponitur Iy 'asinus', ergo oportet quod sit pars praedicati, et hoc
non quia precedit copulam, vel pars subiecti, et sic habetur intentum. Respondetur quod
est pars predicati, et dico quod non est inconvelliens partem predicatiali quando precedere
copulam, immo aliquando totale predicatum precedit copulam, sicut in ista
propositiolle: 'homo allimal est'; tUIlC dicelldum est quod in predicta propositione solum
Iy 'hominis' sit subieetum." Albert of Saxony, Sophismata, sophisma viii. In the
Perutilis Logica, however, Albert seems to defend the opposite thesis: "4" suppositio.
Cum syllogieamus ex obliquis, non oportet quod maior aut minor extremitas sit
subiectum vel predicatum alicuius premisse, nee oportet quod medium syl/ogisticum
sit subiectum vel predicatum in anteeedente. Unde aliquando valet syllogismus ex
obliquis et medium syl/ogisticum Ilec est subiectum Ilec predicatum ill maiore, nee pars
subieeti nee pars predicati; similiter nee maior extremitas nec minor est subiectum vel
predieatum in eonclusiolle nee in premissis." (f. 30ra). "Si vero terminus obliquus
preeedat terminum rectum, tunc nihil est subiectum secundum logieum nisi terminus
obliquus, et totum residuum se tenet ex parte predicati." Vincent Ferrer, Tractatus De
Suppositionibus, Stuttgart - Bad Canstatt: Frommann-Holzboog 1977, p. 138.
16"Dubitatur hic primo an de obliquis syl/ogizari possit vel non. Videtur enim quod
non: eadem enim est materia propositionis et enunciation is, quia ipse eadem sunt in
subiecto, sed enunciationis principia materialia sunt scilicet 110m en et verbum, quare
proposition is similiter; sed nomen solum rectum est et verbum similiter, quare ex
obliquis non fiet propositio syllogistica sicut nec enunciatio." R. Kilwardby -
attributed to Aegidius Romanus, In Libros Priorum Analyticorum Expositio, f. 46rb.
17"Ad aliud dieendum quod si propositio et enuntiatio sint idem in substantia, differunt
tamen ill esselltiis sive ratione. Quod partes enulltiationis sunt nomen et verbumfiniti
modi exigit quod nomen ei supponens sit in reetitudine, hoc aUfem est subieetum; unde
in enuntiatione semper subiicitur rectus. Aliter autem est de propositione, quia verbum
non est pars ipsius proposition is, quia quicquid est pars ipsius propositionis est
primum vel medium vel postremum in sillogismo. Subieetum ergo et predicatum sunt
partes propositionis, sed subiectum et predicatum sUIIt nomen et verbum, et verbum
predicatur tam in rectitudine quam in obliquitate et tam <nomen> rectum quam nomen
obliquum potest subici in propositione, in enuntiatione autem non." Sophismata
Parisius Determinata, ff. 14va-b, quoted in L.M. De Rijk, "Each man's ass is not
everybody's ass. On an important item in 13th-century semantics", Historiographia
Linguistica, 7 1/2, 1980, p. 227. "Alii solvunt aliter, et dicunt quod prima est duplex
ex eo quod potest iudicari penes subiectum propositionis vel penes subiectum
enulltiationis. Subiectum vero propositionis appel/am subiectum illud <sub> quo sicut
sub medio potest aliquid sumi [ ...J Subiectum vero enumiationis nomen rectum
appellatur (ex solo enim verba finito et nomine recto componitur enumiatio, sicut dicit
Aristotiles)." Tractatus Florianus de solutionibus sophismatum, f. 42va. "Et
respondendum ad primum quod enunciationis secundum quod huius(?) et propositionis
secundum quod propositio non est necesse eadem esse principia materialia. Propositio
enim in ratione propositionis potest habere pro subiecto quod enunciatio secundum
quod huius(?) habet pro determinatione subiecti [...J Si eonsiderentur principia eius
secundum quod propositio est et secundum quod enunciatio est potest enim propositio
 HOMINIS ASINUS/ASINUS HOMINIS 387

This reply, however, is not completely satisfactory because it

presupposes the admission of a double doctrine of the enunciation and
therefore it does not achieve the full articulation of the doctrine of the
enunciation and the syllogistic doctrine. Hence medieval logicians tried a
second procedure which breaks with the dyadic analysis of propositions
and introduces a triadic analysis which recognizes a double subject in such
propositions - both when considered as an enunciative whole and when
considered as a part of a syllogistic complex. Such an analysis appears in
tum in different forms according to the remaining conflicts which arise
from such propositions.

b) 'Subiectum locutionis / attributionis / distributionis'

The second conflict arises between the different functions assigned to

the subject of the proposition. The parts of complexes made up of
nominative and oblique terms have different meanings and can both be
affected by syncategoremata of a varied nature. How then can we
determine what the proposition is talking about, which is the term that
gives the suppositum to the predicate and the enunciation, and what is the
quantity of the proposition?IS These three questions lie at the root of the
distinction between the 'subiectum locutionis' (what is being spoken of),
the 'subiectum attributionis' (that which gives the suppositum to the
predicate, to which the predicate is attributed), and the 'subiectum
distributionis' (that which gives the proposition its quantity), but none of
these questions seems to have an unequivocal answer. Hence we can find
in medieval logical texts different formulations of these distinctions, as
well as great variations in their usage. These variations become even
greater when these distinctions concur with the previously mentioned
distinction between the 'subiectum enuntiationis' and the 'subiectum
propositionis. '19

In order to determine the function which each of the parts of the

complex term (nominative and oblique) play in the proposition, medieval

habere subiectum secundum obliquum et secundum rectos. sed enunciatio solum

secundum rectum. et hoc est quod solet dici. et bene. quod duplex est subiectum.
scilicet. propositionis et enunciationis." R. Kilwardby, In Libros Priorum
Allalyticorum Expositio, f. 46rb.
ISThese questions were already clearly formulated in the Ars Meliduna: "Dehinc queritur
de quantitate earum quarum subiecti ex obliquo constant et recto." (f. 227rb)
"Notandum iterum quod non semper definiendum de quo fiat sermo indefinita
propositione vel particulari per id totum quod ponitur in subiecto." (f. 237va), quoted
in L.M. De Rijk, Logica Modernorum, vol. II, part I, Assen: Van Gorcum 1967, p.
327 and p. 368.
19"Sciendum tamen quod in talibus solebat distingui duplex subiectum. scilicet.
subiectum propositionis et subiectum locutionis. Subiectum propositionis est illud
quod est subiectum quo ad logicum. et est illud sub quo debet fieri sumptio in
syllogismo perfecto. Subiectum vero locutionis est subiectum quo ad grammoticum. et
est illud quod redit suppositum verbo. Ullde in ista 'cuiuslibet hominis asinus currit' li
'hominis' est subiectum propositionis et distributionis. sed iste terminus 'asinus' est
subiectum locutionis. tamen se tenet ex parte praedicati loquendo de praedicato
propositionis." W. Burley. De Puritate Artis Logicae. Tractatus Longior, St.
Bonaventure, N.Y.: Franciscan Institute 1955, p. 41.
388 ANGELD'ORS

logicians first focused their attention on the function which each of these
parts plays in the complex. Here is where the two questions which would
henceforth pervade the later history of the discussion arise. In the first
place, conflict arises between grammar and logic. From the grammatical
point of view, the nominative term always behaves as the determinable part
and the oblique term behaves as the determination, while from a logical
point of view it is also possible to consider the oblique term as the
determinable part and the nominative term as the determination. 20 Thus,
regarding the complex' Socratis asinus' from a logical point of view, we
can say both that this complex talks about donkeys, though only those
donkeys that belong to Socrates, and that it talks of Socrates' belongings,
though only about the donkeys among them. In some cases, the analyses
amount to the same thing because in both cases the predicate is attributed to
the donkeys and only to those that belong to Socrates; but in order to
determine what the proposition is talking about and what is the quantity of
the proposition, the difference does hold important consequences.

If, in agreement with the grammatical point of view, it is claimed that

the nominative term 'asinus' is the determinable part and the oblique term
'Socratis' is the determination, and - deriving from this - that by means of
such terms we are talking about donkeys, though only those that belong to
Socrates, then we will say that in the proposition 'Socratis asinus currit'
the term 'asinus' (or perhaps the whole complex 'Socratis asinus') is at
one and the same time 'subiectum locution is " 'subiectum attributionis' and
'subiectum distributionis', and that the proposition, therefore, is an
indefinite one. If, on the contrary, we stick to the alternative viewpoint, we
will say that in such a case the different functions are shared out between
the two terms, and that the nominative term carries out the function of
'subiectum attributionis', given that the predicate' curri!' is attributed, after
all, to donkeys, but that it is the oblique term that assumes the functions of
the 'subiectum locutionis' and of the 'subiectum distributionis'. We will
say that the proposition 'Socratis asinus currit' talks about Socrates'
belongings, and that it is a singular proposition. 21

The second question has to do with the syncategorema which affects

the oblique term. Is it a part of the complex, or does it rather determine the
oblique term from outside the complex? So, for example, regarding the
complex •cuiuslibet hominis asinus', must we say that it is talking about
the donkeys that are common property of all human beings, or rather that it
talks in general about the donkeys that have an owner, whoever that may
be? In the first case, in so far as the syncategorema is made part of the
determination of the nominative term, we will say that it has a

20This twofold possibility was already suggested in the Ars Meliduna: "Hoc etiam non
nisi ab his dicendum videtur qui dicere consueverunt substantivum terminum
supponere. verbo adiectivo circa eius rem determinante proprietatem." (f. 237va, quoted
in De Rijk. Logica Modernorum. vol. II, part I, p. 368). See also the text of R.
Kilwardby quoted in n. 17.
2 I"/deo litem de medio tollentes dicimus quod suppositio potestjieri tum per obliquum,
ut sit propositio singularis, tum per orationem ex ob/iquo constantem et recto, ut <sit
propositio> indejinita. Singulari autem agitur de a/iquo discrete, ut de Socrate, de quo
dicitur sui asinum esse album." Ars Meliduna, quoted in De Rijk, Logica Modernorum,
vol. II, part I, p. 369).
 HOMINIS ASINUS/ASINUS HOMINIS 389

categorematic function, or that it has a collective use, and therefore we will

tend to claim that the nominative term is what provides the 'subiectum
locutionis' and 'subiectum distributionis' to the proposition. In the second
case, on the contrary, we will say that the syncategorema has a purely
syncategorematic and distributive usage, and that the oblique term is what
gives the two subjects to the proposition.22

The different questions which we have just examined, both those

which have to do with the parts of complexes and those which have to do
with the parts of propositions or of syllogisms, usually appear mixed and
confounded with each other. Medieval logicians were conscious of the
relevance of such distinctions to the justification of different inferences and
towards the solution of different sophisms, but each author adopted one
criterion or another according to the nature of the difficulties that he had to
deal with. Each author therefore used one set of notions or another without
paying attention to their origin and strict sense. This mixing up is the origin
of many difficulties of interpretation, which become multiplied when the
problem of the value of the order of the parts, which I will now examine,
becomes complicated with these questions.

2. On the Logical Value of the order of the parts

Medieval discussion of the logical value of the order of the parts of the
compounds ofa nominative term and a oblique term constitutes an
interesting chapter in the complex history of the doctrine of the suppositio
terminorum, in which medieval logicians found a powerful instrument for
unifying the fundamental formal logical doctrines (opposition, conversion,
equipollence, syllogistic) capable of justifying the soundness of very
diverse forms of consequence, and of detecting and undoing a great
number of fallacies.

Within the realm of this doctrine, medieval logicians assigned a genus

and - where appropriate - a species of suppositio to each of the
categorematic terms which constitute a proposition. According to this
doctrine, the species of suppositio had by a certain categorematic term
obeys and is recognized by the function which such a term carries out in
the proposition, and the mode in which it is affected by the different
syncategorematic terms that occur in it. In view of the species of suppositio
which its terms have, medieval logicians assigned to propositions different
forms of resolutio (ascensus and descensus). These forms of resolutio
permit the manifestation of the conditions of truth of the different types of
proposition and - derivatively - the manifestation of the relations of
opposition and consequence between them.

One of the features that characterizes the peculiar mode in which

medieval logicians approached the formulation of this doctrine of the

22"Et dixi 'manens syncategorematica', quia si dictio syncategorematica fieret pars

extremi, quod contingit quando disponit partem extremi, tunc taUs dictio non tenetur
syncategorematice nee manet ut syncategorema, et tunc non habet virtutem confundendi
terminum communem mediate sequentem confuse tantum." W. Burley, De Puritate
Artis Logicae. Tractatus Longior, p. 21.
390 ANGELD'ORS

suppositio terminorum is the attention they gave to a highly normalized

language in which the function that the terms carried out in propositions
and the mode in which they are affected by the syncategoremata - on
which their species of suppositio depends - are narrowly linked to the
position which the terms occupy in the proposition. 23 Attention to this
highly normalized language allowed medieval logicians to draw up a very
simple formulation of the doctrine of suppositio which made possible a
certain mechanization of logical analysis. Any change in the position of a
term in a proposition carries along with it a change of its logical properties,
of its function, or of the mode in which it is affected by the
syncategorematic terms. The analysis of the suppositio of terms in
universal affirmative propositions ('omnis homo est albus') or particular
negative propositions ('quidam homo non est a/bus') provide us with a
paradigmatic example of this manner of proceeding. 24

By linking the function and properties of the categorematic terms to

their position in the proposition and by thus assigning logical relevance to
the order of the terms, medieval logicians, on the one hand, found a
powerful and simple instrument for the control and solution of multiple
fallacies. But, on the other, they came into conflict with characteristic
grammatical features of the Latin language, which in some manner put the
value of such a solution in question. The solution encountered contains,
then, at the same time, great naturalness and great artificiality. The fact is
that, sometimes, at the same time that a sentence has different senses -
giving rise to multiple confusions and fallacies - different sentences
express those very same different senses, and there is nothing more natural
- in order to avoid such confusions and such fallacies - than to proceed to
distribute the different senses and to assign a particular sense to each of
these sentences. And if such sentences differ among each other as per the
order of their parts, there is nothing more natural than to assign a logical

23The linkage of the functions of terms to their position in the proposition seem to find
its root in the framework of the doctrine of conversion, a doctrine which is of
extraordinary logical importance, both for its relevant role in the development of
syllogistic doctrine, and for the important role which, without doubt, it played in the
constitution and development of the doctrine of the suppositio.
24To the subject term 'homo' there corresponds, in the universal, a confused and
distributive suppositio, and, in the particular, a determinate suppositio, by virtue of the
different syncategoremata ('omnis 'r quidam') which in each case, and by virtue of the
fact of their concordance, affect it. To the predicate term 'albus' there corresponds, in the
universal, a suppositio con/usa tantum, and, in the particular, a confused and
distributive suppositio, but, ordinarily, in order to justify the attribution of this species
of suppositio, we no longer appeal, as we might have done, to its function as a
predicate affirmed or denied with respect to a subject taken in distributive or determinate
manner, but rather presupposing the order subject-copula-predicate to be the normal
order, we appeal to its position with respect to the syncategoremata, 'omnis' or 'non',
that precede it. In this manner, we attribute to the sign 'omnis' a distributive effect over
the term that immediately follows it and a confusive power over all the terms that
follow it in a mediate manner, and to the sign of negation 'non' a distributive power
over all the terms that follow it. Thus the way is open for new forms of propositions;
for example, in the proposition 'quidam homo albus non est' - which initially could
have been considered as a synonym with the negative particular - in so far as the
predicate 'albus' does not follow the sign of negation, we can no longer say that its
predicate is distributed, but rather determinate, and therefore the latter constitutes a new
type of proposition which is distinct from the four types A, E, I, 0 that are usually
considered.
 HOMINIS ASlNUS/ASlNUS HOMINIS 391

value to that order which makes possible the recognition of the sense
assigned to it. Such a distribution, however, is conventional, and in so far
as it refers to an already constituted language which in its own right does
not obey such rules, that is to say in so far as it is not a constituent
distribution within that language, it becomes contrary to the nature of the
language with regard to which such a convention is established.25

The conflict we have just pointed out between the conventions

established and the grammatical rules of the language to which such
conventions refer, is one of the conflicts around which medieval
discussion of the logical value of the order of the parts of the compounds
of nominative and oblique terms revolved. It is not, however, the only
conflict, nor the most important one to which we must attend in order to
understand such a discussion. If this is a conflict which arises - so to
speak - from below, from grammar, which declares the procedure to be
unjustified, there is a no less important conflict which is raised from
above, when we detect the limits which this solution runs against. The
assignation, to the syncategorema, of a virtue on the terms which follow it
immediately or mediately, and in a derived manner of a logical value to the
position of the terms in the proposition, seem to be a very effective
solution to the problems raised within the realm of Two-Term
Propositions, but serious difficulties are encountered when it is extended to
Three-Term Propositions. 26

Medieval discussion of the logical value of the order of the parts of the
compounds of a nominative term and an oblique term, in my opinion, must
be understood in the light of this double conflict. Such conflicts of a
diverse nature also have different logical relevance and for this reason they
must be distinguished and analyzed separately. Their confusion, as we
shall see, has given rise to interpretations which are not adequate to the
sense of such a discussion. In my opinion, the second of these conflicts is
the one having the greatest logical relevance, and the one which has more
clearly marked the history of this discussion. More than a defence of the

25"Ad hoc dicendum quod haec 'cuiuslibet hominis asinus est albus' duplex est, ex eo
quod suppositio huius termini 'asinus' potest multiplicari vel non. Non autem
ostenditur causa ex vi locutionis quare debeat simpliciter multiplicari, sed quod
multiplicetur hoc est ab intelllione loquelllis." Tractatus De Propietatibus Sermonum,
quoted in De Rijk, Logica Modernorum, vol. II, part II, p. 719. "Unde ad hoc est
magna diligentia adhibenda utrum rectus precedat obUquum vel e converso. Et
quandocumque ponitur aliqua propositio ab aliquo philosopho vel doctore, recto
precedellle obliquum vel e converso, semper trahenda est ad bonum sensum intellectum
[. ..J Et ideo quando aliqua propositio alicuius philosophi vel doctoris allegatur colllra
regulas scielllie logicalis, communiter solet distingui, et bene, quod talis propositio
dupliciter potest sumi, quia vel de virtute sermonis vel de bonitate intellectus, vel quod
idem est, potest dupUciter sumi, scilicet, vel in sensu quem /acit vel in sensu in quo
At." Vincent Perrer, Tractatus De Suppositionibus, p. 139.
6Por example, it is not enough to say that the difference between the propositions
'cuiuslibet hominis asinus est albus' and 'asinus cuiuslibet hominis est albus' is located
in the fact that 'asillus', in the first proposition, has a suppositio con/usa tantum,
because it mediately follows the universal syncategorema 'cuiuslibet', whereas in the
second proposition it has a determinate suppositio because it precedes it, since 'albus'
also follows, in both cases, the syncaregorema 'cuiuslibet', and yet the suppositio
cOil/usa talllum which in virtue of such a criterion would correspond to it, is not
assigned to it in both cases.
392 ANGELD'ORS

relevance of the order, what is sought is a solution of the difficulties that

arise when applying to Three-Term Propositions those criteria which were
elaborated within the realm of Two-Term Propositions.

In a synthetic manner, it could be said that there exist two moments or

stages in the history of this discussionP In the first, which is dominated
by the first conflict, what is sought is the consolidation of positional logical
criteria, that is, criteria aLa Polaca in the face of grammatical objections. 28
In the second, dominated by the second conflict, that is once the positional
criteria have been detected as being insufficient,29 what is sought is to
remedy such deficiencies. The solution is found in recourse to criteria of
grouping, that is criteria a La Russell, with parentheses or points.30 These
two moments or stages in the history of this discussion, the two conflicts
in which it is framed, are clearly reflected in the problems, notions and
examples which are brought to bear in each case.

In the first stage, attention is focused upon the parts of the complex,
upon the problems that arise due to ascent and descent under these parts,
and upon change in their order. Attention is preferentially focussed upon
examples which have a sole universal syncategorema, which rests upon the
determination in an oblique case. What is sought is a distribution of senses
allowing us to assign a sole sense to each proposition, and in order to do
this recourse is had to the distinction between the "subiectum Locutionis" or
the "subiectum distributionis" and the "subiectum attributionis", and
between "suppositio corifusa" and "suppositio determinata."31 When the
determination in the oblique case precedes the nominative term, it is said
that the syncategorema which rests on the oblique term is not a part of the
complex, that is to say that it is used syncategorematically; that the oblique

21Such moments or stages possess, without doubt, a clear chronological 'iQmponent,

but they could also have geographical components or be linked to certain schools.
28"Quando haec dictio 'omnis' preponitur orationi constanti ex recto et obliquo, tunc
comprehelldit omllia appel/ata illius orationis, ut 'omnis filiUS hominis currit'. Si vero
interponatur recto et obliquo ita quod precedat rectum, tunc comprehendit omnia
appel/ata illius recti, lion absolute tamen, sed respectu ullius appellati obliqui, ut cum
dicitur 'homillis omllis filius currit'. Si vero precedat obliquum, comprehelldit omllia
appellata illius obliqui, ut 'videns omnem hominem currit'." Tractatus Anagnini,
3uoted in De Rijk. Logica Modernorum, vol. II, part II, p. 299.
2 The limits detected are not limits that are intrinsic to the positional criteria in
themselves, but rather they derive from the particular and restricted usage which
medieval logicians made of them. Due to requirements of grammatical concordance,
they could not apply positional criteria with all the freedom that is required in order to
solve the problems they were faced with.
30"Et secundum istum intellectum debet fieri pllnctuatio et pausa inter ly 'pater' et ly
'patris filii', ita quod pUlictuetur sic 'omnis pater. patris filii est pater'." Albert of
Saxony, Sophismata, sophism VII. See also n. 22.
31"Solutio: Ad primum dicelldum quod verum est, sicut probatum, cum ille terminus
'asillus' respiciatur mediate ab hoc signo ulliversali 'cuiuslibet', unde confunditur
exiliter, et ita, secundum exigentiam locutionis, stat pro pluribus particularibus asinis
pro quibus locutio vera est. Ad hoc autem quod postea illfertur 'ergo asinus cuiuslibet
hominis currit', dicendum quod non valet <quia> cum ille termillus 'asillus' precedat
signum, nOli confullditur ab ipso, et indeterminate tenetur pro aliquo, pro locutio est
falsa. Et in tali argumentatiolle est fallacia figure dictiollis, scilicet, univocatio ex
mutata suppositione termini, quia ille termillus 'asinus' primo cOllfusam habet
suppositiollem, ill conclusiolle vero determillatam." Dialectica MOllacellsis, quoted in
De Rijk, Logica Modernorum, vol. II, part II, p. 614.
 HOMlNIS ASlNUS/ASlNUS HOMINIS 393

term is the' subiectum locutionis' or the 'subiectum distributionis', and that

the nominative term, confused, is the 'subiectum attributionis'. When, on
the contrary, the determination follows the nominative term, it is said that
such a syncategorema is part of the complex, that is to say, that it is used
categorematically or collectively and that it is the nominative term,
determinate, that carries out the functions of both the 'subiectum
locutionis' and the 'subiectum attributionis.'32

Problems soon crop up. The predicate is not confounded by the fact
that it is placed behind the universal syncategorema,33 the determinations
and syncategoremata in plural number seem to demand a collective
consideration no matter what their position with regard to the nominative
term may happen to be,34 the plurality of syncategoremata gives rise to
networks which seem to paralyze the positional criteria, and so on and so
on. There thus arises a second stage in which, on the contrary, attention is
focused upon the whole proposition and attention is given to both the
relationships between the parts of the complex and the relationships
between these and the other parts of the proposition. Now the most
interesting problems are those raised by the syllogistic mediation; the
examples in which both parts are affected by syncategoremata of universal
sense are those which receive the greatest attention. The distribution of
senses no longer is of interest; each proposition is assigned the sense
which makes it true; the distinction between confused and determinate
suppositio is no longer seen as the solution to all the problems, and the
problems of grouping dominate over the problems of order.

Obviously, the questions raised here require a more profound

analysis, both logical and historical, with the object of more clearly
establishing the relationship between the different problems that have
arisen and the different notions that have been introduced, the chronology
of their unfolding and the identity of their promoters. But this takes us
beyond the scope of this paper.

3. William of Sherwood (ca. 1200/1210-1266/1272)

The preceding analyses allow us to give a better interpretation of an

interesting passage in William of Sherwood's Syncategoremata,35 which
has already been treated by Kretzmann 36 and De Rijk,37 in which

32See n. 22.
33"Verbi gratia sic dicendo 'videns omnem hominem est animal', in ista propositione
non stat iste terminus 'animal' confuse tantum sed determinate; sequitur enim 'videns
onl1lem hominem est animal, ergo animal est videns omnem hominem', et econverso;
et in ista 'animal' supponit determinate, et ideo supponit determinate in alia." W.
Burley, De Puritate Artis Logicae. Tractatus Longior, p. 21.
34This is the reason why the sophisms 'ab omni homine enuntiatum est verum' and 'ab
utroque istorum enuntiatum est verum' cannot have the same solution. See Peter of
Spain, Tractatus, called afterwards Summule Logicales, ed. L.M. De Rijk, Assen: Van
Gorcum 1972, pp. 222-4.
35J.R. O'Donnell, "The Syncategoremata of William of Sherwood", Mediaeval Studies
3, 1941, pp. 46-93. Sherwood's texts will be quoted from this edition.
36N. Kretzmann, William of Sherwood's Treatise on Syncategorematic Words,
Minneapolis: University of Minnesota Press 1968, pp. 32-4.
394 ANGELD'ORS

Sherwood examines the two sophisms, 'cuiuslibet hominis asinus currit'

and 'ab omni homine enuntiatum est verum'. Sherwood's text is as
follows: 38

A 1.- Item sit quod Sortes dicat Deum esse et Plato dicat aliud verum,
et sic de aliis; item dicat quilibet te esse asinum.
2.- Deinde:
ab omni homine enuntiatum est verum,
sed quodcumque est ab omni homine enuntiatum est te esse asinum,
ergo te esse asinum est verum.
3.- a) Solutio: si sumatur prima secundum quod est vera incidit
fallaciafigurae dictionis in processu, eo quod commutatur 'Quale quid' in
'hoc a/iquid'.
i) Quia ly enuntiatum respicit hoc quod dico 'ab omni' ita quod pro
uno suorum suppositorum respicit unum suppositum eius quod est' ab
omni', et pro alio aliud, et sic deinceps, et sic pro multis respicit ipsum; et
sic, cum multa simul sint sicut unum commune et quale, stat ly enuntiatum
hic sicut •Quale quid' .
ii) Sed in minori respicit ly enuntiatum pro uno aliquo supposito
totam multitudinem eius quod est '<ab> omni', et sic respectu eius est
sicut 'hoc aliquid' .
4.- Eodem modo est si in minori diceretur:
sed nihil est enuntiatum ab omni homine nisi te esse asinum.

B 1.- Eodem modo hic est: sit quod quilibet homo habeat asinum unum
et currat, et Brunellus sit asinus communis et non currat.
2.- Deinde:
cuiuslibet hominis asinus currit,
sed quicquid est cuiuslibet hominis asinus est Brunellus,
ergo Brunellus currit.
3.- Similiter enim mutatur suppositio huius dictionis 'asinus'.

C 1.- Dicunt tamen quidam quod hae orationes sunt multiplices eo

quod potest iudicari locutio penes subiectum locutionis vel attributionis,
subiectum attributionis vocantes ipsum nominativum, subiectum
locutionis ipsum obliquum (quidam vero converso modo nominant), sed
hoc nihil est, quia cum ly cuiuslibet praecedit ly asinus habet potestatem
supra ipsum, et sic ab ipso iudicanda est locutio.
2.- Item quod sic vel sic iudicetur non est ex parte sermonis, sed ex
parte nostra tantum.

Both Kretzmann 39 and De Rijk40 interpret this text in the light of the
rule "non tenet processus a terminG postposito distributioni affirmativae ad
eundem praepositum."41 According to Kretzmann's interpretation - which
De Rijk accepts - in parts A and B of this text, Sherwood proposes, in the
light of this rule, his own solution to the difficulties raised by the
sophismata 'ab omni homine enuntiatum est verum' and 'cuiuslibet

37L.M. de Rijk. "Each man's ass is not everybody's ass" (see n. 17 above). pp. 221-30.
38 p. 52.
39 p. 33. n. 61.
40p.222.
41 p . 51.
 HOMINIS ASINUS/ASINUS HOMINIS 395

hominis asinus currit', while in part C he rejects an alternative solution

which does not take account of this rule. The key to Sherwood's solution,
according to this interpretation, would lie in a recourse to positional criteria
of interpretation, which allow us to assign a univocal sense to the
premisses of each of the arguments here examined. The alternative solution
would be rejected precisely because it admits the equivocity of such
premisses and ignores the positional criteria. 42 Kretzmann goes to the
extreme of characterizing Sherwood's logical doctrine as "directed toward
deciding logico-semantic questions on the basis of the structure of the
discourse itself as far as possible. "43 This characterization is also accepted
by De Rijk.44 In the light of this interpretation, it is not surprising that both
Kretzmann and De Rijk are unhappy with Sherwood's texts45 and that they
run up against difficulties when they come to interpret the final clause of
this text: "Item quod sic vel sic iudicetur non est ex parte sermonis, sed ex
parte nostra tantum", in which Sherwood seems to admit a double sense of
these expressions. 46

In my opinion, such an interpretation meets multiple and

insurmountable difficulties. It is true that in parts A and B of this text
Sherwood proposes his own solution to the difficulty, and in part C he
rejects an alternative solution, but the sense of the one and the other do not
seem to be those which Kretzmann and De Rijk attribute. In my opinion,
Sherwood does not reject, but rather admits, the plurality of senses of the
premisses of these arguments. Ifhe rejects the alternative solution it is not
because it presupposes that same multiplicity of senses, but rather because
the distinction which it uses to effect the separation of such senses is not
relevant to the question which is now being solved. In my opinion,
Sherwood is not here defending - as Kretzmann and De Rijk believe - the
semantic value of the positional criteria, but rather - on the contrary - he is
pointing out the limits of such criteria. Kretzmann and De Rijk do not seem
to have noticed the difference between the two conflicts which I have
previously referred to, nor that Sherwood is now primarily interested in the
second conflict. Sherwood's problem is not the problem of order, but
rather the problem derived from the existence of two syncategoremata
which annul the value of the positional criteria. But let us look at
Sherwood's text.

What is the difficulty which Sherwood is facing in this text? The

difficulty is obvious. Sherwood is faced here with two syllogisms which
appear to be logically correct, whose conclusions are obviously false, but
whose premisses - given the supposed case - appear to be true. This
constitutes, undoubtedly, a logical difficulty which demands a solution. In

42"Sherwood's fundamental objection is that the alternate analysis supposes that there
are two admissible readings of these expressions. His own position is, in effect, that
Rule [IV] makes only one reading admissible" (Kretzmann, op. cit., p. 34, n. 63); "For
that matter, when dealing with the above sophism Sherwood rejects the opponents'
analysis simply by appealing to the plain structure of the sentence (i. c., word order)
which admits of one interpretation only." (De Rijk, op. cit., p. 221).
43p. 34, n. 64.
44p. 221.
45Kretzmann, p. 33, n. 60; De Rijk, p. 222.
46Kretzmann, p. 34, n. 64; De Rijk, pp. 221-5.
396 ANGELD'ORS

the face of such a difficulty, only two solutions are logically admissible:
either we deny the soundness of the consequence, or else we deny the truth
of some of the premisses. What is Sherwood's proposed solution?

Despite what Kretzmann seems to believe,47 it is not the first premiss

that is at stake. Moreover, there is no sense in appealing to positional
criteria in order to undo the possible ambiguity of this first premiss (and to
assign it its sole and authentic sense) because the sense which such criteria
would assign to it would be, precisely, the one according to which the
premiss is true, and this can in no way serve as a solution to the difficulty.
On the other hand, Sherwood explicitly admits that the first premiss has
several senses ("si sumatur prima secundum quod est vera" [A, 3, a],
although he takes it for granted that the difficulty only arises with one of its
senses, namely that which the positional criteria would assign to it and
according to which the premiss is true.

Neither is the truth of the second premiss at stake here. The second
premiss is true according to either of its two possible senses, and therefore
to distinguish senses according to positional criteria would be useless.
Moreover, Sherwood explicitly affirms that the question of the order is
here completely irrelevant ("Eodem modo est si in minori diceretur: 'sed
nihil est enuntiatum ab omni homine nisi te esse asinum' " [A,4] ). It
would be of no use, therefore, contrary to what De Rijk suggests,48 to
propose an emendation of the text and to replace, in this second premiss,
'Cuillslibet hominis asinus' by 'asinus clliuslibet hominis'; the problem,
according to Sherwood, would remain the same. Neither is it possible,
therefore, against what both Kretzmann49 and De Rijk50 propose, to
reduce the problem that Sherwood is facing here to the problem of the
change of suppositio which derives from the change of position with
respect to a universal syncategorema.

What is then the problem which Sherwood faces? In my opinion, the

problem is clear: in these syllogisms there is not a defect of matter, but
rather a defect of form, .. defect of the syllogistic middle term, which
obstructs the consequence. And what is the root of this defect of form? The
defect does not lie, as Kretzmann51 and De Rijk52 would seem to think, in

47"The sophisma requires that 'each man's ass is running' be taken in two senses: [a] 'for
each man x there is an ass y such that x owns y and y is running'; [b] 'there is an ass y
such that each man owns it and it is running.' Sense [a] is the one supported by the
hypothesis, but sense [b] is the one that supports 'Brownie is running', which is false;
and in sense [a] 'ass' has merely confused supposition while in sense [b] it has
determinate supposition." (Kretzmann, p. 33, n. 61)
48"It is therefore tempting to assume that our MSS wrongly read 'cuiuslibet hominis
asinus' (rather than 'asinus cuiuslibet hominis') in the assumption." (De Rijk, p. 222)
49See n. 47.
50"Sherwood says that our sophism falls within the scope of the previous ones, because
the supposition of the word ass is changed in a similar way as there. Indeed, in the first
premiss of the syllogism ass (asinus) is taken in merely confused supposition, confuse
tantum, in the assumption in determinate supposition. [ ... ] Sherwood most certainly
must refer to the word order (see his Rule IV) as supporting the. determinate
supposition of ass in the assumption." (De Rijk, p. 222)
51"The point of Sherwood's solution could, it seems, have been made at least as readily
and perhaps more in keeping with his previous solutions if he had put it in terms of
 HOMlNlS ASlNUS/ASlNUS HOMlNlS 397

a change of suppositio conjusa tantum to a determinate suppositio of the

terms 'enuntiatum' / 'asinus', in virtue of a (non-existent) change of
position with respect to the syncategoremata 'ab omni' / 'cuiuslibet', but
rather it lies in the peculiar distribution of the terms 'enuntiatum' / 'asinus'
which derives from the presence of a second syncategorema also having
universal value. It is in the syncategoremata 'quodcumque' and 'quicquid',
respectively, and not in the order of the terms, that we find the root of the
difficulty. The problem lies in the fact that by virtue of the syncategoremata
'quodcumque' and 'quicquid', the terms 'enuntiatum' / 'asinus' - no
matter what their position with respect to the determination in the oblique
case - acquire a distributive suppositio which in virtue of the supposed
case is equivalent to a discrete or singular suppositio. This is precisely the
reason why Sherwood solves the difficulty in terms of the notions of
'quale quid' and' hoc aliquid', and not - as Kretzmann would prefer - in
terms of 'suppositio confusa tantum' and 'suppositio detenninata' .

In my opinion, part C of Sherwood's text is better understood in this

manner. Contrary to what De Rijk maintains,53 Sherwood does not reject
the alternative solution on the grounds that it admits a plurality of senses in
the premisses, but rather because it is of no use for solving the difficulty.
The clause 'hoc nihil est' is not a declaration of meaninglessness, but
rather a mere declaration of inappropriateness. What Sherwood defends is
not that the distinction between the subiectum locutionis and the subiectum
attribution is has no value at all, but rather what is here defended is that in
this case such a distinction does not help to solve the problem, which is to
say that it is no solution.

Universidad Complutense de Madrid

detenninate and merely confused supposition rather than in the roughly corresponding
tenns of hoc aliquid and quale quid." (Kretzmann, p. 33, n. 60)
52See n. 50.
53"Sherwood, however, rejects energetically this latter interpretation - he even calls it
nonsense (hoc nichil est) - giving in fact two reasons for his rejection: first, a reference
to the rule given: 'when the phrase "each man's" precedes the word "ass" (i. e. in
"cuiuslibet hominis asinus") the sign "each" has power over the nominative case "ass",
and so the phrase "cuiuslibet hominis asinus" is to be judged starting from the
distributive sign'; in other words, taking asinus, too, as confused by cuiuslibet, not
only man; and so any detenninate supposition of asillus is excluded, and, accordingly,
any ambiguity of the phrase 'cuiuslibet hominis asinus' as well." (De Rijk, p. 223)
 Solving the lnsolubles: hints from Ockham and Burley

by Claude Panaccio

In his approach to the "insolubilia" as elsewhere in his semantics,

Ockham firmly sticks to the Bivalence Principle: each apparently
paradoxical sentence is "solved" by him by being given one and only one
of the two traditional truth-values. The Liar's 'What I am now saying is
false' (uttered by a speaker who says nothing else) is reckoned to be false,
while 'What I am now saying is not true' is in the same circumstances
considered true. The well-known contradictions which threaten are avoided
by the use of a restriction rule which applies to such semantical terms as
'true' or 'false' and prohibits them - at least in some contexts - from
suppositing in a sentence s for s itself.

Thus, the reason why the sentence:

(l) What Socrates says is false

is false on Ockham's view when it is said by Socrates and is the only thing
Socrates says, is that the predicate 'false' is kept by the restriction rule
from suppositing in (1) for (1) itself. The sentence should be read as:

(1 a) What Socrates says is a falsehood different from the present

sentence,

which, in the case under consideration, is obviously false since Socrates

says nothing else. And it is easy to see that basically the same solution
applies - but with a different result - to what is nowadays often called the
Strengthened Liar:

(2) What Socrates says is not true,

which is by similar reasoning shown to be true.

Unfortunately, Ockham is rather sparing of explanations about this

solution. It occupies less than 3 out of the 900 pages of the Summa
Logicae in the St. Bonaventure edition; and the corresponding passage in
the Expositio super Libros Elenchorum is even sketchier.I In particular,
the Summa's text - on which I will concentrate here - raises two delicate
puzzles for the interpreter. First: what is the exact formulation of the
restriction rule Ockham has in mind? He does not say. Of course, he
carefully describes the sort of sentences that are liable to generate

ISee William of Ockham, Summa Logicae 111-3, ch. 46, in Opera Philosophica
(hereafter: OPh.) I, ed. P Boehner, O. Oal and S. F. Brown, St. Bonaventure, N.Y.:
The Franciscan Institute 1974, pp. 744-6; and Expositio super Libros Elenchorum II,
ed. F. del Punta, St. Bonaventure, N.Y.: The Franciscan Institute 1979, ch. 10 (OPh.
III, p. 268). There is also a very allusive reference to the illsolubilia in Summa Logicae
III-I. ch. 4 (OPh. I, p. 368).

398
 SOLVING THE INSOLUBLES: OCKHAM AND BURLEY 399

contradictions: they always include, he says, a (semantical) term such as

'true' or 'false'; if this term is negative (such as 'false'), the dangerous
sentences are the affirmative ones (such as (1»; and if the semantical term
is affirmative (such as 'true'), they are the negative ones (such as (2».2
But does the restriction rule apply to all such contexts and no other? And to
which terms exactly does it apply? Does it affect only the supposition of
the semantical predicate itself or is the non-semantical subject of such a
sentence also restricted? Ockham is content to apply his solution to
examples (1) and (2), and he provides no general answers to these
questions.

The second puzzle is that although on the second page of his three-
page chapter, Ockham very clearly attributes a definite truth-value to
sentences such as (1) and (2), he nevertheless writes on the third page:

"when Socrates begins to speak by uttering 'Socrates utters a

falsehood' and it is asked whether Socrates utters a truth or a
falsehood, it must be said that Socrates neither utters a truth nor a
falsehood ... "3

Doesn't that very much sound like a denial of the law of the excluded
middle, which apparently Ockham strongly believed in? How can both
things be reconciled in a coherent view?

I will attack these problems here from a basically philosophical point

of view. What I will be trying to do is argumentatively to find out the best
answers Ockham could logically have brought to these puzzles, given what
he says in the relevant chapter on the one hand and his general tenets in
semantics and logic on the other hand. The key-question will be: what is
the best Ockhamistic formulation of the restriction rule? An answer to that
automatically yields, as we shall see, a natural solution to the puzzle about
Ockham's treatment of Bivalence.

1. Burley's rule

It is to be presumed at the outset (in virtue of the respectable Principle

of Charity) that Ockham's best theory in these matters would be quite close
to the one he in fact had in mind. Our philosophical inquiry, then, can start
with a historical question: what restriction rule did Ockham have in mind
when he wrote the Summa (around 1325)? The St. Bonaventure critical
edition gives us a precious clue to this in a footnote about the chapter on
insolubles as a whole: "Ockham," the editors write, "probably had before
his eyes Walter Burley's tract (on insolubles)."4 And this is also the line

2See OPh. I, p. 745, II. 22-33.

30Ph. I, p. 746, 11. 50-52: ..... quando Sortes incipit sic loqui 'Sortes dicit faIsum', et
quaeritur 'aut Sortes dicit verum aut falsum', dicendum est quod Sortes neque dicit
verum neque falsum ...." All English translations from the Latin in the paper are mine.
40Ph. I, p. 744 n. 1.
400 CLAUDE PANACCIO

followed by Keith Simmons in a very interesting paper about Burley and

Ockham on insolubles published in 1987.5

Burley's Insolubilia is commonly supposed to have been written

around 1302.6 It is much longer and much more explicit than Ockham's
treatment. It distributes truth-values among insolubilia in exactly the same
way as Ockham does for sentences (1) and (2). And it too rests on a
restriction rule, which has the advantage of being stated explicitly.
Moreover it displays exactly the same tension as we found in Ockham
between the attribution of a definite truth-value ('true' or 'false') to each
insoluble sentence7 and the apparently anti-bivalentist claim that "if
somebody says that he is not uttering a truth, it must not be conceded that
he utters a truth nor that he utters a falsehood."8

All this strongly suggests that Ockham may have had in mind exactly
Burley's version of the restriction rule when he so confidently assured his
readers that the same principles he had just used in solving (1) and (2)
would enable the zealous and clever student (the 'studiosus' and
'ingeniosus') to solve all the insolubilia, a strong claim indeed!9

So here is Burley's formulation of the rule:

(RR) " ... a part never supposits for a whole of which it is a part
when, should the whole be posited in the place of the part, there
would occur a reflection of the same upon itself through a privative
determination," 10

which I propose to read as:

(RRI) A term T cannot supposit (personally) for an expression E

of which T is a part (e.g. a sentence) if the substitution of a
demonstrative designation of E for T in this context reveals a
reflection of E upon itself through a negative determination (such as
'false' or 'not true').

Burley is not very explicit about what exactly is a reflection of an

expression upon itself through a negative determination, but a study of his
examples leads to the following interpretation. First, let us stipulate that a

5Simmons, Keith, "On a medieval solution to the Liar paradox", History and
Philosophy o/Logic 8,1987, pp. 121-40.
6Burley's lnsolubilia (hereafter: lnsol.) has been edited by Marie-Louise Roure in "La
problematique des propositions insolubles au XIIIc siecle et au debut du XIve, suivie de
l'edition des traites de W, Shyreswood, W. Burleigh et Th. Bradwardine", Archives
d' histoire doctrinale et litteraire du Moyen Age 37, 1970, pp. 205-326; Burley's treatise
can be found on pp. 262-84.
7See lnsol., 3.05, p. 272: "Et insolubile affirmativum semper est falsum, et insolubile
negativum semper est verum ......
8ldem, 3.04, p. 272: " ... si aJiquis dicat se non dicere verum, nee est eoneedendum quod
dicit verum nee quod dicit falsum."
9See OPh. I, p. 746, II. 69-71: "Per praedicta potest studiosus respondere ad omnia
insolubilia (... ). Quod relinquo ingeniosis ......
lOll/sol., 3.03, p. 272: " ... nunquam supponit pars pro toto cuius est pars, quando,
posito toto loco partis, aecidit reflexio eiusdem supra se ipsum cum deterrninatione
privativa."
 SOLVING THE INSOLUBLES: OCKHAM AND BURLEY 401

sentence s reflects upon itself if and only if s authorizes a descent to one of

its own singulars in which s itself is singularly designated (example: if (1)
is unrestricted, it authorizes a descent to 'What Socrates says is this
falsehood' where the demonstrative designates (1) itself; therefore (1) is
self-reflexive). Secondly, a self-reflection is said to occur with a
determination of a certain sort (e.g. negative, affirmative, etc.) if and only
if the self-reflection occurs through a demonstrative which is itself
determined in the relevant way (by an accompanying word such as in 'this
falsehood' or by the copula itself) (example: (1) allows a descent to 'What
Socrates says is this falsehood: what Socrates says is false', in which the
demonstrative designating (1) itself is determined by the negative term
'falsehood', therefore the reflection of (1) upon itself will be said to be
accomplished through a negative determination).

Many of Ockham's requirements are satisfied by RRI thus

understood. Combined with his own theory of truth-conditions, RRI
yields exactly the truth-values he wants to attribute to the Liar sentences (1)
and (2) (e.g.: (1) turns out to be false because the term 'false' in it cannot,
because of RRl, supposit for sentence (1) itself and therefore there is
nothing for which the subject and the predicate of (1) both stand,
Ockham's necessary condition for the truth of a singular proposition -
namely, that there be a single thing for which they both stand - thus being
violated). 11

Moreover, RRI succeeds in explaining Ockham's and Burley's

apparent denials of the Principle of Bivalence. In virtue of RRl, the term
'false' in sentence (1) cannot stand for (1) itself. So not only Socrates, but
even we - the logicians - have to deny (1) (since it is false). And on the
other hand, we certainly do not want to admit:

(3) What Socrates says is true

because what he said was (1) and (1) is counted by us as false. So we have
to deny both (1) and (3). But the crux of the matter is that this is not a
denial of Bivalence at all because, in virtue of the restriction rule, (1) and
(3) are not strict contradictories. They are not the exact negations of one
another: there can exist a situation in which they would both be false.
Remember that if RRI is in force, (1) cannot be uttered without being
restrictively interpreted as:

(4) What Socrates says is a falsehood different from 'what

Socrates says is false' .

And this, of course, can be false while (3) is also false if what Socrates
says is precisely 'What Socrates says is false'. In this sole case, (1) (as
restricted by RRl) and (3) are both to be denied - and this is exactly what
Ockham says - but each sentence nevertheless receives one and only one
of the two old truth-values. That is, for each sentence s, one of these has to
be admitted: 'This is true' (pointing at s) or 'This is false' (pointing at s).

lIOn Ockham's truth-conditions for singular sentences, see Summa Logicae II, ch. 2
(OPh. I, pp. 249-54).
402 CLAUDE PANACC/o

The important thing here is that for Burley and Ockham, sentence (1) is
itself false and consequently must not be admitted. But this does not
prevent the truth of a different sentence saying that (1) is false; such as:
'This is false' (pointing at (1», which in effect is certainly not the same
sentence as 'What Socrates says is false'.

It can be checked that this explanation also fits Ockham's

developments about certain (apparently natural) inferences he wants to ban,
such as: 'Socrates utters this falsehood, therefore Socrates utters a
falsehood.' 12 In this case, for example, the inference is not formally good
because 'falsehood' in the consequent is prohibited by RRI from standing
for a certain sentence (namely: 'Socrates utters a falsehood'), which, one
never knows, could precisely tum out to be the very one that is designated
by the singular expression 'this falsehood' in the antecedent, in which case
the antecedent would be true and the consequent false. As Keith Simmons
notices, 'true' and 'false' in this approach behave like indexicals: their
extension varies with the context of utterance; and that is why some
apparently plausible inferences are not to be accepted asformally valid.I 3

So Burley's rule fits Ockham well and, as far as we can see up to

now, could very well have been exactly what he had in mind (and even, as
the editors suggest, before his eyes) when he came to write about
insolubles. But couldn't there be an even better fit? I mean: in view of what
the restriction rule is expected to accomplish and given Ockham's general
nominalism, can Burley's rule be amended and tightened up in order to be
both more satisfying theoretically and at least equally acceptable to
Ockham? That is what I will be exploring in the remainder of this paper.
Four amendments will thus successively be incorporated into RRl. I
should stress that in so doing, I do not want to defend the historical thesis
that Ockham had precisely these amendments in mind (or even that he had
any amendment in mind at all), but rather the philosophical thesis that the
revised restriction rule in its final form RR6 improves significantly on
Burley's rule itself as for its philosophical merits, while still being in
principle acceptable to Ockham (and even, as we shall see, more acceptable
than RRI itself). And I also want at least to suggest another (much more
important) philosophical thesis, namely that a restrictionist approach of the
Burleian sort revised in a more stringently nominalist spirit is still a serious
candidate for solving, as Ockham hoped, "all the insolubles".

2. Focus on semantical terms

First of all, RRI does not explicitly stipulate that it applies only where
a semantical term occurs. And it certainly should. All the insolubles Burley

120Ph. I, p. 746, II. 67-8.

I3See K. Simmons, op. cit., p. 134. Simmons draws a suggestive parallel between this
approach and some modern indexical solutions to the Liar, such as Charles Parsons'
and Tyler Burge's (see pp. 138-9). He also rightly rejects Roure's interpretation of the
Burley-Ockham solution as implying in Tarski's fashion a hierarchy of levels of
language and metalanguage. Nothing prevents the Burleian indexical truth-predicate and
falsehood-predicate from belonging to the same language as (some of) the sentences
they are applied to.
 SOLVING THE INSOLUBLES: OCKHAM AND BURLEY 403

wants to solve make essential use of terms like 'true', 'false', 'supposit
for' and so on. Moreover, even if such a clause was added to it, RRI
would still be ambiguous about which term exactly should be restricted. In
sentence (1), the rule can be indifferently applied to the subject ('what
Socrates says') or to the predicate ('false') without any change in the
resulting truth-value. As far as I can see, all this can be straightened out by
an explicit limitation of the restriction rule to affect only the supposition of
purely semantical terms. Let us, for the needs of this paper, define the
class of semantical terms as including only: 'true', 'false', 'true of', 'false
of', and any other expression which incorporates one of these, such as
'not-true', 'true-or-false', 'true-of-itself', and so on. The rule can then be
amended in the following way:

(RR2) A semantical term T cannot supposit (personally) for an

expression E of which T is a part if the substitution of a
demonstrative designation of E for T in this context reveals a
reflexion of E upon itself through a negative determination.

This amendment concentrates on the precise locus of the paradoxes and it

leaves untouched the harmless negative self-reflections such as:

(5) The present sentence is not in French.

On the other hand, it should be totally acceptable to Ockham who neatly

specifies that all the insolubles contain a term like 'false' or a term like
'true.' 14

3. Close encounter with the abominable Truthteller

But our rule is still far from perfect. One problem it does not deal with
is the so-called Truthteller, whose only statement is:

(6) What I am saying is true.

In such a case, no negative reflection occurs at any stage. So RR2 just does
not apply. But it should. Although often neglected by paradox hunters, the
Truthteller is as much an intellectual scandal as the Liar is. Of course, it
generates no contradiction: if it is true, it is true, and if it is false, it is false.
That is precisely why it has more often than not been thought to be
innocuous. But the trouble here is that nothing in the world, nothing in
language, nothing in logic determines which truth-value the Truthteller is
supposed to have. Here is a well-formed contingent statement about a
precise fact of the world (my own speaking at a certain moment). Nothing
prevents it from being true or false. It must have one and only one truth-

140Ph.I, p. 745, II. 22-6: "Et ad solutionem istius et aliorum omnium est sciendum
quod talis propositio eontingens, ex qua debet inferri sua repugnans, vel habet hune
terminum 'falsum' vel aliquem eonsimilem, vel hune terminum 'verum' vel aliquem
eonsimilem."
404 CLAUDE PANACCIO

value. And yet nothing whatsoever determines which truth-value it has!

This is to me as deeply irrational as the plain contradiction of the Liar. 15

Neither Burley nor Ockham ever thought of the Truthteller as an

insoluble. On the contrary, their general characterizations of what an
insoluble is quite immediately exclude anything as positive as the
Truthteller. But nevertheless another amendment is required. What has to
be relaxed in RR2 in order to cope with the newcomer is simply the part
about negative self-reflection. The rule may be drastically simplified to:

(RR3) A semantical term cannot supposit (personally) for an

expression of which it is a part.

This neutralizes the Truthteller (which turns out to be false!) and all its
variants without affecting, as far as I can see, any of the harmless reflexive
statements (such as (5», which we should certainly prefer to leave
untouched if possible.

We may now seem to be heading further and further away from

Burley and Ockham. But their basic drive is maintained: their idea of
imposing a restriction on the denotation of some terms in certain
(semantical) contexts is precisely what we are exploring. And it is a tribute
to the restrictionist solution that it can be generalized (and even simplified)
to handle both the Truthteller and the Liar with a single blow. Bradwardine
and Buridan's solution to the Liar, by contrast, cannot readily be adapted
to the Truthteller. 16

15J.L. Mackie nicely stresses this point about the Truthteller in Truth, Probability, and
Paradox, Oxford: Clarendon Press 1973; see esp. pp. 240-1.
16The principle of the Bradwardine-Buridan solution is that every sentence has as its
truth-conditions not only that the world be as it says it is, but also that it itself be true.
The resulting truth-value in the Liar case thus turns out to be the same as in the
Burley-Ockham approach, but for different reasons. The truth of a paradoxical sentence
such as 'The present sentence is false' requires that (1) this sentence itself be false (as it
says it is), and (2) the very same sentence be true; its two truth-conditions, then, are
contradictories to one another, and that is why the sentence has to be false after all. Yet
its own truth does not perversely follow from this very falsity, since in this case the
falsity of the sentence under consideration is a necessary but not a sufficient condition
for its own truth. As can readily be seen, this type of solution leaves utterly
undetermined the truth-value of the Truthteller's statement. In this case, the two truth-
conditions coincide: for the sentence to be true, it has to be true! And nothing more can
be said. Thomas Bradwardine's Insolubilia have been edited by Roure (op. cit., pp. 285-
325); see also P.V. Spade, The Mediaeval Liar: A Catalogue of the Insolubilia-
Literature, Toronto: The Pontifical Institute of Mediaeval Studies, 1975, pp. 105-10;
and idem, "Insolubilia and Bradwardine's theory of signification", Medioevo 7, 1981,
pp. 115-34. As regards John Buridan, his main developments on the insolubilia are to
be found in chapter 8 of his Sophismata, ed. by T.K. Scott, Stuttgart: Frommann
Holzboog 1977; the relevant passages have been translated in English and commented
upon both by T.K. Scott himself: John Buridan. Sophisms on Meaning and Truth,
New York: Appleton Century Crofts 1966, and by the well-known logician G.E.
Hughes: John Buridan on Self-Reference, Cambridge: Cambridge U.P. 1982; see also
Fabienne Pironnet's very interesting dissertation, Le paradoxe du menteur dans la
logique medievale. Edition des Sophismata de Jean Buridan (partim), Universite de
Liege,1986-7.
 SOLVING THE INSOLUBLES: OCKHAM AND BURLEY 405

4. Types and tokens

Unfortunately, RR3 as it now stands is still ambiguous. It could mean

that a semantical term cannot supposit in a sentence s for s itself considered
as a type. Or it could simply prevent the token-term from standing for the
very token-sentence in which it appears. These are two quite different
rules. I will call them type-restrictionism and token-restrictionism
respectively. So far we have in fact tacitly adopted a type-restrictionist
interpretation ofRR, RRl, and so on. This is what Ockham himself seems
to be doing when he writes, for instance, that "'Socrates is not uttering a
truth' is equivalent to 'Socrates is not uttering a truth different from this (ab
isto): Socrates is not uttering a truth. "'17 It is clear that the demonstrative
'this' in this quotation points to a different token of 'Socrates is not
uttering a truth' from the one the whole sentence started with. And
therefore the restriction-rule is not in practice seen by Ockham as excluding
only the original sentence-token from the supposita of the term under
consideration. Ockham leans toward type-restrictionism rather than token-
restrictionism.

Despite appearances, this is not inconsistent with his nominalistic

ontology. Of course, linguistic types in themselves are universals. So
strictly speaking, no term can be admitted in Ockhamist semantics to really
supposit for them (supposita have to be singular right across the board).
But to say that a term supposits for a type can be, in a nominalist
framework, interpreted as an abbreviated way of saying that the term
supposits for every existing token of that type. Thus a type-restrictionist
interpretation ofRR3 which would be acceptable to Ockham's nominalism
has to read as follows:

(RR4) A semantical token-term cannot supposit (personally) for a

token-expression of the same type as one of which it is a part.

Token-restrictionism, by contrast, is encapsulated in the following rule:

(RR5) A semantical token-term cannot supposit (personally) for a

token-expression of which it is a part.

RR4 and RRs are not equivalent to each other. For instance, if Plato and
Socrates both simultaneously say 'What Socrates says is false', then
according to RR4 they are both saying something false, while according to
RRs, only Socrates' token is false, Plato's one being counted as true.

What then is the best interpretation? Historically, as I said, Ockham

does seem in his treatment of examples to have type-restrictionism in mind.
But this might not be a considered decision on his part. The type idiom
comes much more naturally than the token one, even to a medieval thinker
and even to a nominalist. Ockham, for example, often expresses himself in
logic as if he was talking about linguistic types even in situations where,

170Ph. I, p. 745, II. 41-2: " ... ista propositio 'Sortes non dicit verum' aequivalet isti
'Sortes non dicit aliud verum ab isto: Sortes non dicit verum'."
406 CLAUDE PANACCIO

according to his own ontology and semantics, he should have been

speaking only of tokens. In general, although tokens are clearly posited by
him as the real bearers of semantical properties, he does not seem to be
clearly aware of the requirements, difficulties, and possibilities of the
systematic sort of token-semantics which his nominalism calls for.18

Logically, type-restriction ism as formulated in the framework of

token-semantics by way of RR4 (that is, as excluding all the tokens of a
certain type from the supposita of a certain token-term) is unnecessarily
strong. The best restriction-rule is surely the weakest efficient one: in
semantics as in real life, restrictions should not be imposed without
necessity! It can be checked that RRs is strong enough to solve the
insolubles Burley wants to solve and much more (such as the Truthteller
and, as I have shown elsewhere, Grelling's heterologicality's paradox).l9
As far as I can see, RR4 has no special virtue of its own. So token-
restriction ism as expressed by RRs has to be deemed philosophically
preferable. And even if it is not what Ockham more or less clearly had in
mind, it is certainly deeply congenial to his general nominalist outlook.

5. Odd Couples, Infernal Trios, and indirect reflexivity

A classical objection to restrictionism (found in Buridan, for example)

is that it cannot cope with situations involving two speakers. Consider, for
instance, the case of what I will now call the Asymmetrical Twins. Let us
suppose that it is Plato not Socrates who utters only sentence (1): 'What
Socrates says is false'; and let Socrates say nothing but:

(7) What Plato says is true.

RRs here prevents 'false' in (1) from standing for (1) and 'true' in (7)
from standing for (7). But in this case, this is not enough to neutralize the
threatening contradiction. To do it, either 'false' in (1) should be prevented
from suppositing for (7) or 'true' in (7) should be prevented from standing
for (1). Maybe Burley's original formulation RR could be expected to deal
with such a case. But it was unduly indeterminate about the circumstances

18It must be noticed, though, that the type/token distinction had been quite clearly
drawn in explicit relation with the formulation of a restriction rule for solving the
insoiubilia by an anonymous (l3th century?) logician in a treatise edited by H.A.G.
Braakhuis: "The second tract on insoiubilia found in Paris, B.N. Lat. 16.617. An
edition of the text with an analysis of its contents", Vivarium 5, 1967, pp. 111-45.
See esp. p. 134: " ... distinguo cum Aristotile quod enuntiationum quaedum sunt eedum
numero, quedam sunt heedem specie. Et dico quod hec oratio: 'aJiquid dicitur a me' bis
dicta non est eadem numero sed specie ( ... ). Et si hoc est, dico quod quamvis
impossibile sit quod terminus supponit pro eadem oratione numero cuius est pars,
potest tamen supponere pro oratione que est eadem specie cum oratione cuius est
pars .... " This is a striking instance of token-restrictionism (but its rule is not limited
to seman tical terms as in RRS).
19See C. Panaccio, Guillaume d' Occam et ies paradoxes semantiques, Cahiers
d'epistemologie no. 8705, Montreal: Universite du Quebec 1987. This paper - which
had a very limited circulation - was an earlier and, on the whole, less satisfying
attempt of mine at reconstructing Ockham' s solution to the paradoxes in a theoretically
promising way.
 SOLVING THE INSOLUBLES: OCKHAM AND BURLEY 407

under which negative self-reflections threaten to occur. RRs at any rate is

beaten by the Asymmetrical Twins and we have no choice but to amend it.

What is needed is a generalization of RRs. It should take care not only

of the Asymmetrical Twins but of some other Odd Couples as well, such
as (the appellations are mine): the Polite Neighbours, who both utter a
token of:

(8) What my neighbour says is true

(a variation on the Truthteller theme); and the Hostile Brothers, who both
utter a token of:

(9) What my brother says is false

(at most one of them could logically be right, but nothing determines which
one). But this still would not be enough. The rule we are looking for
should also settle all cases involving three speakers (let us call them
Infernal Trios!}20 or, for that matter, 4, 5, or n speakers. Suppose that
they all form a circle and each one of them only utters a token of:

(10) What the person to my right says is true;

we then have a generalized Truthteller situation. If, instead, each speaker

only utters a token of:

(11) What the person to my right says is false

then if the number of ~peakers is odd, the familiar contradiction threatens

under any attribution of truth-values and we have a generalized Liar
situation; if the number of speakers is even, what we have is a generalized
Hostile Brothers situation: at least two incompatible truth-valuations are
logically possible, but nothing whatsoever determines which one is the
right one. And similar problems also arise for any combination of tokens
of (I O) and tokens of (II) among the speakers in the circle, whatever their
number may be.

To settle all these cases with a single rule, a few intermediate

theoretical notions first have to be introduced. Let us lay down the
following definitions:

- A suppositional chain of sentences =df any series of sentences SI ...

Sn such that for any Sj in the series up to Sn_l, at least one of the

20/nsolubilia involving three speakers were not unknown to medieval authors. See, for
example, the anonymous treatise called Insolubilia Monacensia, ed. L.M. de Rijk,
"Some notes on the mediaeval tract De insolubilibus, with the edition of a tract dating
from the end of the twelfth century", Vivarium 4, 1966, pp. 83-115: "Unde si ita sit
quod hic sint tres homines: Sor, Plato, Cicero; Sor dicat Platonem mentiri, Plato
Ciceronem mentiri, Cicero Sortem mentiri, sequitur circularis deductio ..... (p. 109).
Other references to medieval "Infernal Trios" can be found in a recent paper by S.
Ebbesen and P.V. Spade, "More Liars", Cahiers de l'Institut du Moyen-Age Grec et
Latin 56, 1988, pp. 193-227 (see pp. 193-4).
408 CLAUDE PANACCIO

tenns of Sj (the subject or the predicate) supposits (personally)

for Sj+I'
- A circular suppositional chain of sentences = df any suppositional
chain of sentences in which SI =So.

- A semantical sentence = df any sentence in which one of the tenns

(the subject or the predicate) is a semantical tenn (in the sense
defined above in the paper).

Thus the following series of sentences fonns a suppositional chain:

(SI) S2 is an English sentence,

(S2) S3 is a French sentence,

(S3) The cat is on the mat.

The subject of SI and its predicate both supposit for S2, and the subject of
S2 supposits for S3' The fact that in this case the predicate of S2 does not
supposit for S3 (and consequently that S2 is false) does not break the chain.
On the other hand, the last member in the chain does not have to be
metalinguistic in any sense, but if we change S3 for a metalinguistic
sentence in which one of the tenns supposits for SI (for example: 'SI is a
five-word sentence'), the result is a circular suppositional chain.

It should by now be clear that what we want to ban in order to avoid

the semantical paradoxes we have introduced so far is all circular
suppositional chains of semantical sentences.21 The original Liar and
Truthteller sentences are but special cases of such circular chains, cases,
that is, where all members of the chain are identical to one another (SI = S2
= Sn).

The most appropriate rule I can think of, then, is the following one:

(RR6) A semantical token-tenn cannot supposit (personally) in a

token-sentence S for a token-sentence S' if S and S' both belong to
the same circular suppositional chain of semantical sentences.

If this rule is in force, the Liar's, the Truthteller's, the Asymmetrical

Twins', the Polite Neighhbours', the Hostile Brothers', and the Infernal
Trios', Quartets' or n-tets' statements all turn out to be false, since in all
these cases the semantical predicate of an affinnative sentence is prevented
(by RR6) from suppositing for what its subject supposits for.

21 If at least one sentence in the chain is not a semantical sentence, troubles are avoided,
even if the chain is circular. Think of the following situation: Socrates says 'What
Plato says is false', Plato says 'What Cicero says is true', and Cicero says 'What
Socrates says is in French'. Here Cicero is just plainly wrong since what Socrates says
is in English, and consequently Plato is also wrong, and Socrates is right.
 SOLVING THE INSOLUBLES: OCKHAM AND BURLEY 409

In the generalized Liar or Hostile Brothers situations, for example,

where each member in a circle utters a token of (11), each one of these
tokens has to be interpreted as:

(11 a) What the person to my right says is a falsehood which does not
belong to a circular suppositional chain of semantical sentences
to which the present sentence also belongs.

In order for (11a) to be true, the sentence for which its subject supposits
must not only be false, but it also must not belong to a certain chain to
which, in this case, it does belong. Thus (11a) is false after all, and so is
(11). And the same reasoning applies to all these similar cases where the
sentences with the truth- or the falsehood-predicate are affIrmative.

In the corresponding Strengthened cases, where the suspect sentences

are negative, all these sentences tum out to be true. Suppose that
everybody in the circle utters a token of:

(12) What the person to my right says is not true.

RR6 forces us to interpret each one of these statements as:

(l2a) What the person to my right says is not a truth which does not
belong to a circular suppositional chain of semantical sentences
to which the present sentence also belongs.

In order for (12a) to be true, it must be the case either that the sentence
which its subject refers to is not a truth or that this sentence belongs to a
certain circular chain. Since in the case under consideration, the second of
these conditions is fulfilled, (l2a) - and consequently (12) - is true. And
so is - for the same reason - the corresponding sentence:

(13) What the person to my right says is not false.

6. Conclusion

What we now have is a generalization of Ockham's rule of thumb: in

the typical paradoxical situations, all the affirmative sentences are false,
and all the negative ones are true. His own rule has been generalized by
taking into consideration the Truthteller on the one hand, and all the
situations involving more than one speaker on the other hand. But the
original tack has been maintained: that of precisely restricting the
supposition of semantical terms in order to avoid certain sorts of (direct or
indirect) self-reflections. No doubt the rule could still be refined. In
particular, we have, in the passage from RRs to RR6, replaced
'expression' by 'sentence'. That was good enough for our present
purposes. But there are some semantical paradoxes - such as Grelling's
heterologicality paradox - in which the problematic circular reference is to
expressions which are not themselves sentences (to the adjective
'heterological' itself, for example). These puzzles could be taken care of by
reformulating RR6 and the intermediate definitions that were preliminary to
410 CLAUDE PANACCIO

it in order to reintroduce' expression' instead of 'sentence'. But that would

bring about certain technical complications which are not necessary here.

However it may be, the general trend is clear. So I will for the time
being rest with RR6. It has many virtues. It solves in a very economical
way Burley's typical examples of insolubles, plus the Truthteller, the
Asymmetrical Twins and all these other cases we have considered. It also
allows us to counter Tarski's severe verdict about the inconsistency of
ordinary languages: a language containing its own truth-predicate avoids
inconsistency if it is equipped with a restriction-rule such as RR6.
Moreover, it nicely conforms to the strictest nominalist requirements as
well as, it must be stressed, to the rest of Ockham's semantics. Is it
intolerably artificial and ad hoc or can it be founded on some independent
motivation? This is an important point of course, which I have not
discussed here. My sole aim has been to formulate the most efficient
Ockhamistic restriction rule I could think of. Might it not be, after all, that
under Bradwardine's, Heytesbury's, and Buridan's attacks, restrictionism
as a way of solving the insolubles has been despaired of too soon and that
it still deserves a serious try?22

Universite du Quebec a Trois-Rivieres

22Thanks are due to the Canadian Social Sciences and Humanities Research Council for
financial support and to a number of persons for helpful remarks and discussions on
earlier variations of mine on the same theme, especially to Paul Gochet, Hubert
Hubien, Elizabeth Karger, Daniel Laurier, Alain de Libera, Joanna Pasek, Fabienne
Pironnet, Graham Priest, and Stephen Read.
Indice s
 Index of Manuscripts

Towns and Libraries Manuscript Pages

BAMBERGStiftsbibliothek
lat. HJ V.1 225,262

BARCELONA Archivo de la Ripo1l109 196, 232, 236,

Corona de Aragon 241,260

BASEL Universitiitsbibliothek BVillI4 262

BRUGES Stadsbibliothek 497 144-5

506 6

509 45-62

CRACOW Biblioteca Jagiellonica 25 117

621 103-15,117

642 117

649 187, 189

662 116

703 116

711 116

712 116

713 116

719 116

736 116,126

750 116

1587 123

1893 117

2178 127

2205 120-6

2215 120

2216 120

413
414 INDEX OF MANUSCRIPTS

Towns and Libraries Manuscript Pages

(CRACOW) 2231 120

2330 116

2459 120

2591 127

2660 117
Biblioteka Ksiezy 827 116
Misjonarzy

CAMBRIDGE Gonville & Caius 182/215 6

611/341 45,57

Peterhouse 191 239-41,244-5,

247,260,263

University Library KID20 236,238-41,

243-7,260

COPENHAGEN Royal Library fragm.1075 144

ERFURT Wissenschaftliche Amplon. 40 220 262

Allgemeine Bibliothek

Amplon. 40 276 194, 196

Amplon. 40 328 187, 193, 199

Amplon. 80 10 262
FABRIANO Biblioteca Comunale 34 308

FLORENCE Biblioteca Mediceo- S. Croce 12 sin., 6, 45-62, 194,

Laurenziana 3 198

Fiesol. 161 89-102

Biblioteca Nazionale Conv. Soppr. 262

Centrale D.II.45

Conv. Soppr. 47,62

E.1.252

V43 128-40

LONDON British Library Royal 8 A VI 233,239,260

Royal Fxix 12 144-84

 INDEX OF MANUSCRIPTS 415

Towns and Libraries Manuscript Pages

(LONDON) Add. 8167 233

Lambeth Palace 221 140

LAON 465 222

MANTUA Biblioteca Comunale D III 19 (445) 94,97-100,102

MUNICH Bayerische Clm 7205 233

Staatsbibliothek

Clm 14522 51,53-4,62,

185-201

Clm 19672 128-43

ORLEANS Bibliotheque 266 64-85

Municipale

OXFORD Bodleian Library Digby 2 5,144-84

Digby 24 5,144-5

Canon. lat. 278 129

Corpus 293B 144-6

Magdalen 92 107

Oriel 15 349

PADUA Biblioteca Universitaria 1123 31,35,37-8,42,

129-40

1570 129

1589 194

PALERMO Biblioteca Comunale 2 QqD142 102

PARIS Bibliotheque de l'Arsenal 711 224

Bibliotheque Nationale 1374 46,63

3572 219

11412 373,376

14069 144

14927 219
416 INDEX OF MANUSCRIPTS

Towns and Libraries Manuscript Pages

(PARIS) 14947 46

15005 46

15135 220

16089 45-6,48,50-1,
58,62

16134 288-303

16135 51,53-4,62,
187, 191-2, 194,
196,198,200-1,
219-30,259-60

16149 188

16160 47,62

16221 219-20,222

16617 406

16618 187,194,196,
219-24,230,
231-60

18528 221

PI STOIA Archivio Capitolare 61 128-40

ROME Biblioteca Casanatense 85 (D.IV.3) 31

98 129

Biblioteca Angelica 1017 103-15

ST FLORIAN Stiftsbibliothek XI.632 262

SEVILLE Biblioteca Capitular Colomb. 5-1-14 31

Colomb. 5-5-9 262-74

UPPSALA Universitetsbiblioteket C599 3-30

C600 4

C601 4

V ATICAN Biblioteca Apostolica Vat.1at. 298 236, 238-9, 243-

7,251,260
 INDEX OF MANUSCRIPTS 417

Towns and Libraries Manuscript Pages

(VATICAN) Vat.lat. 2136 31

Vat.lat. 2189 31

Vat.lat. 2520 219

Vat.lat. 3061 47,63,87-102

Vat.lat. 6768 87-102

Vat.lat. 7678 186, 195-6,219,

232-3,250-1,
254,260

Vat.lat. 14718 6, 45-6, 48-50,

58,63

VENICE San Marco lat. Z 310 129

(1577)

VIENNA Osterreichische Pal. 4698 129-40

N ationalbibliothek

WARSAW Biblioteka Narodowa akc.1819 117

WORCESTER Cathedral Library FIl6 31

FIl8 128-42

Q13 52,63,187,219-
20,224-5

WROCLAW Biblioteka 380 120

Uniwersytecka

ZWETTL Stiftsbibliothek 338 232,238-9,247-

8,251-2,256,
260, 262, 264-6,
268,270-1,274
 Index of Names

Abelard (Abaelard, Abailard) Boethius of Dacia (Boethius of

x, 69, 72-4,195,219,226-7, Denmark, Boethius de
358-9,361-2 Dacia) x, xiv, 46, 49-50, 55-
Adam of Buckfeld 188 6,59
Adam of Whitby 188, 192n Boethius, A.M.S. 19n, 28n, 64-
Adams, M.M. 334 70,73-4,81,84,112,146-7,
Aegidius Romanus, see Giles 168,194,274,358-61,367
of Rome Boh, I. 12
Albert of Saxony (Albert de Bourges, see Joscelin of
Saxe, Albert von Sachsen, Soissons
Albertus de Saxonia) xi, xiii, Braakhuis, H. x, 199
4-8, 11, 16-20,22, 24-5nn, Bradwardine, see Thomas B.
116, 187, 189,288-303,385, Brinkley, see Richard B.
392n Bruges, see Bartholomew of B.
Albert the Great (Albert Ie Bruni, L. ix
Grand, Albertus Magnus) Buckfeld, see Adam of B.
5n, 97, 242n Burge, T. 375
Alessandro Sermoneta 108n Buridan, see John B.
Alexandre de Villedieu 250 Burleigh, see Walter B.
Andree de Neufchatel 35n Caraccioli, see Landulf C.
Andrews, R. xv Chisholm, R. 366
Anselm (Anselmus) 65 Cicero 5n
Aquinas, St Thomas x, 258-9, Clarembald of Arras 360, 367
280-I, 283, 285, 335n Coombs, J. xi
Aristotle (Aristote, Curry, H.B. xi
Aristoteles) xiii, 5, 18-20nn, Davidson, D. 15
22-3, 26, 28n, 39,43,47, 64, de Libera, A. x, xiv, 54, 105,
67, 100, 112, 145,213,297, 192,196,227,231
315,317-8,357-9,361,363 de Rijk, L.M. x, xv, 4, 5,12,
Armstrong, D.M. 71, 74 130,144,358,393-7
Arnauld, see Pseudo-Giles of Donatus 222
Rome Dry ton, see John of Secheville
Arnold of Strelley 349n Durandus 232, 260
Ashworth, E.J. 12 Ebbesen, S. x, xv, 12, 14,93,
Augustine (Augustin) 15, 256, 94n,110
355 Eudemus 358
A verroes 100 Faes de Mottoni, B. 187
Avicenna 90 Ferrybridge, see Richard F.
Blick, A. xi Fletcher, J. 104-5, 139
Bacon, Robert, see Robert B. Franciscus Mayronis 350
Bacon, Roger, see Roger B. Gaetanus de Thienis 6
Bartholomew of Bruges xiii Galerannus, master 65
Bermingham, see William B. Garlandus Compotista 359
Bernardino del Pietro Landucci Gauterus, master 65
108n Gauthier d'Ailly 87
Bertagna, M. x-xi Gauthier, R.-A. 188
Biard, J. xiii Geach, P.T. 71
Billingham, see Richard B. Gelber, H. 349, 353

418
 INDEX OF NAMES
Gentilis de Cingulo 95
Gibson, S. 106-7
Gilbert of Poitiers 360
Giles of Rome 4, 188, 382n,
386n
Goodman, N. 366
Goslenus, see Joscelin of
Soissons
Gosvin of Marbais 220, 223,
225-7,230
Grabmann, M. x, 45
Gregoire de Rimini 290-2, 294,
297-9,303
Grosseteste, see Robert G.
Guido, master 65
Guillaume Arnauld, see
Pseudo-Giles of Rome
Henri de Harclay 298
Henricus, master 65
Henry Hopton (Heinrich
Hopton) 32n, 33n, 103-15,
116
Heytesbury, see William H.
Hieronymus of Hangest 27n
Hintikka, J. 213-16
Hopton, see Henry H.
Hughes, G. 369
Inghen, see Marsilius of I.
Jandun, see John of J.
Jaques of Vitry 362
Jeronimo Pardo 319-332
Johann von Glogau 117
Johannes Buser 131
Johannes de Osswyancim 125n
Johannes Konigsbeerg de
Oppavia 122n
Johannes Ie Rus 232, 233n, 260
Johannes Muntzinger 130
Johannes Venator 114
Johannes von Monickedam 117
John Buridan (Jean Buridan,
Johannes Buridan) x, xi, 4,
116, 130-1,288-91,293-4,
300,302,322,333-47,351,
358,367-9,404,406,410
John Duns Scotus 307, 370
John Ie Page (Jean Ie Page) xiv,
194-6,220,227-30,234
John Major 7-8
John of Jandun 4, 86, 89n, 91,
100
419
John of Salisbury 74
John of Secheville (John de
Seccheville, John Dry ton)
53-4, 192
John Wyclif 307-9
Jordanus de Tridentia 94n
Joscelin of Soissons 64-85
Kilvington, see Richard K.
Kilwardby, see Robert K.
Knuuttila, S. 202, 208
Kretzmann, B. x
Kretzmann, N. x, 12,393-7
Kripke, S. 357
Lambert of Lagny 195
Landucci, see Bernardino del
Pietro L.
Landulf Caraccioli 349
Laurentius Aicher of St
Emmeram 189n
Lehtinen, A.I. 202, 208
Lewis, D. 352
Lewry, P.O. 257
Libera, see de Libera, A.
Mackie, J.L. 366
Magister Abstractionum
(Ricardus Sophista) xi, 5n,
144-84, 186, 195
Maier, A. 350
Maieril, A. x, 12, 94n
Major, see John M.
Marenbon, J. ix
Marsilius of I.(Marsilius von
Inghen) 4, 16, 116, 340n
Marsilius of Padua 86-102
Martin de Dacie 245n
Martinus Anglicus 130
Matthew, master 195
Maulevelt, see Thomas M.
Michael von Biestrzyk6w 117
Michalski, K. 348
Minio-Paluello, L. 64
Muntzinger, see Johannes
Muntzinger
Neufchate1, see Andree de N.
Nicholas de Pressoir 46n
Nicholas of Normandy
(Nicolaus de Normandia)
46,56,59
Nicholas of Paris (Nicolas de
Paris) 70, 186,234,236n
Normore, C. 362
420 INDEX OF NAMES
Nuchelmans, G. 234, 257
Ockham, see William of O.
Panaccio, C. xii
Pappus 206
Pardo, see Jeronimo P.
Paul of Pergula 12n
Paul of Venice (Paolo Veneto,
Paulus Nicolettus Venetus)
10, 34n, 304-318, 319,327
Peter Candia (Petrus Candie)
130
Peter Helias 219, 224
Peter Lombard 348
Peter of Auvergne 6, 45-7, 50,
52-60,92,194,198
Peter of Mantua 108n
Peter of Spain (Petrus
Hispanus, Pierre
d'Espagne) 5-7, 18n, 116,
150,229,292
Peter, master 64
Petrus de Colonia 87
Petrus de Insulis 87
Petrus Olai 3-30
Petrus Rogerii 350
Piltz, A. 3
Pinborg,J. 12,45, 197,219,234
Pironet, F. xi
Porphyry 64, 95, 99, 101, 194
Priscian (Priscien) 69, 75, 77-8,
82,219,222,226,236,252,
255, 263, 272-4
Pseudo-Giles of Rome 92
Pseudo-Grosseteste 222, 237,
241, 243n, 246n,261
Pseudo-Johannes Ie Rus 239n,
250,260
Pseudo-Kilwardby 236, 239n,
240,244,260
Pseudo-Sherwood 362, 368n,
370
Quine, W.V. 15
Radulphus Brito 47-8, 56, 87,
92,95, 186, 199,220n
Ralph Strode 110
Ricardus Sophista, see
Magister Abstractionum
Richard Billingham (Bilingam)
5,9, 10-11, 16, 24n, 26-29nn,
27n, 28n, 29n, 32n, 33, 109-
ll, 112n, 130,202-5,210
Richard Brinkley 33, I11n
Richard Ferrybridge xi, 31-44,
111
Richard Kilvington 103, 116,
204, 353, 363n
Riehard of Campsall 354
Richard Rufus 149-51
Rijk, see de Rijk, L.M.
Robert Alyngton 129n
Robert Bacon 235, 261, 288
Robert Caubraith 319-21, 326-
32
Robert Grosseteste 100
Robert Holcot 348-356
Robert Kilwardby 8, 193, 199,
219-20,222-3,225-6,231-2,
236, 238, 245, 246n, 248n,
251n, 252n,253,255-8,260-
1,263-74, 382n, 386n, 387n
Robert of Paris (Robert de
Paris) 70, 71, 219, 226-7,
229n
Robertus 64
Roger Bacon 5, 7, 149-50, 194,
199,219-20,222,231-6,240,
241-2,245,248-9,253-8,
261, 262-3nn
Roger Roseth 349-51
Roger Swyneshed 354
Roos, H. 45
Roscellinus 65
Rosier, I. x
Russell, B. 212
Saxony, see Albert of S.
Scardeburh, see William S.
Scott, T.K. 337
Scotus (Scoto), see John Duns
Scotus
Sermoneta, see Alessandro S.
Sextus Empiricus 348
Sherwood, see William of S.
Siger of Brabant 89
Siger of Courtrai 89, 219
Simmons, K. 400, 402
Simon of Faversham 92, 95,
186, 199
Spade, P.V. 362-4
Stanislaus of Znaim 354n
Strawson, P.F. 71
Streveler, P. xi, xv
Stump, E. 362-3
 INDEX OF NAMES
Sylla, E. 104, 107
Tabarroni, A. x, xii, xiv, xv
Tarski, A. 15-16, 112, 402n, 410
Thierry of Chartres 360
Thomas Bradwardine 109,404,
410
Thomas d'Aquin, see Aquinas,
St Thomas
Thomas de Moston 108
Thomas Maulevelt 4
Thomas of Erfurt 247n
Villedieu, see Alexandre de V.
Vincent Ferrer, St. 385, 391n
Virgil (Virgile) 255
Walentin von Krakau 117
Walter Burleigh (Burley) xii, 6,
8,2In, 199,277,284-6,307,
334,337,362, 385,386-7nn,
389n,393n,398-410
Walterus, master 65
Weisheipl, J. 104-6
Whitby, see Adam of W.
Willelmus, master 64
421
William Bermingham 33, 111
William Chatton 35n
William Heytesbury
(Guillaume Heytesbury) xi,
6,32n, 103-4, 109, 128-43,
289-96,299,303,319,
410William of Champeaux
65
William of Conches 69
William of Ockham x-xii, 6, 70,
112,116,147,149,210-11,
277,281-7,307,333-7,340,
345-6,351-3,355-6,398-410
William of Sherwood
(Guillaume de Shireswode)
xi,5, 7, 12n, 17n, 186, 195,
234,249,277-81,284,288,
292,362,384,393-7
William Scardeburh 219-20,
223-4,230
William, master 72-4, 76-7
Wilson, C. 128
Wyclif, see John W.
 Nijhoff International Philosophy Series

35. R.L. Epstein: The Semantic Foundations of Logic. Volume 1: Propositional

Logics. With the assistance and collaboration of W.A. Carnielli, LM.L.
D'Ottaviano, S. Krajewski and R.D. Maddux. 1990 ISBN 0-7923-0622-8
36. A. Pavkovic (ed.): Contemporary Yugoslav Philosophy. The Analytic Ap-
proach. 1988 ISBN 90-247-3776-1
37. A. Winterbourne: The Ideal and the Real. An Outline of Kant's Theory of
Space, Time and Mathematical Construction. 1988 ISBN 90-247-3774-5
38. K. Szaniawski (ed.): The Vienna Circle and the Lvov-Warsaw School. 1989
ISBN 90-247-3798-2
39. G. Priest: In Contradiction. A Study of the Transconsistent. 1987
ISBN 90-247-3630-7
40. J. Wolenski (ed.): Kotarbinski: Logic, Semantics and Ontology. 1990
ISBN 0-7923-0865-4
41. P. Geach (ed.): Logic and Ethics. 1991 ISBN 0-7923-1044-6
42. U. Wybraniec-Skardowska: Theory of Language Syntax. Categorial Approach.
1991. ISBN 0-7923-1142-6
43. J. Bransen: The Antinomy of Thought. Maimonian Skepticism and the Relation
between Thoughts and Objects. 1991 ISBN 0-7923-1383-6
44. J.T.J. Srzednicki, S.J. Surma, D. Barnett and V. F. Rickey: S. Lesniewski's
Collected Works. 1992 ISBN 0-7923-1512-X
For 'The Lesniewski's Collection' see also Volumes 13 and 24.
45. P. Simons: Philosophy and Logic in Central Europe from Bolzano to Tarski.
Selected Essays. 1992 ISBN 0-7923-1621-5
46. J. T. J. Srzednicki and D. Wood (eds.): Essays on Philosophy in Australia.
1992 ISBN 0-7923-1695-9
47. B. Dziemidok: The Comical. A Philosophical Analysis. 1993
ISBN 0-7923-2103-0
48. S. Read (ed.): Sophisms in Medieval Logic and Grammar. 1993
ISBN 0-7923-2196-0

Further information about our publications on Philosophy are available on request.

Kluwer Academic Publishers - Dordrecht I Boston I London