100% found this document useful (3 votes)
640 views369 pages

Early Modern Philosophy Universals

This document provides an introduction to the problem of universals in modern philosophy. It discusses how the problem originated in antiquity and was extensively debated in the medieval era in scholastic literature. While many modern philosophers were skeptical of scholastic concerns and did not explicitly address universals, the topic remained important to modern philosophy through reconceptualization. The chapters that follow examine how specific modern philosophers such as Gassendi, Hobbes, Spinoza, Descartes, and others addressed questions regarding universals.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (3 votes)
640 views369 pages

Early Modern Philosophy Universals

This document provides an introduction to the problem of universals in modern philosophy. It discusses how the problem originated in antiquity and was extensively debated in the medieval era in scholastic literature. While many modern philosophers were skeptical of scholastic concerns and did not explicitly address universals, the topic remained important to modern philosophy through reconceptualization. The chapters that follow examine how specific modern philosophers such as Gassendi, Hobbes, Spinoza, Descartes, and others addressed questions regarding universals.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 369

THE PROBLEM OF 

UNIVERSALS IN
E A R LY M O D E R N P H I L O S O P H Y
THE PROBLEM OF
UNIVERSALS IN
EARLY MODERN
PHILOSOPHY

Edited by Stefano Di Bella

and Tad M. Schmaltz

1
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.

Published in the United States of America by Oxford University Press


198 Madison Avenue, New York, NY 10016, United States of America.

© Oxford University Press 2017

All rights reserved. No part of this publication may be reproduced, stored in


a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.

You must not circulate this work in any other form


and you must impose this same condition on any acquirer.

CIP data is on file at the Library of Congress


ISBN 978–​0–​19–​060804–​0

9 8 7 6 5 4 3 2 1
Printed by Sheridan Books, Inc., United States of America
CONTENTS

Acknowledgments vii
List of Contributors ix
1.   Introduction to Universals in Modern Philosophy—
Stefano Di Bella and Tad M. Schmaltz 1
2.   Gassendi on the Problem of Universals—Antonia LoLordo 13
3.   Hobbes, Universal Names, and Nominalism—Stewart Duncan 41
4.   Spinoza on Universals—Samuel Newlands 62
5.   Descartes on Universal Essences and Divine
Knowledge—Lawrence Nolan 87
6.   Platonism and Conceptualism among the
Cartesians—Tad M. Schmaltz 117
7.   Universals and Individuals in Malebranche’s
Philosophy—Mariangela Priarolo   142
8.   Universals in English Platonism: More, Cudworth,
Norris—Brunello Lotti 166
9.   Some Perspectives on Leibniz’s Nominalism
and Its Sources—Stefano Di Bella 198
10.  Locke’s Essay and Leibniz’s Nouveaux Essais: Competing Theories
of Universals—Martha Brandt Bolton 220
11.   Locke on General Ideas—E. J. Lowe 252
v i   •  Contents

12.   Berkeley on Abstraction, Universals, and Universal


Knowledge—Tom Stoneham 267
13.   Hume on Abstraction and Identity—Donald L. M. Baxter 285
14.   Kant and Abstractionism about Concept
Formation—Alberto Vanzo   305

Works Cited 325


Index 345
ACKNOWLEDGMENTS

Several chapters in this volume were first presented at a workshop on the problem
of universals in early modern philosophy, held at the Scuola Normale Superiore
in Pisa, Italy, in November 2011. This workshop was part of a cycle of conferences
on the topic of universals, the central aim of which was to combine a historical
treatment of different periods of this problem (ancient, medieval, and modern)
with a consideration of theoretical perspectives drawn from the contemporary
philosophical debate. The idea of this ambitious project came from Francesco
Del Punta, a historian of medieval philosophy, who provoked us to explore the
continuities and transformations of treatments of this topic in modern philoso-
phy, outside the limits of the medieval philosophical agenda with which it is usu-
ally associated. While working on this volume, we lost Francesco and another
friend and colleague, Jonathan Lowe, who also contributed to the general project
from its very beginning and who is the author of a chapter of the present volume.
We dedicate this volume to the memory of Francesco and Jonathan.
LIST OF CONTRIBUTORS

Donald L. M. Baxter is Professor and Department Head in the Department of


Philosophy at the University of Connecticut. He specializes in Metaphysics and
Early Modern (Western) Philosophy. He authored Hume’s Difficulty: Time and
Identity in the “Treatise” (Routledge, 2008) and coedited (with A. J. Cotnoir)
Composition as Identity (Oxford University Press, 2014).
Stefano Di  Bella is Professor of History of Philosophy at the University of
Milan, Italy. His research interests are focused on the history of philosophy,
especially early modern, and metaphysics. He is the author of numerous articles
in early modern philosophy, and of the volumes Le “Meditazioni metafisiche” di
Descartes: Introduzione alla lettura (1997) and The Science of Individual: Leibniz’s
Ontology of Individual Substance (2005).
Martha Brandt Bolton is a Professor of Philosophy at Rutgers University. She
works on the history of seventeenth-​and eighteenth-​century philosophy and is
currently interested in theories of cognition and their metaphysical implications.
Stewart Duncan is Associate Professor of Philosophy at the University of Florida.
He is the author of several articles on Hobbes, Leibniz, and other seventeenth-​
century philosophers.
Antonia LoLordo is Professor of Philosophy at the University of Virginia. She
is the author of Pierre Gassendi and the Birth of Early Modern Philosophy and
Locke’s Moral Man, as well as papers on Descartes, Edwards, Gassendi, Locke,
Malebranche, Shepherd, and other early modern philosophers.
Brunello Lotti is Associate Professor of History of Philosophy at the University
of Udine, Italy. He is author of Ralph Cudworth e l’idea di natura plastica (Udine,
2004) and of L’iperbole del dubbio. Lo scetticismo cartesiano nella filosofia inglese
tra Sei e Settecento (Florence, 2010).​
E. J. Lowe (1950–​2014) studied in Cambridge and Oxford and taught for many
years at the University of Durham. He was an original and influential philosopher.
His most important contributions were to philosophy of mind, philosophical
x   •  List of Contributors

logic, and especially metaphysics, but his scholarship extended to early modern
philosophy, especially Locke. Among his numerous books and articles are Kinds of
Being: A Study of Individuation, Identity and the Logic of Sortal Terms (1989), Locke
on Human Understanding (1995), Subjects of Experience (1996), The Possibility of
Metaphysics (2001), Locke (2005), The Four-​Category Ontology (2006), Personal
Agency: The Metaphysics of Mind and Action (2008).
Samuel Newlands is William J. and Dorothy K. O’Neill Associate Professor of
Philosophy at the University of Notre Dame. He has received an NEH fellowship
for his work on Spinoza, and he has published more than two dozen pieces on
early modern philosophy. Most recently, he is the author of Reconceiving Spinoza
(Oxford University Press).
Lawrence Nolan is Professor of Philosophy at California State University, Long
Beach. He is the author of numerous articles in early modern philosophy and
the editor of Primary and Secondary Qualities: The Historical and Ongoing
Debate (Oxford University Press, 2011) and The Cambridge Descartes Lexicon
(Cambridge University Press, 2016).
Mariangela Priarolo has been a research fellow in History of Philosophy at the
University “Ca’ Foscari,” Venice. She is working on a project devoted to Leibniz’s
perspectivism, in particul Leibniz’s conception of toleration. Her publications
include two books (Visioni divine: La teoria della conoscenza di Malebranche tra
Agostino e Descartes, 2004; Il determinismo: Storia di un’idea, 2011) and several
articles on Malebranche, Locke, and Leibniz.
Tad M. Schmaltz is Professor of Philosophy at the University of Michigan,
Ann Arbor. He has published articles and book chapters on various topics in
early modern philosophy, and is the author of Malebranche’s Theory of the Soul
(Oxford University Press, 1996), Radical Cartesianism (Cambridge University
Press, 2002), Descartes on Causation (Oxford University Press, 2008) and Early
Modern Cartesianisms (Oxford University Press, 2017).
Tom Stoneham is Professor of Philosophy and Dean of the Graduate Research
School at the University of York. He is the author of Berkeley’s World (Oxford
University Press, 2002) and has edited Causation and Modern Philosophy (with
Keith Allen, Routledge, 2011) and Locke and Leibniz on Substance (with Paul
Lodge, Routledge, 2015). He also writes on analytic metaphysics and mind, and
was Associate Editor of Mind 2005–2015.
Alberto Vanzo is an independent scholar. He works on Kant’s philosophy, early
modern natural philosophy, and the history and methodology of philosophical
historiography.
1 INTRODUCTION TO UNIVERSALS
IN MODERN PHILOSOPHY

Stefano Di Bella and Tad M. Schmaltz

The set of issues debated under the heading of the “problem of univer-
sals” originated in antiquity and received explicit and extensive treat-
ment in the scholastic literature during the medieval era. However, one
could be excused for thinking that these issues had lost their importance
in the modern era. After all, during this period there was widespread
skepticism of scholastic preoccupations, and many modern thinkers
left the topic of universals off their explicit philosophical agendas. Even
so, overt breaks can sometimes conceal deeper continuities, and there
is reason to think that the topic of universals remained important to
modern philosophy, albeit not without undergoing some significant
reconceptualization. In the following introductory remarks, we offer
a preliminary case for this claim and try to disentangle the elements
of continuity and change.1 But first we need to consider some relevant
historical background for the early modern treatment of universals.

1.1. Ancient and Medieval Problems


In its ancient context, the problem of universals has both epistemo-
logical and metaphysical aspects. With regard to epistemology, the
problem is how universal cognition of sensory particulars is possible.
The metaphysical problem concerns the precise ontological status of
the universal features that we grasp by means of such cognition. On a
familiar account, Plato suggested that universal cognition is possible
because such cognition involves a purely intellectual grasp of universal

1. For an excellent introduction to the issue of universals in early modern philos-


ophy, highly sensitive to both continuity and changes in the framework of the
problem, see Bolton 1998.
2   •  The Problem of Univer sals in Early Modern philosophy

Ideas or Forms that subsist apart from both intellects and sensible objects. In
contrast, his student Aristotle held that we could explain universal cognition by
appealing to a process of intellectual abstraction from what is received in sensa-
tion from sensible objects. However, Aristotle also suggested that our universal
cognition has an extramental basis in the natures of sensible objects. Though he
rejected a radical form of realism that posits universals as Platonic Ideas, he none-
theless accepted a kind of “moderate realism” that takes universals to have a kind
of existence in the natural world.
The classic source for the discussion of the problem of universals in the medi-
eval period is the Isagoge of Porphyry (ca. 234–​305). Porphyry belonged to the
school of later Platonism, but his work is a commentary on Aristotle’s Categories,
and he expresses the problem in terms of an Aristotelian framework. In particu-
lar, universals are described in terms of the Aristotelian predicables of genus, spe-
cies, difference, property, and accident, which are organized in the hierarchical
structure of the so-​called Tree of Porphyry. In his text, Porphyry brackets the
question of whether the members of this tree exist in sensibles, as separated from
sensibles, or in thought alone. However, Boethius (ca. 480–​524) later did address
these sorts of ontological questions in his commentary on the Porphyrian text.
There he offers the Aristotelian proposal that though universals subsist in sen-
sibles, they nonetheless can be distinguished from sensibles only in the intellect.
This version of moderate realism became the dominant position after the
rediscovery of the Aristotelian corpus in the thirteenth century. This view was
typically expressed in connection with a psychological account of universal
thought in terms of an abstraction of “intelligible species” from “sensible species”
received from external objects via sensation.
However, there also was a source for a revised Platonist account of universals
in the work of the Church Father St. Augustine (354–​430). In later Platonism,
Forms were placed in an Intellect that constitutes the first emanation from
the One, the source of all being. Augustine “Christianized” this view by iden-
tifying this sort of Intellect with the Word, the second person of the Trinity,
which serves as the realm of divine Ideas. Though the proper interpretation of
Augustine’s views was a matter of considerable controversy during the medieval
period and after, one can nonetheless find in his writings a basis for the view that
our cognition of universal and necessary features of the world derives not from
our abstraction from the contents of our sensory experience but rather from a
divine “illumination” of our intellect that has its source in God’s own Ideas.
Given the authority of Augustine, it was important not to directly contra-
dict his views. Nonetheless, there was an Aristotelian attempt to “naturalize”
Augustinian illumination. We see this sort of naturalization, for instance, in the
work of Thomas Aquinas (1225–​1274). Thomas proposes that God illumines our
Introduction to Universals in Modern Philosophy  •  3

intellect not by directly operating on it, but rather by providing the mind with
an “agent intellect” that is able to abstract out universal features that correspond
to what is contained in divine Ideas. In this way, one can admit the existence
of divine Ideas (as everyone had to in the post-​Augustinian Christian tradi-
tion) without holding that our universal cognition involves any direct access to
these ideas.
In the later medieval period, however, the greatest challenge to the Aristotelian
consensus came not from an Augustinian form of Platonism, but rather from a
“nominalism” that seeks to explicate universality in terms of the signification of
words, without any commitment to “universal” entities. The nominalist via mod-
erna, as represented for instance in the work of William of Ockham (ca. 1280–​
1350), was contrasted with a more realist via antiqua that remained committed
to the position that universals have a basis in concrete sensible objects. Ockham
does admit that our universal terms correspond to universal concepts. However,
he insists that as in the case of all existing things, such concepts are fully particular
mental acts, and he explains the universality of such acts in terms of the fact that
they indifferently signify several different particulars. The universality is there-
fore in the signification, and not in the content of the act.
The distinction between the via moderna and the via antiqua is somewhat
obscured by the fact that both sides subscribed to a robust ontological partic-
ularism. There was the common assumption in the later medieval period that
only individual substances and their properties exist, with universal knowledge
depending in some way on the mental activity of abstraction and generalization.
What remains is a more subtle distinction based on the answer to the question
of whether there is any basis for universality apart from our acts of signification.
According to realist proponents of the via antiqua, there must be some “common
nature” within particulars that provides the truthmaker for our true universal
cognition. However, it was this feature of the world that the nominalist propo-
nents of the via moderna sought to eliminate.

1.2. The Modern Turn
Both sides of the later medieval debate between the two viae remained committed
to some version of Aristotelian hylomorphism as well as to a broadly Aristotelian
theory of cognition. During the so-​called Scientific Revolution, however, the
former gave way to a new mechanical philosophy. This new theory replaces the
appeal in explanations of natural change to substantial and accidental forms
that unite with or inhere in “primary matter” with the appeal to the size, shape,
and motion of material parts. Moreover, during this period the old Aristotelian
theory of cognition gave way to a new “way of ideas” (to borrow Thomas Reid’s
4   •  The Problem of Univer sals in Early Modern philosophy

famous characterization of the modern view). According to this new position,


the direct objects of our knowledge are ideas that are sharply distinguished from
features of external objects.
The modern turn made it impossible simply to continue to debate the status
of universals in scholastic terms. With the emergence of the mechanical philoso-
phy, for instance, one could no longer speak of universals as immanent in sensible
objects, at least insofar as those universals were conceived in terms of Aristotelian
hylomorphism. Even so, there were ways of reconceiving the issue of universals in
modern terms. One option here, reflected in Descartes’s philosophy, involved the
grounding of a mechanistic conception of nature in “true and immutable” math-
ematical natures. Another option, however, is Locke’s view that the mechanistic
conception of nature supports an account of our general concepts according to
which they depend on “nominal essences” that we construct rather than on the
“real essences” that we discover in nature.
Likewise, the emergence of a new way of ideas did not so much eliminate the
issue of universals as shift the focus in a consideration of this issue away from sen-
sible objects or language and toward the contents of our thoughts. Even with this
shift, there remained questions concerning the possibility of abstracting universal
concepts from ideas of sensory particulars, as well as questions concerning the
possibility of and conditions for a universal and necessary knowledge of nature.

1.3. Modern Universals: Mind and Ideas


The standard way of treating universals during the early modern period—​in line
with the new conception of cognition—​was to consider them only as “ideas” or
“concepts.” Given this fact, it would be tempting to classify the modern view in
general as “conceptualist.” Yet there is an important distinction within the mod-
ern conceptualist camp between those who accepted a form of conceptualism
that utilizes the Aristotelian notion of abstraction from sensation, on the one
hand, and those who insisted that our universal cognition derives from purely
intellectual innate ideas, on the other. Moreover, an intellectualist form of con-
ceptualism must sometimes be distinguished from the Platonist position that
our intellectual understanding has its source in universal ideas in the divine
mind. Finally, there is a relevant distinction in early modern thought between
an abstractionist form of conceptualism and a stricter form of nominalism that
rejects the existence of genuine universality even in thought.

1.3.1. Modern Conceptualism
There were different versions of the new way of ideas that had different conse-
quences for the issue of universals. In the form of conceptualism that we find in
Introduction to Universals in Modern Philosophy  •  5

Locke, universality is a feature of general ideas that we form by abstracting from


the particular ideas we receive directly from sense experience. The former are held
to have a content intrinsically different from the latter. In particular, the ability
of general ideas to represent a plurality of objects is given by their indeterminate
character, which is the product of a mental act of abstraction. The mind, after
comparing several complex particular ideas, is able to single out and generalize a
single aspect (or set of aspects) in which they agree.2 For Locke, there is no foun-
dation in reality for our general ideas beyond a primitive sort of resemblance that
particular sensory ideas bear to each other.
We have the basis for a very different form of modern conceptualism in
Descartes. For Descartes insists that universal cognition depends not on abstrac-
tion from particular sensory ideas, but rather on the deployment of purely intel-
lectual ideas that are “innate” as opposed to “adventitious” or “factitious.” We do
not receive or construct these ideas, but rather find them ready-​made (at least
potentially) within our own mind. Moreover, in place of a Lockean antireal-
ism that denies any further basis for resemblances among particulars, Descartes
offers his famous appeal in the Fifth Meditation to universal “true and immu-
table natures” that are “not made by me or depend on my mind.” Though there
is some suggestion in Descartes that at least some of these “natures” can simply
be identified with our innate intellectual ideas, nonetheless such ideas provide a
basis for universal cognition that is itself antecedent to our perception of resem-
blances among particular sensory objects. In this way, Descartes offers a more
thoroughly conceptualist form of realism about universals than we find in the
medieval period, since we have seen that medieval moderate realists tended to
place common natures in the sensory objects themselves.

1.3.2. Modern Platonism
It is perhaps tempting to understand Descartes’s appeal to true and immutable
natures in terms of the Platonist view in Augustine that uncreated ideas in the
divine intellect provide the source for our universal cognition. However, it is
clear that Descartes himself could not accept this form of Platonism given his
doctrine that immutable natures and eternal truths concerning them derive from
God’s free and indifferent will. During the early modern period, however, these

2. It is important to note, however, that the precise nature of Locke’s account of abstraction is
a matter of some dispute. The reading offered here is most in line with a traditional reading of
Locke, assumed by Berkeley, on which he embraced a robust view of abstraction as separation.
However, others have insisted that Locke conceived of abstraction as selective attention to an
idea with a fully particular content. In this volume Stoneham (chapter 12) and Baxter (chapter
13) take Locke to identify abstraction with separation, whereas Bolton (chapter 10) suggests
that he identified it rather with selective attention.
6   •  The Problem of Univer sals in Early Modern philosophy

natures and truths typically were founded in the divine understanding in a more
Augustinian manner. Interestingly enough, some authors, while endorsing this
typical brand of Christianized Platonism, combine it with the acceptance of anti-
realist tenets in the analysis of concrete objects and of human cognition. Already
in medieval philosophy, the recourse to divine understanding as the locus for
Platonic ideas allowed even nominalistically minded authors to avail themselves
of some resources of Platonism. Moreover, in early modern thought the recogni-
tion of the archetypal role of ideas is sometimes accompanied by a particular-
ist ontology and a conceptualist account of human cognition. This double-​level
account can be found, for instance, in the English Platonists as well as in Leibniz.
Within this context, Leibniz advances an interesting terminological distinc-
tion: he prefers to reserve the label of “ideas” for the intelligible contents, origi-
nally present in God’s intellect and derivatively implanted in us, while qualifying
as “concepts” our ways of grasping these contents. This décalage between idea
and concept is reinforced also by his readiness to accept the lesson of Hobbes
(considered further below) concerning the constitutive role of signs for human
thought. Given the essential connection of our universal ideas to general words,
the former must be distinguished from the contents of God’s own understand-
ing. Thus, we are faced in Leibniz with an interesting attempt to combine themes
from radical nominalism with Platonic intuitions into a unitary account of uni-
versal cognition.
A distinctive interpretation of Augustinian themes, within a Cartesian con-
text, can be found in Malebranche’s philosophy, where “ideas in God” become
the direct object of our intellectual knowledge. In Malebranche’s view, our abil-
ity to generalize is not explained on the basis of the traditional abstraction from
particulars. Rather, generalization involves the application of the supreme intel-
lectual idea of infinity, identified with the idea of “being in general.”

1.3.3. Modern Nominalism
In the early modern period, Hobbes is perhaps the first one to explicitly endorse
a “nominalist” stance in a strict and literal sense, apparently exclusive of univer-
sal concepts. In a well-​known passage, he says: “The word ‘universal’ is never the
name of any thing existent in nature, nor of any idea or phantasm formed in the
mind, but always the name of some word or name.”3 To be sure, this stance is far
from univocal even within the Hobbesian corpus. In the same works, in fact, such
strong nominalist statements are accompanied by others that recognize a form

3. OL 1:20; emphasis added.


Introduction to Universals in Modern Philosophy  •  7

of nonlinguistic “internal speech” composed of concepts or images. Even when


admitting this sort of internal speech, however, Hobbes continues to insist that
only the intervention of linguistic signs allows us to gain access to generality.
A different form of modern nominalism is reflected in Hume’s acceptance
of Berkeley’s critique of a central thesis of Locke’s conceptualism: namely, that
general ideas have a distinctive content that is shaped by abstraction. Though
Berkeley and Hume continue to admit that generality is a property of some of
our ideas, they nonetheless deny that general ideas differ from particular ones in
virtue of their content. In their view, a general idea is general simply in virtue of
serving as a sign of a class of similar things, or more precisely, of ideas, given that
we are moving within the new way of ideas.
The representative role of ideas combines here with the linguistic model of
“standing for,” already exploited in the medieval theories of suppositio. As a matter
of fact, such a linguistic reading of concepts had already permitted Ockham to
give an account of the reference of concepts to the world without involving him
in problems deriving from the theory of abstraction. One could say that in the
anti-​abstractionist theories of Berkeley and Hume the functional homogeneity
of concept and word—​vividly present in Ockham’s approach, but largely lost by
the start of the modern era—​was somehow rediscovered.

1.4. Modern Universals: Classification and Science


In the Aristotelian paradigm, science was necessary and universal knowledge; as
such, it was held to be essentially about universals. Moreover, in the classification
of the relevant universals, a privileged role was assigned to substantial universals
and their hierarchical organization in genera and species. More precisely, in the
Aristotelian framework the key role was played by the lowest species, roughly
corresponding to what are nowadays called ‘natural kinds’. Already during the
Renaissance, however, there were serious challenges to the claim that this frame-
work carves nature at its joints. One line of objection was that the Aristotelian
framework is merely a projection of our logical classifications onto the luxuriant
variety of natural reality.
The breakdown of this whole framework and of the assumptions underlying it
threatened both common classificatory concepts for natural beings and the very
conception of science as necessary and universal. One option during the early
modern period was to couple the Aristotelian view that science is founded on
knowledge of universal essences with an anti-​Aristotelian account of the nature
of those essences. We see this option, for instance, in Descartes, who took the
new mathematical science of nature to require a new kind of mathematical
essentialism. Thus Descartes contrasts the knowledge of the “true nature” of a
8   •  The Problem of Univer sals in Early Modern philosophy

mathematical object, such as a triangle, with the confused notion we possess of a


natural species, such as a lion.4 There is a similar sort of contrast in Spinoza, who
distinguishes universals such as “horse” or “man,” which are confused products of
the imagination, from “common notions” that pertain to the attribute of exten-
sion. Thus, the main shift in the modern consideration of universals may concern
the view of what is to count as a true universal. Traditional universals are to be
replaced by notions more in line with a mathematical understanding of material
nature. Such notions are “true and adequate” ideas that capture reality as it is in
itself. While such notions do play some of the roles of the ancient universals, the
former nonetheless are sharply distinguished from the later by modern authors
such as Descartes and Spinoza.
Hobbes offered a different way of reconfiguring the Aristotelian ideal of
science. Rather than appealing to an alternative set of real essences, Hobbes
grounded his universal and necessary science in purely conventional definitions.
In his discussion with Hobbes, Descartes simply could not understand how it
is possible to think of our judgments as connections of names, rather than of
ideas. Among the supporters of an antiempiricist understanding of the way of
ideas, perhaps Leibniz alone was sensitive to Hobbes’s challenge. This sensitivity
was made possible by the fact that Leibniz shared Hobbes’s conviction regard-
ing the constitutive role of signs for human thought. At the same time, though,
Leibniz was eager to avoid the conventionalist consequences Hobbes draws from
this conviction.
Another option during the early modern period was simply to abandon
the Aristotelian ideal of science, at least as far as knowledge of the natural
world is concerned. This was the option chosen by empiricist-​minded thinkers
such as Gassendi and, later, Locke. In particular, Locke focuses on the pecu-
liar role of substantial concepts (in his jargon, of “sortal terms”), but he inter-
prets them in a descriptivist manner, as “bundles” of simple ideas that we take
directly from experience. Thus, our terms for natural kinds are tied to “nomi-
nal essences” that we construct on the basis of general ideas that derive from
sensation, rather than to “real essences” that we find in the objects themselves.
Locke is eager to emphasize the dependence of our classificatory scheme on
the “workmanship of understanding,” and its consequent relativity to our cog-
nitions and interests.

4. See First Replies, AT 7:117. In the Fifth Replies, he rejects Gassendi’s criticism by contrasting
his own view about essences with the “universals of the logicians”: “What you are opposing
to the universals of the logicians [universalia Dialecticorum] does not concern me, because
I understand them in a quite different manner. As far as the essences are concerned, however,
these are known in a clear and distinct way, such as the essence of the triangle” (AT 7:380).
Introduction to Universals in Modern Philosophy  •  9

The rejection of the Aristotelian notion of a universal and necessary science of


nature in Locke can finally be contrasted with the reconfiguration of this notion
in Kant’s transcendental philosophy. As is well known, the Kantian framework of
a priori knowledge, embracing both pure intuitions and pure concepts, is located
at a deeper level than our ordinary empirical concepts. The genesis and working
of the latter continue to be explained by the traditional procedures of abstraction
and generalization applied to sensation. Their import is empirical, while their
formal treatment is a task of ordinary logic. However, the true secret of univer-
sality and necessity is to be discovered at the deeper level of transcendental con-
stitution. Whereas previous modern thinkers required a translation of talk about
universal essences into the language of ideas or notions, Kant requires a new tran-
scendental understanding of the basis of our universal cognition of nature.

1.5. Chapter Summaries
The contributions to this volume serve to further develop some features of the
argument we have sketched for the importance of the issue of universals in early
modern philosophy. Chapters 2 to 4 concern early modern views that reflect
in varying degrees an “antirealist” perspective on universals. Antonia LoLordo
focuses on the main features and the possible motivations of Gassendi’s antirealist
polemics: a historically influential one, although its real target is not so appar-
ent. LoLordo also draws our attention to some interesting tensions internal to
Gassendi’s account of general knowledge, bearing on his critical confrontation
with the Cartesian philosophy of mind, on one hand, and his will to preserve a
view compatible with the Christian faith, on the other.
Stewart Duncan considers Hobbes’s “ultranominalism” (as distinct from more
common conceptualist versions of antirealism), testing its capacity to face some
main objections, both historical and conceptual. Duncan emphasizes in particu-
lar the need to distinguish the issue of the semantics of kind terms, which Hobbes
attempted to explain in terms of a sort of primitive similarity relation, from the
issue—​commonly raised in contemporary discussions of Hobbes’s nominalism,
but of less interest to Hobbes himself—​of the universal nature of qualities.
It may seem that Samuel Newlands’s chapter on Spinoza is misplaced since it
is out of chronological order, coming as it does before the chapter on Descartes.
Although the influence of Descartes on Spinoza can be taken for granted, the
focus in this chapter is on nominalist aspects of Spinoza's criticism of universal
concepts. For this reason it is grouped with the discussions of nominalism in
Gassendi and Hobbes. Newlands emphasizes the strategic import that Spinoza’s
critique of universal concepts has for his treatment of an array of metaphysical,
psychological and ethical issues. Moreover, Newlands shows that in Spinoza’s
1 0   •  The Problem of Univer sals in Early Modern philosophy

case there is a puzzling general tendency to appeal to universality (in the case of
common notions and essences), while at the same time severely criticizing tradi-
tional universal notions.
Chapters 5 to 7 address issues concerning conceptualism and Platonism in the
work of Descartes and his French followers. Lawrence Nolan defends an updated
version of a “conceptualist” reading of Descartes that he has offered in previous
work. Nolan explicitly faces a sophisticated version of a competing “Platonist”
reading, finally dismissing it by appealing to Descartes’s strong view of divine sim-
plicity and his radical doctrine of the creation of eternal truths.
This radical doctrine is also central to Tad Schmaltz’s exploration of some
lesser-​known debates within seventeenth-​century French Cartesianism. Antoine
Arnauld’s “conceptualism,” on the one hand, and the highly original ontology of
Robert Desgabets and Pierre-​Sylvain Regis, on the other, provide two possible
answers to the dilemma posed by the alleged eternity of truths. These views consti-
tute Cartesian alternatives to an Augustinian form of Platonism in Malebranche
that is itself premised on the rejection of created eternal truths and essences.
Mariangela Priarolo’s chapter brings to a close the exploration of the Cartesian
milieu, focusing on the treatment of universality in Malebranche. She addresses
a classic problem in Malebranche scholarship, namely, how to account for our
(and God’s) knowledge of particular things within a framework of his theory
of our vision of universal ideas in God. Her discussion of this problem relates
Malebranche’s views to those of his scholastic predecessors. Moreover, Priarolo
highlights the distinctive connection in Malebranche of general knowledge to
the positive idea of the Infinite.
The issue of universal knowledge in intellectual intuition provides a bridge
to Brunello Lotti’s discussion of English Platonism (chapter 8). In particular,
Lotti explores the treatment of this topic in the work of Henry More, Ralph
Cudworth, and John Norris. We are faced here with a quite different histori-
cal and conceptual background, directly connected to ancient and Renaissance
Neoplatonism. Still, Lotti can reconstruct in these philosophers a multilayered
ontological and epistemological model, according to which a frankly Platonic
view of the “archetypal” world is combined with a more conceptualist account of
human knowledge in terms of abstractive procedures.
A similar delicate blend of Christianized Platonism and conceptualism can be
found also in Leibniz, according to Di Bella’s reconstruction of his stance (chap-
ter 9). Most characteristic of Leibniz, however, is his emphasis on nominalistic
features of his ontology, as well as the persistent influence of Hobbes with respect
to the constitutive role of signs for human thought.
Leibniz turns out to be more sensitive to realistic concerns when he is con-
fronted with Locke’s decidedly antirealist form of conceptualism. His exchange
Introduction to Universals in Modern Philosophy  •  11

with Locke is the focus of Martha Brandt Bolton’s chapter (chapter 10). Whereas
there is an attempt in Locke to relativize genera and species to human ideas and
languages, Leibniz insists on the fact that kind terms have a real basis in the inner
natures of particular objects. Moreover, Leibniz is concerned to defend against
Locke’s attack the traditional view that knowledge of objects in the human mind
is grounded ultimately in God’s understanding of relations among essences.
The consideration of the position in Locke with which Leibniz takes issue
introduces us to a group of chapters devoted to classic British empiricism, where
the issue of universals is perhaps most prominent in the early modern period.
Chapter 11, from the late Jonathan Lowe, addresses Locke’s theory of sortal terms
more in a conceptual than in a purely historical fashion. In particular, Locke’s
theory is considered in terms of a contemporary account of the semantics of natu-
ral kind terms deriving from the work of Kripke and Putnam. Focusing on Locke’s
example of water and ice, Lowe defends the central intuitions of a Lockean-​style
theory against criticisms based on this account.
Berkeley’s argument against abstraction is the theme of Tom Stoneham’s
chapter (chapter 12). Against a settled interpretative tradition, Stoneham tries to
reconsider this argument for its own sake, and not only as part of an antimateri-
alist strategy. Moreover, he emphasizes that Berkeley offers a positive “semiotic”
solution to the epistemic problem of how, on a nominalist view, one can be said
to make universal claims. Though some have claimed that Berkeley’s theory of
signs is strongly formalistic, Stoneham concludes that formal properties in fact
have no fundamental role to play in this theory.
Hume’s view of abstraction is central to Donald Baxter’s chapter (chapter 13).
His analysis of Hume’s arguments aims to show that Hume’s account of the idea
of identity stands or falls with a kind of abstractionism that Hume himself was
committed to rejecting. This is so because Hume’s rejection of abstraction is
based on his argument that the mind cannot separate what is inseparable in real-
ity, whereas his account of identity requires that we can represent the same thing
as two things that are distinct, and thus separable. Given the depth of Hume’s
commitment to a Berkeleian critique of abstraction, Baxter’s argument would
seem to show that it is the account of the idea of identity that must go.
The final chapter in this collection is devoted to Kant. Alberto Vanzo focuses
on a less studied aspect of Kant’s thought, that is to say, his account of empiri-
cal concepts. An accurate comparative study of Kant’s logical courses allows one
to reconstruct a quite traditional view of the psychological process of concept
formation. Vanzo confronts this account with some present-​day objections
raised against abstractionist theories in general, and shows how a Kantian-​style
theory could have the resources to counter them. Interestingly enough, in so
doing, Vanzo comes to defend the view that there is a preconceptual stage to our
1 2   •  The Problem of Univer sals in Early Modern philosophy

cognitive activity, which might be appear at odds with other important tenets
of Kant’s theory of knowledge. This would invite a further consideration of the
relation of this layer of empirical concepts to the underlying deeper constitutive
activity that is the object of transcendental analytics.
As this final comment intimates, these chapters do not exhaust the topic indi-
cated by the title of this volume. Rather, they serve to indicate the persistent and
multifaceted presence of the issue of universals in early modern philosophy. They
therefore serve to illustrate that the ancient topic of universals was reconfigured
rather than merely rejected during the early modern period.

Abbreviations
AT = Descartes 1964–​74; cited by volume and page.
OL = Hobbes 1966; cited by volume and page.
2 GASSENDI ON THE PROBLEM
OF UNIVERSALS

Antonia LoLordo

2.1
There are many problems of universals, but the most basic problem is
the metaphysical one. Are there elements of reality that can be multi-
ply instantiated? If so, what are these universal elements and how do
they relate to particulars? Gassendi does not see this problem as wor-
thy of consideration. It’s simply obvious, he thinks, that every exist-
ing thing is particular. However, he does recognize the existence of
universal concepts—​“general ideas” that apply to many things despite
being in themselves particular.1 Since he thinks that all ideas are com-
posed of the materials we acquire through the senses, he thus owes
us an account of the formation of general ideas. And since he thinks
that our general ideas track sharply delineated natural kinds, he must
explain how they succeed in doing so and what kind membership
consists in.

2.2
Gassendi insisted on nominalism throughout his career. I  won’t say
that he argued for nominalism throughout his career, since he never
offers much argument for it. Of course, he says a great deal about

1.  Gassendi also recognizes the existence of general terms. But although at one
point he misleadingly insists that nihil esse aliud grandia haec universalia, quam
quae Grammatici vocant nomina Appellativa (Exercitationes 2.2.3; OO 3:159a),
language plays almost no role in his discussion. For general terms do no significant
explanatory work and require no special explanation: they derive their intention-
ality from the intentionality of general ideas, in exactly the same way particular
terms derive their intentionality from particular ideas.
1 4   •  The Problem of Univer sals in Early Modern philosophy

universals, especially in the early Exercitationes … adversus Aristoteleos. But what


he offers is polemic, not argument.
Gassendi’s main tactic is portraying realism about universals as closely
entwined with Aristotelianism. Then, by rejecting both, he positions himself as
one of the “new philosophers” offering a radical, nominalist alternative to the
realist, Aristotelian orthodoxy of the day:

What, you will ask? Do you then accept the crazy opinion of the nomi-
nalists who recognize no universality except the universality of concepts
or names? So it is; I accept this, but I hold that I accept a completely sane
opinion. (Exercitationes 2.2.3; OO 3:159a)2

For, Gassendi insists, “God is most particular and all his works are particulars”3
(Exercitationes 2.2.3; OO 3:159a).
This seems like an odd way to proceed. Nominalism was hardly an obscure or
radical view at the beginning of the seventeenth century. Gassendi’s contempo-
raries, whether Aristotelians or not, were almost all nominalists of one kind or
another: there’s no necessary connection between Aristotelianism and realism,
or even a natural association. Gassendi—​who held the chair in philosophy at the
University of Aix while writing the Exercitationes—​was perfectly well aware of
this. So why didn’t he engage with the more sophisticated nominalist views cur-
rent in the early seventeenth century, like Suárez’s?
One could ask a parallel question about many of Gassendi’s discussions.4 In the
Syntagma, he spends a lot of time analyzing and arguing against the views of vari-
ous ancients: not just the Epicureans and their critics, but also various Platonists
and Stoics. In contrast, he says very little about Aristotelian views. Sometimes
he tells us what Aristotle himself said, but he almost never discusses other scho-
lastic philosophers, and certainly no seventeenth-​century scholastics. If you got
your knowledge of the history of philosophy just from reading the Syntagma, you
wouldn’t even realize there was an Aristotelian tradition. This is, of course, delib-
erate. It implies that the Aristotelian tradition doesn’t merit rebuttal: we should

2. Quid? Inquies, accedes ergo ad vesanam illam opinionem Nominalium, qui universalitatem
aliam non agnoscunt, quam conceptuum aut nominum? Ita sane est; accedo, sed puto me accedere
ad opinionem admodum sanam (OO 3:159a).
All translations are my own, but I have benefited greatly from consulting existing translations
in CSM and Gassendi 1959, 1962, 1972, and 1981.
3. singularissimus est Deus, singularia omnia eius opera (OO 3:159a).
4. Cf. Bloch 1971, 115.
Gassendi on the Problem of Universals  •  15

simply reject it and start over. For Gassendi, this means starting over on the basis
of a different ancient model, but the basic move is hardly unique to him.
Thus, although associating realism with Aristotelianism is an important rhe-
torical strategy for Gassendi, it cannot be what really motivates his nominalism.
So, what is the underlying motivation? Here are three possibilities, which are not
mutually exclusive.
First, perhaps Gassendi’s nominalism is motivated by an intuition. Perhaps,
that is, he simply finds it incomprehensible that anything could be multiply
instantiated. In the background here is what Tom Lennon calls “the localization
pattern”: the assumption that everything that is, is in some place at some time.5
Consider passages like the following:

What is this nature that is in you and me at the same time? Perhaps if
you were outside a wall and I  were inside it, the wall would not fully
divide whatever is between your nature and mine? If someone sliced the
air between you and me with a sword, they would not be able to divide
this whole common thing? Indeed, I would like to know how you con-
ceive of a common thing of this sort. Like a vault that encloses us both?
Like a gas that surrounds or penetrates us both? Now, really, how do you
conceive it? You say, and with general approval, that human nature is in
many things even if nobody is thinking of it? And isn’t what is really in
many things really universal? I grant that human nature is in many things,
even if nobody is thinking of it, but I add that it is multiple. You wished
to conclude that it is one, in order to establish a universal, but I say that
it is multiple, in order to preserve particulars. (Exercitationes 2.2.3; OO
3:159a–​b)6

For how could a sword divide a universal unless it’s located in space and time?7

5. Lennon 1993, 118, citing, e.g., OO 1:182a.


6. Ecquaenam tandem est haec natura, quae in te simul, ac meipso sit? Nisi forte sit tu ultra, ego
citra parietem simus, paries abunde non discernat quidquid inter tuam, meamque naturam est, si
quispiam autem te inter & me ensem duxerit, abscindere non valeat totam hanc communitatem?
Certe enim concipere velim quomodo tu hanc huiusmodi communitatem concipias? An ut fornicem
ambos includentem, an ut aerem utrumque vel ambientem, vel penetrantem, quo enim tandem
modo concipis? Dicis & magno quidem cum applausu; nonne nemine cogitante natura humana
est in multis? Quae revera autem est in multis, nonne revera universalis est? Ego fateor quidem
naturam humanam nemine cogitatante in multies, sed adiicio multiplicem. Tu volebas inferre
unam, ut statueres universalem, at ego multiplicem dico, ut singulares tuear (OO 3 :159a–​b).
7. Here it looks like Gassendi has something a bit stronger than localization in mind: material-
ity. In his physics, absolute space separates us, but space can’t be cut with a sword. However, it’s
1 6   •  The Problem of Univer sals in Early Modern philosophy

Notice that this argument assumes that universals exist in bodies, not just
in the mind. As a result, it doesn’t affect the “moderate realist” who holds that
forms as they exist in the intellect are universal but that forms as they exist in
bodies are particular, because they are particularized by the matter they inform.
However, although there’s no necessary connection between realism in general
and Aristotelianism, this moderate realist view is hard to reconstruct without
hylomorphism. So, Gassendi may well think that once we have rejected the met-
aphysics of form and matter, we have ruled out this kind of moderate realism.
A second possibility is that Gassendi rejects realism about universals because
he fears that realism places unacceptable constraints on God’s creative power.
Margaret Osler—​who argues that “[v]‌oluntarism was the unifying thread that
bound [Gassendi’s] natural philosophy, theory of knowledge, and ethics into a
coherent whole” (1994, 48)—​emphasizes this line of explanation:

Gassendi believed that God’s absolute power is in no way constrained by


the creation, which contains no necessary relations that might limit God’s
power … there are no essences in the world. There are no necessary con-
nections linking fire and heat or whiteness and snow. (Osler 1994, 53)8

Gassendi does argue that universals independent of God are theologically unac-
ceptable in both the Exercitationes and the Disquisitio. (We’ll see these arguments
in section 2.5.) However, I find it implausible that such theological worries are
what motivates Gassendi’s nominalism. Allowing universals to be entirely inde-
pendent of God and denying their existence altogether are hardly the only two
options. And although the theological worries explain why Gassendi chose nomi-
nalism over realism, they do not help explain why he chose the particular form of
nominalism he did.
A third possibility is that Gassendi rejects the existence of universals out of
a preference for simplicity. Several different variants of this are suggested in dif-
ferent passages. Perhaps universals are unhelpful because even if they exist, they
could not be grasped. Perhaps even if we grasped them, they would not help us
explain anything. Or perhaps universals are unnecessary because although they
have explanatory power, there are simpler ways to explain whatever it is that needs
explaining. I do not claim that simplicity considerations, in any of these variants,
provide a complete explanation of Gassendi’s nominalism. Nor do I claim that

not clear that Gassendi accepts the general principle that every existing thing is material even
in the Exercitationes.
8. Osler is just talking about universal natures: she agrees that Gassendi accepts that particulars
have natures (Osler 1994, 115).
Gassendi on the Problem of Universals  •  17

they provide the only interesting explanation. However, I will focus on these con-
siderations because they help us see why Gassendi adopted his particular form of
nominalism.

2.3
Gassendi’s discussion of universals is almost entirely about one particular subcat-
egory of universals: the natures of species or natural kinds. The same is true of the
theory of general ideas he introduces to take the place of a theory of universals.
Thus, we can simplify matters by just looking at various ways species natures are
used in explanation.
In the tradition, species natures are sometimes thought to explain kind
membership by being constituents of particular beings. (So, Socrates and Plato
are both men because they both have the nature man.) They are sometimes
thought to explain how we can conceive of a kind and not just the particular
members of the kind we have experienced. (We conceive of mankind by using
our idea of the nature man.) They may explain why there are sharp boundaries
between species and how we track those boundaries. (Something either has the
nature man or not, so no vague cases are possible, and we track the boundary
between men and monkeys by grasping the nature man.) They may serve as
truthmakers for the eternal truths. (Man is an animal is eternally true because
the nature man exists eternally.) And they are the patterns God used to create
the world.
Gassendi thinks most of these things can be explained without universals,
appealing only to particulars and the brute similarities between them. Socrates
and Plato are members of the same kind because they each have a particular
nature, and there is a brute similarity between their particular natures. We can
think about men we have never met because we have a general idea of man that
we constructed out of the particular ideas of men we acquired in sensation. There
are sharp species boundaries because of the way these brute similarities cluster.
And we can track these boundaries because our ideas of particulars, like the par-
ticulars themselves, are similar.
He deals with the last two things in a slightly different way, by simply deny-
ing that there is anything that needs explanation. There are no eternal truths and
hence no need for them to have truthmakers. And God is perfectly capable of
creating the world directly, without relying on a pattern.
The details of Gassendi’s explanations evolved over time. Thus, I’ll look at the
views expressed in four different texts written at different points in his career:

(1) The early Exercitationes … adversus Aristoteleos.


1 8   •  The Problem of Univer sals in Early Modern philosophy

This is mainly a critical work, but one can elicit a sketchy positive view from it.

(2) The Objections and Counter-​Objections to Descartes’s Meditations, which


were published together as the Disquisitio Metaphysica.

This is also mainly negative, and we have to tease out the positive view.

(3) The Institutio Logica, a section of the Syntagma’s Logic intended to be a


textbook.

Here, Gassendi distinguishes two types of general ideas:  aggregate ideas and
abstract ideas. The aggregate idea of a man is just a collection of ideas of particular
men, while the abstract idea of a man is the idea of a man with all irrelevant deter-
minations set to one side. The distinction itself is straightforward, but it raises
a number of questions. What is the evidence that we have both types of general
ideas? What is the point of having abstract ideas as well as aggregate ideas? And
what cognitive faculties are required to form abstract ideas?

(4) The discussion of the immateriality of the soul in Syntagma 2.3b.9.9

Here, Gassendi presents an argument that might surprise readers of his earlier work.
He argues that the soul must be immaterial because it can grasp universal notions
and their ratio universalitatis. Hence there must be an immaterial intellect along
with the material imagination. This is surprising because Gassendi argues against
the distinction between intellect and imagination at some length in the Disquisitio.
Like the distinction between abstract and aggregate ideas, the distinction
between intellect and imagination raises a number of questions. I’ll focus on how
the two distinctions relate. Does Gassendi think that the intellect uses abstract ideas
while the imagination uses aggregate ideas? Or does he have something else in mind?

2.4
Even in his most skeptical work, the Exercitationes, Gassendi accepts that particu-
lar things have natures:

You say, for example, that there is a human nature which is universal. But
where is this universal nature seen? I do indeed see this human nature of

9. That is, book 9 (“On the Intellect, or Mind”) of the second “member” of section 3 of part 2
(Physics). Bloch 1971, 140, shows that this was written significantly earlier than the Institutio Logica.
Gassendi on the Problem of Universals  •  19

Plato, that human nature of Socrates, but these are all particular natures.
(Exercitationes 2.2.3; OO 3:159a)10

Gassendi later glosses a particular nature as “the principium of all properties,


faculties, and operations” (Disquisitio 2.6.3; OO 3:306a).11 But what this turns
out to be is something material: the inner texture of corpuscles that determines
the properties and behavior of a macroscopic body (Disquisitio 2.8; OO 3:311a).
Thus, in principle, the particular natures of bodies are visible in the same way
bodies are, because they are bodies.
Various particular natures are more or less similar to each other:

When it is said that all men are of the same nature … the sense of
this can only be … that each has a nature that is similar to the others.
(Exercitationes 2.2.4; OO 3:159b)12
… it is commonly said that these natures are one … because they are
conceived in one and the same way on account of their similarity and can
be expressed by one and the same concept. But isn’t the similarity founded
in unity? So indeed says Aristotle; but nevertheless, if you consider, unity
is founded in similarity. (Exercitationes 2.2.5; OO 3:160a)13

Both Gassendi and his opponent agree that the various members of a kind are
similar in nature. What they disagree about is whether this similarity requires
further explanation. His opponent argues that it does, and offers the following
explanation: particular men are similar because they all instantiate the universal
man. In contrast, Gassendi holds that the similarity need not be grounded in
anything, so that universal natures are unnecessary:

If all these particular natures exist, you say, then there must be one nature
common to all of them. But to me, it is enough that I have a particular

10. Dicis exempli causa humanam dari naturam, quae universalis sit. At ubinam visitur universa-
lis haec natura? Ego quidem video hanc naturam humanam Platonis, illam Socratis, at hae omnes
sunt natura singulares (OO 3 :159a).
11. . . . quid intelligis nomine Naturae? An-​non principium omnium proprietatum, facultatum,
operationum? (OO 3 :306a).
12. At, inquies, nonne dicuntur omnes homines eiusdem esse naturae, vel habere eandem naturam?
Quo sensu ergo istae loquutiones, & aliae id genus sint verae, ac admittendae ut paucis accipias
(OO 3:159b).
13. . . . ac una dicitur vulgo natura? Profecto non quod revera una, eademque in illis sit, sed quod
concipiatur per modum unius, eiusdemque propter similitudinem, ac uno eodemque conceptu
exprimi valeat; at nunquid similitudo fundatur in unitate? Ita quidem Aristoteles; si tamen
attendas potius unitas in similitudine fundatur (OO 3:160a).
2 0   •  The Problem of Univer sals in Early Modern philosophy

nature. … I see nothing that is the same in you and me and common to us
both. (Exercitationes 2.2.3; OO 3:159a)14

Gassendi admits that he risks making a statement like Plato is a man into a tautol-
ogy: on his view, “it will predicate something of itself.”15 However, he responds,

[E]‌very proposition, if it is true, must be identical, because it must predi-


cate nothing of the thing that is not either the same as it, or in it. …
And so, when it is said that Plato is a man, by the first term, namely Plato,
I understand just this particular thing, and by the second, namely man,
I understand the same thing as it has some similarity to Socrates, etc. …
this particular thing, which is conceived according to its particular differ-
ences and called Plato, may also be conceived in a more universal way on
account of its similarity to other things, and hence described by a more
universal term. (Exercitationes 2.2.5; OO 3:160b)16

This does not mean that all true propositions are what we would now call analytic.
To say that Plato is a man is to say that he bears certain similarity relations with
the other things we call men, and it could be the case that we knew of nothing
similar to Plato.17 Thus, general ideas embody statements about the relationship
between various particulars. This leads Gassendi to argue that if we knew the spe-
cies and difference of one thing, we would know everything else in the universe:

In fact, we cannot say that this is the difference between one thing and
another, unless we know that there is something not in the former that

14. Cum dentur, inquis, tot singulares, reperitur ergo una communis in omnibus. Ita-​ne dicis?
Quomodo probas? Mihi quidem satis est, ut habeam unam singularem; & tibi etiam quidquid
dicas singularis una satis est; quod me attinet, ego nullam, quae sit in te, & me eadem, commu-
nisque perspicio (OO 3 :159a).
15. . . . at inquies, ergo nugatoria, & identica erit propositio, cum dicetur Plato est homo; praedi-
cabitur enim idem de seipso (OO 3:160b).
16. Respondeo omnen propositionem, ut sit vera, debere esse identicam, quia scilicet nihil de re
praedicari debet, quod vel eademmet, vel in eadem non sit. … Cum itaque dicitur Plato est homo,
per priorem vocem scilicet Platonis intelligo solum hanc particularem rem, at per posteriorem
nempe hominis, intelligo eandem rem ut habentem quid simile cum Socrate, &c. … haec res par-
ticularis, quae propter suas particulares differentias concipitur, & dicitur Plato, eadem propter
similitudinem cum aliis concipiatur quoque sub universaliore ratione, ac proinde exprimatur uni-
versaliore voce (OO 3 :160b).
17. Cf. Disquisitio 3.2.2 (OO 3:319b): Nam quando dicis Ego sum Res, aut conceptus Rei est sin-
gularis, ut universalis. Si singularis, ergo tui ipsius. … At si tui ipsius, identica est, ut loquuntur,
& nugatoria propositio … Si universalis, ergo includit comparationem tui ad alia, quibus idem
attributum conveniat; idemque est ac si dices, Ego sum certa res, seu una ex numero rerum.
Gassendi on the Problem of Universals  •  21

is found in the latter. But how could we know that something is not in
the former unless we have revealed its inner recesses? It is claimed that
man, for instance, differs somehow from all the other things there are
in the world. You attribute the fact that he is rational as his specific and
sufficient difference … but to sufficiently distinguish man from all the
other animals, you must know all the animals perfectly and intimately.
(Exercitationes 2.5.4; OO 3:184b–​85a)18

And again,

[I]‌f any proposition is universal, it cannot be inferred except by induction,


as is clear from Aristotle … and in any case, it is obvious. And a universal
proposition cannot be inferred by induction because all the particulars
… cannot be looked through and counted … if you wished to infer by
induction that, for instance, every man is an animal, who cannot see that
it is impossible to look through and count all particular men, not only
however many exist now but also however many have existed, will exist
later, and even can exist? (Exercitationes 2.5.5; OO 3:187b–​88a)19

Since we cannot examine every case, including all the merely possible cases, we
cannot know differences by induction. And since even Aristotle admits there is
no other way to know differences, the system of genus and difference is useless.
Although Gassendi keeps many of the positive claims from the Exercitationes
throughout his career, he drops this last one—​not because he comes to doubt
that general ideas embody claims about relations but because he comes to doubt
that induction requires examining every particular. In fact, he comes to reject
skepticism in general in favor of an epistemology where certainty is hard to find
but probability is good enough for most purposes.

18. Dicere certe non possumus hanc esse huius rei differentiam ab illa, nisi in illa cognoscamus
aliquid non esse, quod reperiatur in hac. Quomodo vero cognoscemus non esse aliquid in illa, nisi
omnes ipsius recessus etiam intimos evolverimus? Proponatur homo v.c. is revera aliquid differt
ab omnibus aliis rebus, quae in Mundo sint. Assignas tu illi pro differentia specifica & sufficienti
quod sit rationalis, dicisque hominem per illam differre ab omnibus aliis animalibus. Esto, at ut
sufficienter inde discernatur ab omnibus animalibus debes perfecte atque intime cognoscere omnia
Animalia (OO 3 :184b–​85a).
19. Primum igitur si sit aliqua propositio universalis illa non potest alio modo, quam inductione
colligi, ut patet ex ante citato Aristotelis loco … & res manifesta est. Atqui Inductione colligi non
potest Universalis propositio, siquidem percurri prius & ennumerari non possunt omnia singu-
laria … si velis inductione colligere hanc exampli causa propositionem, omnis homo est animal,
quis non videat factu impossibile ut percurras & enumerares homines omnes singulares non modo
quotquot iam sunt, sed etiam quotquot pridem fuerunt, quotquot erunt in posterum, & quotquot
tandem esse possunt? (OO 3:187b–​88a).
2 2   •  The Problem of Univer sals in Early Modern philosophy

2.5
We can learn a bit more about Gassendi’s positive views by looking at the
Objections and Counter-​Objections to the Meditations. Their general strategy is
much like the Exercitationes. Gassendi begins by portraying his opponent as a
realist (again with rather questionable accuracy), elucidates some unattractive
implications of that kind of realism, and uses this to motivate his own position.
Consider his discussion of the eternal truths, whose real focus (like that of similar
discussions in the Exercitationes) is on the various features that go along with
eternality:

From the fact that propositions are universal they are said to be necessary,
per se, de omni, sempiternal, and so on—​which are no less empty than uni-
versality itself. They are said to be prior, better known, truer, the cause of
conclusions—​but who cannot see that these [labels] apply to particulars
better than to universals? (Exercitationes 2.5.5; OO 3:189a)20

In other words, the eternal truths are just those propositions that can serve as
first principles in a demonstrative science:  propositions that are eternally true,
necessary, universal, and better known by nature than the propositions deduced
from them.21
Like many later readers, Gassendi sees Descartes’s claim that the eter-
nal truths are created as incoherent (Disquisitio 5.1.2; OO 3:377a). But he
sets that problem aside and devotes most of his attention to what Descartes
“speciose, atque Platonice” says about the eternal truths (Disquisitio 5.1.4;
OO 3:378b). This is because he sees the eternal truths as intertwined with
Descartes’s true and immutable natures. And it’s essential to Descartes’s proj-
ect that there are true and immutable natures, but not that the eternal truths
are created.
Gassendi’s critique of the Cartesian doctrine of eternal truths is more or less
the same as his critique of the Aristotelian doctrine of eternal truths. They cannot
exist because the true and immutable natures—​that is, universals—​that they’d
be grounded in do not exist. But why must eternal truths be grounded in uni-
versals in the first place? Why can’t man is an animal, for instance, be made true

20. Ex eo enim quod propositiones sint universales dicuntur esse necessariae, per se, de omni, sem-
piternae, &c. quae quidem vanae non minus sunt, quam universalitas ipsa. Dicuntur esse priores,
notiores, veriores, causae conclusionis, sed quid non videat ista magis singularibus convenire, quam
universalibus? (OO 3:189a).
21. Cf. Exercitationes 2.4.4 (OO 3:178b): res existentes seu particulares non admittunt
Propositiones sempiternae veritatis.
Gassendi on the Problem of Universals  •  23

by the fact that at some time, men exist and are animals? Gassendi answers that
this would give the eternal truths insufficient robustness. His claim is not that
propositions like man is an animal are false: he agrees that men are animals. But
he thinks such propositions have less ontological commitment than is commonly
supposed:

When someone says, while no rose exists, a rose is a flower, this is


like he said a rose is a flower when it exists. Indeed, if this were not
his meaning, the proposition would be false. (Exercitationes 2.4.4;
OO 3:178b)22

Eternal truths, read in this deflationary way, do not require universals and hence
are unobjectionable.
Margaret Osler claims that Gassendi rejected universals for theological rea-
sons. The best evidence for this claim is a dilemma which Gassendi poses in the
Disquisitio. Although Descartes denies “that the essences of things, and the math-
ematical truths which we can know concerning them, are independent of God,”
he insists “that they are immutable and eternal, since the will and decree of God
willed and decreed that they should be so” (CSM 2.261, AT 7:380).23 If universal
natures are independent of God, then—​as Descartes himself recognizes—​they
limit God’s creative power, which is unacceptable. But if they depend on God,
then they cannot be immutable (because God could change them), necessary
(because God could have created different ones), or eternal (because created
things exist in time). The only way out is to avoid the dilemma altogether by
rejecting universal natures:

[I]‌t seems better to recognize that there is nothing beyond God the thrice-​
great except that which is created by him, that which exists, that which is
particular; and to think that things that have not yet been created and do
not exist but are merely possible have no reality or truth. Or, if they are
thought to have some reality and truth, it should be understood as future
reality and truth. And so it is not so much that they actually have reality
(since they do not exist and hence do not have anything) as that they will

22. Nam & cum rosa non existente, dicit quis Rosa est flos, perinde est ac si diceret; Rosa cum existet
est flos: nisi quippe hic fieret sensus, Propositiones falsae fierent (OO 3:178b).
23. Exercitationes 2.4.4 (OO 3:177b) ascribes to the Aristotelians the similar view that eternal
truths result when subject and attribute are connected so tightly ut sit nexus insolubilis etiam
virtute divina.
2 4   •  The Problem of Univer sals in Early Modern philosophy

have reality at some time (since they will exist at some time). (Disquisitio
5.1.2; OO 3:377b)24

Hence, universals should not be admitted

in any other way than from an operation of the intellect, in so far as it, as
it were, educes them from the many particular things it has observed or
understood as possible, and forms certain general notions which can be
attributed to particulars because they were educed from particulars …
these universal natures depend on God, in so far as the particulars from
which they are formed and from which they do not actually differ depend
on God. (Exercitationes 5.1.2; OO 3:377b)25

Thus, there are no universals and, in consequence, no eternal truths.26 And


thus—​since Gassendi sees the project of the Meditations as dependent on true
and immutable natures—​the argument of the Meditations is fatally flawed.

2.6
The Institutio Logica assumes what the Exercitationes and Disquisitio argue for:
everything that exists is particular. It develops a theory of cognition that is compat-
ible with what’s said in earlier works but which goes substantially beyond them.
Gassendi’s project in Part I of the Institutio is to explain how we acquire ideas:

[S]‌ince all the things that exist in the world that can affect the senses are
particulars … it surely must be the case that the ideas that pass from them

24. Quam satius itaque videtur nullam rem veram, praeter Deum ter. Max. agnoscere, nisi ab eo
reipsa creatam, reipsa exsistentem, reipsa singularem: Existimare vero res nondum creatas, neque
exsistenteis, sed duntaxat possibileis, nullam neque realitatem, neque veritatem habere; aut si
quam habere censeantur, intelligere futuram oporteat; adeo proinde, ut revera realitatem non tam
habeant (utpote quae ipse non sint, neque proinde aliquid habeant) quam habiturae aliquando
sint (utpote quando fuerint) (OO 3:377b).
25. Denique res, seu potius Ideas non alia ratione admittere, quam ex operatione intellectus, qua-
tenus ex pluribus rebus singularibus observatis, aut quasi possibilus habitis intellectus quasi educit,
ac format generaleis quasdam notiones, quae ut ex singularibus educuntur, ita singularibus tribui,
ut ipsis convenientes, possunt. Hoc certe modo dici poterunt ipsae res, seu naturae universales esse
dependentes a Deo, quatenus a Deo dependent, singularia, ex quibus formantur, & a quibus reipsa
non differunt … (OO 3:377b).
26. Gassendi notes that logical space allows for eternally (i.e., sempiternally) true propositions
without genuine universals: God could will that particulars always exist. But this is a mere pos-
sibility: in fact, nothing created exists through all time (Disquisitio 5.1.2; OO 3:377b).
Gassendi on the Problem of Universals  •  25

into the mind and adhere there are particulars. (Institutio Logica 1.4;
OO 1:93a)27

He describes two different ways we can make general ideas from the ideas acquired
through the senses. The first produces aggregate ideas:

[T]‌he mind, as it were picking out similar ideas, collects them together
into one aggregate which, containing all of them, is an idea of all taken
collectively and hence is called universal, common, and general. … Such,
for example, is the aggregate of the ideas of Socrates, Plato, Aristotle, and
all other similar things, which, because the name man is appropriate to
the particulars, is usually called the genus man. (Institutio Logica 1.4;
OO 1:93a–​b)28

The second method produces abstract ideas:

Although these particular ideas are similar in some way or agree


with each other, nevertheless they also have many discriminations
by which they differ from each other. Thus the mind—​considering
them separately and thereby as it were abstracting out everything in
which they all agree and removing or not considering everything in
which they differ—​takes what is thus considered, having nothing that
is not common to all, for a common, universal, general idea … . For
example, when the mind notices that the ideas of … Socrates, Plato,
and Aristotle agree with each other and are similar because each one of
them represents an animal with two legs, head erect, reasoning, laugh-
ing, able to be taught, and so on, it as it were abstracts this (namely
being an animal with two legs, head erect, etc.), forms an idea from
which all the discriminations in which they differ (like that … one
is an old man, another a man, the last a boy …) have been removed,

27. Quippe, cum res omnes, quae in Mundo sunt, incurrereque in Sensus possunt, singulares sint
… non possunt profecto ideae, quae ex illis in Mentem transeunt, inque ipsa haerent, singulares
non esse (OO 1:93a).
28. Priore enim modo Mens simileis ideas veluti seponens in unam cogit aggeriem, quae omnes
proinde continens, universarum Idea sit, ac universalis proinde, et communis, generalisque
dicitur. … Talis est, v.c aggeries ex Ideis Socrates, Platonis, Aristotelis caeterorumque omnium
similium, quae ob commune Hominis nomen accomodatum singulis dici solet Hominum Genus
(OO 1:93a–​b).
2 6   •  The Problem of Univer sals in Early Modern philosophy

and takes this idea for the universal or general idea of man. (Institutio
Logica 1.4; OO 1:93b)29

This passage raises two issues. First, what exactly is left after we have removed every
respect in which the particulars differ? Gassendi blithely remarks that

It is difficult, not to say impossible, to imagine a man in general so accurately


that he is neither large nor small nor medium height, neither old nor young
nor intermediate in age, neither pale nor dark nor another particular color.
But it is necessary to at least bear in mind that the man we want to use for
general consideration should be free of all these discriminations. (Institutio
Logica 1.8; OO 1:95b)30

But how, exactly, do we bear this in mind? One possibility is that the abstract idea of
a man is not something we imagine at all—​it’s an idea that is not an image. However,
Gassendi does not seem to be using the term “imagination” in a technical sense in
the Institutio. He certainly does not contrast imagination with some other cognitive
faculty. Rather, he explains,

Here we use the word imagination for that cogitation or action of the
mind that results in an image of the thing cognized hovering before
the mind. This should be noted because the term is sometimes used
for the imaginative faculty, which some people call by the Greek
word phantasia and attribute to the inferior part of the soul, which

29. Posteriore modo: cum licet Ideae illae singulares in aliquot similes sint, seu mutuo conve-
niant, multa tamen simul discrimina habeant, quibus inter se differant, ideo Mens seorsim
spectando, ac ideo veluti abstrahendo ex omnibus id, in quo omnes conveniunt, et detractis,
seu non spectatis discriminibus, quibus differunt; illud sic abstracte spectatum, nihilque non
commune habens, pro Idea communi, universali, generali habet … Exempli enim gratia;
dum Mens eorundem Socratis, Platonis, Aristotelis, in eo convenire, simileisque esse attendit,
quod unaquaeque earum repraesentet Animal bipes, eracta facie, ratiocinans, ridens, discipli-
nae capax, &c. istud (nempe esse Animal bipes, eracta facie, &c.) velut abstrahit, inque ideam
format, a qua sint detracta discrimina omnia, quibus illi mutuo differunt (ut, quod … iste
senex, ille vir, alius adolescens …) ac talem rursus ideam habet pro Idea universali, seu generali
hominis (OO 1:93b).
30.  Ac difficile quidem est, ne dicam impossibile ita pure hominem in commune imaginari, ut
neque magnus, neque parvus, neque mediocris statutae sint; ut neque senex, neque infans, neque
intermediate aetatis; ut neque albus, neque niger, neque alterius specialis coloris: At Mente saltem
tenere oportet, hominem, quem communiter consideratum volumus, debere esse his omnibus dis-
criminibus absolutum (OO 1:95b).
Gassendi on the Problem of Universals  •  27

is common to men and animals. (Institutio Logica 1, Introduction;


OO 1:92a)31

And he lets us know that he is using the term “image” broadly:

The image that hovers before the mind and is, as it were, presented when
we think about something, is customarily given many other names. For it is
called idea, species, and—​adapting the name of the action—​even notion,
preconception, anticipation, [etc.]. (Institutio Logica 1, Introduction;
OO 1:92a)32

Thus, it’s implausible that Gassendi is relying on a distinction between imagina-


tion and some other cognitive faculty when he notes, “It is difficult … to imag-
ine man in general so accurately that he is neither large nor small nor medium
height.” Rather, he is just reminding us that height is irrelevant.
The second issue stems from Gassendi’s admission that forming an abstract
idea requires not just noticing that various particular ideas are similar but notic-
ing that they are similar in certain respects:

[A]‌fter the mind has formed general ideas of men, horses, lions, and bulls by
abstraction, then, noticing that they agree with each other in some ways and
differ in others (for they agree in that each one represents a sentient body;
they disagree in that one represents a laughing thing, one a whinnying thing,
one a roaring thing, one a lowing thing), it removes all the ways in which
they differ and select only the way in which they agree, namely, represent-
ing a sentient body, which is called by the one name animal, and thus forms
from them a more general idea. (Institutio Logica 1.5; OO 1:93b)33

31. Imaginationis vocem haec accipimus pro Cogitatione, seu actione Mentis, quae ad rei cogitatae
imaginem Menti obversantem terminatur. Hoc autem notandum, quia talis vox sumitur inter-
dum pro facultate imaginatrice, quae nonnullis & Graeca voce appellatur Phantasia, & tribuitur
parti Animae inferiori, quae est Homini communis cum Brutis; quippe ipsa quoque Bruta imagi-
nantur (OO 1:92a).
32. Imago porro illa, quae nobis rem quampiam cogitantibus Menti obversatur, ac veluti obiicitur,
plerisque etiam aliis donari nominibus solet. Dicitur enim etiam Idea ac Species, & accommodato
nomine actionis, etiam Notio, Praenotio, Anticipatio, seu anticipata notio (prout nempe fuit prius
acquisita) ac rursus Conceptus, itemque Phantasma, prout sedem habet in Phantasia, facultative
imaginatrice (OO 1:92a).
33. Posteriore, sive abstrahendo, postquam Mens Ideis generaleis Hominis, Equi, Leonis, Tauri,
abstractione illa formavit, tum attendens ipsas convenire in aliquo, dissidere in alio (convenient
2 8   •  The Problem of Univer sals in Early Modern philosophy

This is an account of the formation of abstract ideas; the account of the forma-
tion of aggregate ideas is roughly similar but leaves out the step where the mind
recognizes the respects in which its ideas are similar and dissimilar in various
respects.
The issue is how the mind can recognize such similarities. Does noticing
that the abstract ideas of a man and a lion are similar require using our aggre-
gate idea of a sensing body? Or do we simply recognize that men and lions are
similar in some respect, without needing any idea of the respect in which they
are similar?
Gassendi has independent reasons for holding that we have a brute capacity to
recognize similarities. Consider his claim that general ideas can be formed more or
less well:

A general idea is more perfect if it is more complete and more accurately repre-
sents that in which the particulars agree. Since, first, an idea is called general by
aggregation, because it is an aggregate containing all things of the same kind,
it will be more perfect if fewer of the things that would make it complete
are left out. In this way, if some idea of man included not only Europeans,
Africans, and Asians but also Americans, it would be more perfect than if,
as customary for the ancients, it only represented Europeans, Africans, and
Asians. (Institutio Logica 1.8; OO 1:95b)34

This is true for abstract ideas as well as aggregate ideas:

Since, second, an idea is also called general by abstraction—​that is,


because it was, as it were, chosen to represent something common to all
the particulars—​if it has something mixed in that does not belong to all
of them, it will be less general and so less perfect. This is like the idea of a
man if it represents an animal four cubits tall, with a white face, a straight

nimirum in eo, quod unaquaeque repraesentet corpus sentiens; dissident in eo, quod una reprae-
sentet ridens, alia hinniens, alia rugiens, alia mugiens) ideo illa omnia, quibus dissident, detra-
hit, seligensque solum id, in quo conveniunt, nempe Corpus sentiens, quod nomine uno dicitur
Animal, Ideam ex eo facit, ideis illis generaliorem (OO 1:93b).
34.  Idea generalis tanto est perfectior, quanto est completior, ac repraesentat purius id,
in quo singularia conveniunt. Cum generalis enim dicatur, primum quidem aggregatione,
quod sit aggeries continens omneis, quae eiusdem sunt generis, tanto sane perfectior erit,
quanto pauciores deerunt, quibus veluti completa reddatur. Quo pacto si quis hominum Idea
non modo Europaeos, Africanos, Asiaticos, seu Americanos etiam complectatur, perfectiorem
illam habebit, quam si Veterum more, solos Europaeos, Africanos, Asiaticos complecteretur
(OO 1:95b).
Gassendi on the Problem of Universals  •  29

nose, and so on, since these and other properties of this kind are common
to some but not all men. (Institutio Logica 1.8; OO 1:95b)35

Americans are men, for instance, whether or not anyone’s general idea of a man
includes them. Our ideas do not determine the boundaries of the kind man. So
what does? Gassendi tells us that

An attribute agrees with and belongs to a subject either inseparably, and is


called necessary, or separably, and is called contingent. When I say insepa-
rably, I mean that the attribute agrees and belongs in such a way that the
subject cannot exist without it. Such is animal in respect of man, for man
cannot exist without animal existing at the same time. (Institutio Logica
2.3; OO 1:101b)36

Attributes can be necessary in various different ways:

A necessary attribute is either a genus, or a quality implanted in the subject


by nature. For whatever is inseparable from a subject is either its genus,
whether proximate or remote—​like being an animal, being a living thing,
or being a body, in the case of man—​or it is a quality implanted in the sub-
ject by nature, either one proper to the subject—​like reason or the ability
to laugh in the case of man—​or a quality common to it and other things—​
like the faculty of sensing in the case of man, who has it in common with
all other animals. (Institutio Logica 2.4; OO 1:101b)37

35. Cumque deinde etiam abstractione generalis dicatur, quod selecta veluti sit, ut repraesentet
quidpiam commune omnibus singularibus, si quid sane admistum habeat, quod omnibus non
competat, eo minus erit generalis, atque adeo minus perfecta. Quo se habet modo Hominis idea, si
repraesentet Animal staturae quadricubitalis, candida facie, naso recto, &c. quippe haec; aliaque
id genus aliquorum propria, non omnium hominum sunt communia (OO 1:95b).
36. Attributum congruit Subiecto, ipsique cohaeret; aut inseparabiliter, & Necessarium dici-
tur; aut seperabiliter, & dicitur Contingens. Cum Inseperabiliter dico, ita congruere, cohaer-
ereque Attributum intelligo, ut Subiectum esse sine ipso non valeat. Tale est Animal respectu
Hominis; neque enim esse Homo, quin simul sit Animal, potest (OO 1:101b).
37. Attributum Necessarium aut Genus est, aut Qualitas a Natura insita Subiecto. Nimirum,
quia quicquid inseparabile est a Subiecto, aut ipsius est genus, idque tam proximum, quam remo-
tum, ut respectu Hominis esse Animal, esse Vivens, esse Corpus; aut est qualitas eidem Subiecto a
Natura insita, eaque sive illius propria ut Hominis Ratio, aptitudo ad risum; sive ipsi, aliisque com-
munis, ut est Homini facultas sentiendi, quam habet communen cum omnibus aliis Animalibus
(OO 1:101b). In contrast, a contingent attribute id fere duorum alterum est. Et primum qui-
dem perspicuum est seperabileis esse Qualitates, quae non a Natura insunt, sed externe adveniunt,
Accidentaliaque dicuntur, quia subiecto sic accident, ut abesse ab illo, sine eius interitu possint …
Deinde perspicuum est quoque separabileis esse Denominationes, quae ob relationes ad res externas
3 0   •  The Problem of Univer sals in Early Modern philosophy

To say that being an animal is an essential property of particular men means that
it follows from what’s essential to those particular men that they are similar to all
other animals. This similarity obtains in virtue of the essential properties of various
particular men, without requiring any property to be shared by all the particulars.
What we need, then, is the ability to recognize similarities in respect to essen-
tial properties. This makes us able to grasp kinds that exist in nature and thus able
to form both abstract and aggregate ideas:

[A]‌ll the evidence and certainty that general propositions have depends
on what’s been inferred by induction from singulars. … We know, for
example, that every man is an animal because we perceive by the senses
that [Plato and Socrates and other particular men] are animals. (Institutio
Logica 3.16; OO 1:116b)38

Presumably, just as it is a brute fact about the world that the various particu-
lars are grouped into categories in virtue of their similarities and differences, it
is a brute fact about us that we can recognize these similarities and differences.
After all, we need this ability to get around in the world God created, and so an
omnibenevolent God gave us this ability.
Notice how far Gassendi has come here from the Exercitationes’s claim that to
know that men are animals you would have to examine every past, present, future,
and merely possible man.

2.7
In Syntagma 2.3b.9—​“On the Intellect, or Mind”—​Gassendi introduces what
some readers have seen as a radical change in position from the Exercitationes and
Disquisitio. This is part of his attempt to establish “what is stipulated by the holy
faith,” namely, “that the mind, or the higher part of the soul … is an incorporeal
substance” (OO 2:440a).39 On his view, this is best established by establishing the

attribuuntur: ut pote, cum quibus definentibus, aut immutatis pereant, neque amplius congruunt.
… Sic at mortem filij, aut uxoris, Pater, aut Maritus esse desinit (OO 1:101b–​102a).
38. . . . evidentia, & certitudo omnis, quae de generali propositione habetur, dependet ab ea, quae
ex singularium inductione collecta est … nosceamus omnem Hominem, v.c. esse Animal, ex eo
est, quia prius percepimus sensu & Platonem, & Socratem, & singulatim caeteros esse Animalia
(OO 1:116b).
39. . . . superest de eo dicamus, quo, adstipulati Sacrae Fidei, dicimus Mentem, seu partem illam
superiorem Animae (quae & proprie Rationalis est, & in unoquoque homine singularis) substan-
tiam esse incorpoream, quae a Deo creetur, infundaturque in corpus (OO 2:440a).
Gassendi on the Problem of Universals  •  31

incorporeality of the intellect.40 For, he says, if you deny the existence of a distinct
intellect and identify the mind with the corporeal imagination, like Epicurus and
Democritus, then human minds and animal minds will differ only in perfection,
not in substance (OO 2:425a). And thus, you will have to conclude either that
we are mortal or that animals are immortal, both of which conclusions are unac-
ceptable. Thus, we have evidence that the human soul is immaterial, albeit not the
same degree of evidence that we have for the truths of mathematics (OO 2:627a).
Gassendi tries to show that intellect and imagination are distinct by argu-
ing that the intellect performs three actions that could be performed only by an
incorporeal thing:

(1) The intellect can form notions of things that cannot be represented by
images (2.440b), like God and other immaterial substances (OO 2:451a).
(2) The intellect can reflect on itself and its operations, although no material
thing can act on itself directly (OO 2:441b, 2:451a).41
(3) The intellect can form universal notions and recognize their ratio universali-
tatis, which requires that the intellect is immaterial (OO 2:441b, 2:451b).

These claims are puzzling. For one thing, some of them seem to be the modus
tollens forms of modus ponens arguments given in the Disquisitio. Compare the
following to (1) and (2):

as for the ideas of things that are believed to be immaterial, like God,
angels, or the human soul or mind, it is clear that even our ideas of them
are corporeal or quasi-​corporeal, since the ideas are derived from the
human form or from other very rarefied, simple and insensible things.
(Disquisitio 6.1; OO 3:386a)42
When I think about why sight does not see itself or intellect under-
stand itself, it occurs to me that nothing acts on itself … in order for
knowledge of some thing to be elicited, it is necessary that the thing acts
on the faculty that cognizes it, namely by transmitting its species to it or

40. Will and intellect are distinct faculties. The intellect is more promising because it is pri-
mary: the will simply pursues what the intellect presents to it as good (OO 2:440b).
41. The move from self-​reflective ability to immateriality is commonly found in the Neoplatonic
tradition. See, for instance, Ficino, Theologia Platonica VIII 15. Thanks to Brunello Lotti for
this reference.
42. Quod spectat ad Ideas rerum immaterialium creditarum, ut Dei, Angeli, Animae humanae
seu mentis; constat etiam quascumque habemus de ipsis Ideas, esse vel corporeas, vel quasi cor-
poreas, ex forma scilicet humana, & ex rebus alias tenuissimis, simplicissimis, insensibilissimis
3 2   •  The Problem of Univer sals in Early Modern philosophy

informing it with its species. It seems clear that, because the faculty is not
outside itself, it cannot transmit its species to itself and consequently can-
not elicit knowledge of itself or, what is the same, cannot perceive itself.
(Disquisitio 3.6; OO 3:332b–​33a)43

This should not be surprising. In the Disquisitio Gassendi argues that the distinc-
tion between intellect and imagination is just a difference of degree. Descartes
claims that intellect must be distinct from imagination because the intellect can
clearly and distinctly perceive things that imagination can only grasp confusedly,
like the chiliagon. Gassendi replies:

It will not follow from this that you have reason to add more than one
kind of internal cognition. It is only accidental and a matter of degree
whether you contemplate a certain figure distinctly or confusedly, intently
or in a slack way. And indeed, when we wish to run through the heptagon,
octagon, and other figures all the way up to the chiliagon or miriagon and
always continually attend to the distinctness or slackness, we will not be
able to say where, or with what figure, imagination stops and only intellect
remains. (Disquisitio 6.1; OO 3:385b–​86a)44

Why did Gassendi change his mind about whether we can conceive of universals
and other things that cannot be represented by images and about whether there is
a distinct faculty of intellect? He doesn’t say, but here is one possible explanation.
Gassendi felt certain that the human soul is immaterial. This certainty was not
based on a philosophical argument, but, given his conception of the relationship
between faith and reason, he thought that the immateriality of the soul should

(OO 3:386a). Cf. 6.4 (OO 3:399b), speaking of an idea: Alioquin certe si partibus careat,
quomodo partes repraesentabit? Si extensione, quomodo rem extensa? … Non ergo videtur idea
extensione prorsus carere.
43.  Cogitanti certe mihi quorsum fieri possit, ut neque visus seipsum videat, neque intellectus
seipsum intelligat; illud in mentem subiit, quod nihil agat in seipsum. … Cum aliunde vero ad
notitiam alicuius rei eliciendam, necesse sit, rem agere in facultatem cognoscentem; immittere
nempe in illam sui speciem, sive sui specie illam informare: perspicuum videtur ipsam facultatem,
cum extra seipsam non sit, non posse illam sui speciem in seipsam transmittere, neque sui notitiam
consequenter, elicere, sive, quod idem est, percipere seipsam (OO 3:332b–​33a).
44. . . . at non erit propterea, cur plusquam unum genus internae cognitionis adstruas, cui acci-
dentarium solummodo sit, ut secundum magis, & minus distincte vel confuse, intente vel remisse
figuram quampiam intuearis. Et certe cum Heptagonum, Octagonum, caeterasque porro figuras
ad Chiliagonum, aut Myriagonum usque percurrere voluerimus & ad maiorem, minoremque dis-
tinctionem, vel remissionem, semper continuoque attenderimus, dicere-​ne poterimus, ubinam, seu
in qua figura imaginatio desinat, intellectio sola remaneat? (OO 3:385b–​86a).
Gassendi on the Problem of Universals  •  33

be established on the basis of reason. Hence, he provided the best arguments he


could come up with for the immateriality of the soul, even though he did not find
those arguments entirely convincing.
In any case, let us look at the details of Gassendi’s claim that the intellect can
form universal notions:

We not only form universals or universal notions but also perceive their
ratio universalitatis. This is because—​since universals are the kind of thing
that prescinds from all material conditions and from the differences of
particulars, such as size, shape, color, and the like—​it is in fact necessary
that the intellect which brings about this precision from matter and con-
siders it is separate from matter and of a condition higher than every mate-
rial circumstance. (OO 2:441a–​b)45

This is a difficult passage. What exactly is the ratio universalitatis? And why does
Gassendi think that perceiving it requires an incorporeal intellect? It cannot just
be that images are incapable of representing universals:  (3)  is presented as an
argument in its own right, not a special case of (1).
Looking at how Gassendi responds to two possible objections may help. The
first objection is that animals can grasp universals but nobody thinks they have
an immaterial intellect:

It seems that brutes also make universals, in their own way. For example,
a dog makes a species of man, according to which, when an erect, bipedal
animal appears, it infers that this animal is a man rather than a lion or a
horse. (OO 2:441b)46

Gassendi responds:

But first, in any case animals do not recognize universality itself, or a


universal nature, for example humanity, as prescinded and distinguished

45.  Tertium est earum, quibus non modo universalia, universaleisve notiones formamus; sed
percipimus quoque ipsam rationem universalitatis. Siquidem, cum universalia eiusmodi sint, ut
praecidantur ab omnibus conditionibus materialibus, discriminibusque singularitatis, ut mag-
nitudine, figura, colore, & similibus; opertet sane Intellectum, qui hanc praecisionem a materia
facit, & considerat, absolutum esse a materia, conditionisque esse omni circumstantia materiali
eminentioris (OO 2:441a–​b).
46. Nam quod posset obiici, videri quoque Bruta suo modo universalia fabricari, veluti speciem
hominis, iuxta quam, quoties occurrit bipes erectumque Animal, canis v.c. coniiciat esse hominem,
non vero leonem, aut equum (OO 2:441b).
3 4   •  The Problem of Univer sals in Early Modern philosophy

from every degree of particularity. And next, animals do not apprehend


abstract things, but only concrete things: not color, but a colored thing,
not taste, but a thing with a taste, etc. And, it can be understood that
there is nothing in a dog other than the bare memory of those particulars.
(OO 2:441b)47

Gassendi does not tell us what, in the dog’s behavior, is evidence that it cannot
apprehend color itself. (Nor does he give us evidence that we can apprehend color
itself; he seems to think this is simply obvious.) But at least this passage makes
clear that the ability to categorize things is not evidence of grasping the ratio
universalitatis.
Now for the second objection, together with Gassendi’s reply:

And, lest you insist that in our case too, whenever we conceive a univer-
sal there is always mixed in some element of particularity, like a certain
size, a certain shape, a certain color, etc., nevertheless we find by experi-
ence that—​if not all at once, at least in succession—​we can strip this
nature of any special shape, any special color, and so on. (OO 2:441b)48

Gassendi cannot just mean that for any particular determination, we can think
of something without that determination. Even Berkeley allows that for any par-
ticular weight, we can think of a dog without that weight. Rather, he must mean
that for any particular determinable, we can think of something without any of
its determinations—​if not something without any determination of any of the rel-
evant determinables. (We can think of a dog without any weight or without any
color or without any size, even if we can’t think of a dog lacking all three.) Thus, he
argues, we must possess a faculty of intellect, distinct from the imagination. Indeed,
we must possess a rational soul, distinct from the corporeal sensitive soul which
imagines and has all the other cognitive capacities we share with other animals.
It’s tempting to think that the distinction between imagination and intellect
maps onto the distinction between aggregate and abstract ideas. This would make

47.  At imprimis, saltem Bruta non agnoscunt ipsam universitatem, seu universalem naturam,
v.c. humanitatem, tanquam praecisam, & ab omni gradu singularitatis discretam; ac deinde, ut
Bruta non ipsa abstracta apprehendunt, sed concreta solum; ut non colorem, sed coloratum; non
saporem, sed sapidum, &c. ita licet intelligi nihil aliud in cane esse, quam memoriam solam eorum
singularium (OO 2:441b).
48. Et, ne instes in nobis quoque, dum universale concipimus, admiscere semper aliquid singulari-
tatis, ut certae magnitudinis, certae figurae, certi coloris, &c., experimur tamen, nisi simul, saltem
successive spoliari a nobis naturam qualibet speciali magnitudine, qualibet speciali figura, quolibet
speciali colore; atque ita de caeteris (OO 2:441b).
Gassendi on the Problem of Universals  •  35

aggregate ideas the objects of imagination and abstract ideas the objects of intel-
lect. And it would mean that aggregate ideas are common to animals and humans
while only human beings have abstract ideas. Such a reading has its advantages.
Given the way Gassendi answers the objection about animal cognition just dis-
cussed, he should say, in Syntagma 2.3b.9, that animals have aggregate ideas.
Moreover, aligning abstract ideas with the intellect would help explain both what
the special function of the intellect is and why we need a second type of general
ideas in addition to aggregate ideas.
Unfortunately, this doesn’t work. There are two reasons that Gassendi cannot
think that abstract ideas are the objects of the intellect. We saw the first reason
in section 2.6 above: when Gassendi presents the distinction between two types
of general ideas in the Institutio, he doesn’t distinguish imagination from intel-
lect. He contrasts his use of the term ‘imagination’ with one on which it stands
for a faculty belonging to the inferior part of the soul, and says that ‘idea’, ‘spe-
cies’, ‘notion’, and so on are synonyms for ‘image’ (1.92a). This is hardly what you
would expect from someone who thinks that abstract ideas are the object of intel-
lect and aggregate ideas the object of imagination.
The second reason is that Syntagma 2.3b.9 doesn’t mention abstract ideas.
In fact, it denies that the immaterial intellect has its own objects, at least in this
life. Instead, it uses the material ideas of the imagination—​something which,
Gassendi remarks, explains why the intellect’s function is disturbed when the
imagination is disturbed by wine or sickness (OO 2:454b). The intellect’s reli-
ance on material ideas is made clear in Gassendi’s discussion of (1), the argument
that the soul must be immaterial because it cognizes things that cannot be repre-
sented by images:

because the intellect uses no other species than phantasms, it seems espe-
cially clear that we understand nothing by a species other than a corporeal
species, which is the sort of thing a phantasm is. Indeed, if the intellect
ever used any incorporeal species, it would do so especially when we were
striving to understand God, the most incorporeal thing at all. And we can
deduce from this that it is certain that God is not understood by us, while
we are in this mortal life, except under some species which the imagina-
tion supplies. (OO 2:447b)49

49.  Quod autem non aliis speciebus, quam Phantasmatibus Intellectus utatur, vel ex eo vide-
tur imprimis patere, quod nihil prorsus intelligamus alia specie, quam corporea, cuiusmodi
Phantasma est. Sane vero si ulla unquam incorporea specie uteretur, tum faceret maxime, cum
Deum intelligere, rem maxime omnium incorpoream, adnitimur; ac ex suo loco deductis, constat
non intelligi Deum a nobis, donec in vita mortali sumus, nisi sub specie aliqua, quam Phantasia
suppeditet (OO 2:447b).
3 6   •  The Problem of Univer sals in Early Modern philosophy

And again,

Intellect is distinct from Imagination so that although imagination has


material species under which a thing is imagined, the intellect does not have
[species] under which it understands things. For instance, intellect has no
[species] of the size of the sun it grasps, but only by its own power, or by
reasoning, comprehends that there is such a size in the sun … from this it
follows that the intellect, understanding a thing without a material species,
must be immaterial … and indeed, the intellect uses species perceived by the
imagination in such a way that, by reasoning from them as if by degrees, it
finally arrives at that which can be understood without species or phantasms.
(OO 2:440b–​41a)50

What is crucial here is how the intellect uses the corporeal species of the imagina-
tion. Gassendi is not saying that the intellect uses material species to produce some
other kind of species, some sort of immaterial representation that can serve as the
object of the intellect. His claim is that the intellect can somehow use material spe-
cies to think other things, without producing a distinct mental representation of
those things.51 While explaining how we can know that God and other incorporeal
substances exist without having mental representations of them, Gassendi says:

although the intellect uses corporeal species, still when it reasons it does
not necessarily content itself with them. Rather, by its liberty and energy,
it reasons that there is something further which cannot be represented by
a corporeal species. And although the intellect may not perceive what sort
of thing this is in itself, nevertheless it concludes and understands that it
really exists. (OO 2:442b)52

50. Intellectum ita esse distinctum a Phantasia, ut cum Phantasia habeat materialeis species, sub
quibus res imaginetur, non habeat tamen Intellectus, sub quibus res intelligat; neque enim ullam,
v.c. habet illius magnitudinis, quam in Sole intelligit; sed tantum vi propria, seu ratiocinando,
eam esse in Sole magnitudinem comprehendit; ac pari modo caetera. Nempe ex hoc efficitur, ut rem
sine specie materiali intelligens, esse immaterialis debeat. … Ac utitur quidem etiam Intellectus
speciebus Phantasia perceptis, tanquam gradibus, ut ratiocinando assequatur ea, quae deinceps
sine speciebus, Phantasmatisve intelligit (OO 2:440b–​41a).
51. Cf. also OO 2:448b: Videtur itaque Mens nostra, donec degit in corpore, non aliis uti intelli­
gibilibus speciebus quam ipsis Phantasmatibus, iisque seu meris, seu ipsa vi Mentis veluti modi-
ficatis, applicitis, in habitum versis. The kind of modification he has in mind here is the way in
which we think of something incorporeal by thinking of progressively less solid things: first
earth, then, fire, then air, etc. (OO 2:448a).
52. Declarat nimirum, tametsi speciebus utatur corporeis, non tamen necessario se in illis conqui-
escere, sed ea libertate, ac energia esse, ut ratiocinetur, esse praeterea aliquid, quod specie corporea
Gassendi on the Problem of Universals  •  37

This allows the intellect to apprehend incorporeal substances negative or abstrac-


tive, although it cannot have positive or intuitive knowledge of them while it is
housed in the body (2.442b). Recognizing that some readers may find it strange
that the intellect uses phantasms, he points out that it’s no more mysterious for
an incorporeal faculty to use corporeal species than for an incorporeal soul to be
joined immediately to the body, as people typically assume (OO 2:449a).53
To explain how the intellect uses corporeal species to think about universals,
Gassendi reverts to a familiar explanation:

the intellect, which … perceives by its own power not only universals but
also the nature or condition of universality, apprehends a universal thing,
or a thing considered in a universal way, more absolutely if it understands
it as more separate from the special marks and properties of the particu-
lars and, as they say, from individual differences. And so, for example, of
particular men one is young, another old, one large, another small … and
so on; the intellect knows that it should separate all the things of this sort
and only select what is common to all or in which they are all similar;
of this sort is being an animal capable of understanding, born to laugh,
etc. so that man is reckoned to be considered universally or to be some
universal thing … and for that reason the intellect considers not only, for
example, the aggregate of men, in so far as it differs from the aggregate of
horses … but considers it so carefully that, besides the proper marks by
which the particular men comprehended in it differ from one another,
something similar in all of them is manifest. As a result, the intellect as it
were selects that similarity and considers it separately and so recognizes
that there is a common human nature that can be predicated of the par-
ticulars. (OO 2:459a)54

repraesentari non possit; quodque utcumque non pervideat cuiusmodi secundum se est, concludat
nihilominus, intelligatque revera esse (OO 2:442b).
53. Non debet nos vero imprimis morari, quod praeter omnem proportionem sit, ut Intellectus,
facultas incorporea, Phantasmate, specie incorporea, immediate utatur. Nempe hoc minore pro-
portione non fit, quam quod vulgo admittunt incorpoream Animam jungi immediate corpori, &
corporeis membris ad movendum uti (OO 2:449a). This is not Gassendi’s own view: he holds
that the incorporeal soul is joined to the body through the intermediary of the corporeal soul
(cf. OO 2:627b).
54. Heinc ergo Intellectus, qui ex antedictis percipit sua vi non ipsa modo universalia, sed etiam
naturam, conditionemve universalitatis, deprehendit rem universalem, universeve spectatam,
tanto esse absolutiorem, quanto intelligitur esse a specialibus singularium notis, proprietatibusve,
&, ut appellant, differentiis individualibus secretior. Sic cum, exempli gratia, Hominum singu-
larium alius sit iuvenis, alius senex; alius magnus, alius parvus … atque ita de caeteris; novit
Intellectus debere huiusmodi omnia seponi, ac id solum deligi, quod commune sit omnium, seu
3 8   •  The Problem of Univer sals in Early Modern philosophy

He says less than one might like about how this works, but it sounds like selective
attention. The intellect is not constructing a new idea on the basis of the aggre-
gate idea but simply considering the aggregate idea in a special way. The continu-
ation of the passage is also reminiscent of the Institutio:

Indeed, it is rare, if not perhaps impossible, that when we think or speak


of a man in general, a multitude of such men does not hover before us …
obviously, man is thought of with some size, some color, and similar
adjuncts, which, if we pay attention, we will not find in all things. But
although this happens to us on account of the imagination acting as a
henchman, nevertheless it is enough that the intellect itself also wants to
set such adjuncts aside or understands them as things that should be set
aside, and does not judge that they should be affirmed of man taken uni-
versally. (OO 2:459a)55

But although a lot of material from the Institutio Logica is reused in Syntagma
2.3b.9, talk of abstract ideas does not appear there. In the Institutio Logica, it
sounded like selective attention produced a new type of idea—​a genuinely abstract
idea—​distinct from ideas of particulars and their aggregates. In Syntagma 2.3b.9,
it sounds like selective attention is simply a new way of considering the aggregate
ideas we already have and produces nothing new.
Seeing the formation of abstract ideas as selective attention helps explain
several things. It helps explain what Gassendi means by the ratio universali-
tatis: it is the “something similar” that is “manifest” in every member of the
aggregate. Something similar—​a respect in which they are all similar, not
a genuine universal. Thus, the metaphysical picture from the Exercitationes
remains in place. At the same time, seeing the formation of abstract ideas as

in quo omnes sint similes; cujusmodi est, esse Animal intelligentiae capax, natum ad risum, &c.
ut Homo spectari universe, seu esse res quaepiam universalis reputetur. … Quare Intellectus non
modo spectat Hominum, v.c. aggeriem, quatenus est discreta ab aggerie equorum … sed spectat
eam maxime, quatenus praeter notas proprias, quibus singuli homines ea comprehendi inter se
different, elucet in omnibus quidpiam simile, ut illus veluti seponat, seorsimque consideret, indque
esse agnoscat naturam humanam commune, quae enunciari de singulis possit (OO 2:458b–​59a).
55. Et rarum est quidem, nisi forte impossibile, ut cum cogitamus, dicimusve universe hominem,
non obversetur nobis multitude quaedam hominum, seu species quaedam hominis aliquibus notis
singularitatis affecti; cogitatur scilicet Homo cum aliqua magnitudine, aliquot colore, consimili-
busque adiunctis, quae si attendamus, non reperientur in omnibus. At cum istud nobis contin-
gat ob satellitium Phantasiae; sufficit tamen, ut Intellectus haec quoque, si alidqua sunt, seposita
velit, seponendave subintelligat; neque ipsa de Homine universe spectato affirmanda censeat (OO
2:458b–​59a).
Gassendi on the Problem of Universals  •  39

selective attention helps with how we recognize the ratio universalitatis. Saying
that something similar is “manifest”—​elucet, shines forth—​in the members of
the aggregate suggests that the process is passive. The world is doing the work,
not us. But, in any case, we should not expect Gassendi to explain the mecha-
nism by which the ratio universalitatis is grasped, since his point in Syntagma
2.3b.9 is precisely that there is no mechanism. That’s why it requires an incor-
poreal intellect.

2.8
When I started thinking about Gassendi’s remarks on universals I was puzzled
by several things. Why isn’t he more interested in universals? Why does he argue
against universals only in re? Why does he distinguish two different kinds of gen-
eral ideas—​abstract and aggregate? Do the two different kinds belong to two
different faculties, intellect and imagination? And what does it mean to grasp not
only universals but also their ratio universalitatis?
I’ll end by summing up my admittedly rather provisional answers. Gassendi
is not particularly interested in writing about universals because he finds
it obvious that there are no universals and because he has alternate ways of
doing the philosophical work universals had been taken to do. He only argues
against universals in re because his purposes in discussing universals are rhe-
torical, not argumentative:  thus, he picks the easiest target. What look like
two different kinds of general ideas are in fact two different ways of using one
kind of ideas: the ordinary way and a special, abstract one. Positing these two
different uses accords with the results of introspection and provides a func-
tion for the immaterial intellect, which Gassendi needs in order to establish
the immateriality and hence immortality of the human soul. When we use
aggregate ideas in the ordinary way, they are objects of the imagination; what
the Institutio calls abstract ideas are simply the intellect’s selective attention to
those aggregate ideas. Finally, grasping universals and their ratio universalitatis
is just grasping the respects in which the various members of the aggregate are
similar.56

56.  Notice that if these suggestions are right, then the Institutio’s presentation of gen-
eral ideas is rather misleading. This is not as bad as one might think, given that Gassendi
intended the Institutio as a manual to be used in teaching logic to young students. Given
that purpose, the more complex material from Syntagma 2.3b.9 would simply have been
out of place.
4 0   •  The Problem of Univer sals in Early Modern philosophy

Abbreviations
AT = Descartes 1964–​74; cited by volume and page.
CSM = Descartes 1984–​85; cited by volume and page.
OO = Gassendi 1658; cited by volume and page.
3 HOBBES, UNIVERSAL NAMES, AND
NOMINALISM

Stewart Duncan

Thomas Hobbes was, rather famously, a nominalist. The core of that


nominalism is the belief that the only universal things are universal
names: there are no universal objects, or universal ideas. Hobbes’s nomi-
nalism prompted notable objections from his contemporaries. Leibniz
referred to Hobbes as an ultranominalist, someone who went well
beyond the position of previous nominalists, and he and other contem-
poraries objected that Hobbes’s nominalist views would have the conse-
quence that people could not say the same thing in different languages.
Hobbes’s nominalism is prominent enough that it is regularly
mentioned in accounts of his philosophy, but there is relatively
little extended discussion of this topic in the secondary literature.1
However, despite the brevity of Hobbes’s treatments of the issue, and
the seeming clarity of his central view that universal names are the
only universals, a variety of questions do arise.
This chapter has three main sections. The first and longest looks at
what Hobbes’s views about universal names were, how they evolved
over time, and how Hobbes argued for them. The remainder of the
chapter investigates Hobbes’s view further by looking at two objec-
tions to it, one from the seventeenth century and the other from the
twentieth. Thus the second section examines a criticism made by
several of Hobbes’s contemporaries, that Hobbes’s view could not
account for people saying the same thing in different languages. Then
the third section looks at a more recently popular criticism of Hobbes,
that his nominalism’s reliance on similarity implicitly (and inconsis-
tently) involves reliance on a universal.

1. That said, for some discussions see Bernhardt 1985; Callaghan 2001; Jesseph 1999,
205–​19; Laird 1934, 147–​49; Peters 1956, 126–​37; and Watkins 1965, 103–​09.
4 2   •  The Problem of Univer sals in Early Modern philosophy

3.1. Hobbes’s Nominalism
The central reason to call Hobbes a nominalist is that he says that universal
names are the only universals: “there is nothing universal but names” (EL 5.6).
Thus, understanding his nominalism is largely a matter of understanding his
view about common or universal names. There are three main textual sources for
Hobbes’s views about these matters. The first two are the chapters on language
in The Elements of Law and Leviathan, two general accounts of the workings of
language written about a decade apart. The third is the slightly later De Corpore,
which contains Hobbes’s longest discussions of the workings of language.
The details of the views and arguments change from text to text, but there
are certain common features. Discussions of universals are situated in discus-
sions of the distinction between proper and common names. Proper names each
name one thing, but common names each name more than one thing. Those
common names are, Hobbes says, the only universals. In supporting that idea,
that names are the only universals, Hobbes argues that there are no universals in
rerum natura, no universal things, we might say. He also argues that there are no
universal ideas.
Hobbes was thus opposed to the great majority of his contemporaries on
this issue. The rejection of universal things was widespread in the early modern
period. However, that tended to be accompanied by an acceptance of what we
might call universal ideas (though the terminology and the details of the views
varied widely). This was true of Hobbes’s seventeenth-​century “modern” contem-
poraries. It had also been true of Hobbes’s most famous nominalist predecessor,
Ockham.2 Ockham argued in his Summa Logicae that “it ought to be said that
every universal is one particular thing and that it is not a universal except in its
signification, in its signifying many things.”3 Moreover, he went on to argue there
that “no universal is a substance existing outside the mind,” but that “every univer-
sal is an intention of the mind.”4 As Panaccio summarizes Ockham’s view: “spe-
cies and genera are concepts. Which is to say they are but natural signs within the
mind. Their generality does not amount to a special ontological feature they have
but to a semantical one: a general concept, after all, is simply a singular mental
occurrence that signifies several other singular entities” (Panaccio 2000, 65).5

2. On the comparison of Hobbes’s nominalism and Ockham’s, see Bernhardt 1985.


3. Summa Logicae I.15; Ockham 1998, 78.
4. Summa Logicae I.16; Ockham 1998, 79.
5. It might seem mistaken to compare Hobbes to Ockham, on the grounds that the early moderns
were engaging in different debates about universals and nominalism than their medieval prede-
cessors were. But although Hobbes ignored or abandoned much of the apparatus of medieval
Hobbes, Universal Names, and Nominalism  •  43

Hobbes sympathizes with something of the spirit of this: talk of universals such


as genera is not to be explained by the existence of special universal entities that
correspond to them, but by a better understanding of the semantics of the talk.
And that better understanding is, for Hobbes, one according to which there are
only particular things, which come to represent multiple other particular things.
For Hobbes, however, the particulars that do that representing are words, not
concepts or anything else that exists in the mind. The mental language that plays
such an important role in Ockham’s account is a much less important aspect of
Hobbes’s story. Hobbes does say that the role of language is to make our thoughts
known, talking of speech, “consisting of names or appellations, and their connex-
ion, whereby men register their thoughts, recall them when they are past, and also
declare them one to another for mutual utility and conversation” (L 4.1). This sug-
gests perhaps that there is a sort of mental language that is translated into an exter-
nal one.6 But our power of thought would, in Hobbes’s picture, be severely limited
if we lacked public, external language. In particular, we would lack the power of
universal expression, indeed cognition, which is only enabled by language.7

3.1.1. The Elements of Law (1641)


Already by the early 1640s, Hobbes believed that there were no universal things,
only universal names. In chapter 5 of The Elements of Law, having introduced names,
Hobbes distinguishes between universal and singular names: singular names name
one thing, while universal names name more than one thing.8 As examples of singu-
lar names Hobbes gives ‘Socrates’ and ‘he that writ the Iliad’, so definite descriptions
count as names. As an example of a universal name he gives ‘man’, which is a name
given “to every particular of mankind” (EL 5.5), that is, to every individual man.
Having distinguished the two sorts of name, Hobbes goes on to note that
some people have thought there to be universal things, as a result of noticing that
there are universal words (EL 5.6). On this view

besides Peter and John, and all the rest of the men that are, have been, or
shall be in the world, there is yet somewhat else that we call man, (viz.)

discussions of language, there are still strong echoes of it. He was still, for instance, traditional
enough to talk of “the five names that Porphyry expounds in his Isagoge” (Hobbes 1994a, 108).
6. For a useful discussion of Hobbes on mental language, see Pécharman 1992. For a reading
of Hobbes that emphasizes the importance of “mental discourse,” see Hacking 1975, 15–​25.
7. It is therefore misleading to say that “Hobbes’s view about what is a universal is orthodox
Ockhamism” (Martinich 1981, 357).
8. On Hobbes’s views about the signification of singular names, see Duncan 2011.
4 4   •  The Problem of Univer sals in Early Modern philosophy

man in general, deceiving themselves by taking the universal, or general


appellation, for the thing it signifieth. (EL 5.6)

That is, on this view there is some sort of universal thing man, as well as the uni-
versal name ‘man’.9 But this view is, Hobbes thinks, wrong. The name is univer-
sal, but the thing signified is not universal, for there are no universal things to
signify.10
In support of his view, Hobbes imagines someone asking a painter for “the
picture of a man, which is as much as to say, of a man in general” (EL 5.6). All that
is being asked for, Hobbes says, is a painting of some man or other. There is no
other thing, man in general, of which one wishes a picture. Hobbes’s opponents
might however handle the example without much difficulty. To ask for a picture
of a man, they might say, is just to ask for a picture of some man or other. There
is another thing, the universal man, but it is not what “paint me a picture of a
man” is asking for a picture of. You might do that with another request: “paint
me a picture of the universal man.” But we don’t do that, perhaps just because the
universal man is not the sort of thing that can be pictured.
Hobbes frames this discussion in Elements of Law 5.6 in a somewhat limited
way. The existence of universal names “hath been the cause that men think” that
there are universal things. Thus, Hobbes is trying to undermine a certain motiva-
tion for a sort of realism about universals. The realists can find a way to deal with
his particular example. But still, he might ask, why should we ever get to the point
of believing and defending the view that there are universal things? What is the
good initial motivation?11
In the background of Hobbes’s discussion—​and underlying his doubts
about motivations for realism—​is, perhaps, his basically imagistic view about
thought. If we think using images that resemble the objects thought about,
then there will be no way to think about something of which we can produce
no resembling image. So there will be no way to think about any alleged uni-
versal object that cannot be pictured. And if you thought this, you might con-
sider using it as the basis of an argument against belief in such things. It might
go as follows: we can only think using images; but we can have no images of
alleged universal objects; so we cannot think of alleged universal objects; but
you should not believe in things you cannot think about; so you should not

9. Hobbes leaves it rather vague as to what sort of view he had in mind to oppose. The most
obvious candidate though is the “realism” of scholastic Aristotelian debates.
10. In this text Hobbes appears to talk interchangeably of naming and signifying.
11. This would presume of course that realism requires motivation, and is not the default to
which we should return, absent other motivation.
Hobbes, Universal Names, and Nominalism  •  45

believe in universal objects.12 Such an argument for nominalism is not explicit


in this text. But Hobbes’s thinking along these lines would explain why he
thinks that the example of what can be painted should persuade us that there
are no universal things, only universal names. A  painting that resembles the
painted object is not, for Hobbes, just one example of representation. Rather,
it is an example that is very much like the most basic sort of mental representa-
tion, mental images resembling the things they represent.13
Related to that, it is notable and curious that there is no argument against
universal ideas here. Indeed it is not completely clear, from The Elements of Law
alone, that Hobbes was by this point committed to the denial of universal ideas.
That said, just as there is no argument against them, there is also no explicit
acknowledgment of their existence. An imagistic argument might perhaps,
indeed, be extended to provide an argument against universal ideas, as well as
against the universal things that are his explicit target. Hobbes appears to have
believed that representation was resemblance, showing this with his argument
that that there is no idea of God, because we have no image of him.14 So one
might think it was open to Hobbes to argue that there is no universal idea man.
Pictures of men all resemble particular men. There is no universal image that
resembles men in general, without resembling any particular one. So there is no
such mental image, no such idea. Mind you, that is far from a watertight argu-
ment. If nothing else, we should note that one image can resemble more than
one thing. A picture of a black cat might well look like several different black
cats, and thus be said to represent them, given a resemblance theory of represen-
tation. If a physical picture can do that, why not a mental one? And if a mental
picture can do that, why can it not function as a universal idea? Even with a
theory of ideas based on images and resemblance, there appears to be room for
universal ideas.15

12. That argument works similarly to the way in which it is sometimes suggested Hobbes argued
for materialism, starting from the fact that unextended incorporeal objects cannot be pictured.
See Duncan 2010 on Leibniz reading Hobbes this way. On the nominalism-​materialism con-
nection, see also Zarka 1985.
13. Hobbes does allow that there is at least one way of thinking about things that does not
involve resembling images. This is the point of his repeated story about how we can think
about God as the cause of the world, by analogy with the way in which a man born blind can
think of fire as the case of heat (Descartes 1984–​85, 2:127, L 11.25).
14.  See the Third Objections, and Duncan 2005. The imagism was persistent, not just a
feature of the Third Objections. Thus Hobbes wrote to Mersenne in 1648 that “This little
book also contains some expressions which do not produce any mental images of things”
(Hobbes1994a, 167).
15. See Laird 1934, 147.
4 6   •  The Problem of Univer sals in Early Modern philosophy

3.1.2. Other Texts of the 1640s


Though The Elements of Law was Hobbes’s one systematic work of philosophy of
the 1640s, there are other works of the time in which he discussed the relevant
workings of language.16 For example, Hobbes briefly presented his nominalist
view in a 1641 letter to Mersenne. There he says that

one must know that although the name ‘man’ is a common name (one, in
fact, of the five names that Porphyry expounds in his Isagoge), every man
is either Peter or Socrates or some other individual. … Socrates and man
are not two men, nor two things, but one man described by two names
(since it is the same thing which is named ‘Socrates’ and named ‘man’).
(Hobbes 1994a, 108)

That is, the objects named by ‘man’ are all individuals. There is no universal
object man, though there is a universal name. This is the same line as taken
in The Elements of Law: acknowledging the existence of universal names, and
denying the existence of universal objects that correspond to them. The letter
again leaves open, however, the question of the existence of universal ideas of
some sort.
Another useful source is Chatsworth manuscript A10 (Hobbes 1973, 463–​
513), which has sometimes been thought to be Hobbes’s own work, but which
Malcolm (2002, 99–​101) has argued is in fact a set of Robert Payne’s notes on
Hobbes’s work.17 This text thus needs to be treated carefully, especially in its
details. But it does appear, in one way or another, to be a source of information
about what Hobbes thought at this time. And it contains the following highly
relevant passage.

Common names are attributed to many things, proper names to single


things. Common names, as they are the names of many things one by one,
not taken collectively, are called universals.
Therefore a universal is a name, not some thing existing in rerum
natura, or a phantasm of something formed in the soul. A universal is a
word, or a name of names. Thus when animal is said to be universal, it is

16. Hobbes did publish part of his Elements of Philosophy in the 1640s (De Cive) but the other
two parts did not appear for several more years.
17. On the dating of the text, see Malcolm 2002, 101–​03.
Hobbes, Universal Names, and Nominalism  •  47

the name itself that is universal, and ‘animal’ does not signify some uni-
versal thing.18

Here we see, as in The Elements of Law, the observation that there are univer-
sal names, and the denial that there are any corresponding universal things. We
also find something new, the denial that there is any universal phantasm (or idea).
We also see here, perhaps, the suggestion of an argument for Hobbes’s view,
as the denial of universal things and ideas is apparently supposed to follow
somehow from the views in the first paragraph. There are two ways in which
we might understand the connection between the two paragraphs. One is just
to take Hobbes to be pointing out that if his view that common names are the
only universals is correct, then alternative views that countenance universal ideas,
or universals in rerum natura, must be mistaken. That is no doubt a helpful
thing to have clarified, but clarifying it gives us no particular reason to believe
Hobbes’s view. Alternatively, perhaps we might see more of an argument here.
But why might common names being “the names of many things one by one,
not taken collectively” show us that there are no universal ideas or things? After
all, that observation alone does not rule out the existence of universal ideas or
things. Perhaps there is a sort of appeal to simplicity here. We know that there
are common or universal names. If we see that they work as Hobbes says they
do—​naming each and every one of the several things that they apply to, not the
collection of them or any other single thing—​we will see that we can explain the
phenomena of language with reference to only the things that everyone already
agreed existed, without needing to suppose or appeal to extra things.

3.1.3.  Leviathan (1651)
Chapter 4 of Leviathan provides another discussion of names. Again Hobbes dis-
tinguishes proper or singular names from common or universal names. Singular
names are each the name of one thing, while common names are each the names

18. Hobbes 1973, 465, my translation. In the original:

Nomen est, commune, quod pluribus rebus; proprium, quod rei singulari, attribuitur.
Commune, cum sit plurium rerum singillatim, non collective sumptarum nomen, uni-
versale dicitur.
Est igitur nomen hoc universale, non rei alicujus in rerum natura existentis neque
ideae, sive phantasmatis alicujus in animo formati; sed alicujus vocis, sive nominis
nomen. Sic cum animal dicatur universale, non res aliqua voce universalis significata,
sed ipsum nomen animalis est universale.
4 8   •  The Problem of Univer sals in Early Modern philosophy

of many things.19 And again Hobbes says that names are the only universal things.
A common name names several things:

some [names] are common to many things, as man, horse, tree; every of
which, though but one name, is nevertheless the name of divers particular
things; in respect of all which together it is called an universal, there being
nothing in the world universal but names; for the things named are every
one of them individual and singular. (L 4.6)

In Leviathan we have, then, a restatement of the view that names, not things, are
universal. With the statement that there is “nothing in the world universal but
names” we also have a clear denial of universal ideas, as well as of universal things.
In addition, Hobbes says a little about why a universal name is imposed on
those things it is imposed on. The reason lies in a similarity between those named
things.

One universal name is imposed on many things for their similitude in


some quality or other accident: and whereas a proper name bringeth to
mind one thing only: universals recall any one of those many. (L 4.7)

This view will be important in later discussions, for here we find some barrier
against the view that all use of language in Hobbes’s picture is completely arbi-
trary and we also find the reason why some have thought there is a hidden uni-
versal in Hobbes’s system.

3.1.4. Common Names in De Corpore


There is yet another account of common or universal names in chapter 2 of De
Corpore. In that chapter, Hobbes argues that names are signs of conceptions, but
names of things (DeCo 2.5–​6). The thing named is often a body, but could also
be an accident, a phantasm, a name, or another piece of language (a speech). After
laying out these views, Hobbes works through a variety of distinctions between
kinds of names, such as positive and negative, first intention and second inten-
tion, and absolute and relative.
Among the distinctions discussed is that between common and proper names
(DeCo 2.9). The distinction is made in a way very similar to that in which it

19. Universal names are said in L 4.6 to name many things. In L 4.9 the extent of application of
different terms is described in terms of larger signification. This appears to be another case in
which signification language is used interchangeably with naming language.
Hobbes, Universal Names, and Nominalism  •  49

was made in previous texts. Common names name many things, whereas proper
names name only one thing.

But, because a common name is the name of many things taken one by one
but not of all things collectively at the same time (as ‘man’ is not the name
of the human race, but of each member, such as Peter, John and the other
men separately) it is called universal for that reason. Therefore the name
‘universal’ is not the name of some thing existing in nature, nor of an idea
or of some phantasm formed in the mind, but is always the name of some
vocal sound or name. (DeCo 2.9)20

Here we have perhaps Hobbes’s clearest statement of the view that only names
are universal, and that no idea, and no other thing, is universal. We also have a
suggestion of an argument, very similar to the suggestion of one that was seen in
manuscript A10. In both cases, we are first told that universal names are names
of several things singly, but not of the collection of them. Then we are told that
names are therefore the only universals, and there are no universal ideas, or uni-
versals in rerum natura.
As with the passage from the earlier manuscript, there appear to be two
ways one might understand this connection between claims in De Corpore. One
involves Hobbes simply making explicit a consequence of his view: if only names
are universal, then neither ideas nor anything else is universal. Alternatively, per-
haps this is supposed to be more of a persuasive argument than a clarification. But
again, how might this go? Well, again, we might speculate that Hobbes is offering
an argument from simplicity of explanation. If this is how universal names work,
there is no need to suppose there are universal ideas or things in order to explain
the phenomena of language. If ‘horse’ names this horse and that horse, and so
on, why suppose there are universal horses in the mind or elsewhere? The only
reason to suppose they existed, one might suspect, was to give “horse” something
to name. But we already believe in the horses themselves, and they turn out to be
what ‘horse’ names.

20. “Nomen autem commune, cum sit plurium rerum sigillatim sumptarum nomen, non autem
collective omnium simil (ut homo non est generis humani nomen, sed uniuscujusque ut Petri,
Johannis, et caeterorum hominum seorsim), vocatur ob eam rem universale. Est ergo nomen
hoc universale non rei alicuijus existentis in rerum natura neque ideae sive phantasmatis alicu-
jus in animo formati, sed alicujus semper vocis sive nominis nomen.” I depart from Martinich’s
(1981) translation in beginning the passage “But, because” rather than “Moreover, when.” It’s
not that some common names sometimes work in the way described—​rather, this is the way
that they all work in general.
5 0   •  The Problem of Univer sals in Early Modern philosophy

If that is what Hobbes was arguing—​and it is far from explicit that he was—​
how persuasive might it have appeared, and how ought we to judge it? Hobbes’s
audience was full of people who rejected universals in rerum natura, so his argu-
ing for that rejection would have been unsurprising, if hardly novel. The rejection
of universal ideas would have been more controversial. Indeed, they might have
looked like something we know by experience to exist, not like a theoretical pos-
tulate that could be rejected as unnecessary. Thinking that way, then, even if the
universal idea horse is not the thing named by ‘horse’, we still have good reason to
believe there is such an idea. Hobbes, of course, rejects that, even if his argument
being discussed here does not give others much reason to follow him in doing so.
And the text of De Corpore continues with what may be Hobbes’s most explicit
argument against universal ideas.

3.1.5. Another Argument in De Corpore


In the continuation of the above passage from De Corpore, Hobbes explicitly con-
siders the issue of what there is in our minds corresponding to common names.
There are, he says, “conceptions answering to” universal names in our minds,
though as in Leviathan common names are not said to signify those conceptions.
The conceptions answering to ‘animal’, for instance, are “images and phantasms
of individual animals.” When we hear a universal name, “we remember that
vocal sounds of this kind sometimes evoked one thing in the mind, sometimes
something else.”21 So when a common name is used it brings an idea of an ani-
mal to mind, and at different times it might well bring different ideas to mind.
Sometimes when I  hear ‘animal’ I  think of this cat, sometime I  think of that
giraffe. The most important point for Hobbes, perhaps, is that there is no univer-
sal idea that is brought to mind (or signified, or denoted, or named) by ‘animal’.
Nor indeed is an idea of a universal object brought to mind. Rather, only particu-
lar ideas of individual things come to mind. When we use a universal name, all
we think about are particular things, giving us reason to think that aside from the
name there are only particular things.
Why should one agree that no universal idea is brought to mind? Hobbes
seems to appeal to experience and introspection. If you think, he says, of what
comes to mind when you hear the word ‘animal’, you will realize it is only the idea

21. More fully: “voces eas animal, saxum, etc. esse nomina universalia, id est, nomina pluribus
rebus communia, et respondentes ipsis in animo conceptus sunt singularium animalium vel
aliarum rerum imagines et phantasmata. Ideoque non est opus ad vim universalis intelligen-
dam alia facultate quam imaginativa, qua recordamur voces ejusmodi modo unam rem, modo
aliam in animo excitasse.”
Hobbes, Universal Names, and Nominalism  •  51

of some animal or other. But why not suspect that, sometimes at least, the univer-
sal idea of animal comes to mind? Certainly, that is not visualized, as one might
mentally picture some particular animal. But its not being visualized is a problem
only if you think that visualization, and more generally imagination, is all there
is to the world of ideas. So Hobbes might well be appealing to experience, but
that appeal is bolstered by his commitment to an imagistic, imagination-​based,
account of our mental life.
What are particular and universal ideas in this discussion? What exactly is
Hobbes denying? Each idea, after all, is particular, one particular state of one per-
son’s mind.22 Any distinction between particular and universal ideas will have to
be made in terms of what and how the ideas represent. Certainly Hobbes thinks
there are no ideas that represent universal objects, because there are no universal
objects. But the denial of universal ideas means more than that—​to use a certain
terminology, Hobbes is denying conceptualism as well as realism. In Hobbes’s
view, the name is associated with ideas, but only with particular ideas of particu-
lar things. There is no special sort of idea that is associated with general names but
not with proper ones.23
One might suspect that Hobbes’s argument here is grounded in his imagism,
like the argument in The Elements of Law. In discussing his views there (see section
3.1.1 above) I noted the possibility that one image—​and thus one mental image—​
might resemble and represent multiple things. In De Corpore Hobbes explicitly
denies the existence of universal ideas: “they err … who say that the idea of some
thing is universal, as if there might be in the mind an image of some man which
is not that of any one man, but of man simpliciter; but this is impossible, for every
idea is both one and of one thing” (DeCo 5.8). The denial of universal ideas is
here. However, it is not clear whether Hobbes’s view had the resources to support
a complete denial of universal ideas. If representation is resemblance, then Hobbes
needs some way to rule out one image’s resembling (well enough), and thus repre-
senting, more than one thing. How that is to be done is apparently left unsaid.24
Indeed, what Hobbes argues against in the quoted passage is just the view that
there is an idea representing a universal thing. He appears not to engage with the
view that ideas are universal because they themselves represent multiple things.

22. “In the first sense a particular is that which is one and not many. Those who hold that a
universal is a certain quality residing in the mind which is predicable of many (not suppositing
for itself, of course, but for the many of which it is predicated) must grant that, in this sense
of the word, every universal is a particular” (Summa Logicae 1.14, 78; Ockham 1998, 77–​78).
23. Contrast, for example, Locke’s claim that each general word is “a sign of an abstract Idea in
the mind” (Locke 1975, 3.3.12).
24. On this problem, see again Laird (1934, 147–​48).
5 2   •  The Problem of Univer sals in Early Modern philosophy

As we have seen, Hobbes seems to argue from his views about ideas to his
nominalism. However, Hobbes goes on in De Corpore 2.9 to argue in more or
less the opposite direction, from his nominalism to there being only the imagi-
nation:  because the conceptions we have that relate to general names are only
particular ones, then there is no need to appeal to any faculty other than the
imagination here. Thus he says:

the conceptions answering to those things [universal names] in our minds


are the images and phantasms of individual animals or other things. And
for this reason there is no need to understand the force of a universal with
any faculty other than the imaginative one, by which we remember that
vocal sounds of this kind sometimes evoked one thing in the mind, some-
times something else. (DeCo 2.9)

This is in part an anti-​Cartesian point, in that Hobbes takes himself to show we


do not need to appeal to an ability to have intellectual insight, via clear and dis-
tinct perception, into the natures and essences of things. If universal thought,
via universal names, can be explained using the imagination, Hobbes can com-
plete his project of explaining all the workings of the mind on a corporeal basis,
given his view of (and a fair amount of agreement about) the material basis of the
imagination.
This is almost the exact opposite of using imagism about ideas to argue
for nominalism. That uses the fact that we can only conceive using the imagi-
nation as a reason for nominalism. This comment in De Corpore takes the
nominalism as a reason for there being only the imagination. But this and the
argument of The Elements of Law show how three of Hobbes central claims
fit together. His nominalism has us thinking only of particulars. His view
that the imagination is the only mental faculty has us thinking only with
images (which can only be of particulars, given some views about representa-
tion). And his materialism has us denying the existence of the sort of further
mental faculty that Descartes, for instance, thought must be housed in an
immaterial soul.

3.2. Ultranominalism
3.2.1. Three Critics of Hobbes
Hobbes was criticized in the seventeenth century for having gone, in a problem-
atic way, beyond nominalism. This criticism is associated with Leibniz, and his
claim that Hobbes is an ultra-​or supernominalist. Consider Leibniz’s “Preface to
Hobbes, Universal Names, and Nominalism  •  53

an Edition of Nizolius,” where Hobbes’s nominalism arises in a discussion of the


nominalist principle “the simpler a hypothesis is, the better it is.”25

From this principle the nominalists have deduced the rule that everything
in the world can be explained without any reference to universals and
real forms. Nothing is truer than this opinion, and nothing is more wor-
thy of a philosopher of our own time. So much so that, I believe, Occam
himself was not more nominalistic than is Thomas Hobbes now, though
I confess that Hobbes seems to me to be a super-​nominalist. For not con-
tent like the nominalists, to reduce universals to names, he says that the
truth of things itself consists in names and what is more, that it depends
on the human will, because truth allegedly depends on the definitions
of terms, and definitions depend on the human will. This is the opinion
of a man recognized as among the most profound of our century, and as
I said, nothing can be more nominalistic than it. Yet it cannot stand. In
arithmetic, and in other disciplines as well, truths remain the same even
if notations are changed, and it does not matter whether a decimal or a
duodecimal number system is used.26

Leibniz makes more than one point here. Hobbes, he says, has gone beyond the
views of earlier nominalists:  thus “super-​nominalist.”27 Leibniz is clearly con-
cerned that Hobbes’s view is too conventionalist, such that “the truth of things …
depends on the human will.” And he also makes the—​presumably related—​
critical point that “truths remain the same even if notations are changed.”
Though that sort of criticism of Hobbes’s nominalism is most famously asso-
ciated with Leibniz, we can find very similar points in the works of earlier critics
of Hobbes, namely Descartes and More. All three agree in thinking that Hobbes’s
view can be shown to be mistaken by reflecting on the fact that we can say the

25.  Leibniz 1969, 128. “Hypothesin eo esse meliorem, quo simpliciorem” (Leibniz 1923–​,
6.2:428). This is itself offered as an explanation of the principle that “entities must not be mul-
tiplied beyond necessity,” “Entia non esse multiplicanda praeter necessitatem.”
26. Leibniz 1969, 128. The final two sentences were added by Leibniz in a handwritten note
to his printed copy (see Leibniz 1923–​, 6.2:398). “Sed quae tamen stare non potest. Uti in
Arithmetica, ita et in aliis disciplinis manent eaedem veritates etsi notae mutentur, nec refert
decadica, an duodenaria progressio adhibeatur” (Leibniz 1923–​, 6.2:429).
27. On Leibniz’s claim of ultranominalism, see among others M. B. Bolton 1977 and Hübener
1977. For further discussion of Leibniz’s view in the preface of the Nizolius edition, see chapter
9 in this volume.
5 4   •  The Problem of Univer sals in Early Modern philosophy

same thing in different languages. Descartes’s version of this criticism is found in


his 1641 Replies to Hobbes’s Third Set of Objections to the Meditations.

As for the linking together that occurs when we reason, this is not a link-
ing of names but of the things that are signified by the names, and I am
surprised that the opposite view should occur to anyone. Who doubts
that a Frenchman and a German can reason about the same things,
despite the fact that the words that they think of are completely dif-
ferent? And surely the philosopher refutes his own position when he
talks of the arbitrary conventions that we have laid down concerning
the meaning of words. For if he admits that the words signify some-
thing, why will he not allow that our reasoning deals with this some-
thing which is signified, rather than merely with the words? And surely
on his account, when he concludes that the mind is a motion he might
just as well conclude that the earth is the sky, or anything else he likes.
(Descartes 1984–​85, 2:126)

Descartes is here responding to the following curious passage in Hobbes’s


Objections:

Now, what shall we say if it turns out that reasoning is simply the join-
ing together and linking of names or labels by means of the verb “is”? It
would follow that the inferences in our reasoning tell us nothing at all
about the nature of things, but merely tell us about the labels applied
to them; that is, all we can infer is whether or not we are combining the
names of things in accordance with the arbitrary conventions which we
have laid down in respect of their meaning. If this is so, as may well be the
case, reasoning will depend on names, names will depend on the imagi-
nation, and imagination will depend (as I believe it does) merely on the
motions of our bodily organs; and so the mind will be nothing more than
motion occurring in various parts of an organic body (Descartes 1984–​
85, 2:125–​26).

Hobbes here suggests several views he explains in more detail elsewhere:  that
much reasoning requires language; that language involves propositions that have
the basic structure “A is B,” where ‘A’ and ‘B’ are names; and that we can explain
the workings of the mind without reference to an incorporeal intellect if we
properly understand what imagination and language can do. Mixed in with that,
however, is the view that much of reasoning is arbitrary, because the use of names
is arbitrary. This is a view that could be controversial or trivial, depending on how
Hobbes, Universal Names, and Nominalism  •  55

it is understood. Moreover, there is a good question as to how the various claims


are supposed to be linked in the passage.28 What exactly is an argument for what?
Descartes makes two points in response to Hobbes’s claim. One is that we
are still reasoning about things, not just names, even though we are using names
to think about things. The other involves the example of the Frenchman and the
German. They can talk and reason about the same things, despite their use of
different words, which is in some sense arbitrary. Descartes thinks, however, that
Hobbes must deny that they can reason about the same thing.
A similar line of criticism is present in Henry More’s 1659 The Immortality of
the Soul. He too finds it a strange and problematic consequence of Hobbes’s views
that people talking in different languages cannot say the same thing. In the pas-
sage below, More attacks the Hobbesian view that we can explain the workings of
the mind using only imagistic thought and the power of language, thus needing
to appeal to no faculty other than the (corporeal) imagination.

Here Mr. Hobbs, to avoid the force of this Demonstration, has found out
a marvelous witty invention to befool his followers withall, making them
believe that there is no such thing as these Secundae Notiones, distinct
from the Names or Words whereby they are said to be signified; and that
there is no perception in us, but of such Phantasmes as are impressed from
externall Objects, such as are common to Us and Beasts: and as for the
Names which we give to these, or the Phantasmes of them, that there is the
same reason of them, as of other Markes, Letters, or Characters, all which
coming in at the Senses, he would beare them in hand that it is a plain case,
that we have the perception of nothing but what is impressed from cor-
poreall Objects. But how ridiculous an Evasion this is, may be easily dis-
covered, if we consider, that if these Mathematicall and Logicall Notions
we speak of be nothing but Names, Logicall and Mathematicall Truths
will not be the same in all Nations, because they have not the same names.
For Example, Similitudo and ὁμοιότης, ἀναλογία and Proportio, λόγος and
Ratio, these names are utterly different, the Greek from the Latine; yet
the Greeks, Latines, nor any Nation else, doe vary in their conceptions
couched under these different names: Wherefore it is plain, that there is a
setled Notion distinct from these Words and Names, as well as from those
corporeall Phantasmes impressed from the Object; which was the thing to
be demonstrated. (More 1659, 133–​34)

28.  A  significant part of the examination of Hobbes’s nominalism in M.  B. Bolton 1977 is
structured round how to understand Hobbes’s argument here.
5 6   •  The Problem of Univer sals in Early Modern philosophy

Hobbes argues that our thought can be explained in terms of phantasms


(mental images) and names. Such a view, More objects, cannot account for the
fact that Latin and Greek geometers are talking about the same thing when they
are talking about proportion, for instance. They do not possess the same sensory
images or corporeal phantasms. And they do not use the same words. Their abil-
ity to talk about the same thing is explained by the presence in their minds of
another sort of thing, a notion.29 The Greek and Latin geometers have the same
notion, which is related to each of their different vocabularies.
Sometimes we want to say that speakers of different languages are saying
the same thing, or talking about the same thing. On at least one occasion, as
Descartes noted in the Third Objections, Hobbes appeared to deny this. Leaving
that possible denial aside for now, one might wonder, as Leibniz and More did,
about how Hobbes could say that those two speakers were saying the same thing.
Hobbes clearly usually thought they could. Thus we sometimes find him in
Leviathan using his central semantic term “signify” to express a translation rela-
tion. Taking just one of many examples, we find him talking of “versura, which
signifies taking money at usury for the present payment of interest” (L 8.12). So
Hobbes generally agreed with the usual view that speakers of different languages
could say the same thing. But did his views about language have enough resources
to explain how this could be the case? This is what More and Leibniz followed
Descartes in contesting.
The passage from More shows one way in which Hobbes could have
explained this phenomenon, but did not. The French and German speakers
might be using different names, but associating those names with the same
notion. Thus, even in cases in which there is no external object they are both
naming and discussing, and despite the obvious differences in what they say,
there would be something that their talk had in common.30 But this is just the
sort of thing that Hobbes denies, when he argues that there are no universal
ideas. There are no universal ideas—​be they of proportion or elephant. Rather
there are only particular ideas of particular things, and different speakers, even
of the same language, might well have different particulars in mind when using
the same general name.31 Speakers of the same language at least have the name
in common, but speakers of different languages lack even that. So is there really

29. Compare also the concepts of common reasons invoked by Ward in response to Hobbes
( Jesseph 1999, 217).
30. This presumes, of course, that we can make good sense of different speakers possessing the
same notion, presumably by possessing different token notions of the same notion type.
31. Indeed, Hobbes thinks that one individual speaker need not have the same particular idea
in mind each time she uses a universal name.
Hobbes, Universal Names, and Nominalism  •  57

nothing that their use of language has in common, that Hobbes can point to, in
order to explain how it is, on his view, that speakers of different languages can
say the same thing?
Consider an example: cat, Katze, and chat are all terms for the same sort of
animal, but in different languages. Is the fact that speakers of English, German,
and French all appear to succeed in talking about the same animals, cats, just a
giant accident, according to Hobbes? Not really. For Hobbes thinks that there
are objective similarities in the world, and our grouping objects together and
applying general names to them depend on recognizing those similarities. As he
puts the point in Leviathan, a “universal name is imposed on many things for
their similitude in some quality or other accident” (L 4.7). So the speakers of
French, German, and English have two things in common. Their names apply
to the same things, and they do so because the speakers of the three languages
have all recognized the same objective similarity in the world. The same story
can be told about proportions: one instance of proportion is relevantly similar
to another, and speakers of many languages are recognizing this, even as they use
different terms to talk about it.

3.2.2. Similarity
On Hobbes’s view, then, the application of a universal name to multiple objects is
explained, at least in part, by the similarities between those objects. This view can
be traced through several works.
Consider first the account in The Elements of Law. “The appellations that be
universal, and common to many things, are not always given to all the particu-
lars, (as they ought to be) for like conceptions and considerations in them all”
(EL 5.7). Hobbes is here discussing equivocation. But he also has a picture of
how things ought to be done. A nonequivocal word, say ‘horse’, is applied to all
the things it is applied to, “for like conceptions and considerations in them all.”
We see that the horses all look like one another, and as a result we give them a
common name.
Something similar is said in Leviathan. “One universal name is imposed on
many things for their similitude in some quality or other accident” (L 4.7). This
might appear to change the story a little. Whereas The Elements of Law talks
about a similarity in our conceptions, that in Leviathan talks about a similarity
in the objects. But in the passage from The Elements of Law, the relevant “con-
ceptions and considerations” are said to be “in” the things the name is applied
to, so perhaps Hobbes is not really, there, talking about mental representa-
tions. Moreover, knowledge of similarity of things will be based on knowledge
of phantasms. And Hobbes apparently has a pretty simple conception of the
5 8   •  The Problem of Univer sals in Early Modern philosophy

relationship, such that if we judge the phantasms to be similar, we judge the


objects to be similar.32
Finally, there is a related account in De Corpore, presented when Hobbes
describes his distinction between positive and negative names.

The first distinction of names will be that some are positive or affirmative,
others negative, which are usually called privative and infinite. Those are
positive which are imposed because of the similarity, equality or identity
of the things thought; those are negative which are imposed because of the
diversity, dissimilarity or inequality of the things thought. Examples of
the former are ‘man’ and ‘philosopher’, for ‘man’ denotes any one of many
men and ‘philosopher’ denotes any one of many philosophers because of
the similarity of all of them. ‘Socrates’ likewise is a positive name because
what it denotes is always one and the same. (DeCo 2.7)

Here, of course, the context is slightly different from that of the above two
passages, as Hobbes is describing all positive names, including positive proper
names. The role of identity here relates to the case of positive proper names, and
the distinction between ‘Socrates’ and ‘not-​Socrates’. Still, the relations of “simi-
larity, equality or identity” appear to play roughly the same role that “similitude
in some … accident” plays in the account in Leviathan. In De Corpore as in
the earlier works, Hobbes thinks of common names as imposed on the objects
it names because of a preexisting similarity between the objects on which it is
imposed.
There is, then, a consistent account in Hobbes’s writings of why a com-
mon name is applied to several objects:  it is because of a similarity between
them. What is similarity? Hobbes addresses this issue briefly in chapter 11 of
De Corpore, which looks at issues of identity and difference.

We must not, however, think about relation, as if it were an accident dif-


fering from all the other accidents of the relatum, but as one of them,
namely, that [accident] with respect to which the comparison is made. For
example, the similarity of one white to another white, or its dissimilarity
to black, is the same accident as its white; and equality and inequality are
the same accident as the magnitude of the thing compared, under differ-
ent names; for that which is called white or great when it is not compared

32. See DeCo 25.8, which switches back and forth between knowing similarity of phantasms
and knowing similarity of things.
Hobbes, Universal Names, and Nominalism  •  59

with something else, is called like, unlike, equal, or unequal when it is


compared.33

On this account, the world contains (1) accidents expressed by one-​place predi-


cates, (2) the things that are the subjects of accidents, and (3) minds that make
comparisons between the things. But the things themselves do not contain
(4) further accidents that are their relations. Similarity is among those relations.
So while Hobbes grounds the application of universal names in similarity, simi-
larity itself is grounded in—​indeed is nothing in the world in addition to—​the
accidents of the individual similar things.

3.3. Another Objection
The above section considered an objection to Hobbes’s nominalism that was com-
mon in the seventeenth century. Hobbes’s views about similarity and accidents,
and their role in the workings of general names have prompted more recent vari-
ous commentators to object that his view is inconsistent. I would like to conclude
by looking briefly at this objection.
Watkins, for example, thought that Hobbes was not consistent in his nomi-
nalism, because he admitted accidents into his ontology.

[Hobbes] not only admitted resemblances between things, but resem-


blances with respect to certain properties or accidents; and he said, in effect,
that a common name gets extended to new objects, not arbitrarily, but in
accordance with such objective resemblances. (Watkins 1965, 148)

Thus, accidents play an essential role in Hobbes’s system. But accidents are,
Watkins thinks, a sort of universal.34 So Hobbes was inconsistent in his nominal-
ism, because he had to accept these universals to make his view work.
This sort of criticism has been made repeatedly. There are two slightly dif-
ferent versions. One is the above one, that the similarity or likeness involves

33. DeCo 11.6, my translation. “De relatione non ita censendum est, tanquam ea esset accidens
aliquod diversum ab aliis relati accidentibus, sed unum ex illis, nempe illud ipsum, secundum
quod fit comparatio. Exempli causâ, similitudo albi cum alio albo vel dissimilitudo cum nigro
est idem accidens quod albedo ejus, et aequalitas vel inaequalitas idem accidens quod rei com-
paratae magnitudo, sub diversis nominibus; nam quod album vel tantum vocatur, quando non
comparator cum alio, idem comparatum dicitur simile, dissimile, aequale vel inaequale.”
34. At least, such universals play this role for “descriptive names” (Watkins 1965, 150), though
with moral language there is no objective basis to counteract the ultraconventionalist tendency
in Hobbes’s thought.
6 0   •  The Problem of Univer sals in Early Modern philosophy

comparison with respect to accidents, which are universals. The other version
alleges that the similarity or likeness itself is a universal: “to grant that speakers
can grasp the resemblance between two particulars sorted under the same name
seems, at a minimum, to commit the nominalist to the existence of an abstract
property of similarity, likeness, or resemblance” ( Jesseph 1999, 209).
In thinking about these objections, it is useful to remember that there have
been several debates and views under the heading of ‘nominalism’. We have already
seen this, to some extent, with Leibniz’s argument that Hobbes goes (problem-
atically) beyond Ockham. But there are other differences to note.35 The central
question for Hobbes concerns the status and semantics of universal terms, with
primary examples being the likes of ‘man’ and ‘animal’. Despite Hobbes’s avoid-
ance of much of the apparatus of medieval debates about realism and nominal-
ism, his questions and views have fairly clear connections to those earlier debates.
Less immediately obvious is the connection Hobbes’s view has to later debates
about the status of properties or qualities.
Callaghan (2001) emphasizes the importance of seeing there are multiple
debates about nominalism involved here. In a second debate, nominalism is “the
rejection of properties (attributes, characters, features, qualities—​the name doesn’t
matter)” (Callaghan 2001, 37). And in this debate too, some of the “nominalists”
try to explain things by invoking resemblance. Thus, Price (1953, 7–​32), for exam-
ple, contrasts a Philosophy of Universals with a Philosophy of Resemblances.36
This resemblance-​invoking view would seem to have connections to Hobbes’s.
It is a sort of nominalism, based on the resemblances between everyday objects.
But the question Price is describing answers to—​about the ontological status of
recurring qualities in the world—​is not the question that Hobbes was directly
addressing when presenting his nominalism. Whereas Hobbes is primarily wor-
ried about the semantics of kind terms, and focuses on examples such as ‘man’,
Price is primarily worried about the nature of qualities, and focuses particularly
on colors.
How is this distinction between debates relevant to the objections made
to Hobbes’s view by Watkins and others? Well, Hobbes is addressing the first
debate, but the objection is considering Hobbes’s views in the terms of the second

35. There are yet other senses of ‘nominalism’ and ‘universal’. The more recent use of ‘nominal-
ism’ to mean the denial of abstract objects is not relevant here. For that use see for example
Field (1980), which is described as a “Defense of Nominalism” because it shows it how we
can have “Science without Numbers,” or Burgess and Rosen (1997, 3), where a “nominalist,
in the most common contemporary sense of the term” is described as “a disbeliever in abstract
entities.”
36. For another resemblance-​based nominalist answer in this second debate, see Rodriguez-​
Pereyra (2002).
Hobbes, Universal Names, and Nominalism  •  61

debate. Hobbes is said to be an inconsistent nominalist because he believes in


accidents, which are universals. But they are universals in the terms of the second
debate, about the status of qualities, not the debate that Hobbes was explicitly
addressing. Thinking in terms of the second debate gives rise to the two versions
of this objection to Hobbes. One might note that he believes in accidents with-
out attempting to reduce them to resemblances, and conclude he was not a nomi-
nalist. Or one might acknowledge some attempt in Hobbes’s system to explain
things in terms of similarity, but think that the similarity on which this depends
is itself a universal.37 Either way, in the terms of the second debate, there is a lin-
gering universal in Hobbes’s system. But that second debate was not the debate
Hobbes was addressing. It is no coincidence that Hobbes himself did not refer
to accidents as universals, and showed no signs of having thought that accidents
(nor, indeed, similarity or likeness) were, in a sense he cared about, universals.
And as Callaghan says, “it is not entirely clear that someone who rejects proper-
ties [in the debate Hobbes was in] is thereby bound … to reject universals [in
the other debate]” (Callaghan 2001, 37). Given this distinction between the two
debates, it is tempting to conclude that Hobbes’s critics have confused two senses
of ‘nominalism’. Hobbes is being criticized for failing to be a nominalist in a sense
in which he never tried to be one.38 No doubt Hobbes’s nominalism has its prob-
lems. But it is not so clear that this relatively recently emphasized difficulty is one
of them.

Abbreviations
EL = Hobbes 1990; cited by chapter and paragraph.
L = Hobbes 1994b; cited by chapter and paragraph.
DeCo = Hobbes 1981, 1999; cited by chapter and section.

37. This relates to an objection that Russell made to nominalist views that invoke resemblance
in chapter 9 of The Problems of Philosophy (Russell 1990, 95–​97). Rodriguez-​Pereyra (2002,
105–​23) is one nominalist attempt to respond to that criticism, discussed by MacBride (2004).
38. If we must attribute to Hobbes a position in the second debate, which should it be? Various
objectors, as we have seen, take Hobbes to believe in universals. But this is not necessarily cor-
rect. Callaghan (2001) argues that Hobbes was a sort of trope theorist. And we should also
consider whether we might understand Hobbes as a sort of resemblance nominalist.
4 SPINOZA ON UNIVERSALS

Samuel Newlands

Like many prominent early moderns, Spinoza espouses a brand of


nominalism about “abstractions and universals,” and he frequently
warns against confusing universals with real things. While many
of his conclusions about the status and origins of universals were
increasingly common in the seventeenth century, Spinoza insists that
the consequences of falsely reifying universals reach farther than his
contemporaries recognized. Spinoza also tries to integrate his criti-
cisms of reified universals into distinctive tenets of his own meta-
physics, epistemology, psychology, and even ethics. At the same time,
however, Spinoza employs universal-​like categories in very reifying-​
sounding ways, raising concerns about whether Spinoza fully abides
by his own admonitions. This too is part of an increasingly com-
mon pattern in early modern discussions of universals: reject mind-​
independent universals in one domain while appearing to tacitly
accept them in others.
In this chapter, I will begin by looking at Spinoza’s account of uni-
versals and focus on what he takes to be their ontological status (sec-
tion 4.1) and psychological origins (section 4.2). Although Spinoza
is not always clear on the metaphysical details, he is a kind of concep-
tualist (to use older terminology) and a resemblance trope nominalist
(to use more contemporary terminology). I  also examine Spinoza’s
more distinctive accounts of the confused origins of universal notions
and the limited positive role they can play in our cognitive lives. In
section 4.3, I  turn to Spinoza’s critique of universals and highlight
what he takes to be the dangerous and widespread consequences of
falsely reifying abstractions. In the final section, I raise a worry about
internal consistency. I  focus on Spinoza’s account of attributes and
common notions, and I  suggest ways to mitigate some—​but only
some—​of the tension between these doctrines and Spinoza’s claims
about universals.
Spinoza on Universals  •  63

4.1. The Ontological Status of Universals


In his early writings, Spinoza lumps universals together with other “abstractions”
like species and numbers, and he categorizes them all as entia rationis, things
whose existence depends on the existence and activity of a (finite) mind.1 He
contrasts these mind-​dependent entities with particular or singular things, entia
reale.2 Spinoza claims that universals are nothing but modes of thought, the same
conclusion that Descartes had reached in the Principles.3
While this rules out so-​called Platonic or “extreme realist” accounts of univer-
sals, it leaves unclear which of the other traditional positions on universals Spinoza
would have endorsed. After all, most prominent scholastics agreed that universals
were mind-​dependent beings, but they thought such dependence was consistent
with a kind of realism about universals.4 Spinoza does not show much interest in
the details of the long history of medieval disputes about universals in the Latin,
Jewish, and Islamic traditions. Like many other early moderns, Spinoza’s occa-
sional references to medieval sources on universals are vague and fairly general.
Admittedly, classifications of medieval views on universals, including those made
by later commentators, are themselves frustratingly inconsistent and ambiguous.
Hence, it will probably be more illuminating to compare Spinoza’s views to those
of particular historical figures, rather than to employ amorphous categories like
“nominalism,” “conceptualism,” “Aristotelianism,” and the like5—​with the caveat

1. I discuss Spinoza’s early views on universals and, more generally, abstracta and abstract think-
ing in much more detail in Newlands 2015. Some of the material in that paper overlaps with
parts of this one.
2. For some examples, see TIE 99–​100, G 2:36; KV II.16, G 1:82–​83; CM I:1, G 1:234–​36.
3. CM I.1, G 1:233–​34; Descartes, Principles I.58, CSM 1:212, AT 8(1):27.
4. For a survey of examples, see DM VI.ii.1, VI.vi.5, VI.vi.12, and VI.vii.2.
5.  The term ‘nominalism’ suggests a view according to which universals are only names or
words, but it is very hard to find a premodern philosopher who endorsed that view. As it is
more often used, ‘nominalism’ is a catchall term that refers to the large range of views that reject
realism about universals, though exactly what counts as “realism about universals” also varies
across classification schemes. “Conceptualism” is often presented as a more moderate version
of nominalism, one that accepts the existence of universal concepts but denies that universals
exist in things. “Aristotelianism” is often presented as a form of moderate realism about univer-
sals, one that accepts universal names, concepts, and universalia in rebus, but denies the exis-
tence of universalia ante res (i.e., “Platonism”). However, depending on how one cashes out the
difference between conceptualism and scholastic Aristotelianism, it can easily seem like most
scholastic Aristotelians, including Aquinas, Ockham, and Suárez (not to mention Aristotle
himself ) are conceptualists—​and therefore not Aristotelians! (On this classification, Scotus
looks like the closest to an “Aristotelian” among the major scholastics.) Even worse, when con-
ceptualism is understood as a form of nominalism, the charge of conceptualism can become
toxic, though often this is the result of terminological slipperiness and innuendo, rather than
careful reconstruction of views.
6 4   •  The Problem of Univer sals in Early Modern philosophy

that there is little reason to think Spinoza actually knew or cared much about
particular premodern views of universals.
Nevertheless, for those who value such classificatory schemes, there is a strong
case to be made that Spinoza was a conceptualist about universals, that is, someone
who believes that universals are only mental states (i.e., concepts that can denote
multiple particulars) and are, at most, merely occasioned by mind-​independent,
particular things.6 Spinoza claims in the metaphysical appendix to his book on
Descartes that “there is no agreement [convenientiam] between an ens reale and
the ideata of an ens rationis” (CM I.1, G 1:235.30–​31), meaning that the represen-
tational content of these mental states does not directly correspond to the mind-​
independent nature of the represented objects. Spinoza also asserts that things
like universals are not “in nature” (KV I.10, G 1:49.5), nor are they “inferred from
anything real” (TIE 99; G 2:36.19). Instead, they are “only our own work” (KV
I.10, G 1:49.5–​6). In the Ethics, Spinoza adds that universals do not “indicate
the nature of anything [real]” (E Iapp, G 2:83.12–​13), and they “indicate noth-
ing positive in things, considered in themselves, nor are they anything other
than modes of thinking, or notions we form because we compare things to one
another” (E IVpref, G 2:208.8–​11).
While these passages indicate Spinoza’s distance from realists of all stripes,
Spinoza also avoids going as far as some nominalists. He claims that entia rationis
like universals are not “fictitious beings,” which means that they do not “depend
on the will alone, nor do [they] consist of any terms connected with one another”
(CM I.1, G 1:237.13–​16). I take this to be Spinoza’s way of saying that universals
are not merely verbal or “nominal” entities and that their content is not purely
conventional. This rules out a position like the one endorsed by Hobbes (whom
Leibniz famously called a “super nominalist”) and places Spinoza in the company
of virtually all other prominent early modern philosophers.7
Although this brand of conceptualism may seem like a sharp departure from
premodern views about universals, it was already the predominant view among
late scholastics by the start of the seventeenth century. For illustration, consider
the closeness of Spinoza’s ontology of universals with that of Suárez, the great
sixteenth-​century scholastic metaphysician. Like Spinoza, Suárez accepts that
“everything which exists is necessarily singular and individual” (DM VI.ii.2).

6. Like Spinoza himself sometimes does, I will set aside his doctrine of parallelism for ease of
expression, but everything I say in this chapter can be recast in parallel-​friendly terms, if one
wishes.
7. See Hobbes 1994b, chap. 4, and Leibniz 1969, 128, and other essays in the latter volume. For
Hobbes’s position, see chapter 3 in this volume, and for Leibniz’s view of Hobbes, see chapter 9.
Spinoza on Universals  •  65

Universals, according to Suárez, exist only in the mind, and universal natures are
distinct from particulars only by a distinction of reason.
To be sure, Suárez repeatedly claims that although the unity of a universal
across particulars “arises through the activity of the intellect,” nonetheless “the
ground or occasion is taken from the singular things themselves [ex ipsis rebus sin-
gularibus fundamento seu occasione]” (DM VI.v.1; emphases mine).8 This appeal
to an in rebus ground for universals might sound more realist than Spinoza’s posi-
tion, but the italicized phrase hints at just how deflationary this grounding is
for Suárez. Universality is grounded in things in the sense of being occasioned
by things, a very weak kind of dependence that early modern conceptualists like
Descartes and Spinoza could accept.
Furthermore, when Suárez spells out what it is in things that grounds or occa-
sions the mind’s creation of universals, he claims, “there is merely something in
this [particular nature] to which something is similar in the other nature; how-
ever, this is not real unity but similarity” (DM VI.ii.13). As Suárez emphasizes a
little later, “they are grounded in the things themselves, not insofar as the nature
has any universality in the things, but insofar as there is in the individuals them-
selves agreement and similarity in essence and its properties” (DM VI.v.3). In
other words, objective similarities among particulars are what, in things, ground
the content of universal concepts. Hence for Suárez, universals, as mental con-
cepts, succeed in denoting the natures of individuals only extrinsically, in virtue of
“the non-​repugnance of the singular things themselves to having it possible that
other things be like them” (DM VI.v.3).9 As we will see, this sort of resemblance-​
based conceptualism is the position that Spinoza adopts as well. Not for nothing
does Suárez admit, “the nominalists … speak otherwise, although in reality they
do not differ much from us” (DM VI.v.3; see also DM VI.ii.1).
However, even within the family of conceptualists, Spinoza stands on the
more deflationary and radical end of the spectrum. Spinoza claims that only finite
minds use universals. God directly knows only singular, concrete things, though
God can know universals in virtue of knowing the representations of finite minds
(CM II.7, G 1:262–​63; KV I.6, G 1:43).10 And while Spinoza believes that uni-
versals are just confused finite mental states, he denies that universals are even
ideas (CM I.1, G 1:234.29–​30). From this Spinoza infers that predications of uni-
versals are neither true nor false: “Still these modes of thinking cannot be called

8. For other passages in which Suárez emphasizes the ground [fundamentum] of universals
(as well as common natures and genus or species) in things, see DM VI.ii.8, VI.iii.7, VI.v.3,
VI.ix.8, and VI.ix.21. For the most conceptualist sounding passage in Suárez, see DM VI.vii.2.
9. Spinoza makes a very similar claim about the universals good and evil in E IVp37s2, G 2:39.
10. Suárez makes a similar point (DM LIV.ii.23).
6 6   •  The Problem of Univer sals in Early Modern philosophy

ideas, nor can they said to be true or false, just as love cannot be called true or
false, but [only] good or bad” (CM I.1, G 1:235.17–​19).11 This is one reason why
Spinoza concludes that universals are not the proper objects of scientific investi-
gation, pace even conceptualist-​leaning scholastics like Suárez.12 In later sections,
we will see Spinoza distancing himself even further from mainline conceptualism.
If it is difficult to map Spinoza’s scattered and elliptical claims onto historical
camps, it is even more difficult to situate him in the expansive array of contempo-
rary positions on universals. I do not think Spinoza’s texts provide enough detail
to give us much confidence here, but if forced to speculate, I would put Spinoza’s
ontology of universals close to that offered in D.  C. Williams’s resemblance
trope nominalism (Williams 1953).13 I say this because I think that, for Spinoza,
(1)  properties, or modes, exist; (2)  all modes, including modes of modes, are
tropes (i.e., particularized properties); (3)  universals are nothing but modes of
modes (i.e., modes of a mind, which is itself a mode); and (4) similarities between
modes constitute the in rebus ground of universal concepts. I take (1) to be obvi-
ous, (2) to have been aptly defended by John Carriero (Carriero 1995), and (3) to
be clear from the preceding. The interpretive point most in need of defense is (4),
and I offer the following thin textual reed on its behalf.14
According to Spinoza, universals like “man, horse, dog, etc.” are formed when
the mind distinctly considers “only what [a group of particulars] all agree [conve-
niunt] in” (E IIp40s, G 2:121.13–​23). Admittedly, this might sound like Spinoza
is positing some additional thing, a property or universal, that several particu-
lars all share in common, pace his overarching conceptualism about universals.
However, Spinoza often uses “agreement” in a thinner sense that does not require
literal sharing or multiple instantiation.15 On this reading, singular things agree

11. Hobbes makes a very similar point (Hobbes 1994b, chap. 4, 11).


12. Cf. TIE 99, G 2:36, and DM VI.v.3.
13. Spinoza would surely deny the bundle theory of substances attached to Williams’s trope
theory, but that is distinct from Williams’s account of universals in terms of resemblance
among particularized properties, which I think Spinoza could/​would/​should affirm.
14. Descartes offers a similar account of universals in terms of resemblance in Principles I.59,
and it isn’t too much of a stretch to think that Spinoza intentionally echoes Descartes on this
in CM and in the Ethics. As Michael Istvan helpfully pointed out in correspondence, Leibniz
sometimes offers a similar analysis (Leibniz 1996, III.iii.12–​14).
15. See esp. E IIp13l2, G 2:98; E IIp37, G 2:118; and, more controversially, E IVp18s, G 2:223.
Suárez also used the language of agreement to denote mere similarity:  “the natures which
are denominated universals should be in singulars and the singulars themselves should have
among themselves something in which they agree or are alike [conveniant vel similia sint] and
something in which they differ or are distinguished” (DM VI.ii.1). Much more difficult for my
reading are passages like E IIp31, in which Spinoza appeals to a “common property of singular
Spinoza on Universals  •  67

with one another in the sense of having highly similar particularized properties or
modes.16 They agree in virtue of resembling one another.
In a similar vein, Spinoza claims that we employ universal notions when we
“recall something else familiar to us, which agrees with it, either in name or in
reality” (CM I.1, G 1:234.5–​7, emphasis mine). I think we should again under-
stand “agreement in reality” in terms of resemblance. In the very next sentence,
Spinoza appeals to collections of such resembling things as the true bases of pre-
modern universals: “Similarly, the Philosophers have reduced all natural things
to certain classes, to which they recur when anything new presents itself to them.
These they call genus, species, etc.” (CM I.1, G 1:234.8–​10).17
In short, some of the particular aspects of singular things more exactly resem-
ble aspects of other things, and collections of such similar aspects or things are
the basis of universal concepts—​which is just to attribute (d) to Spinoza. In con-
temporary metaphysics, admitting that the content of universals rests on objec-
tive similarities among tropes commits Spinoza to a nominalist position. In the
minds of some late medieval Aristotelians, the same admission would commit
him to a more realist position. This again says more about the plasticity of these
categories than it does about the looseness of Spinoza’s views.

4.2. The Origins of Universals


Spinoza provides a twofold account of the origin of universals that clarifies and
reinforces his deflationary ontology. In the first part, Spinoza explains the causal
source of universals, understood as finite modes of thought. In the second part,
Spinoza explains how universals nonetheless play an important role in our cogni-
tive lives, despite their lowly ontological status and disreputable origin. In both
parts, Spinoza retains some premodern claims, but he uses them to reach novel
and fairly radical conclusions.
In very general terms, Spinoza thinks the content of universals arises from
mental activities of abstraction and comparison. This is certainly not an original

things,” and E IIp39d, in which Spinoza claims that something may be common to multiple
bodies in a way that it is “equally in the human body and in the same external bodies.”
16. To make this consistent with Spinoza’s earlier claim that “there is no agreement” between
singular things and the ideata of universals and other entia rationis (CM I.1, G 1:235.30–​31),
we should understand him in CM to be denying that the content of universal concepts repre-
sent real things as they are in themselves (as literally sharing properties), not that real things are
objectively similar to one another.
17. See also Ep 12, G 4:57.3–​6 and CM I.1, G 1:235.22–​26: “But [Plato] referred man to a
certain class so that, when he wished to think about man, he would immediately fall into the
thought of man by recalling that class, which he could easily remember.”
6 8   •  The Problem of Univer sals in Early Modern philosophy

idea. Hobbes, Gassendi, and Descartes make very similar claims, and the role
of intellectual abstraction and comparison in forming universal concepts has a
rich heritage since at least Aquinas.18 For example, according to the influential
Thomistic account, when forming universals the active intellect extracts intel-
ligible species from material species (the “phantasms”) by stripping away indi-
viduating information from sensory representations via selective attention. This
combined effort of the senses and the intellect, when habitual, produces a con-
cept in the intellect whose content is nonindividuating, that is, general or “uni-
versal.” Three details of this account are worth highlighting. (1)  For humans,
universals are ultimately abstracted from sensory input, in accordance with gen-
eral scholastic empiricism. (2) Abstraction is an act of giving selective attention
to features of singular things that are only conceptually distinct from the singular
thing itself (ST I, q. 85, art. 1, ad 1 and ad 2). (3) Although the resulting universal
concept exists only in the intellect, it is nonetheless grounded in the natures of
mind-​independent, singular things (ST I, q. 85, art. 2, ad 2). As the slogan runs,
universals are formaliter in mente, but fundamentaliter in re.
Spinoza rejects many aspects of the Thomistic account, from the general
form/​matter empiricism to the more specific appeal to intelligible species and
phantasms (which Spinoza calls “bits of nonsense”) (Ep 56, G 4:261.34–​35).19
More radically, Spinoza also came to reject Aquinas’s view that universals are
formed through the activity of the intellect, and claims that they arise solely from
the imagination. They are, Spinoza writes, “only modes of imagining [that] do
not indicate the nature of anything, only the constitution of the imagination …
I call them beings, not of reason, but of the imagination” (E Iapp, G 2:83.15–​16).20
In the previous section, I claimed that Spinoza also rejects (3), at least on a
suitably strong reading of “founded in”—​though I also claimed that some late
scholastics deny (3) as well on such a strong reading. (I also indicated a very weak
sense that Spinoza could accept: universals are grounded in things in the sense
that objective similarities between particulars occasion the formation of univer-
sal notions by the mind.)
Spinoza does accept both (1) and (2) from the Thomistic account, though he
draws very different conclusions from them. Spinoza thinks that the mind forms
universal notions from bodily impressions caused by singular things. The mind
abstracts from the particulars by paying selective attention to what a collection of

18.  See especially Aquinas, ST I, q.  85, and Descartes, Principles I.59, CSM 1:212–​213, AT
8(1):27–​28. For an early version of this in Spinoza, see KV I.1, G 1:16–​17, note 3. For Hobbes
and Gassendi, see chapters 2 and 3 in this volume.
19. Translated by Samuel Shirley.
20. See also Ep 12, G 4:57.
Spinoza on Universals  •  69

those singular things “all agree in,” that is, their objective similarities (E IIp40s1,
G 2:121). But whereas Aquinas presents this activity as an ennobling feature of
humans, Spinoza takes the generation of universals to be inevitably full of confu-
sion and error.

These notions they call Universal, like man, horse, dog, etc. have arisen
from similar causes, viz. because so many images (e.g., of men) are formed
at one time in the human body that they surpass the power of imagining—​
not entirely, of course, but still to the point where the mind can imag-
ine neither slight differences of the singular [things] … and imagines
distinctly only what they all agree in, insofar as they affect the body.
(E IIp40s1, G 2:121.13–​20)

Spinoza thinks that when the body is so bombarded with sensory images of
individual bodies that it lacks the ability to process them distinctly, the mind
compensates by ignoring slight differences and focusing selectively on perceived
similarities. In other words, our mind represents this confusing array by abstract-
ing away from particular features of individuals and representing their similarities
as a single, more distinct image, such as humanity.
Spinoza also believes that the representation of similarities of individuals is
always more powerful and distinct than the representation of their respective dif-
ferences. He thinks this based on a somewhat crude mechanistic principle: “For
the body has been affected most [NS:21 forcefully] by this [viz. what is com-
mon], since each singular has affected it” (E IIp40s1, G 2:121.20–​21). In other
words, when I see ten cups that are similar in color but very different in sizes, my
abstracted representation of their color will be more affecting than my represen-
tations of their individual sizes, since I have ten sensory impressions of a nearly
identical color but only one of each size.
Spinoza concludes that universals are a kind of mental crutch, a way of coping
with the fact that “there are many things in nature whose difference is so slight that
it almost escapes the intellect” (TIE 76, G 2:29.9–​11). Linguistically, we “express
this [confusion] by the word ‘man’ and predicate it of infinitely many singulars”
(E IIp40s, G 2:121.22–​23). So although universals are formed through abstrac-
tion from sensory impressions, the confused nature of all such bodily impressions
means that universals will be only slightly more clear and distinct versions of con-
fused representations. Hence, not only are universals mind-​dependent for Spinoza,

21. NS = Nagelate Schriften, the Dutch edition of Spinoza’s Ethics, which includes material not
in the Latin edition.
7 0   •  The Problem of Univer sals in Early Modern philosophy

they are also the products of confused and inadequate images and representations.
In the Ethics, Spinoza singles out good, evil, confusion, warm, cold, beauty, ugliness,
will, and intellect as examples of universals formed in this way, a list that indicts
not only Aristotelian science, but also moral realism, traditional theism, and even
Cartesian philosophy of mind. In Spinoza’s hands, what had seemed like a fairly
standard starting point about the role of abstraction in forming universals turns
into a wide-​ranging critique of several central seventeenth-​century beliefs.
Although Spinoza thinks false accounts of universals have given rise to a host
of philosophical and theological errors, he admits that they play an ineliminable
role in our cognitive lives. Spinoza emphasizes that, in addition to helping us
avoid cognitive overload, universals are important mnemonic devices:

That there are certain modes of thinking which help us to retain things more
firmly and easily, and when we wish, to recall them to mind or keep them
present to mind, is sufficiently established for those who use that well-​known
rule of memory, by which to retain something very new and imprint it on
the memory, we recall something else familiar with it, which agrees with it,
either in name or in reality. Similarly, the Philosophers have reduced all natu-
ral things to certain classes, to which they recur when anything new presents
itself to them. These they call genus, species, etc. (CM I.1, G 1:234.1–​10)22

This explains how universals can be “good or bad” without being true or false: some
mnemonic devices work better than others. For example, grouping things together
by color is a better aid to recollecting particulars than grouping them by distance
from the sun.
The content of universal notions will also vary from person to person, given
the variability of our cognitive structures and circumstances. “But it should be
noted that these notions are not formed by all in the same way, but vary from one
to another, in accordance with what the body has been more often affected by,
and what the mind imagines or recollects more easily” (E IIp40s1, G 2:121.24–​
27). In slogan form, “each will form universal images of things according to the
disposition of his body” (E IIp40s1, G 2:121.33–​34).
From subject-​variability Spinoza infers that apparent disagreements over uni-
versal notions and real definitions aren’t actually disagreements at all:

So when Plato said that man is a featherless biped, he erred no more than
anyone else who said that man is a rational animal. For Plato was no less

22. See also TIE 82, G 2:31; Ep 12, G 4:56–​57; and CM I.1, G 1:233. Hobbes makes a similar
point (Hobbes 1994b, chap. 4. 3).
Spinoza on Universals  •  71

aware than anyone else that man is a rational animal. But he referred man
to a certain class so that, when he wished to think about man, he would
immediately fall into the thought of man by recalling that class, which he
could easily remember. (CM I.1, G 1:235.19–​26)

Spinoza repeats this example in the Ethics and concludes, “Hence it is not sur-
prising that so many controversies have arisen among the philosophers, who
have wished to explain natural things by mere images of things” (E IIp40s1,
G 2:121.31–​35).
This is an easy point to overlook, but it highlights one of Spinoza’s deepest
interests in the topic of universals. Spinoza thinks that once we see the confused
source of most predications of universals, we can correctly interpret what had
seemed like substantive disagreement over, say, the nature of God and human
beings as really just differences in the particular constitutions of our bodies—​a
difference hardly worth mob violence, war, and institutionally sponsored sanc-
tioning, to name some of Spinoza’s more pressing practical concerns.

And most controversies have arisen from this, that men do not rightly
explain their own mind, or interpret the mind of the other man badly.
For really, when they contradict one another most vehemently, they either
have the same thoughts or they are thinking of different things so that
what they think are errors and absurdities in the other are not. (E IIp47,
G 2:129.3–​8)

Realizing that universals do not carve the joints of reality but instead describe
the contours of different physiological and psychological persons helps rob them
of their power over us and promises a way of discussing divergent scientific, reli-
gious, and philosophical opinions in the congenial spirit in which physicians
discuss pathologies. Here we find a vivid example of just how tightly Spinoza
intertwines his ethics and metaphysics.

4.3. The Dangers of Universals


For Spinoza, the main danger of universals is that we tend to forget their true
ontological status and origin, and we reify them in ways that lead to all sorts of
confusion, misunderstandings—​and worse. In his early writings, Spinoza repeat-
edly warns against confusing universals and other entia rationis with real things.
The failure to do so, he claims, “interferes with the true progress of the intellect”
(TIE 99, G 2:36.19–​20), leads to “great errors, as has happened to many before us”
(CM I.1, G 1:236.4–​5), and is “something a true philosopher must scrupulously
7 2   •  The Problem of Univer sals in Early Modern philosophy

avoid” (KV II.4, G 1:60.31–​32). Spinoza warns that even Aristotelians who admit
that universals are entia rationis are guilty of confused reification in practice:

But this objection arises from ignorance, from the fact that men have
formed universal ideas. … They maintain, then, that these [universal]
ideas are in God’s intellect, as many of Plato’s followers have said, viz. that
these universal ideas (such as rational animal, etc.) have been created by
God. … And though Aristotle’s followers say, of course, that these things
are not actual, but only beings of reason, nevertheless they very often
regard them as things. (KV I.6, G 1:42.26–​35)

Spinoza offers three reasons why reification is common and easily done. First, we
are prone to reify abstractions like universals because they “arise from the ideas of
real beings so immediately that they are quite easily confused with them by those
who do not pay close attention” (CM I.1, G 1:234.31–​33). That is, the process
of abstraction from confused sensory representations is so easily missed that we
readily treat these mental constructions as real things:

For when things are conceived abstractly, as all universals are, they always
have a wider extension in our intellect than their particulars can really
have in nature. And then, since there are many things in nature whose
difference is so slight that it almost escapes the intellect, it can easily hap-
pen, if they are conceived abstractly, that they are confused. (TIE 76, G
2:29.7–​11)

The second reason is that, as we saw in the previous section, confused bodily rep-
resentations based on similarities often stand out more clearly and vividly than
do the representations of the discrete individuals. Third, Spinoza points out that
natural language easily misleads us into thinking that the referent of universal
terms like “man” has the same ontological status as the referent of singular terms
like “Peter,” since both terms seem to function in syntactically similar ways. This
is one reason to avoid reading ontology from our predications, a practice Spinoza
denounces as “judg[ing] the things from the words, not the words from the
things” (CM I.1, G 1:235.8–​9).23
Spinoza is hardly alone in warning against reifying abstractions like universals,
of course. Descartes concludes a letter about universals with a similar warning: “It

23. See also E IIp40s1, G 2:121; E IIp49s, G 2:132; and E Iapp, G 2:83.


Spinoza on Universals  •  73

seems to me that the only thing which causes difficulty in this area is the fact that
we do not sufficiently distinguish between things existing outside our thought
and the ideas of things, which are in our thought” (CSMK 280, AT 4:349). Still,
Spinoza’s rhetoric far surpasses that of his contemporaries. He claims that failing
to heed this warning leads one to “absurd fantasies” (CM I.1, G 1:236.5), “the
most absurd absurdities” (Ep 12, G 4:57.12), and “nonsense, not to say madness”
(Ep 12, G 4:55.13). More importantly, Spinoza takes the consequences of treating
universals, species, and other beings of reason as real things to be far more wide-​
ranging than Descartes and others appreciated.
Spinoza claims that a wide range of philosophical, theological, and scientific
views arise from confused reification of abstractions. Many of these views were
widely held by his fellow early moderns, all of whom claimed to affirm something
close to Spinoza’s conceptualist position. Spinoza offers a challenge to his fel-
low nominalists. They claim to accept the mind-​dependent, conceptualist status
of universals and to heed the warnings against reifying them. And yet, in other
domains, they tacitly accept the very same reification of abstractions. Spinoza’s
challenge is this: either reject all such reified abstractions, even if that means
rejecting popular and entrenched views, or else admit to being inconsistent. In
other words, Spinoza’s fellow early moderns claim to avoid reifying universals and
abstractions, but Spinoza thinks they do not do so consistently, or else they would
have rejected far more than the “easy cases”, such as Platonism about species and
numbers.
Here is a list of the philosophical problems and positions that Spinoza explic-
itly names as arising from false and confused reification of abstractions, starting
with more specific mistakes and moving to more general views:

• materialism about the soul (TIE 74, G 2:28);


• Zeno’s paradox (Ep 12, G 4:58–​59);
• privation theory of evil (Ep 19, G 4:91–​92);
• misunderstandings of infinity (Ep 12, G 4:59);
• realism about secondary qualities (E Iapp, G 2:81);
• incorrect views of Divine providence and knowledge (KV I.6, G 1:42–​43;
CM II.7, G 1:162–​63);
• false mechanistic physics (Ep 12, G 4:55–​56);
• libertarian accounts of human freedom (KV II.16, G 1:82; Ep 2, G 4:9; E
IIp49s, G 2:135);
• false views of perfection and imperfection (E IVpref, G 2:207);
• the problem of evil (KV I.6, G 1:43; E Iapp, G 2:83; E IVpref, G 2:206);
• forms of theological anthropormorphism (E Iapp, G 2:79);
74   •  The Problem of Univer sals in Early Modern philosophy

• faculty psychology (KV II.16, G 1:81–​3; E IIp48s, G 2:129; E IIp49s);24



blame, praise, sin, and merit (E Iapp, G 2:81);

objective aesthetics (E Iapp, G 2:82);

divine and natural teleology (E Iapp, G 2:81–​82; E IVpref, G 2:206);

moral realism (CM I.6, G 1:248; KV I.10, G 1:39; KV II.4, G 1:60; KV I.6, G
1:43; E IVpref, G 2:208);
• skepticism (E Iapp, G 2:82).

The most striking thing about this list is how wide-​ranging it is, applying to cen-
tral positions in science, theology, metaphysics, epistemology, aesthetics, and
ethics. Spinoza is rightly regarded as a systematic philosopher, someone who
uses a small set of principles to generate a comprehensive account of the world.
Here we see another aspect of his systematicity: he also tries to show how many
alternative philosophical views stem from violating a core set of principles.
Spinoza makes many of these charges in passing, so for the sake of space and inter-
est, I will focus on one of his more developed examples.25 Spinoza claims that those
who, like Descartes, postulate real and distinct faculties of will and intellect often do so
on the basis of a confused reification akin to what we are prone to do with universals:

[T]‌here is in the mind no absolute faculty of understanding, desiring,


loving, etc. From this it follows that these and similar faculties are either
complete fictions or nothing but metaphysical beings or universals, which
we are used to forming from particulars. So intellect and will are to this or
that idea, or to this or that volition as “stone-​ness” is to this or that stone,
or man to Peter or Paul. (E IIp48s, G 2:129.20–​26)

His basic point is clear. There exist singular volitions, desires, and so forth, just
like there exist singular bodies. These particular volitions have similarities to each
other, and they are often so similar that it is easy for the mind to form a represen-
tation of their perceived similarities. These representations constitute universal
notions like will and desire. Philosophers then attribute causal powers to these
reified universals, which remain mere entia rationis.

For because man has now this, now that volition, he forms in his soul a
universal mode which he calls the Will, just as he forms the idea of man

24. Lodewijk Meyer picks up on this point in his preface to Spinoza’s summary of Descartes’s
Principia Philosophiæ (G 1:132).
25. This appears in both KV and the Ethics. The KV version is considerably more detailed, but
the arguments are otherwise so similar that I will cite them interchangeably.
Spinoza on Universals  •  75

from this and that man. And because he does not sufficiently distinguish
real beings from beings of reason, it comes about that he considers the
beings of reason as things that are truly in nature, and thus posits himself
as a cause of some things. … For if you ask someone why man wills this
or that, the answer is: because he has a Will. (KV II.16; G 1:82.8–​1:83.2)

This is nonsense, according to Spinoza.

But since, as we have said, the will is only an idea of this or that volition
(and therefore only a mode of thinking, a being of reason, and not a real
being), nothing can be produced by it. For nothing comes from nothing. So
I think that when we have shown that the Will is no thing in Nature, but
only a Fiction, we do not need to ask whether it is free or not. (KV II.16;
G 1:83.2–​8; emphases in original)

Herein lies the origin of faculty psychology, Spinoza thinks. While engaging in
a perfectly good inquiry about the source of particular volitions, philosophers
confusedly reify a mere concept, ascribe to it causal powers and even a capacity
for freedom—​all while forgetting that, extra mentem, there is no such thing as a
will in the first place.
The most interesting thing about this argument against faculty psychology is
that it isn’t really an argument at all, at least in the sense of having premises that
are antecedently more convincing than is the denial of the conclusion. It is highly
unlikely that Descartes, for instance, would accept Spinoza’s premises that the
faculty of willing is posited because he reifies an abstraction, or that we should
be more confident in our lack of freedom than in the real existence of a source of
causal power within us. The whole thing reads more like a declaration of Spinoza’s
views rather than a defense of them.
This is a general feature of Spinoza’s charges of reification: they are wielded
as diagnoses, rather than as conclusions of an argument. Spinoza does not try to
prove that reification occurs in these cases; instead, he tries to show that a range
of positions that disagree with his own systematic conclusions plausibly origi-
nate from a confused reification like those committed by realists about universals.
However, if one does not already agree with Spinoza that, for example, “the par-
ticular willing [of ] this or that … must proceed from some external cause” (KV
II.16, G 1:82.1), one will not find his charge very worrisome nor his diagnosis
very convincing.
This may sound like a veiled criticism of Spinoza, but I do not intend it that
way. Spinoza’s approach here is precisely what we should expect from a systematic
philosopher. Very often, the plausibility of individual pieces of Spinoza’s views
7 6   •  The Problem of Univer sals in Early Modern philosophy

require accepting other, controversial claims he makes, which are themselves plau-
sible only if one accepts yet further controversial claims, and so on. (Coherence
is no substitute for correspondence, but it’s certainly a move in the right direc-
tion!) While there may be some basic and prima facie plausible entry points into
his system—​the Principle of Sufficient Reason, the contours of metaphysical
perfection, and perhaps a denial of extra mentem reality to universals—​many of
Spinoza’s claims and criticisms come as a package. It should be unsurprising that
Spinoza’s charges of reification are no exception.
Furthermore, although Spinoza does not put the matter quite this way, he
does raise an interesting challenge for his fellow early moderns. As I noted above,
many would have cheered his rejection of mind-​independent universals. Spinoza
points out, however, that the very same sort of reification error may lie behind
other, more cherished doctrines. For those who want to preserve such doctrines,
Spinoza issues the following sort of challenge: just try to defend, say, moral realism
without tacitly reifying goodness in the same way that Platonists reify numbers.

4.4. Questions of Consistency
Spinoza’s challenge in the previous section highlights a worrisome pattern in dis-
cussions of universals by prominent early moderns. When seventeenth-​century
philosophers focused explicitly on the general topic of universals, or on certain
instances of universals such as species and numbers, they were quick to denounce
even moderate realist accounts. Yet in other contexts, when their focus is else-
where, they appear to implicitly endorse a mind-​independent realism about enti-
ties that seem very similar, such as transcendentals (e.g., goodness, being, thing),
geometrical forms, essences, “true and immutable natures,” and so forth.
In short, early moderns do not treat all mind-​ independent universals
equally: some appear more acceptable than others. One reason for this uneven
tendency is that by and large the most prominent early moderns were not espe-
cially interested in the traditional problem of universals, especially in compari-
son to earlier scholastics or twentieth-​century analytic metaphysicians. Often an
attack on realist accounts of universals seems merely instrumental, performed in
the service of advancing a new mechanistic physics, for example. This is one rea-
son why their attacks on realist theories of universals often focus on examples
like natural kinds and infima species, essential elements in the superstructure of
Aristotelian science. There is also little evidence that these early moderns fully
grasped the difficulties of embracing an exceptionless form of nominalism, dif-
ficulties that arguably were not fully appreciated until the twentieth-​century
revival of questions about universals. It turns out that thoroughgoing nominal-
ism is very difficult to maintain, and so perhaps it is unsurprising that when their
Spinoza on Universals  •  77

philosophical energies were directed elsewhere, seventeenth-​century philoso-


phers regularly slipped back into more realist mindsets.
While that might help explain, even if not excuse, the tacit division of “good”
versus “bad” universals in philosophers like Descartes and Hobbes, what about
Spinoza? After all, he was well aware of potential inconsistencies on this topic,
and he tried to use it as leverage against rival views. So if it turns out that he too is
inconsistent, if he fails to abide by his own universal admonitions against reifying
“abstractions and universals,” if he too tacitly accepts some realist universals as
“good” without argument—​that would be especially problematic.
Unfortunately, Spinoza sometimes looks guilty of the same sort of inconsis-
tency he scorns in others. At times, he is even upfront about this. For example, in
his early treatise on philosophical method, Spinoza claims that we should investi-
gate nature “in such a way that we do not pass to abstractions and universals, nei-
ther inferring something real from them, nor inferring them from something real”
(TIE 99, G 2:36.17–​19). Just two paragraphs later, however, Spinoza confesses
that he needs something like universals after all. “So although these fixed and eter-
nal things are singular, nevertheless, because of their presence everywhere, and
most extensive power, they will be to us like universals, or genera of the defini-
tions of singular, changeable things, and the proximate causes of all things” (TIE
101, G 2:37.5–​8). In other words, Spinoza admits that his own ontology con-
tains singular, concrete things that function like reified universals and that can be
legitimately used in philosophical and scientific inquiries. In short, he appeals to
ersatz universals.
It is to Spinoza’s credit that he admits his need for something that is (1) one-​
over-​many, (2) distinct from existing concrete singulars, and (3) legitimately used
in the investigation of the world. However, these were among the very roles that
universals played in scholastic accounts! In the Ethics, Spinoza invokes three cate-
gories that correspond to these traditional functions for universals: (1) attributes,
(2) formal essences, and (3) common notions. However, to maintain consistency,
Spinoza needs to show either that the reification of these categories is consistent
with his conceptualist framework for universals, or else that he can capture the
functions of reified universals without actual reification. For the sake of space,
I will focus on two of these examples, one metaphysical and one methodological.26

26. For a bit on the concern with formal essences and a line of reply on behalf of Spinoza, see
Newlands 2015. These are not the only points of tension. In section 4.1, I pointed to similar
concerns about common natures or properties of bodies. Additionally, in part two of the Ethics,
Spinoza’s physics seems inconsistent with anything like nonconventional natural kinds, yet in
part four, Spinoza invokes what look like realist versions of natural kinds, forms, and natures
(see esp. E IVpref, E IVp29, E IVp36d, and E IVp37s1). Also in part two, Spinoza claims that
notions like being and thing “signify ideas that are confused in the highest degree” (E IIp40s1,
7 8   •  The Problem of Univer sals in Early Modern philosophy

4.4.1. Attributes
Spinoza’s theory of attributes has been subjected to centuries of interpretive
puzzlement and debate. I  will take up only a small subset of these issues here.
One concern is that Spinoza’s attributes function very much like universals. They
are fundamental ways of being that characterize multiple singular things. But if
Spinoza rejects realist theories of universals, it seems like he should also reject
realist theories of attributes. As Wolfson acutely puts it, “what is true of univer-
sals is also true of attributes” (Wolfson 1934, 1:153). Spinoza either needs to dis-
tinguish universals from attributes or accept the mind-​dependence of attributes
as well.
Wolfson famously accepts the latter option, and uses Spinoza’s theory of
universals to motivate his subjectivist reading of the attributes (Wolfson, 1934,
1:142–​56). Most recent interpreters reject the subjectivist reading of the attri-
butes, though without giving as convincing a defense of the difference between
universals and attributes in Spinoza as one might like.27 Steven Nadler, for exam-
ple, argues, pace the subjectivist interpretation, “Spinoza regards the attributes as
real and essential features of Nature. They represent objective kinds or categories
of things, and not merely phenomenal or subjective ways of regarding things”
(Nadler 2006, 130). That’s true, but the trickier question is whether Spinoza is
entitled to such mind-​independent, objective categories, given his conceptual-
ism about universals. In this section, I will suggest one line of defense that also
respects the mind-​dependent passages emphasized by Wolfson and other subjec-
tivist interpreters.
There can be little doubt that Spinoza links his theory of attributes to the intel-
lect. He defines an attribute as “what an intellect perceives of a substance as con-
stituting its essence” (E Id4, G 2:45.17–​19). He later claims, “outside an intellect
there is nothing except substances and their affections. Therefore there is noth-
ing outside an intellect through which a number of things can be distinguished
from one another except substances, or what is the same, their attributes, and their
affections” (E Ip4d, G 2:47.25–​2:48.2; emphasis mine). The connection is even
more explicit in a letter from 1663: “I understand the same by attribute [as I do by
substance], except that it is called attribute in relation to an intellect, which attri-
butes such and such a definite nature to substance” (Ep 9, G 4:46.22–​23).28 This

G 2:121.11–​12). Yet the metaphysics of part one is full of appeals to these transcendentals. For
other points of concern, see Gueroult 1968–​74, 1:417–​22 and, more unevenly, Haserot 1950.
27. The locus classicus of the nonsubjectivist reading is Gueroult 1968–​74, 1:428–​61.
28. In all these, I have rendered the article in front of ‘intellect’ as indefinite, whereas Curley
renders it as definite.
Spinoza on Universals  •  79

suggests at a minimum that intellects play a role in distinguishing a substance


from its attributes. Independent of the mind, a substance and its attributes are
not distinct, though they can be distinguished by an intellect. That is, to borrow
terminology from Suárez and Descartes, the distinction between a substance and
its attributes is a distinction of reason, and not a real distinction, for Spinoza.
The intellect also appears to play a role in distinguishing attributes from each
other. In E Ip10, Spinoza claims that each attribute is conceived through itself,
which means it is conceptually and explanatorily independent of every other
attribute. He claims in E Ip10s that the conceptual independence of attributes
does not entail that there is a real distinction between any attributes. But given
his substance monism, Spinoza must also accept the stronger claim that there
is no real distinction between any attributes. If attributes are not really distinct
from each other, yet are distinct enough to be conceived independently of one
another, then perhaps attributes are distinguished from each other only by a dis-
tinction of reason as well.
Spinoza explicitly accepts these conclusions in CM.29 He writes, “that distinc-
tion is said to be of reason which exists between substance and its attribute” (CM
II.5, G 1:258.1–​2).30 Several paragraphs later, he adds, “And from this we can
now clearly conclude that all distinctions we make between the attributes of God
are only distinctions of reason—​the attributes are not really distinct from one
another” (CM II.5, G 1:259.3–​5).31 There is nothing especially remarkable about
this view. Descartes makes the same claims in the Principles, and it was commonly
held that God’s attributes are distinct from each other and from God’s essence
only by a distinction of reason.32
However, within Spinoza’s system, this admission seems to make the plural-
ity of God’s attributes too mind-​dependent, especially if one also thinks that,
according to Spinoza, God lacks an intellect altogether (E Ip17s). Tying the
diversity of attributes to intellects appears to transform their distinctness and
multiplicity into a mental projection onto what is, in itself, a homogeneous
and indistinct divine nature. Indeed, this mental projection looks very much

29. Gueroult is surely right that we should be careful in projecting too much of what Spinoza
says in CM onto him (Gueroult 1968–​74, 1:446), but as I tried to show, these points have a
textual basis in the Ethics and correspondence as well. I should emphasize that I do not mean
to collapse what Spinoza says in the Ethics about conceptual dependence into Suárez’s or
Descartes’s “conceptual distinction”; these are somewhat orthogonal categories, a point Tad
Schmaltz helpfully pressed me on in discussion.
30. See also CM I.3, G 1:240.16–​18.
31. See also KV I.22, G 1:23.14–​16.
32. See Descartes, Principles I.62–​63, and Suárez, DM VI.i.4–​5.
8 0   •  The Problem of Univer sals in Early Modern philosophy

like what Spinoza decried as the psychological projection of unity across dis-
tinct singulars that occurs when we reify universals—​exactly as Wolfson had
charged. The mind-​dependent character of attributes, in turn, seems to imply
that substance, in itself, does not have any attributes at all, much less a multi-
plicity of attributes, a conclusion that would make Spinoza a kind of attribute
nihilist.33
What’s a nonsubjectivist interpreter to do? Here is one way to proceed while
still admitting that (1) distinctions between a substance and an attribute and dis-
tinctions among attributes of a single substance are merely distinctions of reason
for Spinoza, and (2) Spinoza’s appeal to an intellect in his discussions of attributes
is his way of capturing (1). Suárez and Descartes both recognize that distinctions
of reason come in two varieties: distinctio rationis ratiocinantis (usually translated
as “a distinction of reasoning reason”) and distinctio rationis ratiocinantae (“a dis-
tinction of reasoned reason”). A distinction of reasoning reason is a mental dis-
tinction that lacks any basis in mind-​independent things.34 As Suárez explains, “it
arises exclusively from the reflection and activity of the intellect” (DM VII.i.4).
This is the sort of purely projectionist/​constructivist distinction that nonsubjec-
tivist interpreters rightly reject as inconsistent with Spinoza’s other claims about
attributes.
By contrast, a distinction of reasoned reason is one that has some “basis” or
“foundation” in mind-​independent things. Suárez emphasizes that while this dis-
tinction is still mind-​dependent—​“actually and formally it is not found in real-
ity, but has its origins in the mind”—​it nonetheless has an important ground in
reality such that “it arises not entirely from the sheer operation of the intellect,
but from the occasion offered by the thing itself on which the mind is reflecting”
(DM VII.i.4). When a thing is the foundation for this kind of distinction of rea-
son, Suárez claims the thing must have a special “eminence” over the relata in two
ways: “Although the same object is apprehended in each concept [of the thing],
the whole reality contained in the object is not adequately represented, nor is its

33. In Newlands 2012, I explain the havoc this conclusion would wreak elsewhere in Spinoza’s
metaphysics. Note that nothing I have said here turns on whether the distinction is made by
a finite or an infinite intellect. There are good textual reasons to think that Spinoza had the
infinite intellect in mind in E Id4, but the problems with mind-​dependence remain even if a
finite mode is not the ground of attribute multiplicity.
34.  The stock example is the distinction between Peter and himself in the claim that Peter
is identical to himself (DM VII.i.4 and DM LIV.vi.5). Descartes mentions the two kinds of
distinction of reason in correspondence, only to put the category of reasoning reason aside: “I
do not recognize any distinction made by reason ratiocinantis—​that is, one which has no foun-
dation in reality—​because we cannot have any thought without a foundation” (CSMK 280,
AT 4:349).
Spinoza on Universals  •  81

entire essence and objective notion exhausted” (DM VII.i.5).35 First, the essence
of the thing itself is more real or perfect than the reality expressed by any one of
the relata. Second, the concept of each relata does not exhaustively capture every-
thing that is contained in the concept of the thing itself. And the most prominent
example Suárez has in mind is the distinctions among the divine attributes and
between the divine attributes and the divine essence itself.36
Suárez also emphasizes that distinctions of reasoned reason are not made only
between entia rationis. That is, a reason-​dependent distinction does not require
reason-​dependent relata. This helps alleviate the worry that attributes and sub-
stances would somehow become beings of reason in virtue of being only ratio-
nally distinct. “As is clear from the instances cited [including God’s attributes],
things said to be [rationally] distinct are real entities, or rather, a single real entity
conceived according to various aspects” (DM VI.i.6; emphasis mine). Suárez’s
last phrase nicely summarizes what I take attributes to be for Spinoza: ways of
conceiving the essence of substance, each of which is an extensionally adequate
expression of God’s essence (E Ip10) but no one of which expresses that essence
exhaustively (E Id6).37 On this analysis, Spinoza’s God has the kind of eminent
perfection required to ground distinctions of reasoned reason without admitting
of any distinctions among attributes independently of being distinguished by an
intellect.
Although Spinoza never distinguishes these two types of rational distinc-
tions, I think he surely intends the latter type when he claims that substance and
its attributes are distinct only in intellectu. This can help break the worrisome link
between universals and attributes. I claimed in section 4.1 that Spinoza thinks
universals are only very loosely grounded in mind-​independent things, in the
sense of being merely occasioned by their interaction with our bodies. However,
Spinoza can consistently claim that, unlike universals, distinctions among attri-
butes do have a stronger ontological foundation in mind-​independent things and

35.  Suárez describes these incomplete concepts as “inadequate,” but we should not under-
stand inadequacy to entail falsity in this case. It is inadequate in the sense of being incom-
plete. Spinoza accepts that although extension is a way of conceiving the essence of God that
is wholly self-​contained and extensionally adequate, it does not exhaustively represent the full
essence of God, a being with infinitely many attributes (E Id6, E Ip9).
36.  DM VII.i.5; DM LIV.vi.5; Descartes, CSMK 280. One interesting question is whether
Spinoza would agree with Suárez’s further claim that the different attributes “in an ineffably
eminent manner are found united in the absolutely simple virtue of God,” a point that turns
on Aquinas’s doctrine of analogy that, to my knowledge, Spinoza nowhere discusses. Certainly
Spinoza’s rationalism seems at odds with any sort of divine “ineffability,” and he is critical of
appeals to eminent containment in other contexts (e.g., E Ip15s).
37. For more on Spinoza’s attributes as ways of conceiving, see Newlands 2012.
8 2   •  The Problem of Univer sals in Early Modern philosophy

so do “indicate the nature” of something real, namely God’s essence. Admitting


that God’s attributes are distinguished only by a distinction of reason neither
undermines the mind-​independence of substance nor eliminates a real, mind-​
independent basis in substance for these differences. In other words, whereas
Spinoza rejects moderate realism for universals, he could accept something close
to it for attribute distinctions: formaliter in mente, fundamentaliter in re.
This is only a partial solution, however. For if Spinoza accepts that attributes
are robustly grounded in re, even if only distinguished in mente, one wonders
why other universals couldn’t be assigned a similar status. Put differently, why are
attributes like extension and thought grounded in the essence of substance and
only rationally distinct from substance and each other, while traditional univer-
sals like man and good are dismissed as merely occasioned products of a confused
imagination? Rather than eliminating the tension between “good” and “bad” uni-
versals, this account seems to highlight it all the more clearly!
Spinoza’s reply turns on his case for why thought and extension are attributes
in the first place, whereas man and good cannot be. Although it would take us too
far afield to pursue this much further here, the key is found in Spinoza’s claim that
“there belongs to God an attribute whose concept all singular thoughts involve,
and through which they are also conceived” (E IIp1, G 2:86.15–​16). This occurs
in the middle of Spinoza’s proof that thought is an attribute of God. Spinoza
claims that attributes are explanatorily prior to particular finite things, that
“through which” these singular things are conceived.38 This, I take it, is not the
case with universals for Spinoza. The universal man does not explain the nature
of particular men, much less everything else that exists. The explanation runs in
the opposite direction, in fact.39 Unsurprisingly, the demonstration of that relies
on yet more Spinozistic claims about the natures of thought, extension, and men.

4.4.2. Reasoning via Common Notions


Instead of philosophizing based on “abstractions and universals,” Spinoza claims
in his early treatise on method that we should “deduce all our ideas from physical
things, or from the real beings, proceeding, as far as possible, according to the

38. Descartes makes a similar claim about attributes being the “principle property” of a sub-
stance “to which all its other properties are referred” (Principles I.53, CSM 1:210, AT 8[1]‌:25).
39. Spinoza sometimes cites the disagreement over the concept of man as evidence for why
being a man is not an explanatorily fundamental way of being of a thing (e.g., CM I.1, G 1:235;
E IIp40s, G 2:121), though this line of argument is not very convincing. Furthermore, by the
same reasoning, Spinoza should admit that seventeenth-​century disagreements over the nature
of extension provide evidence that extension is not an attribute.
Spinoza on Universals  •  83

series of causes, from one real being to another real being” (TIE 99, G 2:14–​17).
In the Ethics, Spinoza devotes the bulk of E IIp40s to showing how transcenden-
tal universals, like being and thing, as well as less abstracted universals, like man
and dog, arise from confused bodily impressions.40 We expect Spinoza to repeat
his earlier methodological admonition. Down with reasoning based on abstrac-
tions and universals! On to the real things!
At the start of E IIp40s, just before he provides his discrediting account of
the origins of universals, Spinoza admits that “the foundations of our reasoning”
rest on “notions which are called common” (E IIp40s1, G 2:120.15–​16).41 A bit
later, Spinoza adds, “we have common notions and adequate ideas of the prop-
erties of things … this I  shall call reason and the second kind of knowledge”
(E IIp40s, G 2:122.12–​13). Reasoning based on the common affections of bodies
(as per E IIp13l2 and EIIp37c) sounds like reasoning based on abstractions from
the bodily impressions made by distinct singular things, the sort of universals-​
based reasoning Spinoza had previously rejected. Spinoza even acknowledges
that common notions are based on ideas that neither constitute nor explain the
essence of any singular thing.42 So we should avoid them at all costs, right?
Not according to part two of the Ethics. Far from challenging this method
of reasoning, Spinoza emphasizes how it is guaranteed to be adequate and true!
Although he admits in passing that there is a better form of knowledge in the
offing (intuitive knowledge), nothing he says in part two suggests that there is a
deep flaw in using this method to study the world.43
The puzzling feature here isn’t that Spinoza admits that we have common
notions, understood as ideas of exactly similar bodily properties (E IIp13l2).
Being a conceptualist about universals, Spinoza already accepts that we have uni-
versal notions or concepts. Nor is the concern that Spinoza seems to admit that
things have “common” or shared properties. I claimed in section 4.1 that the sense
of “common” here is consistent with his resemblance nominalism (though I also
noted some passages in tension with this.)
Rather, the puzzle surrounds Spinoza’s proposed methodology in the Ethics.
He seems to admit that progress can be made via the use of some abstractions
from bodily impressions, contra his earlier blanket warnings against inferring

40. As mentioned above, I do not know how to reconcile Spinoza’s claim that “being, thing,
etc.” are terms that “signify ideas that are confused in the highest degree” with his own prolific
use of such terms throughout the Ethics, including in the very scholia of E IIp40!
41. See also E IIp44c2d, G 2:126.
42. See E IIp37, G 2:118; E IIp44c2d, G 2:126.
43.  See also TTP VII.6, G 3:102, for another seemingly pro-​
universal statement of
methodology.
8 4   •  The Problem of Univer sals in Early Modern philosophy

anything from “abstractions and universals.” In fact, Spinoza opens his attack on
universals in EIIp40s by saying that he will examine “which notions are more
useful than others, and which are of hardly any use at all” (G 2:120.18–​19). Is
Spinoza now conceding that reasoning via some abstractions can be useful and
appropriate, after all?
In reply, one might note that Spinoza eventually points out the limits of
reasoning based on common notions. The reasoning that Spinoza describes as
providing a “second kind of knowledge” prepares the reader to make the leap to
intuitive knowledge in part five.44 This intuitive, third kind of knowledge moves
directly from God’s essence to the essences of singular things in just the way TIE
prescribed, proceeding from one real thing to another (E IIp40s2, G 2:122; TIE
99, G 2:36). Hence, Spinoza’s endorsement of reasoning via common notions in
part two might be understood as merely provisional, much like the provisional
morality in the early part of TIE or the provisional model of the “free man” in
part four.45 Perhaps Spinoza thought he needed to send a methodological ladder
down to his readers in the early parts of the Ethics, even if he ultimately kicks that
ladder away.
There is surely something correct in that reply, but I  remain puzzled by
Spinoza’s about-​face in E IIp40s itself. Once again, we are left wondering why
some universal notions are even provisionally better than others, even if intuitive
knowers avoid them all. Spinoza hints at one answer: the “bad” universal notions
are acquired only via the senses and imagination, whereas the “good” universal
notions are acquired via reason, even though all are occasioned by bodily affec-
tions.46 This difference in source helps explain differences in the representational
features of these different classes of universal notions: those drawn only from the
senses and imagination are confused and mutilated, whereas those drawn from
rational insight are adequate and guaranteed to be true.47
Spinoza’s interpreters often stop here, as if showing that Spinoza claims
that common notions are distinct from traditional universals because they are
acquired through different sources suffices for showing that Spinoza is entitled to
his division.48

44. See E Vp25 and following.


45. I am grateful to Michael LeBuffe for the expression “provisional morality.”
46. See esp. E IIp44c2d, G 2:126.
47. For the first class, see E IIp28–​29, G 2:113–​14; for the second, see E IIp38, G 1:118–​19.
48. For examples, see Gueroult 1968–​74, 2:581–​82 and Wolfson 1934, 2:124–​25.
Spinoza on Universals  •  85

But once again, this just pushes the bump in the carpet back a bit. Why
should we accept Spinoza’s claim, for instance, that abstracted ideas like “being”
can be acquired only via the imagination and bodily impressions? Why accept
his groupings of “good” and “bad” universals in the first place? More generally,
why accept Spinoza’s account of the different sources of universal notions and
the corresponding representational clarity or confusion he attaches to them? As
usual with Spinoza, whether one thinks he has satisfying answers will depend in
large part on whether one accepts a host of other Spinozistic claims whose scope
extends well beyond the topic of universals.49

Abbreviations
Frequently cited works have been identified by the following abbreviations,
which are grouped by author.

Aq u i na s

ST = Aquinas 1947; cited by part, question.

D es c a rt e s

AT = Descartes 1964–​74; cited by volume and page.


CSM = Descartes 1984–​85; cited by volume and page.
CSMK = Descartes 1991; cited by page.

Spinoza

C = Spinoza 1985.
CM = Cogitata Metaphysica; cited by part and chapter
E = Ethica; cited by the standard part, type, number:
app = appendix
c = corollary
d = definition (when not after a proposition number)
d = demonstration (when after a proposition number)

49.  I  would like to thank audience members at the 2011  “The Problems of Universals in
Modern Philosophy” conference in Pisa, Italy, for their helpful suggestions and questions. I am
also grateful to Michael Istvan for a series of written comments on an earlier draft. Thanks as
well to Colin Chamberlain, Richard Cross, Stefano Di Bella, Liz Goodnick, Marcy Lascano,
Tad Schmaltz, Eric Stencil, and Aaron Wells for helpful discussions of these topics.
8 6   •  The Problem of Univer sals in Early Modern philosophy

p = proposition
pref = preface
s = scholium
Ep = Epistolae; cited by letter number in G.
G = Spinoza 1925; cited by volume and page, or volume, page, and line
number.
KV = Korte Verhandeling; cited by part and chapter.
PP = Renati des Cartes Principiorum Philosophiae, cited by page number in G
TIE = Tractatus de Intellectus Emendatione; cited by paragraph number.
TTP = Tractatus Theologico-​Politicus; cited by chapter and paragraph
number.

Suá r e z

DM = Suárez 1965; cited by disputation, section, paragraph.


DM VI = Suárez 1964.
DM VII = Suárez 1947.
5 DESCARTES ON UNIVERSAL ESSENCES
AND DIVINE KNOWLEDGE

Lawrence Nolan

5.1. Introduction
It is sometimes said that early modern rationalism can be charac-
terized as an attempt to vindicate Plato’s philosophy over against
Aristotle’s thought, which informed much of scholasticism.1 As a gen-
eral statement, this assertion has much to recommend it, but, as one
might expect, the relation between seventeenth-​century rationalism
and Platonism is rather complex. Descartes’s philosophy provides a
striking example of this complexity. There are many Platonic doctrines
that Descartes accepts or at least adapts for his own purposes, such
as the theory of innate ideas and the doctrine of mind-​body dualism.
But his philosophy also contains a potent strain of anti-​Platonism in
its account of the ontological status of universal essences and eter-
nal truths. This strain runs deep in his thought and can be found as
far back as the Regulae, where he insists that number is not distinct
from the thing numbered and cautions readers against the tempta-
tion to reify mathematical objects—​in effect, to fall prey to Platonic
realism—​a temptation to which he thinks every mathematician suc-
cumbs (Rule 14, AT 10:445–​4 6; CSM 1:61).2 Later, in a series of
articles in the Principles of Philosophy devoted to laying out his meta-
physics in a systematic manner, Descartes says that universals such as
a triangle or the number two are merely ideas or ways of thinking and,
further, that eternal truths have a “seat within the human mind” (AT

1. See, e.g., Jolley 1990, introduction.


2. Any divergences from the translations given in CSM are my own.
8 8   •  The Problem of Univer sals in Early Modern philosophy

8–​1:27–​28, 23–​24; CSM 1:210, 209). Far from endorsing Platonism, Descartes
explicitly affirms a conceptualist account of abstract entities.
Despite these and other such passages, some English-​speaking commentators
going back to Anthony Kenny (1970) and Alan Gewirth (1970, 1971), have been
mesmerized by the Fifth Meditation, where, as a preface to the ontological argu-
ment, Descartes asserts that objects such as a triangle, even if they do not exist
anywhere outside our thought, have “immutable and eternal” natures, essences,
or forms that do not depend on the human mind (AT 7:64, CSM 2:45). On
the basis of these remarks, Kenny (1970, 685–​700) reads Descartes as positing
a third realm of abstract entities—​distinct from created substances and distinct
from God—​and declares him to be “the father of modern Platonism.”3 In a previ-
ous essay, I argued that contrary to Kenny the apparent Platonism of the Fifth
Meditation is just that—​apparent—​and that Descartes’s claims about true and
immutable natures are easily reconciled with the explicit conceptualism of the
Principles (Nolan 1997a; also see Nolan 1998). I was not alone. Focusing largely
on the Principles, Vere Chappell (1997) drew the same conclusion, namely, that
Descartes reduces mathematical essences, and the eternal truths concerning
them, to the objective being of innate intellectual ideas in the human mind.
But the Platonist reading of Cartesian universals dies hard. Since Kenny wrote
on this topic, a couple of notable commentators have proposed that Descartes
locates essences, and the eternal truths concerning them, in God. Let us refer
to this general interpretation of Cartesian essences as “Theological Platonism”
(to borrow a term from Émile Bréhier). Most recently, Marleen Rozemond
(2008) has argued that it is open to Descartes to embrace Duns Scotus’s view that
essences have objective being in the divine intellect.4 Rozemond’s interpretation
harks back to an essay published by Tad Schmaltz (1991), who argues that eternal
truths and essences are divine decrees.5 Both essays are inspired by Descartes’s
claim that essences are “immutable and eternal”; Rozemond and Schmaltz main-
tain that Cartesian essences can possess these attributes only if they exist in God.

3.  Cf. Kenny 1968, where he compares true and immutable natures, or their objects, to
Meinongian pure objects, which are subjects of true predication whether or not they exist.
4.  In another recent paper, Raffaella De Rosa (2011) expresses qualified support for differ-
ent aspects of both Theological Platonism and the Conceptualist Interpretation but does not
endorse either position on the grounds that neither fully accounts for all of the strands in
Descartes’s treatment of created essences. De Rosa’s goals, as she describes them, are mainly
negative:  she claims that neither of the two main interpretations satisfies all of Descartes’s
commitments, but that is because she sees him as holding inconsistent views. I maintain that
Descartes’s view is consistent.
5. Schmaltz has told me in conversation that he is no longer committed to the Platonist reading
of Descartes. Also see Schmaltz 2002, 83, fn. 21 and chapter 6 in this volume.
Descartes on Universal Essences and Divine Knowledge  •  89

Besides lacking direct textual support, the main problem with Theological
Platonism is that it violates the Christian doctrine of divine simplicity, at least
as it was understood by Descartes. While many medieval philosophers held
that creaturely essences reside in the divine understanding as ideas, Descartes’s
strict conception of divine simplicity—​which entails that in God understand-
ing and willing are the same—​forces him to reject the traditional account of
essences as divine ideas. This brings me to the topic of this chapter. I intend to
develop a new defense of the conceptualist account of Cartesian universals that
places special emphasis on the simplicity of Cartesian substances, both created
and divine. I argue that Descartes’s account of universals, while conceptualist in
character, is motivated by one of Plato’s most seminal insights. Plato held that
the ultimate constituents of reality—​the most fundamental entities in the meta-
physical universe—​are perfectly simple. Descartes’s account of universals is best
read in the context of this Platonic intuition.6 As a consequence of the simplic-
ity of Cartesian substances, universal essences cannot exist as discrete entities
in God nor as distinct constituents of created minds and bodies. They can exist
only as ideas in human minds. There is a notable irony here: Descartes is led by a
Platonic premise to reject all forms of realism about universals, Platonic or other-
wise. Within a Christian philosophical context, one form of Platonism supplants
another. The relation that Cartesianism bears to Platonism is complex indeed.
In the next section, I  begin by arguing for the simplicity of all Cartesian
substances, both created and divine, and then use this doctrine to uncover the
philosophical motivation for Descartes’s conceptualism regarding universals.
This discussion further develops the conceptualist reading and explains why even
moderate forms of realism, which locate universals in created things, are barred
to Descartes. I also argue that the doctrine of simplicity precludes Descartes from
locating creaturely essences and eternal truths in God, contrary to Theological
Platonism. In section 5.3, I consider Rozemond’s recent attempt to accommo-
date divine simplicity within Theological Platonism and show why it fails. This
discussion raises an important question about the nature of divine knowledge. As
an omniscient being, God knows created things, but how does he know them?
A standard medieval account of divine knowledge held that God knows creatures
by cognizing their ideational archetypes in his understanding. But if, as I claim,
Descartes rejects the traditional theory of divine ideas, then God must know
creatures in some other way. In section 5.4, I argue that Descartes is committed
to a view like that of Ockham, who holds that the objects of divine cognition
are created things themselves (“Divine Direct Realism”). In section 5.5, I answer

6. The simplicity of being is of course a theme running throughout ancient Greek philosophy.
9 0   •  The Problem of Univer sals in Early Modern philosophy

objections to the Conceptualist Interpretation, including the objection that it


cannot account for the sense in which Cartesian essences and truths are immu-
table and eternal.

5.2. The Simplicity of Cartesian Substances


I begin by characterizing what I take to be Descartes’s most fundamental meta-
physical intuitions. The simplicity of substances is one of these intuitions, but
not his only or most basic one. Once we have an appreciation of his most funda-
mental intuitions, we can gain a better purchase on where universals and eternal
truths fit within his metaphysics and will also be able to understand what moti-
vates his conceptualism.

5.2.1. Descartes’s First Intuitions


A common refrain of many early modern philosophers is that only particulars
exist. This principle is of course an affirmation of antirealism about universals.
Descartes subscribes to this principle himself, but it is not one of his first intu-
itions. Instead, he takes as his starting place what particular things or substances
exist or, more exactly, could exist. His clear and distinct ideas reveal the limits of
what is conceptually possible with respect to the particular inhabitants of the
metaphysical universe. Here we can take our bearings from the Meditations. If
one reflects on the clear and distinct ideas that he enumerates over the course
of this work, it is notable that they are mostly ideas of particular substances.
First and most importantly, there is the idea of God or the infinitely perfect sub-
stance. This is not the idea of being in general, as it is for Aquinas or even for
Malebranche, Descartes’s successor. The idea of God that one discovers in the
Third Meditation—​what Descartes calls the “mark of the craftsman stamped on
his work”—​is the idea of something that is actually infinite, but nevertheless par-
ticular. Second, there is the idea of the mind as a thinking thing. This is not the
idea of thought in general but the idea of one’s own mind or self, as Descartes
makes clear in the Third Meditation (AT 7:51, CSM 2:35). We can of course
form the idea of thought in general or universal thought, but this is not one of
our most fundamental ideas. And finally, there is the idea of body. On this topic,
there is much controversy. Does Descartes think there are many corporeal sub-
stances or only one, namely, the whole material plenum? Fortunately, we do not
need to resolve this debate. It suffices for our purposes to note that whether there
is one such body or many, the substance or substances in question are particular,
and hence the idea (or ideas) of corporeal substance(s) that one entertains in the
Descartes on Universal Essences and Divine Knowledge  •  91

context of the Meditations is (are) also particular.7 So, to summarize, Descartes’s


first intuitions are about what things there are or could be, and these all turn out
to be particular substances.

5.2.2. Divine Simplicity
With this list of particular substances in hand, Descartes’s strategy is to reduce
most other putative entities to one of them or, in the case of finite substances,
to one of their modes. This reductionist strategy is motivated by a second intu-
ition of his metaphysics, namely that all substances, whether created or divine, are
simple. Most Christian philosophers endorse divine simplicity in some form or
other, but Descartes understands this doctrine strictly and draws out its implica-
tions for the nature of divine creation in a way than no philosopher had done
before. As we shall see, he also extends the doctrine of simplicity to finite, created
substances, at least with respect to their attributes. I begin, however, with the case
of divine simplicity, since God constitutes the paradigm of a simple substance.
Descartes’s statements of this doctrine are few in number in his published
writings but explicit and unequivocal. Here are two such passages, from the Third
Meditation and the Principles of Philosophy, respectively.

[1]‌The unity, the simplicity, or the inseparability of all the attributes of


God is one of the most important of the perfections which I understand
him to have. (AT 7:50, CSM 2:34)
[2]‌There is always a single identical and perfectly simple act by means
of which he simultaneously understands, wills and accomplishes every-
thing. (AT 8–​1:14, CSM 1:201)

In the second passage, we begin to get a sense of just how strictly Descartes
understands divine simplicity:  God does everything by one perfectly simple
act, without any priority between understanding and willing. Here we might
wonder why, of all of God’s faculties or attributes, Descartes stresses intel-
lect and will in his effort to illustrate divine simplicity. The answer is not dif-
ficult to discern, given Descartes’s philosophical inheritance. In the Middle
Ages, it became standard among Christian philosophers to explain creation

7. Jonathan Bennett (2001, vol. 1, 136) asserts that if Descartes is a monist regarding corporeal
substance, then matter is a mass noun, not a count term. It would thus be a mistake on this view
to say that corporeal substance is particular. I think I can accept this proposal without doing
violence to my general point, which is that Descartes’s first intuitions are not about kinds.
9 2   •  The Problem of Univer sals in Early Modern philosophy

in quasi-​Platonic terms. Inspired by Plato’s Timaeus, in which the Demiurge


or divine craftsman creates the universe according to rational, purposive prin-
ciples, many medieval philosophers going back to Augustine maintained that
God creates the universe via ideas of possible creatures, which exist in his intel-
lect prior to his creative activity. This is the doctrine of exemplary causation,
according to which God creates the universe via ideas or archetypes in his intel-
lect. The point of this doctrine is to make creation intelligible and rational.
Philosophers who endorse this view of creation are sometimes called “divine
intellectualists” because of their emphasis on the priority of God’s intellect over
his will. Descartes is clearly rejecting divine intellectualism when he says that
God simultaneously wills and understands everything in one perfectly simple
act. Indeed, it is often noted that Descartes is a divine voluntarist in that he
makes everything depend on God’s will. One reason that Descartes stresses
divine simplicity in this context is because he thinks that the Platonic account
of creation anthropomorphizes, and thus debases, God. Just before the sen-
tence quoted in passage [2]‌, he writes: “And even his [God’s] understanding
and willing does not happen, as in our case, by means of operations that are in a
certain sense distinct one from another” (AT 8–​1:14, CSM 1:201). In creating
the world, God is not like a human architect or craftsman, who must look to an
idea or blueprint in his intellect prior to creation.
When the doctrine of divine simplicity appears in Descartes’s correspon-
dence, it is typically linked more explicitly to his creation doctrine, which states
that everything finite is created, including essences and eternal truths. The fol-
lowing passages from two early letters to Mersenne are representative.

[3]‌As for the eternal truths, I say once more that <they are true or possi-
ble only because God knows them as true or possible. They are not known
as true by God in any way which would imply that they are true indepen-
dently of him>. If men really understood the sense of their words they
could never say without blasphemy that the truth of anything is prior to
the knowledge which God has of it. In God willing and knowing are a sin-
gle thing in such a way that <by the very fact of willing something he knows
it and it is only for this reason that such a thing is true> (To Mersenne, May
6, 1630, AT 1:149; CSMK 24; emphasis added; angle brackets indicate
Descartes’s use of Latin in a French context).
[4]‌You ask me <by what kind of causality God established the eternal
truths >. I reply: <by the same kind of causality> that he created all things,
that is to say, as their <efficient and total cause>. For it is certain that he is
the author of the essence of created things no less than of their existence; and
this essence is nothing other than the eternal truths. …
Descartes on Universal Essences and Divine Knowledge  •  93

You ask also what necessitated God to create these truths; and I reply
that he was free to make it not true that all the radii of the circle are
equal—​just as free as he was not to create the world. And it is certain that
these truths are no more necessarily attached to his essence than are other
created things. You ask what God did in order to produce them. I reply
that <from all eternity he willed and understood them to be, and by that
very fact he created them>. Or, if you reserve the word <created> for the
existence of things, then he <established them and made them>. In God,
willing, understanding and creating are all the same thing without one being
prior to the other <even conceptually> (To Mersenne, May 27, 1630, AT
1:151–​53; CSMK 25–​26; emphasis added).

Many readers have found Descartes’s doctrine of the creation of the eternal truths
to be bizarre and even incoherent. What would possess him to hold this doctrine,
which is unprecedented in the history of philosophy and which seems to have disas-
trous consequences for the possibility of knowledge and for the status of necessary
truths? Some recent scholarship has attempted to answer this question by point-
ing to the doctrine of divine simplicity.8 These passages provide strong evidence that
Descartes did indeed see the creation doctrine as a direct consequence of divine sim-
plicity. Note that in both of them he follows his assertion of the creation doctrine
with an affirmation of divine simplicity, understood in the strictest sense. The reason
everything is created, including essences and the eternal truths concerning them, is
because in God willing, understanding, and creating are the same, which is just to
say that whatever he understands he wills (and vice versa), without any priority or
distinction between them in re. My aim in section 5.4 will be to show that the doc-
trine of divine simplicity also has important consequences for the nature of God’s
knowledge of creation.
Before moving on, I  want to draw attention to one other point in passage
[4]‌that is crucial for understanding Descartes’s account of essences. In the first
few lines he identifies the eternal truths with the essences of created things (“this
essence is nothing other than the eternal truths”).9 Descartes’s statement is impor-
tant because it means that any claims that he makes about the status of eternal

8. See, most notably, Nelson 1993, 686f., and Walski 2001 and 2003.
9. What explains this statement is that eternal truths are generally truths about essences, but
not just any truths. In at least some cases they are definitional. The scholastics spoke here of
essential predications, such as man is a rational animal or, to pick a more Cartesian example, a
triangle is a three-​sided polygon. This is why Descartes sometimes treats essences and the eter-
nal truths concerning them as merely rationally distinct. See Nolan 2015 for further discussion
of this latter point.
9 4   •  The Problem of Univer sals in Early Modern philosophy

truths are also claims about the status of essences and vice versa. We should keep
this in mind as we move forward.10

5.2.3. The Simplicity of Created Substances


Now that we have a sense of Descartes’s conception of divine simplicity, let us
turn to the simplicity of created substances. Created substances are of course not
perfectly simple in the way that God is. They admit of modal variation both syn-
chronically and diachronically. My body, for example, has a certain shape, size,
and motion (or rest) at any given time, and these properties can vary with diet,
age, exercise, and so on. My mind too thinks of different things over time and, by
Descartes’s lights, could be entertaining two thoughts simultaneously.11 By con-
trast, God—​being immutable—​admits of neither type of modal variation (AT 8–​
1:26, CSM 1:211). But for the present purposes, the fact that created substances
have diverse modes is irrelevant. For my aim in this section is to show that these
substances do not have common constituents. There are no universals in created
substances because they do not admit of composition. Modes are not constitu-
ents of finite substances but ways of being those substances. And, as a result of the
intimate relation between a substance and its modes, it is impossible for modes to
be shared by, or common to, multiple substances. Descartes confirms this in his
definition of a modal distinction in the Principles, where he says that the modes
of any two substances are really distinct (AT 8–​1:30, CSM 1:214).
Descartes’s austere substance-​mode ontology marks a stark contrast to the
ontologies of the scholastic Aristotelians. According to the latter, a created sub-
stance is the subject of various forms of composition—​for example, matter and
form, substantial form and accidental forms, act and potency, essence and exis-
tence, and so on. By banishing all such composition, and conceiving substances as
simple, Descartes leaves no room for universals within the created universe. The
only things that exist are particular substances and their attributes and modes. As
was just noted, modes are simply ways of being those particular substances and,
as we shall see in what follows, the attributes of a substance are merely rationally
distinct from it. What motivates this move away from scholastic composition?
Part of the answer is Descartes’s new conception of matter in terms of geometri-
cal extension alone, shorn of the faculties, qualities, forms, and potentialities of
his predecessors. But a more basic answer, which precludes universals from being

10. Incidentally, the defenders of Theological Platonism and I agree on this point. See, e.g.,
Rozemond 2008, 41.
11. See Conversation with Burman, AT 5:149; CSMK 3:335.
Descartes on Universal Essences and Divine Knowledge  •  95

constituents of either bodies or minds, is the Platonic intuition that the most
fundamental entities—​namely, particular substances—​are simple.
In a series of articles devoted to setting forth his metaphysics in the Principles
of Philosophy, Descartes stresses this simplicity in a couple of different ways. The
strongest statement of it appears in his discussion of the theory of distinctions,
particularly in his definition of the notion of a rational distinction in part  1,
article 62. There, he says that a substance and each of its attributes, or any two
attributes of a single substance, are merely rationally or (as translated by CSM)
conceptually distinct (distincta ratione). The force of this claim is not fully clear
until one, first, understands what he means by it and, second, considers his exam-
ples of attributes. To address the first point, I have argued elsewhere that what it
means to say that two things are merely rationally distinct is that they are numeri-
cally identical in reality and distinguished only within our thought; hence the
term “distinction of reason” (distinctio rationis).12 The scholastics spoke here of a
“real identity” between the terms of a rational distinction. We have already seen
Descartes affirm this identity thesis with respect to the divine attributes, specifi-
cally divine will and intellect. Immediately after defining the notion of a rational
distinction in article 62 of the Principles, he also affirms this identity thesis in the
case of a created substance and its essence or principal attribute:

[5]‌Thought and extension can be regarded as constituting the natures


of intelligent substance and corporeal substance; they must then be
considered as nothing else but thinking substance itself and extended sub-
stance itself—​that is, as mind and body. (AT 8–​1:30–​1, CSM 1:215;
emphasis added)

So corporeal substance is not a substratum that underlies its extension. It just is


its extension. Likewise, my mind just is its thinking. We will encounter further
affirmations of the identity thesis in a moment, but I  would like to turn now
to the other issue I posed regarding a rational distinction, namely, the extension
of the term “attribute” for Descartes. What counts as an attribute in Descartes’s
technical sense of something that is merely rationally distinct from a substance?
As it turns out, thought and extension are not the only attributes of minds and
bodies, respectively. In other articles in the Principles, part 1, he lists four generic
attributes—​viz., existence, duration, order, and number or unity—​which every
substance possesses. This means that the essence and existence of any substance,
whether created or divine, are merely rationally distinct, a thesis that Descartes

12. See Nolan 1997b and 1998.


9 6   •  The Problem of Univer sals in Early Modern philosophy

explicitly confirms in a letter written to an unknown correspondent shortly after


publishing the Principles.

[6]‌I do not remember where I spoke of the distinction between essence


and existence. However, I make a distinction <between modes, strictly so
called, and attributes, without which the things whose attributes they are
cannot be; or between the modes of things themselves and the modes of
thinking [modos cogitandi]> … <Thus shape and motion are modes, in
the strict sense, of corporeal substance; because the same body can exist
at one time with one shape and at another with another, now in motion
and now at rest; whereas, conversely, neither this shape nor this motion
can exist without this body. Thus love, hatred, affirmation, doubt, and so
on are true modes in the mind. But existence, duration, size, number and
all universals are not, it seems to me, modes in the strict sense; nor in this
sense are justice and mercy, and so on modes in God. They are referred
to by a broader term and called attributes, or modes of thinking. …
Accordingly I say that shape and other similar modes are strictly speaking
modally distinct from the substance whose modes they are; but there is
a lesser distinction between the other attributes. [. . .] I call it a rational
distinction … > (1645 or 1646, AT 4:348–​49; CSMK 279–80)

In this passage, Descartes draws a distinction between two senses of the term
“mode.” On the one hand, there are modes in the strict sense, such as shape and
motion in the case of body, or an act of doubt in the case of the mind. On the
other hand, there are modes of thinking, by which he means not modes of mind
in the strict sense of acts of thought but ways of thinking about a substance. He
indicates that the term “modes,” in this second sense, means the same as “attri-
bute.” With this distinction in hand we can now appreciate what else he is say-
ing in the passage. First, notice that he reaffirms that the attributes of God are
merely rationally distinct and offers examples of divine attributes other than
will and intellect. Second, he notes that with respect to finite substance, dura-
tion, size, number, and all “universals” are merely rationally distinct from their
respective substances. So, again, the rational distinction applies not just to a sub-
stance’s essence or existence but to all of its attributes. The term “universals” is
a bit slippery here, as Descartes draws a distinction between the attributes of a
substance—​which are merely rationally distinct from it and from each other—​
and universals properly speaking, which he says in the Principles are merely ideas
in finite minds. In other words, he wants to reduce the attributes of a substance
to the substance itself, outside our thought, whereas he wants to reduce universals
(or what one might think of as general attributes considered in abstraction from
Descartes on Universal Essences and Divine Knowledge  •  97

all substances outside the mind) to mere ideas.13 Third, in the continuation of
this letter Descartes reaffirms the identity thesis when he notes that outside our
thought the essence and existence of a triangle are “in no way distinct” (AT 4:350,
CSMK 280).
Although finite minds and bodies are simple with respect to their attributes,
it is important to note that they are not absolutely simple in the way that God
is. We observed this point already at the beginning of this subsection:  unlike
God, created substances are capable of modal variation. It turns out that this sort
of complexity is innocuous vis-​à-​vis the status of universals, but there are other
forms of diversity as well. For example, bodies are capable of being divided into
spatial parts (AT 7:85, CSM 2:59). In fact, when articulating his physics in the
Principles, he claims not just that bodies are divisible but that at least some of
them are actually divided ad indefinitum.14 Although created minds do not admit
of spatial parts, one might think that they manifest another form of complex-
ity. Descartes sometimes speaks of the mind’s various “faculties”—​imagination,
memory, sensation, and so on. In the Principles, he insists that these reduce to
two—​intellect and will—​and that all other so-​called faculties are simply differ-
ent forms of them:  “Sensory perception, imagination and pure understanding
are simply various modes of perception; desire, aversion, assertion, denial and
doubt are various modes of willing” (AT 8–​1:17, CSM 1:204). Nevertheless, one
might argue that if there are two faculties, the mind is still divisible—​not spatially
divisible, to be sure, but divisible in some sense. But even if this were correct,
it would not matter for our purposes. Despite the simplicity of their attributes,
bodies may be complex in one sense and created minds in another, but neither
case involves the sort of complexity that would provide fertile ground for a mod-
erate form of realism about universals, whereby universals exist in particular cre-
ated things. It would be absurd to suppose that Descartes regarded the parts of
corporeal substances as their metaphysical constituents on the order of genera,
species, and other categories of universal recognized by the scholastics. Likewise,
created minds might have distinct faculties of intellect and will, but, whatever
such faculties amount to, they are not metaphysical constituents that are shared
by multiple minds.
To summarize the results of this section, Descartes maintains like other con-
ceptualists that only particular substances exist. His first philosophical intuitions,
however, are not about that claim but about which particulars there are—​viz.,

13. See Nolan 1998 for further discussion of this point.


14.  AT 8–​1:59–​60, CSM 1:239. Descartes argues that such division is necessary to explain
vortical motion in a plenum.
9 8   •  The Problem of Univer sals in Early Modern philosophy

God, the self or one’s own mind, and body. His second main intuition is about
the nature of these particulars. Following the Platonic or Neoplatonic tradition,
he maintains that substances are simple. This simplicity is reflected in his general
claim that a substance and each of its attributes are merely rationally distinct,
which entails that a substance is numerically identical with each of its attributes,
outside of our thought. As a result of this simplicity, there is no room within
substances for universals of any kind, neither in God (as ideas or archetypes for
creation) nor in created things. Given his central intuition about simplicity,
Descartes is compelled to treat universals merely as ideas in finite minds.

5.3. Divine Simplicity Revisited


Contrary to the argument of the previous section, some defenders of Theological
Platonism have attempted to reconcile Descartes’s strict understanding of divine
simplicity with the thesis that creaturely essences reside in God as ideas. My pri-
mary aim in this section will be to show why such efforts fail. In the process,
we will deepen our understanding of the role that divine simplicity plays in
Descartes’s philosophy.
There are actually two ways that ideas in God would violate his simplicity,
understood in Descartes’s strict sense, but so far we have discussed only one of
them. First, positing ideas in God presupposes an ontological distinction (and
priority) between divine intellect and will. Second, the theory of divine ideas
presupposes a multiplicity of ideas or essences in what is supposed to be a per-
fect unity. Either one of these violations would provide Descartes with a suf-
ficient reason for rejecting the traditional theory of divine ideas. As we have
seen, Descartes’s formulation of the doctrine of divine simplicity, especially as it
appears in his correspondence, is intended to combat the first of these violations.
He identifies God’s intellect and will and rejects any priority between them, as
would be required if there were ideas in his understanding logically prior to his
creative activity. Interestingly enough, however, it is the second of these violations
that so enthralled the scholastics, many of whom tried to reconcile the theory
of divine ideas with the doctrine of divine simplicity. In what follows, I briefly
explore this scholastic attempt at reconciliation in order to explain why Descartes
cannot accept it. This discussion provides the necessary background for discuss-
ing Rozemond’s recent defense of Theological Platonism on this issue.
In her magisterial book William Ockham, Marilyn Adams provides a use-
ful discussion of what she calls the “Simplicity Problem” in the Middle Ages.
Adams notes that early medieval philosophers who endorse the doctrine of
exemplary causation, such as Augustine and Anselm, did not confront the
problem. They posited discrete ideas or essences in God while also maintaining
Descartes on Universal Essences and Divine Knowledge  •  99

that God is perfectly simple, without acknowledging the inconsistency of their


position. But most thirteenth-​and fourteenth-​century scholastics appreciated
the force of the problem and attempted to solve it. The standard proposal was
that creaturely essences enjoy a nonreal mode of existence in the divine under-
standing. Because their existence is nonreal, they do not violate the simplicity
of God’s essence, which is fully real (the ens realissimum.). Adams writes: “The
really existent divine essence is one simple thing that involves no real internal
complexity. … But the activity of the divine intellect produces a possibly infi-
nite plurality of things in some non-​real mode of existence. It is this non-​real
plurality that provides the rationale for speaking of a plurality of ideas” (Adams
1987, 2:1037).
Without evaluating the merits of this proposed solution to the Simplicity
Problem, it suffices to observe that Descartes cannot accept it. Since divine intel-
lect and will are the same, ideas in God cannot possibly enjoy a nonreal mode
of existence. If God understands them, then he wills or creates them, and, if he
creates them, then they have being. In the Conversation with Burman, Descartes
reportedly says that God cannot “incline to nothingness”; he always tends toward
being (AT 5:147, CSMK 334). Among other things, this means that God cannot
create something that lacks being.
This might seem like the end of the debate, but in her recent defense of
Theological Platonism, Rozemond has attempted to trace Descartes’s thought
back to Duns Scotus, whose views often deviated from those of other scholastics.
Instead of affirming that creaturely essences enjoy a nonreal mode of existence
prior to creation, Scotus urges that they possess objective being in God’s under-
standing. Rozemond suggests that this position better befits Descartes’s view that
creaturely essences are created (Rozemond 2008, 48). As such, these essences
must possess some sort of being, even if only attenuated (what the scholastics
called esse diminutum). In the Third Meditation, Descartes says that objective
being is a diminished form of being in comparison with what he calls “formal
being” or actual existence outside the mind; nevertheless, it is “not nothing”
(AT 7:41, CSM 2:29). Although Descartes’s remarks about objective being are
confined to human cognition, Rozemond thinks that we can extrapolate from
what he says there to account for divine thought. She also notes that according
to Scotus the presence of a plurality of objective beings in the divine understand-
ing is consistent with God’s simplicity because the beings in question are merely
“formally distinct” in his technical sense of that expression (2008, 54–​55).
There are at least two problems with this attempt to accommodate divine sim-
plicity within Theological Platonism. First, it lacks direct textual support and is
based exclusively on extrapolations from what Descartes says about the role of
objective being in human cognition. It is not clear how Descartes’s statements
1 0 0   •  The Problem of Univer sals in Early Modern philosophy

concerning the latter, which incidentally are notoriously difficult to interpret,


commit him to anything with respect to divine cognition.
Second, and more importantly, Rozemond’s proposal does not solve the
Simplicity Problem, at least not for Descartes who, again, has a stricter under-
standing of divine simplicity than perhaps any other Christian philosopher.
Objective beings would still have some form of being in God and hence would
violate his simplicity. Indeed, many of Scotus’s critics have made this very point,
most notably Ockham. To understand the problem, we must make a brief excur-
sus into Scotus’s metaphysics. Commentators have found Scotus’s notion of a for-
mal distinction to be extremely difficult to understand, but one thing is clear: it
is something less than a real distinction, which obtains between two things (res),
and something more than a distinction of reason, which obtain solely in the
mind. Since the formal distinction obtains prior to thought, commentators note
that it can be called “real” by contrast with the rational distinction. But herein
lies the problem for divine simplicity. As Peter King writes in an essay surveying
Scotus’s metaphysics:

Given that the formal distinction is real in the broad sense, must there not
then be some degree of complexity in its subject? The formal distinction
holds in reality prior to the operation of the intellect. Even if there are not
distinct thinglike property bearers in a subject, then, it nevertheless seems
as though no thing to which a formal distinction applies can be simple.
This would rule out any formal distinction in God (2003, 24).

Scotus thinks that the formal distinction is consistent with divine simplicity, but
that is only because he holds an idiosyncratic view of real composition. Scholastic
philosophers oppose simplicity to real composition: something is simple if and
only if it is not a composite (Adams 1987, 2:903). On Scotus’s view, real composi-
tion requires that one entity be in potency with respect to another and that the
activity of the latter serves to perfect it (see Adams 1987, 2:932–​33, and King
2003, 37). King offers the example of a genus (e.g., animal) that is in potency to
a differentia (e.g., rationality). Since there is no potency in God, who according
to the scholastics is “pure act,” the presence of a multiplicity of diverse objec-
tive beings in his understanding poses no threat to his simplicity. For Scotus, this
solution to the Simplicity Problem is intended to be general, for it also purport-
edly explains how there can be three persons in one God as according to the doc-
trine of the Trinity (King 2003, 24).
The relevant point for our purposes is that Scotus can account for divine
simplicity only by appealing to an idiosyncratic account of real composition
that Descartes would not accept. Descartes rejects the Aristotelian act-​potency
Descartes on Universal Essences and Divine Knowledge  •  101

distinction. Moreover, it is clear that Scotus has a weak notion of divine sim-
plicity: something is simple if and only if it is not composite in this restricted
sense. He is willing to bend the notion of simplicity in a way that Descartes is
not because his priorities and philosophical motivations are different. One of his
main priorities is to explain God’s knowledge of creatures and he thinks this can
be done only on the Platonic model by appealing to objective beings in the divine
understanding. Descartes would likely say that he is not unconcerned to explain
divine knowledge, but that one should do so without anthropomorphizing God.
Scotus’s appeal to objective beings in God does just that. As we saw in section 5.2,
Descartes lays stress on the doctrine of divine simplicity, particularly as it applies
to the relation between divine intellect and will, largely because he wants to avoid
treating God on the model of a human artisan who must look to a blueprint prior
to creating. In his view, the latter conception of God diminishes his perfection,
which he sees himself as trying to vindicate. This is one of the key points that
defenders of Theological Platonism have failed to appreciate.
At one point in her essay, Rozemond tries to solve the Simplicity Problem by
saying that for Descartes, unlike Scotus, the objective beings in God’s intellect
are merely rationally distinct from his essence (2008, 54–​55). But as I  noted
above (and have argued at length elsewhere15), to say that x and y are merely
rationally distinct is to say that they are identical in reality, outside our thought.
It is only we, with our finite minds, who draw such a distinction, which obtains
solely in our thought. Rozemond concurs with this interpretation, at least on
the main point: “And the distinction of reason, for Descartes, is a distinction
that does not correspond to a distinction within God” (2008, 55). The problem
is that by saying there are objective beings in the divine understanding and they
are merely rationally distinct, Rozemond is trying to have it both ways: there
are discrete creaturely essences in the divine understanding and they are not dis-
crete. Scotus recognized that he could not consistently affirm both claims, and
so adopts the formal distinction. But there is no textual basis for attributing this
type of distinction to Descartes.16 On the contrary, in the Conversation with
Burman, Descartes reportedly tells his interviewer that “Whatever is in God is
not in reality separate from God himself; rather it is identical with God himself

15. See Nolan 1997b and 1998.


16.  In the letter cited in passage [6]‌, Descartes says that the rational distinction between
essence and existence can be called a formal distinction. However, as the context makes clear,
this is only a verbal point. He goes on to say that the rational distinction can also be called a
real distinction, which shows that in both cases Descartes is attempting to accommodate the
two main rival accounts of the relation between essence and existence within his own as a way
of neutralizing them (AT 4:349–​50, CSMK 3:280–​81). Cf. the First Replies, where Descartes
collapses the modal and formal distinctions (AT 7:120–​21, CSM 2:86).
1 0 2   •  The Problem of Univer sals in Early Modern philosophy

(imo est ipse Deus)” (AT 5:166, JC 32). This statement serves to confirm the
Conceptualist Interpretation and shows why Theological Platonism cannot find
a foothold within his system.
In affirming that for Descartes everything in God is merely rationally dis-
tinct, including divine intellect and will, Rozemond is forced to acknowledge yet
another important difference between him and Scotus:

Descartes firmly rejects the idea that God’s understanding of essences


precedes his creative activity. This is clearly a major departure from what
we found in Scotus. But this difference does not rule out the possibility
that Descartes thought they have objective being in God’s mind. He
might hold that God’s understanding/​willing/​creating the truths gives
them objective being in God as a result of efficient causality. (2008, 50)

I take this concession to be fatal to Rozemond’s interpretation. It is not just that


the proposed analogy with Scotus is now extremely tenuous. The deeper problem
is that by acknowledging Descartes’s view that in God understanding and willing
are the same, she is also acknowledging that he lacks any motivation for locating
essences in God. Such essences would be philosophically idle or otiose. Recall
that the scholastics posited creaturely essences in God to explain his knowledge
of finite beings and to show how divine creation is rational. As an interpreta-
tion of Descartes, Theological Platonism is motivated in part by just these con-
siderations, especially the former. But, as the scholastics understood, essences can
serve these functions only if they are in his intellect logically prior to his will. But
if Descartes rejects all such priority, and identifies God’s intellect and will, how
can such an account explain God’s knowledge of creation? It could explain how
he knows objective beings, for those are the ones he understands/​wills/​creates on
Rozemond’s proposal. But what about actually existing substances? Rozemond
is forced to that God understands/​wills/​creates them by a separate act. So that
means he understands or knows actually existing creatures by that second act,
not the first! The first act and the objective beings it produces are thus otiose. If
Rozemond wishes to insist that there is only one divine act, as she should, given
the doctrine of divine simplicity and given Descartes’s remark to Burman cited
above, then she is committed to the view I defend in the next section, that God
understands/​wills/​creates actually existing substances. In other words, she is
committed to reading Descartes as a Divine Direct Realist. That is the dilemma
for the Theological Platonist Interpretation: either it posits creaturely essences in
God that are metaphysically idle and violate his simplicity or it collapses into the
view I favor.
Descartes on Universal Essences and Divine Knowledge  •  103

5.4. Divine Knowledge of Creation


The medieval theory of divine ideas was intended to serve multiple functions. In
addition to explaining how God creates the world, it was also enlisted to explicate
how God knows creation. It delivers what might be termed a representational or
indirect realist account of the latter. God knows created substances by knowing
himself or, more specifically, by knowing the essences or exemplars that reside
in his intellect. As we have seen, Descartes rejects the theory of divine ideas on
the grounds that it violates God’s supreme simplicity. In fact, I have argued that
Descartes formulates the latter doctrine in the way that he does (by identifying
divine intellect and will), and draws consequences from it about the dependence
of all things on God’s will, because he sees himself as rejecting the traditional
theory of divine ideas as well as the accounts of exemplary creation and divine
knowledge that accompany it.
But if God does not know creatures through ideas then how does he know
them? As an omniscient being he must know them in some way and yet Descartes
says virtually nothing about this issue. One might be tempted to ascribe this
silence to the doctrine of divine incomprehensibility, which he takes very seri-
ously, and leave matters at that. However, Descartes typically invokes divine
inscrutability in contexts where God’s infinitude or his purposes in creation
are at issue. That is not the case here. Moreover, as we have seen, Descartes is
quite willing to draw consequences about the nature of God’s creation from the
doctrine of divine simplicity. What if further consequences followed from this
and related doctrines about the nature of God’s knowledge of creation? In this
section, I shall argue that at least one consequence does so follow. Specifically,
I  claim that Descartes is committed to “Divine Direct Realism”:  God knows
finite substances immediately without the mediation of ideas. Rozemond has
urged that Descartes’s theory of universal essences be traced to Scotus, but if
I am right then it makes more sense to look to Ockham, who was a direct real-
ist about divine cognition and, not coincidentally, a nominalist or conceptual-
ist about universals. Now, admittedly, the assertion that Descartes is a Divine
Direct Realist does not have the same status as interpretive claims for which
there is explicit textual evidence. But I think we can at least say that Descartes
ought to have held this view given his other commitments. One virtue of attrib-
uting Divine Direct Realism to Descartes is that it explains how God can know
a diversity of things without himself being diverse. In other words, it provides an
elegant solution to the Simplicity Problem as applied to divine knowledge: the
objects of God’s knowledge are complex, but the divine essence remains per-
fectly simple.
1 0 4   •  The Problem of Univer sals in Early Modern philosophy

Here then are two simple and straightforward arguments for this interpreta-
tion. These are “Cartesian” arguments in the sense that the premises are drawn
from Descartes’s philosophical commitments.

Argument I

(1) In God, willing and understanding are the same.


(2) Therefore, what he wills and understands are the same.
(3) What he wills or creates are actually existing substances (what Descartes
sometimes calls “existing essences”; see passage [8]‌below).
(4) Therefore, what he understands or knows are actually existing substances.

The first premise is a statement of Descartes’s version of the doctrine of divine sim-
plicity (see passages [2]‌–​[4] above). Premise (3) is a piece of Christian doctrine that
Descartes clearly accepts. The conclusion in step (4) is asserting that the immediate
objects of divine cognition are created substances themselves, not ideational proxies.
I have saved step (2) for last, since it is the hinge upon which the whole argument
turns. What is being claimed is that the objects of God’s will and understanding are
the same. Descartes clearly takes the inference from (1) to (2) to be valid, for he draws
the same inference when he derives the doctrine of the creation of the eternal truths
from divine simplicity. According to his reasoning, if divine intellect and will are the
same, then there is nothing in his intellect (such as ideas) prior to his will. Whatever
he understands he wills and vice versa. Thus, absolutely everything depends on God’s
will, including creaturely essences and the eternal truths concerning them.
The other argument for Divine Direct Realism can be formulated as an argu-
ment from elimination.

Argument II

(1) God knows creaturely essences either


(a) indirectly, by knowing himself, or
(b) directly, by knowing created substances themselves.
(2) Not (1a).
(3) Therefore (1b).

Let us take each premise in turn. One common objection to arguments from elim-
ination is that the enumeration of possible alternatives is not exhaustive. In this
case, however, the first premise is stated very generally, such that the two options are
jointly exhaustive (and mutually exclusive): God knows creaturely essences either
directly or indirectly. Premise (1a) is also stated broadly so that it is consistent
Descartes on Universal Essences and Divine Knowledge  •  105

with different versions of Theological Platonism. God might know creatures on


this view by having ideas of them in his understanding (Rozemond 2008) or by
knowing his volitions (Schmaltz 1991). We have already spent much of this chap-
ter considering the main grounds for accepting premise (2), namely, Descartes’s
strict conception of divine simplicity, which precludes the presence of multiple
ideas or volitions in God or a distinction between divine will and understanding,
as such a theory of divine knowledge would require. There are two further sup-
porting reasons for premise (2) that I discuss in what follows. I also consider and
reject objections to these reasons from advocates of Theological Platonism.
First, given his view that the essences of finite beings are created, Descartes
cannot countenance creaturely essences in God on pain of heresy. Recall that
Descartes maintains that whatever is in God is identical with him, but God can-
not be identical with (part of ) creation. I noted this problem in my original paper
on this topic (see Nolan 1997a, 188n13). Rozemond and Schmaltz have both
tried to address it by distinguishing divine causation of essences and eternal truths
from divine creation of existing things. According to them, Descartes draws a dis-
tinction like this in the May 1630 letter to Mersenne (from passage [4]‌):

[7]‌You ask what God did in order to produce them [i.e., the eternal
truths]. I reply that <from eternity he willed and understood them to be,
and by that very fact he created them>. Or, if you reserve the word <cre-
ated> for the existence of things, then he <established them and made
them>. (AT 1:152–​3, CSMK 25)

To explain and further support this distinction, Rozemond and Schmaltz appeal
to the fact that Descartes sometimes suggests that eternal truths are not actual
existents and thus not produced by God in the same way as finite substances
(Rozemond 2008, 50, 56; Schmaltz 2002, 83–​84). In the Sixth Replies, for exam-
ple, Descartes compares the eternal truths to the laws produced by a king and
says that such laws do not actually exist but are “moral entities” (ens morale) (AT
7:436, CSM 2:294).17 The proposed solution to the heresy problem is thus that
eternal truths and essences—​although caused by God—​are not creatures in the
same sense as finite minds and bodies. So there is no heresy involved in the claim
that they reside in God qua objective beings.
I have two rejoinders. First, I suggest that Descartes makes this point about
eternal truths being “established” and “made” rather than “created” only to satisfy

17.  Cf. earlier remarks to Mersenne in 1630 (AT 1:145, CSMK 3:23). The talk of a “moral
entity” lends itself to multiple interpretations and thus, in my opinion, cannot be used to adju-
dicate the debate.
1 0 6   •  The Problem of Univer sals in Early Modern philosophy

Mersenne. Note that he says “if you reserve the word <created> for the existence
of things.” Mersenne might have been shocked by the claim that the eternal truths
are created, given the skeptical consequences it seems to have and the lack of prec-
edent in the history of Christian thought. Descartes is best read as trying to soften
the blow. So this is merely a rhetorical point. Descartes is not really committed
to any robust distinction between two kinds of divine causation, which might
itself violate God’s simplicity. This suggestion is confirmed by the first sentence
of the letter, where Descartes writes: “You ask me by what kind of causality God
established the eternal truths. I reply: by the same kind of causality as he created
all things, that is to say, as their efficient and total cause” (AT 1:151–​2; CSMK
25). Descartes could not be clearer about his view here. Rather than distinguish-
ing two kinds of divine causation, he insists that God creates the eternal truths by
the same kind of causality that he created all things. It strains credulity to believe
that he would begin the letter with this unqualified statement and then, in the
very next paragraph, draw anything more than a nominal distinction between the
creation of the eternal truths and the creation of existing substances.
Second, even if creaturely essences were to have an attenuated form of being
in God as objective beings, they would still be “not nothing,” as Descartes says in
the Third Meditation concerning the objective being of human ideas. So to say
that God causes, rather than creates, eternal truths seems to be splitting hairs or
engaging in a verbal dispute—​something Descartes abhorred. The upshot is that
if Descartes held that God produces them in his intellect by an act of will, then
he would be guilty of heresy.
A second piece of evidence for premise (2)  of Argument II derives from
Descartes’s account of the essence-​existence distinction. If the essences of finite
beings were in God’s intellect prior to creation, then there would be a distinction
in reality (i.e., either a real or modal distinction in Descartes’s senses) between the
essence and existence of an actually existing created substance. This is so because
essences would have some form of (eternal) being prior to the actually existing
things in which they were instantiated; indeed, some scholastics spoke here of
esse essentiae. But as we observed in section 5.2, Descartes maintains that there is
merely a distinction of reason between essence and existence in all things.18 Thus,
there can be no essences in God’s intellect prior to his will.
Rozemond anticipates an argument like this and questions its soundness on
the following grounds: the scholastics treat the following as separate issues: (1) the
nature of the distinction between essence and existence within the created world,
and (2) the status of creaturely essences in God as knower and creator. She then

18. See esp. passage [6]‌.


Descartes on Universal Essences and Divine Knowledge  •  107

claims that in the passages from the Principles in which Descartes treats uni-
versals as ideas he is concerned with the first of these topics and not the second
(2008, 58).19 According to her, he simply means to deny that universal essences
are distinct entities in the objects to which we attribute them. But this position,
she argues, has no philosophical import for the status of essences in God. “Even
if Descartes thinks there is only a distinction of reason in the created world
between the thing and the universal, that leaves open the possibility that essences
have objective being in God’s mind” (2008, 59).
Let me begin by agreeing with one part of Rozemond’s claim. It is true that
the scholastics treat claims (1) and (2) as separate philosophical issues, and the
reason they do so is because these issues are motivated by different problems.
As we have noted, the scholastics posit ideas or essences in God, prior to cre-
ation, in order to show that creation is intelligible and rational, and to explain
God’s knowledge of creatures. But they draw a distinction between essence and
existence in order to mark the theological divide between God (the sole infinite
being) and finite, created beings. Traditionally, Christian philosophers mark this
divide by distinguishing God’s supreme simplicity from the composite nature of
finite beings. As noted in section 5.2, the scholastics countenance various forms
of composition, but foremost among them is the distinction between matter and
form. However, some thinkers such as Aquinas reject the doctrine of universal
hylomorphism (according to which all finite substances are composites of mat-
ter and form), for what is one to say about the case of purely spiritual substances
such as angels, which lack matter? The distinction between essence and existence
is thus enlisted to provide another form of composition to account for the con-
tingent and finite nature of creatures.20
Although there are two separate issues here, motivated by different problems,
Rozemond is wrong to conclude from this fact that these issues are not related
or that the position a philosopher stakes out on one does not have consequences
for the other. If one affirms a distinction in reality between essence and existence,
then one can—​without contradiction—​posit creaturely essences in God prior to
existence. There are multiple ways of drawing such a distinction:  for example,
Aquinas and others, like Giles of Rome, affirmed a real distinction between
essence and existence; Scotus and his followers drew a formal distinction; and so
on. The important case, however, is the distinction of reason. If one affirms this
type of distinction then one cannot, on pain of contradiction, posit creaturely
essences in God. To understand why this is so, we must turn to Francisco Suárez,

19. The main passage in question is at AT 8–​1:27–​8, CSM 1:212–13.


20. See Wippel 1982a and 1982b.
1 0 8   •  The Problem of Univer sals in Early Modern philosophy

the late scholastic thinker who seems to have influenced Descartes’s theory of
distinctions.
Like Descartes, Suárez argues that there is merely a distinction of reason
between essence and existence in all things and devotes a lengthy disputation to
this doctrine as applied to finite beings. One of his first moves is to establish the
following principle, which he sees as providing the necessary groundwork for the
doctrine of rational distinction: prior to being created by God, the essences of
finite beings are absolutely nothing. Suárez writes: “the essence of a creature, or
the creature of itself, and before it is made by God, has in itself no true real being
and in this precise sense of existential being, the essence is not some reality, but
it is absolutely nothing” (omnino nihil) (1983, 57). Let us call this the Omnino
Nihil Principle. As translator Norman Wells argues in his commentary on this
disputation, Suárez defends this principle as a way of dispensing with the two
competing traditions on the relation between essence and existence, as found
in the Thomists and the Scotists, who again posit a real and formal distinction,
respectively (Suárez 1983, Introduction, 7–​10). Both of these other traditions
take it as a premise that, prior to God exercising his efficient causality, the essence
of a creature is a res or thing in its own right and even enjoys some form of eter-
nal being (esse essentiae) in God. They also assume that these essences serve as
exemplars for divine creation.21 By arguing that essences are omnino nihil prior
to creation, Suárez sees himself as depriving these traditional accounts of their
foundation. He understands that there is a deep connection between the status
of essences prior to creation and the nature of the essence-​existence relation in
finite beings. In particular, he recognizes that he can establish that essence and
existence are identical, or merely rationally distinct, in finite beings only by first
showing that creaturely essences are omnino nihil prior to God’s creative activity.
Wells writes: “with the alleged eternal essences of his adversaries reduced from a
res to a nihil, there is no longer any sound basis for [a real or formal distinction] to
bedarken future discussions of essence and existence. In principle, with this des-
patching of the creature’s essence as an actual eternal res, the identity between an
actual essence and its actual existence is secured” (Suárez 1983, Introduction, 10).

21. Wells writes: “However cast, the prevailing feature [of the general position that there is a
distinction of reality between essence and existence] is that of an essence which in and of itself
enjoys a being or reality apart from any creative efficient causality of God. So much is this the
case that essences continue to perdure, though the actual things which embody them have
ceased to exist. In this light, the essences of finite beings, unlike their existence, are eternal,
necessary and uncaused by an efficient cause.” Wells notes that it is a very short step from this
premise to the conclusion that there is a real or formal distinction between essence and exis-
tence in creatures (Suárez 1983, 8).
Descartes on Universal Essences and Divine Knowledge  •  109

Contrary to Rozemond, at least one scholastic philosopher held that the status
of essences prior to creation and the essence-​existence distinction are intimately
related. And among scholastic philosophers, Suárez was of particular impor-
tance to Descartes. As Étienne Gilson (1979) proposes, he is the likely source of
Descartes’s theory of distinctions. Descartes even endorses Suárez’s Omnino Nihil
Principle. In his Conversation with Burman, Descartes’s youthful interviewer asks
him whether essence is prior to existence and whether in creating things God
“merely gave them existence.” Descartes reportedly replies:

[8]‌We are right to separate the two in our thought, for we can con-
ceive of essence without actual existence, as in the case of a rose in
winter. However, the two cannot be separated in reality in accordance
with the customary distinction; for there was no essence prior to exis-
tence, since existence is merely existing essence. So one is really not
prior to the other, nor are they separate or distinct. (AT 5:164, JC 24,
emphasis added)

Here, Descartes not only claims that “there was no essence prior to existence,”
thus endorsing the Omnino Nihil Principle, but he also underscores the infer-
ential link between this issue and the essence-​existence relation. Given this
principle, essence and existence can be separated in thought, but not in reality
according to the “customary distinction.” In reality, there is no distinction; “exis-
tence is merely existing essence.” By “customary distinction,” he likely means a real
distinction, which was favored by the Thomists, but whatever the case he clearly
intends to contrast a distinction in reality with a distinction in thought (or rea-
son), and to endorse the latter on the grounds that there was no essence prior to
actual existence.

5.5. General Objections to the Conceptualist


Interpretation and Replies
I turn now to two general objections to the Conceptualist Interpretation. The
main objection is that it cannot account for the sense in which Cartesian essences
are, in Descartes’s words, “immutable and eternal” (AT 7:64, CSM 2:45). The
conceptualist reading asserts that Cartesian essences, insofar as they are distin-
guished from actually existing substances, are nothing more than the objective
being of innate ideas in finite minds. But although they are immortal, finite
minds are not eternal and their existence does not extend even to the begin-
ning of time. They are also highly mutable. One is not always thinking about the
1 1 0   •  The Problem of Univer sals in Early Modern philosophy

essences of God, mind, body, or geometrical figures, and one’s thought moves
frequently from idea to another.
In my original paper on this topic (Nolan 1997a), I offered a complex and
multifaceted response to this objection, but critics focused on my ancillary
remarks and overlooked the main point. So I  begin by restating it in a differ-
ent way. According to divine voluntarism, which Descartes endorses, God cre-
ates all finite things, including essences and the eternal truths concerning them.
Thus, even before we ask about the ontological status of these essences, we can
ask a prior question: how can anything created be eternal? Created things have a
beginning in time. Indeed, it is an article of Christian faith that God created the
world in time.22 Thus, so-​called eternal truths might have an everlasting existence
in the future but not in the past. As for the immutability of such truths, if God
created them then it would seem that he could also change them. Descartes him-
self confronts this issue directly in an early letter to Mersenne where he debates an
imaginary interlocutor (the latter’s remarks are in quotation marks):

[9]‌It will be said that if God had established these [mathematical] truths
he could change them as a king changes his laws. To this the answer is: Yes
he can, if his will can change. “But I understand them to be eternal and
unchangeable.”—​I make the same judgment about God. “But his will is
free.”—​Yes, but his power is beyond our grasp. In general we can assert
that God can do everything that is within our grasp but not that he cannot
do what is beyond our grasp. It would be rash to think that our imagina-
tion reaches as far as his power. (April 15, 1630; AT 1:145–​4 6; CSMK 23)

Descartes does not assert positively that God could change the eternal truths.
But he also does not deny it. His position here, as in some other statements of the
creation doctrine, is nuanced and cautious. Given the innate ideas that God has

22.  Descartes sometimes distinguishes creation “from eternity” (or “from all eternity”) and
creation “in time.” In the Sixth Replies, he says that God creates the world in time but created
the eternal truths from eternity (AT 7:432–​36, CSM 2:291–​94). The claim that God creates
the eternal truths “from all eternity” appears in the correspondence as well (AT 1:152, CSMK
25). But one must be cautious in drawing any substantive conclusions from this distinction,
for elsewhere in the correspondence Descartes says that God wills the thoughts that enter a
person’s mind “from all eternity” and also decides from eternity which of our prayers he will
answer (AT 4:314, CSMK 272; AT 4:316, CSMK 273). In these contexts at least, saying that
something is willed from eternity makes a claim not about the temporal status of the thing
willed but about the nature of God’s will, namely that it is indifferent and inalterable (for the
latter attribute, see AT 5:166, CSMK 348). This is only a conjecture, but Descartes may have
wished to affirm that God creates all things from eternity (in the sense just explained), but
feared that doing so would give the appearance that he rejects the Christian doctrine that the
world has a beginning.
Descartes on Universal Essences and Divine Knowledge  •  111

implanted in us, we cannot conceive of the eternal truths as being other than they
are. However, given what we know about God’s omnipotence, nothing is beyond
his power. So, as he says elsewhere, we should “not dare to say that God cannot”
alter the eternal truths (AT 5:224, CSMK 358–​59). This doubly negative state-
ment does not assert a positive modal claim to the effect that God could change
the eternal truths.23 But it does not rule out this possibility, or so it has seemed to
many readers, and that is the source of the problem concerning the immutability.
My main reply to the objection, then, is that the problem of explaining how
Cartesian essences can be eternal and immutable has its source in Descartes’s doc-
trine that the eternal truths are created. The problem arises independently of the
Conceptualist Interpretation, which does not raise any new problems. Defenders
of Theological Platonism might acknowledge this point but still try to claim as
an advantage of their view that it solves the problem: if essences are in God then
there is a straightforward sense in which they have these properties, a sense that
is parasitic on God’s own immutability and eternality. As the arguments in the
previous sections of this chapter show, however, this proposal comes at a very
steep price for it violates Descartes’s strict understanding of divine simplicity and
saddles him with heresy. A solution that Descartes could accept only on these
terms is no solution at all.
This concludes the main part of my reply to this objection. What I  say in
what follows is secondary to that. Given the remarks above, it turns out that
the defender of conceptualism is not under any special obligation to provide a
solution to this difficulty. Nevertheless, I would like to propose one. In my orig-
inal paper on this topic I  offered a deflationary account of the sense in which
Cartesian essences are “immutable.” In what follows I  review that account, to
which I remain committed, and then show how a similar account can be offered
of their eternality. When Descartes first presents his theory of “true and immuta-
ble natures, essences, or forms” in the Fifth Meditation, he draws a sharp contrast
between them and invented ideas (or invented natures.). He claims to know, for
example, that the essence of a triangle is “not invented by me or dependent on my
mind” but “true and immutable” because

[10] various properties can be demonstrated of [it], for example that its
three angles equal two right angles, that its greatest side subtends its great-
est angle, and the like; and since these properties are ones which I now
clearly recognize whether I want to or not [velim nolim], even if I never
thought of them at all when I previously imagined the triangle, it follows
that they cannot have been invented by me. (AT 7:64, CSM 2:45)

23. I owe this insight to Nelson and Cunning 1999.


1 1 2   •  The Problem of Univer sals in Early Modern philosophy

In further elucidations of this issue in the First Replies, Descartes affirms that
invented ideas are composed by us. Like his empiricist contemporaries, he holds
a compositional theory of fictitious ideas. Because such ideas are composed by us,
they depend on our mind for their content. If, for example, I decide to fashion
an idea of a fictitious beast it is up to me to determine how many heads it has,
whether it breathes fire, can fly, and so on. By contrast, the content of innate
(or “true and immutable”) ideas imposes itself on my thought. Because the lat-
ter were created by God, and not by me, I  am compelled to think of them in
certain prescribed ways (velim nolim, as he says above). This is not to say that
I must ever attend to these ideas. On the contrary, I can think of them at will,
but I cannot alter them at will (AT 7:64, CSM 2:44–​45). They are incorruptible
by my thought. These remarks complement Descartes’s claim at the end of the
causal argument for God’s existence in the Third Meditation, where he says that
he knows that his idea of God is innate and not invented because he cannot “add”
anything to it or “subtract” anything from it (AT 7:51, CSM 2:35). This then is
the sense in which Cartesian essences, qua innate ideas, are immutable.
Defenders of Theological Platonism insist that we should understand the
eternality of Cartesian essences and truths in a strict and very literal sense. But
Descartes never says anything that requires this. On the contrary, given the prob-
lem posed by the creation doctrine, it seems impossible for him to understand
the eternality of essences in any but a deflationary sense. As creatures distinct
from God, they cannot be eternal either in the everlasting sense (sempiternity)
nor in the timeless sense that one associates with medieval accounts of God and
to which Descartes most likely subscribed. Eternality in the strict sense is an attri-
bute exclusive to God. Moreover, as I argued above, the texts suggest very clearly
that Cartesian essences are immutable in a deflationary sense. If that is so, then it
seems likely that he also conceived their eternality in a deflationary sense. In my
original paper on this topic, I argued that Descartes employs Platonic language
in the Fifth Meditation when he speaks of “true and immutable natures” and the
“determinate nature, essence, or form of the triangle” because he sees himself as
transplanting essences from Plato’s third realm to the minds of human beings
(Nolan 1997a, 184; AT 7:64, CSM 2:44–​45). If this is correct, then what I am
claiming in effect is that Descartes reduces Platonic essences to mind-​dependent
entities. Thus, it should not be surprising that he also has a reductive analysis of
the sense in which they are immutable and eternal.
Jonathan Bennett (1994, 663–​65) has claimed that Cartesian truths are eter-
nal in the same sense that they are immutable—​namely, they are unchanging.
Chappell (1997, 126) also endorses this view. While I think this suggestion moves
in the right direction, it is not quite right. As we have seen, Descartes conceives
the immutability of essences in terms of the fact that they, unlike invented natures,
Descartes on Universal Essences and Divine Knowledge  •  113

cannot be altered by our thought. Their content is causally independent from our
mind. That is saying something more than merely that they are “unchanging.” But
the property of being eternal is akin to being “unchanging.” A more precise way
of putting it would be to say that whenever we are clearly and distinctly perceiv-
ing a created essence, such as that of a triangle, we always perceive it as having the
same properties. That is not to say that we can never discover new properties in
our idea of it. Rather, it means that once having discovered some property, for
example, having angles equal to two right angles, we cannot exclude that property
from it in our thought, at least not clearly and distinctly. Descartes provides very
strong evidence for this reading in the Sixth Replies: “you cannot deny that many
truths can be demonstrated of these essences; and since they are always the same,
it is right to call them immutable and eternal” (AT 7:381, CSM 2:262).
It is interesting to note that one of Descartes’s followers, Pierre-​Sylvain Regis,
articulates a view like this. Regis’s remarks are telling because, although he has
a different conception of the nature of finite substance, he is one of the few
Cartesians who subscribed to the doctrine of the creation of the eternal truths.
But Regis explicitly denies that the so-​called eternal truths are eternal in any
strict or literal sense and suggests that the immutability of these truths admits of
a reductive analysis.

[11] It remains therefore that the numerical, geometrical, and metaphys-


ical truths are not at all eternal, but only that they are immutable, since
substances can always be compared with one another, and that God
willed that all souls be determined to conceive the same thing when they
compared them in the same manner. He willed, for example, that they
be determined to conceive this two, when they compared one unit with
another unit; that they be determined to conceive a triangle, when they
consider extension as bounded by three sides; and finally, that they be
determined to conceive a cause, when they consider a subject as produced
by another. This shows that the immutability itself of the so-​called eter-
nal truths is not absolute but dependent. (Regis 1690, 1:179–​80; trans.
Patricia Easton)

As Patricia Easton writes of this passage: “Thus, on Regis’s account, the impor-


tant feature of the eternal truths is their immutability and the consequence that
they are always conceived by the soul in the same way. The so-​called eternality of
truths is derived from the fact that the soul is determined to conceive the same
truths, so that in this sense, though not eternal in themselves, these truths are
forever the same” (2009, 354). Descartes, then, was not the only Cartesian who
saw the need to offer a deflationary account of the eternality and immutability of
1 1 4   •  The Problem of Univer sals in Early Modern philosophy

truths created by God. I have argued that it is precisely because they are created
that Descartes conceived of them in this way.
In recent years, Schmaltz has raised a new objection to the Conceptualist
Interpretation pertaining to the nature of God. The objection is that, if true, the
interpretation must be restricted to finite essences on the grounds that God’s
essence is identical with his existence and thus cannot be merely an innate idea in
the human mind:

Descartes wanted to set God’s essence apart from the other created
essences . . . Given the ontological argument the divine essence cannot be
identified with any feature of our mind, however enduring or innate. The
indication in Descartes is that this external essence is, in fact, identical to
the supremely perfect being that exists external to our mind. (Schmaltz
2002, 85)24

There is a straightforward answer to this objection. Given Descartes’s view that


there is a rational distinction between essence and existence in all things, the
term “essence” is systematically ambiguous.25 Sometimes this term is intended to
mean an actually existing substance (sense 1), but at other times it means an idea
in the human mind (sense 2). This dual usage can be found at various places in
Descartes’s writings. For example, in passage [8]‌, cited above, Descartes employs
both senses of the term “essence” in the course of three sentences. He says that we
are right to separate essence and existence in our thought

for we can conceive of essence without actual existence, as in the case


of a rose in winter. However, the two cannot be separated in reality in
accordance with the customary distinction; for there was no essence prior
to existence, since existence is merely existing essence. So one is really not
prior to the other, nor are they separate or distinct. (AT 5:164, JC 24;
emphasis added)

In the first line quoted here, Descartes uses the term “essence” in sense 2, since we
are conceiving of the rose in abstraction from its existence, but when he speaks
toward the end of the passage of “existing essence,” the term “essence” clearly
means an actually existing substance (sense 1). This double meaning of the term

24. Cf. Schmaltz 2014, 209.


25. The expression ‘systematically ambiguous’ recalls Bertrand Russell, but I am using it solely
in the sense indicated here.
Descartes on Universal Essences and Divine Knowledge  •  115

“essence” is not confined to finite substances. According to Descartes’s ontologi-


cal argument, we cannot separate or exclude necessary existence from our clear
and distinct idea of God, but we can regard God’s essence in abstraction from his
existence, just as we can in the case of finite substances. One of the places where
Descartes acknowledges the latter point is in passage [6]‌. Having noted that all
of the divine attributes are merely rationally distinct, Descartes adds, “because we
do indeed understand the essence of a thing in one way when we consider it in
abstraction from whether it exists or not, and in a different way when we consider
it as existing; but the thing itself cannot be outside our thought without its exis-
tence” (AT 4:349, CSMK 280). Here too the term “essence” is used in the two
different senses distinguished above. We can consider God’s essence in abstrac-
tion from his existence, in which case we are regarding an idea, or we can con-
sider it as existing, that is, as it is outside our thought. Contrary to Schmaltz, it is
consistent with Descartes’s claim that in reality God’s essence just is his existence
that we are able to distinguish the divine essence from existence in our thought.
In fact, that is what it means to say that they are merely rationally distinct.
One final point on this topic: when scholastics philosophers drew a distinc-
tion between essence and existence, they followed up by asking about the status
of the essence that is so distinguished. The Conceptualist Interpretation is mak-
ing a claim about that status. Here, then, is a way of formulating the interpreta-
tion that clarifies the relevant ontological issue: all essences, insofar as they are
distinguished from actually existing substances, are merely ideas in finite minds.

5.6. Conclusion
In previous work on the ontological status of Cartesian universals, I argued that
Descartes is a conceptualist who locates essences in the minds of human beings
as innate, intellectual ideas (Nolan 1997a). I also offered a reconstruction of his
general theory of universals (the details of which are only partially articulated in
the few short texts devoted to this topic) and revealed it to be a corollary to his
theory of attributes (Nolan 1997b and 1998). I developed a new defense of this
interpretation in this chapter by locating the source of Descartes’s conceptual-
ism in his view that all substances—​both created and divine—​are simple. As a
result of this simplicity, there is no room in the created world for universals as
shared properties. There is also no place for universal essences or eternal truths
within God.
The latter point raises an important question about the nature of God’s
knowledge of creation. Many medieval philosophers held that God knows crea-
turely essences by knowing ideas in his understanding, and that he uses these
ideas as exemplars for creation. But I have shown that Descartes formulates the
1 1 6   •  The Problem of Univer sals in Early Modern philosophy

doctrine of divine simplicity in the way that he does—​by identifying God’s intel-
lect and will—​because he sees himself as rejecting these traditional accounts of
divine knowledge and creation, which in his view anthropomorphize God. The
doctrine of divine simplicity, together with a few other central doctrines, commit
Descartes to what I call Divine Direct Realism: God knows creaturely essences
by knowing created substances themselves, without the mediation of ideational
archetypes in his intellect. Unlike the traditional theory of God’s knowledge,
Divine Direct Realism complements Descartes’s conceptualism concerning the
status of universals.26 This chapter has also afforded the opportunity to answer
objections to the Conceptualist Interpretation and to identify failings of rival
Platonist readings.

Abbreviations
AT = Descartes 1964–​74; cited by volume and page.
CSM = Descartes 1984–​85; cited by volume and page.
CSMK = Descartes 1991; cited by page.
JC = Descartes 1976; cited by page.

26.  I  am grateful to several people for their suggestions on previous drafts of this chapter,
most notably Stefano Di Bella, Nicholas Jolley, Cathay Liu, Alan Nelson, and Al Spangler.
I also thank Tad Schmaltz for many lively exchanges over the years, including at the confer-
ence on “The Problem of Universals in Modern Philosophy” (Scuola Normale Superiore, Pisa,
Italy, November 7–​9, 2011), where he was the commentator. The many participants at that
conference are owed a debt, including Don Baxter, Stefano Di Bella, Antonia LoLordo, Sam
Newlands, Mariangela Priarolo, and Tom Stoneham. In addition, I received helpful feedback
at the conference on “Theories of Ideas in Early Modern Philosophy” (University of North
Carolina, Chapel Hill, May 13–​15, 2011), especially from Patrick Connolly, Cathay Liu, Alan
Nelson, Lex Newman, David Owen, Raphaella De Rosa, and Martha Brandt Bolton.
6 P L AT O N I S M A N D C O N C E P T U A L I S M A M O N G
THE CARTESIANS

Tad M. Schmaltz

6.1. Introduction
A primary question raised by the traditional “problem of univer-
sals” concerns the ontological status of universal features of reality.
One of the early modern offshoots of this problem stems from the
views of René Descartes (1596–​1650) regarding “true and immutable
natures” and eternal truths. In broad terms, the ancient debate over
universals divided Platonists, who held that universals exist in a sepa-
rate “third realm,” from Aristotelians, who attempted to place uni-
versals in nature. In comparison, the division in the recent literature
on Descartes is between those who see him as adopting the broadly
Platonic view that immutable natures are independent both of the
human mind and of the particular objects in the created world that
exemplify those natures, on the one hand, and those who see him as
endorsing a conceptualist reduction of such natures to features of the
human mind, on the other. There is a related controversy regarding
the ontological implications of Descartes’s famous (some would say
infamous) doctrine of God’s free creation of eternal truths. Though
there is some support in Descartes’s texts for the identification of these
truths with the ideas concerning them that God has imprinted in our
mind, there also is reason to worry that such an identification cannot
fully accommodate everything Descartes has to say concerning eternal
truths and immutable natures.
The main concern in recent commentary on Descartes (and here
I include my own past work) has been to eliminate these tensions by
taking his thought in either a Platonist or a conceptualist direction.
My aim here, however, is not to eliminate the tensions, but indeed
to emphasize the importance of their role in later debates among
1 1 8   •  The Problem of Univer sals in Early Modern philosophy

Descartes’s followers. I take these tensions to be relevant in particular to debates


over the nature of ideas, immutable natures, and eternal truths that involve three
prominent French Cartesians:  Nicolas Malebranche (1638–​ 1715), Antoine
Arnauld (1612–​1694), and Pierre-​Sylvain Regis1 (1632–​1707). Malebranche
takes Descartes’s doctrine of the creation of eternal truths to undermine the
eternality and immutability of such truths, and he argues that these features can
be preserved only if one adopts the version of Platonism in Augustine, accord-
ing to which we know truths concerning essences only by means of God’s own
eternal and uncreated ideas. Though he doesn’t make an issue of Descartes’s
doctrine, Arnauld nonetheless ultimately responds to Malebranche and “the
Platonists” by insisting on the conceptualist position that we have access to truths
concerning essences only by means of our own ideas. One implication of this
conceptualism—​which Arnauld explicitly embraces in his later work—​is that
these truths do not have the sort of eternality and immutability that truths in
the divine intellect possess. In contrast, Regis takes the Cartesian doctrine of the
creation of eternal truths to require that created eternal truths are grounded in
something that is atemporal and, therefore, immutable. Though he recognizes
that this doctrine does not allow for Malebranche’s explanation of these features
of the truths in terms of God’s uncreated ideas, Regis proposes that these features
can be explained in terms of mind-​independent features of the created world. In
Regis, then, we have a turn away from Arnauld’s Cartesian conceptualism that
nevertheless does not involve a return to Malebranche’s Cartesian Platonism.

6.2. Descartes on Immutable Natures and


Eternal Truths
Let us start with the classic text with respect to Descartes’s view of immutable
natures, from the Fifth Meditation:

I find within me innumerable ideas of things that even though they may
not exist anywhere outside me, nonetheless cannot be called nothing; and
although they can be thought in some manner by will, nonetheless they
are not made by me [a me figuntur], but have their own true and immu-
table natures. Thus when, for example, I imagine a triangle, even if perhaps
no such figure exists outside my thought, nor has ever existed, there is
still some determinate nature, or essence, or form, immutable and eternal,

1. Aka Régis; here I follow the convention—​which to my knowledge derives from the work
of Geneviève Rodis-​Lewis—​of indicating the Latin form of the name by omitting the accent.
Platonism and Conceptualism among the Cartesians  •  119

which is not produced by me [a me non efficta est], nor depends on my


mind [nec a mente mea dependet]. (AT 7:64)

There is an explicit distinction here of the “true and immutable nature” of a trian-
gle from particular triangles external to mind. However, such a nature might also
seem to be distinct from the mind that considers it, given Descartes’s insistence
that the nature does not depend on his mind. And indeed, primarily on the basis
of this Fifth Meditation passage, Anthony Kenny has claimed that “Descartes’s
philosophy of mathematics [. . .] is thoroughly Platonic; indeed he is the founder
of modern Platonism.” In particular, Kenny takes the triangle that “does not exist
anywhere outside me” to be “an eternal creature of God, with its own immutable
nature and properties” (Kenny 1970, 692–​63).
However, there is a serious problem with Kenny’s claim that Descartes’s view
of immutable essences is “thoroughly Platonic.” In particular, there is a strong con-
ceptualist bent to Descartes’s discussion of “eternal truths” in the later Principia
Philosophiœ. Thus, for instance, there is the claim in this text:

When we recognize that it is not possible that something come from


nothing, this proposition, Nothing comes from nothing, is considered not
as some existing thing, nor also as a mode of a thing, but as some eternal
truth, which resides in our mind. (AT 8[1]‌:23–​24)

Martial Gueroult has argued that passages such as these suffice to rule out any
Platonic reading of Descartes view of immutable natures.

Nothing is more contrary to Cartesianism than the realism of Platonic


ideas and the exemplarism of essences. If one is allowed to speak of the
realism of essences, it is to the extent that, within my mind, clear and dis-
tinct ideas are presented as realities against which my own thought cannot
do anything, since it is powerless to annihilate them or to modify them—​
in brief, it is to the extent that they are true and immutable natures,
implanted in me by God. (Gueroult 1984, 1:277)

More recently, other commentators have argued that the conceptualist read-
ing finds support even in the very discussion of immutable natures in the Fifth
Meditation that Kenny takes to reveal a clear Platonist bias. Thus, both Vere
Chappell and Lawrence Nolan have proposed that the true and immutable
natures invoked in this text are to be identified with the ideas that we have of
them. More specifically, the proposal is that the natures are identical to the
“objective being” of the objects of innate ideas (Chappell 1997 and Nolan
1 2 0   •  The Problem of Univer sals in Early Modern philosophy

1997a). To understand this proposal, we need to say more about both objective
being and innateness.
In the Third Meditation, Descartes takes the objective being of an object to
be a mode of being “by which a thing exists in the intellect by way of an idea” (AT
7:41). The proposal that true and immutable natures are to be identified with
objective being is conceptualist insofar as it requires that the natures are merely
certain features of our ideas of those natures. Though this requirement may seem
to conflict with Descartes’s insistence that the natures do not depend on the
mind, both Chappell and Nolan claim that this insistence is to be understood in
terms of Descartes’s distinction between innate and “factitious” ideas. In the case
of a factitious idea, such as that of a winged horse, the objective being has a struc-
ture that has been constructed, and can be deconstructed, by the mind. But in the
case of an innate idea, such as that of a triangle, the structure is preset and cannot
be altered by the mind. Thus Descartes notes in the First Replies—​in the course
of addressing his Fifth Meditation account of true and immutable natures—​that
the fact that the idea of a triangle is not merely “made” by the mind is revealed by
the fact that “I cannot deny that this property [of having three angles equal to two
right angles] applies to the triangle by a clear and distinct intellectual operation”
(AT 7:117–​18). And in the Fifth Meditation itself, he claims that such mathemati-
cal truths are “so open and so much in harmony with my nature that on first dis-
covering them … it seems like noticing for the first time things which were long
present within me although I had never turned my mental gaze on them before”
(AT 7:64).2
According to Chappell and Nolan, then, when Descartes says that true and
immutable natures do not depend on his mind, he means merely that his ideas
of those natures are not constructed, but rather reveal truths that are innate to
his mind. Just as Chappell claims that Descartes’s true and immutable natures
are simply “the ideas that God makes to be innate in us” and that “are constant
and never change” (Chappell 1997, 125), so Nolan holds that these natures are
innate ideas that “impose their content on our thought, compelling us to think
of them in certain prescribed ways.” Nolan concludes: “There is nothing in this
claim that commits Descartes to a transcendental realm of extra-​mental objects”
(Nolan 1997a, 183–​84).
There is much to be said for this conceptualist reading of the passage from
the Fifth Meditation. Especially telling here is Descartes’s claim in this text that

2. There has been the claim in the literature that Descartes offers competing views of the related
distinctions between innate and factitious ideas, on the one hand, and immutable and con-
structed natures, on the other. For a further discussion of this literature, and an attempt to
reconcile and develop Descartes’s views on this issue, see Schmaltz 2014.
Platonism and Conceptualism among the Cartesians  •  121

the properties that he can demonstrate of triangles “are certainly true, since I am
clearly aware of them, and therefore they are something, and not merely noth-
ing; for it is obvious that whatever is true is something” (AT 7:65). Earlier, in
the Third Meditation, he notes that the being by which something exists objec-
tively in the intellect, “imperfect though it may be, is certainly not nothing” (AT
7:41). In light of this comment concerning objective being, it would be natural
to take the claim in the Fifth Meditation to be that the properties of triangles
are “something” in the sense that they are present objectively in the innate idea
of a triangle.
However, there is a real question whether a conceptualist reading can fully
accommodate what Descartes says in this text about the true and immutable nature
that is of most interest to him there, namely the nature of God as a “supremely per-
fect being.” For this nature is presented as something that is identical not to the
objective being of an innate idea, but rather to the actually existing being, God.
Indeed, Descartes emphasizes that our conception of the necessity of God’s existence
is imposed on our mind from this external source. Thus he concludes:

And from the fact that I cannot think of God except as existing, it follows
that existence is inseparable from God, and thus that he really exists; not
because my thought produces this [hoc efficiat], or imposes any necessity
on anything, but on the contrary the thing itself, namely the existence of
God, necessarily determines me to think this; (AT 7:67)

So in the case of God, at least, it is ultimately not God insofar as he exists in our
thought that compels me to think of God as existing, but rather God himself as
an actual and external nature. In at least one case, then, a conceptualist account of
true and immutable natures is unacceptable.3
But perhaps God is a special case. One could argue that in cases where natures
do not include existence—​namely, in all other cases4—​there seems to be no bar-
rier to identifying those natures with features of our innate ideas. There could
still be an external source of the ideas insofar as God imposes them—​or better,

3. For a similar objection to Nolan’s conceptualist reading, see Cunning 2010, 158–​60. On
Cunning’s view, the true and immutable nature of an object is identical to the actually exist-
ing object itself in general, and not just in the special case of God. However, this view seems
to conflict with the suggestion in the Fifth Meditation that an object such as a triangle—​
or indeed, any created thing—​would have a true and immutable nature even if it didn’t
actually exist.
4.  Indeed, the argument for the existence of God in the Fifth Meditation stresses that the
essence of God is unique in requiring the actual existence of its object.
1 2 2   •  The Problem of Univer sals in Early Modern philosophy

perhaps, the faculty of thought from which they arise5—​on us. In contrast to the
case of the innate idea of God, the ideas do not derive their content from any
extramental natures.
Even so, there is a difficulty for a conceptualist understanding of true and
immutable natures of creatures that is broached by Descartes’s claim in the Fifth
Replies that

I do not think that the essences of things, and the mathematical truths
which we can know concerning them, are independent of God; but nev-
ertheless I think they are immutable and eternal, because God willed and
arranged [disposuit] that they should be so. (AT 7:380)

Descartes is here presupposing his doctrine of the creation of eternal truths,


according to which such truths derive from God’s indifferent will. A conceptu-
alist can perhaps accommodate the claim that God freely willed and arranged
that the natures and truths be immutable by stressing the constancy of our innate
ideas (or the structure of our faculty of thinking). However, Descartes himself
suggested that the immutability of the natures and truths is to be explained in
terms of the fact that they are eternal. I find such a suggestion in the 1630 cor-
respondence that introduces his creation doctrine, in which Descartes offers the
following dialogue with an imagined critic:

One will say that if God has established these truths, he could change
them as a king changes his laws; to which it is necessary to respond, yes,
if his will can change.—​But I understand them to be eternal and immu-
table.—​And I judge the same of God.—​But his will is free.—​Yes, but his
power is incomprehensible. (AT 1:145–​4 6)

What seems difficult to accommodate on a conceptualist interpretation is the


implication here that the created truths are immutable in virtue of the fact that
they have a sort of eternity that is akin to the eternity of the divine will.
It might be thought that given Descartes’s own emphasis on the incompre-
hensibility of God’s creation of eternal truths, one should not expect an under-
standing of the eternity of these truths.6 However, when Descartes stresses—​as
in the passage just quoted—​the incomprehensibility of the power involved in

5. See Descartes’s claim in the Notae in programma quoddam that he never held that innate
ideas “are some sort of ‘forms’ that are distinct from our faculty of thinking” (AT 8[2]‌:366).
6. As proposed, for instance, in Nolan 1997a, 184–​86.
Platonism and Conceptualism among the Cartesians  •  123

the creation of the truths, he does not have in mind problems concerning the
ontological status of these truths. Rather, his point is that we cannot compre-
hend the indifference involved in God’s act of creation. Thus in the Sixth Replies,
Descartes indicates that even though it is unintelligible that mathematical truths
are false, still it follows from divine indifference that God could have brought this
about from eternity (AT 7:436). What is unintelligible here is not the fact that
the truths are eternal, but rather the fact that God was completely indifferent in
instituting such truths.
The ontological problem of the eternity of created eternal truths seems to
remain with us. But one possible conceptualist response is to deny that we can
take Descartes’s claims that these truths are eternal at face value; so, for instance,
we have Chappell’s conclusion that “the objects and truths of mathematics are
not, for Descartes, strictly and literally eternal” (Chappell 1997, 127).7 Indeed, we
will see that this is the sort of conceptualist route that Arnauld embraced. But in
suggesting that created truths and natures have something akin to divine eternity
that distinguishes them from temporal creatures, Descartes himself introduces a
roadblock for Arnauld’s conceptualism.
In earlier work, I  attempted to accommodate Descartes’s insistence on the
eternity of the truths and natures by identifying them with the decrees by which
God establishes them (Schmaltz 1991). More recently, Marleen Rozemond has
defended the related “moderate Platonic” view that Descartes’s created truths and
natures are to be identified with the objective being of God’s thoughts concern-
ing them (Rozemond 2008).8 As conceptualist critics have noted, however, this
Platonist alternative also is problematic for Descartes.9 In particular, this view
does not seem to fully accommodate Descartes’s insistence that created eternal
truths are distinct from God. Thus, we have his claim, again from the 1630 cor-
respondence on his created truths doctrine, that

these [eternal] truths are no more necessarily conjoined to [God’s] essence


than other creatures are. You ask what God has done to produce them.

7. Descartes allows that though the idea of God is formally finite, it is nonetheless objectively
infinite. One might think that he could say in the same way that an idea that is formally tempo-
ral could nonetheless be objectively eternal. However, in the case of an idea that is objectively
infinite, there is a corresponding object that is at least potentially formally infinite. Moreover,
when Descartes tells Mersenne that God establishes eternal truths, he suggests that God cre-
ates something that is eternal not merely objectively but also formally. What Chappell denies,
however, is that Descartes’s ontology allows for any eternal object that corresponds to created
eternal truths.
8. Cf. the endorsement of a “quasi-​Platonic” reading of Descartes in Wilson 1978, 171.
9. See, e.g., Nolan 1997a, 171–​72.
1 2 4   •  The Problem of Univer sals in Early Modern philosophy

I say that from all eternity he willed and understood them to be, and by that
very fact created them, or better (if you reserve the word created for the
existence of things) he disposed and made them. (AT 1:152–​53)

What seems difficult to conceive, however, is that divine decrees, or the objec-
tive being of his thoughts, could be disjoined from God’s essence in the way that
Descartes requires.
Just as I  suggested that one could save the conceptualist interpretation of
Descartes by downplaying his claim that created truths and natures are eternal,
so perhaps one could save a moderate Platonic interpretation of his views by
qualifying his claim that the truths and essences God creates are not “necessarily
conjoined” to his essence.10 As I have indicated, however, my purpose here is not
to save any particular interpretation of Descartes.11 Rather, the goal is to high-
light certain tensions in his thought that played a crucial role in later Cartesian
debates. And the particular tension I  have highlighted is between Descartes’s
claim that created truths and natures are robustly eternal (and therefore immu-
table), on the one hand, and the claim that such truths and natures are creatures
distinct from God, on the other. In later Cartesian debates, Malebranche offers
a Platonist position that emphasizes the former claim, whereas Arnauld offers
a conceptualist position that emphasizes the latter claim. But there also is the
view of Regis, who offers a non-​Platonist but also anticonceptualist position that
attempts to accommodate both of Descartes’s claims.

6.3. Malebranche’s Platonism
In my discussion of Descartes I have drawn attention to the importance of his
doctrine of the creation of eternal truths (hereafter, the created truths doc-
trine). This doctrine also plays a prominent role in the initial critical response to
Malebranche’s Recherche de la vérité. In the first volume of the first edition of this
text, published in 1674, we find the following:

There are two truths, the first are necessary and the others contingent.
I  call truths necessary that are immutable by their nature, and because

10. I earlier proposed (but have come to doubt) that Descartes could accommodate the dis-
tinction from God by distinguishing between God’s “strong” and “weak” attributes (Schmaltz
1991, 156–​58). Rozemond proposes that Descartes can hold that the objective being of God’s
thoughts differs “in reason” from his essence (Rozemond 2008, 53–​56).
11. But see Nolan’s further defense in chapter 5 of this volume of his “conceptualist” interpreta-
tion of Descartes.
Platonism and Conceptualism among the Cartesians  •  125

they have been fixed by the will of God, which is not subject to change.
(OCM 1:63n)

In the first response to the Recherche, the Critique de la Recherche de la vérité,


which was written prior to the publication of the second volume of this text and
published in 1675, the French Academic skeptic Simon Foucher objects to this
claim on the following grounds:

[I]t seems that he [Malebranche] considers the immutability of the


will of God as the cause of the necessity of these [necessary] truths.
But if this is so, he proves too much in the place where we have pro-
tested that he proves nothing at all. For if what God wills is immutable
because his will is not subject to change, does it not follow that all
that he wills must have an equal immutability, since it is the cause?
However, it is certain that he wills things that are subject to change
when he determines creatures to exist or to cease to be in the vicis-
situde of time. (CRV 30)

The objection here is that the immutability of the divine will cannot explain
the immutability of necessary truths, since this same immutable will is the
cause of mutable creatures. Notice that this objection applies not only to
Malebranche’s suggestion in the first edition of the Recherche that necessary
truths are immutable because God’s will “is not subject to change,” but also
to the indication in Descartes’s 1630 correspondence that the eternity and
immutability of created eternal truths follows directly from the immutability
of God’s eternal will.
Malebranche subsequently took care to distinguish himself from Descartes
on the issue of the status of necessary truths. Thus, in the final version of the pas-
sage that Foucher cited, Malebranche indicates that there are necessary truths
that do not derive from God’s will.

There are two truths, the first are necessary and the others contingent.
I call necessary truths those that are immutable by their nature, and those
that have been fixed by the will of God, which is not subject to change.
(OCM 1:63)

And in an Éclaircissement appended to later editions of the Recherche,


Malebranche insists that necessary truths that are immutable by their nature can-
not also be immutable by means of a free act of the divine will. For any depen-
dence on such an act would deprive these truths of their intrinsic necessity, and
1 2 6   •  The Problem of Univer sals in Early Modern philosophy

render them as dubitable as any contingent truth. His argument there, which is
directed explicitly against Descartes, is as follows:

Surely, if eternal laws and truths depended on God, if they had been estab-
lished by a free volition of the Creator, in short, if the Reason we consult
were not necessary and independent, it seems evident to me that there
would no longer be any true science and that we might be mistaken in
claiming that the arithmetic and geometry of the Chinese is like our own.
For in the final analysis, if it were not absolutely necessary that twice four be
eight, or that the three angles of a triangle are equal to two right angles, what
assurance would we have that these kinds of truths are not like those that are
found only in certain universities, or that last only for a certain time? Do we
clearly conceive that God could not have willed certain things, for a certain
time, for a certain place, for certain people, or for certain kinds of beings—​
given, as some would have it, that He was entirely free and indifferent in
His willing? As for me, I can conceive no necessity in indifference, nor can
I reconcile two things that are opposite. (OCM 3:132)

It is more than a little ironic that Malebranche here emphasizes the very point
that Foucher had earlier raised against his own view of necessary truths. It turns
out that Foucher and Malebranche agree that since the divine will can immuta-
bly produce changing objects, the immutability of this will cannot guarantee the
eternity and immutability of necessary truths.12
Malebranche’s argument would also apply to the view that God creates eter-
nal truths and immutable natures by creating the human mind with a particular
sort of innate structure. In this case, the objection would be that God could create
a structure limited to certain people, or lasting only for a certain amount of time.
For Malebranche, the eternity and necessity of eternal truths and immutable
natures can be guaranteed only if they derive from some feature of reality inde-
pendent of our minds that is itself eternal and necessary. Malebranche holds that
the only such feature of reality is the set of divine ideas that serve as the eternal
and immutable archetypes for God’s creation of the mutable world.
Malebranche also argues that the universality of certain ideas requires their
identification with divine archetypes that cannot themselves be created. Thus he
notes in the Recherche that

it is clear that the idea, or immediate object of our mind, when we think
about limitless space, or a circle in general, or indeterminate being, is

12. On the history of the exchange between Foucher and Malebranche, see Watson 1966 and
Gouhier 1976.
Platonism and Conceptualism among the Cartesians  •  127

nothing created. For no created reality can be either infinite or even gen-
eral, as is what we perceive in these cases. (OCM 1:432)

Because the reality that we perceive with respect to general objects or truths is
itself general, it cannot be created. But whereas the traditional Platonic argument
is that universal Forms must exist apart from all particulars, Malebranche’s claim
is that the general reality we perceive must be contained in the only uncreated
being there is, namely, God.13
This claim of course broaches the doctrine—​which Malebranche introduced
in the Recherche—​of “the Vision in God.”14 The inspiration for this doctrine is
Augustine, who Malebranche takes to have shown “in an infinity of passages
[that] we already see God in this life through the knowledge we have of the eter-
nal truths” (OCM 1:444). As far as I can determine, Malebranche himself does
not acknowledge the Platonic roots of the Augustinian view. However, the asso-
ciation of Malebranche with Platonism emerges in the work of one of his defend-
ers, Henri de Lelevel. In response to Regis’s critique of Malebranche’s distinction
between ideas and perceptions (see below), Lelevel appeals in his 1694 La Vraye
et fausse métaphysique to the position in Plato that “the ideas we have of corporeal
beings are the exemplars of these beings” (VFM 1:5).15 Lelevel endorses in partic-
ular the version of this Platonic alternative in Malebranche, on which the object
of our perception of the material is an “intelligible extension” in God that serves
as the eternal and immutable archetype for his creation of material objects.16
According to Lelevel, this Platonic position in Malebranche blocks a sort
of skepticism that derives from the Cartesian created truths doctrine. Just as
Malebranche had earlier objected to this doctrine that “there can be no necessity
in indifference,” so Lelevel holds that the doctrine “reverses religion and the per-
fect sciences” (VFM 1:42). For both, the only way to safeguard our knowledge
of necessary truths concerning extension is to ground such knowledge directly in
God’s eternal and immutable idea of extension. One commentator has correctly
noted that “the desire to avoid Cartesian voluntarism and to defeat skepticism in

13. For further discussion of Malebranche’s account of universals, see Mariangela Priarolo’s dis-
cussion in chapter 7 of this volume.
14.  For more on this doctrine and its relation to Malebranche’s theory of ideas, see
Schmaltz 2000.
15. For more on the views in Regis to which Lelevel is responding, see §6.5 below. For further
discussion of Lelevel’s exchange with Regis, see Schmaltz 2002, 251–​56.
16. For more on Malebranche’s account of intelligible extension, see Reid 2003. Reid is con-
cerned to respond to the view in the literature that “ultimately, Malebranche has no satisfac-
tory answer to any of the ontological questions regarding ideas” (Nadler 1992, 150; cf. Jolley
1990, 78–​80).
1 2 8   •  The Problem of Univer sals in Early Modern philosophy

a way Descartes could not is one of the major motives for Malebranche’s whole
theory of ideas” ( Jolley 1990, 11). We can add that this same desire plays a major
role in Lelevel’s defense of Malebranche’s theory.
In a comment particularly germane to Descartes’s discussion in the Fifth
Meditation of mathematical natures, Malebranche counts himself among those
who believe “that geometrical and numerical truths, such as 2 times 2 is 4, are
eternal, independent, preliminary to the free decrees of God” (OCM 17[1]‌:308).
Of course, Descartes is committed to rejecting the claim that the eternal truths
of mathematics are “preliminary to the free decrees of God.” As Lelevel’s defense
makes clear, Malebranche’s Platonism leaves no room for the Cartesian created
truths doctrine. But we have seen that Descartes nonetheless takes these truths
to be genuinely eternal. And for Malebranche and Lelevel alike, the eternity of
mathematical truths requires that they be grounded in ideas in God’s eternal
mind rather than in perceptions in our own temporal mind. If any Cartesian was
a representative of what—​to borrow Kenny’s phrase—​was a “thoroughly Platonic
philosophy of mathematics,” Lelevel’s discussion indicates that it is Malebranche
rather than Descartes.

6.4. Arnauld’s Conceptualism
Malebranche’s Platonism did not go unchallenged. Among the Cartesians,
the most prominent challenge was from Arnauld. Indeed, the debate between
Arnauld and Malebranche was one of the major intellectual events of the early
modern period.17 In light of Malebranche’s firm opposition to Descartes’s doc-
trine of the creation of the eternal truths, one might anticipate that this doc-
trine would play a central role in their debate. This expectation would only be
increased by the claim in the recent literature that Arnauld was deeply commit-
ted to this doctrine.18 However, in his initial responses to Malebranche in the
1680s—​viz., Des vraies et des fausses idées (1683) and Défense … contre la Réponse
au livre des vraies et des fausses idées (1684)—​Arnauld sidesteps this doctrine.
Indeed, the ontological status of the eternal truths is not an issue for him at all in
these texts.19 What he emphasizes rather is the question of whether we perceive
external objects in the material world by means of “representative beings” distinct

17. This debate concerned not only the nature of ideas, but also—​and in later stages of the
debate, especially—​theological issues concerning the distribution of grace and divine provi-
dence. For discussion of the various issues involved in this debate, see Moreau 1999.
18. See, for instance, Moreau 1999, chap. 6, and Nadler 2008.
19.  Though there are some anticipations of the line he develops in his later writings. See,
for instance, his appeal in the Défense to the principle, Universalia sunt tantum in mente
Platonism and Conceptualism among the Cartesians  •  129

from our perceptual modifications, or whether we perceive them rather by means


of these modifications themselves.
The debate over ideas may seem to be a battle for the soul of Descartes. Thus,
Arnauld opens the debate by citing Descartes explicitly in support of the posi-
tion that our idea of an external object is simply a perceptual modification that
represents—​or is “of ”—​that object (OA 38:198–​99, 205–​06). In response,
Malebranche initially insists that Descartes himself did not claim that “the
modalities of the soul are essentially representative” (OCM 6:172). In a later
response, he concedes that Descartes did say that “ideas are modalities of minds,”
though he adds that Descartes said this only because “unlike me, he does not take
the word ‘idea’ to signify exclusively the ‘representative reality’, but for those sorts
of thoughts by which one perceives a man, an angel, etc.” (OCM 6:217).
Given Arnauld’s considerable intellect and intimate knowledge of Descartes’s
system, there was little chance that Malebranche would show him up in Descartes
exegesis. Indeed, the clear indication in Descartes is that the objective reality of
a perception of an external object is simply the internal “form” of the percep-
tion that serves to relate it to that object.20 Yet it must be said that Malebranche
was not overly concerned to connect his account of ideas to Descartes’s writings.
Rather, as I have indicated, the main source for Malebranche’s account of ideas
was Augustine. Thus, in responding to Arnauld’s charge that his account is dan-
gerously novel, Malebranche insists that “it is principally [Augustine’s] authority
which has given me the desire to put forth the new philosophy of ideas” (OCM
6:80). It was also the authority of Augustine that gave Malebranche the desire to
reject Descartes’s created truths doctrine.
In order to counter Malebranche’s appeal to the authority of Augustine
in support of his doctrine of the Vision in God, Arnauld argues in his initial
exchanges with Malebranche that one who reads Augustine in light of the writ-
ings of Thomas Aquinas will attribute to the former the position that we perceive
external objects by means of our own ideas, and not God’s. On the view—​central
for Malebranche’s Augustinian Platonism—​that the eternal truths and immu-
table natures we know are themselves eternal and immutable, Arnauld is simply
silent.
The silence on the Cartesian created truths doctrine continues in Arnauld’s
later writings. However, in some of these writings he begins to address other
questions concerning the ontological status of these truths. Though the writings

(OA 38:394). However, he stops short of saying that the truths and essences we know by means
of universal ideas are only features of our mind.
20. See, for instance, Descartes remarks on objective reality at AT 7:102–​03, cited by Arnauld
at OA 38:200.
1 3 0   •  The Problem of Univer sals in Early Modern philosophy

I have in mind—​from the 1690s—​are not direct responses to Malebranche, they


do more than Arnauld’s initial critique of Malebranche to provide a systematic
alternative to Malebranche’s Platonism.
The first text—​the 1692 Dissertatio bipartite—​is a response to De veritate
aeterna, sapientia et justitia aeterna, a thesis of Henri van den Sanden defending
the Augustinianism of his teacher, the Louvain theologian Gommaire Huygens.21
One of the claims in this text is that anyone who has knowledge of necessary and
immutable truth sees that truth “in the first and uncreated truth, which is God”
(Arnauld 2001, 49). No doubt reminded of Malebranche’s doctrine of the Vision
in God, Arnauld found it necessary to respond. In his response, Arnauld appeals
to the view in Thomas’s Summa Theologiae22 that truth exists in the understand-
ing alone, and that truth does not reside in a created understanding in the same
way in which it resides in the divine mind. Arnauld draws from this view the
consequence that “created truth”—​that is, the truth of propositions as they
exist in a created understanding—​cannot be called eternal “properly speaking,”
because this sort of eternality pertains to God alone. Rather, the truth as present
in such an understanding can be called eternal only in the “improper sense” of
“not depending on a determined place and time.” Here, created truth is compared
to a universal, which according to Thomas is “always and everywhere, because the
universal abstracts from the here and now,” but is not “eternal unless in an under-
standing, if there is one, that is eternal” (Arnauld 2001, 54–​55). Since the truths
that we know are not eternal, there is no need to place them in the divine mind.
One might think that Arnauld’s talk of “created truth” in us is a veiled ref-
erence to the Cartesian created truths doctrine.23 However, the most immedi-
ate source is Thomas, who does not link this kind of talk to anything like this
Cartesian doctrine. Indeed, in the passages Arnauld discusses in the Dissertatio,
Thomas makes clear that eternal truth in God is identical to God’s uncreated
nature, and nothing Arnauld says in this text seems to me to conflict with this
Thomistic position. There is no clear opposition here to Malebranche’s claim
against Descartes that the eternal truths are uncreated.24 Where Arnauld differs
from Malebranche, as well as from Huygens, is in embracing the consequence of

21.  On Huygens, see Ceyssens 1974, 52–​53. For the background to the dispute between
Huygens and Arnauld, see Moreau 1999, chap. 6.
22. More specifically, Summa Theologiae [= Aquinas 1947] Ia, q. 16, art. 1, 6, 7, and 8.
23. As claimed, for instance, in Moreau 1999, 173. Cf. the claim in Faye 2005—​in line with
my view here—​that the account of created truth in the Dissertatio is Thomistic rather than
Cartesian.
24. So also in the Règles du bons sens, to be discussed presently, Arnauld refers without chal-
lenge to la vérité souveraine et incréée in God; see, e.g., Arnauld 2001, 97,
Platonism and Conceptualism among the Cartesians  •  131

his conceptualism that the reality that grounds the universal truths we know is
simply a feature of our minds, and thus properly speaking is not eternal.
Arnauld’s friend Pierre Nicole was troubled by the apparent conflict of
the Dissertatio with Augustinian orthodoxy, and he called on the Benedictine
François Lamy to respond to Arnauld.25 Lamy’s response—​now lost—​prompted
Arnauld to further defend his conceptualism in the 1693 Règles du bons sens. It is
clear from Arnauld’s remarks in the Règles that Lamy adhered to the Augustinian
line that “we see necessary and immutable truths in sovereign and uncreated
truth” (Arnauld 2001, 97). Even more clearly than in the Dissertatio, Arnauld
distances himself from Augustine in his 1693 text, claiming there: “I abandon
St. Augustine to follow St. Thomas, thus preferring the opinion of the Disciple
to that of the Master” (Arnauld 2001, 98).26 Though Augustine continued to be
an authority for him in theological matters, Arnauld saw no need to follow his
philosophical views on our knowledge of eternal truths, which he took to be sim-
ply borrowed from “the philosophy of Plato” (Arnauld 2001, 99). In opposition
to Augustine’s Platonism, Arnauld insists again on the conceptualist position in
Thomas that universal reality is simply a created feature of our mind, and thus
cannot be identified with uncreated and eternal features of the divine mind.
In the 1702 edition of the Conversations chrétiennes, Malebranche argues
against this conceptualist alternative to his own view by appealing to the fact that
our true perceptions must correspond to a reality distinct from our own mind. In
this text, he has his representative Theodore claim:

[I]‌t is evident that nothingness is not visible. From which it is easy to con-
clude not only that all that the mind perceives immediately and distinctly
is true, but also that it is always such as it is perceived, in the sense that it is
perceived… . [T]here is a contradiction that creatures are not conformed
to the idea of the Creator, the eternal model on which he has made them.
To ruin Pyrrhonism therefore there is no more to be proved than that the
idea we have of extension from which all bodies are formed is the same
that God has of it. (OCM 4:72)

Thus, Malebranche would endorse Descartes’s claim in the Fifth Meditation


that all that we clearly and distinctly perceive to be true is something, though he
also would require that the “something” here be external to our mind. Indeed,

25. On Lamy and his relation to Arnauld, see Zehnder 1944, 91–​94.


26. There is certainly a shift from Arnauld’s 1641 comments on Descartes’s Meditations, which
emphasize the affinities of Descartes’s views with the teaching of Augustine; see AT 7:197, 205.
1 3 2   •  The Problem of Univer sals in Early Modern philosophy

Malebranche’s view is we can overturn Pyrrhonian skepticism only by identifying


this something with the idea in God that serves as the archetype for his creation
of the world. In line with Descartes’s own claim that truths concerning triangles
hold even if no triangle actually exists external to mind, Malebranche allows
that our perceptions of eternal truths concerning extension do not require the
actual existence of the created ideatum of the idea of extension. However, what
Malebranche must reject is the implication of Arnauld’s conceptualism that what
we know is a created truth that is distinct from the eternal and uncreated truth
in the divine mind.
We have noted the apparent tension in Descartes between the claim that eter-
nal truths can be characterized in terms of something similar to divine eternity,
on the one hand, and the claim that these truths are creatures distinct from the
divine essence, on the other. We also have seen that Malebranche rejected the
latter claim in favor of the former one in a manner incompatible with Descartes’s
created truths doctrine. What we find in Arnauld’s 1690 writings is a clear rejec-
tion of the former claim in favor of the latter one, though no explicit stand on
the Cartesian doctrine. What we will discover in Regis, however, is an attempt
to accommodate both claims by appealing to this Cartesian doctrine in a manner
that deviates from Arnauld’s strict conceptualism.

6.5. Regis’s Non-​Platonic Anti-​Conceptualism


Regis was a popularizer of Cartesian physics whom the virulently anti-​Cartesian
Pierre-​Daniel Huet called “the Prince of the Cartesians” (for Huet, not a com-
pliment).27 More than a decade after the initial exchange between Malebranche
and Arnauld over the nature of ideas, Regis entered into this controversy in his
Système de philosophie (1691) by defending Arnauld’s side in this debate. Thus,
Regis argues there for the conclusion that we see bodies not by means of a “union”
with features of God that serve to represent them, but rather “by means of ideas
that are in us, and that depend on the bodies that they represent” (SP 1:188).
Moreover, he insists there on the distinction between the finite “formal being” of
our idea of God, on the one hand, and the infinity this idea possesses “according
to the property that is has of representing its object,” on the other (SP1:194–​
95). Arnauld weighed in on Regis’s side, noting to Malebranche that “you are not
happy with this distinction; too bad for you” (OA 40:88–​89).
With respect to the issue of the relation between ideas and perceptions, Regis’s
contributions seem to offer merely a footnote to the earlier dispute between

27. For this and more on the Huet-​Regis connection, see Schmaltz 2002, chap. 5.
Platonism and Conceptualism among the Cartesians  •  133

Malebranche and Arnauld. In his 1693 Réponse à Regis, however, Malebranche


draws attention to aspects of Regis’s position that have no clear counterpart in
Arnauld. In particular, in a passage I cited previously, Malebranche contrasts his
own conclusion that eternal truths are “eternal, independent and preliminary to
the free decrees of God” with the claim in Regis that “they consist in the sub-
stances that God has created, being that the soul considers these substances in a
certain manner” (OCM 17[1]‌:308–​09). Such a claim is found not in the chapter
in the Système that concerns Malebranche’s doctrine of the Vision in God, but
rather in an earlier chapter on the nature of “the truths that one calls eternal.”
In that chapter, Regis distinguishes between the matter of the eternal truths and
their form. He holds that the matter “consists in substances and modes,” whereas
the form consists “in the action by which the soul considers substance and modes
in a certain manner.” For instance, external extended substance provides the mat-
ter for eternal truths concerning triangles, whereas the action internal to our
mind by which we consider this substance provides their form (SP 1:178). Regis
concludes that this account of eternal truths “is very different from that of certain
philosophers who believe that we see these truths in God, because all souls con-
ceive them in the same way” (SP 1:180).
Regis’s explanation of eternal truths in terms of the form “by which the
mind considers substance and modes in a certain manner” may seem to be in
line with Arnauld’s conceptualist view that “created truth” exists in our mind.
The reading of Regis as a conceptualist seems only to be reinforced by his later
remarks in the 1704 Usage de la raison et de la foi. For in this text he explicitly
endorses the “famous maxim,” Universalia sunt tantum in mente: universals are
only in the mind. Since all that exists is particular, universals can exist only in
the mind, by means of abstraction. Thus, the universal human nature that com-
prises Pierre, Paul, Jean, and Jacques is merely a mental entity, and not a feature
of the created world (URF 217). We have seen the argument in Malebranche
that universal reality can exist only in God. Regis here would seem to endorse
the conceptualist response in Arnauld that this reality can exist also in our
created mind.
However, in both of Regis’s works there are elements that go beyond what is
found in Arnauld’s own form of conceptualism. Thus, in the Système Regis draws
attention to the fact that the matter for eternal truths consists in substances and
modes external to mind. Moreover, in the Usage his conceptualist remarks are
followed by the claim that “all human certainty is founded on ideas depending on
their objects as their exemplary causes” (URF 235). Eternal truths are ultimately
grounded not in our ideas, but rather in the “objects” that serve as their exem-
plary causes. Thus, for Regis, the reality that grounds the eternal truths cannot
consist simply in features of our own mind.
1 3 4   •  The Problem of Univer sals in Early Modern philosophy

To understand why this further appeal to the external objects of our ideas
is required, we need to consider the distinctive version of the Cartesian created
truths doctrine in the work of the Lorraine Cartesian Robert Desgabets (1610–​
1678), whom Regis called “one of the greatest metaphysicians of our century”
(URF 639).28 This version emerges in the course of the dispute between Foucher
and Malebranche that I mentioned earlier.29 Before he had read Malebranche’s
response to Foucher, Desgabets wrote a response on Malebranche’s behalf,
the Critique de la critique. In this text, Desgabets assumes with Foucher that
Malebranche accepts the Cartesian created truths doctrine, and he takes himself
to be defending Malebranche’s purported acceptance of this doctrine in citing
the “great principle of saint Augustine: The nature of a thing is one and the same
as the will of God” (CdC 74). This appeal to Augustine is particularly unfortunate
since, as we have seen, Malebranche later invoked Augustine precisely to under-
mine the Cartesian created truths doctrine.30 Yet Desgabets’s appeal to Augustine
is also misleading insofar as it suggests that Desgabets himself takes the immuta-
bility of eternal truths to be grounded directly in the immutability of the divine
will. Indeed, in his Critique Desgabets makes clear that the immutability of these
truths has a different proximate source. Thus, he there responds to Foucher’s
objection that the divine will cannot guarantee the immutability of eternal truths
by claiming that “the immutability of essences and the necessity of truths does
not come precisely from the immutability of the divine decree, but rather the
immutability of all these things comes from the indivisibility of their existence,
which has no extension” (CdC 84). For Desgabets, then, the immutability of cre-
ated eternal truths derives most directly from the fact that they are grounded in
created objects that have an “indivisible existence.”
In his Critique, Desgabets suggests at one point that truths are to be identi-
fied with “the manner in which the understanding is determined to know things,
whether we think of them actually or simply have the power to think of them”
(CdC 75). Taken in isolation, this sort of comment may seem to indicate a con-
siderable sympathy for conceptualism. However, Desgabets precludes a concep-
tualist interpretation of his account of eternal truths when he warns following

28. Desgabets and Regis were among the few Cartesians who explicitly accepted Descartes’s
created truths doctrine. On the reception of this doctrine among the Cartesians, see
Rodis-​Lewis 1981.
29. See the discussion above at note 12.
30. Malebranche in fact distanced himself from Desgabets: In an “avertissement” appended
to the second and third editions (published in 1676 and 1678, respectively) of the second vol-
ume of the Recherche, Malebranche notes with respect to Desgabets’s Critique that though one
might believe “that I had some part in his work, I believe that I must say that although I am
very satisfied with his person, I am not extremely content with his book” (OCM 2:500).
Platonism and Conceptualism among the Cartesians  •  135

this comment that “it is necessary to add the correspondence that the object has
with the thought and the thought with the object” (CdC 75). And elsewhere he
claims that eternal truths in our mind “are true only by the relation of conformity
that it has to the thing that [is as] it is really and immutably” (RD 6:232).
In his earlier and unpublished Traité de l’indéfectibilité des creatures, Desgabets
derives from the fact that our true ideas correspond to their objects the strik-
ing conclusion that “the clear and distinct knowledge that one has of substance
necessarily supposes its actual existence” (RD 3:76). In line with Descartes’s
own remarks in the Fifth Meditation, Desgabets admits that our conception of
modes—​such as triangles—​does not require the actual existence of the modes,
since the modes need only be “contained in a pure power” of the substance that
can possess them—​such as extended substance, in the case of triangles (RD 3:76–​
77). In the Fifth Meditation, the fact that there need be no real triangles in order
for us to conceive them seems to show equally that there need be no extended
substance in order for us to conceive it. But Desgabets’s argument cannot be gen-
eralized to extended substance in this way; if this substance does not exist, then
neither that substance nor its modes can be conceived to be possible. Desgabets
in fact inveighed against “pure possibilities” that are not rooted in any existing
substance. For him, the mere fact that we can conceive possibilities concerning
triangles reveals the existence of extended substance.
As the remarks in his Critique make clear, moreover, Desgabets requires not
only that extended substance exist, but also that this substance have an “indivisible
existence.” For this existence serves to guarantee the immutability of the neces-
sary truths we know concerning extension. In the Traité, Desgabets distinguishes
the existence of “substantial and permanent beings,” which has “no relation to
time,” from the existence of “successive and modal beings,” which is divisible into
temporal parts (RD 2:34–​35). Immutable essences are to be identified with these
substantial and permanent beings:

[E]‌ssence will be nothing other than substance considered in itself accord-


ing to its intrinsic and essential attributes without relation to time or to
other creatures that one does not consider when one regards substance in
this state of abstraction. … Thus matter considered in itself will be the
essence of corporeal things, which will receive its existence when it will be
clothed in its modes, which give it a particular and determinate manner
of being. (RD 2:27)

Thus, for instance, God creates the immutable essence of matter by creating
a material substance that “considered in itself ” has no relation to time, and
so is not subject to temporal change. In Desgabets’s terms, this substance is
1 3 6   •  The Problem of Univer sals in Early Modern philosophy

“indefectible,” and in virtue of being so, truths concerning it are necessary and
immutable.
We have seen that Malebranche insisted, in opposition to Arnauld’s concep-
tualism, that in order for our conception of necessary and universal features of
created extension to be true, it must conform to an uncreated and eternal essence
external to our mind. However, he identified this essence with God’s own uncre-
ated and eternal idea of extension, and not with that substance as it exists in the
created world. Desgabets’s commitment to the Cartesian created truths doctrine
was too deep to allow him to accept this sort of identification. But instead of
adopting the conceptualist position that this essence exists merely in our mind,
and thus is not eternal in a strict sense, he claimed that it is to be identified rather
with an indefectible feature of created reality.
We find in Regis the same appeal to indefectible substances and the same
rejection of pure possibilities that are not grounded in such substances.31 Thus
Regis notes in the Système:

If I want to go back as far as the origin of the possibility and impossibil-


ity of modal things, I would perceive evidently that God is the sole and
unique cause, and that he has produced them by the same action by which
he has created body and mind capable or incapable of receiving certain
modes. (SP 1:103)

God creates the possibility and impossibility of modal things—​or what Regis
calls “modal essences”—​by creating substances that can be modified in certain
ways and cannot be modified in other ways. These possibilities and impossibilities
are stable because the substances themselves are “indefectible” in a manner that
precludes change (SP 1:101).
In his Usage, Regis makes clear his view—​in line with Desgabets’s earlier
position—​that truths concerning possible modes of created substances require
the actual existence of those substances:

The idea of modes can represent them as possible, since modes, even when
they do not exist, are contained in the power of substances; on the con-
trary the idea of substances can never represent them as possible, and it
always represents them as actually existing; the reason for which is that
substances being as if drawn from nothing by creation, they can be con-
tained in the power of no subject. (URF 259)

31. For a more detailed consideration of Desgabets’s influence on Regis, see Schmaltz 2002.


Platonism and Conceptualism among the Cartesians  •  137

Even in the case where a particular triangular shape does not exist as a mode of
something external to mind, it is possible in virtue of the fact that there is an
extended substance that can take on that mode. But the possibility of extended
substance cannot be conceived in the same way in terms of the existence of some-
thing else, since substances, unlike modes, do not exist in a subject. Whereas ideas
of modes can represent merely possible objects, ideas of substances must repre-
sent actual objects in order to represent anything at all.
I have noted the claim in Desgabets that indefectible substances have an atem-
poral existence that ensures the immutability of the necessary truths concerning it.
Similarly, Regis denies that such substances have a temporal duration. He empha-
sizes in particular that whereas modal beings have a duration that is divisible into
parts, the existence of both modal essences and the substances that ground them
is “simple and indivisible,” and thus involves no succession or change. As Regis
recognizes, this feature of his system places him in opposition to the claim of “a
very considerable philosopher”—​indicated in the margin to be Descartes—​that
created substances require a power that conserves them in existence since they
have a duration that is divisible into independent parts (URF 321–​22). This claim
of course recalls Descartes’s insistence in the Third Meditation that his duration
is divisible into various distinct parts, and thus that he requires something that
conserves him in existence at each moment (AT 7:48–​49). Regis’s response to
Descartes is that since

the existence of substances is simple and indivisible, it is not necessary


to seek successive conservation, but simple creation the entire action of
which is contained in an indivisible point; which properly establishes the
indefectibility of created substance. (URF 325)

Here the relation between substance and its modes is no longer, as it is in


Descartes, a relation between a temporal subject and its varying modifications.
Rather, as in Desgabets, it is the relation between an atemporal essence and its
temporal instantiations. This implication helps to explain Regis’s identification—​
found also in Desgabets—​of “the essence of modal things” with “the substance
itself, insofar as it is capable of receiving certain modes, without any relation to
time and place, which one never considers when one regards things with a sim-
ple degree of essence” (URF 263). For Desgabets and Regis alike, the true and
immutable essence of a triangle is nothing other than an indefectible extended
substance that exists “without relation to time and place,” and so is not subject
to change.
To be sure, Regis is more concerned than Desgabets to deny that created sub-
stances are eternal in the same sense that God is. For in the Systéme de philosophie,
1 3 8   •  The Problem of Univer sals in Early Modern philosophy

he refers to the perpetuité of the existence of such substances, whereas in the Usage
he speaks of their eviternité.32 However, his consistent position is that the exis-
tence of created substances is similar to the existence of God in not being subject
to temporal duration. What distinguishes the existence of God is simply the fact
that it is absolutely necessary and independent of any external cause, in contrast
to the existence of created substances, which is only hypothetically necessary and
dependent on God’s indifferent will.
On the view that Regis inherited from Desgabets, then, the relation of modes
to the substance in which they inhere consists in the relation of temporal instan-
tiations of an essence to its atemporal ground. In some respects, this relation may
seem to be similar to the relation of particulars to the universal Platonic form
in which they participate. However, we have seen the conceptualist insistence
in Regis that universals can exist only in the mind. The atemporal essence that
exists external to mind must itself be a particular, and so distinct from a universal
Platonic form. Thus, the atemporal essence of extension just is the one extended
substance, and the atemporal essences of the thought that constitutes different
minds just are those individual mental substances.33
Regis’s account of the substance-​mode relation is certainly unusual from a
more orthodox Cartesian perspective. But given Descartes’s own remarks con-
cerning true and immutable natures, one advantage of this account is that it pro-
vides an immutable and eternal foundation for created mathematical truths. The
mathematical essences comprehended in extended substance are not subject to
change since the substance has an atemporal existence. As in the case of truths
concerning God in the Fifth Meditation, moreover, the necessity of truths con-
cerning extended substance is not something that our mind imposes on reality,
but rather is imposed on our mind by something external to it. Nevertheless, this
result is consistent with Descartes’s created truths doctrine since it allows for the
fact that truths concerning extended substance depend on God’s indifferent will
insofar as the existence of this substance so depends.
Regis takes his account of truths and essences to position him in the old debate
between Platonists and Aristotelians. In the Usage, he distinguishes between
“the disciples of Plato” who hold that God can know creatures prior to seeing

32. Cf. SP 1:108, and URF 345.


33. But cf. Lennon’s claim that Regis “is remarkably explicit in treating individual minds as
modes of a single thinking substance” (Lennon 1994, 26). However, at one point Regis explic-
itly distances himself from the Spinozistic view that all minds are modes of a single thinking
substance. In an appendix to his Usage devoted to a “refutation” of Spinoza, Regis insists that
whereas thought is unified generically, it is distinguished into different concrete attributes that
constitute different individual minds (URF 906).
Platonism and Conceptualism among the Cartesians  •  139

them in his essence, on the one hand, and “the disciples of Aristotle”—​identified
with scholastics such as Thomas Aquinas—​who claim that it is due to God’s will
that creatures participate in his essence in a particular manner, on the other. In
response to the Platonists, Regis claims:

There is nothing that accords less with the notion that we have of God as
a perfect thought than the manner of knowing things that the disciples of
Plato attribute to him. I continue to agree completely with them that God
can learn nothing about creatures, but it does not follow that God must
know creatures before producing them; because it has been proved that
in God, knowledge does not precede will. In effect, how can God know
things before willing to produce them; and if those that he does not pro-
duce according to their essence or their existence do not differ from pure
nothingness? (URF 209–​10)

Here the rejection of Platonism consists in the denial that necessary truths con-
cerning creatures and creaturely essences are grounded in features of God’s essence
that precede his act of creation. Regis’s remarks reflect Descartes’s own view that
it follows from the indifference of the divine will that prior to the act of creation,
there are no possibilities or impossibilities concerning creatures.34
Though Regis is more sympathetic to the Aristotelian position of the scholas-
tics, he nonetheless disputes the scholastic view that creatures can participate in
God’s essence. He takes Descartes’s created truths doctrine to show that

God does not see creatures in his perfections, because it has been proved
that the perfections of God have nothing in common with creatures, and
by consequence they cannot represent them; we must say only that God
sees creatures in his will, insofar as it is by his decree that he produces them
and conserves them. (URF 169)

There is an obvious disagreement here with Malebranche, who identified the


essences of creatures with eternal archetypes in the divine intellect. But Regis
seems to disagree as well with Descartes himself. For the latter was led by his
axiom that the “reality or perfection” of the effect is contained in its cause “for-
mally or eminently” to conclude that God “eminently” contains creatures in his

34. Though Descartes suggests at time that his created truths doctrine includes all truths, and
thus also those concerning God himself, there are good reasons for him to restrict the scope
of this doctrine—​as Desgabets and Regis do explicitly—​to truths concerning creatures. For a
defense of this position, see Schmaltz 2011.
1 4 0   •  The Problem of Univer sals in Early Modern philosophy

infinite perfection.35 It is interesting that Malebranche also insisted on the emi-


nent containment of extension in God when responding to the objection that his
view of intelligible extension has the implication that God is formally extended.36
But for Regis it is an implication of the Cartesian created truths doctrine that
God is merely an “analogous cause” that contains his effects neither formally nor
eminently (URF 406–​07).
Regis does take seriously the objection in Malebranche and Lelevel that this
Cartesian doctrine yields a kind of Pyrrhonian skepticism, given that “all would
be reversed in the natural sciences” that depend on created truths. His response in
the Usage is that the mathematical sciences cannot be reversed in this way because
the “objects of these sciences are immutable and necessary with a hypothetical
necessity and immutability, that is to say, with a necessity and immutability
depending on the will of God” (URF 277). This hypothetically necessary object
can be found neither in our mind nor in God’s; rather, it can be only the created
but atemporal extended substance that imposes its necessity on our thought. We
have here an alternative not only to Malebranche’s Platonism, but also to a con-
ceptualism in Arnauld that precludes the view that the essences and truths we
know are atemporal in any literal sense.

6.6. Conclusion
I have noted two seemingly conflicting claims that emerge from Descartes’s dis-
cussion of immutable natures and eternal truths:  (1)  Such natures and truths
are eternal in a manner that distinguishes them from temporal creatures;
and (2)  Such natures and truths are nonetheless creatures distinct from God.
Malebranche’s Platonic solution to this tension involves the complete rejection
of the latter claim, which he takes to derive from a created truths doctrine that
results in skepticism. According to Malebranche, we can overcome this sort of
skepticism by identifying mathematical natures with the uncreated and eternal
ideas in God that serve as the archetypes for his creation of the world. On this
solution, the mathematical natures are aspects of reality external to our mind that

35. Descartes suggests that bodies are eminently contained in God, for instance, in his proof
in the Sixth Meditation of the existence of the material world; see AT 7:79. In the Second
Replies, Descartes defines formal containment as “whatever is in itself such as we perceive,” and
eminent containment as “whatever is not such [as we perceive], but greater, such that it can
take the place [of what we perceive]” (AT 7:161). These definitions are less than clear, to say the
least, but I try to make some sense of them in Schmaltz 2008, 64–​71.
36. For this objection, see Arnauld’s remarks in his 1683 Vraies et fausses idées, at OA 38:246–​
47, and the remarks of Dortous de Mairan in 1714 correspondence with Malebranche, at OCM
19:904–​906. For Malebranche’s responses, see OCM 6:118–​19 and 19:909–​10, respectively.
Platonism and Conceptualism among the Cartesians  •  141

impose their necessity on us, and the truths deriving from these natures share in
the eternity and immutability of God.
Arnauld sought to free our thoughts from God’s, and so was concerned
to deny that the natures and truths that we know could be found in God. But
since he held that the only eternity is divine eternity, he was led to deny that the
natures and truths that we know are eternal in any straightforward way. Having
identified these truths and natures with features of our mind, however, he could
affirm that they are creatures distinct from God. Whereas Malebranche sacrificed
Descartes’s second claim to save his first, Arnauld sacrificed Descartes’s first claim
to save his second.
For Regis, however, there was no need to sacrifice either claim. The Cartesian
created truths doctrine—​which is absent from Arnauld’s discussion—​is front
and center in Regis as the source of his acceptance of Descartes’s second claim.
Yet Regis also saw a need to provide an atemporal and, therefore, immutable
ground for mathematical truths concerning the external world. This ground he
took to be provided not by the form of the truths in our temporal mind, but
rather by the matter of the truths that he identified with the atemporal extended
substance that particular temporal bodies modify. Despite his clear disagree-
ments with Malebranche’s form of Cartesian Platonism, then, Regis stood with
his fellow Cartesian in opposing the sort of conceptualist account of immutable
natures and eternal truths that Arnauld embraced, and that some commentators
continue to see as the most viable option for Descartes.

Abbreviations
AT = Descartes 1964–​74; cited by volume and page.
CdC = Desgabets 1675; cited by page.
CRV = Foucher 1969; cited by page.
OA = Arnauld 1967; cited by volume and page
OCM = Malebranche 1958–​67; cited by volume and page.
RD = Desgabets 1983–​85; cited by volume and page.
SP = Regis 1970; cited by volume and page.
URF = Regis 1996; cited by page.
VFM = Lelevel 1694; cited by volume and page.
7 UNIVERSALS AND INDIVIDUALS
IN MALEBRANCHE’S PHILOSOPHY

Mariangela Priarolo

In early modern philosophy the problem of universals, far from disap-


pearing, became central to philosophical debate, but with a very impor-
tant change. In fact, although in the scholastic debate the point at issue
was above all the ontological status of universals, in the modern age phi-
losophers seem more interested in epistemological than in metaphysical
questions. It is known that, from Descartes to Hume, the discussion on
the nature of universals was strictly related to research on the modalities
of human knowledge, in particular the definition of ideas. In Descartes,
for instance, it is only through the ideas—​which represent the universal
essences of bodies, i.e., innate ideas—​that we can get true knowledge of
the mathematical structure of the world and then build an a priori phys-
ics; by contrast, the particular and individual ideas reached by the senses,
that is, adventice ideas, do not possess any epistemic content.1
Indeed, in Malebranche’s philosophy the connection between
ideas and universals becomes much stronger than in Descartes, since
Malebranche, at least in his mature writings, thinks that all ideas,

1.  The shift from the ontological to the epistemological treatment of universals
in Descartes is attested, I think, by the difficulty in giving a definitive answer as to
Descartes’s opinion of the status of universals, as shown by the wide range of inter-
pretations on this subject that we can find in the literature. Some scholars think that
Descartes is a Platonist (e.g., Kenny 1968), others a conceptualist (e.g., Chappell
1997), others a nominalist (e.g., Cunning 2003). The problem arises from the differ-
ence between some places in the Meditations on First Philosophy in which Descartes
defines universal essences represented by ideas as “true and immutable natures”
(Fifth Meditation, AT 7:64), and two articles in the Principles of Philosophy in which
Descartes says that “number, when it is considered simply in the abstract or in general,
and not in any created things, is merely a mode of thinking; and the same applies to all
the other universals, as we call them. … These universals arise solely from the fact that
we make use of one and the same idea for thinking of all individual items which resem-
ble each other: we apply one and the same term [nomen] to all the things which are
represented by the idea in question, and this is the universal term” (1:58, 59, AT 8:27).
Universals and Individuals in Malebranche’s Philosophy  •  143

properly speaking, are universal and infinite. This is one of the main reasons pro-
vided by Malebranche to support the thesis according to which man cannot own
ideas, but can only attain God’s ones—​this is the core of Malebranche’s theory of
the “vision in God of ideas.” This is because no finite and particular idea is able
to represent something that is universal and infinite.2 According to Malebranche,
since science—​that is, a true, universal, and necessary knowledge of the world—​
exists and, as Descartes showed, we attain knowledge of the world through ideas,
then these ideas can belong to the only reason that satisfies the requirements of
truthfulness, universality, and necessity:  the reason of God.3 But God’s ideas,
which are the archetypes of the world, possess the same features of God’s rea-
son:  universality, infinity, and necessity. Hence, the ideas attained by man are
the same ideas of God, and the certainty of our knowledge—​the science we can
have about the world—​depends on these ideas being the model that God used to
create the world. As we will see, by drawing on a medieval model of knowledge
that has been identified in Aquinas’s beatific vision,4 Malebranche built the new
science on divine grounds. But a question then arises: if the ideas through which
we know the world are universal, general, and infinite, how is it possible to know
a single and particular body—​for instance, the sun?
In the following, we try to answer this question by analyzing the relationships
between Malebranche’s vision in God and its medieval source, in order not only to
clarify Malebranche’s conception of the universal but also to show that this scho-
lastic source can be very helpful for understanding the relationships between the
conception of universals and knowledge of individuals in Malebranche. This road
will show that, in contrast to his contemporaries, for Malebranche the concep-
tion of universal cannot be separated at all from its metaphysical background—​
an impossibility that seems to lend further evidence to Paul Hazard’s suggestive
description of Malebranche: “Il eût été capable d’inventer la métaphysique si elle
n’eût existé avant lui” (Hazard 1995, 93).

7.1
As previously mentioned, since Malebranche placed a strict connection between
universals and ideas, in order to understand his view about universals we must plainly

2. The emphasis on the “and” is not casual because, as we will see, Malebranche, in contrast with
Descartes, saw a strict connection between the two features of ideas, giving to both of them
the same, positive, value.
3. See especially Tenth Elucidation of The Search after Truth, OC 3: 128–​31.
4. See Scribano 1996 and Priarolo 2004.
1 4 4   •  The Problem of Univer sals in Early Modern philosophy

turn to his definition of idea. The better place to find Malebranche’s opinion on this
subject is the book explicitly devoted to the theory of knowledge: La Recherche de la
verité. Written between 1668 and 1673—​after four years in which Malebranche had
studied Descartes and the new scientific discoveries—​the Search after Truth should
be considered a “work in progress.” Seven different editions were in fact published
during Malebranche’s life: the first volume appears in 1674, followed the next year by
a second one and by the second edition of the first volume. Two new editions were
published in 1678 when the Elucidations, the third volume of Search, appeared; the
fifth edition in 1700, at the end of the querelle des idées with Arnauld;5 and the last
one in 1714, one year before Malebranche’s death. The fundamental critical edition
directed by André Robinet allows us now to follow the evolution of Malebranche’s
theory of knowledge, an evolution that some scholars, as Robinet himself, see more
as a series of radical changes than a progressive adjustment of the theory.
In the introduction, Malebranche—​who, like Descartes, thinks that man is
composed of two distinct realities: body and soul—​states that man’s soul is char-
acterized by a “double union,” one with his body, another with God:

The mind of man is by nature situated, as it were, between its Creator and
corporeal creatures, for, according to Saint Augustine, there is nothing but
God above it and nothing but bodies below it. But as the mind’s position
above all material things does not prevent it from being joined to them,
and even depending in a way on a part of matter, so the infinite distance
between the sovereign Being and the mind of man does not prevent it
from being immediately joined to it in a very intimate way. (SAT Preface
xxxiii, OC 1:9)

While the union with the body is the main source of our errors—​since the
two faculties of the soul that are strictly related to the body, i.e., sensibility and
imagination, are given to us not for knowing the world but only for preserving
ourselves—​they have, in current words, an adaptive function—​union with God
provides us with the truth. In this sense, Malebranche explicitly refers to Saint
Augustine’s doctrine of divine illumination, which states that the truth attained
by man derives from the connection between our intellect and God.6 It is impor-
tant to note that Malebranche’s sympathy toward Augustine’s doctrine of divine

5. On this debate, see Moreau 1999 and Nadler 1989.


6. See SAT xxxvii, OC 1:16–​17: “This is what Saint Augustine teaches us with these elegant
words. ‘Eternal wisdom’ he says, ‘is the source of all creatures capable of understanding, and
this immutable wisdom never ceases speaking to His creatures in the most secret recesses of
Universals and Individuals in Malebranche’s Philosophy  •  145

illumination depends on a thesis that plays a pivotal role in his philosophy: the


radical impotence of creatures. This thesis is one of the bases of so-​called occasion-
alism, the theory shared by Malebranche according to which the only real and
true cause is God, whereas the actions of finite beings are just the occasions for
God to carry out his decrees.7 As we will see, it is precisely by appealing to the the-
sis of the structural impotence of creatures that Malebranche in the third book of
the Search—​where he discusses different theories of ideas—​refuses every theory
that considers ideas as mental modifications. This is an essential step for affirming
the vision in God of ideas. Malebranche’s argument is as follows.
Malebranche first states that we know the external world through ideas, by
invoking the consensus omnium:

I think everyone agrees that we do not perceive objects external to us by


themselves. We see the sun, the stars, and an infinity of objects external to
us; and it is not likely that the soul should leave the body to stroll about
the heavens, as it were, in order to behold all these objects. Thus, it does
not see them by themselves, and our mind’s immediate object when it
sees the sun, for example, is not the sun, but something that is intimately
joined to our soul, and this is what I call an idea. Thus, by the word idea,
I mean here nothing other than the immediate object, or the object closest
to the mind, when it perceives something. (SAT 217, OC 1:413–​14)

Since the objects of knowledge, ideas, have many properties, but “nothing has
no properties,” we must conclude that ideas not only are “something,” but are
real in a strong sense; they are more real than external bodies that could not
exist even if we perceived them, since what we perceive is the idea of a body, not
the body itself: “It often happens that we perceive things which do not exist,
and that even have never existed” (SAT 217, OC 1:414).8 But to know the func-
tion of the ideas, that is, their being the primary object of perception, and to
state that they possess some kind of reality is not sufficient to understand what
an idea actually is. According to Malebranche, in order to fully grasp the role of
ideas in human knowledge, we shall need to detect their origin. There are five

their reason so that they might be inclined toward Him, their source, because only the vision
of eternal wisdom gives minds being, only eternal wisdom can complete them, so to speak, and
give them the ultimate perfection of which they are capable.’ ”
7. See Nadler 2011.
8. In his late writings, Malebranche will say that, for the above reasons, even if God would
destroy the material world, we could continue to perceive the same things we perceive now.
See, for instance, Entretiens d’un philosophe chrétien et un philosophe chinois, OC 15.
1 4 6   •  The Problem of Univer sals in Early Modern philosophy

different possible answers to this question, five hypotheses that can be divided
into three subsets:

(a) ideas could come from bodies (first hypothesis);


(b) they could come from the mind (second and fourth hypotheses);
(c) they could come from God (third and fifth hypotheses).9

The first hypothesis, a rough interpretation of the scholastic theory of the species
intellegibilis,10 is quickly refused by Malebranche, who reads this theory in a mate-
rialistic way. According to his interpretation, the species are a kind of Epicurean
simulacra, composed by matter and then impenetrable, and above all similar to
the bodies they are supposed to represent, a likeness that Malebranche, by means
of an argument already used by Descartes,11 refuses:

[W]‌hen we look at a perfect cube, all the species of its sides are unequal,
and yet we see all its sides as equally square. And likewise when we look
at a picture of ovals and parallelograms, which can transmit only species
of the same shape, we see in it only circles and squares. This clearly shows
that the object we are looking at need not produce species that resemble it
in order for us to see it. (SAT 221, OC 1:420)

But the nonsimilarity between the object represented and the representation is
not a sufficient sign of the correctness of the theory. In fact, the second hypoth-
esis, an empiristic model of knowledge according to which the mind produces
ideas—​which are here considered as dissimilar from their objects—​from the sen-
sible impressions it receives from bodies,12 is also refused by Malebranche, because
it gives man too much power. Since the ideas are real, Malebranche explains, if we

9.  “We assert the absolute necessity, then, of the following:  either (1)  the ideas we have of
the bodies and of all other objects we do not perceive by themselves come from these bodies
or objects; or (2) our soul has the power of producing these ideas; or (3) God has produced
them in us while creating the soul or produces them every time we think about a given object;
or (4) the soul has in itself all the perfections it sees in bodies; or else (5) the soul is joined to
a completely perfect being that contains all intelligible perfections, or all the ideas of created
being” (SAT 219, OC 1:417).
10. On this subject see Spruit 1995.
11. See Descartes, Dioptrics, Discours IV, AT 6:113. For a discussion of Malebranche’s denial of
likeness in representation see Scribano 2003.
12. Emanuela Scribano showed that this hypothesis could be found in Arnauld’s Logique de
Port-​Royal (I, 1). See Scribano 2006, 201ff.
Universals and Individuals in Malebranche’s Philosophy  •  147

say that the mind could produce ideas, we are then saying that the mind is able to
create something and, hence, that man has the same power as God:

According to them, it is in this that man is made after the image of God
and shares in His power. … This share in God’s power that men boast of
for representing objects to themselves … is a share that seems to involve
a certain independence… . But it is also an illusory share, which men’s
ignorance and vanity makes them imagine. (SAT 222, OC 1:422)

With a Platonic argument, Malebranche explains that we cannot represent some-


thing if its idea is not already present to our minds:

[A]‌s a painter, no matter how good he is at his art, cannot represent an


animal he has never seen and of which he has no idea—​so that the paint-
ing he would be required to produce could not be like this unknown
animal—​so a man could not form the idea of an object unless he knows it
beforehand, i.e., unless he already had the idea of it, which idea does not
depend on his will. (SAT 223, OC 1:424–​25)

Nor can it be said that man has the power to produce particular and distinct ideas
from general and confused ones, as shown by “the painter’s example”:

For just as an artist cannot draw the portrait of an individual in such a


fashion that he could be certain of having done a proper job unless he
had a distinct idea of the individual, … so a mind that, for example, has
only the idea of being, or of an animal in general cannot represent a horse
to itself, or form a very distinct idea of it, or be sure that the idea exactly
resembles a horse, unless it already has an initial idea against which it com-
pares the second. (SAT 223–​24, OC 1:425)13

The reason is that, according to Malebranche, the intellect of man is only a pas-
sive faculty that can only receive ideas and, thus, having no power at all, cannot
produce by itself any representative content:

[T]‌he faculty of receiving different ideas and modifications in mind


is entirely passive and contains no action; and I  call that faculty or that

13. We will come back to this passage because, as we will see, Malebranche in the following will
say that the idea of God, which is the basis for everything we could know about the world, is
precisely l’idée de l’être en general.
1 4 8   •  The Problem of Univer sals in Early Modern philosophy

capacity which the soul has of receiving all these things understanding
[SAT 3, OC 1:41, my emphasis]
… I  understand by this word understanding that passive faculty of
the soul by means of which we receive all the modifications of which it is
capable. (SAT 3, OC 1:41)

The same argument is invoked for refusing the third hypothesis—​the thesis
according to which we know the world through “ideas created with us” (SAT
226, OC 1:429)—​and the fourth theory—​which states that “the mind needs
only itself in order to see objects, and that by considering itself and its own
perfections, it can discover all external things” (SAT 228, OC 1:433). These
are two different versions of innatism—​the first is very close to the innatism
present in Descartes’s Fifth Meditation, and the second is possibly contained
in The Port-​Royal Logic, written by the Cartesians Antoine Arnauld and Pierre
Nicole.14 Whereas the fourth hypothesis is rejected by means of the “pow-
erlessness argument,” the third is not considered valid for a reason that will
become very important in Malebranche’s metaphysics, the so-​called economy
principle—​that is, the conviction in which God does not make complex that
which he can make in a simpler and more economical way. As Malebranche
writes:

To see the implausibility of this view, it should be considered that there are
in the world many totally different things of which we have ideas. But to
mention only simple figures, it is certain that their number is infinite, and
even if we fix upon only one, such as an ellipse, the mind undoubtedly con-
ceives of an infinite number of different kinds of them… . The mind, then,
perceives all these things; it has ideas of them; it is certain that it will never
want for ideas should it spend countless centuries investigating even a single
figure… . It has, then, an infinite number of ideas—​what am I saying?—​it
has as many numbers of ideas as there are different figures; consequently
since there is an infinite number of different figures, the mind must have
an infinity of infinite numbers of ideas just to know the figures. Now, I ask
whether it is likely that God created so many things along with the mind
of man. My own view is that such is not the case, especially since all this
could be done in another, much simpler and easier way. (SAT 226–​27, OC
1:430–​31)

14. See Nadler 1994 and Scribano 2006.


Universals and Individuals in Malebranche’s Philosophy  •  149

This “way” is precisely the vision in God of ideas, the fifth hypothesis described
by Malebranche, the only hypothesis that satisfies the three conditions previously
mentioned: (1) the reality of ideas; (2) the impotence of the human intellect—​
since, once provided with ideas, it would be independent from God; and (3) the
“economics” of God’s action.
Let us try to summarize the relevant points of Malebranche’s conception of
idea as it emerged in the argument of the Search:

(1) We cannot know the bodies directly but we can know them through ideas
(premise);
(2) ideas are the direct object of our knowledge of the world (consequence of
the premise);
(3) ideas are real (consequence of the thesis “nothing has no properties”);
(4) ideas can be neither material nor similar to material objects (first
hypothesis);
5) ideas shall already be present to our minds when we know something (sec-
ond hypothesis);
(6) ideas can be in us neither as dispositions (third hypothesis) nor contents
(fourth hypothesis), due to the thesis of the impotence of creatures;
(7) ideas can be created neither by us (second hypothesis) nor by God (fourth
hypothesis);
(8) ideas through which man knows things are not in human minds but in God
(fifth hypothesis and conclusion).

Given the tenets presented by Malebranche in the discussion of the second and
the fourth hypothesis—​that is, the “painter example” and the claim about “an
infinite number of ideas” for understanding a single figure—​it seems that we shall
conclude that, for Malebranche, in God we see an infinite number of particular
and distinct ideas. In the first edition of the Search we might find an answer that
confirms this suggestion:

But the strongest argument of all [for supporting the vision in God] is the
mind’s way of perceiving anything. It is certain, and everyone knows from
experience, that when we want to think about some particular thing, we
first glance over all beings and then apply ourselves to the consideration
of the object we wish to think about. Now, it is indubitable that we could
desire to see a particular object only if we had already seen it, though in a
general and confused fashion. As a result of this, given that we can desire
to see all beings, now one, now another, it is certain that all beings are
present to our mind; and it seems that all being can be present to our mind
1 5 0   •  The Problem of Univer sals in Early Modern philosophy

only because God, i.e., He who includes all things in the simplicity of His
being, is present to it. (SAT 232, OC 1:440)

But three years later, in the Tenth Elucidation, Malebranche will explicitly deny
this conclusion:

It should not be imagined that the intelligible word is related to the sen-
sible, material world in such a way that there is an intelligible sun, for
example, or an intelligible horse or tree intended to represent to us the
sun or a horse or a tree. (Tenth Elucidation, SAT 627, OC 3:153)

This difference between the two writings has been remarked upon by
Malebranche’s contemporaries, such as Arnauld, and by several of Malebranche
scholars. These scholars see a clear sign of a radical change that occurred in
Malebranche’s description of knowledge,15 a change that Malebranche on the
contrary has always denied. He therefore aimed to clarify, not modify, his opin-
ions. One possible way to shed light on the question, in my opinion, would be
to look to Malebranche’s conception of universal.
In order to better understand this point we must first remember that since his
first writings Malebranche has agreed with Descartes that the essence of bodies
is extension. As Malebranche writes in a chapter dedicated to “The essence of
matter” in the Search:

[E]‌xtension is the essence of matter … with extension alone we can cer-


tainly form the heavens, an earth, and the entire world we see as well as an
infinity of others. (SAT 245, OC 1:463)

To know the essence of bodies is then equivalent to knowing their extension, not
directly, but through the idea of extension:

[O]‌ur idea of extension suffices to inform us of all the properties of which


extension is capable, and we could not wish for an idea of extension, fig-
ure, or motion more distinct or more fruitful that the one God gives us.
(SAT 237, OC 1:450)

Since every body is nothing but a particular configuration of extension, the knowl-
edge of the essence of the general idea of extension, “the intelligible extension” in

15. Cf. Gueroult 1955 and Robinet 1965.


Universals and Individuals in Malebranche’s Philosophy  •  151

Malebranche’s words, is also the knowledge of the essence of the singular body. As
we read in the Tenth Elucidation:

[A]‌ll intelligible extension can be conceived of as circular, or as having the


intelligible figure of a horse or a tree, all of intelligible extension can serve to
represent the sun, or a horse or a tree, and consequently can be the sun or a
horse or a tree of the intelligible world. (SAT 627, OC 3:153–​54)

Malebranche explains that when he had said that we know the ideas, and not the
idea, of bodies in God,

I did not exactly mean that there are in God certain perfections that repre-
sent each body individually, and that we see such an idea when we see the
body; for we certainly could not see this body as sometimes great, sometimes
small, sometimes round, sometimes square, if we saw it through a particular
idea that would always be the same. (SAT 627–​28, OC 3:154)

Why then does Malebranche, in his critiques of innatism, speak about an infinite
number of ideas? It becomes clear in the third edition of the Search when, at the end
of the critiques, he concludes that if the ideas were created they have to be particular,
since every creature is particular:

[I]‌t is clear that the idea, or immediate object of our mind, when we think
about limitless space, or a circle in general, or indeterminate being, is nothing
created. For no created reality can be either infinite or even general, as is what
we perceive in these cases. (SAT 227, OC 1:432)

Indeed, this statement is consistent with another definition of ideas already present
in the first edition of the Recherche, where Malebranche defines the particular ideas
as “participations” of the idea of God:

[A]‌ll these particular ideas are in fact but participations in the general idea
of the infinite. (SAT 232, OC 1:441)16

16. It must be noted that properly speaking for Malebranche we do not possess any idea of
God, but we see God directly: “Only God do we know through Himself… . Only God do
we perceive by a direct and immediate perception. Only He can enlighten our mind with His
own substance. Finally, only through the union we have with Him we are capable in this life of
knowing what we know” (SAT 236–​37, OC 1:449).
1 5 2   •  The Problem of Univer sals in Early Modern philosophy

The entire passage is especially interesting: here, Malebranche is saying that the


last argument for supporting the vision in God is the proof of the existence of God
derived from the idea of the infinite. Recalling what Descartes has written in the
Third Meditation, Malebranche explains that our idea of infinite is primitive—​
that is, not derived from the finite. But unlike Descartes, Malebranche states that
(a) we know this idea only because we are “united” to God, and (b) the idea of
infinite, although “very distinct” coincides with the “general notion of being”:

[T]‌he proof of God’s existence, the loftiest and most beautiful, the pri-
mary and most solid (or the one that assumes the least) is the idea we have
of the infinite. For it is certain that (a) the mind perceives the infinite,
enough it does not comprehend it, and (b) it has a very distinct idea of
God, which it can have only by means of its union with him… . But not
only does the mind have the idea of the infinite, it even has it before that
of the finite. For we conceive of infinite being simply because we conceive
of being, without thinking whether it is finite or infinite. In order for us
to conceive of a finite being, something must necessarily be eliminated
from this general notion of being, which consequently must come first.
Thus, the mind perceives nothing except in the idea it has of the infinite,
and far from this idea being formed from the confused collection of all
our ideas of particular beings (as philosophers think), all these particular
ideas are in fact but participations in the general idea of infinite. (SAT
232, OC 1:441)

For Malebranche, therefore, the general idea of being, which is equivalent to the
idea of infinite, is positive—​that is, nonnegative, primary—​and unitary as well
as the idea of infinite, so that it cannot be seen as a mere combination, a sum,
of particular ideas, but the (general) principle from which the different (par-
ticular) ideas derive (“participate”). In this sense, we must underscore that the
relationship between the unity of God and the plurality of ideas is invoked by
Malebranche also in order to explain the origin of universals:

[I]‌t seems that all being can be present to our mind only because God, i.e.,
He who includes all things in the simplicity of His being, is present to it.
It even seems that the mind would be incapable of representing universal
ideas of genus, species, and so on, to itself had it not seen all beings con-
tained in one. (SAT 232, OC 1:441)

Hence, it is because of the connection between our minds and the infinite, that is,
God, that we can have general or universal ideas. This point will be clearly stated
Universals and Individuals in Malebranche’s Philosophy  •  153

in the later Dialogues on Metaphysics (Entretiens sur la Métaphysique, 1688),


in which Malebranche defends the thesis of the priority of the infinite on the
finite by arguing against an empiricist objection to the origin of universals. In
the II Dialogue, Ariste—​a young and talented man not yet fully convinced by
Malebranche’s philosophy—​declares that it is impossible for him to admit that
the general idea of being is not “un amas confus d’idées” (II Dialogue, art. viii,
OC 12:57). “Let us see what is true and what is false in this thought to which you
are so strongly predisposed,” answers Malebranche-​Theodore. Ariste’s reasoning,
according to Malebranche, is the following: one thinks of a circle that is one foot
in diameter, then a circle that is two feet, then three feet, and so on, and at least
he does not determine the diameter and thinks of a circle in general. This idea, if
Ariste were right, would be a confused set of the “different and particular circles”
thought. But this consequence is wrong, concludes Malebranche, since “the idea
of the general circle represents infinite circles and applies to them all, whereas you
have thought only of a finite number of circles” (DM 27, OC 12:58). Therefore,
the general idea of a circle cannot be the sum of n-​particular circles. But, con-
tinues Malebranche, there is a sense in which we can say that the general idea is
the result of a mental process that goes from the particular to the general. When
Ariste first considers the particular circles and second the circle in general, what
he does is to extend “the idea of generality over the confused ideas of circles you
have imagined” (DM 27, OC 12:58). This means that it is not Ariste who makes
the process of generalization, but “the idea of infinity” that transforms the par-
ticular ideas into a general one. As Malebranche writes,

I maintain you could form general ideas only because you find enough
reality in the idea of the infinite to give the idea of generality to your ideas.
You can think of an indeterminate diameter only because you see the infi-
nite in extension and can increase or decrease extension to infinity. I hold
that you could never think of these abstract forms of genera and species
were the idea of infinity, which is inseparable from your mind, not entirely
naturally joined to the particular ideas you perceive. You could think of a
particular circle, but never of the circle. (DM 27, OC 12:58)

Hence, according to Malebranche, the only reason why we can have general con-
cepts is that we are already and originally connected to the idea of the infinite,
which is then the only and real cause, metaphysical and cognitive, of what we
think of being our mental process of generalization. In other words, generality
is at the beginning and not at the end of our knowledge of general ideas, since it
coincides with the infinity of God. In this sense, God is the source of all general
ideas, not only the “good” ideas, but also the “bad” ones. In fact, with an amazing
1 5 4   •  The Problem of Univer sals in Early Modern philosophy

argument present both in the Search and in the Dialogues, Malebranche states
that the general idea of being, which as we have seen is equivalent to the idea of
the infinite, gives rise to many traditional, and wrong, notions:

Although this idea that we receive through the immediate union we have
with the Word of God, sovereign reason, never deceives us by itself as do
those we receive because of the union we have with our bodies, which rep-
resent things to us other than as they are, yet I have no hesitation in saying
that we make such ill use of the best of things that the ineradicable pres-
ence of this idea is one of the main causes of all of the mind’s disordered
abstractions, and consequently, [it is one of the main causes] of all that
abstract and chimerical philosophy that explains all natural effects with
the general terms act, potency, causes, effect, substantial forms, faculties,
occult qualities, and so on. For it is certain that all these terms and several
others arouse in the mind only vague and general ideas, i.e., those ideas
that present themselves to the mind with no difficulty or effort on our
part, those ideas that are contained in the ineradicable idea of being. (SAT
241–​242, OC 1:457)

Consequently for Malebranche, not all general ideas have a positive value in
knowledge: the general ideas of mathematics and the idea of extension are treated
very differently from the general concepts quoted above. The reason is that act,
forms, cause, and so on, do not denote real essences—​as in, for instance, the gen-
eral idea of extension—​but are mere logical and abstract terms.17 As Malebranche
writes,

[These ordinary philosophers] pretend to explain nature through their


general and abstract terms—​as if nature were abstract; and they would
absolutely have it that the physics of their master Aristotle is a true
physics, explaining the foundation of things, and not simply a logic.18
… [T]‌hey are so obstinate about these imaginary entities and the vague
and indeterminate ideas that spring naturally from their mind that they
are incapable of pausing long enough to consider the real ideas of things in
order to recognize their solidity and clarity. (SAT 243, OC 1:459)

17. In this sense it seems that for Malebranche 'general' is negative when equivalent to 'abstract'
and positive when equivalent to 'universal'.
18. The editor of the Search, Geneviève Rodis-​Lewis, notes that the same definition can be
found in Francis Bacon, Novum Organum, 1, I, aph. 63: Aristotelis physica nihil aliud quam
Dialecticae voces plerumque sonet.
Universals and Individuals in Malebranche’s Philosophy  •  155

“The real ideas of the things” are nothing but the general and infinite idea of
extension—​that is, the essence of the matter from which we can derive, through
an effort of attention, all the properties that we can find in the bodies: figure,
impenetrability, divisibility. Therefore, to know the nature in a distinct way is
equivalent to knowing the idea of extension—​a knowledge that, as we have seen,
corresponds to directly seeing the idea of extension in God.
But we could ask: if there are no particular ideas in God, how can we know an
individual body, for instance, the sun? Three answers may be given.
The first—​suggested by Martial Gueroult, which can be defined as a “con-
structive theory of individuals”—​consists in saying that man’s mind derives
a particular and distinct idea from the general idea of extension by means of a
consideration of the different relationships of distance that it can infer from sen-
sible experience and from God’s general idea of extension.19 As Gueroult himself
acknowledges, the main problem of this interpretation is that, if true, we shall
conclude that man has a knowledge of the sun that God, who knows the sun
only through the idea of extension, does not. Moreover, even if the constructive
theory could resolve the problem of the knowledge of individuals for man, it does
not explain how God, provided with only a general idea of the essence of bod-
ies, can know the individual bodies he creates before creating them—​a condition
invoked by Malebranche in supporting the vision in God of ideas:

God must have within Himself the ideas of all the beings he has created
(since otherwise He could not have created them), and thus He sees all
these beings by considering the perfections He contains to which they are
related. (SAT 230, OC 1:437)

The second answer, defined by André Robinet as the theory of “efficacious idea”
(idée efficace) and deepened more recently by Jean-​Christophe Bardout,20 links
the knowledge of individuals to the causality of God. In this account, man can

19. “Le monde des existences, qui dépend de la volonté de Dieu, se révèle à moi par un ensemble
de sensations dont le cours est réglé par la combinaison de lois immuables: loi du monde des
corps, lois de l’union de l’âme et du corps. Le jeu régulier de certaines perceptions permet de
découvrir entre elles des rapports, lesquels sont mesurables en tant qu’ils sont rendus possibles
par l’étendue intelligible, celle-​ci renfermant tous les rapports intelligibles des idées entre elles,
toutes les espèces possibles de rapports de distance, c’est à partir de l’interpretation mathéma-
tique de certains rapports observés que j’ai pu construire l’idée du soleil astronomique, c’est-​à-​
dire du soleil en soi” (Gueroult 1955, 224).
20.  Robinet dates to 1695 the birth of this theory (see Robinet 1965, 259–​72). See also
Bardout 1998.
1 5 6   •  The Problem of Univer sals in Early Modern philosophy

know a particular being because God applies the idea of extension to human
minds in different ways giving rise to a sensation, void of epistemic content, and a
particular idea, whose “intelligibility” is then an effect of God’s almightiness and
not a cause related to God’s mind.21 Furthermore, the dependency of the indi-
viduals on God’s causality invests God himself, which, according to this inter-
pretation, can know the individuals only in his will and then (logically) after the
creation.22 Hence, this answer also cannot explain how God can know individual
bodies before the creation of the world.
The third answer, which I  would like to propose here, needs a preliminary
consideration on what is the real problem that Malebranche must solve. What
I mean is that in Malebranche the problem of the knowledge of individuals is not
the problem of determining how we know existing bodies, since, in Malebranche’s
view, properly speaking we do not know existing bodies, but perceive them.23 Now,
the perception of a sensible object is obtained through a combination of a cogni-
tive element, the idea, and a particular sensation:

When we perceive something sensible, two things are found in our per-
ception: sensation and pure idea. The sensation is a modification of our
soul… . As for the idea found in the conjunction with the sensation, it is
in God, and we see it because it pleases God to reveal it to us. God joins
the sensation to the idea when objects are present. (SAT 234, OC 1:445)

In the first book of the Search, in which Malebranche discusses the sensations and
the errors caused by them, he explains that the function of the sensations, pain,
pleasure, and colors, is precisely to distinguish the different bodies:

[W]‌hile feeling pleasure and pain … we more easily distinguish the


objects that occasion them. (SAT 52, OC 1:128)
We need these colors only to know objects more distinctly, and that
is why our senses lead us to attribute them solely to objects. (SAT 60,
OC 1:142)

21. “Toute se passe comme si l’intelligibilité était désormais suspendue à l’efficience de l’idée”


(Bardout 1998, 112).
22.  “Dieu lui-​même ne semble pas en mesure de concevoir les possibles en leur singularité,
antérieurment à leur effectuation par sa volonté” (Bardout 1998, 105).
23. The confusion of knowing and perceiving is caused by Malebranche’s ambiguity in using
the two terms often interchangeably. For an exhaustive account of this subject, see Nadler
1992, 60ff.
Universals and Individuals in Malebranche’s Philosophy  •  157

As we have seen, because sensations have only an adaptive function, and not a
cognitive one, they allow us to discern bodies, since to survive we need to quickly
discern them; for this reason, God has established some laws of the mind-​body
union that produce this kind of experience in us. But perceiving is not know-
ing and knowing is the task of the “pure intellect,” a faculty that does not have a
corporeal equivalent.24 Therefore, the problem of the knowledge of individuals
shall not be related to the knowledge of existing things, as Robinet and Bardout
seem to think, but to the problem posed by Gueroult of the relationship between
the general idea of the essence of bodies and the notion of idea as archetype of
creation.
Now, I think that a possible solution can be found if we consider Malebranche’s
definition of God’s ideas as God’s “perfections” and as “participations of the gen-
eral idea of infinity.” But in order to understand this assertion we must take a
step back and turn to the traditional conception of idea, analyzing one of the
“standard” Christian theories of God’s knowledge: Thomas Aquinas’s theory of
divine ideas.25

7.2
The opportunity to dwell on Aquinas’s conception of divine ideas lies with the
strict connections that have been noted between Malebranche’s vision in God
of ideas and Aquinas’ description of the beatific vision.26 A  brief recalling of
Malebranche’s relevant positions can be useful here.
As we have seen, Malebranche firmly denies that human ideas are mental
modifications, not only because of the mentioned thesis of the radical impotence
of creatures, but also because, as Malebranche will write in answering Arnauld’s
objections, if ideas were properties of human minds, they would be as finite and
particular as everything belonging to creatures. Moreover, mental modifications

24. As Malebranche writes at the beginning of the third book of the Search: “We shall first
discuss the mind as it is in itself and without any relation to the body to which it is joined.
Accordingly, what will be said about it could be said as well about pure intelligences and a
fortiori about what we have here called pure understanding, for by the expression pure under-
standing, nothing is meant but the mind’s faculty of knowing external objects without form-
ing corporeal images of them in the brain to represent them” (SAT 198, OC 1:380–​81). The
problem of the pure intellections is one of the targets of the first critics of Malebranche, Simon
Foucher. See on this subject Favaretti Camposampiero 2010.
25. Despite Malebranche’s criticisms of scholastic philosophy, several scholars have shown how
deep Malebranche’s debt is to it. See, for instance, Connell 1967, Scribano 1996 and 2003, and
Priarolo 2004.
26. See the previously cited Priarolo 2004, Scribano 1996 and 2003, and Trottmann 1998.
1 5 8   •  The Problem of Univer sals in Early Modern philosophy

cannot have the features of clearness and distinction that characterize the ideas
through which we know the essence of the world, since the knowledge we have
of our minds is not clear and distinct, but obscure and confused. As Malebranche
explains in the Search after Truth, we do not know our mind through an idea, but
through an “inner sensation,” or “consciousness,” so that “our knowledge of it is
imperfect” (SAT 238, OC 1:451). If ideas were mental modifications, the finitude
and obscurity of our mind would, then, also invest them, and this would make
it impossible to represent anything as clear and infinite as God—​whose knowl-
edge is for Malebranche undeniable—​or the idea of extension—​whose clearness
and distinctness is considered by Malebranche as unquestionable. Hence, for
Malebranche we know things because of God’s ideas, and we have access to God’s
ideas since we are strictly united to God and we see him directly. As we read in the
first edition of the Search after Truth:

Only God do we know through Himself… . Only God do we perceive by


a direct and immediate perception. Only He can enlighten our mind with
His own substance. Finally, only through the union we have with Him are
we capable in this life of knowing what we know. (SAT 237, OC 1:449)

Now, what we do “in this life,” according to Malebranche, is what we will do in


another life according to Thomas Aquinas.
As it is well known, for Aquinas the viator cannot know the essence of God,
but can only attain his existence, starting from the effects of God’s action in the
world. Since our minds are embodied—​and for this reason we know things
through species abstracted by the sensible impressions that are the ground for
every cognitive experience we have of the world—​in order to know God we
would need a species of God—​that is, a finite medium representing God’s essence.
But, in Aquinas’s view, knowing the essence of God in this way is impossible: first,
because there would be no proportion between the representation and the repre-
sentatum; second, because of the indistinction of essence and existence in God,
which makes impossible that something which is created, whose essence is always
distinct from existence, could be similar to God and thus able to represent him;
and thirdly, because the infinite essence of God cannot be circumscribed by a
finite species.27 But if knowledge of God in himself is the ultimate scope of the

27. “[T]‌he essence of God cannot be seen by any created similitude. First, because as Dionysius
says (Div. Nom. i), ‘by the similitudes of the inferior order of things, the superior can in no way
be known;’ as by the likeness of a body the essence of an incorporeal thing cannot be known.
Much less therefore can the essence of God be seen by any created likeness whatever. Secondly,
because the essence of God is His own very existence, as was shown above (q. 3, art. 4), which
Universals and Individuals in Malebranche’s Philosophy  •  159

human soul, its “beatitude,” how is it possible to deny that “any created intellect
can see the essence of God”? (ST, I, q. 12, art. 1). Against the negative theology
of Dionysius, noted in many other parts of his writing, Aquinas explains that
man must know the essence of God, since denying this knowledge to man would
mean depriving him of achieving “his highest function, which is the operation of
his intellect”:

[I]‌f we suppose that the created intellect could never see God, it would
either never attain to beatitude, or its beatitude would consist in some-
thing else besides God, which is opposed to faith. (ST, I, q. 12, art. 1, resp.)

We must underscore the word “see,” since, as we have mentioned, according


to Aquinas, knowledge of the essence of God cannot be mediated, occurring
through a finite species that functions as a medium between human minds and
God. Hence, the only way a created mind has for knowing God is to see him. Of
course, this kind of knowledge is not available to sinners or during this life, but
only to those who deserve to see God “face to face”: the blessed, which are pure
souls until the resurrection of their bodies strengthened by the lumen gloriae. But
a new question arises: if the blessed are pure souls, how can they know material
things? The answer given by Aquinas is that the blessed know things through
God’s essence “as in an intelligible mirror”:

The created intellect of one who sees God is assimilated to what is seen in
God, inasmuch as it is united to the Divine essence, in which the simili-
tudes of all things pre-​exist. (ST, I, q. 12, art 9 ad contra, resp.)

The similitudes of all things are nothing other than God’s ideas, which are required
for the intentional creation of the world.28 Hence, similar to Malebranche’s men,
Aquinas’s blessed know material things through God’s ideas.

cannot be said of any created form; and so no created form can be the similitude representing
the essence of God to the seer. Thirdly, because the divine essence is uncircumscribed, and con-
tains in itself super-​eminently whatever can be signified or understood by the created intellect.
Now this cannot in any way be represented by any created likeness; for every created form is
determined according to some aspect of wisdom, or of power, or of being itself, or of some like
thing. Hence to say that God is seen by some similitude, is to say that the divine essence is not
seen at all; which is false” (ST, I, q. 12, art. 2, resp.).
28. “As then the world was not made by chance, but by God acting by His intellect … there
must exist in the divine mind a form to the likeness of which the world was made. And in this
the notion of an idea consists” (ST I, q. 15, a. 1, resp.).
1 6 0   •  The Problem of Univer sals in Early Modern philosophy

It is important to note that, according to Aquinas, ideas are not different


from the essence of God—​they are not creatures or Platonic models external to
God—​and must be many because God knows every aspect of what he creates
and because every created thing participates in some respects of God’s being. In
this sense, ideas—​which are defined by Aquinas as the same essence of God as
“capable of imitation” by creatures—​are not only the cognitive principles of the
divine knowledge of things, but also the ontological principles of creation:

[M]‌any ideas exist in the divine mind, as things understood by it; as can
be proved thus. Inasmuch as He knows His own essence perfectly, He
knows it according to every mode in which it can be known. Now it can
be known not only as it is in itself, but as it can be participated in by crea-
tures according to some degree of likeness. But every creature has its own
proper species, according to which it participates in some degree in like-
ness to the divine essence. So far, therefore, as God knows His essence as
capable of such imitation by any creature, He knows it as the particular
type and idea of that creature; and in like manner as regards other crea-
tures. (ST I, q. 15, art. 2 ad 2, resp.)

Notice that the plurality of ideas does not conflict with the simplicity of God, an
attribute that Aquinas strongly defends in many places,29 precisely because the
differentiation of ideas lies not in God’s essence considered by itself, but in the
relationships that the knowledge of God has with the plurality of things. This
does not mean that the plurality of ideas is caused by the plurality of things cre-
ated, but rather that the plurality depends on an act of the divine intellect, that is,
“by the divine intellect comparing its own essence with these things” (ST I, q. 15,
art. 2, ad 3). Hence, the relations that give rise to the multiplicity of God’s ideas
“are not real relations, such as those whereby the Persons are distinguished, but
relations understood by God” (ST I, q. 15 art. 2, ad 4), so that we could say that
the essence of God is the unique model of the entire creation. As Aquinas will
write in Question 44,

[D]‌ivine wisdom devised the order of the universe, which order consists in
the variety of things. And therefore we must say that in the divine wisdom
are the types of all things, which types we have called ideas—​i.e., exemplar
forms existing in the divine mind. And these ideas, though multiplied by
their relations to things, in reality are not apart from the divine essence,

29. On these questions, see Wippel 2000.


Universals and Individuals in Malebranche’s Philosophy  •  161

according as the likeness to that essence can be shared diversely by differ-


ent things. In this manner therefore God Himself is the first exemplar of
all things. (ST I, q. 44, art. 3, resp.)

In brief, for Aquinas the multiplicity of ideas is a consequence of an act of knowl-


edge, which grounds different relations between the essence of God and creatures
and is not an ontological feature of ideas, whose reality is, then, not different
from the reality of the essence of God. Moreover, as we have mentioned, ideas
can be considered both as cognitive principles and ontological principles of the
creation. In the first sense, Aquinas calls them “notions” (rationes), in the second
sense, “exemplars.”30 As we read in Question 15,

As ideas, according to Plato, are principles of the knowledge of things and


of their generation, an idea has this twofold office, as it exists in the mind
of God. So far as the idea is the principle of the making of things, it may
be called an “exemplar”, and belongs to practical knowledge. But so far as
it is a principle of knowledge, it is properly called a “notion” (ratio), and
may belong to speculative knowledge also. As an exemplar, therefore, it
has respect to everything made by God in any period of time; whereas as
a principle of knowledge it has respect to all things known by God, even
though they never come to be in time; and to all things that He knows
according to their proper notion, in so far as they are known by Him in a
speculative manner. (ST I, q. 15, art. 3, resp.)

As exemplars, ideas pertain to the domain of God’s practical knowledge, that is,
the knowledge that God has about things that he will create “in any period of
time,” and for this reason their objects are also the singulars, since singular objects
are what God creates. As Aquinas has explained in the Summa contra Gentiles,
since God, who knows himself perfectly, is the cause of everything, and “when
the cause is known, the effect is known” (SG bk. 1, chap. 60, art. 2), then God
knows perfectly everything he created “not only universally, but also in the sin-
gular” (SG bk. 1, chap. 65, art. 2). But “as principle of knowledge,” that is, as an
object of the divine intellect, ideas denote “proper notions,” which are the same
essence of God considered as “capable of imitation by the creatures”:

[T]‌he divine essence comprehends within itself the nobilities of all beings,
not indeed compositely, but, as we have shown above, according to the

30. See Doolan 2008.
1 6 2   •  The Problem of Univer sals in Early Modern philosophy

mode of perfection. Now, every form, both proper and common, consid-
ered as positing something, is a certain perfection; it includes imperfec-
tion only to the extent that it falls short of true being. The intellect of
God, therefore, can comprehend in His essence that which is proper to
each thing by understanding wherein the divine essence is being imitated
and wherein each thing falls short of its perfection. Thus, by understand-
ing His essence as imitable in the mode of life and not of knowledge, God
has the proper form of a plant; and if He knows His essence as imitable
in the mode of knowledge and not of intellect, God has the proper form
of animal, and so forth. Thus, it is clear that, being absolutely perfect,
the divine essence can be taken as the proper notion [ratio] of singulars.
Through it, therefore, God can have a proper knowledge of all things. (SG
bk. 1, chap. 54, art. 4)

Hence, the divine perfections that represent the things, which are God’s ideas,
contain “virtually” all their possible “specifications” and then can provide the
knowledge of the singulars without being in themselves singulars. In other words,
the singularity typical of things that can be created by God is not a real prop-
erty of God’s ideas—​which are the same infinite essence of God—​but, rather, a
relational property that arises from the divine act, logically prior to the creation,
of comparison between his essence and the different perfections he possesses.
Therefore, from an epistemological point of view, the ideas are but different ways
of looking at the same source of knowledge, the essence of God—​an outlook that
gives rise to different models for practical knowledge; in other words, the act of
creating the world. The epistemological aspect of ideas seems then to precede the
ontological feature of ideas, the causality of God following his understanding. In
Aquinas’s words, “God causes things through His intellect, since His being is His
understanding” (SG bk. 1, chap. 50, art. 3).

7.3
Let us now come back to Malebranche. As we have seen, since the first edition of
the Search after Truth, Malebranche has defined God’s ideas as “the perfections
of God that represent them” (SAT 68, OC 1:157). With words very similar to
Aquinas’s, in refusing the fourth hypothesis on the origin of ideas, Malebranche
explains that

It cannot be doubted that only God existed before the world was created
and that He could not have produced it without knowledge or ideas; con-
sequently, the ideas He had of the world are not different from Himself,
Universals and Individuals in Malebranche’s Philosophy  •  163

so that all the creatures, even the most material and terrestrial, are in God,
though in a completely spiritual way that is incomprehensible to us. God
therefore sees within Himself all beings by considering His own per-
fections, which represent them to Him… . He sees all these beings by
considering the perfections He contains to which they are related. (SAT
229–​30, OC 1:434–​35)31

The connection with Aquinas will become explicit in the Tenth Elucidation in
the Search after Truth, where Malebranche writes

God’s ideas of creatures are, as Saint Thomas says, only His essence, insofar
as it is participable or imperfectly imitable, for God contains every crea-
turely perfection, though in a divine and infinite way. (Tenth Elucidation,
SAT 625, OC 3:149)

By saying that we see God’s ideas, Malebranche means then to say that we see
“the divine substance … as relative to creatures and to the degree that they can
participate in it” (SAT 231, OC 1:439). God reveals to us “what in Him is related
to and represents these things” (SAT 231, OC 1:439), a “revelation” that in our
minds give rise to particular representations: “He is no being in particular[, but]
what we see is but one or more particular beings, and we do not understand this
perfect simplicity of God” (SAT 231, OC 1:439). Therefore, the constructive
process proposed by Gueroult to explain how it can be possible that man knows a
particular object as the astronomical sun is realized by God himself, who, looking
at his essence in different ways, gives rise to the different relationships that can
represent any object.
The undeniable ambiguity that characterizes Malebranche’s description of
ideas—​sometimes defined as particular and sometimes as genera—​seems then
caused by the same twofold consideration of ideas present in Aquinas, as ontologi-
cal exemplars and as notions. But with regard to the first meaning of ideas, it should
be noted that Malebranche does not have the same problems as Aquinas because,
following Descartes, for Malebranche God creates one and only one material sub-
stance: extension. Singular bodies are only modifications—​that is, particular con-
figurations of extension—​resulting from the laws of motion that God has given to

31.  Gueroult concludes that Malebranche’s theory of knowledge is not fully consistent and
cannot resolve once and for all the problem of the knowledge of individuals—​above all because
Malebranche’s conception of idea tries, and fails, to keep together the traditional Platonic defi-
nition of idea as the archetype of creation and the new Cartesian notion of extension. See, for
instance, Gueroult 1955: 249.
1 6 4   •  The Problem of Univer sals in Early Modern philosophy

the world;32 consequently, for creation only one exemplar is needed. Instead, with
regard to the second meaning, the general idea of extension requires a multiplica-
tion in order to give rise to a real knowledge of the world, even if, as we have seen,
knowledge of the particular depends on knowledge of the universal. But this multi-
plication, caused by an act of God’s understanding, does not produce new realities—​
that is, new (particular) ideas—​but only different relationships between the general
idea of extension and the universal ideas of numbers present in God. In this sense,
as Malebranche explains, knowing the relationships present in the ideas is nothing
more, nothing less, than knowing the truth—​a truth, however, that does not possess
reality by itself:

We are of the opinion … that truths (and even those that are eternal, such
as that twice two is four) are not absolute beings, much less that they are
God himself. For clearly, this truth consists only in the relation of equality
between twice two and four… . The ideas are real, whereas the equality
between ideas, which is the truth, is nothing real. (SAT 234, OC 1:444)

If my interpretation is correct, we understand, then, why Malebranche so firmly


denies having changed his opinion about the presence of particular ideas in
God: in fact, as an ontological model of creation, as archetypes, in God there is
only one idea, the idea of extension. But considered as cognitive objects, God’s
ideas are several. This is not because of an act of man’s understanding—​which,
as we have seen, is considered by Malebranche to be a very limited faculty, as
Gueroult suggested—​but because of an act of the understanding of God, which,
regarding his essence, “multiplies” the possible cognitive objects; this is the foun-
dation for the various relationships between them that man can know.33 Secondly,
it should be noted that these ideas maintain a strict connection with its source,
the essence of God. This is because, as we have seen, for Malebranche, only ideas
that represent real essences are real ideas and not mere logical or, in his words,
abstract concepts which denote nothing that is real. In this sense, the ontological
ground of ideas—​their being the essence of God as exemplar of the world—​is
what guarantees the reality and effectiveness of our knowledge. For this reason,
as we suggested at the beginning of this chapter, according to Malebranche it is

32. See SAT bk. 6, pt. 2, chap. 4, 453–​66, in which Malebranche resumes the fourth part of
Descartes’s Principia philosophiae.
33. The above-​mentioned “painter’s example” can be read, in this sense, as an element of proof
of the thesis here proposed, since it shows that according to Malebranche even if we possess
general ideas we cannot derive from them particular ones. We cannot, because only God has
this power.
Universals and Individuals in Malebranche’s Philosophy  •  165

very difficult to separate the epistemological conception of universals from their


metaphysical notion. It is precisely this ambivalence that gives rise to the several
problems of interpretation that we have detected.34

Abbreviations
AT = Descartes 1964–​74; cited by volume and page.
DM = Malebranche 1999; cited by page.
OC = Malebranche 1958–​67; cited by volume and page.
SAT = Malebranche 1997; cited by page.
SG = Aquinas 1955–​57; cited by book, chapter, and article.
ST = Aquinas 1947; cited by part, question, and article.

34. I would like to thank Elisa Angelini, Stefano di Bella, and Emanuela Scribano for their help-
ful suggestions on different topics related to these pages, and Joseph Raho and Tad Schmaltz
for their kind assistance with the English text.
8 U N I V E R S A L S I N   E N G L I S H P L AT O N I S M

More, Cudworth, Norris

Brunello Lotti

It is well known that the problem of universals is not central to early


modern philosophy, and English Platonism1 is not exceptional in this
regard. However, the English Platonists’ lack of interest in the scho-
lastic disputes on this subject does not imply that they neglected the
basic issues involved in the traditional question of universals. The
positions of More, Cudworth, and Norris can be reconstructed from
their explicit statements on this residual problem and, beyond that,
from the main traits of their epistemologies and ontologies. More and
Cudworth can be understood to be metaphysical conceptualists inso-
far as their epistemological conceptualism is linked to a metaphysi-
cal archetypal theory of creation. As Christian Platonists they reject
essentialism,2 while as Platonists partially influenced by Descartes’s
ontology they reject the immanentistic conception of universals that
stemmed from Aristotle. Henry More regards universals as abstract
concepts and adopts the ontological principle that whatever is is indi-
vidual. Cudworth too restricts universals to the mental realm, stating
that “whatsoever exists without the mind [is] singular.” However,
for More and Cudworth universal concepts are not abstract ideas
inductively constructed by the human mind, but are primarily divine

1. Though the term ‘Cambridge Platonism’ is perhaps more familiar, and though I do
use the term to speak of More and Cudworth, it is not applicable to John Norris, given
that he studied at Oxford. I therefore apply the term ‘English Platonism’ as a label for
the three philosophers. There is a precedent in the literature for the use of this term;
see, for instance, A. O. Lovejoy’s “Kant and the English Platonists” (Lovejoy 1908),
as well as the references to “English Platonism” in Muirhead 1931, 210, 212, and 422.
2. Bolton 1998, 182f. I use the term ‘essentialism’ to designate “the doctrine, in
Plato, of the existence of Forms, that is, of really existing abstract entities of which
physical objects are imperfect copies” (Flew 1984, “Essentialism,” 112).
Universals in English Platonism  •  167

thoughts and secondarily a priori ideas in the human mind, which is a finite copy
of the divine Intellect. The archetypal theory of creation and the connection of
finite minds to God’s Mind ensure the objective validity of universal concepts.
Since a priori universal ideas are principles of divinely guaranteed true knowl-
edge, the question arises of how this is compatible with More’s and Cudworth’s
ontological assumption that whatever exists outside the mind is individual. The
answer is provided by the Platonic view that the created world is pervaded by
the order impressed in it by God—​an order understandable in terms of univer-
sal essences and relations—​but that ordering universal patterns are manifested in
bodies only in an imperfect manner (by way of imitation and participation). As
for John Norris, he shares with the Cambridge Platonists the archetypal theory
of creation, but rejects innatism, and his discussion of the problem of universals
is framed in terms of his theory of the ideal world inspired by the Malebranchean
doctrine of the vision in God and by the Augustinian conception of divine illu-
mination. A further difference is that for Norris, universals are eternal essences
existing in the divine mind as constituents of the infinite essence of the eternal
wisdom, while Cudworth had defined them as noémata (thoughts in the human
mind corresponding to those in God’s mind), which do not have any actual exis-
tence apart from the thinking of human minds and the divine mind.

8.1. Henry More
More deals with the issue of universals in one of his major works, Enchiridion
Metaphysicum (first edition: 1671), where a conceptualist theory is stated:

First, then, is to be noted that all universals of whatever genus, of which


kind are body, animal, tree, man, and the like, are indeed, insofar as they
are abstracted from singulars, not things themselves, but certain modes
according to which our human reason conceives the things themselves.3

Things are only singular, while genera are merely “modes which we employ in
contemplating things.”4
More distinguishes the general concepts employed in the various disciplines
(philosophy of nature, metaphysics, and mathematics) from the most universal
ones we make use of in thinking of any subject whatsoever. These last he defines
as logi dialectici, that is, logical notions. They concern singular beings in so far as

3. EM II.3, 143; cf. MM 6.


4. EM II.4, 143 (MM 7).
1 6 8   •  The Problem of Univer sals in Early Modern philosophy

we consider them not as members of a determinate genus (e.g., as a single body


or tree, or animal or man), but only as res, as substances or beings as such, that
is, when we take each singular being into account from the most common and
general point of view (sub generalissima communissimaque ratione).5 The nega-
tion of natural kinds as universals in re is coupled with the claim of the objec-
tive validity of logical notions, “according to which we consider Being as Being,
namely when we compare the thing itself either with its causes or principles, or
with any other thing.” So, although universal concepts do not correspond to any
universal entity, they illustrate the common relations of things and their mutual
connections, which More designates with the Greek word skeseis, translated as
habitudines logicae, logical dispositions.6 More’s theory of universals is based on
the distinction between logic and metaphysics, and, at the same time, takes into
account the ontological relevance of logical notions. I will try to clarify this deli-
cate balance in the following discussion.
Many years before the Enchiridion Metaphysicum, in his letter to Descartes
of March 5, 1649, More stated that the essences of substances are inscrutable
(“the root and essence of every thing lies buried in eternal darkness”: radix rerum
omnium ac essentia in aeternas defossa latet tenebras) and therefore that in order to
define a substance it is indispensable to refer to its relations and properties. This
skeptical statement implies that universals are not to be considered as the essences
of substances, but only as properties or relations. In the same letter, debating with
Descartes the controversial issue of extension, More states that universal notions,
in so far as they belong to logic, have to be considered independently from their
adherence to real being. Among the universal notions of logic are listed “part and
whole, subject and adjunct, cause and effect, opposites and relates, contradicto-
ries and privatives,” and the idea of extension as well, since it means only “parts
which are outside other parts.” All these notions concern indifferently both
being and not being, and when we forget their logical nature, we get involved
in a sequence of deceptive mental games. It is a dangerous mistake to forget that
logical notions, by which we consider external things, are our modes of thinking
and to regard them as something existing in the things themselves separately from
human mind.7

5. EM II.3–​4, 143 (MM 7). See also More 1662, Preface General, xiii: “And Universals are not
things, but notions.”
6.  EM II.4, 143 (MM 7). Cudworth too will use the term skesis with the meaning of ideal
and logical relation (see EIM: IV.ii.4, 154; IV.ii.5, 156f.; IV.ii.7, 161; IV.ii.8, 163; IV.ii.9, 165;
IV.ii.10, 166; IV.ii.11, 172; IV.ii.12, 174; IV.ii.13, 176; IV.iii.1, 188). The main source is Enneads
VI.i.6–​9, where Plotinus discusses the Aristotelian theory of relations.
7. More to Descartes, Cambridge, March 5, 1649, AT 5:299, 307.
Universals in English Platonism  •  169

In the Enchiridion Metaphysicum More remained faithful to these initial


views. He claims that the distinction between universal and singular—​as well as
between formal and objective concept, or between simple and composed, iden-
tical and different, and the notions of existence, entity and perfection—​must
be treated by logic and not by metaphysics.8 Even the discussion of the princi-
pium individuationis belongs to logic and not to metaphysics.9 Logic deals with
universal logoi, while metaphysics is the science of res incorporeae, whose reality
is ascertained through arguments based on natural philosophy.10 The fact that
More regards universals as a logical topic shows his intention of reforming the
traditional conception of metaphysics. To him, metaphysics is a discipline that
enables us to rise “from the more accurate knowledge of nature or the world …
into a sufficiently clear knowledge of God and of the other incorporeal enti-
ties.”11 Metaphysics allows us to know “the incorporeal universal causes,” so called
because their “influence … is indeed most universal” and their power of act-
ing (efficiendi vis) is “the most noble and most simple.”12 The res incorporeae are
spirits or immaterial substances, of which metaphysics gives us “true and intel-
ligible notions.”13 More highlights his intention of detaching himself from the
Aristotelian tradition in which metaphysics deals not only with forms and prop-
erties of incorporeal substances, but also with general concepts like being qua
being, which should instead belong only to logic.14 If metaphysics is a kind of
knowledge that emerges from natural philosophy, the ens quatenus ens cannot be
its object, “because it is so general that it antecedes physical things in the order
both of nature and of doctrine.” Therefore, as a “most universal notion,” its con-
sideration must be assigned to logic.15
Since metaphysics is the science of incorporeal entities, we may wonder if it
also provides a knowledge of universal aspects pertaining to the nature and action
of incorporeal substances. More denies this:  cause, form, matter, universally
regarded, are logical notions by which we refer in actual reality to singular beings,
to this determinate cause, to this determinate form and matter which make a

8. EM III (title), 149 (MM 19).


9. EM III.5, 150 (MM 21).
10. EM I.1–​5, 141f. (MM 1–​3).
11. EM I.1, 141 (MM 1).
12. EM I.2, 141 (MM 1f.).
13. EM I.5, 142 (MM 3).
14. EM I.7, 142 (MM 3).
15. EM II.1–​2, 143 (MM 6).
1 7 0   •  The Problem of Univer sals in Early Modern philosophy

thing what it is.16 The attribution of the problem of universals to the domain of
logic means that universals are only “the universal modes that the human mind
uses in the contemplation of things.” From an ontological point of view, every
being, as such, is singular, because it exists only hic et nunc, and whatever is not
located in a time and place is nothing.17 Consequently, the principium individu-
ationis of the scholastic tradition is made redundant and More disposes of it with
an ironical simile:

Every being … is singular immediately and by itself, apart from any other
superadded thing which makes it singular. … So that there is no need of
any principle of individuation as a clasp for detaining a thing lest it fly into
the universal, but, from the fact that it is, every thing is singular.18

More’s thesis that whatever beings are individuated by their spatio-​temporal prop-
erties should not be regarded as an example of those theories which identified the
principle of individuation with accidental features. One traditional example of
such a theory, which More might have known, is that of Boethius in De Trinitate,
according to which the numerical difference between individuals is ascribed to
accidents, in particular to their position in space; two bodies cannot occupy the
same place and so space is a sufficient condition to differentiate two individual
beings.19 More’s thesis differs from that of Boethius because for the Cambridge
Platonist space and time are not extrinsic accidents but essential constituents of
being. If in Boethius the principle of individuation is provided by accidents, in
More this principle is redundant in so far as it would be applied pointlessly to
beings that are individuated by an existence that is intrinsically characterized by
extension and duration, two attributes which God shares with his creatures.20
To understand better More’s thesis on individuality, the passages quoted
from Enchiridion Metaphysicum should be read in connection with the thesis of
the unknowability of the essence of substances, which he asserts in his letter to

16. EM II.5, 144 (MM 7).


17. EM III.5, 150. See also EM XXVII.9, 311. A possible source of this thesis is B. Keckermann’s
Scientiae Metaphysicae compendiosum systema (Hanau 1609): see Thiel 1998, 222.
18. EM III.5, 150 (MM 21).
19. Boethius 1968, 6 (De Trinitate, I, 23–​31).
20. More’s notion of space and its connection with his idea of God is too well known to need
documentation here (see on the subject Reid 2012, chaps. 4–​6, 103–​236). As for the divine
attribute of duration see More 1662, An Antidote against Atheism, bk. 1, chap. 4, §1, 14; An
Appendix to the foregoing Antidote, chap. 4 §1, 155; chap. 7, §2, 164. See also More’s letter to
Descartes, March 5, 1649, AT 5:306.
Universals in English Platonism  •  171

Descartes of March 5, 1649 (quoted above). This is indeed a thesis that More
maintains throughout his life. We find it stated again in An Antidote against
Atheism (1653) and then as an axiom in The Immortality of the Soul. In the former
More wrote that “the very essence or bare substance” of anything, either corporeal
or incorporeal, is unknowable, though “the essential and inseparable properties”
of both kinds of substances are “intelligible and explicable”; in the latter it is said
that the bare essence of a substance is inconceivable and that substance vanishes
if we subtract its aptitudes, operations, properties, and modifications.21 Finally, it
is repeated in the late Annotations upon the Discourse of Truth (1683). Here More,
having defined his ontology in terms of substances and modes, says that “’tis these
modal differences of Substances which we only know, but the specifick substance
of any thing is utterly unknown to us.”22 The assumption that the “intimate Form”
of a substance is in itself obscure, and that what is knowable is only “the insepara-
ble fruits or results thereof,” is proclaimed as a self-​evident principle that does not
need any demonstration.23 More’s thesis is clear in this sentence: “[the substance
is known] only by its essential modes, but the modes are not the substance it
self of which they are modes, or every substance would be more substances than
one.”24 More thus separates the epistemological approach to the knowledge of
the substance from the basic ontological assumption that substance and modes
are distinct. Each substance, being individual, cannot be reduced to its essential
properties, and therefore the knowledge of the modes (which in themselves are
universal concepts) does not exhaust the meaning of the individual substance. An
obscure and unique essence remains, which constitutes the individuality of each
being. Given More’s metaphysics, time and space are conditions of individuation
in the sense that for each being its position in space and time relates it in a distinc-
tive way to the omnipresence and duration of God.
Examining the Aristotelian distinction between primary and secondary sub-
stances, More judges it “the most absurd,” because the secondary (or universal)
substance is not a substance at all, “but only a certain logical kind or mode of
conceiving the entire multitude of singular substances under a separate similitude
or idea in which all suit.” Only singular substances exist while the so-​called sec-
ondary substances (species and genera) are only resemblances, that is, common

21. More 1662, An Antidote against Atheism, bk. 1, chap. 4, §3, 15; The Immortality of the Soul,
bk. 1, chap. 2, axiom VIII, 19. These statements have probably influenced Locke’s theory of
substance.
22. More 1683, 208. The specific form, that is, the specific nature, of a substance is unknown,
while only the “essential or inseparable Attributes” are known (229). See also 231.
23. Ibid., 231.
24. Ibid., 208.
1 7 2   •  The Problem of Univer sals in Early Modern philosophy

qualities conceived by our intellect that we call universals: “All universals are in


the intellect, singulars are in nature.” The secondary substance, as it represents a
resemblance in individual things, “is an operation of the intellect, and therefore
not a substance, but an accident.” In conclusion, the universal is an accident of the
incorporeal substance and is not an object existing outside the mind.25
More’s conception of universals exhibits two features that are typical of the
nominalistic approach, namely, the superfluity of a principium individuationis,
given that individuality is a primary property of each being, and the antireal-
ist warning not to confuse general concepts with concrete features of the real-
ity outside the mind.26 However, it is not likely that More was influenced either
by the Medieval nominalists or, still less, by the most notorious nominalist of
his own age, the much-​hated Thomas Hobbes. As a possible source of More’s
view on the principle of individuation I  would suggest Suárez, at least insofar
as the Spanish philosopher had stated that substances are individuated through
their entity with no need of an added principle of individuation.27 As for the
conception of universals, More might have been influenced by Descartes, who in
Principia Philosophiae I.58 and 59 had defined universals as “manners of think-
ing” (modi cogitandi), that is, as general and abstract ideas formed by the mind.28
Even if, for both More and Descartes, universals are only concepts, they are none-
theless productive of true cognition, since even if they must not be confused with
real things, they still provide us with knowledge of real things. The knowledge
obtained by means of universal concepts is rooted in the a priori structure of
the human intellect. In Enchiridion Metaphysicum II §3, as we have seen, More
talks of universal concepts as “abstracted” from singular things, without specify-
ing how the process of abstraction is developed, but he seems to refer only to
natural genera (“body, animal, tree, man, and similar things”), while universal
concepts include also relational categories of a logical kind, which are certainly
not abstracted from sense impressions.29
Already in A Platonick Song of the Soul (1647), More referred to “th’innate
idee /​essentiall forms created with the mind,” asserting that, when God creates

25. EM III.4, 150 (MM 21). See Aristotle, Categories, 2b 15–​20.


26. On these defining features of Medieval nominalism, see Galluzzo 2011, 86. The criticism
of the reification of universals is widespread also in the seventeenth century: see Bolton 1998,
189f. and n. 56.
27. See Suárez, Disputatio Metaphysica V.vi.1; cf. Thiel 1998, 217.
28. The status of universals in Descartes is controversial: see chapters 5 and 6 in this volume.
29. The coexistence of universal concepts, obtained a posteriori by abstraction, with a priori
innate logoi is attested in Neoplatonic epistemologies:  see Steel 1997, 293–​309, especially
300–​305.
Universals in English Platonism  •  173

a human soul, he fills it “with hid forms and deep idees innate.”30 In An Antidote
against Atheism (1653), More upheld an aprioristic epistemology, stating that the
soul possesses “actual knowledge” that is not acquired by means of the senses.
Such knowledge involves geometrical, mathematical and logical notions, which
“cannot be the impresses of any material object from without,” but “are from the
soul her self within, and are the natural furniture of human understanding.”31
This kind of epistemology is propounded again in The Immortality of the Soul
(1659). Sense is only one source of knowledge; in addition to it we have the “com-
mon notions” and the “second notions,” perceived only by the mind. “Common
notions” are axioms, that is, propositions whose truth is apprehended by the sim-
ple understanding of the component terms; “second notions” are those logical
and mathematical concepts that we use to reason about the objects of the senses
(the so-​called relative ideas of An Antidote against Atheism).32
This array of general concepts, which provide a purely intellectual knowledge,
enables More to criticize Hobbes’s materialism and nominalism. More reads
Hobbes’s nominalism as a trick to defend a thoroughgoing materialism:

Here Mr. Hobbs … has found out a marvellous witty invention to befool
his followers withall, making them believe that there is no such thing as
these Secundae Notiones, distinct from the names or words whereby they
are said to be signified; and that there is no perception in us, but of such
phantasmes as are impressed from external objects, such as are common
to us and beasts.33

Since the names and images of external objects are material, insofar as they are
signs apprehended by the senses, Hobbes feels entitled to conclude “that we have
the perception of nothing but what is impressed from corporeal objects.” More
attacks Hobbes’s nominalism from a linguistic point of view: if universal notions
are only names, then the meaning of logical and mathematical notions should

30. More 1998, 463ff., 465f., 468 (Antipsychopannychia, Canto II, stanzas 22ff., 29, 36 and 37).
See also The Interpretation Generall (that is, the Glossary appended to A Platonick Song of the
Soul), where “innate idees” are said to be “the soul’s nature it self, her uniform essence, able by
her Fiat to produce this or that phantasme into act” (617).
31. More 1662 (An Antidote against Atheism I.vi.3, 18).
32. On the “common notions,” see More 1662, The Immortality of the Soul, bk. I, chap. II, axiom
III, 17; on the “secundae notiones,” ibid., bk. II, chap. 2, §9, 69. Among the “relative ideas”
More enumerates cause and effects, whole and part, like and unlike, equality and inequality,
proportion and analogy, symmetry and asymmetry (More 1662, An Antidote against Atheism
I.iv.4, 19).
33. More 1662, The Immortality of the Soul, bk. II, chap. 2, §10, 69.
1 7 4   •  The Problem of Univer sals in Early Modern philosophy

vary with the different languages, while, on the contrary, we all know that they
remain the same in different languages and cultures.34
This refutation unveils a weak point in Hobbes’s extreme nominalism. Both
in Leviathan and in De corpore, Hobbes restricts universality to names without
clarifying the conceptual basis on which the same name should be applied to dif-
ferent particulars (apart from a vague reference to the similarity in quality or in
some accident among the things which are designated by the same word). More
takes Hobbes’s statements at face value and so finds an easy way to refute him.35
More’s criticism of Hobbes shows that the purely intellectual nature of logical
notions has a metaphysical implication, since it proves that soul is an incorporeal
substance. In More’s writings we find further evidence of the metaphysical impli-
cations of universal logical notions. In the Preface to the Reader of Conjectura
Cabballistica (1662), More illustrates the thesis that the activity of the human
intellect is rooted in that of the divine mind through participation in the same
wisdom. More connects and compares the ratio stabilis of the divine mind to the
ratio mobilis of the human mind:

For what is the Divine Wisdome but that steady comprehension of the
Ideas of all things, with their mutual respects one to another, congruities
and incongruities, dependences and independences? … And what is this
but Ratio stabilis, a kinde of steady and immutable Reason discovering the
connexion of all things at once? But that in us is Ratio mobilis, or Reason
in evolution, we being able to apprehend things only in a successive man-
ner one after another. But so many as we can comprehend at a time, while
we plainly perceive and carefully view their ideas, we know how well they
fit, or how much they disagree one with another, and so prove or disprove
one thing by another: which is really a participation of that Divine Reason
in God.36

The all-​encompassing intuitive wisdom of God is reflected in the demonstra-


tive procedures of the human intellect, and this link guarantees that the order

34. Ibid.
35. For Hobbes’s nominalism see Hobbes 1991, chap. 4, 26; Hobbes 1994, I.20 (Elements of
Philosophy: The First Section Concerning Body I.ii.9). See also the discussion of Hobbes in chap-
ter 3 in this volume.
36. More 1662 (Conjectura Cabballistica, Preface to the Reader, §3, 2–​3). Cf. Crocker 2003,
70ff.: “without the prior stability of these [sc. ideas] attained in the mind of God, it would not
be possible for the human mind to relate coherently to its own notions or external ‘things’ at
all” (73).
Universals in English Platonism  •  175

of things expressed in divine ideas is understood, even if gradually, partially and


with effort, by the human mind when it develops its logical and mathematical
notions. The universal concepts that the human mind actively discovers and
elaborates within itself, and that are immutable and indelible, reflect the general
structures of the world, that is, the aspects of things and their relations.37
The nexus between human minds and the divine mind is documented also
in Enchiridion Ethicum (1667), where More theorizes the recta ratio, that is, the
transcription in the human mind of the eternal moral laws originally existing in
the divine mind. The recta ratio, which reveals to us the content of the eternal
moral laws, is communicated to men by God, or, more precisely, is insita et con-
genita in our nature.38
Finally, of importance for reconstructing More’s conception of universals
is Annotations upon the Discourse of Truth (1682). Commenting favourably on
George Rust’s criticism of voluntaristic theology, More upholds the archetypal
theory of creation and grafts onto it his view of the connection between human
and divine intellect. In the mind of God there are

certain and immutable Ideas of such and such determinate things, as


suppose of a Cylinder, a Globe and a Pyramid, which have a setled and
unalterable nature, as also immutable properties, references and aptitudes
immediately consequential thereto, and not arbitrariously added unto
them.39

This ideal world of unchangeable universal essences includes not only geometri-
cal and mathematical ideas, but also physical, moral, and metaphysical notions.
All of them have “determinate natures, with properties and aptitudes immedi-
ately issuing from them”; “their habitudes and respects” are “fixedly, determi-
nately and unalterably represented in their Ideas.” In short, what More endorses is
the conception of  “eternal Omniformity which the Platonists call the Intellectual
World”: in this realm of essences, all the determinate natures and their various
relations are “necessarily and immutably exhibited.”40 This Platonic idealism is
framed in terms of the Christian idea of creation: Platonic ideas “are the patterns

37. More 1662, 4 (Conjectura Cabalistica, The Preface to the Reader, §5).


38. More 1675–​79, 2:15 (Enchiridion Ethicum, bk. I, chap. iii, §5). On intellect and reason
as the elements which make man similar to God see ibid., 2:51 (Enchiridion Ethicum, bk. II,
chap. v, §§5–​7).
39. More 1683, 179.
40. Ibid., 179f., 254f., 265.
1 7 6   •  The Problem of Univer sals in Early Modern philosophy

or paradigms according to which every thing is made” and “all created things are
but the copies of these.”41
The same kind of relation between the archetype and the copy governs the
connection between the divine mind and the human intellect: “For as much as
the Intellect of man is as it were a small compendious Transcript of the Divine
Intellect, and we feel in a manner in our own Intellects the firmness and immuta-
bility of the Divine, and of the eternal and immutable Truths exhibited there.”42
This derivative nexus—​conceived in terms of an imperfect relation, namely, the
one a copy bears to its archetype—​accounts for the existence in the human mind
both of universal concepts as abstractions from experience and of universal con-
cepts as a priori innate ideas: the first are the result of the limitations of human
understanding, the last are witnesses to the divine origin of the human mind.
More’s Annotations also deals with the problematic relation between the
divine Mind and the objective truth of universal essences. More denies the
hypostatization of the essences, insisting instead on the priority of the active
understanding of God over objective truth. Truth cannot exist before the Divine
Understanding:

The Divine Understanding being before all things, how could there be any
truth before it, there being neither Understanding nor Things in which
this Truth might reside? Or the Divine Understanding be a mere passive
Principle actuated by something without, as the eye by the Sun. [On the
contrary] it is an eternal, necessary and immutable Energy, whose very
essences is a true and fixt Ideal Representation of the natures of all things,
with their respects and habitudes resulting eternally from the Divine foe-
cundity at once.43

To complete this exposition of More’s theory of universals we must ask if his


“nominalistic” thesis (each being is individual and universals are only modi cogi-
tandi) is compatible with his archetypal theory of ideas as eternal objects and
patterns of creation in the mind of God. My answer is that the compatibility
lies in a constitutive original aspect of Platonism: the relation between ideas and
sensible objects is at the same time a relation of méthexis (participation) and of
chorismós (separation). The objects existing in the spatial and temporal frame of

41. Ibid., 266f. (italics in the original text). See also 263: “the eternal and immutable reasons of
things are originally and paradigmatically in the divine understanding, of which those in the
creatures are but the types and transitorie shadows.”
42. Ibid., 257f.
43. Ibid., 259.
Universals in English Platonism  •  177

the world cannot mirror perfectly the archetypal models: therefore particulars


cannot be interpreted as bundles of universals, since the eide are not immanent
and perfectly instantiated in the sensible world. Universals provide us with the
conceptual apparatus by means of which the world becomes intelligible, but
the individuality of beings is rooted in the ontological gap between eide and
phaenomena. This is especially so in Christian Platonism, in which the eide have
become noémata of the divine Mind, and the particulars are conceived as singular
creations of a personal Being.

8.2. Ralph Cudworth
Cudworth is a conceptualist, too, and holds an aprioristic view of knowledge
based on the assumption that universal essences conceived by the human mind
correspond to thoughts in the mind of God. The metaphysical foundation of
Cudworth’s gnoseology lies in the thesis that imperfect human minds are ectypic
derivations of the eternal and uncreated mind of God.44 This assumption is
required to explain the origin of human mind: “since no mind could spring out
of dead and sensless matter, and all minds could not possibly be made, nor one
produced from another infinitely; there must of necessity be an eternal unmade
mind, from whence those imperfect minds of ours were derived.”45 The human
mind, as the creative manifestation of the divine archetypal mind, is endowed
with the faculty of acquiring knowledge by means of conceptions of universals
virtually contained within the soul and unfolded on the occasion of external
sensations.46
Cudworth’s aprioristic theory of knowledge is directed against Hobbesian
sensism and nominalism:  knowledge, even of sensible and material objects,
is obtained not by sensory images (phancies) and adventitious ideas, passively
received,

but by intelligible ideas exerted from the Mind itself, that is, by something
native and domestick to it. … Wherefore besides the phantasms of sin-
gular bodies, or of sensible things existing without us, … it is plain that

44. Cf. EIM IV.iv,7, 292: “all particular created minds being but derivative participations of one
infinite eternal mind, which is antecedent to all corporeal things”; cf. also EIM IV.i.5, 134f.;
IV.iv.11–​12, 256f.
45. TIS 729f.
46. “[Cudworth’s] view as to the ‘potential omniformity’ of the mind presents the mind as
a substance, whose modes as they gradually unfold, embrace the essential reality of all being”
(De Boer 1931, 142; cf. also 87).
1 7 8   •  The Problem of Univer sals in Early Modern philosophy

our humane mind hath other cogitations or conceptions in it, namely the
ideas of the intelligible natures and essences of things, which are universal,
and by and under which it understands singulars. It is a ridiculous conceit
of a modern atheistick writer, that universals are nothing else but names,
attributed to many singular bodies, because whatsoever is is singular. For
though whatsoever exist without the mind, be singular, yet it is plain, that
there are conceptions in our minds, objectively universal. Which universal
objects of our mind, though they exist not as such any where without it,
yet are they not therefore nothing, but have an intelligible entity for this
very reason, because they are conceivable, for since non-​entity is not con-
ceivable, whatsoever is conceivable, and an object of the mind is therefore
something.47

Cudworth shares with More the idea that nominalism is a device to which
Hobbes resorted in order to support his materialistic monism. To deny the exis-
tence of immaterial substances, Hobbes had to reduce all knowledge to the pas-
sive impressions of the senses; for this reason he denied that there are in the mind
universal ideas and claimed that “those things which are called universal, are
nothing else but names applied to several individuals.”48 This implies that science
can never go beyond the sphere of sensible individual things and consists only “in
making use of common names to express several individuals by at once.”49 Such
extreme nominalism is untenable because, if we consider geometrical knowledge,
we have evidence of a set of axioms and theorems universally true, that is, per-
manent and immutable, which cannot be derived from any sensible image and
must be the product of the mind. Any attempt to explain mathematical notions
as originating from sensory impressions is bound to fail “because there never was
any material or sensible straight line, triangle, circle, that we saw in all our lives,
that was mathematically exact.”50
In rejecting nominalism, Cudworth appeals to the Aristotelian thesis that we
have science only of the universal, while individual things are objects of sense

47. TIS 731; cf. EIM IV.iii.15, 223–​26. On the proleptical nature of human knowledge (singu-
lars are understood by means of universal conceptions which are not derived from singulars “in
way of ascent”): TIS 732. See also EIM IV.iii.13, 218f. See De Boer 1931, 139.
48. EIM IV.iii.15, 223.
49. EIM IV.iii.15, 225f.
50. EIM IV.iii.7, 203. Cudworth regarded mathematics as the paradigm of clear intelligibility,
certainty, universality, necessity, and perfection: see De Boer 1931, 146. The same argument
appears in Norris 1722, 1: 51–​56 (pt. I, chap. 2, §§ 30–​32), to demonstrate the existence of an
ideal world as the object of mathematical knowledge.
Universals in English Platonism  •  179

perception, and he combines this thesis with Descartes’s aprioristic rationalism.51


In A Treatise Concerning an eternal and immutable Morality, this antithetical
scheme is repeated over and over again:  A.  sense—​individual bodies—​things
existing outside the mind—​a posteriori apprehensions; B. intellect (or mind or
ratio)—​universal essences—​inwardly present in the mind—​a priori knowledge.
This antithesis is coupled with the Cartesian view that only the mind by its a
priori ideas provides proper knowledge of individual bodies: “[the soul] exerts
from within the intelligible ideas of things, virtually contained in its own cog-
noscitive power, that are universal and abstract notions, from which, as it were
looking downward it comprehends individual things.”52
Although Cudworth rejects Hobbes’s nominalism, it is noteworthy that
he accepts it from a limited ontological perspective with regard to corporeal
substances:

There is nothing in the world (saith a late Author [viz., Hobbes]) univer-
sal, but names; for the things named are every one of them individual and
singular. Now indeed this is true, and no body denies it, of things existing
without the mind; but this author’s meaning herein is to deny all universal
conceptions and reasonings of the mind.53

Existing bodies are singular while universals are only concepts; however, univer-
sal concepts are both human and divine, and not obtained by abstraction from
sensory perceptions, but virtually contained in the minds of men because these
reflect divine wisdom; therefore, against this background, universal concepts
possess a prominent ontological relevance.
Cudworth identifies as “universal conceptions” a wide array of ideas, includ-
ing ethical values, ideas of geometry and mathematics, logical categories, and
all sorts of “abstract notions” concerning the activity of the mind and relations
occurring between corporeal or incorporeal substances.54 In The True Intellectual
System of the Universe, Cudworth remarks that from the point of view of the
human mind and with regard to worldly beings, universal concepts describe a

51.  EIM IV.iii.17–​20, 229–​35. Cudworth quotes extensively from Posterior Analytics I, 25,
for the opposition between episteme, which regards katholou, and aisthesis, which concerns
kath’ekasthon (EIM IV.iii.18, 230f.); he refers also to Descartes, Meditatio Secunda (AT 7:34),
for the thesis that, even if bodies are perceived by the senses, only the intellect understands
them with clear and distinct ideas (EIM III.iii.3, 97).
52. EIM IV.iii.13, 217f.; see also III.iii.2, 192; IV.iii.3, 194f.; IV.iii.13, 219; IV.iii.18, 232.
53. EIM IV.iii.15, 223f. Cudworth quotes De corpore, I.9 (Hobbes 1966, 1:17–​18).
54. For a list of universal concepts, see TIS 732; EIM IV.i.8, 140f.; IV.ii.1, 149.
1 8 0   •  The Problem of Univer sals in Early Modern philosophy

realm of possibilities. But what is possible, he argues, implies an actual reality as


its ultimate foundation. This means that ideal possibilities subsist in the divine
mind. The human mind is capable of conceiving them, transcending prolepti-
cally empirical reality, thanks to its being an ectypal derivation of the archetypal
divine mind, which guarantees the eternal verity of ideas. “Eternal verities,” which
Cudworth defines as “common notions,” have a degree of reality that goes beyond
their conceivability by the human mind, because they subsist in God’s mind as
rationes rerum. The immutability of eternal verities is proof of their necessary
existence in a sphere beyond sensory experience. With regard to the world of the
senses, “eternal truths” are “no where,” but they exist in a mind that knows them
eternally, “since these truths and intelligible essences of things cannot possibly be
any where but in a Mind.” Eternal truths are thoughts of the divine mind, that is,
of “Infinite Power and Understanding.”55
Cudworth examines the ontological import of universal ideas repeatedly in
A Treatise Concerning an eternal and immutable Morality. Starting from con-
cepts of “relations,” such as “cause and effect, whole and parts and the like,” he
admits that they should be regarded as “meer notions of the mind and modes of
conceiving in us”; however, this admission does not amount to saying that these
intellectual relations “had no reality at all, but were absolute non entities.” For
Cudworth, “intellect [is] a real thing,” and is ontologically superior to matter;
therefore, modifications of intellect are even more real than modifications of
matter.56 Besides this purely metaphysical argument, based on the hierarchical
dualism of incorporeal and corporeal substances, Cudworth offers a pragmatic
argument, which refers to the causal efficacy of art and science, to show that there
is a correspondence between our logical relational concepts and extramental real-
ity. Human art and wisdom “beget real effects of the greatest moment and conse-
quence in nature and human life,” and this proves that intellectual notions are not
“figments of the mind, without any fundamental reality in the things themselves.”
It is because science concerns “the relations, proportions, aptitudes of things to
one another, and to certain ends” that we acquire power over external natural
things.57 The power of science is made possible by the fact that “even the strength
and ability of corporeal things themselves depends upon the mutual relations and
proportions of one thing to another.” Science has causal efficacy because it grasps
an ideal order that explains the functioning of mechanical powers.58 Explaining

55. TIS 734 and 736.


56. EIM IV.ii.5, 156.
57. EIM, IV.ii.5, 157f.
58. EIM IV.ii.7, 161f.
Universals in English Platonism  •  181

how the automatism of a clock depends upon its ideal structure, Cudworth
stresses the objectivity of universal concepts to the point of apparently contra-
dicting his acceptance of the nominalistic principle that all corporeal entities are
individual:

I say … , returning to our former instance of an Automaton, or Horloge,


that though those several relative ideas of cause, effect, symmetry, pro-
portion, order, whole and part, and the like, considered formally as
conceptions of the mind, be only in the intellect it self (as the ideas and
conceptions of all other things likewise are;) yet notwithstanding the
intellect doth not forge or falsify any thing in apprehending of them, in
that material self-​mover represented to it by sense, because all the several
relations are fundamentally and really in the same, though they could not
be stamped upon sense materially, and received passively from it. And
therefore that the true nature, formal reason, essence and idea of this self-​
mover, watch or horloge, is really compounded and made up of those sev-
eral relations, as ingredients into it, so that it cannot possibly be understood
without them.59

Here it would seem that universals are in re, as forms ordering sensible matter,
and not simply ante rem, as thoughts in the mind of God or as a priori ideas in
the human mind. Cudworth makes recourse to the old metaphor of the book
of nature, that is, the material and visible universe printed with the characters
of divine wisdom; this book is not legible to the senses, but only to the intel-
lect, since intellect “hath an inward and active participation of the same divine
wisdom that made it.” The book of nature and the human mind derive from the
same “archetypal seal,” which has impressed objective features on the things in the
world and the corresponding a priori ideas on the human mind. This metaphori-
cal picture confirms the impression that universals are both ideas in the mind and
immanent forms in the world. However, following the standard Platonic view
Cudworth claims that material bodies do not correspond perfectly with mathe-
matical notions, but exhibit only rough resemblances to them:

when we look upon the rude, imperfect and irregular figures of some
corporeal things, the mind upon this occasion excites from within it self
the ideas of a perfect triangle, square, circle, pyramid, cube, sphere, and
the like, whose essences are so indivisible, that they are not capable of the

59. EIM IV.ii.8, 162f. (my italics).


1 8 2   •  The Problem of Univer sals in Early Modern philosophy

least additions, detraction or variation without the destruction of them,


because there was some rude and bungling resemblance of these regular
figures in those material objects that we look upon, of which probably the
Maker had the ideas in his mind.60

Although the natural order is conceivable in terms of geometrical universal con-


cepts, material bodies maintain their individual nature since they are not perfect
expressions of mathematical essences.
In a later chapter, Cudworth openly poses the question: “where [do] these
immutable entities … exist?”61 His answer is typical of Christian Platonism: uni-
versal intelligibles do not exist as immanent forms in bodies, because bod-
ies are transient and ever-​changing realities; they exist as innate ideas in the
human mind and as noémata in the mind of God. Aristotelian immanentism is
rejected thus:

I answer, first, that as they [sc. the universal rationes] are considered for-
mally, they do not properly exist in the individuals without us … because
no individual material thing is either universal or immutable. And if these
things were only lodged in the individual sensibles, then they would be
unavoidably obnoxious to the fluctuating waves … in which all individ-
ual material things are perpetually whirled. But because they perish not
together with them, it is a certain argument that they exist independently
upon them.62

Cudworth denies universals in re, while asserting that the material universe is
pervaded with an ideal structure intelligible only to the understanding as an
imperfect manifestation of the universal rationes rerum. Universals are inside
matter and constitute nature but only in a derivative and imperfect way, and mat-
ter is not their unique or proper “place” of existence. From a formal point of view,
that is, with regard to their metaphysical essence, universals exist independently

60. EIM, IV.iii.6, 201f. In a couple of passages (EIM IV.iii.8, 204; IV.iii.17, 228f.) Cudworth
does not exclude that divine wisdom might have created perfect geometrical shapes in some
suitable materials, but he draws the conclusion that, whatever the case, knowledge of the invari-
able notions of mathematics and geometry cannot be drawn through abstraction from sense
experience, but is discovered by the a priori reflection of the mind, since “sense could not at
all reach to the discerning of the mathematical accuracy of these things.” These passages have a
Cartesian antecedent in Fifth Replies (AT 7:381).
61. EIM IV.iv.4, 243.
62. EIM IV.iv.4, 243f.
Universals in English Platonism  •  183

of individuals. Having thus excluded, following Plato himself,63 the position that
universals are perfectly instantiated in re, Cudworth also rejects the essentialism
that Aristotle criticized in Plato: “Neither in the next place, do they exist some-
where else apart from the individual sensibles, and without the Mind, which is
that opinion that Aristotle justly condemns, but either unjustly or unskilfully
attributes to Plato.” To deny the separate existence of universals, Cudworth
argues that mind knows intelligible essences not looking outward, in a passive
contemplation of external truths, but exerting actively its own inward ideas. From
this psychological consideration he feels entitled to conclude that “these intel-
ligible ideas or essences of things, those forms by which we understand all things,
exist no where but in the Mind it self.”64 As soon as Cudworth has stated that
universals are noémata, and so has rejected their substantial hypostatization, he
worries about dispelling a possible subjectivistic interpretation. Even if noémata
exist only in the mind, they are not “figments,” that is “they are not arbitrarily
framed by the mind, but have certain, determinate, and immutable natures of
their own, which are independent upon the mind.”65 If universals are, on the one
hand, noémata and, on the other, “immutable and eternal objects of science,” their
constant and necessary being cannot be bound either to the material world or to
the particular human minds that think them. To assert the eternal being of uni-
versals, Cudworth uses an annihilatory hypothesis antithetical to the famous one
employed by Hobbes in De Corpore for his phenomenalistic approach to science:

These things have a constant being, when our particular created minds
do not actually think of them, and therefore they are immutable … , not
only because they are indivisibly the same when we think of them, but
also because they have a constant and never-​failing entity; and always are,
whether our particular minds think of them or not. … all these ratio-
nes and verities had a real and actuall entity before, and would continue
still, though all the geometricians in the world were quite extinct, and no
man knew them or thought of them. Nay, though all the material world
were quite swept away, and also all particular created minds annihilated
together with it; yet there is no doubt but the intelligible natures or
essences of all geometrical figures, and the necessary verities belonging to
them, would notwithstanding remain safe and sound. Wherefore these

63. For the thesis that eide cannot exist in the mutability of the sensory world and that the par-
ticipation of bodies in eide implies also their separation, see Plato, Symp. 211a–​b, Ti. 51d–​52a,
Prm. 130b, 133c–​d.
64. EIM IV.iv.4, 244.
65. EIM IV.iv.4, 245.
1 8 4   •  The Problem of Univer sals in Early Modern philosophy

things had a being also before the material world and all particular intel-
lects were created. 66

Cudworth’s theory has to harmonize two strong constraints: the objective immu-


table content of intelligible essences and at the same time their being “nothing
but noemata,” that is, objective notions that cannot exist alone, but stand always
“together with that actual knowledge in which they are comprehended.” To sat-
isfy these two requirements, Cudworth asserts that universals existed before the
creation of the world, including human souls, as thoughts conceived ab aeterno in
the mind of God, a solution perfectly consistent with the main goal of Cudworth’s
apologetic philosophy:

It is all one to affirm, that there are eternal rationes, essences of things,
and verities necessarily existing, and to say that there is an Infinite,
Omnipotent and Eternal Mind, necessarily existing, that always actually
comprehendeth himself, the essences of all things, and their verities; or
rather, which is the rationes, essences, and verities of all things.67

Cudworth’s solution is nothing new, since it interprets Platonic eide as thoughts in


the mind of God according to a well-​established tradition that stems from Philo
Judaeus, and runs via Augustine, through Medieval philosophy, to Renaissance
Christian Platonism. Cudworth interprets this tradition in a particular way,
emphasizing the priority of the thinking mind as active energy, whose objects
are its own inner reflections. The first principle in Cudworth’s metaphysics is the
divine mind as a living active energy, and this rules out the conception of eter-
nal truths as reified essences, existing in themselves. Echoing Plotinian themes,68
Cudworth binds eternal essences to the thinking activity of the divine mind:

To prevent all Mistake, I shall again remember … that where it is affirmed


that the Essences of all things are eternal and immutable; … this is only
to be understood of the intelligible essences and rationes of things, as they
are the objects of the mind: and that there neither is nor can be any other
meaning of it, than this, that there is an eternal knowledge and wisdom, or
an eternal mind or intellect, which comprehends within it self the steady

66. EIM IV.iv.5, 248f. Cf. TIS 734. On Hobbes’s annihilatory hypothesis, see De corpore II.vii.1.
67. EIM IV.iv.7, 251. See also EIM IV.i.5, 134; TIS 734.
68. Plotinus, Enn. 5.9.8, where being and intellect (nous) are conceived as unified under the
concept of activity (energheia). See Breteau 1995, 19; Emilsson 2007, 137f.
Universals in English Platonism  •  185

and immutable rationes of all things and their verities, from which all par-
ticular intellects are derived, and on which they do depend. But not that
the constitutive essences of all individual created things were eternal and
uncreated, as if God in creating of the world, did nothing else, but as some
sarcastically express it, sartoris instar rerum essentias vestire existentia, only
cloathed the eternal, increated, and antecedent essences of things with a
new outside garment of existence, and not created the whole of them.69

Founded on this metaphysics, Cudworth’s theory of knowledge highlights the


dynamism of mind and is opposed not only to sensism but also to extreme inna-
tism, since both these conceptions share the common view of the passivity of
mind on which, as on a blank slate, either sensible data or ideal notions would be
impressed.70
To complete our understanding of Cudworth’s approach to universals, we
also need to consider the influence of Descartes’s philosophy. In the latter’s writ-
ings, Cudworth found statements that he easily adopted: in the Fifth Meditation,
Descartes stated that a priori ideas have “true and immutable natures,” and are
not invented by the human mind nor, in this sense, are dependent on the human
mind;71 from the beginning of the Sixth Meditation, Cudworth takes the ratio-
nalistic thesis that geometrical ideas, being general, are understood by the intel-
lect independently from their being represented by the imagination:

The mind can clearly understand a triangle in general, without determin-


ing its thought to any particular species, and yet there can be no distinct
phantasm of any such thing; for every distinct phantasm or sensible pic-
ture of a triangle must of necessity be either equilateral or equicrural,
or inequilateral, uneven-​legged. And so as we can in like manner clearly
understand in our minds a thing with a thousand corners, or one with ten

69. EIM IV.iv.2, 284f. Cf. TIS 736. Here Cudworth echoes a sentence by Gassendi, who, in his
Fifth Objections (AT 7:319), had mistakenly attributed to Descartes what he believed to be the
scholastic doctrine of the eternal essences of things “independent of God.” (I thank Stefano
Di Bella for pointing out this passage to me.)
70. See Muirhead’s judgement of Cudworth’s position: “while essentially Platonic in insisting
on the presence of universals in all knowledge, he sees the futility both of mediaeval discus-
sions as to their separate existence and of facile appeals in his contemporaries to innate ideas”
(Muirhead 1931, 46). On Cudworth’s virtual innatism and his criticism of Plato’s doctrine of
anamnesis, see Lotti 2004, 152.
71. Fifth Meditation, AT 7:64: Cudworth refers to this passage writing that “Descartes affir-
meth that the essences of things were eternal and immutable” (EIM I.iii.3, 31).
1 8 6   •  The Problem of Univer sals in Early Modern philosophy

thousand corners, though we cannot possibly have a distinct phantasm of


either of them.72

In the Fifth Replies, Cudworth read that Descartes reasserted the “eternal and immu-
table” nature of geometrical truths, that geometrical ideas are not drawn (desumptas)
from singular existing things, and do not enter into our mind through the senses, but
are conceived a priori in our intellect.73
Cudworth uses the dualistic pattern of Cartesian ontology, centered on the
opposition between mind and matter and on the two categories of substance
and mode, to refute the hypostatization of essences. Within a Cartesian onto-
logical scheme, Cudworth argues that intelligible essences of things and eternal
truths cannot exist without the existence of a substantial mind. He reasons thus:
essences and eternal truths “must of necessity be either substances, or modifica-
tions of substance—​for what is neither substance nor modification of substance,
is a pure non-​entity”; “if they be modifications of substance, they cannot possi-
bly exist without that substance whose modifications they are”; “[this substance]
must either be Matter or Mind”; “they are not modifications of matter as such,
because they are universal and immutable”; “therefore they are the modifications
of some mind or intellect.” Conclusion: “these cannot be eternal without an
eternal mind.”74 Cudworth does not even examine the possibility that rationes
rerum are substances. He excludes this from the start, assuming that they “are
nothing but noemata,” and so are “modifications of some mind or intellect.”75 The
Cartesian ontological scheme has served to illustrate that, on the assumption that
universal essences are notions, they can only be the modes of immaterial sub-
stance (ideas in the Mind of God and in the human mind). In fact, the real theo-
logical motivation for denying that Platonic ideas might be substances is that, if
they were substances, this would imply a fragmentation of God, and thus would
provide the philosophical justification for polytheism, a charge which Cudworth
brings against some Neoplatonists in The True Intellectual System.76

72. EIM IV.iii.12, 212: cf. AT 7:72f., with the same references to the chiliogonum and the
myriogonum. Here Cudworth offers an argument that could be used as a rationalistic reply to
Berkeley’s critique of abstract ideas.
73. AT 7:380, ll. 23–​382, l. 24. See EIM IV.iii.6–​10, 199–​208, which repeats the same examples
provided by Descartes (compare AT 7:382, ll. 13–​19 and EIM IV.iii.6, 201).
74. EIM IV.iv.9, 252f.
75. EIM IV.iv.7, 250f.
76. TIS 553–​54. See also TIS 736, where the issue is formulated in more general terms: eternal
truths cannot be substances because, if the quality of being eternal were conceived of their
Universals in English Platonism  •  187

Whatever affinity Cudworth might have glimpsed between the Platonic and
the Cartesian conceptions of the a priori ideas of immutable essences and of
mathematical truths, there is undoubtedly a topic upon which Cudworth and
Descartes diverge: the Cambridge Platonist was hostile to every form of volun-
taristic theology, and was extremely concerned about the Cartesian doctrine of
the creation of eternal truths, which he defined as a “metaphysical way of trans-
formation of essences, by meer will and command.”77 One statement summarizes
Cudworth’s opposition to voluntaristic theology: “things are what they are, not
by Will but by Nature.”78
Cudworth presents a series of arguments to support his critique of volunta-
ristic theology. He remarks that divine omnipotence cannot be invoked to over-
throw the stability of essences, because even God’s will cannot transgress the
principles of identity and contradiction. This principle is deemed to stand firm
despite God’s omnipotence, because it “is a truth fundamentally necessary to all
knowledge,” without which “nothing would be certainly true or false.”79 The inde-
pendence of eternal truths and essences from the will of God does not imply “that
something that was not God was independent upon God”—​as the voluntarists
insinuated critically—​but means simply that “there is an eternal and immutable
wisdom in the mind of God, and thence participated by created beings indepen-
dent of the will of God.”80 Theologically, the wisdom of God has to be regarded
as superior to God’s will, since wisdom, as a rule and measure, is determined and
unchangeable, while will, in itself, is indefinite, blind and dark, and therefore
apt to be regulated by wisdom. The Cartesian doctrine of the creation of eternal
truths destroys “the definite natures or notions of things” and so makes all kinds
of knowledge impossible. If God could change the essences of things arbitrarily,
not only human science and demonstration would vanish—​since mathematical
and metaphysical knowledge would be at the mercy of irrational ways like fanati-
cal faith and enthusiastic inspiration—​but God’s knowledge and wisdom, having
been made dependent on God’s mutable will, would be lost.81 In his sustained
criticism of the Cartesian doctrine of the creation of eternal truths, Cudworth
did not take into account two qualifications that Descartes had made in order to

substantial esse, and not simply of their esse cognitum, then “everything were in it self eternal
and uncreated.”
77. EIM II.iii.4, 32.
78. EIM I.ii.1, 14. See also EIM I.ii.1–​2, 16.
79. EIM I.ii.1, 15.
80. EIM I.iii.1, 27, and I.iii.7, 34f.
81. EIM I.iii.4–​6, 32–​34.
1 8 8   •  The Problem of Univer sals in Early Modern philosophy

neutralize its potential dangerousness: Descartes had written to Mersenne that


the eternal and immutable truths remained steady because of the immutability
of God’s will and that in God there is no distinction between will and intellect.82
I take Cudworth’s view of universals to be a kind of metaphysical concep-
tualism, and thus distinguish it from the epistemological conceptualism that
was to be elaborated by Locke a few years later. Cudworth has several polemical
targets according to which he highlights different features of his theory: against
the hypostatization of essences, he emphasizes that universals are exclusively noé-
mata; against the Aristotelian theory of immanence, he claims that bodies can be
only individual because the metaphysical properties of universals and of bodies
are incompatible; against materialism and nominalism, he shows that universals
exist as intelligible immutable essences conceived by the human mind and corre-
sponding to the ontological structure of reality; against voluntaristic theology, he
maintains that universals are rooted in divine wisdom and are immutable facets
of eternal truth.

8.3. John Norris
Norris discusses universals in a Letter concerning the true notion of Plato’s Ideas,
and of Platonic Love (1684), an early essay in which he adopts an approach very
similar to that of Cudworth, denying that Plato had upheld the existence of uni-
versal ideas separated from the divine mind:  a monstrous opinion ascribed to
Plato by Aristotle who “wanted a shadow to fight with.” Norris interprets Platonic
philosophy in a Christianized way, along the lines of Augustine: he attributes to
Plato the doctrine that the world is the product of an intellectual agent, and that
ideas are “eternal forms, models or patterns” existing in the mind of God before
creation.83
Ideas are the causes of things and the paradigms of truth, but this does not
confer upon them an absolute independent reality. There are no ideal essences dis-
tinct from the divine essence, and the existence of Platonic ideas consists in their
being objects of God’s thinking as facets of the divine essence. Consequently, cre-
ated beings, by participating in the ideal forms, participate also in divine essence.

82.  See TIS 646 and Descartes’s letters to Mersenne of April 15, May 6 and 27, 1630:  AT
1:145–​4 6, 149–​50, 151–​53. Cudworth did not realize or, better, was not convinced that—​
as modern scholarship suggests—​Descartes’s strategy consisted in positing an “all powerful
but non-​threatening God, i.e., in insulating claims about God’s power from what might be
thought to be their standard implications concerning the features of created beings” (Della
Rocca 2002, 73.)
83. Norris 1687, 436–​38; cf. MacKinnon 1910, 5; Acworth 1975, 1:32–​37; Mander 2008, 51.
Universals in English Platonism  •  189

In the Letter Norris introduces a distinction between the mind of God as “exhibi-
tive” (i.e., intelligible) and as “conceptive” (i.e., intelligent), which he takes from
More and which he will maintain also in later writings:

By the Mind of God Exhibitive is meant the Essence of God as thus or


thus imitable or participable by any Creature, and this is the same with
an Idea. By the Mind of God Conceptive is meant a reflex act of God’s
Understanding upon his own Essence as Exhibitive, or as thus and thus
imitable.84

The purpose of this distinction is to articulate the relations among God’s mind,
its object and the truth, and so answer the question of whether an ideal truth
is such because God was pleased to understand it as such or, vice versa, if God
understands it as a truth because it is a truth of its own nature. If we consider
the mind of God as conceptive, the object is not created but presupposed by the
divine mind and the truth depends on the conformity of the mind to its object,
that is, to the ideas in the mind as exhibitive. If, on the other hand, we consider
divine understanding as exhibitive, the truth depends on the conformity of the
object to the mind, in the sense that the object (meaning by “object” the distinct
natures of things) does not impose its being on the divine mind as if it preexisted
independently of it, but exists only because the essence of the divine mind can be
imitated in various ways in the world of creation. In the Letter and in his major
work, An Essay Towards The Theory of The Ideal or Intelligible World (1701–​04),
Norris employs this distinction to set forth what he considers a correct and bal-
anced conception of eternal truths, which are neither entirely independent of the
divine mind, nor, at the same time, its products, in contrast with the Cartesian
doctrine of the creation of eternal truths.
In Reason and Religion (1689), Norris upholds, against the peripatetic tradi-
tion, the view that all universals “or abstract essences are really distinct from, and
exist out of those singulars whose essences they are.” Their separate existence is
an ideal (not a natural) existence and has to be acknowledged to account for the
stability of science: “Things must exist in Idea before they do in Nature, other-
wise ’twill be impossible to give an intelligible account of the stability of Science,
and of propositions of Eternal Truth.”85 Propositions stating immutable truths,

84. Norris 1687, 440. Cf. Norris 1722, 1:357–​59. Henry More has presented this distinction
in one of his annotations to the Discourse of Truth (1682) by George Rust: see More 1682,
178–​79, 246–​49. Cf. Mander 2008, 53.
85.  Norris 1689, 23f. (cf. 76f.). For an analysis of “eternal truths” in Norris, see Mander
2008, 22–​34.
1 9 0   •  The Problem of Univer sals in Early Modern philosophy

of which science consists, would become meaningless if we did not postulate a


world of intelligible essences eternally subsisting.86
To say that the eternal permanence of the simple essences of things is ideal
and not natural is to say that eternal essences do not exist outside of an intellect.
Eternal essences constitute an archetypal world in which “all the Rationes rerum,
or simple essences of things have an eternal and immutable existence, before ever
they enter upon the Stage of Nature.”87 The intellect in which essences subsist is
eternal as they are, that is, is the divine intellect. Since God is a simple being, there
is nothing in God that is other than God; therefore eternal essences cannot exist
in God either as accidents inhering to God’s nature, or as beings distinct from
God. Norris concludes that “these simple essences of things … can be nothing
else but the Divine Essence it self.”88 The diversity of essences (their difference in
species) indicates the different ways in which divine essence may be participated
in or imitated. Resorting to the Platonic concepts of mimesis (imitation) and
méthexis (participation), which Malebranche too had embodied in his doctrine of
the vision of ideas in God, Norris tries to explain the multiplicity of the intelligible
world in the divine mind, that is, the various possibilities in which finite beings
manifest the divine essence.89
This Neoplatonic approach to the problem of universals persists in An
Essay, in which Norris blends Malebranche’s philosophy with his Platonic and
Augustinian views. The metaphysical system of the ideal world, reinstated in its
pristine glory against Locke’s subjective theory of ideas, is the system of universal
reason and truth: “A world simple in its variety, and various in its simplicity, infi-
nite in its store and fullness, and stored with incorruptible and unfading treasures,
universal in its presence, and uncircumscribed by any limit of time or place, the
genuine country of truth, and its proper native soil, the place of spirits, the living
and ever springing fountain of intelligence.”90 In An Essay we again find the epis-
temological argument: the certainty of science requires immutable objects, that
is, “ideal reasons or essences” and only intellectual intuition and reasoning lead us
to universal truth, while sensible perception is limited to singular and individual

86. See Cassirer 1922, 220f. For an exposition of the several arguments employed by Norris to
demonstrate the existence of the ideal world, see Mander 2008, 18–​37.
87. Norris 1689, 80f.; Norris 1722, 1:231: “by the divine ideas we are to understand originary
and archetypal forms, representing things not as images do their originals … but contrarywise
as originals do their images or pictures.”
88. Norris 1689, 81f.
89. Ibid., 82.
90. Norris 1722, 1:10.
Universals in English Platonism  •  191

matter.91 Contemplative sciences, like geometry and metaphysics, obey the prin-
ciple that “contemplation does not make, but suppose its object.” Therefore the
ideal system that these sciences outline is a system of objective truths irreducible
to the subjective science of the mind.92 From Aristotle’s assumption that the uni-
versal is the only object of science, Norris draws the negation both of Locke’s
thesis that universals are merely human concepts and of Aristotle’s own thesis
that they are forms immanent in the natural world. Universals, as immutable
objects of science, cannot be “only Abstractions or Inadequate Conceptions in
the Mind,” because, if they were, they would be “as Contingent and Mutable, as
the Mind it self is wherein they are form’d.” They cannot exist in natural things
either, because, if they had this sort of existence, in that case too “they will be as
Contingent and Mutable as those things, and so no more capable of being the
Objects of Science than the things themselves are.”93 The only option left is that
universals exist not in a natural, but in an ideal, state. Of what precisely does this
ideal existence consist?
Ideas, Norris upholds, are not “Abstract and Universal Forms or Essences
separately existing from the divine nature,” but are “Exemplaria rerum in mente
divina, the Original Forms or Pattern of things in the Divine Understanding.”94
As God and the Ideal World share identical properties—​they are eternal, neces-
sary and omnipresent—​it is impossible to admit the existence of Ideas outside of
God. The presupposition of this argument is that, if we assume an Ideal World
separate from God, we would face a theologically unacceptable dilemma: either
God would be imperfect (that is, he would be deprived of some of the proper-
ties of the Ideal World) or we should admit two Gods, two realities endowed
with the same divine attributes.95 Thus Norris concludes that the ideal state of
universals or eternal essences must coincide with the eternal being of God. To
this conclusion Norris links the Malebranchean doctrine of the vision in God
and the Augustinian doctrine of divine illumination.96 Norris’s argument is sum-
marized thus:  “The Ideas whereby we understand are Necessary, Eternal and

91. See ibid., 1:192–​95. For an explicit reference to Cudworth, ibid., 1:401.


92. Ibid., 1:127–​29. See also 1:70f.
93. Ibid., 1:130f. See also 2:406.
94. Ibid. 1:138f. Norris attributes his position to Plato and quotes Steuchus as a reliable source
on this topic. See Steuchus 1542, l. I, cap. XII, 33.
95. See Norris 1722, 1:137–​39; cf. Mackinnon 1910, 36ff.; Acworth 1975, 599–​602.
96. Norris 1722, 2:406f. See also 1:233 on the Augustinian definition of Ideas as creaturarum
rationes incommutabiles. For more on Malebranche’s doctrine of the vision in God, see chapter 7
in this volume.
1 9 2   •  The Problem of Univer sals in Early Modern philosophy

Immutable: but there are no ideas so qualified but the divine; therefore the ideas
whereby we understand are the divine ideas.”97 This doctrine is linked also to
Norris’s rejection of innatism, which differentiates his theory of universals from
that of the Cambridge Platonists.98
Norris is in accord with both Cudworth and Malebranche in offering a
criticism of the Cartesian doctrine of the creation of eternal truths; that eternal
essences are ideas of the mind of God does not imply that they are dependent on
the arbitrary creative will of God. Eternal truths

are no creatures of ours nor yet of God’s neither. God is what he is and
they are what they are, and they can neither not be, nor be otherwise, but
remain in their steady and immutable order, the same yesterday, to-​day,
and for ever, without any Variableness, or shadow of turning. For they
depend not upon any Will, nor any Understanding, not even the Divine
itself.99

Divine ideas are not products but aspects of the eternal wisdom of God. Norris
formulates the thesis that “Eternal Truths are really the same with the Divine
Essence.”100 This strong statement places universals at the core of Norris’s philo-
sophical theology and, as a consequence, when the soul apprehends eternal truths,
it sees the mind of God and is united with the Creator, though, to preserve the dif-
ference between the divine mind and human intellect, Norris takes care to specify
that we do not see God as he sees himself. This distinction, which avoids a mysti-
cal outcome, is justified on the assumption that divine ideas “are not the essence
of God purely and absolutely as it is in itself; but as it is in relation to Creatures,
according to the several degrees of its participability whereby it is communica-
ble to them.”101 With the same conceptual device Norris tries to resolve another
theological difficulty:  How is the simplicity of God’s essence to be reconciled

97. Ibid., 2:426. See Malebranche 1700, 134 (The Illustration upon the Nature of Ideas); cf.
Malebranche 1962–​6 4, 3:131 (Éclaircissement X).
98. On Norris’s refusal of innatism, see Mander 2008, 79–​80, 171–​75.
99. Norris 1722, 1:314. For extensive criticism of the Cartesian doctrine of the creation of
eternal truths, see 1:339–​6 4 (pt. I, chap. 6, Sect. III, §§ 23–​43). Cf. Cudworth, TIS 646f.,
EIM I.iii.1–​6, 27–​34; Malebranche 1700, 119f. (The Illustration upon the Seventh Chapter of
the Second Book), 134ff. (The Illustration upon the Nature of Ideas); Malebranche 1962–​6 4,
3: 84–​86 (Éclaircissement VIII), 132–​34 (Éclaircissement X).
100. Norris 1722, 1:331f.; see also 1:155: “Ideas … are really coessential with and indistinct
from the essence of God.”
101. Ibid., 2:503. On this point, see Mander 2008, 47f.
Universals in English Platonism  •  193

with the multiplicity of the “archetypal forms or Representations of things in the


Divine Mind.”102 Norris answers that there is no real distinction between God
and his ideas, and so there is no real composition in God threatening the sim-
plicity of the divine nature: “ideas in God are in reality no other than the very
essence of God himself as it relates to things out of himself.”103 The possibility of
distinguishing various degrees of being and perfection in the simple and unitary
essence of God depends on the way in which the human mind, in its limitations,
conceives the relation between God and the objects in the created world that
participate in different forms of the divine essence.104 However, this possibility
is conceivable also with regard to the “inadequate” mode of knowledge that God
has of himself when he considers himself “in relation to the Creatures”: “when
[God] knows himself absolutely speaking, he considers himself adequately, but
in the knowledge of the creatures, he considers himself inadequately, according
to those degrees of Being or Perfection wherein his Essence is imitable or par-
ticipable by them.”105 Ideas in the mind of God are not necessarily connected
to their worldly and material instantiations. So the imitability and participabil-
ity are to be intended as possibilities which do not imply any binding connec-
tion between the ideal divine world and the sensible created world: “the Ideas in
God having no necessary relation to any Thing actually out of him.” As Norris
claims, in discussing the distinction between essence and existence, ideal essences
are distinct from existence not because they lack existence absolutely speaking
(being endowed with their own ideal existence), but in the sense that their exist-
ence does not require “the existence of things ad extra,” that is, “the existence of
those things in nature” that would instantiate them.106 Consistent with a Platonic
approach, Norris holds that universals are independent of their sensible instances
and that the ideal world enjoys its own archetypal reality exceeding both in exten-
sion and intension the existence of material, sensible, particular things. Even the
hypothetical annihilation of natural things would not affect divine ideas in any
way, since as they existed before creation, so they would continue to exist after its
disappearance.107

102.  Norris 1722, 1:231. This difficulty was widely debated in Medieval philosophy. For
instance, see Aquinas, Summa contra Gentiles, bk. 1, chap. 51, art. 1.
103. Norris 1722, 1:294 (my italics). See also 1:232.
104. See Norris 1689, 56–​58, where the issue is posed with regard to the difference between the
divine essence and its attributes.
105. Norris 1722, 1:167.
106. Ibid., 1:424–​25. See also 1:413–​15, and especially 1:416.
107. Ibid., 1:234.
1 9 4   •  The Problem of Univer sals in Early Modern philosophy

The centrality of the issue of universals in Norris’s philosophical theology goes


even deeper than the equation of divine ideas with God’s essence, since the very
idea of God, before any consideration of the plurality of divine ideas, is equated
with the most universal feature, that of Being in general. Quoting the famous pas-
sage of Exodus 3:14, Norris asserts that God is “Being it self, or Universal Being,
or Being in General, Being in the Abstract, without any restriction or limita-
tion.”108 Norris states that God is the genus summum from which every particular
being derives its own being by participation, and defines the identity of God with
Universal Being as “the formal conception and idea of God”:

I consider that as in every particular order or kind of Being there is a uni-


versal nature, under which all singulars are comprehended, and whereof
they all partake: as for instance, there is the Nature or Essence of a circle
or triangle in common, as well as this or that particular circle or triangle.
So in Beings consider’d as Beings, there is Being in general, Universal
Being, being it self, or the Essence of Being, as well as this or that Being in
particular. 109

Norris’s thesis, which echoes claims by Malebranche, radicalizes a theme of the


Platonic tradition, in opposition to Aristotelian and Thomistic line of thought.110
It is easy to see how this ontotheology, molded according to a realistic theory of
universals in which God himself is the most universal genus, risks rendering the
Christian idea of creation redundant and obfuscates the transcendence of God,

108. Norris 1689, 22. Norris echoes Malebranche 1700, 118 (III.2.5); 129 (III.2.9, §IV). See
Mander 2008, 38.
109. Norris 1689, 23.
110.  For the identification of God with Being in Plato, the Medioplatonists, Philo Judaeus,
Porphyrius, and in some Eastern Greek Fathers, see Berti 2009, 41–​73, 92–​93. A  notable
exception is Plotinus, who defined the One as “beyond essence” (epekeìna tes ousìas) (Enn.,
V.4.2, ll. 37–​42). On the other hand, against Plato, Aristotle denied that being is a genus
(Posterior Analytics 92b14, Metaphysics B.3, 998b22–​26). Augustine was the author who most
of all emphasized the theological relevance of Exodus 3:14 and who identified God with ousìa
but without stating that God is therefore the summum genus (De Trinitate V.ii.3; VII.v.10).
Aquinas, on the contrary, devoted a chapter of his Summa contra Gentiles (bk. 1, chap. 26) to
confuting the concept that God is esse commune omnium, highlighting the pantheistic danger
of such a thesis. But the doctrine of God as the esse summum and esse primum reappeared
in Ficino: see Kristeller 2005, 173 ff. Malebranche, who is a direct source for Norris, defined
God as “that Being without restriction, the immense Being, the universal Being” (Malebranche
1700, 122 [III.ii.7, §2]), as well as “Being without any particular Limitation” and “Being infi-
nite,” and identified the idea of God with the “indeterminate and general Idea of Being” (ibid.,
124 [III.ii.8, §1]).
Universals in English Platonism  •  195

opening the way to pantheism.111 However, this bold and dangerous identifica-
tion of God with being as summum genus is rectified in An Essay towards the
Theory of the Ideal or Intelligible World, where Norris takes care to state that God
is “All-​Being” in the sense that “the perfections of all things are in God,” meaning
that God “has in himself the Essential Reasons, exemplary Forms, or Ideas of all
things.” To state, in Ficinian terms, that God is omniform, that is, that “there is no
reality in the Creature but whereof there is the representative in the Creator,”112
is indeed quite different from identifying God with the universal genus of being.
In the first case the distinction between essence and existence is preserved, since
God by creation confers existence in the material world upon the creatures that
reflect in their essences the being of the archetypal world; in the second, when
God is identified with the universal genus of being, the way is open to an inclu-
sion of the finite beings inside the infinite being.
Flora Mackinnon pointed out the convergence of Thomas Aquinas,
Cudworth, Malebranche, and Norris in conceiving universals as ideas in God’s
mind.113 For my purpose it is particularly interesting to note not only the affinity
but also the differences among Norris, Cudworth, and Malebranche. Cudworth
and Norris agree in not conceiving ideas as humanly constructed concepts, and in
excluding the possibility that they were immaterial objects subsisting in a “third
realm.”114 The difference between them is that Cudworth emphasizes that eternal
truths are noémata, thoughts entirely dependent on the thinking activity of God.
In tune with the spiritual dynamism that characterizes his metaphysics, universals
are not a system of ideal objects, but first and foremost “living things,” that is,
modifications of the divine intellect existing in “that actual knowledge in which
they are comprehended.”115 On the contrary, Norris, following Malebranche,

111. On the difficulty of avoiding a pantheistic outcome in Norris’s philosophy, see Mander
2008, 58–​60.
112.  Norris 1722, 1:148; see also 1:141–​43, where Norris quotes Aquinas and Dionysius
approvingly to illustrate the concept that God “contains in himself whatever is in the Creature,
at least after a more eminent manner” (see Aquinas, Summa Theologiae I, q. 4 art. 2; Dionysius
Pseudo-​Areopagite, De divinis nominibus chap. V). For a further evidence that the “universal
being” of God is interpreted in the sense that God has “in its own essence, the ideas of all
things,” see Norris 1722, 2:258.
113. MacKinnon 1910, 43f., with reference to: Thomas Aquinas, Summa contra Gentiles, bk. 1,
chaps. 34, 48, 51; EIM IV.vi.2, 284; Malebranche 1871, 20, 102.
114. Norris 1722, 2:268f. There is a remarkable coincidence of views and arguments between
Norris’s Essay and Cudworth’s Treatise on Eternal and Immutable Morality, which Norris could
not have known since it was published in 1731. A possible common source is the Discourse of
Truth (1682) by George Rust.
115. EIM IV.iv.7, 250f.
1 9 6   •  The Problem of Univer sals in Early Modern philosophy

assumes that they are identical to the divine essence, which contains in itself the
supremely perfect archetypal world. This identity also explains why Norris dis-
cusses at length and rejects Suárez’s thesis that divine ideas lack “true real actual
being” and possess only a potential being, an aptitudo ad existendum. To this
thesis Norris opposes the argument that the simple acceptance of eternal truths
requires the admission of the reality of eternal essences: eternal truths presup-
pose relations among the essences of things; but to be related is a modification of
being and modal being implies pure being. Norris observes that Suárez and other
scholastic theologians wanted to preserve “the doctrine of eternal truths” without
admitting “the eternal existence of things.” To achieve this result they devised the
expedient of considering eternal truths as propositions asserted in a hypothetical
way that does not imply the actual existence of the related terms. Norris flatly
rejects this attempt at disconnecting logical truths from ontological reference
and does not admit a middle way, in terms of esse possibile, between actual reality
and nonexistence: “The short is, whatever truly is, is actu, and whatever is not
actu, really is not. These I take to be clear propositions, and therefore if in neces-
sary and eternal truths the relations of things are only ex hypothesi, or not actual,
then as yet those relations are not, nor consequently those eternal truths which
are founded in those relations, contrary to what they themselves [sc. Suárez and
others] suppose.”116
Distinguishing, as we have seen, the Mind of God as intelligent (or concep-
tive) from the Mind of God as intelligible (or exhibitive), Norris theorizes both
the independence and the dependence of the ideas from God’s Mind. With
regard to the Divine Mind as intelligent, eternal truths are presupposed to the
acts of God’s will “as their measure, as to all the acts of his Understanding, as
their object.” The priority of the object to the act of understanding is a principle
rooted in the very nature of understanding, which cannot create its object: as a
consequence, divine ideas are the necessary presuppositions of the Act of divine
understanding “as the terminative forms of it.”117 On this view, universals are
indeed objects (and not simply thoughts, as Cudworth maintained), but not self-​
subsisting objects, being encapsulated in the divine essence. Indeed, with regard
to the Divine Mind as intelligible, Norris recognizes that ideas may be said to be
dependent on it. How does Norris interpret this nexus? He says that divine ideas

116. Norris 1722, 1:92; see also 1:76, 89. The position of Suárez is discussed at length in 1:72–​
122, while Norris clarifies his “Platonic Heresy” of the real, actual, and eternal existence of the
essences of things (1:72). Cf. F. Suárez, Disputationes Metaphysicae, Disp. XXXI, §§ 2 (nn.
10–​11) and 12 (n. 45). On Suárez’s theory of the ontological status of divine ideas see Priarolo
2004, 78–​96.
117. Norris 1722, 1:234f.; cf. 1:36, 314.
Universals in English Platonism  •  197

are really in the mind of God, “not by way of inhaesion [inherence],” but “existing
by a kind of substantial emanation from it.”118 The introduction of the concept of
“emanation” differentiates Norris from Malebranche, who had never employed it
to describe the relation between God and divine ideas and had criticized its use
in characterizing the relation between God and the created world.119 The concept
raises a theological problem:  if divine ideas emanate from God’s essence, they
cannot be strictly identical to it and must constitute a derivative and inferior
stage, inserting therefore a gradation within the Deity. Norris tries to avoid this
unpleasant consequence by reading the concept of emanation against the back-
ground of the Trinitarian doctrine: divine ideas are identifiable with the Logos or
second person of the Trinity, which Norris does not hesitate to conflate with the
Plotinian Nous. The Neoplatonic concept of emanation is interpreted as synony-
mous with the Trinitarian procession of Christ (“the Light of the World”) from
the invisible God.120 Here we see again how the issue of universals, though not
discussed in logical terms, but diluted in the general question of the universality
of reason, lies at the core of Norris’s philosophical theology.

Abbreviations
AT = Descartes 1964–​74; cited by volume and page.
EIM = Cudworth 1731; cited by book, chapter, section, and page number.
EM = More 1679, Enchiridium Metaphysicum; cited by chapter, section
number, page.
MM = More 1995; cited by page .
TIS = Cudworth 1678; cited by page.

118. Ibid., 1:235. See also 1:191.


119. See Malebranche 1965, 137, 199 (Entretiens sur la Métaphysique VI §5; IX §2). On Norris
as a Platonic idealist more than a Malebranchean, see Mander 2008, 5–​9, 12f., 16, 183.
120. Norris 1722, 1:240. For the identification of the ideal world to the logos of the Gospel of
John see ibid., The Preface to the Reader, 1:vi–​xiv. The recourse to the concept of emanation
in Trinitarian theology is deeply rooted in Medieval philosophy: see Friedman 1999, 13–​25.
9 SOME PERSPECTIVES ON LEIBNIZ’S
NOMINALISM AND ITS SOURCES

Stefano Di Bella

9.1. The Nizolius Preface: A Genealogy


for Nominalism
In 1670, the young Leibniz reedited a book of Mario Nizolius, an
Italian humanist of the sixteenth century, and added a Preface that
provides a true manifesto for the renewal of philosophical method
and style.1 Nizolius’s De veris principiis et ratione philosophandi contra
pseudophilosophos is a typical example of vehement antischolastic and
anti-​Aristotelian criticism of the Renaissance age. The main targets of
Nizolius’s criticism were the realistic assumptions widely shared in the
tradition. In his Preface, Leibniz also focuses on this crucial aspect,
emphasizing its connection to the nominalist tradition:

since Nizolius did not hesitate openly to call himself a nominalist


… and since the nerve of his argument consists in his destruction
of the reality of forms and universals, I consider it worthwhile to
present certain facts about the sect. (A 6.2:427; GP 4:157, L.127–​28)

To show the roots of Nizolius’s stance, Leibniz reconstructs a geneal-


ogy going back to medieval nominalism, especially in its Ockhamist
variety.2 Most of all, he makes clear how much he appreciates this

1. See De veris principiis, etc., Dissertatio praeliminaris (A 6.2:401–​32; GP 2:127–​


74). For a modern edition of Nizolius’s work, see Nizolius 1956. Leibniz also left a
series of remarks on Nizolius’s work, now in A 6.2:445–​76.
2. Leibniz’s brief historical survey of medieval nominalism begins with Roscelin,
and focuses on Ockham as the most skillful representative and author of the great
revival of the school. See A 6.2:427–​28 (GP 4:157–​58, L I.158). Leibniz’s teacher
Jakob Thomasius had published an historical dissertation on nominalism, De secta
nominalium (1658).
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  199

philosophical attitude: “How greatly inferior the insights of the Scholastics of


this and the preceding centuries are to the earlier ones can be shown by the nom-
inalist sect, the most profound of all the scholastics, and the most consistent with
the spirit of our modern philosophy” (GP 4:157, L I.127; emphasis added). And
again: “Nothing is truer than this opinion, nothing is more worthy of a philoso-
pher of our own time” (GP 4:158, L I.128).
In his Preface, Leibniz makes the surprising claim that the nominalist atti-
tude is “all but extinct.”3 Most plausibly, he emphasizes the ideal connection
between that line of thought and the antirealistic attitude popular among
the “new philosophers” of his own time. In any event, Leibniz manages to
sketch a unified account of the nominalist tradition, capable of preserving
some core intuition through major changes in the philosophical background,
from medieval scholasticism, to humanistic thought, to the seventeenth-​
century philosophy shaped by the scientific revolution. Moreover, this his-
torical reconstruction reflects Leibniz’s own philosophical development. As
a pupil educated in scholastic philosophy, he had concluded his training in
1663 with a Disputatio on individuation in purely scholastic style. The intro-
duction of his teacher Jakob Thomasius, however, already showed a deflation-
ary attitude toward scholastical ontological subtleties.4 In the Disputatio,
Leibniz had endorsed the thesis, found in the work of the later scholastic
Francisco Suárez, of the individuation tota sua entitate (“by its own entity”).
Though Suárez was officially a moderate realist, he nonetheless was strongly
influenced by the nominalist view.5 Indeed, the Suarezian thesis that Leibniz
endorses can be seen more as a dissolution of than as a solution to the tradi-
tional issue of individuation. To say that an individual is such by its own entity
amounts to saying that it is such by itself—​which is exactly the nominalist
intuition. For the nominalist, individuality is a primitive fact, standing in no

3.  “This sect, once very prosperous, is now extinct, certainly among the Scholastics”
(A 6.2:427). This historical judgement is hard to accept, at least in this blunt form.
Admittedly, few scholastic authors of the seventeenth century would have perhaps
accepted the label of nominales; moreover, several technical aspects of the semantic analy-
sis applied by Ockham and his followers had been actually lost. Still, nominalist theses and
attitudes were well present, also in authors officially far from the nominalist “school”—​
Francisco Suárez among them. For an analogous (historically as untenable) assimilation
of the whole of scholastic teaching to the realist stance, see Gassendi’s judgment, reported
in chapter 2 in this volume.
4.  See Disputatio metaphysica de principio individui, A  6.1:9–​19 (GP 4). For Thomasius’s
Preface, A 6.1:5–​8. On the Disputatio, which is a significant text for the understanding of the
young Leibniz’s philosophical background, but often overinterpreted in the scholarship, see
Ariew 2001.
5. Suárez’s classic text on the universals problem is his Disputatio VI.
2 0 0   •  The Problem of Univer sals in Early Modern philosophy

need of explanation. Given that there is no “common” or “universal” reality, each


(possibly or really) existing thing is particular, and is individuated only by itself.
This “particularist claim” is now indicated by Leibniz, in his Preface to
Nizolius, as the core of the nominalist stance:

Nominalists are those who are committed to the existence of singular sub-
stances only, taking all the rest for mere names; hence, they do away with
the alleged reality of any abstract or universal item. (A 6.2:427; GP 4:15)

This antirealist assumption is grounded for Leibniz in a requirement of ontologi-


cal parsimony, which is, in its turn, an application of the general epistemological
rule known as Ockham’s razor. This characterization of nominalism emphasizes
the elimination not only of universal things, but also, and more generally, of every
abstract item.6 It is quite austere insofar as it seems to admit only particular things
(substances) in its ontology.7 Moreover, it may seem to be strictly “nominalist”
insofar as here only names are taken into account. However, Leibniz’s statement
should be given a “conceptualist” reading (as was the case for Ockham himself ).8
If Leibniz is correct in saying that this antirealistic assumption is largely shared by
the representatives of the “new philosophy” of his age, at the same time he is eager
to circumscribe its import by arguing that it does not have certain controversial
implications. He pursues this task by criticizing the details of Nizolius’s nominal-
istically minded solutions.9

9.2. Nominalism: Its Force and Limits


9.2.1. Platonism Conserved?
In considering the main aspects of Leibniz’s critical assessment of Nizolius’s
stance, one can gain a better insight into the sense and limits of his persistent

6. More precisely, he excludes “any reference to universals and real forms” (A 6.2:428; L I.158).
The translation “forms,” however, does not entirely capture the Latin formalitates: a technical
term designating the metaphysical constituents of concrete things (such as “common natures,”
e.g., humanity), whose ontological status of real “quasi-​things” had been especially emphasized
in the Scotistic tradition.
7. Actually Leibniz also admitted particular accidents in his ontology, along with many of his
contemporaries, both realistically or nominalistically minded. On this, see more below.
8. Concepts are at the center of Leibniz’s logical ontology. At the same time, he will be eager
to emphasize the relevance of their linguistic expression. See below about his critical reception
of Hobbes’s lesson.
9. See A 6.2:428 (GP 4:158, L I.128): “The general rule which the nominalists frequently use is
that entities must not be multiplied beyond necessity. … From this principle the nominalists
have deduced the rule that everything in the world can be explained without any reference to
universals and real forms.”
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  201

endorsement of a “nominalist” stance in his later career.10 Leibniz distances him-


self from Nizolius’s antirealist polemic chiefly on three points: (1) a historical
one, involving the consideration of the roots of realism in ancient philosophy,
with the important philosophical corollary of a quite different assessment of
Plato’s heritage; (2)  a logico-​ontological one, regarding the interpretation of
generality; and (3) an epistemological one, concerning the relationship between
realist assumptions and the traditional framework for science. I consider these
points in turn.
While Nizolius did not hesitate to attribute the realist mistake to Aristotle
himself, Leibniz—​in the spirit of his teacher Thomasius—​tends to distinguish
the Stagirite from the later scholastic tradition, and credit him with a decidedly
antirealist reading of universals: “On the contrary, for Aristotle there is hardly
any trace of universals, unless within singular things, or in our minds and words”
(A 6.2:451 note).11 More interestingly, however, he does not share Nizolius’s
rejection of Platonic ideas, contrary to what one would expect, given his posi-
tive appreciation of Aristotle’s alleged antirealist stance: “Our author [Nizolius]
seems not to have understood their [i.e., Plato’s ideas] nature and function”
(A 6.2:447). How can these apparently contrasting claims be reconciled?
The answer probably lies in a traditional theological view that Leibniz
endorses. According to this view, going back at least to Augustine and shared
in different ways by the generality of scholastic thinkers, Plato’s ideas had been
located in the divine mind, as the paradigms according to which God created
the world.12 This view allowed even nominalist thinkers to avail themselves of
the epistemological resources provided by a Plato-​style theory, without subscrib-
ing to a Platonist ontology and theory of knowledge. In order to account for
the metaphysical constitution of concrete things, and for our activity of concep-
tualization, the same scholastic thinkers could rely, instead, on the Aristotelian
model in its moderately (or even frankly) antirealist version. We should bear in

10. The presence in Leibniz of a nominalist train of thought has been fully appreciated only
in the last decades, especially thanks to the pioneering works of Benson Mates and Massimo
Mugnai, which have explored the influence of nominalistic themes and attitudes in several
Leibnizian topics, from the ontology of predication to the theory of relations to the work for a
philosophical language and the general conception of philosophical analysis. See Mates 1978,
1980, 1986; Mugnai 1976, 1992.
11. To be exact, Leibniz alludes to the traditional reading of Aristotle as a “moderate realist,”
admitting only immanent universals within singular things. But he seems to suggest also a more
radically antirealist reading of Aristotle, such as had been proposed by medieval nominalists.
12.  For Leibniz’s explicit endorsement of the Platonic-​Augustinian tradition, see his well-​
known statement in the New Essays, bk. IV, chap. 11. For a balanced assessment of the ontologi-
cal import of this theory as far as the reality of possibilia and the relationship with nominalistic
assumptions are concerned, see Mugnai 1990; Mondadori 1990.
2 0 2   •  The Problem of Univer sals in Early Modern philosophy

mind that Leibniz also is always consciously operating within a kindred two-​layer
framework, working out a quite sophisticated version of it.

9.2.2. The Criticism of Nizolius’s Extensionalism


In Leibniz’s view, Nizolius was quite right in rejecting the realist’s reification of
universals, but did not manage to provide an adequate alternative account of gen-
erality. According to the Italian humanist, universal items should be interpreted
as standing for “collective wholes”: for example, the species “man” would simply
amount to the sum of all existing men. For Leibniz, this reading is unsatisfying
for two reasons. First and foremost, it cannot provide a correct account of predi-
cation. In fact, a universal concept is not a “collective” but a “distributive whole”
that designates disjunctively the (infinite) elements of a collection.13 Second, in
Nizolius’s view the reference of the universal term would depend on the existence
of its instances, with all of the problems related to the constantia subjecti (that is,
the assumption of the existence of the object designated by the subject term of a
sentence). As Leibniz objects: “If the genus of roses is the multitude of all past,
present, and future roses taken together, then it will never exist: because past and
future things never exist together” (A 6.2:454 note 26).
As a matter of fact, Nizolius admits that if all individuals of a species were
to perish, the species itself would perish. Leibniz’s reply points to the need to
quantify over not only existing (actual) but also possible individuals. This is why
one could state truths about men, or elephants, even if all individual men, or ele-
phants, perished:

[Nizolius] is mistaken here. If all singular instances in the world were


suppressed, nevertheless a universal proposition would maintain its truth
value with respect to the possibilia. Even if all elephants were killed, the
proposition “Every elephant is an animal” would remain true, given that
it can be reduced to this conditional proposition: “If something is an ele-
phant (whether it actually exists, or not), then it is an animal.” (A 6.2:448
note 6)

This remark introduces us to the epistemological implications of the debate over


universals.

13. See A 6.2:430–​31; A 6.2:453–​54 note 24. I use ‘collection’ in a loose way, to avoid the termi-
nology of “class,” which is for us much more technical and theory-​laden, of course. The relation
of Leibniz’s criticism of Nizolius to class-​theoretical notions is a complex and controversial
issue discussed by interpreters. See for this Angelelli 1965; Rauzy 2001, 177–​202.
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  203

9.2.3. Universals and the Foundation of Science


According to the traditional Aristotelian framework, science is based on univer-
sal concepts, which confer on its statements their universality and necessity. But
if universal concepts are deprived of their alleged ontological import, then the
whole building of science seems to lack adequate foundations, or at least loses its
connection to reality. Nizolius does not refrain from drawing this consequence
by rejecting the traditional ideal of science. Moreover, he thinks that the abolition
of real universals brings with it the breakdown of both logic and metaphysics.
In contrast, Leibniz is adamant in rejecting this consequence. As far as logic is
concerned, he holds that the interpretation of its subject matter in purely linguis-
tic terms—​as a science bearing on words or signs—​does not prevent its fruitful
exercise, as the old nominales themselves show.14 More generally, even if one gives
up the alleged ontological basis of logic in real universals, one is not commit-
ted to giving up as well the traditional ideal of science as universal and necessary
knowledge—​an ideal that Leibniz wants to firmly maintain.
Once real universals are abandoned, in fact, one is not confined to the inspec-
tion of particular existing things, insofar as one can avail oneself, again, of the
resources of possibilia:

On the contrary, this [breakdown of science] does not follow. Science, in


fact, is not only about existing things, but also about possible ones. It is
not concerned with the question, whether the triangle actually exists, but
only with the question of what follows … if it exists. Science, therefore,
is not about real universals, but about all singular things, the possible ones
included. (A 6.2:461 note)

This seems to provide a quasi-​extensional interpretation of modal claims, in tune


with the spirit of much present-​day philosophy of logic. And if one is struck by
the luxuriant ontology of possibilia (in the precise sense of possible individuals)
that is presupposed, nonetheless room for them might be provided in principle,
in Leibniz’s view, by the divine understanding.
If general concepts were to be interpreted, instead, in Nizolius’s way, then epis-
temological bankruptcy could not be avoided. The “collective” view of universals,
taken as multitudes of actual things, would commit us to look for a justification
of universal statements only by way of induction, something certainly doomed to

14. “[N]‌ominalists made usage of Aristotle’s dialectics no less than others did. … Although
that which was attributed to the things is true only of names, nevertheless the import of our
knowledge is assured, provided we employ names to understand things” (A 6.2:451 note).
2 0 4   •  The Problem of Univer sals in Early Modern philosophy

failure.15 Leibniz is eager to subscribe to the ancient idea that no scientific knowl-
edge can be grounded on the “bad infinity” of individuals. In a well-​known pas-
sage from his Preface to Nizolius, Leibniz shows how even the simplest inductive
reasoning needs to rely on some a priori rational assumptions in order to be con-
clusive.16 What is more relevant for us, his intuition is clearly that the universality
of a statement cannot be captured in a purely extensional way, by embracing a
list or an infinite conjunction of cases, even if this were practically possible. True
universality is grounded, instead, on a kind of conceptual necessity.
For this reason, we should refrain from thinking that Leibniz’s extensionaliza-
tion move in terms of possible individuals provides his ultimate explanation of
the truth of necessary universal statements. The logical form of necessary propo-
sitions, remember, is conditional: “If there were a man, he would be rational.” But
for Leibniz, the truth of these conditionals is ultimately grounded in conceptual
inclusion, considered from a firmly intensional perspective, rather than in the
ideal inspection of possible cases.17

9.2.4. Against Hobbes: Supernominalism and


the Conventionalist Challenge
Leibniz’s idea of the conditional analysis of necessary propositions can be found
not only in Suárez, but also in Hobbes. In contrast to Nizolius (and, among
seventeenth-​century thinkers, Gassendi), Hobbes, like Leibniz himself, attempts
to combine a form of radical antirealism with the traditional ideal of science. For
Hobbes, as is well known, the necessity of scientific propositions is ultimately a
linguistic one, based on definitions, which are stipulated by human beings.18 But
this strategy is exactly what Leibniz, in his Preface to Nizolius, dubs “ultranomi-
nalism,” and he is eager to distance himself (and “good” nominalism) from it:

Occam himself was no more nominalist than is Thomas Hobbes now,


though I confess that Hobbes seems to me to be a supernominalist. For
not content like the nominalists to reduce universals to names, he says
that the truth of things itself consists in names and what is more, that it

15. See A 6.2:452 note 19; A 6.2:431.


16. See Dissertatio praeliminaris, A 6.2:431–​32.
17. Consider Leibniz’s well-​known foundation on conceptual containment of syllogistic rela-
tions and hypothetical propositions, in his mature theory of concepts and truth. The founda-
tion of necessary truths on conceptual containment was already present in some scholastic
accounts of “eternal truths” like Suárez’s. See Disputationes Metaphysicae, Disp. XXXI.
18. For a discussion of Hobbes’s view, see chapter 3 in this volume.
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  205

depends on the human will, because truth allegedly depends on the defini-
tion of terms, and definitions depend on the human will. This is the opin-
ion of a man recognized as among the most profound of our century, and
as I said, nothing can be more nominalistic than it. (A 6.2:429; L I.128)

But he annotates in the margin: “Yet it cannot stand …”. To falsify Hobbes’s con-
clusion, Leibniz points here to the fact that different systems of signs can express
the same truths. The fact of translatability was in fact the standard argument that
Descartes and his followers offered against Hobbes, to show the irreducibility of
our reasoning to the mere manipulation of conventional signs.19
Given his robust conceptualist intuition, rooted in the Platonistic strand of
his thought, one might expect that Leibniz also would insist on the irreducibil-
ity of meaning to “signs,” as well as on the reality of our “ideas.” In fact, he does
emphasize the need to rely on the deeper level of ideas or thoughts, underlying
that of language or signs. Things are not so simple, however. In the years follow-
ing the Preface, Leibniz takes Hobbes’s conventionalist challenge much more
seriously than Descartes and the other opponents of Hobbes did. The key point is
that Leibniz increasingly acknowledges the constitutive role of signs with respect
to our thought.20 Hence, the ideas/​signs distinction does not imply, in his view, a
total autonomy of ideas themselves. A culminating point in this reflection is found
in his well-​known 1677 Dialogus.21 In this fascinating text, Leibniz concedes that
(1) our access to ideas and truths must be mediated by signs; but (2) signs are
arbitrary, that is to say conventional. How, then, could one avoid Hobbes’s con-
clusion that (3) truths themselves would depend on our conventions?
Certainly, Leibniz points to the fact that (4) the same ideas and truths can
be expressed by different systems of signs (the standard counterexample to the
Hobbesian thesis). Taken together with (1)–​(3), however, (4) seems to result in
an aporetic situation. But now Leibniz manages to give an account of (4) from
within his “semiotic” view of human thought, expressed in (1)–​(2). More exactly,
he can reconcile the truth of (4)  with that of (2)  while avoiding the radically
conventionalist conclusion (3). This is possible if one recognizes that (3´) all dif-
ferent expressions of one and the same proposition, though being conventional,
must respect a certain order or share a common structure. This isomorphism—​
captured by the key technical notion of expression—​preserves the objective

19. The argument is advanced also in the Logique of Port-​Royal. On this Cartesian criticism,
see chapters 3 and 8 in this volume.
20. See on this Dascal 1987.
21. A 6.4:20–​25.
2 0 6   •  The Problem of Univer sals in Early Modern philosophy

value of the expressed truth and works as explanatory ground for (4), that is, for
mutual translatability among different languages.
I have already noted that several supporters of the Platonic-​Augustinian
framework of ideas in the divine mind were prepared to apply an Aristotelian
theory of abstraction at the level of human knowledge. The role Leibniz is pre-
pared to assign to signs introduces a further relevant aspect into this type of
two-​level view. According to him, there are important differences between ideas
in themselves (that is to say, in the mind of God) and ideas insofar as they are
grasped by human minds. Leibniz often marks this difference terminologically
by reserving the venerable name of ‘ideas’ for the intelligible contents of the
divine mind, which provide the ground for our intellectual knowledge, while
using the label of ‘concepts’ for the same contents insofar as they are grasped by
human minds. More exactly, ideas become somehow present in our minds both
as dispositional structures and as concepts actually grasped through the medium
of language and/​or images.22 This inseparability of concept and sign suggests
that the conceptualist/​nominalist divide is not as relevant for Leibniz as one
might imagine. And his usual ambiguity in the usage of ‘term’—​sometimes
standing for a word, sometimes for a concept—​far from simply being a mark of
inaccuracy, could be a clue for the close relation he maintains between thought
and sign.

9.2.5. Fundamentum in re: Concepts and Things


Leibniz holds that although we have an access (mediated by signs) to divine ideas,
still we are bound—​at least in most cases—​to draw our concepts from things, by
way of abstraction.23 In any event, we are assured that our concepts have some
foundation in reality. When we consider whichever group of things, “even if
our mind does not find the common genus, God or angels will know it; there-
fore a foundation of all these abstractions will preexist” (De arte combinatoria,
A 6.1:192; GP 6:61).

22. The classic text for this is the Quid sit idea, A 6.4:1370–​71.
23. This seems to be true, at least, in the case of our empirical concepts. Things are different
as far as mathematical notions and truths are concerned (on this difference, see also chap-
ter 10 in this volume). Here, innatism means having access to some ideal patterns which are
not abstracted from sense experience. This contrast between an abstractionist/​conceptualist
(“Aristotelian”) approach, at the level of ordinary empirical concepts, and a truly Platonic one,
at the level of the basic notions of the new mathematical science of nature, is also shared by
many of Leibniz’s contemporaries. For our empirical knowledge, however, Leibniz emphasizes
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  207

This minimal requirement of a foundation in re for our conceptualizing activ-


ity was shared by nominalist (or conceptualist) authors as well as by (moder-
ate) realists. The two groups parted company with respect to the further task of
specifying what the relevant foundation amounts to. While the former stopped
at the recognition of some objective resemblance relation taken as a primitive
fact, the latter claimed the need to ground resemblance on the sharing of a com-
mon element. As a matter of fact, Leibniz will always be satisfied with the appeal
to resemblance as the last objective foundation of our classificatory activity. In
his later years, when criticizing the conventionalist drift of Locke’s nominalism
in the treatment of the concepts of natural kinds, Leibniz defends the objective
reality of these concepts, insofar as they are grounded on some real (or possible)
resemblance.24
So far, I  have tried to reconstruct the complex and nuanced blend of
Platonistic, conceptualist and nominalist motives underlying Leibniz logical and
metaphysical research during the 1670s. Having traced the main lines of Leibniz’s
“conceptualism,”25 I want to focus in the next section on his development of a
different critical line starting from Hobbes, which is found already in the Preface
to Nizolius. This line seems to provide the resources for handling some of the
semantic and ontological problems most impervious to nominalist analysis,
namely, the logico-​linguistic problem of abstract reference and the related issue
of the ontology underlying predication.

the need to presuppose also a framework of a priori notions, which are not drawn from expe-
rience by way of abstraction. In any event, for both empirical and mathematical concepts,
Leibniz rejects their reification: we are dealing in this chapter with his criticism of traditional
realism, but also his criticism to the handling of space, time, and other ideal notions as “things”
is well known.
24.  “I do not see the consequence [Leibniz challenges Locke’s claim that the species of
things are a ‘workmanship’ of our understanding]. For generality consists in the resem-
blance of separate things among themselves, and this resemblance is a reality” (Nouveaux
Essais, bk. 3, chap. 3, GP 5:271). Also, Ockham’s conceptualism emphasized the “natural”
character of our conceptualizing, based on objective resemblances. In another passage from
the New Essays, Leibniz speaks about the “possibility of resemblance” as the ground of the
general concepts of essences: “It does not follow that if general essences are not this [i.e.,
something existing as such], they are merely signs … for I have many times remarked to
you that they are possibilities in the resemblances” (New Essays, bk. 3, chap. 6, GP 5:303).
Leibniz, that is, as usual, is reinforcing a typical nominalist solution (resemblance nominal-
ism) by relying on the space of possibility. On the discussion between Leibniz and Locke,
see chapter 10 in this volume.
25. For a balanced appreciation of the significance of Leibniz’s conceptualism, and its relation-
ship with the Ockamist tradition, see Schepers 2014.
2 0 8   •  The Problem of Univer sals in Early Modern philosophy

9.3. With (and Beyond) Ockham and Hobbes: The


Ontological Program and the Criticism
of Abstract Terms
9.3.1. The Background: Modifying the Ontological Square
The standard ontological paradigm of Aristotle’s Categories, represented by the
“ontological square,” was framed around the crossing of the two pairs of par-
ticular/​universal and substance/​accident. The medieval nominalist approach
excluded from the ontological inventory the allegedly universal entities (in the
categories of substance and accident). It maintained, instead, the substance/​acci-
dent distinction. Hence, this approach admitted within its ontology also particu-
lar properties, that is, individual accidents.26
Sometimes, however, a further sharp program of ontological reduction was
applied also to the so-​called accidental categories, where these accidents were
ranged. Admittedly, Ockham himself did recognize at least some types of items
in the category of qualities as irreducible.27 Yet he was careful to deny that these
individual accidents constitute the proper reference of the terms that pick them
out. In this case, the appropriate semantic relation was not for him designation,
but connotation. Thus, for example, the term “white” properly designates a con-
crete white thing (a ball of snow), but it connotes the relevant individual quality
(this whiteness).
Moreover, in Leibniz’s age there is, among the “moderns,” an uneasiness with
the ontological status of accidents.28 For instance, Hobbes seems a bit embar-
rassed with accidents:  he oscillates between a view of them as relative to our
knowledge and a more realist understanding of them as real properties within the
things. Moreover, he is unsatisfied with the traditional definition of the relation
of inherence holding between accidents and their ontological subject. He agrees
with Aristotle that accidents are not in the subject as if they were a part of it, but
he has trouble finding a positive way of making sense of this type of nonmereo-
logical inclusion within his physicalist ontology.29
In the following I try to reconstruct Leibniz’s stance on these topics, starting
once again from his reception of Nizolius’s and Hobbes’s lesson.

26. Therefore, this form of traditional nominalism is less economical than standard present-​day
“austere nominalism,” which admits only particular things as entities. See also above, note 7.
27. See on this Adams 1987, 143–​67.
28. This attitude finds also a terminological expression in the option for the terminology of
“modes” instead of “accidents.”
29. See De corpore, part II, chap. 8, OL 1:91–​93.
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  209

9.3.2. Nizolius’s Ontological Program and the Substantive/​


Adjective Distinction
For his own part, Nizolius basically follows the standard ideas of nominalist
tradition on the ontological square, though oversimplifying them as usual. His
square is articulated on the grammarian’s classification of names, hence on the
two pairs of substantive/​adjective and proper/​common.30 According to him,
however, the second distinction does not correspond to a true ontological divide,
insofar as no universal item corresponds to the “common” (usage of ) names: in
the world there are for him only individuals and their groups. Instead, a categorial
difference that corresponds to the grammatical substantive/​adjective distinction
is admitted (as in the Ockhamist tradition) between things, or substances, on one
hand, and accidents or qualities, on the other.
Interestingly enough, Leibniz seems to be even more radical than Nizolius
on this point. Already in his remarks on the De veris principiis, we find the idea
that the grammatical substantive/​adjective distinction is not so relevant, from
the philosophical point of view. In advancing this view, Leibniz is in tune with
Hobbes’s attitude in the De corpore. Hobbes, in fact, holds that “man” or “cold”
are both “names” by which we refer to a thing. This move does not mean, how-
ever, that the distinction between the thing and its characters is suppressed;
rather, it reflects the weakening of the essentialist distinction between essential
and accidental properties. In this way, all characters or qualifications are leveled,
and opposed as such to the “thing” or “body.”
Translated from the physicalist language of “body” to the more neutral onto-
logical one of “being” (Ens), Leibniz adopted this analysis of our ways of referring
to concrete things in his remarks on Nizolius’s book:

From the adjectival expression ‘animal’ [Latin: animalis]—​where “animal


being” is meant—​the substantive ‘animal’ [Latin: animal] is made. And
substantives and adjectives differ in the fact that the ‘Being’ or ‘Thing’
(the prime substantival term) inheres in the former by way of ellipsis, and
must be mentally supplied. (A 6.2:449 note 11)

The reference to “being” (or “thing”) is indirectly implied (subintelligitur) by


the meaning of the term. Thus the order of priority in semantic relations turns
out to be reversed with respect to the Ockhamist model of connotatio, according
to which adjectival terms directly designate the thing and only indirectly sig-
nify its (particular) property. The thesis of the dispensability of the substantive/​

30. Nizolius 1956, bk. I, chap. 4–​5, 41–​58.


2 1 0   •  The Problem of Univer sals in Early Modern philosophy

adjective distinction in a logically regimented language, and the analysis of every


concrete term as implying a reference to Ens, will constantly occur in Leibniz’s
later work (mostly from the middle period of his intellectual career) devoted to
the philosophical analysis of language (the so-​called characteristica verbalis).31

9.3.3. Hobbes and the Concrete/​Abstract Distinction


The other linguistic dichotomy of concrete/​abstract was handled more periph-
erally by Nizolius. It received, instead, more attention by Hobbes. Leibniz also
focuses mainly, and not incidentally, on this polarity.32 Recall his remark in the
Preface to Nizolius, where he emphasizes the rejection of “every abstract item.”
We know that the standard “moderate realism” in late medieval and early mod-
ern scholasticism, while not committed to the existence of real universal things,
insisted on the presence within the things of real constituents, corresponding to
abstract terms, in order to explain attribute sharing and predication. As a con-
sequence, the discussion of the semantics of abstract terms had a crucial place
within the wider linguistic approach of both Hobbes and Leibniz to the onto-
logical problems. This is also a nice antecedent of the present-​day debate over
nominalism and realism, where linguistic considerations play a seminal role,
and where the phenomenon of abstract reference has proved to be one of the
main challenges posed to the nominalist, who tries to eliminate it by appropriate
paraphrases.
In the De Corpore, Hobbes shifts the study of the concrete/​abstract polarity
from the chapter on names to that devoted to proposition. This move is justi-
fied by the fact that abstract terms are rooted in predication; hence, they receive
their significance only in a propositional context, and not as autonomous refer-
ring devices. They designate the “cause for which (concrete) things are referred to
by way of the respective (concrete) names,” where “names,” according to Hobbes’s
general usage, embraces both substantival and adjectival forms; and “cause” is
not something having a properly causal, explanatory role, but rather indicates the
“state of affairs” corresponding to the abstract property. Thus, ‘justice’ is para-
phrased as ‘being just’, ‘sweetness’ as ‘being sweet’, ‘heat’ as ‘being hot’.33 This
Hobbesian analysis will assume quite a seminal value for Leibniz’s reflection on
the issue.

31. See, for instance, Characteristica verbalis, A 6.4:334–​35. On the general analysis of concrete
terms, see Rutherford 1988.
32. On Leibniz’s treatment of abstract terms, see Rauzy 1993; Di Bella 1998.
33. See De corpore, part I, chap. 3, OL 1:28–​31.
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  211

Hobbes, for his own part, makes use of this analysis as a kind of linguistic ther-
apy directed against the practice of metaphysicians, who unduly reify abstract
terms: “Hence come the crass mistakes of some metaphysicians: so from the fact
that one could consider thought without considering body, they claim that there
is no need for the thinking body” (OL 1:30). Of course, this is a clear allusion to
his well-​known objection to Descartes’s proof of mind/​body distinction in the
Meditations.34 However, Hobbes also is ready to recognize that abstract talk is a
linguistic device that is well suited for the manipulation of concepts:

The use [of abstract terms] consists in this, that without them we cannot,
for the most part, either reason, or compute the properties of bodies; for
when we would multiply, divide, add, or subtract heat, light or motion,
if we should double or add them together by concrete names, saying …
hot is double to hot, light double to light, or moved double to moved,
we should not double the properties, but the bodies themselves that are
hot, light, moved, etc, which we would not do. (EW 1:33; compare OL
1:29–​30)

9.3.4. Leibniz on Abstract Terms 1: The “Hobbesian” Paraphrase


and the Logic of Concepts
In his Preface, Leibniz is willing to develop a program of linguistic therapy as a
way of eliminating abusive entities from philosophical language. This elimination
matches well with his overall antirealistic attitude:

One further warning seems worthwhile here, since its opposite is com-
monly held: in philosophizing accurately, only concrete terms should be
used … this passion for devising abstract words has almost obfuscated
philosophy for us entirely; we can well enough dispense completely with
this procedure in our philosophizing. For concretes are really things;
abstractions are not things but modes of things. (A 6.2:417)35

34. Interestingly enough, Leibniz will later employ this sort of semantic criticism not only in an
anti-​Cartesian way, as Hobbes did against mind/​body distinction, but also and more generally
against the new “abstract beings” of the moderns, like Descartes’s “extended substance” or the
matter cherished by Hobbes himself.
35. These lines are immediately followed by a classic antirealist argument, based on the infi-
nite proliferation of entities, to which the realist would be committed. Interestingly enough,
Leibniz here also contrasts the scholastic abuse of abstract terms and Aristotle’s sounder usage
of concrete expressions: 'the white' (tò leukon) instead of 'whiteness'.
2 1 2   •  The Problem of Univer sals in Early Modern philosophy

What about the positive appreciation of the advantages of abstract talk, in order
to submit the properties of things to our calculation? Surprisingly enough,
Leibniz shows himself more radical than Hobbes in his antirealistic zeal. We
know well how Hobbes’s idea of knowledge as calculation has been an inspir-
ing one for Leibniz’s lifelong project. But Leibniz thinks that abstract talk is not
required for this. Accordingly, Hobbes is criticized in the Preface for his (relative)
defense of abstractions:

I do recall that the penetrating Hobbes ascribes some usefulness to abstract


terms, by the argument that it is one thing, for example, to double some
hot object, quite another to duplicate heat. But this duplication of heat
can itself be expressed in concrete terms, for if I say that the same thing has
been made twice as hot, or that the effect by which the heat is measured
is double, everyone will understand that it was not the hot thing but the
heat that was doubled. (A 6.2:417; GP 4:147)

This critique of abstract terms is a constant one for Leibniz. Though mitigated,
it reemerges in his later Generales Inquisitiones (1686), one of his most impor-
tant achievements in logic. The draft—​a development of a kind of algebra of
concepts—​provides a good insight into Leibniz’s mature treatment of abstract
terms in a logical language. It opens with the advice that all terms employed in
the calculi are meant to be “taken in concreto.”36 Toward the end of this work,
Leibniz comes again to the example of the “double heat.”37 This time, how-
ever, he makes room for a positive role of abstract terms within his logical lan-
guage and calculus. And this is possible because he reinterprets them through
their Hobbesian-​style paraphrase, leading them back to their propositional
nature according to the following scheme: the B-​ness of A is translated as the

36.  “Let us, for the present at any rate, omit all abstract terms, so that all terms are under-
stood to refer to concrete things alone—​whether these are substances … or phenomena” (A
6.4:740; P 47). The proposal of doing without abstract terms in his ideal language can be found
in several occurrences in Leibniz’s drafts. See, e.g., the Characteristica verbalis: “In the lingua
philosophica abstract terms can be dispensed with, and in this way we will eliminate many com-
plications.” (A 6.4:337)
37. “How shall we explain quantity in abstract terms—​for example, when A is twice as hot as
B, i.e., when the heat of A is twice the heat of B? A’s being hot is the heat of A; so if A’s being
hot is to B’s being hot as 2 is to 1, then the heat of A will be twice that of B. But it must be
seen further how A’s being hot can be to B’s being hot as one number to another. This happens
owing to the fact that the cause which, with uniform action, makes A hot, makes B hot if the
same action is continued; or the sign by which we recognize that something is continuous, and
in the one case is double the other. But much care is needed in this” (Generales Inquisitiones,
§141, A 6.4:778; P 79).
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  213

“A-​being-​B.”38 Taken in this way, abstract terms play a pivotal role in the gen-
eral project of a unified treatment of propositions and concepts pursued in the
Generales Inquisitiones: in fact, they allow the reduction of categorical proposi-
tions to concepts as well as the reduction of hypothetical propositions to cat-
egorical ones. According to this analysis, “If A is B, then C is D” becomes “The
‘A-​being-​B’ is [i.e., implies] the ‘C-​being-​D’.”39 In this way, the mature Leibniz
rehabilitates Hobbes’s intention of using abstract terms for the manipulation of
concepts; moreover, this is done by developing the paraphrase originally sug-
gested by Hobbes himself, which properly dispenses with abstract reference.

9.3.5. Leibniz on Abstract Terms 2: The “Hobbesian” Paraphrase


and a Deflationary Ontology of Predication
In contexts more oriented toward a regimentation of natural language in view
of categorial inquiry, a similar deflationary attitude toward abstract terms marks
the first step in Leibniz’s search for concrete basic particulars, or substances. On
the one hand, his analysis of the synctactic and semantic properties of abstract
substantival terms (like “heat”) shows their irreducibility to the corresponding
concrete terms; on the other, this analysis confirms that they do not add any new
content, but simply express an operation of reification of our concepts.40
When abstract substantive terms are used in the analysis of predication, how-
ever, they seem to stand for some entities that entertain the (oblique) relationship

38. Paragraphs 138–​43 of this work are devoted to the analysis of abstract terms according to
the propositional paraphrase, and they test systematically the capacity of this paraphrase to
capture the semantical and inferential import of ordinary abstract terms. Leibniz adopts some
graphic and syntactic device to express his “logical abstracts”: the “A-​esse-​B” is written with a
trait on it, and preceded by the Greek article tò.
39.  Sometimes Leibniz goes as far as to suggest an interpretation of his abstracta as if they
expressed what we would call “states of affairs.” See the draft De illatione et veritate: “Let us call
‘L’ the state, by virtue of which A is B, and ‘M’ the state, by virtue of which C is D” (A 6.4:863)
40. “So far I am not able to explain the abstract and the concrete other than in the following
manner. Suppose A and B are one and the same thing; then consider two items L and M, which
do differ from A and B respectively only from the fact that L and M are (contrary to A and
B) two different things. Then I call L and M abstract items and A and B concrete ones. E.g.
let something be hot and dry; now, heat and dryness do not differ from the hot and the dry
respectively, if not because what is hot and what is dry are one and the same thing (which is said
to be a subject), whereas heat and dryness are two different things which are said to be inherent
in the subject. This is why, ‘hot’ is called concrete, and ‘heat’ abstract, ‘dry’ concrete, ‘dryness’
abstract.” (A.6.4: 400, note) In the background there is the view we are already familiar with,
according to which each meaningful unit of our language (term) embodies the reference to a
thing (Ens), together with some character of it. Often this polarity is expressed by Leibniz as
the relationship between the one “thing” and the plurality of “terms”: the latter being concepts,
2 1 4   •  The Problem of Univer sals in Early Modern philosophy

of inherence with respect to the ontological subject. Thus, according to a com-


mon reading, to say that Socrates is wise is to say that wisdom inheres in Socrates.
Can this way of expressing be dispensed with, or does it maintain its legitimacy?
In other words, if abstract talk can be in principle dismissed (or reinterpreted)
in logical and epistemological contexts, can it also entirely be given up when one
is concerned to give an account of the ontological basis of predication? A classic
argument for realistic assumptions has always been the alleged need to rely on
them in order to make sense of the truth conditions for our predicative statements.
This question—​clearly coinciding with that of the ontological status one
is ready to concede to the accidents—​lies at the center of the most important
text devoted by Leibniz to the topic of abstract talk, the De abstracto et concreto,
from the end of the 1680s.41 He begins here by illustrating the technical work-
ing of oblique predication of abstract terms, without committing himself to any
ontological interpretation. But the further development of reflection imposes
a decision. Leibniz works out in this context a general definition of inherence
in mereological terms, trying to overcome the Aristotelian limitation that had
embarrassed Hobbes: A is in B iff its reality is part of the reality of B.42 But then,
the clarification of the ontological import (the “reality”) to be attributed to the
accidents cannot be further avoided.
Leibniz is well aware that the oblique predication of abstract terms, by itself,
strongly suggests a realist interpretation of accidents. However, he emphasizes
that this way of expressing predication is more typical of philosophical jargon
than of ordinary language. And finally, he dismisses the realist interpretation by
appealing, again, to the paraphrase rooted in Hobbes’s analysis, which he had
already used in the Generales Inquisitiones to provide a unified theory of con-
cepts and truths. This paraphrase is now expressly finalized for the purpose of
ontological deflation. If one systematically substitutes the forms “to be just,” “to
be white” and so on—​now called “logical” (or “conceptual”) abstracta—​to the
corresponding “philosophical” (or “real”) ones, like “justice” or “whiteness,” then
one can preserve the same semantic and expressive power without the undesired
ontological costs:

Two kinds of abstract terms can be conceived of, the former being prior
to the corresponding concrete terms, the latter posterior. The former are

which simply express our ways of referring to the same thing. In contrast, in abstract talk these
concepts or ways of referring are treated as if they were things.
41. De abstracto et concreto, A 6.4:987–​96.
42. On the Leibnizian analysis of the inesse relationship, see Di Bella 2014.
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  215

commonly known, and they are taken as some realities in the subjects, of
which the concrete things partake. … But I have introduced in philoso-
phy some abstract terms that are posterior to the corresponding concrete
ones, e.g., 'the being wise' … This type of abstract terms is very useful to
my way of reducing philosophical reasoning to a calculus.43

For a present-​day reader, it is tempting to see here an antecedent of the Quinean-​


style ontological deflationary moves by way of linguistic paraphrases. We know
also the methodological objections that can be raised against the adequacy of
these techniques in order to solve ontological disputes. Yet although the para-
phrase plays an important role in Leibniz’s deflationary strategy, it is important
to understand that its success is not, by itself, the reason for the final restatement
of his antirealist option. In the De abstracto et concreto, Leibniz puts forward some
independent metaphysical reasons that would decide the game against realism.
They bear basically on the way of understanding (1) the particularity of accidents,
and (2) the phenomenon of change.

9.3.6. Tropes and Change: Squaring Counts


with Realistic Assumptions
In the last decades, interpreters have recognized the presence of the traditional
theme of individual accidents also in modern philosophers. Leibniz, especially,
has been credited as one of the most conscious supporters of this doctrine.
Although this opinion is basically true, I think that it requires some important
precisions. Consider the explicit remarks that Leibniz devotes to this topic in the
De abstracto et concreto:

It is very doubtful whether abstract items are true real Beings. If they are,
they are never complete even within a single individual. If this wisdom,
which is in Seneca, involves a relation to his riches, by the very fact it does
not involve the formal concept of wisdom only, that has been brought
to our attention when we have coined such an abstract term. If, then, we
were to abstain from accidental considerations, in this case it would not be
clear at all how two individual wisdoms of the same grade, or two equally
intense heats, or two numbers “two” differ. Or better still, we shall com-
pare these Metaphysical beings with Mathematical ones, e.g., the Circle

43. De abstracto et concreto, A 6.4 :992. The contrast between “real” and “logical” abstract terms
returns later in the New Essays. See bk. 3, chap. 8 (GP 5:314).
2 1 6   •  The Problem of Univer sals in Early Modern philosophy

insofar as it is abstracted from matter, or the parts of time insofar as they


are separated from their contents; exactly as we cannot explain how two
equal circles differ, which are separated from matter, or two hours, so we
cannot explain how two similar and equal instances of wisdom, or two
instances of heat having the same nature and grade differ. … If one were
to claim that numerically the same wisdom, or numerically the same heat
are in one subject and in another one, one will be confuted by the fact that
the wisdom of the former can be said to perish, while the wisdom of the
latter still persists. (A 6.4:991)

The realist view, according to which two objects can partake the same constitu-
ent, is dismissed. But also the standard view of tropes is challenged:  contrary
to the accepted view concerning individual accidents, Leibniz is saying that an
expression like “this red(-​ness)” is not able to pick out something like a “trope”
of red, taken in isolation from other tropes. Such a trope would not be a truly
individualized property. It could be individuated only through the copresence of
all other accidents belonging—​at least at a determinate temporal position—​to
this concrete individual object.
As a matter of fact, we know that for Leibniz the accidents of a substance
amount to its successive states. In the more concrete psychological terms already
used in the Discourse on Metaphysics (1686), and then developed in the later mon-
adological account, they are properly perceptual states, which are infinitely com-
plex. But then, every aspect or property, if cut off and isolated from this complex
unity, turns out to be something abstract, hence general. Therefore, it is—​as is
shown by the comparison to certain mathematical notions—​a purely conceptual
and/​or linguistic device, apt for the sake of economy in expression and of reason-
ing, but deprived of any referential force.
The idea of the successive states of a substance leads to the second problem,
which turns out to be a crucial test for the ontology of predication. This prob-
lem was already present in the earlier nominalist tradition. In fact, it was the all-​
pervading phenomenon of change that imposed a limit to Ockham’s reductionist
program applied to categories, by providing a criterion for distinguishing the
items in the category of Quality which can and cannot be dismissed: those quali-
ties must be recognized as real, which are required in order to give an account of
change. The reality of change presupposes in its turn that while the same subject
persists, something real arises and perishes within it.
Leibniz shares the same conviction: accordingly, in the De abstracto et concreto,
he recognizes a logical connection among (a) the reality of change (a matter of
experience); (b) the “hypothesis of real accidents” (an ontological assumption);
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  217

and (c) the usage of “philosophical abstract terms” (the indispensability of a cer-


tain class of linguistic expressions, with the semantics they seem to incorporate).
So far, the test of change seems to work also for Leibniz as a powerful obsta-
cle to ontological deflation and nominalistic analysis. But in the same text its
significance is dramatically overturned, and this, again, as a by-​product of the
“mereological” interpretation of inherence put forward there. If real accidents
are somehow “parts” of substantial reality, the substance itself will be open to the
mereological apories to which material objects are committed because of their
mereological composition, namely, the type of apories symbolized by the case of
Theseus’s ship, a well-​known puzzle for Leibniz. The way out will be to give up (b),
and to replace (c), also in these contexts, with (c´) Leibniz’s “logical abstracta.”

9.3.7. Leibniz’s Precautionary Nominalism: Between Austere


Nominalism and Ontological Epoché
If the semantic strategy adopted is clear, with what ontology are we left? The
question is taken up again in the draft entitled De realitate accidentium,44 where
the same aporia of change is illustrated in a more detailed way than in De abstracto
et concreto. The solution of giving up real accidents is qualified here by the sugges-
tive label of “precautionary nominalism”:

So far, I do not see any way of avoiding these difficulties, except by consid-
ering abstract terms not as standing for things, but as abbreviations [com-
pendia loquendi]: in the sense that, when speaking about heat, there is no
need to refer to some undetermined subject … ; and in this sense I am
a nominalist, at least for precaution [per provisionem]. Therefore, I  will
say that substance does change, i.e., that it has different attributes at dif-
ferent times; this cannot be doubted. Whether some reality perishes and
arises with change, however; or whether there are different realities within
substance, that are the foundations of different predicates: all this needs
not be asked, and if it is, it is not easy to determine. But it is enough to
take substances as things and to tell truths about them. Geometers do not
make use of definitions of abstract concepts, but they reduce them to con-
crete terms: thus Euclid does not make use of the definition of proportion
[ratio] … , but he relies on a definition where he explains which things

44.  De realitate accidentium, A  6.4:994–​96. For a detailed analysis of this argument, see
Mugnai 2005.
2 1 8   •  The Problem of Univer sals in Early Modern philosophy

are said to have one with the other the same, or less, or bigger proportion.
(A 6.4:996)

What are the ontological commitments of this type of attitude? Certainly, they
are more austere than those of standard nominalism in the Ockhamist tradition,
and closer to what is currently labeled as “austere nominalism.” No reference to
tropes is made, but only to concrete things, whereas abstract terms are taken as
linguistic devices suited for expressive economy.
This economical function was already stated in the De abstracto et con-
creto. Recall that Leibniz assimilates there the abstract consideration of an
isolated property (this wisdom) to the case of mathematical notions. He con-
tinues by showing that the comparison pointed to the functional role they
played:

Meanwhile, however, we can avail ourselves of such notions, to make our


reasoning easier, in the same way as we make use in Algebra of imaginary
roots, and in Geometry, maybe, of both infinite and infinitely short lines.
(De abstracto et concreto, A 6.4:991)

To give an account of the formation of these predicative concepts, Leibniz could


always rely on an explanation in terms of “resemblance classes.”45
Admittedly, the explanation is not entirely satisfying, insofar as, according to
an entrenched ontological intuition to which Leibniz wholeheartedly subscribes,
relations in general (hence, also resemblance) and their changes must have, in
their turn, a foundation within the related things. This is why this radical nomi-
nalistic option is not felt as a final positive solution, but rather as a kind of onto-
logical epoché.
I have translated per provisionem as “for precaution.” The Latin term, however,
can be rendered also as “provisionally.” The two senses are compatible. Avoiding
stronger ontological commitments is a prudential strategy, at least insofar as we
are not able to further specify the truthmakers for our true predications con-
cerning concrete particular objects. In any event, Leibniz’s endorsement of this
strategy reflects a sympathy for nominalism and a confrontation with Hobbes’s
program that accompanied his thought constantly from the Preface to Nizolius
to his later years.

45. Leibniz explicitly subscribes to the explanation of universals in terms of resemblance. See


note 23 above.
Some Perspectives on Leibniz’s Nominalism and Its Sources  •  219

Abbreviations

A = Leibniz 1923–​; cited by series (Reihe), volume (Band), and page.


EW = Hobbes 1962; cited by volume and page.
GP = Leibniz 1875–​90; cited by volume and page.
L = Leibniz 1969; cited by page.
OL = Hobbes 1961; cited by volume and page.
P = Leibniz 1966; cited by page.
10 L O C K E ’ S E S S AY A N D L E I B N I Z ’ S
NOUVEAUX ESSAIS

Competing Theories of Universals

Martha Brandt Bolton

Locke, in his Essay concerning Human Understanding, maintains that


everything that exists is particular, and Leibniz, in the Nouveaux Essais,
concurs. With regard to universals, Locke urges a form of nominalism:
“General and Universal, belong not to the real existence of Things; but
are the Inventions and Creatures of the Understanding, made by it for
its own use, and concern only Signs, whether Words, or Ideas” (Essay
3.3.11).1 Unlike less radical nominalist formulations, this one stresses
the dependence of universals on specifically human practices. This is
where Leibniz disagrees. Nouveaux Essais repeatedly objects to the view
that species depend on human ideas and languages. Leibniz’s theory of
general natures is not succinctly stated in the dialogue but it emerges
as the conversation moves from one topic to another. Its implications
are pervasive because it is integral to the doctrine of innate knowledge
which runs through most issues in dispute between the two philoso-
phers. The nativist debate will not be discussed here, but several closely
related controversies are central to the topic of this chapter. In what fol-
lows, four topics are discussed: (1) the nature of general ideas and their
role in perception and thought; (2) the foundation of the classification
of substances and how, if at all, it differs from that of the classification
of mathematical, moral, and other sorts of entities; (3) dimensions
and their diversification; (4) the place of general truths in the order of
explanation and justification of assent.

1. Quotations and citations of the Essay are taken from Locke 1975. They are iden-
tified by book, chapter, and section. Quotations from Nouveaux Essais are based
on Leibniz 1996 with minor changes.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  221

10.1. General Names, General Ideas, and Abstraction


In view of his doctrine that human ideas originate in experience of wholly par-
ticular things, Locke needs to explain how general names come to be used. On
his theory of language, a name immediately signifies an idea in the mind of the
one who uses it. The name denominates whatever is represented by the idea but
in using the name a speaker’s primary intention is to communicate the idea (Essay
3.1.1–​3). General names signify general ideas formed by abstraction from particu-
lar ideas received in experience:

Ideas become general, by separating from them the circumstances of


Time, and Place, and any other Ideas, that may determine them to this
or that particular Existence. By this way of abstraction they are made
capable of representing more Individuals than one; each of which, hav-
ing in it a conformity to that abstract Idea, is (as we call it) of that sort.
(Essay 3.3.6)

An idea that somehow abstracts from time, place, and other circumstances is not,
so far, general. It becomes general because a contingent act of the mind assigns it a
representative function: “their general Nature [is] nothing but the Capacity they
are put into by the Understanding, of signifying or representing many particulars.
For the signification they have, is nothing but a relation, that by the mind of Man
is added to them” (Essay 3.3.11). The idea is made to signify just those things
which conform to it in some way fixed by the mind which may, but need not, be
an easily discerned similarity in appearance.
The experience of similarities among several particular things stimulates a
mind to abstract the qualities of one thing in respect of which it resembles the
others. The ability to recognize similar things is taken for granted; the sole aim
is to explain the semantic generality of ideas and names. At first, a child has only
ideas of particular individuals and only particular names:

Afterwards, when time and a larger Acquaintance has made them observe,
that there are a great many other Things in the World, that in common
agreements of Shape, and several other Qualities, resemble their Father and
Mother, … they frame an Idea, which they find those many Particulars do
partake in; and to that they give, with others, the name Man, for Example.
And thus they come to have a general Name, and a general Idea. Wherein
they make nothing new, but only leave out of the complex Idea they had of
Peter and James, Mary and Jane, that which is peculiar to each, and retain
only what is common to them all. (Essay 3.3.7)
2 2 2   •  The Problem of Univer sals in Early Modern philosophy

What is “common to them all” is not any one element contained in the complex
ideas of different substances, but just that each such complex has its particular
conformity relation to one and the same general idea.2
Although it is clear that selective consideration of some components of the
child’s perceptual idea of, say, her father drives the process, it may not be clear
what the “separation” of qualities amounts to. As George Berkeley understands
it, two qualities are “separated” just in case one is conceived in the absence of
the other; he objects that no quality can be conceived in isolation from all other
qualities; for example, softness is inconceivable without warmth, coldness, or
some other tactual quality. Evidence that Locke holds the untenable view is
drawn from his famous remark that the general idea of a triangle is neither right-​
angled, nor scalene, nor obtuse, “but all and none of these” (Essay 4.7.9). Yet this
description is deliberately designed to show that forming the abstract idea of a
triangle and grasping the basis on which it is to represent other things requires
some skill. Berkeley assumes that “all and none of these” describes the basis of
representation, but it should be understood to describe what is represented—​all
triangles, inasmuch as they have three angles, and none, insofar as its angles have
a certain ratio. In fact, other passages in the Essay make the point that many quali-
ties which are indeed distinct cannot be conceived without others which attend
them in perception:

’Tis true, Solidity cannot exist without Extension, nor can Scarlet-​
Colour exist without Extension; but this hinders not, but that they are
distinct Ideas. Many Ideas require others as necessary to their Existence
or Conception, which yet are very distinct Ideas. Motion can neither be,
nor be conceived without Space, and yet Motion is not Space, nor Space
Motion. (Essay 4.2.13)

There is no reason to think his doctrine of abstract ideas conflicts with this obser-
vation.3 Overall, it seems that Locke is like Berkeley in supposing that the general
idea of a triangle is a particular triangle partially considered.
To Leibniz’s mind, Locke misconstrues both the properties and origin of
abstract ideas. Precisely because abstract ideas subsume many things, they dif-
fer from sensory presentations in kind as well as function. Abstraction “requires
attention to the general apart from the particular, and consequently involves
knowledge of universal truths” (NE 142). The idea of a triangle enables one not

2. Essay 3.3.9.
3. See Ayers 1991, 1:248–​52; also Ayers 2008.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  223

just to think of all and only triangles, but to think of them as subjects of certain
general truths. For this, one needs a range of abilities proper to intellect—​abilities
to reflect on oneself, to use general notions, to grasp propositions, predications,
and logical relations. Beasts share with us the faculty of sense perception, but
abstraction requires an entirely different faculty of cognition.
In human beings, as Leibniz has it, intellect continually affects the guise
in which perceptual objects are presented to awareness. Unlike lower animals,
human beings always have thoughts of the objects they distinctly perceive:

So “understanding” in my sense is what in Latin is called intellectus, and


the exercise of this faculty is called “intellection,” which is a distinct per-
ception combined with a faculty of reflection, which the beasts do not
have. Any perception which is combined with this faculty is a thought,
and I do not allow thoughts to beasts any more than I do understanding.
(NE 173; also 50)

Thoughts have propositional content; they contain two or more general ideas
(concepts, notions) and a connecting relation in virtue of which the whole
is capable of truth or falsity.4 Every distinctly perceived object is conceived as
a thing of some sort. Ideas, as he understands them, are never abstracted from
objects of mere sense perception simply because perception as such is nonconcep-
tual in its mode of representation.5
Leibniz holds that an idea represents the things it does intrinsically and essen-
tially, not as a result of an extrinsic contingent act of a human mind, as for Locke.6
The abstraction is an intentional content incompletely determined relative to

4.  As used in Nouveaux Essais, ‘idea’ refers (with very few exceptions) to modifications of
human minds. In this use, it is ambiguous. (1) It may mean a concept, or notion, i.e., an act of
conceiving an entity; such acts are constituents of thoughts of that thing. Typically a person
conceives a thing with the aid of words, symbols, images, or other signs that represent it (see
NE 77, 185–​86, 488, and especially “Meditations on Knowledge, Truth and Ideas,” A 6.4:590–​
91; AG 26:  NE 254 explicitly refers to this published article). (2)  ‘Idea’ may also mean an
innate disposition to form such concepts (on the distinction between “ideas,” in this sense,
and concepts, see NE 430, 304; “Discourse on Metaphysics,” 27; A  6.4:1572; AG 59). The
word is infrequently used in Nouveaux Essais to refer to entities in God’s understanding (e.g.,
NE 397), although in other works it is often used this way; indeed, God’s understanding is
sometimes called “the region of ideas” (Mugnai 1990). In this chapter, ‘idea’ is used only to
refer to concepts of actual and possible things; native dispositions are called ‘innate ideas’, and
possible things as known by God are indentified by phrases such as ‘God’s understanding’ or
‘region of ideas’.
5. See Bolton 2006.
6. NE 447, 57, 110.
2 2 4   •  The Problem of Univer sals in Early Modern philosophy

anything which might exist; with reference to the actual world, “abstraction is
not an error, as long as one knows that what one is pretending not to notice is
there” (NE 57). It homogenizes endlessly many differences which distinguish
actual things from each other.7 Without this simplification there could be no
theory or explanation according to Leibniz.
Leibniz contends that the morphology of Locke’s general ideas makes them
unsuited to be subjects of definition, reasoning, and demonstration. As he sees
it, general ideas according to Locke can be nothing but images extracted from
sense impressions. They can be matched and manipulated in imagination; they
can support generalizations from an examination of a certain number of similar
things, but nothing more than this empiric’s knowledge.8 They are not subject
to exact definitions, which he takes to be the means for discovering and explain-
ing the properties of kinds. Comparison of a mathematician’s idea of a decagon
with that of a laborer illustrates the difference. The mathematician knows how to
construct a decagon, knows the number of its sides and many of its properties,
but may be unable to tell a nine-​sided from a ten-​sided figure just by looking. By
contrast, the laborer’s idea-​image enables him to recognize a decagon as soon as
he sees it and easily distinguish it from other figures, but he may not know the
number of its sides (NE 261–​3, 137). If Locke purports to attain a precise science
of geometry by reasoning about images, he is badly mistaken, Leibniz contends
(NE 451).9
It is true that Locke’s general ideas are images in part, but they are given an
interpretation. An abstract idea may be made to represent on the basis of any sort
of similarity to the things it represents. In the case of general ideas of warmth, yel-
lowness, bitterness, and the like it may do to say that conformity is an easily recog-
nizable resemblance, but Locke plainly thinks it is more complicated in the case of
a triangle. The idea of a triangle is a simple mode, a complex idea which contains
two or more simple ideas in a relation which unites them.10 To ascertain whether
or not a particular conforms to it, one may need to discriminate its parts and their
relations, count its parts, or compare the figure with other things.11 Locke’s ideas
of figures are not amenable to the demonstrations envisaged by Leibniz, that is,
deductions in logical form. But the critical edge of Leibniz’s remark that there

7. There are no lowest species of physical things; see NE 230–31, 307.


8. NE 50–​51, 271, 475.
9. NE 137, 261–​63, 375.
10. Essay 2.12.4; 2.13.5.
11. See Essay 2.16.5–​7; 2.29.13. Mackie 1976, 114, suggests that recognizing a triangle involves
checking against “something like rules.”
Locke’s Essay and Leibniz’s Nouveaux Essais  •  225

is no image that represents all triangles overlooks resources by which Locke can
explain the representative function of general ideas of geometrical objects and
other simple modes.12 In sum, we might note that although the two philosophers
can agree that lower animals lack the ability to have abstract ideas, Locke ascribes
this to the human capacity for spoken language and the ability to grasp more
complexity than other animals, whereas Leibniz holds that lower animals have
no cognitive ability other than sense perception, a mode of cognition restricted
to things which have no logical structure. He attributes to human minds, in addi-
tion to the sensory faculty, the innate ability to comprehend concepts, concept-​
containment relations, and formally valid inferences.13

10.2. The Foundation of Kinds in Reality


For Locke, abstract ideas, general names, and the kinds they signify are “the
Workmanship of the Understanding”; without minds and their ideas, there
would be no kinds at all. More than that, kinds are “Inventions and Creatures,”
“Fictions and Contrivances” of the mind.14 Still, with a little care, we can form
general ideas which have the basis in reality required for knowledge of “Things as
they really are” (Essay 4.4; 2.30).
Kinds are traditionally supposed to be grounded on essences. Locke invokes
an established view about what the essence of a species does, namely, set the
boundary of the kind, as reason to reduce the essence to the signification of the
name used to stand for it.

For the having the Essence of any Species, being that which makes any
thing to be of that species, and the conformity to the Idea, to which the
name is annexed, being that which gives a right to that name, the hav-
ing the Essence, and the having that Conformity, must needs to the same
thing. (Essay 3.3.12)

What makes a thing a member of a kind, according to Locke, is nothing other


than its being correctly called by the name of the kind. He contends that the
boundary of a kind is defined, or determined, by nothing but conformity to the
general idea signified by a general name. He calls this the “nominal essence” to

12. NE 375.
13. Essay 2.11.5–​10; NE 142–​43.
14. Essay 3.3.11; 4.7.9.
2 2 6   •  The Problem of Univer sals in Early Modern philosophy

emphasize that members of a kind are collected by nothing but the signification
of the name. This is by no means the established view.
In the Aristotelian tradition, the essence of a species has more than one defini-
tion: some definitions specify the effects of the essence; the scientific definition
makes the essence known as the cause of its various effects; and laymen who have
little knowledge of either definition may use the name of the species to refer to
the essence. Locke’s anti-​Aristotelian treatment of kinds is shaped by a very dif-
ferent view of the purpose of sorting things and a much more pessimistic assess-
ment of the cognitive powers of human beings. He is, at the very least, doubtful
of a uniquely natural way of dividing things into species, or kinds. In any case, a
natural science of material things is not possible due to the limits of our cogni-
tive power; so the division of kinds has little if anything to do with potentially
discoverable explanations.15 Far from being bases of explanation, in his view, the
general ideas of species that are most useful for the study of nature are nothing
more than compendia of many painstaking observations.16 The primary purpose
of sorting things is convenience in communication and improvement of knowl-
edge in terms of efficiency.17 This largely explains why Locke maintains that with-
out language, there would be no kinds:

all the great Business of Genera and Species, and their Essences, amounts
to no more but this, that Men making abstract Ideas, and settling them in
their Minds, with names annexed to them, do thereby enable themselves
to consider Things, and discourse of them, as it were in bundles, for the
easier and readier improvement of communication of their Knowledge,
which would advance but slowly, were their Words and Thoughts con-
fined only to Particulars. (Essay 3.6.20; also 3.6.32 and 3.6.25)

Locke accounts for the widespread belief that the names of kinds of substances
stand for unknown real essences, not because they actually do so, but because the
assumption facilitates the use of the same general name by many speakers (Essay
3.6.48–​9).
It is in this spirit that Locke’s arguments against the established view should
be understood. Part of the evidence on offer is that different people use the
same general name to stand for different things: what is covetousness to one

15. Essay 4.3.25, 29.
16. Essay 3.11.24–​25; 3.6.30; 4.10, 12.
17. Essay 3.3.1–​4.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  227

person may not be to another (Essay 3.3.14). Locke’s theory is that a general
name signifies an abstract idea in the mind of the person who uses the name,
and the correct application of the name is determined entirely by this idea. We
do typically try to use a name to signify an idea like the ones other speakers
signify by the same name in order to be understood (Essay 3.2.4–​8). But the
authority of the speaker’s idea is said to be confirmed by discrepant applica-
tions of the same general name. Locke ascribes this to the absence of observ-
able indications of any naturally delineated kinds. As for species of substances,
it has been debated whether an odd issue of human birth is, or is not, in the
species man which, Locke maintains, would not occur if essences were not the
“uncertain and various Collection of simple Ideas” on the part of individuals.
Accordingly, the question how to denominate the creature must be settled with
reference to the idea signified by the name ‘man’. A community may agree about
what this signification is, but the issue cannot be decided on the basis of some
essence which no one knows; rather it is settled by tradition, convenience, civil
law, or the like. A further consequence of the theory is that an effective division
of kinds can be based on nothing but an abstract idea, “every distinct abstract
Idea, is a distinct Essence; and the names that stand for such distinct Ideas, are
the names of Things essentially different” (Essay 3.3.14).
In the Aristotelian tradition, essence is also supposed to be “the very being
of any thing, whereby it is, what it is” (Essay 3.3.15). As such, an essence has
a metaphysical role which Locke does not entirely deny. He allows that every
particular substance has a particular real inner constitution, not known by us,
in virtue of which it has all of its sensible qualities. But this is a far cry from
the “artificial constitution of Genus and Species” (Essay 3.3.15). This refers to
the scholastic view that the form present in the hylomorphic constitution of
an individual substance endows it with a species essence constituted by a genus
and difference (Essay 3.3.17). A  scholastic species essence endows individuals
with a set of essential properties but not their respective accidents, whereas the
inner constitutions Locke ascribes to particular substances mark no distinction
between essential and accidental qualities. This is because any such distinction
presupposes a classificatory scheme which, as Locke has it, is provided by noth-
ing other than abstract ideas: “All such Patterns and Standards, being quite laid
aside, particular Beings, considered barely in themselves, will be found to have
all their Qualities equally essential; … or which is more true, nothing at all”
(Essay 3.6.5). Furthermore, Locke contends that the scholastic theory of insep-
arable properties (propria) that belong to all and, strictly taken, only members
of the same species is effectively refuted by monsters in the species of animals
and changelings in the species man. Changelings manifest the bodily shape
2 2 8   •  The Problem of Univer sals in Early Modern philosophy

but not the rationality that are supposed to be inseparable properties of the
species man.18
By Locke’s non-​Aristotelian lights, the real internal constitution of a particu-
lar substance is the immediate source of all its sensible qualities; if it were known,
it would explain them. It contributes to determining the kind the substance
belongs to only by mediation of an abstract idea and general name.19 It is because
the real internal constitution of a given parcel of matter gives rise, inter alia, to the
qualities specified by the nominal essence immediately signified by “gold” that it
belongs to the species gold. In this capacity, Locke refers to features of the consti-
tutions of certain individual substances as the “real essence” of a kind.20

By this real Essence, I mean, that real constitution of any Thing, which is
the foundation of all those Properties, that are combined, and are con-
stantly found to co-​exist with the nominal Essence. . . . But Essence, even
in this sense, relates to a Sort, and supposes a Species: For being that real
Constitution, on which the Properties depend, it necessarily supposes a
sort of things, Properties belonging only to Species, and not to Individuals.
(Essay 3.6.6)

So although there are real essences of kinds founded on the real constitutions of
particular substances, they are just aspects of constitutions picked out by nominal
essences and therefore dependent on human ideas. The individual real essences
that conform to a nominal essence need not be alike; what matters are the observ-
able similarities selected by the general idea. The whole line of thought depends
on the principle that things cannot be ranked under general names by the stan-
dard of constitutions which are unknown to users of the name.21
For Locke, sorts of things in categories other than substances, that is modes
and relations, also have a real essence as well as a nominal one. Modes include
mathematical entities, spatial, and temporal measures and relations, moral enti-
ties, and social and cultural institutions.22 Locke recognizes nothing independent
of human language, either in the world or among supposed eternal entities, that

18. Essay 3.3.17; 3.6.6, 19.


19. Essay 3.3.13.
20. On individual as opposed to sortal essences, see Owen 1991.
21. E.g., Essay 3.3.1.7; 3.6.9, 14–​19.
22. The former are called “simple modes” and the latter “mixed modes.” Mixed modes have
no special part in issues regarding universals; in this chapter, “modes” refers to simple modes
unless otherwise indicated.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  229

sets the standard for, say, a triangle (Essay 4.11.14). It is just what it is defined
to be. In brief, the real and nominal essences of kinds of substances are always
distinct—​general ideas as opposed to aspects of inner mechanisms these ideas
pick out—​and the real and nominal essences of modes are always the same (Essay
3.3.18). Because we are certain of the real essences of modes, we are in position
to demonstrate indefinitely many general truths in mathematics, and other sci-
ences that deal with modes (Essay 4.3.18–​19). But as long as we are ignorant of the
real essences of kinds of substances, we can aspire to nothing more than probable
opinions about general truths with regard to them (Essay 4.3.23–​29; 4.15.4–​5).
As for what grounds the reality of kinds, Locke maintains that ideas of sub-
stances are intended to represent actual things which subsist by themselves on
the basis of what we know of them, their sensible effects. Ideas of substances are
composites which include ideas of a selection of sensible qualities which we take
to belong to one thing and the idea of substance. The idea of substance is either a
hidden corpuscular constitution from which the observable qualities result or the
“something I know not what” that supports the inferred solidity and extension
of the insensible particles that compose such constitutions.23 Ideas of substances
are real just in case there is one or more actual thing in which the selected com-
bination of qualities exists. By contrast, ideas of modes and relations do not carry
the assumption of representing anything which actually exists, although they may
happen to do so. Their reality consists in the possibility of something’s conform-
ing to them.24
Unconvinced by Locke’s efforts to show that kinds depend on names and their
criteria of application, Leibniz responds: “I cannot see that this follows: general-
ity consists in the resemblance of singular things to one another, and this resem-
blance is a reality.” Since his companion holds that abstract ideas are derived
from perceived similarities among things, “why not look for the essence of gen-
era and species there too?” (NE 292). General natures are similarities among
actual and possible particulars: “Insofar as you conceive the similarities amongst
things, you are conceiving something in addition, and that is all that universality
is” (NE 485). Locke agrees that similarities among actual things are objective,
but not so far universal. According to Leibniz, “Essence is fundamentally noth-
ing but the possibility of the thing under consideration” (NE 293). He suggests
that instead of distinguishing nominal and real essences—​one setting boundar-
ies of kinds, the other being the source of kind-​typical properties—​we do better

23. Essay 2.30.5; 2.23.1–​3; 4.4.12. For a defense of Locke’s account of general ideas of kinds
against contemporary objections, see chapter 11 in this volume.
24. Essay 2.30.
2 3 0   •  The Problem of Univer sals in Early Modern philosophy

to distinguish real and nominal definitions. This focuses attention on the ques-
tion whether a definition specifies a possible entity. The fact is that we some-
times formulate definitions that conceal contradictions, such as the definition
of “decahedron.” When we offer definitions, the possibility of their definienda
has to be proved before we can safely reason from them. This can be done in
three ways: by analyzing the definition back to simple indefinable ideas, as may
happen in mathematics;25 by a priori discovery of the cause, or possible mode
of constructing the purported thing; or by a posteriori observation that such a
thing exists.26
Leibniz maintains that an idea (concept) typically signifies a possible sort of
thing (essence) with the aid of a sensible character or symbol. If the character
implies no contradiction, the idea represents a possible entity. Otherwise the idea
is “false,” which is to say it is not an idea at all.27 The point about definitions
is brought against Locke. His nominal essences are answerable to what is pos-
sible, and he cannot deny that what is, and is not possible, is not up to us (NE
293–​94). Of course Locke is well aware that an abstract idea does not succeed in
specifying a kind unless it is possible for something to conform to the idea. But
for him, the field of possibilities consists of things which are, or might someday
be, represented by ideas formed by a human being. Leibniz’s domain of possi-
bilities is prior to human ideas and presupposed by them: “man’s combining or
not combining such and such ideas—​or indeed their being or not being actually
combined in nature—​has no bearing on essences, genera and species, since they
depend only upon possibilities, and these are independent of our thinking” (NE
293). They are dependent on God, “the source of possibilities and existents alike,
the one by his essence the other by his will” (NE 155).
To sum up, Locke contends that we are apprised of neither divisions made
by nature nor extranatural paradigms, so to set definite boundaries to kinds we
are thrown back on acts of abstraction and construction of ideas assigned to rep-
resent particulars similar to them in specified respects. He locates the reality of
kinds of substances in the actual existence of one or more particular things which
conform to our ideas of substantial kinds and the reality of kinds of modes in
the possibility of things that conform to our ideas of modes. But for Leibniz,
all kinds, ideas of kinds, and significant general names are founded on eternal
essences.

25. NE 268. For more detail, see “Meditations on Knowledge, Truth and Ideas,” A 6.4:589–​90;
AG 25–​6.
26. NE 294
27. NE 398 and especially “Meditations on Knowledge, Truth and Ideas,” A 6.4:589–​90; AG 26.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  231

10.3. Kinds of Simple Ideas and Sensible Qualities


In addition to general ideas of substances and modes, which are complex, the
Essay recognizes kinds of simple ideas. According to Locke’s theory of the origin
of ideas, simple ideas are basic; they can be acquired directly only by sensation or
reflection and all other ideas are combinations of them.28
Simple ideas of sensation, according to Locke, are the immediate effects of
external objects operating on our physiological systems. When perceiving a par-
ticular thing, we typically receive several simple ideas at once, for example, from
a piece of wax, one may have simultaneous sensations of warmth and softness
each of which is “simple,” “unmixed,” and “perfectly distinct” (Essay 2.2.1). To
form the general idea of warmth, a person need only selectively attend to that dis-
tinct appearance, assign it a name, and give it the function of signifying all similar
appearances—​an objective similarity the perceiver is antecedently able to recog-
nize. Names of “simple Ideas are perfectly taken from the existence of things, and
are not arbitrary at all” (Essay 3.5.17). The reality of simple ideas consists in their
being effects of the corpuscular structure of bodies around us; in sense percep-
tion, they represent their regular causes.29 But simple ideas are sorted into kinds
by appearance not by similarities among their mechanist causes.
Locke remarks that unless a general idea marks a precise boundary for a kind,
we can know no general truths about the kind.30 Precise general ideas are thus
more useful than others, but the fact is that on his account, general ideas of colors,
tastes, and the like are imprecise. Although Locke takes no notice of this, Leibniz
remarks on it and addresses it (NE 298, 321, 339). Perhaps Locke supposes the
general idea of yellow, for example, is a determinate shade of yellow abstracted
perhaps from a particular lemon and that it serves to represent all things exactly
similar to it. The problem is that similarity is never exact, as is shown when some-
one produces a third shade intermediate between two samples which, considered
apart from the third, appear exactly alike.
Leibniz does not object to the claim that we divide simple ideas into kinds on
the basis of appearances which are “simple to the senses,” but he maintains that
this is a substitute for definitions of the mechanist sensible qualities the ideas
represent. Definitions of this sort specify kinds which do not suffer from the
same imprecision as kinds of colors, etc., do. On Leibniz’s view, simple ideas of
sensation are confused perceptions of many insensible moving bodies. That is,

28. See Bolton 1990.
29.  Essay 2.30.2; 4.4.4. When abstracted, they are still taken to have external causes, Essay
3.4.2–​3.
30. Essay 4.6.4, 3.6.29, 3.9.3.
2 3 2   •  The Problem of Univer sals in Early Modern philosophy

sensory presentations, as we experience them, are not mere sensory indicators of


the presence of causes to which they have no affinity, as they are for Locke. On
the contrary, simple ideas of sensation have a certain structural resemblance to
the physical qualities they express (NE 131).31 Accordingly, different sorts of sim-
ple ideas are (confused) presentations of sorts of physical configurations which
can be studied empirically; we can frame partially distinct definitions of them
and with enough knowledge, we could, in principle, define kinds of simple sen-
sory ideas without reference to sensations (NE 299, 225). But because the causes
of sensations are infinitely complex, defining yellow or blue as corporeal arrange-
ments is an unattainable ideal. For Leibniz, sensations of yellow, and the like, are
not qualia; they are sources of information which have essences and are capable
of influencing our instinctive behavior.

10.4. Kinds of Material Substances


According to Locke, “Ideas of particular distinct sorts of Substances” unite several
simple ideas (of qualities) which are observed to “exist together, and are therefore
supposed to flow from the particular internal Constitution, or unknown Essence
of that Substance” (Essay 2.23.3). Ideas of man, horse, gold, water, and sun are
examples. But if we knew the inner constitutions of particular substances, our
ideas of their species would be very different—​just as the idea of a workman who
knows the inner mechanism of the clock at Strasbourg differs from the idea of
a countryman who observes only the outward appearances (Essay 3.6.3). But
although we may know the inner mechanisms of artifacts, we have little, if any,
prospect of knowing the unobservable inner constitutions of particular kinds of
substances.
The Essay adduces several arguments purporting to show that we sort sub-
stances into kinds on the basis of arbitrarily made nominal essences rather than
distinct metaphysically independent real essences.32 These arguments are elusive,
but they are the platform for Leibniz’s treatment of kinds of substances in the
Nouveaux Essais. In order to keep the debate in focus, we will stick close to the
texts. The overall line of thought might be analyzed in terms of four intercon-
nected strains of argument.

(1) We do not sort substances on the basis of real essences established indepen-
dently of human ideas, because we are ignorant of them. “Nor indeed can

31. All representations are “expressions,” according to Leibniz. See Swoyer 1995.


32. Essay 3.6.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  233

we rank, and sort Things, and consequently (which is the end of sorting)
denominate them by their real Essences, because we know them not” (Essay
3.6.9). In order to classify on the basis of scholastic species essences, we
would need to know the following: whether all things produced in nature
belong to some species; whether monsters are in species of their own; what
the essences of species are, or at least, all the sensible qualities that are insepa-
rable from each species essence, i.e., their properties. Knowledge fails us on
every point (Essay 3.6.14–​19; also 3.6.9). Nor do we sort substances in accord
with mind-​independent real essences. If we did, there would be no debate
over the species of strange creatures born of a human being (Essay 6.3.25, 27;
also 3.6.8). Locke repeatedly reverts to this example in an effort to explode
the presumption that “rational animal” expresses the known essence of the
species man. We sort humans and other animals as we sort other things,
by reference to the names we use. “If I should ask any one, whether Ice and
Water were two distinct Species of Things, I  doubt not but I  should be
answered in the affirmative: And it cannot be denied, but he that says they
are two distinct Species, is in the right” (Essay 3.6.13).33 It is right because
English speakers give one sortal name to solid and liquid gold.34
(2) Further difficulties for the view that our classification is based on species
constituted by nature are drawn from tradition according to which nature
allows no gaps in the series of forms. The continuity of species of living
things is illustrated by intermediate species which have qualities typical of
two or more other species; there are fish with wings, cold-​blooded birds,
porpoises which can’t live out of water, have warm blood and hog-​like intes-
tines (Essay 3.6.12). Beyond this, there is a continuity pertaining to species-​
typical qualities which “differ but in almost insensible degrees (Essay 3.6.12).
’Tis a hard Matter to say where Sensible and Rational begin, and whether
Insensible and Irrational end: and who is there quick-​sighted enough to
determine precisely, which is the lowest Species of living Things, and which
the first of those which have no Life?” (Essay 4.16.12). This line of argu-
ment is intended to suggest that supposing there are species established by
nature, there are vastly more of them than could be recognized in a workable
system of classification. Moreover, the attributes customarily supposed to
divide species of inanimate and animate, insentient and sentient, irrational
and rational things fail to do so because they make no definite distinctions
among species supposed to be precisely delineated in nature.

33. Also Essay 3.6.8.
34. Also Essay 3.6.22.
2 3 4   •  The Problem of Univer sals in Early Modern philosophy

(3) Propagation does not, as is often supposed, preserve real biological species
because members of different species can interbreed (Essay 3.6.23). Some
pairings between species regularly yield offspring apparently in a third spe-
cies, e.g., an ass and a mare produce a mule. In rare cases, it is said, copulation
by members of different species produces an offspring with a bizarre mixture
of features of both. The suggestion is that either there are creatures which
belong to no species or new species sometimes arise from preexisting ones,
and both alternatives challenge the general rule that lineage determines spe-
cies. Moreover, appeal to lineage to decide the species to which one creature
belongs only postpones deciding biological species by some other means.
(4) Locke’s most fundamental reason for saying that kinds of substances are
voluntary creations of human minds is that no paradigms are provided by
experience. It cannot be denied that “Nature makes many particular Things,
which do agree one with another, in many sensible Qualities, and probably
too, in their internal frame and Constitution” (Essay 4.3.36; cf. 4.3.37). Yet
because the qualities in respect of which many things agree do not recur in
discrete clusters and there are always further qualities to take into account,
no closure can be found in experience: “For the Union in Nature of these
Qualities, being the true Ground of their Union in one complex Idea, Who
can say, one of them has more reason to be put in, or left out, than another?”
(Essay 3.9.13; cf. 3.6.29). Even though we base ideas on repeatedly coinstan-
tiated qualities, we cannot include them all, nor do we even know them all.35
We have to decide where to stop, and to this extent kinds of substances are
human contrivances. The problem recurs at the level of the known inner
constitutions of watches, clocks, and other artifacts. If a species of clocks,
for example, is defined in terms of an inner mechanism, someone can always
add a device with claim to make a different kind.36

The assumption driving the fourth argument is that an archetype would deter-
mine the boundaries of a kind only if it exemplified all and only the qualities nec-
essary and sufficient for membership in the kind. Only the totality of the qualities
possessed by an exemplar could make it a model for all instances of a kind. Hence
Locke’s antirealist remark about genera: “If the number of simple Ideas, that make
the nominal Essence of the lowest Species, or first sorting of Individuals, depends
on the Mind of Man, variously collecting them, it is much more evident, that they
do so, in the more comprehensive Classes, … which are called Genera” (Essay

35. See Essay 2.31.8.


36. Essay 3.6.29, 36, 37; 3.19.13
Locke’s Essay and Leibniz’s Nouveaux Essais  •  235

3.6.32). If there were a natural archetype for the genus metal, there would be a
body which has only the qualities malleability and fusibility.
Nouveaux Essais portrays this portion of the Essay as stressing that substantial
kinds can be defined either by outer marks or inner structures, while insisting
that they cannot be defined by things we don’t know (NE 304). Leibniz argues
for a middle way. His case depends on his version of metaphysical essentialism.
In place of Locke’s antiessentialist claim that particular things are constituted in
nature independently of species, Leibniz maintains that “there are sorts of spe-
cies such that if an individual has ever been of such a sort it cannot (naturally, at
least) cease being of that sort” (NE 305).37 The qualification alludes to a threefold
distinction of the greatest importance to Leibniz: what is true of a thing neces-
sarily (in virtue of its essence alone), naturally, and accidentally. What belongs to
it naturally is “inherently appropriate” to the thing but its denial is not a contra-
diction.38 Nouveaux Essais claims knowledge of four attributes which are neces-
sary to the individuals that have them: thinking in an individual mind, extension
and motion in individual bodies, acting and being acted upon in individual sub-
stances, and rationality in individual human beings.39 The qualities by which spe-
cies of substances can be distinguished are natural, more understandable than a
mere accident.
According to Leibniz, Locke’s second argument is vitiated by an ambiguity in
“species” (NE 307). There are both mathematical and physical species. In the for-
mer, the least difference between instances makes a difference in kind and there
are lowest species. For example, circles are a lowest species because all circles are
exactly alike in respect of geometrical qualities. By contrast, physical species per-
tain to material changeable things. No two physical individuals are exactly alike,
according to Leibniz. So Locke is right that there are no examplars of physical
species, but “no one is asking for patterns of this sort” (NE 323). Locke’s further
point that substantial kinds are abstractions which are nothing but intentional
objects of thought is well taken, but it does not follow that they are dependent
on human thought.
Again, in connection with his first argument, Locke rightly observes that
samples of vitriol have inner constitutions which support not only the several
qualities they share, but also many qualities by which they differ.40 Leibniz main-
tains that the very fact that all samples share a certain number of qualities shows

37. NE 310, 317 suggests this may not be true of all species in nature.
38. NE 433–​34.
39. NE 305, 310, 313–​14; see Jolley 1982.
40. Essay 3.6.8.
2 3 6   •  The Problem of Univer sals in Early Modern philosophy

that vitriol is a real species, an essence established by nature. The fact that some
samples have qualities others lack only shows that the species has subspecies (NE
305). We can choose whether or not to classify things on the basis of the shared
qualities of the oil or shared qualities of species under it. That is, we decide which
kinds and level of specificity to use in our system of classification, but this is not
to say we choose the essences of kinds. Locke is charged with confusing these two
dimensions of choice.
Leibniz envisages such a multiplicity of species established in nature that
human beings are robbed of any opportunity for creating kinds: “No matter what
rules men make to govern how things are to be named … provided that the sys-
tem of rules is orderly (i.e., interconnected and intelligible) it will be founded
in reality”; whatever species we may imagine “have already been made or distin-
guished in nature—​which even encompasses possibilities” (NE 309). One might
question whether such liberality severs the connection between the combina-
tions of outward qualities which are exhibited by many things and a similarity at
the level of their inner constitutions. However, Leibniz subscribes to the general
principle that similar effects have similar causes. This is crucial to his middle way
of deciding the species of particular things. In accord with this principle, the fact
that bats and birds, for example, have a few qualities in common implies that their
inner structures are similar in a causally relevant respect—​abstract as it may be. In
view of the unthinkably many kinds of things established in nature, those recog-
nized in any useful system of classification should be judiciously chosen: “what-
ever we truthfully distinguish or compare is also distinguished or made alike by
nature, although nature has distinctions and comparisons which are unknown to
us and which may be better than ours” (NE 309). Great care and experience are
needed to make divisions among substances which accord with nature as closely
as the present stage of human knowledge allows.
According to Leibniz, empirical means for determining species of actual
things are more diverse than Locke supposes, and more tentative. Physical spe-
cies are decided by human beings, but not without concern for inner natures of
things. “It is for men to say” whether ice and water are in the same species. But
men decide it by a principle different from the Lockean procedure of referring
the question to names in a language. They have adopted the rule that when stuff
of a certain sort undergoes a change that is readily reversible, as happens with
water, gold, quicksilver, and table salt, it remains in the same sort (NE 308–​09;
325). This pits a physical disposition inferred from the observed behavior of a
substance against Locke’s established signification of names.
According to Leibniz, men have also adopted the rule that living things are
in the same species if they could have come from the same seed. In the case of
Locke’s Essay and Leibniz’s Nouveaux Essais  •  237

the human species, we require not just human lineage but also rationality he
claims. Accordingly, we presume the capacity to reason is present in offspring
of human parents even if it is not manifested over a long period of time. The
question about the humanity of a changeling is, then, whether its presumptive
rational capacity is impeded. It is true that there is presently no agreement about
the observable traits which are sufficient to settle the matter. Nevertheless, the
dispute concerns the creature’s nature and it is not out of reach of empirical
evidence (NE 309).
This responds to Locke’s point that in order to classify by scholastic real essences,
we would need to know, inter alia, whether monsters do, or do not, belong to the
species from which they descend. As Leibniz sees it, the question we want to answer
always concerns a particular monster: does it have the inner nature which is common
to its progenitors even though it lacks the outward signs?

But our uncertainty does not affect the nature of things: if there is such a
common nature, then the monster either has it or lacks it, whether or not we
know which. And if the monster does not have the inner nature of any spe-
cies, then it can be in a species of its own. But if the species we are interested
in did not have such inner natures, and if we did not particularly dwell on the
facts of birth either, then the boundaries of a species would be determined
solely by outward signs. (NE 311)

This sort of laxity is what Locke effectively recommends as Leibniz sees it.


“But then,” Leibniz continues, “you would have to prove, sir, that there is
nothing inner which is common to the whole of a species in cases where there
are outer differences.” Locke is countered on the strength of a presumption that
he does not share, that things with similar qualities have constitutions similar in
respect of the cause of these qualities. Further, Locke points out that interbreed-
ing casts doubt on the presumption that a creature is in the same species as its pro-
genitors. Moreover, he maintains that there is no clear meaning to the claim that
the inner constitution of a changeling is the same as that of human beings unless
we know how to identify human beings. Accordingly, he might argue that the
epistemically cautious procedure is to base the decision on observable features
because we can at least be certain of them. For his part, Leibniz takes the evidence
that members of different species produce offspring in a third species to be inclu-
sive (NE 315–​17). He holds that similarity among observable features indicates
a natural, if not necessary, similarity at the level of inner constitution. From his
perspective, Locke takes epistemic risk by classifying human births too casually.
Leibniz’s further efforts to adduce empirical evidence that substances with typical
2 3 8   •  The Problem of Univer sals in Early Modern philosophy

human traits have an inner nature which they pass to their offspring issue in the
same epistemic standoff (NE 311–​14).41 To Locke’s eye, they must look like spe-
cial pleading for a doctrinaire definition derived from an outdated metaphysical
model. To Leibniz, his opponent seems to have unreasonable objections to efforts
to explain observed behavior on the basis of inferred but unobservable constitu-
tions (essences) to the detriment of natural science.42

10.5. The Semantics of General Names


The philosophers’ most fundamental disagreement about the classification of sub-
stances concerns the signification of general names. This is the nub of Leibniz’s
response to the problem posed by the difficulty of knowing the inner constitutions
of things: “It is true that we cannot define a species in terms of something which
is unknown to us; but the outer features serve in place of it, though we recognize
that they do not suffice for a rigorous definition, and that even nominal defini-
tions in these cases are only conjectural and sometimes … merely provisional”
(NE 311). We can offer a less than rigorous definition of the species to which cer-
tain particular things belong even though we do not know, as Locke insists, their
inner constitutions, or essences. Leibniz contends that this is just as true of the
unknown essences of mathematical objects as those of substances.43 If this argu-
ment is to work, in view of Leibniz’s concession that we cannot define a kind in
terms of something we don’t know, he must suppose that a nominal definition
provides a sort of knowledge of the inner essence of a species. Of course, it does this
only if the entity it purports to define is possible. But it is now plain that the fact
that the nominal definition does not show the possibility of what it defines (even
if it is possible) is connected with the way in which it makes an essence known.
As Leibniz puts it, the essence of gold constitutes gold and gives it “the sen-
sible qualities which let us recognize it and which make its nominal definition,”
whereas if we could explain this inner constitution “we would possess the real,
casual definition” (NE 294). The proper distinction between real and nominal
definitions is the distinction between essence and property.44 Leibniz likens the
diversity of nominal definitions of one essence to drawings of the same town from
different perspectives (NE 294). A nominal definition represents, or expresses, an
essence from the perspective of one of its properties, or in the case of gold, its

41. Cf. Essay 3.6.22.


42. Cf. Essay 4.12.13 and NE 450–​51, 455.
43. NE 347.
44. NE 295.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  239

effects—​malleability, ductility, solubility in certain acids, and so on. By contrast,


the real causal definition of gold would provide knowledge of the essence (physi-
cal constitution) as cause of these effects.45 Nouveaux Essais mentions a third sort
of definition: “I once defined ‘adequate idea’ (or ‘perfect idea’) as one which is so
distinct that all its components are distinct; the idea of a number is pretty much
like that” (NE 266).46 A  perfect idea contains nothing but simple indefinable
ideas. All three sorts of definitions, or general ideas, provide information about
the essence of a species but with different degrees of perspicuity, or distinctness.47
The doctrine that one essence, or kind, can be defined in more than one way
is anathema to Locke, who maintains that a general idea is the sole determinate
of the boundaries of the kind it signifies; it is, then, impossible that different gen-
eral ideas, or definitions, signify the same kind. Two definitions are two different
ideas, two distinct nominal essences.48 This fundamental dispute between Locke
and his opponent comes to a head in an exchange about the meaning of a general
proposition about gold.49
Locke contends that the proposition expressed by “All gold is fixed” means
either that being fixed is included in the nominal essence of gold or that fixedness
is a property of the substance, itself. In the former case, the proposition merely
states the signification of the name of the kind. In the latter case, the significa-
tion of the name is rendered obscure and uncertain, because we cannot tell which
things are gold, or which things are supposed to have the quality fixedness (Essay
3.6.50). Without an explicit definition of “gold,” we lack an idea of what the word
signifies, so it has no definite signification. In opposition, Leibniz contends that
the inner constitution of gold is indicated by outward signs in addition to fixity,
so that to say all gold is fixed is like saying that the heaviest of bodies is also one
of the most fixed, which is evidently significant (NE 312). Furthermore, it is cer-
tain because it has been found true countlessly many times (NE 404). Even so,
empirical discoveries could lead us to revise our definitions of this essence (NE
402–​05). Locke would say that a definition cannot be revised but only discarded
for a more useful one.

45. NE 267; also see Leduc 2006.


46. The allusion is to “Meditations on Knowledge, Truth and Ideas,” A 6.4:585–​92; AG 23–​27.
47.  On distinct as opposed to confused ideas, see NE 255–​56, 266–​67; and especially
“Meditations on Knowledge, Truth and Ideas,” A 6.4:585–​92; AG 23–​27.
48. Essay 3.3.14, quoted above; also 3.2.3.
49. The opposed positions might be likened to what are now called “internalist” and “external-
ist” theories of the meaning of kind terms and the content of mental states, but the historical
issue is framed very differently. The historical positions are compared to descriptive and refer-
ential theories of kind terms in Ayers 1981; cf. Jolley 1982, 145–61.
2 4 0   •  The Problem of Univer sals in Early Modern philosophy

The parties to the debate also disagree over the signification of general names
of mathematical objects and other modes. Locke’s theory that the idea signi-
fied by the name of a kind of mode is both its nominal and real essence implies
that the idea is the source and explanation of general truths about the kind.
Knowledge of theorems about triangles is real because the idea of a triangle is
real by his standard of internal consistency.50 He allows that such propositions
are called “Eternal Truths,” but only for the naturalistic reason that “being once
made, about abstract Ideas, so as to be true, they will, whenever they can be sup-
posed to be made again at any time past or to come, by a Mind having those
ideas, always actually be true” (Essay 4.11.14). By Leibniz’s lights, this is a weak
and wholly inadequate theory of necessity. He maintains that necessary truths,
such as we find in geometry, would be true even if nothing contingent existed.
On his theory the things signified by the ideas contained in such truths belong to
the region of ideas: “This … brings us at last to the ultimate foundation of truth,
namely to the Supreme and Universal Mind, who cannot fail to exist and whose
understanding is indeed the domain of eternal truths” (NE 447).

10.6. Dimensions and Determinations


In his second argument against the presumption that things are classified in spe-
cies determined by nature, Locke adopts the ancient doctrine that nature allows
no gaps in the species of things. In response, Leibniz observes that some things
are intermediate between two species in some respects but not others. He says
he has reason to think there are possible species which are incompatible with the
series of things God chose to create, but endorses the Law of Continuity which
states that “nature leaves no gaps in the orderings she follows” (NE 307). Applied
to sensible qualities, the law implies that if there are two things that have a qual-
ity in different degrees, then there is, somewhere in the universe, something that
has it in an intermediate degree (NE 60). The totality of actual things which
have a given quality in various degrees constitutes a continuous ordering but,
Leibniz maintains, some qualities instantiated in our limited part of the uni-
verse exhibit gaps. But although experience does not reveal as much continuity
as Locke reports, Leibniz acknowledges the problem that would present:

there might be insensible transitions from one species to another, and tell-
ing them apart might sometimes be rather like the problem of deciding
how much hair a man must have if he is not to be bald. This indeterminacy

50. Essay 4.4.5–​10.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  241

would hold even if we were perfectly acquainted with the inner nature
of the creatures in question. But I  do not see that this prevents things
from having real essences independently of our understanding, or us from
knowing them. It is true that the names and the boundaries of species
would sometimes be like the names of measures and weights, where there
are fixed boundaries only insofar as we choose them. (NE 321)

Leibniz sees no contradiction in saying that conventionally established measures


have essences which are real independently of human understanding: “ordinarily
the boundaries of species are fixed by the nature of things … I do admit though
that there are some notions which involve a truly arbitrary element: for example, a
one-​foot length, for since a straight line is uniform and indefinite[ly long] nature
does not indicate any boundaries in it” (NE 302). The threat that, say, the con-
cept of gold might be logically tainted in the same way as the concept of a bald
man can be forestalled if we can set the boundaries of kinds with reality grounded
in eternal possibilities as we fix standards of measurement according to Leibniz.51
For Leibniz, the idea of a one-​foot length is the idea of a uniform straight line
in which a cut is arbitrarily made, that is, without basis in the line. According to
him, there is no uniform line in nature. Its uniformity signals its abstraction; it
lacks the diversity of an extension no two parts of which are exactly the same (NE
57, 110). The line is “a simple and uniform continuum,” as is time (NE 152). The
metaphysical status of a straight line is the same as that of space: “[Space] is a rela-
tionship: an order, not only among existents, but also among possibles as though
they existed. But its truth and reality are grounded in God, like all eternal truths”
(NE 149). It does not differentiate distances or positions, but it underlies the pos-
sibility of doing so (NE 154, 155). Such passages sketch a two-​layered structure in
the region of ideas: an undifferentiated ordering of all possible spatial positions at
the highest level of abstraction and a level of lesser abstractness comprising possi-
ble entities which partially realize it.52 As Leibniz puts it: “anything which is con-
tinuous involves an infinity, from which selections must be made” (NE 385). The
idea of a one-​foot line signifies a physical standard of length selected by human
choice, but still the possibility of the length is prior to and independent of the act
of selection. To return to the problem of defining kinds in terms of qualities that
form a continuous series, it is open to us to select a physical standard for a degree
on the continuous series of degrees of a given quality; for instance, a particular

51. On Leibniz’s treatment of concepts with indeterminate boundaries, see parts of NE 302 not
quoted above, NE 311. Also Levey 2002.
52. On ideal space by contrast with its actual realizations, also see NE 127.
2 4 2   •  The Problem of Univer sals in Early Modern philosophy

gold coin could be preserved and designated as the standard for comparative
measures of hardness, heat conductivity, and the like. The suggestion is that if all
went well, ideas of such conventionally established standards might suffice for
nominal definitions of species. In fact, however, Leibniz exploits local gaps in the
continuity of degrees of heaviness and ductility in the nominal essence of gold he
proposes (NE 400).
The Preface to Nouveaux Essais indicates that ideal space and time have a
place among genera; they are not traditional universals, but closely akin to them.
According to the Preface, every subject of qualities or modes has an attribute or
“permanent main predicate.” Following roughly Cartesian lines, there are two
main predicates with this status:  the faculties of perceiving and acting are one;
extension and motion, another. Permanent main predicates are called ‘genera’ but
divided into two importantly different types: logical (or ideal) and physical (or
real). Both genera organize and collect a range of diverse affections that depend
on the main predicates and the subjects they characterize. The difference concerns
the relation between the genus and its variations and the relations among its vari-
ants themselves.

Things which are of the same physical genus, or which are homogenous,
are so to speak of the same matter and can often be transformed from one
into the other by changing their modifications—​circles and squares for
instance. But two heterogeneous things can belong to a common logical
genus, and then their differentiae do not consist in mere accidental modi-
fications of a single subject or of a single metaphysical or physical matter.
(NE 64)

Space and time are ranked as physical genera; relating this to the two permanent
main predicates, space is diversified by modifications of extension, whereas time
is varied by perceptions, actions, and motions. By contrast, continuous quantity is
a logical genus of which space, time, and uniform matter are distinct species. The
logical genus is divided by differentia which are essential to it: three-​dimensional
structure, on one hand, and linear structure, on the other. Neither species can be
transformed into the other because their differentia are logically contradictory
predicates. By contrast, a physical genus is homogeneous, not divided into spe-
cies, but spread out in a way that allows partwise division. Its modifications are
differentiated parts of it, and because the parts contain parts, some modifications
can be transformed into others by internal rearrangement, for example, circles
and squares. We can say that a logical genus is a less than fully specific predi-
cate that is divided into classes defined by simple (rather than partwise-​divisible)
Locke’s Essay and Leibniz’s Nouveaux Essais  •  243

predicates.53 A physical genus is a continuous ordering over a domain of possible


positions which can be realized by parts that are quantitatively, and in the case of
matter, qualitatively, diverse.
For his part, Locke also posits a category of entities that are differential deter-
minations of space, time, number, and sensible qualities, namely, simple modes.54
In his view, ideas of various determinations of space are constructed by modifying
a basic simple idea of a particular spatial quality by means of other simple ideas of
the same sort; for instance, if the basic idea of a particular distance is conjoined
with a repetition of the same idea, the result is the idea of a greater distance. This
apparently reflects the doctrine that a simple sensory idea of a particular spatial
extension contains ideas of extension as parts. None of our ideas of space and
duration “is without all manner of Composition, it is the very nature of both to
consist of Parts.” Again, “the mind is not able to frame an Idea of any Space, with-
out Parts” (Essay 2.15.19). It is not that a person senses the parts a sensory idea of
space contains, but rather that she can’t frame it as lacking parts. This doctrine
needs more discussion than is possible here. But we can say that Locke apparently
wants to hold that all ideas of space have compositional structure like that of
space and, for that reason, the ideas can be arranged in thought (imagination) in
ways parts of space can be configured.
Infinite space and time exist, according to Locke. The Essay strongly suggests
they are attributes of God—​his immensity and eternity—​in which particular
things and events occupy positions relative to each other. Once we have the idea of
a standard unit of length, Locke supposes we can frame the idea of infinite space:

Every one, that has any Idea of any stated lengths of Space, as a Foot, finds,
that he can repeat that Idea; and joining it to the former, make the Idea of
two Foot; and by the addition of a third, three Foot; and so on , without
every coming to an end. … [T]‌he power of enlarging his Idea of Space
by farther Additions, remaining still the same, he hence takes the Idea of
infinite Space. (Essay 2.17.3)

Leibniz accuses him of assuming something his account of the origin of ideas
does not entitle him to, namely, that “since the same principle is always appli-
cable it is impossible that we should ever be brought to a halt; and so the line
can be lengthened to infinity” (NE 158). The emphasis is on the modality that
carries the inference. The point is that Locke assumes all finite lines are similar

53. See diagram, NE 64.
54. Essay 2.12.4–​5, 2.13–​15.
2 4 4   •  The Problem of Univer sals in Early Modern philosophy

in a certain respect without means to prove it is universally or necessarily true.55


Leibniz traces the problem he has in view to a mistaken theory of the origin of
abstract ideas.

10.7. General Maxims as the Foundation of All Other


Knowledge
The Essay repeatedly states that kinds are indispensable for communication and
enlargement of our knowledge. Without them, a common language would be
impossible. There is a hint that this may be due to our mental weakness: “’Tis
true, the Mind in this imperfect state, has need of [general] Ideas. … But yet one
has reason to suspect such Ideas are marks of our Imperfection” (Essay 4.7.9). The
view that kinds are devices of human convenience is briefly allowed to suggest
that they might be dispensable for minds with sufficiently capacious memories
for particulars and, we may suppose, their particular similarities even of higher
order.56 But this is mainly intended to support the contention that general propo-
sitions are not, as is often supposed, necessary for the certainty of knowledge.
Locke undertakes to refute the doctrine that maxims are “the foundations of all
our reasonings” (Essay 4.12.3–​4; 4.2.8).
The target of this attack is broad. It includes Plato’s theory of knowledge and
the Aristotelian doctrine that scientific knowledge is expressed by a demonstra-
tive syllogism (demonstratio propter quid).57 The Essay explicitly mentions that
the laws of identity and noncontradiction are widely supposed to be first prin-
ciples of demonstration. Plainly the assault on the epistemic priority of maxims
over particular truths is intended not just to further discredit scholastic theory of
knowledge, but especially to undermine claims that the principles of the sciences
are innately known.
Locke describes the view he opposes:

The rule established in the Schools, that all Reasons are ex praecognitis, et
preconcessis, seems to lay the foundation of all other Knowledge, in these
Maxims, and to suppose them to be praecognita; whereby, I think, is meant
these two things: First, that these Axioms, are those Truths that are first
known to the Mind; and, secondly, That upon them, the other parts of our
Knowledge depend. (Essay 4.7.8)

55. NE 154; on simple enumerative induction and necessary truths, see 50–​51.


56. Essay 4.17.14; cf. 3.3.2.
57. Posterior Analytics 2.10.1–​2.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  245

Taking the former to refer to temporal priority, he denies it on empirical grounds.


He takes the latter to mean that particular truths are known, or made evident,
on the basis of general maxims. He argues that this is false. Intuitive knowledge,
according to Locke, is immediate (noninferential) perception of a truthmaking
relation among ideas (Essay 4.2.1). Knowledge of this self-​evident sort gives us
the greatest certainty of which we are capable. Principles which are said to be
foundational may be self-​evident, but they are not the foundations of our knowl-
edge because many particular truths are also evident in themselves. A maxim and
its instances may be known independently, each on the basis of nothing but its
intrinsic clarity and evidentness. The logical principle “Whatever is, is” is intui-
tively certain, but so is “A man is a man,” and the like. There is no epistemic prior-
ity among self-​evident truths according to Locke. He grants that maxims have
pragmatic value, but if it seems that they are uniquely certain, it is only because
they are familiar.58
Against this argument, Leibniz maintains that some self-​evident truths can
be demonstrated from others. Two points are brought against Locke.59 One is
that intuitive knowledge which Locke commends, consisting of nothing but a
person’s perception of her own ideas, is unreliable, changeable, and debatable. In
place of it, Leibniz advocates use of formal symbols capable of exhibiting logical
relations among ideas and propositions.60 Locke, by contrast, dismisses reason-
ing in syllogistic form as an artificial distraction from the material relations of
ideas which give evidence of truth.61 Leibniz’s second point has to do with the
epistemic value of subsuming many truths under a few: “to reduce the number of
axioms [is] always something gained” (NE 407).
As Leibniz explains, there is a proper subset of intuitively known propositions
which are the foundations of our knowledge. Truths that are primary are truths
either of reason or of fact. No general propositions are said to be primary truths
of fact.
Primary truths of reason are “identities” (NE 361–​67). Affirmative identi-
ties have subject terms which are repeated in the predicate terms, for example,
‘Whatever is, is’, ‘A is A’, ‘I shall be what I shall be’.62 Negative identities include

58. Essay 4.7.11.
59. See Wilson 1999.
60. The criticism is much more prominent in the little essay Reflexions sur la seconde replique de
Locke (A 6.6:29) than in NE itself; but see NE 360, 478. On the importance of logical form in
reasoning, see NE 478–​84.
61. See Winkler 2003.
62. Cf. Essay 4.8.
2 4 6   •  The Problem of Univer sals in Early Modern philosophy

the principle of noncontradiction which states that “a proposition cannot be


both truth and false at once [and] … it cannot happen that a proposition is
neither true nor false” (NE 362). Also included among negative identities are
“disparities,” which deny that the object of one idea is the object of another;
for example, “warmth is not the same thing as color,” “man and animal are not
the same.” Disparities have terms that are not, at least explicitly, contradictory.
Their inclusion is rather puzzling because their truth is not exhibited by their
logical form.
Nouveaux Essais offers several examples of how identities can be used to dem-
onstrate self-​evident propositions. For one thing, the principle of contradiction
can be used to validate certain forms of syllogism: the validity of Barbara and the
falsity of its conclusion imply the falsity of one of its premises, which shows, for
example, that Bocardo is valid. The law of contradiction is the principle of infer-
ence. The example is especially important, in Leibniz’s view, because it requires
fewer premises than other ways of proving the same result. A second example is
given in response to Locke’s challenge to state the principle needed to prove 2 +
2 = 4. Leibniz produces an argument which purportedly requires nothing but
definitions of ‘2’, ‘3’, and ‘4’, and the principle ‘If equals are substituted for equals,
the equality remains’, which is said to be an axiomatic identity and functions as a
procedural rule.63
Demonstration, according to Leibniz, is a mode of argument that involves
nothing but definitions and one or more identities which serve either as rules of
inference or perhaps as premises.64 Since identities are self-​evident and defini-
tions are strictly speaking neither true nor false, Leibniz supposes it is possible to
demonstrate many self-​evident truths from a few axiomatic identities. He takes
this to show that axiomatic identities are prior in the order of nature, that is, the
order of explanation, which contrasts with the accidental order in which we come
to know things (NE 411, 83). Leibniz also contends for the priority of maxims
which are not primary truths over their instances:

The statement that the body is greater than the trunk differs from Euclid’s
axiom only in that the axiom restricts itself to precisely what needs to be
said; but by exemplifying it—​giving it a body—​we turn something which
can be thought into something which can also be grasped by the sense. …
So we shouldn’t here be contrasting the axiom with the example, as though

63. NE 413–​14. The proof assumes the associativity of addition; see Breger 2006, 57.
64.  Nouveaux Essais offers no example like the well-​known demonstration that a whole is
greater than its part, which reduces the proposition to an identity by substituting the defini-
tions of terms for its terms; see, e.g., “Primary Truths,” A 6.4:1643–​6 49; AG 30–​34.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  247

they were different truths in this respect, but rather regarding the axiom
as embodied in the example and as making the example true. (NE 413)

It is not that the specific case is inferred from the axiom that a whole is greater
than its part. Instead, Leibniz wants to say, although the axiom is not explicitly
known, it is implicit in the example.65 Implicit knowledge influences a person’s
thoughts and judgments without her being aware of it. Even a child who under-
stands “The body is greater than the trunk” grasps, without noticing it, that it is
so because the trunk is a part of the body. The maxim “restricts itself to precisely
what needs to be said.” It is simpler than the example, embodied in it, and por-
trayed as constituting its truth.66
This illustrates the natural order as traditionally understood:  general prin-
ciples are prior to more particular cases in virtue of explaining their truth. But
Leibniz claims more, that the maxim is a constituent of the example; a thought of
the latter is a thought of the former; to judge the latter true is to do the same with
regard to the former without clearly realizing one is doing so. This is meant to
explain why explicit knowledge of the bare-​bones general case enhances a person’s
certainty of an example which is antecedently self-​evident. Now Locke finds the
notion of implicit knowledge unintelligible—​as if one could know something
without knowing it (Essay 1.2.22). His theory of knowledge lacks the psycho-
logical dimension of Leibniz’s. Although Locke does not deny that human beings
are naturally disposed to acquire ideas from experience or that coming to have
an idea involves acquiring certain cognitive dispositions, he takes knowledge of
the mental or physical causes of ideas to be beyond our reach. Leibniz’s implicit
knowledge is a system of more specific active tendencies which, he supposes,
generate human acts of conception and thought by nonconscious means. It is
an operational apparatus posited by the rudimentary psychological theory urged
in Nouveaux Essais. For example, “[W]‌e use the principle of contradiction (for
instance) all the time, without paying distinct attention to it” (NE 76). This is
offered to explain our tendency to make valid inferences, which is a condition
of learning anything at all, and our unthinking tendency to deny a contradiction
when outright confronted with one. There is no counterpart theory in the Essay.
For his part, Locke has an additional argument against the general model of
an explanatory science as founded on first principles better known than more spe-
cific general truths: “There is … a great deal of Talk, propagated from Scholastik
Men, of Sciences and the maxims on which they are built: But it has been my ill

65. NE 448–​49.
66. Also NE 448.
2 4 8   •  The Problem of Univer sals in Early Modern philosophy

luck, never to have meet with any such Sciences” (Essay 4.7.11). His argument has
two prongs. One pertains to general truths about substances which are supposed
to exist; the other concerns modes, which are supposed to be possible existents.
Propositions of the first sort assert that all actual things which have such and
such qualities (collected in the idea of a kind of substance) do (do not) have cer-
tain other qualities. Locke maintains that if we knew the inner constitutions of
particular substances in sufficient detail, we could explain why certain clusters of
qualities which pertain to the interaction of bodies are always combined in the
same thing and even predict them before observing them (e.g., Essay 4.3.25). For
instance, if we know the particulate constitution of opium and that of the human
brain, we might explain why opium causes sleep. But because we have no prospect
of knowing inner constitutions of nature’s making according to Locke, a “science
of body”—​as envisaged by scholastics—​is beyond the cognitive powers of human
beings.67
The situation is different with regard to modes. Because their nominal and real
essences are the same, we are fully cognizant of the sources, or reasons, of general
truths about, say, a triangle. The second prong of Locke’s attack pertains to the
order of knowledge of general and particular truths whose subjects are modes. He
contends that demonstrative knowledge of general truths must be inferred from
demonstrative knowledge of particular truths. So, putting the prongs together,
the established model of science falls either because the causes of general truths
are inextricably hidden from us or because, if they are known, the knowledge of
general truths is derived from knowledge of particular cases.
The latter contention is a consequence of Locke’s theory of what a general idea
is, namely, a particular idea partially considered. We cannot think of a kind with-
out bringing to mind the idea of a particular thing. Although general truths can
be deduced from other such truths if they are known, the question is how such a
truth is known to begin with. It is by proving that a particular thing has a certain
property and selecting the elements of the proof as bases of generalization:

the immediate Object of all our Reasoning and Knowledge, is nothing


but Particulars. [One reasons] only about the Ideas existing in his own
Mind, which are truly, every one of them, particular Existences: and our
Knowledge and Reasoning about other Things, is only as they correspond
with those our particular Ideas. So that the Perception of the Agreement,
or Disagreement of our particular Ideas, is the whole and utmost of all our
Knowledge. Universality is but accidental to it. (Essay 4.17.8)

67. Essay 4.3.14, 16, 26.


Locke’s Essay and Leibniz’s Nouveaux Essais  •  249

Demonstration of a general proposition is a two-​step process.68 First, a person


comes to know a truth about a particular thing, for instance, a particular triangle
drawn in a diagram. She perceives, say, that the longest side subtends the great-
est angle, and knows it either intuitively or demonstratively. Second, she focuses
attention on the features of the figure which show that this is true, such as its
being enclosed by three straight sides which are not equal to each other, and the
location of a maximal side and angle; by doing this, she forms the general propo-
sition which was to be demonstrated.
To be clear, the generalization is licensed only if the reasoner ascertains that
the opposition of side and angle is due to—​biconditionally inseparable from—​
the figure’s having three sides and not being equilateral. Leibniz disputes Locke’s
ability to arrive at certainty of such things by means of general ideas as he
construes them.

10.8. Conclusion
Although both Leibniz and Locke are broadly nominalist in their theories of
universals, their differences are deep and wide-​ranging. They involve basic issues
of metaphysics, theory of knowledge, linguistic signification, the nature of ideas
(concepts), and more. But for purposes of summary, their respective positions
may be characterized as follows.
Locke maintains that, at least as far as we know, particular things are divided
into kinds by nothing other than abstract ideas framed by human beings for the
convenience of thinking and speaking about well-​defined collections of things,
rather than solely about individuals or malleable groupings. General truths of all
sorts depend on the existence of human beings or minds that have the same ideas.
Everything that exists is entirely particular, as he has it, and we have no innate
ideas. Kinds and general ideas must, accordingly, be made from particular ideas
at human discretion. For Locke, a general idea is the idea of a particular thing
considered as having certain selected features and set up by the mind to signify
all things with similar features. Accordingly, it is impossible to think or reason
about a kind of thing without considering the features of one or more particulars.
It is a principle of Locke’s theory that we cannot have an idea of a kind unless
we know what its boundaries are. He vehemently objects to theories on which
natural kinds are delineated by Forms, essences, sets of essential properties, or
constitutions that are unknown to us, as well as theories of innate dispositional,
or implicit, knowledge of essences of kinds. Because observation reveals no

68. Ayers 1991,1:248–​50.
2 5 0   •  The Problem of Univer sals in Early Modern philosophy

fully determinate clusters of qualities that would indicate a natural division of


species, this principle leads Locke to conclude that determinate boundaries can
be fixed only by human practice. For this reason, kinds vary in accord with the
intentions with which we form general ideas. General ideas of substances are sup-
posed to represent actual things by means of their observable qualities, so they
have ground in reality only if there are actual things that conform to them. Ideas
of sensible qualities are meant to represent bodily causes of ideas of sensation,
but are delineated by similarities among sensations. Ideas of modes are intended
only to define kinds of things that might possibly exist, so their ground in reality
is their intrinsic possibility. Each of the provisions mentioned above sets Locke’s
doctrine at odds with traditional nominalist theories of kinds. On each of these
points, Leibniz sides with the tradition.
Leibniz reasons that although kinds are nothing but similarities, neither simi-
larity relations nor their possibility depends on the existence of human beings. He
posits a ground of possible essences in the eternal understanding of God. Beings
in all categories—​substances, qualities, continuous quantities, events, actions,
and so on—​have essences with the same ground in reality. Although Leibniz
does not entirely disagree that we cannot classify by a kind if we are wholly igno-
rant of its essence, he recognizes ways of knowing essences which would not be
acceptable to Locke. It is not just that we have innate ideas which dispose us to
construct definitions of essences, but also that Leibniz offers a method by which
we can inquire about the unknown essence of a kind. This is the method of nom-
inal and real causal definitions. A carefully framed nominal definition is capable
of providing knowledge of an essence from some of its properties or effects. A
well-​constructed real causal definition stands a chance of providing knowledge of
the essence as the cause of the effects mentioned in the nominal definition, and
others as well. If all goes well, we can attain a set of definitions adequate to state
and prove general propositions which are reasonably certain and can be of great
benefit, as Leibniz has it. He suggests that this is the general method by which we
can inquire about essences in mathematics and natural sciences.
Leibniz argues further that knowledge of general truths is required for knowl-
edge of their particular instances. This is because general propositions are true in
virtue of the relation between the concepts they contain, so their instances are
true in virtue of the same relation between the same core concepts. Because a
general truth is embedded in the particular truths it subsumes, knowledge of the
latter is, in effect, knowledge of the former.
In view of the considerable opposition between their theories and differences
in their approaches, it may seem that the two philosophers talk past each other,
not to each other. The subtitle of this chapter is meant to convey that Leibniz
does not enter into a debate with his discussant with the aim of refuting him.
Locke’s Essay and Leibniz’s Nouveaux Essais  •  251

There are few issues on which either participant has a knock-​down objection to
the other. The systematic connections internal to their respective metaphysical
and epistemic doctrines are too tight and the differences between their systems
too great to allow for that. Instead, Nouveaux Essais stages a contest over which
set of interconnected doctrines does the better job of explaining the data, avoid-
ing difficulties, and effectively advancing knowledge and probable opinion.69
Leibniz prevails in the dialogue, but the reader is meant to be the judge.

Abbreviations
A = Leibniz 1923–​; cited by series (Reihe), volume (Band), and page.
AG = Leibniz 1989; cited by page.
G = Leibniz 1875–​90; cited by volume and page.
NE = Leibniz 1996; cited by page.

69. See Bolton 2007.
11 LOCKE ON GENERAL IDEAS

E. J. Lowe

With regard to the so-​called problem of universals, Locke is commonly


described as being a “nominalist,” and this is correct as far as it goes. But
it is also unhelpful, because the term ‘nominalism’ is used in so many
different ways, both in the history of philosophy and in contemporary
philosophical debate. It would be slightly more helpful to describe
Locke as subscribing to a version of “resemblance nominalism,” but
even that is less than fully informative. I prefer to describe him simply as
being a particularist, by which I mean an advocate of the view that only
particulars exist and hence no “general” or “universal” entities. This view
is perfectly consistent with saying that there are general terms and gen-
eral concepts or ideas, but with the caveat that such terms and ideas are
themselves just particular entities. Of course, saying that all things are
particulars is only as clear as the meaning or definition of the word ‘par-
ticular’ permits, and it is unfortunately the case that there is widespread
disagreement about this, both today and in the history of philosophy.
For present purposes, however—​that is, for the discussion of Locke’s
views—​I think we may safely take it to be a hallmark of particularity
that all particulars are confined in their existence to a unique place at
any given time, whereas universals, were there to be such things, would
be items capable of being “present” at many different places simultane-
ously and in that sense “colocated” with many different particulars at
one and the same time. In any case, Locke’s allegiance to particularism
is unmistakable and very strong, being stoutly affirmed in the follow-
ing famous statement of the Essay Concerning Human Understanding:

General and Universal, belong not to the real existence


of Things; but are the Inventions and Creatures of the
Understanding. (Essay 3.3.11)1

1. All future references to the Essay will be from Locke 1975, cited by book, chap-
ter, and section.
Locke on General Ideas  •  253

It will be noted that this is not only an affirmation of particularism, but also serves
to indicate Locke’s view concerning the origin of general terms and ideas, namely,
that they are in some way products of human thought processes in our attempt to
understand and communicate with one another about the nature of the world of
particulars that we inhabit.
Since there is so much to say on the topic of Locke on general ideas, I shall
confine myself in what follows to what is, I think, the most interesting part of his
doctrine on this issue, namely, that concerning our general ideas of the various
sorts of naturally occurring material substances, or what would nowadays more
usually be called “natural kinds,” such as the kinds water and gold. It is this part
of Locke’s doctrine that has continued to stimulate widespread debate, especially
since the appearance of the seminal work on the semantics of natural kind terms
by Saul Kripke and Hilary Putnam in the 1970s. The work of Kripke and Putnam,
which has been highly influential—​to the extent, indeed, of providing what is
now the default view of the semantics of natural kinds terms amongst analytical
philosophers of language and science—​was explicitly critical of Locke, with the
consequence that Locke’s view of these matters is now almost universally regarded
as manifestly mistaken. My own opinion, which I shall attempt to justify in this
chapter, is that Locke’s present-​day detractors have much less right on their side
than they suppose, even if there are indeed deficiencies in Locke’s own account—​
deficiencies, however, which for the most part he cannot be blamed for, given
the state of scientific knowledge in his time. In order to confine my discussion to
manageable proportions, I shall concentrate my attention on one particular natu-
ral kind term—​‘water’—​that figures prominently in current debate, owing to the
notoriety conferred upon it by Kripke and Putnam, but about which Locke him-
self had particularly interesting things to say.
With regard to the science of his day, Locke had, of course, great sympathy for
the “corpuscularian” philosophy of his eminent scientific friends Robert Boyle
and Isaac Newton, thus favoring atomism over the views of the Cartesians where
the nature of matter is concerned. At the same time, as an empiricist who regarded
the scope of human knowledge as being severely restricted by our limited powers
of perception, Locke held only that atomism was a reasonable hypothesis, rather
than a doctrine whose truth we could ever hope to establish as certain. But in
accepting atomism at least as a working hypothesis, Locke was implicitly reject-
ing the Aristotelian scheme of the four elements and thereby rejecting the ancient
idea that water is something fundamental in the constitution of nature. He was
evidently open to the idea that there might be a variety of different kinds of atoms,
perhaps possessing different combinations of primary qualities, such as differ-
ent sizes, shapes, and densities. And he was receptive also to the hypothesis that
such differences could explain the wide variety in the observable characteristics
2 5 4   •  The Problem of Univer sals in Early Modern philosophy

of macroscopic objects, such as animals, plants, metals, and mineral formations.


Indeed, the following passage testifies to his remarkable degree of confidence on
this point:

Had we such Ideas of Substances, as to know what real Constitutions pro-


duce those sensible Qualities we find in them, and how those Qualities
flowed from thence, we could, by the specifick Ideas of their real Essences
in our own Minds, more certainly find out their Properties, and discover
what Qualities they had, or had not, than we can now by our Senses: and
to know the Properties of Gold, it would be no more necessary, that Gold
should exist, and that we should make Experiments upon it, than it is nec-
essary for the knowing the Properties of a Triangle, that a Triangle should
exist in any Matter, the Idea in our Minds would serve for the one, as well
as the other. (Essay 4.6.11)

Locke, then, thought that if we could discover the internal atomic constitution,
or “real Essence,” of any naturally occurring material substance, such as water, we
would be able to deduce the observable properties of that substance, such as water’s
properties of being transparent and liquid at room temperature and its power to
dissolve common salt (at least, if we knew also the “real essence” of common salt).
He was just skeptical about the possibility of our ever finding out the real essence
of any material substance, even though he knew very well about the advances in
microscopy being made in his time by scientists such as Robert Hooke.
It is worth looking more closely here at Locke’s use of the expression ‘real
essence’, to denote the “internal constitution” of a material substance. The word
‘essence’ he gets, of course, from scholastic philosophy and thus ultimately from
Aristotle, even though Locke is vehemently opposed to the scholastic theory of
essence, as he understands it. Locke tells us that the word ‘essence’, in its “proper
original signification”—​by which, I assume, he means its Aristotelian sense—​
denotes “the very being of any thing, whereby it is, what it is” (Essay 3.3.15). Now,
the English word ‘essence’ derives from the Latin word essentia, which was used
by scholastic philosophers to translate a phrase of Aristotle’s whose literal English
equivalent is something like “the what it is to be” or “the what it would be to
be.”2 Thus, what Aristotle and the scholastics understood by an account of the
essence of some kind of substance is an account of what that kind of substance
fundamentally is. And Locke seems to mean much the same, as is indicated by his
characterization of essence as “the very being of any thing, whereby it is, what it

2. See Aristotle, Metaphysics Z, 4 and, for fuller discussion, Lowe 2008.


Locke on General Ideas  •  255

is.” However, where Locke parts company with the Aristotelian tradition is pre-
cisely in identifying the essence of a material substance with its unknown, and (in
his view) probably unknowable, internal atomic constitution. By contrast, in the
Aristotelian tradition, a substance’s essence is revealed by its so-​called real defini-
tion, which is assumed to be knowable by us, at least in many cases. For instance,
man as a kind of substance is famously defined, in the Aristotelian tradition, as
a rational animal. This sort of definition is said to be per genus et differentiam:
that is, a species of substance is defined in terms of the higher genus to which it
belongs and the specific difference which renders it distinct from any other species
of that same genus. But Locke, in common with other empiricist philosophers of
his time, thought that such “definitions” were trifling and could make no substan-
tive contribution to our knowledge of nature or our explanations of natural phe-
nomena. He thought that natural phenomena could ultimately be explained only
mechanically, in terms of the primary qualities of material bodies and hence in
terms of a special class of causal laws governing such bodies. That, fundamentally,
is why he thinks that the only thing that can qualify as the “real essence” of any
kind of substance is its internal atomic constitution, even if we can never hope to
know precisely what this constitution is in any specific case.
Now, however, Locke’s philosophy of nature takes a further skeptical twist.
First of all, since he thinks that we can never hope to discover the real essences of
substances, as he understands them, he thinks that it is impossible for us to clas-
sify kinds of substances on the basis of their supposed real essences. As he puts it:

[T]‌he supposition of Essences, that cannot be known; and the making of


them nevertheless to be that, which distinguishes the Species of Things, is
so wholly useless … [as] to make us lay it by. (Essay 3.3.17)

According to Locke, true to his own empiricist precepts, we can be supposed to


classify different species of substances only on the basis of their observable quali-
ties. And yet how we do that, he thinks, will largely be a matter of human custom
and convention. Furthermore, he believes that even if we did have access to “real
essences,” we would still not have a basis on which to construct a classification
of kinds of substance that was not to a large degree arbitrary or conventional.
This is because he thinks that, at the atomic level, there is probably a more or
less continuous range of variation between one possible “internal constitution”
and another. Internal constitutions, recall, are just configurations of atoms, and
Locke supposes that such configurations can be altered continuously pretty much
without limit. Hence, between any two different possible configurations, there
will be very many others, differing from one another only in rather small respects,
with these differences giving rise to corresponding small differences at the level
2 5 6   •  The Problem of Univer sals in Early Modern philosophy

of the observable qualities of substances. Nature at the atomic level, he thinks,


probably exhibits continuous variety and hence doesn’t divide things up into a
limited number of markedly distinct substantial kinds. Thus, as Locke sees it, our
propensity to regard the natural world as being clearly divided up in this way is
more a reflection of our own ways of perceiving and thinking about things than
it is of the world itself, independently of us. His skepticism on this point was
sustained through his abiding curiosity with travelers’ tales about “monsters” and
“changelings,” since the evidence for such creatures reinforced his conviction that
there are really no sharp boundaries between the species of things. Interestingly
and importantly, this conviction testifies to the presence of a distinctly premod-
ern notion in Locke’s philosophy of nature: that of the “Great Chain of Being,”
according to which there are innumerable gradations, with no “gaps,” between
the most lowly and the most elevated of God’s creatures, with man’s place lying
somewhere in the middle.3
A consequence of all this, where Locke is concerned, is that the word ‘water’,
far from denoting some basic kind of natural substance, according to him merely
serves to signify a collection of “ideas” that we have of certain observable quali-
ties, which we find to be regularly associated with one another. This collection of
ideas Locke calls our “abstract general idea” of water:

Things are ranked under Names into sorts or Species, only as they agree to
certain abstract Ideas, to which we have annexed those Names. (Essay 3.3.15)

But Locke is emphatic that different people may well attach different abstract gen-
eral ideas to the word ‘water’, simply because in their past experience they happen
to have found rather different collections of observable qualities to go together
with one another. Thus, where one person may include the ideas of liquidity or
transparency in his or her abstract general idea of water, another may not. We
may expect different users of the word ‘water’ to share many of the ideas that
they associate with that term, but not necessarily all of them. This is at least part
of the point of the story that Locke tells about a certain visitor to England from
Jamaica, in the following passage (a passage which, significantly, immediately fol-
lows Locke’s expression of his allegiance to the idea of the Great Chain of Being):

If I should ask any one, whether Ice and Water were two distinct Species
of Things, I doubt not but I should be answered in the affirmative: And

3.  See Lovejoy 1936. With regard to Locke, Lovejoy quotes a memorable passage from the
Essay (3.4.12) clearly demonstrating his allegiance to the idea: see 184.
Locke on General Ideas  •  257

it cannot be denied, but he that says they are two distinct Species, is in the
right. But if an English-​man, bred in Jamaica, who, perhaps, had never
seen nor heard of Ice, coming into England in the Winter, find, the Water
he put in his Bason at night, in a great part frozen in the morning; and
not knowing any peculiar name it had, should call it harden’d Water; I ask,
Whether this would be a new Species to him, different from Water? And,
I think, it would be answered here, It would not to him be a new Species,
no more than congealed Gelly, when it is cold, is a distinct Species, from
the same Gelly fluid and warm; or than liquid Gold, in the Fornace, is a
distinct Species from hard Gold in the hands of a Workman. … [T]‌he
ranking of Things into Species … is done by us, according to the Ideas that
we have of them. …[I]f we suppose it to be done by their real internal
Constitutions, and that Things existing are distinguished by Nature into
Species, by real Essences, according as we distinguish them into Species by
Names, we shall be liable to great Mistakes. (Essay 3.6.13)

Locke is saying here that the words ‘water’ and ‘ice’, as they are most commonly
used, signify different collections of ideas, with the idea of “water” including that
of liquidity whereas the idea of “ice” includes that of solidity. And for this reason
he takes it that it is correct to say that ‘water’ and ‘ice’, as they are most com-
monly used, denote different species of substance, since the only way in which
substances can be distinguished into species, according to him, is on the basis of
the abstract general ideas that we associate with the names of those substances.
He allows, however, that it would be perfectly natural and reasonable for his
imagined Jamaican visitor to regard what we call “ice” to be merely “hardened
water,” and thus just a certain form of water, rather than a different kind of sub-
stance altogether. But this is only because the visitor associates a different abstract
general idea with the word ‘water’ from the one that most of the rest of us do, and
has never before experienced the combination of observable qualities, including
those of coldness and hardness, whose idea we associate with the word ‘ice’.
To a modern ear, Locke’s verdict on this story may seem strange and counter-
intuitive. Surely, most modern readers of his text will want to affirm that ice just is
frozen water, not a different kind of stuff, any more than liquid gold is a different
kind of stuff from solid gold. In other words, their sympathies will naturally lie
with the opinion of the Jamaican visitor on this issue. Why is this so? I suspect that
at least in part it may be because some knowledge of the chemistry of water, how-
ever superficially understood, has infiltrated our common-​sense ways of thinking
and talking about it. Few people today are completely unaware of the facts that
water is composed of tiny molecules, which in turn are composed of hydrogen
and oxygen atoms, and that its chemical formula is H2O. And most have at least a
2 5 8   •  The Problem of Univer sals in Early Modern philosophy

superficial grasp of the fact that when water is cooled to zero degrees Celsius these
molecules form crystals, whose structure explains the solidity and density of the
substance that they then compose. But why—​it may be asked—​do they assume
that these changes don’t bring into existence a new kind of substance, ice, rather
than merely transforming water from a liquid into a solid? In fact, I think that
this is a rather difficult question to answer, other than in terms of some degree
of infiltration of modern chemical theory into everyday thinking.4 According to
such theory, a kind of chemical substance is distinguished solely by its chemical
formula, which in this case is H2O. Chemists don’t generally regard different con-
figurations of the atoms or molecules composing a substance to have any bearing
on the kind of substance that it is. This is why, from the chemist’s point of view,
diamond, graphite, and charcoal are just the same kind of substance—​carbon—​
despite the very great differences between their observable qualities. Note that
the ordinary speaker of English with little knowledge of chemistry may be much
more resistant to saying that diamond, graphite, and charcoal really are just the
same kind of stuff than they are to saying this about water and ice. But this, I sus-
pect, is largely because such a speaker is much less likely to know anything at all
about the chemistry of carbon than about that of water.
However, we can’t just leave the matter there and assume uncritically that
chemical science has the last word on such matters of classification. Why, really,
would it be wrong to say that diamond, graphite, and charcoal are different kinds
of stuff ? Why should the fact that they are all wholly composed of carbon atoms
be the sole determining factor with regard to their being, or not being, the same
kind of material substance? Why shouldn’t the configuration of those atoms have a
bearing on the matter? But once we allow that it may legitimately do so, we under-
mine the assumption that chemical science has somehow shown that water and
ice are “really” the same kind of stuff. Questions like these, and the difficulty that
we find in answering them in a principled way, may actually lend some support to
Locke’s skepticism about the thesis that it is nature, rather than our “ideas,” that
divides natural substances into distinct kinds. At the very least, we may be encour-
aged to think that there is no single “right” way to classify natural substances into
distinct natural kinds, but that different and yet equally legitimate systems of clas-
sification may serve different purposes, both in science and in everyday life.5

4. I confess that I have changed my views on this issue somewhat since previously discussing it
in Lowe 1998, 175–​76. There I did not, I now think, give enough weight to the testimony of
Locke’s own words that native inhabitants of England in his day would unhesitatingly judge
water and ice to be different kinds of stuff.
5. Such a view is defended by John Dupré (1993, 6–​7), where he calls his version of the view
“promiscuous realism.”
Locke on General Ideas  •  259

I mentioned earlier the severe criticism that Locke’s account of natural kind
terms received at the hands of Saul Kripke and Hilary Putnam in the 1970s. At
this point I want to look into this sort of criticism more closely, since by examin-
ing it we can, I believe, more easily see what is defensible and what indefensible
in Locke’s views about these matters. Now, it is unquestionable that some impor-
tant developments in metaphysics, logic and the philosophy of language that
took place in the 1970s transformed philosophical thinking about the semantics
of natural kind terms—​that is, terms like ‘water’ and ‘gold’. Prior to that time,
a view about the meaning of such terms that was close to Locke’s prevailed in
philosophical circles. On Locke’s view, such terms are definable in terms of some
combination of observable characteristics—​so, for example, ‘gold’ in terms of yel-
lowness, shininess, ductility, malleability, high density and melting point, solubil-
ity in aqua regia, and so forth. Such a definition captures what Locke called the
nominal essence of gold, as opposed to any “real” essence that it might be supposed
to have. That is to say, it captures the “abstract general idea” that typical users of
the word ‘gold’ associate with it and supposedly employ to classify pieces of stuff
as being, or not being, gold. Kripke and Putnam, independently but at about the
same time, roundly rejected this empiricist doctrine.6 They urged that what deter-
mines the reference of a natural kind term such as ‘water’ or ‘gold’ is precisely a
certain Lockean real essence. Thus, in the case of water, this would be its having
an internal constitution of H2O molecules and, in the case of gold, its having an
internal constitution of atoms possessing the atomic number 79. Locke, of course,
doubted that we would ever be able to discover such real essences, but Kripke
and Putnam assumed that modern chemical science has now actually discovered,
in many cases, precisely what Locke supposed to be indiscoverable. Hence, they
thought, his skepticism about the possibility of our being able to classify sub-
stances according to their real essences could now effectively be overcome.
Indeed, Kripke and Putnam went even further than this. They maintained
that the words ‘water’ and ‘gold’ always have picked out the substances whose
internal constitutions are, respectively, H2O molecules and atoms possessing the
atomic number 79, and thus that they referred to those substances even in Locke’s
day—​indeed, even in the days of Aristotle (allowing for the fact, of course, that
Aristotle used a Greek word that is customarily translated as ‘water’ in English).
Thus, according to Kripke and Putnam, Locke himself was simply referring to the
substance that is composed of H2O molecules when he spoke of water, even though
he had no idea whatever about the existence of such molecules. In so saying,
they directly contradicted what seemed obvious to Locke, namely—​as we noted

6. The seminal texts are Kripke 1980 and Putnam 1975.


2 6 0   •  The Problem of Univer sals in Early Modern philosophy

earlier—​that “the supposition of Essences, that cannot be known; and the making
of them nevertheless to be that, which distinguishes the Species of Things, is so
wholly useless … [as] to make us lay it by” (Essay 3.3.17). For in Locke’s day, at
least, these essences could not be known, and yet Kripke and Putnam nonetheless
insisted that even in his day they did serve to distinguish one species of substance
from another, such as water from gold.
How was this remarkable revolution in philosophical thinking achieved? For
it should be acknowledged that Kripke and Putnam’s doctrine very quickly took
root and soon became accepted orthodoxy regarding the semantics of natural
kind terms. The fact is that it was largely achieved by appeal to what philoso-
phers today call thought-​experiments. A thought-​experiment is an exercise of the
imagination. Putnam, in particular, sought to persuade his readers by recourse
to his imaginary example of Twin Earth. We are to imagine that Twin Earth is
a distant planet very like our own Earth, inhabited by people very much like
us. Indeed, they speak a language which is phonetically indistinguishable from
English—​call it Twinglish. In Twinglish, there exists the word ‘water’, which
refers to the substance filling the oceans of Twin Earth—​the substance that,
on Twin Earth, falls as rain, flows in the rivers, and is drunk by the inhabit-
ants. And this substance, we are to imagine, looks, tastes, smells, and indeed in
every way appears to the senses of Twin Earth’s inhabitants exactly the same as
water on Earth does to us. However, we are now told, the Twin Earth substance
that is there called “water” is not composed of H2O molecules, but instead has
some quite different chemical formula, XYZ. And now we are invited to answer
the following question: is it correct to say that Twin Earth’s water—​twater, as
we might call it—​is the same kind of stuff as our water? Putnam’s answer, with
which we are all expected to agree unhesitatingly, is that twater and water are
different kinds of stuff, despite their exact resemblance at a macroscopic level.
And this is supposed to show that what determines what kind of stuff a word like
‘water’ refers to is the chemical composition of samples of that stuff, even if that
composition is unknown to the users of the word. For, of course, we can easily
build into the Twin Earth story the supposition that the inhabitants of Twin
Earth have no advanced chemical knowledge like ours, but are scientifically
pretty much in the condition of Locke and Boyle in the seventeenth century.
Although there was, initially, a certain amount of resistance to Putnam’s verdict
concerning his Twin Earth thought-​experiment, that rapidly faded away and it
is now almost universally accepted by philosophers who think much about such
matters that he was right.7 And yet, in fact, I think we ought to be very suspicious

7. An honorable exception is Joseph LaPorte: see LaPorte 2004, although I by no means wholly
agree with LaPorte’s position either.
Locke on General Ideas  •  261

about it indeed. To his credit, even Putnam later came to be much more circum-
spect about what could be concluded from the thought-​experiment, as we shall
shortly see. Unfortunately, his change of heart wasn’t taken much notice of by
the wider philosophical community.
Before I come to express my doubts about these matters, I need to say some-
thing about an even more extreme conclusion that was drawn by Kripke and
Putnam in their original arguments concerning the semantics of natural kind
terms. This conclusion relates to their understanding of the word ‘essence’
and its metaphysical implications. In the Aristotelian tradition, the notion of
essence is closely connected with that of necessity. For example, since man is
defined, in that tradition, as being a rational animal, it is concluded that ratio-
nality and animality are necessary features of any man, so that nothing could
be a man without being both rational and an animal, any more than anything
could be a triangle without having three sides. In fact, this association of the
notion of essence with that of necessity was preserved in Locke’s thought, as
is illustrated by the passage from the Essay quoted earlier, in which he implies
that, if we knew the real essence of gold, we could deduce all of the observ-
able properties of gold with as much certainty as we can deduce all of the
properties of a triangle from its geometrical definition, without recourse to
any experimentation or observation. However, it is far from clear that Locke
was warranted in supposing this. The properties of a (Euclidean) triangle, such
as the fact that the sum of its internal angles equals 180 degrees, are indeed
deducible from its definition, in conjunction with the axioms of Euclidean
geometry. And we regard those axioms as expressing absolutely necessary truths
about the geometry of Euclidean space. However, it is much more contentious
to suppose that the observable properties of gold, such as its yellowness and
ductility, are deducible from facts about its atomic constitution in conjunc-
tion with physical and chemical laws. Moreover, it is even more contentious
to suppose that such physical and chemical laws have the status of absolutely
necessary truths, on a par with the axioms of Euclidean geometry. Even if a
kind of necessity attaches to physical and chemical laws, it does not seem to
be the very strong kind of necessity—​called by philosophers “metaphysical”
necessity—​that attaches to mathematical and geometrical truths. In the lan-
guage of contemporary philosophy, a mathematical truth, such as that 2 plus
2 equals 4, is one that obtains in every possible world. But the physical and
chemical laws of our world do not appear to be necessary in this sense: we can
readily suppose that, in “other” possible worlds, the physical and chemical laws
are significantly different from those in ours. Indeed, some cosmologists now
speculate that our universe is just one member of a vast “multiverse,” other
members of which exhibit very different laws, in which different forces are
2 6 2   •  The Problem of Univer sals in Early Modern philosophy

operative or various “constants” of nature, such as the constant of gravitation,


have different values.8
Now, Kripke and Putnam, like Aristotle and Locke, associate the notion of
essence with that of necessity, and they do so despite the fact that their theory of
essence is Lockean rather than Aristotelian. That is to say, they, like Locke, regard
the “real essences” of natural substances as being their internal atomic constitu-
tions, and yet they continue to hold that the essence of a substance, in this sense,
is a strongly necessary feature of it. Hence, they hold that water, for example, is
not just actually composed of H2O molecules, but necessarily so composed—​that
is to say, that in every possible world in which water exists, not just in the actual
world, it is composed of H2O molecules. But what should we say, then, about the
point that the laws of physics and chemistry might be different in “other” pos-
sible worlds? What if it should be the case that, in some other possible world, a
substance composed of H2O molecules has, owing to a difference in those laws,
observable properties quite unlike those of water in this, the actual world? Would
that substance still be water? According to Kripke and Putnam—​or, at least,
according to Putnam’s original view about this issue—​the answer is emphatically
“Yes.” Indeed, on this view, water is H2O in every possible world, and in every
possible world H2O is water, no matter how different its observable properties
might be in different possible worlds. But how, one might wonder, can such an
extreme verdict possibly be warranted? My own suspicion is that it simply cannot.
And I suspect that the very word ‘essence’ is playing an important but surrepti-
tious role here. As we have noted, in the Aristotelian tradition, there is clear and
uncontentious connection between the notions of essence and necessity, arising
from the fact that on the Aristotelian view the essences of substances are revealed
by their real definitions. But it was precisely this theory of essence that was rejected
by Locke when he proposed instead to identify the “real essences” of substances
with their supposed, but unknown, internal constitutions. To assume, as Locke
himself appears to, that the connection between essence and necessity survives
this radical change of view regarding what essences are seems to be quite unwar-
ranted. And yet this assumption is crucial to the Kripke–​Putnam contention that
water is necessarily H2O, or is H2O “in every possible world.”
I mentioned earlier that, Putnam, to his credit, later came to doubt some
of the more extreme aspects of his original views concerning natural kinds,

8. In view of this, it is difficult to see how there could be any warrant in current physical science
for the opinion of some contemporary metaphysicians that there is no real distinction between
“physical” or “causal” necessity and “metaphysical” necessity:  see, for example, Shoemaker
1998. I subject this opinion to extended criticism in Lowe 2006, 141–​73.
Locke on General Ideas  •  263

such as water and gold. The following passage is particularly revealing in this
connection:

I do not think that a criterion of substance-​identity that handles Twin


Earth cases will extend handily to “possible worlds.” In particular, what if
a hypothetical “world” obeys different laws? Perhaps one could tell a story
about a world in which H2O exists (H still consists of one electron and
one proton, for example), but the laws are slightly different in such a way
that what is a small difference in the equations produces a very large dif-
ference in the behavior of H2O. Is it clear that we would call a (hypotheti-
cal) substance with quite different behavior water in these circumstances?
I now think that the question, “What is the necessary and sufficient condi-
tion for being water in all possible worlds?” makes no sense at all. And this
means that I now reject “metaphysical necessity.” (Putnam 1990, 69–​70).

As will be evident from earlier remarks of mine, I think that Putnam was exactly
right to have such doubts about water necessarily being H2O. However, it also
appears from this passage that Putnam did not, in 1990, want to abandon his
Twin Earth thought-​experiment altogether. He says only that it will not “extend
handily to [other] ‘possible worlds.’ ” That suggests that he retains his verdict
regarding the thought-​experiment’s implications for this, the actual world. He
still believes, that is to say, that the thought-​experiment shows that what deter-
mines whether a substance is water anywhere in this world is its molecular com-
position, so that “our” water and Twin Earth’s twater are different kinds of stuff,
simply in virtue of having different molecular compositions, notwithstanding
their supposed exact similarity at the level of observable properties. But now
I  think we need to challenge even this contention. One thing that we should
be suspicious about is this: the very idea that there could exist, in this world, a
chemical substance with all of the observable properties of water, despite hav-
ing a radically different molecular composition. We are never told what XYZ—​
the chemical formula of twater—​is supposed to be, only that it is quite different
from H2O. But if we try to think more specifically what it could be, I suggest that
we shall have to conclude, on the basis of current chemical science, that there is
simply no naturally existing chemical compound, nor any that could be artifi-
cially synthesized from the elements existing in our universe, which could, con-
sistently with the actual laws of physics and chemistry, appear and behave at the
macroscopic level exactly like water, except H2O itself. In other words, the Twin
Earth thought-​experiment is founded upon a supposition that turns out to be
incompatible with the physical and chemical laws of our universe. If Twin Earth
really did have oceans filled with XYZ, where this is a substance not composed of
2 6 4   •  The Problem of Univer sals in Early Modern philosophy

H2O molecules, then that substance—​twater—​would not be remotely like water


on Earth. Hence, the thought-​experiment is incoherent and cannot be used to
establish any useful conclusion whatever. Perhaps we can coherently conceive of
another possible world, in which the laws of physics and chemistry are different
from those of our universe, where there is a water-​like substance composed of
something other than H2O molecules. But that is entirely beside the point for
present purposes, when we are being invited to imagine that Twin Earth is just a
distant planet in our universe.
We have now seen reason to think that anywhere in our universe in which
a water-​like substance exists—​water-​like, that is, at the macroscopic, observ-
able level—​it will be composed of H2O molecules. But in that case, we are by
no means compelled to say that being composed of H2O molecules is what funda-
mentally distinguishes water from any other kind of substance, since being water-​
like—​that is, having the characteristic observable properties typically associated
with the word ‘water’—​is invariably connected with being composed of H2O mol-
ecules in our universe, whence being water-​like can certainly be no less eligible
than being composed of H2O molecules as a distinguishing mark of water. Indeed,
we can now, I  think, turn the tables on the Kripke–​Putnam theory of natural
kind terms and argue as follows. Since, as Putnam belatedly came to recognize,
being composed of H2O molecules is not a good candidate for being an essential and
thus necessary property of water, whereas being water-​like plausibly is, it is prefer-
able to regard the latter as being the primary distinguishing mark of water. As
Putnam acknowledges, we would rightly be loath to describe as water a substance
composed of H2O molecules in another possible world in which, owing to the
different physical and chemical laws obtaining there, this substance had, say, the
appearance and behavior of sticky black tar at room temperature. But that implies
that our criterion for identifying a substance as being water, whether in this world
or another merely possible one, appeals precisely to its possession of a water-​like
appearance and behavior. So it seems that Locke was right, after all, to suppose
that we do, and should, classify different kinds of material substance on the basis
of their macroscopic appearance and behavior, rather than by reference to their
supposed “real essences,” in the shape of their “internal constitutions.”
Where Locke was wrong was in thinking that this makes our classifications of
substances mere “Inventions and Creatures of the Understanding” (Essay 3.3.11)
or that “the ranking of Things into Species … is done by us, according to the Ideas
that we have of them” without any relation to their “real internal Constitutions”
(Essay 3.6.13). For what modern chemistry has revealed, but Locke could not
have known, is that our natural classification of substances into kinds on the
basis of their observable characteristics matches, at least to a considerable degree,
real divisions in nature at the level of atomic and molecular composition and
Locke on General Ideas  •  265

structure. There is not, as Locke supposed, infinite variability of composition and


structure at this level: the idea of the “Great Chain of Being,” to which he so con-
fidently subscribed, has not in fact been borne out by modern science. Even so,
Locke was right to suppose that our natural classifications do not need to be, and
in fact should not be, forced into a single all-​embracing taxonomic scheme. Real
divisions in nature are reflected by our natural classificatory schemes, but they are
often divisions at different levels, allowing for a good deal of cross-​classification.
So, to revert to an earlier example, there is nothing wrong in saying that, for some
purposes, diamond, graphite, and charcoal may be regarded as different kinds of
substance—​with diamond, for instance, but not graphite, being classified as a
species of gemstone—​whereas for others they may all be regarded as different
forms of the same kind of substance, namely, carbon. And the same may be said
with regard to water and ice. In both cases, there are relevant differences both
at the macroscopic level and at the microscopic level. It would be wrong, thus,
to insist that it is just a mistake, borne of scientific ignorance, to say that water
and ice are different kinds of stuff. For some purposes, including some scientific
ones—​in the domain, for example, of meteorology—​this is a perfectly acceptable
thing to say and the relevant macroscopic differences reflect, of course, relevant
differences at the molecular level: differences not in molecular composition, but
in the structural organization of molecules. The latter differences are irrelevant
as far as the chemical, as opposed to the physical, properties of ice and water are
concerned, which is why for chemical purposes we are entitled instead to regard
water and ice as being the same kind of substance.
Where, then, does this leave us, with regard to the debates about the meaning
and reference of natural kind terms like ‘water’ that were started by Locke and
revisited so momentously by Kripke and Putnam? My suggestion is that Locke
had more right on his side than Kripke and Putnam did, unfashionable though it
may now be to say this. It would be wrong to hold that it is just a mistake, revealed
by modern chemical science, ever to say, as Locke did, that water and ice are dif-
ferent kinds of material substance. Rather, for some purposes it is legitimate to say
this and for others it is not. Furthermore, there is no good reason to suppose, with
Kripke and Putnam, that competent speakers of English in Locke’s day were sim-
ply referring to H2O when they spoke of “water,” despite their complete ignorance
of any such molecular structure. Even more emphatically, there is no good reason
to suppose that ‘water’ refers to H2O in “every possible world,” that is, that being
composed of H2O molecules is a “metaphysically” necessary feature of water. If
water has any necessary features, they are macroscopic, observable ones. None of
this is to deny that there is an important connection between water and H2O, at
least in this, our actual universe. The physical and chemical laws that obtain in
our universe explain why, in this universe, water is everywhere composed of H2O
2 6 6   •  The Problem of Univer sals in Early Modern philosophy

molecules and why no kind of substance in it that is differently composed is at all


similar, let alone exactly similar, to water in respect of its observable, macroscopic
properties. But these facts go nowhere at all toward vindicating the distinctive
Kripke–​Putnam theses regarding the meaning of the word ‘water’ and the essen-
tial properties of that substance. On the other hand, Locke was wrong to think
that there is no mind-​independent basis in nature for the distinction between
water and other substantial kinds, because he was wrong to suppose that nature
allows for a continuous qualitative gradation between any one substance and
any other. There are combinations of observable qualities which cannot be found
occurring in nature—​indeed, vastly many more such combinations than there
are ones that can be found. And modern chemical science explains why this is so,
in terms of facts about atomic and molecular composition and organization. It
turns out that the ancient idea of the “Great Chain of Being” is deeply mistaken,
for reasons that Locke could never have anticipated, since they ultimately turn on
certain fundamental features of quantum mechanics that differentiate it radically
from the classical mechanics of Newton.

Abbreviations
A = Alciphron, Berkeley 1948–​57, 3:31–​329; cited by dialogue, section,
and page.
DM = De Motu, Berkeley 1949–​57, 4:31–​52 cited by section.
PHK = P  rinciples of Human Knowledge, Berkeley 1948–​57, 2:41–​113; cited
by section.
DHP = Three Dialogues between Hylas and Philonous, Berkeley 1948–​57,
2:163–​263; cited by dialogue and page.
NTV = Essay towards a New Theory of Vision, Berkeley 1948–​57, 1:171–​239;
cited by section.
12 BERKELEY ON ABSTRACTION, UNIVERSALS,
AND UNIVERSAL KNOWLEDGE

Tom Stoneham

In this chapter I make three claims. The first is that while Berkeley treated
the metaphysical problem of universals as unproblematically resolved in
favor of nominalism (which he interpreted in an extreme form—​see
Stoneham 2002, 238–​44), he recognized the epistemic problem as a
separate issue he needed to engage with and this is the primary positive
contribution of his attack on abstraction. The second is that his solution
to the epistemic problem is semiotic, but his semantics here is anthropo-
centric and pragmatic (in contrast to the semantics of visual language).
This will take up the bulk of the chapter. The third is that this semantic
theory, while it emphasizes the role of signs and thus has some affinities
with formalism, has no special role for formal properties of signs and in
fact makes formalism hard to achieve.

12.1. The Problem of Universals


Berkeley’s most direct engagement with the problem of universals
appears in the discussion of abstract ideas in the Introduction to the
Principles. In the published version of 1710 this appears largely as an
internal debate among nominalists about how best to account for the
meaning of general terms, and in particular, whether it is necessary
to admit a class of abstract ideas. In other words, it appears that “the
universally received maxim, that every thing which exists, is particular”
(Three Dialogues between Hylas and Philonus; DHP1, 192) is taken as
a premise in his discussion, and a more Platonist option is never seri-
ously considered.1

1. Why Berkeley should feel so confident that nominalism is universally received


is not my topic here, though it is an interesting question. One would need one
2 6 8   •  The Problem of Univer sals in Early Modern philosophy

But if the metaphysical question is resolved in this manner, that just makes
more pressing the epistemic question of universals: if we have universal knowl-
edge, knowledge which extends beyond our knowledge of particulars, then
what is that knowledge about? This is not Hume’s epistemological problem of
induction, it is not a problem of how we get to know universal propositions, but
rather the problem of what it is that we know when we do know them. For the
Platonist, to know that man is mortal is not to know something about particular
men (though it entails that each particular man is mortal), but to know some-
thing about the form of Man, the universal nature of Man.2 But the nominalist
must reject this, raising the question of what universal knowledge is about. The
young Berkeley drew attention to this in the Manuscript Introduction version of
Principles of Human Knowledge (PHK Intro 15):

For tho’ it be a point much insisted on in the Schools, that all Knowlege
is about Universals, yet I could never bring my self to comprehend can by
no means See the necessity of this Doctrine. … It is true one thing for a
Proposition to be universally true, and another for it to be about Universal
natures or notions. (Belfrage 1987, 26)

But what is it for a proposition to be universally true? To be true it must be true


of something or other, and to be universally true it seems that it must be true of
something universal. It appears that the abstractionist is offering a nominalisti-
cally acceptable account of what it is to be universally true, namely that it is to
be true of abstract ideas. These are not the widely rejected universals, for they are
merely the “work of the mind” (Locke, Essay 2.5.2 and passim), but they some-
how manage to get us beyond particularity. If Berkeley rejects abstraction, he is
faced with a significant problem of the cognitive content of universal claims.
When we consider this question, it is significant that Berkeley takes Locke
to be his primary antagonist and that he uses a discussion of Locke to introduce
the topic of generality in the Principles (PHK Introd. 11). Locke had noticed the
distinction between merely plural and fully general or universal claims.

to investigate the publications and reception of the 1675 Oxford University Press edition
of Ockham’s Summa Logicae, which appears to have been one of the few editions since the
fifteenth century. And there is, of course, Hobbes’s nominalism as an empiricist precursor,
though Berkeley is hardly like to have included Hobbes within the “received wisdom.” It is
also worth noting that Locke attributed the view that everything that exists is particular to
Malebranche as well (Locke 1823, 240).
2. I hope it is obvious that I am here picking up on the seventeenth-​century usage of ‘man’ to
name our species rather than just one sex. Less controversial examples would be possible, but
I want to keep in the reader’s mind the distinction Locke makes at Essay 3.3.12 (discussed below).
Berkeley on Abstraction, Universals, and Universal Knowledge  •  269

The next thing therefore to be considered, is, what kind of signification


it is, that general words have. For as it is evident, that they do not signify
barely one particular thing; for then they would not be general terms, but
proper names; so on the other side it is as evident, they do not signify a
plurality; for man and men would then signify the same, and the distinc-
tion of numbers (as the grammarians call them) would be superfluous and
useless. (Essay 3.3.12)

Thus “All the men are mortal” involves plural reference to all the particular men.3
But “Man is mortal” does not seem to have the same content: it is not merely
about all the particular men, but about all possible men in virtue of their shared
humanity. When we assert that man is mortal, the proposition we assert is not
just about some collection of men, perhaps all the men we have met or will meet,
but anything whatsoever which is a human. As Berkeley puts it when giving the
example of a geometrical proof, its generality extends to “all particular right lines
that may possibly exist” (PHK Introd. 12). But if, as Locke claims, “All things
that exist being particulars” (Essay 3.3.1), then the nominalist struggles to find a
subject matter for the proposition: there just are the particular men and nothing
else for it to be about.
One might think that the most natural thing for an empiricist to say about
this is that “Man is mortal” has no content beyond “All men are mortal.” But what
Locke is drawing our attention to here is in fact independent of treating ‘man’ as a
noun phrase. For while a twentieth-​century logician might tell you that “All men
are mortal” is equivalent to a huge conjunction of singular propositions, it cannot
be the case that the content of our knowledge is given by this conjunction. Nor
does the model-​theoretic approach help, for telling us that it is true if and only
if each element of the extension of “is a human” is an element of the extension of
“is mortal” does not tell us how we think of the extension of those predicates: do
we think of them by means of plural reference or some other way? The prob-
lem of the content of universal knowledge applies just as much to the proposi-
tion “All men are mortal.” Locke’s contrast between “man” and “men” would be

3.  It is possible that Locke is here alluding indirectly (via the “grammarians”) to De
Interpretatione 7, where Aristotle distinguishes “Every man is white” from “Man is white,”
saying both have universal subjects but only the former has “universal character.” However,
Aristotle’s distinction rests some claims about negation (which seem to ignore the possibility
that a single sentence can have two negations of different scope) and is not really relevant to
what Locke is considering. More significant is Aristotle’s move from a division of things into
universals and particulars to a division of propositions along the same lines, “depending on the
type of thing about which a claim is made” (Whitaker 2002, 83). Thanks to Kenneth Pearce
for suggesting the allusion to Aristotle.
2 7 0   •  The Problem of Univer sals in Early Modern philosophy

better made as the contrast between “All men are mortal” and “All the people in
this room are mortal”: despite the universal quantifier, the latter is not universal
knowledge but knowledge about a plurality. The point Locke is making is really
one about predication: a predicate is not the same as a plural name for everything
it applies to. As Quinton nicely puts the point: “I cannot be introduced to the
entire extension of a predicate” (1973, 261).4 So the epistemic problem arises for
even the most faithful empiricist who accepts predication.
As we have seen, Locke’s abstract ideas provide one solution to this problem,
but one which Berkeley roundly rejects. In the Manuscript Introduction and
PHK Introd. 15, Berkeley shows himself to be concerned with what universal
knowledge is about and that is also one explicit motivation behind Locke’s appeal
to abstraction (Essay 3.3.6). Much of the importance for Berkeley of his alterna-
tive to abstract ideas appears to have been that it removed a “cause of error and
difficulty in the sciences” (Principles, title page). So far, this appears to have noth-
ing to do with immaterialism but to be a separate philosophical problem.

12.2. The Role of the Anti-​Abstraction Arguments


However, some scholars have argued that Berkeley’s rejection of abstract ideas is
primarily there as a crucial premise in his argument for immaterialism.5 Despite
their philosophical inventiveness, these interpretations simply lack plausibility
when we consider the texts as whole works. While the rejection of abstraction is
emphasized, it is not given the structural role we would expect of a major prem-
ise. For example, the Introduction to the Principles is numbered separately from
the main text, making it clumsy to refer back to it, despite the fact that Berkeley
uses paragraph numbers for cross-​reference both within the Introduction and
within the main text. There are four explicit references to the Introduction in
the main text of the Principles, two generic (PHK 97 and 120), where the pos-
sibility of the specific abstract ideas of time and unity is being rejected and the
Introduction is referred to for the general argument against abstraction, and
two references to specific sections, namely PHK 122, which refers to Introd. 19,
and PHK 126, which refers to Introd. 25. The former is to point out that the
account of arithmetic being given is “agreeable” to the account of general terms

4. This is quoted in Margolis 1982, a rich and subtle paper that, in effect, lays the foundations
for what follows. See especially pages 210–​12.
5. The strongest version of this claim is probably Pappas 2000 (see esp. chap. 2). Other versions
can be found in Atherton 1987; Bolton 1987; Bracken 1974, chap. 4; Doney 1982, 274; Tipton
1974, 133, 157; and Warnock 1953, 187.
Berkeley on Abstraction, Universals, and Universal Knowledge  •  271

in the Introduction, and the latter to refer the current discussion of geome-
try back to the earlier discussion. None of these suggest that the rejection of
abstraction is a premise in the argument for immaterialism, though they do
suggest it has some role to play in the wider project of removing the causes “of
error and difficulty in the sciences.” And again in the Three Dialogues the rejec-
tion of abstraction is referred to in general terms but the argument is neither
repeated nor even cited. Compare this to the way that the theory of vision is
treated as an explicit premise in Alciphron (A 4), and An Essay towards a New
Theory of Vision (NTV) was even reprinted with Alciphron. The suggestion that
Berkeley took the arguments of the Introduction to be a crucial premise in the
arguments for immaterialism just does not fit with how he chose to present and
publish the material.
Others, myself included (Stoneham 2005, 154ff.), have tried to show that
the rejection of abstract ideas has this purely negative role to play in Berkeley’s
thought. On this interpretation, Berkeley does not take abstraction to be an
important philosophical error in itself—​after all, it is a form of nominalism—​but
one which leads some philosophers astray into the thickets of materialism. By
rejecting abstraction, Berkeley seeks to remove a crutch which the crippled mate-
rialist might rely upon.
While the latter interpretation has a fair amount of support in the texts and
captures something Berkeley definitely wanted to achieve, it also underestimates
him as a philosopher. Both interpretations have a common fault:  they treat
Berkeley as exclusively concerned with the defense of immaterialism. Of course,
immaterialism will permeate all Berkeley’s thought, but then so does materialism
permeate most other philosophers’ thinking. He was a sophisticated and well-​
read enough philosopher to know that the solutions to some important philo-
sophical problems might be neutral with respect to immaterialism, but that does
not make the problems any less worthy of his interest and attention. So, even if
it had other functions as well, we can legitimately regard his attack on abstrac-
tion and his attempt to find a nominalistic alternative as a direct answer to the
epistemic problem of universals. And treating it like this turns out to make recon-
structing Berkeley’s views a little easier.

12.3. The Semiotic Solution to the Epistemic Problem


Much ink has been spilt over the question of whether Berkeley’s criticisms of
Locke’s theory of abstraction are fair, but our interest here is in reconstructing his
alternative, nominalistic account of the meaning of general terms. Unfortunately,
this is never systematically laid out, but when we regard it as an attempt to solve
2 7 2   •  The Problem of Univer sals in Early Modern philosophy

the epistemic problem of universals, it can be reconstructed from various short


passages (emphases mine):

a word becomes general by being made the sign, not of an abstract general
idea but, of several particular ideas, any one of which it indifferently sug-
gests to the mind. (PHK Intro 11)
an idea, which considered in it self is particular, becomes general, by
being made to represent or stand for all other particular ideas of the same
sort. (PHK Intro 12)
universality, so far as I can comprehend, not consisting in the absolute,
positive nature or conception of any thing, but in the relation it bears to
the particulars signified or represented by it:  by virtue whereof it is that
things, names, or notions, being in their own nature particular, are ren-
dered universal. (PHK Intro 15)
there is no such thing as one precise and definite signification annexed
to any general name, they all signifying indifferently a great number of par-
ticular ideas. (PHK Intro 18)

Words have semantic properties such as signification, denotation, and reference,


so it is easy to see Berkeley as here primarily concerned to give an alternative to
Locke’s philosophy of language (e.g., Stoneham 2002, 224–​37):  general terms
do not denote or signify or represent ideas in the mind of the speaker; rather,
they signify all the particular things of that kind. 6 They are not names for any or
even all of those particulars, for they are not names at all; rather, their semantic
relation to those particulars is one of “indifferent denotation.” What exactly this
amounts to, we are not told, but we can reasonably infer that it is not plural refer-
ence. Rather, it is an entirely different kind of semantic relation. Locke has a view
which effectively requires each term to be the name of an idea, and that forces
him to find ideas for general terms to name, but Berkeley points out, perfectly
correctly, that general terms just do not function like that.
Furthermore, he expresses the account with verbs of action, such as ‘being
made’ and ‘rendered’. This makes clear that he sees generality as a phenomenon of
human language, a phenomenon we actively introduce into a language on top of
its referential semantics.7

6. Locke uses ‘semiotic’ at Essay 4.21.4 and I suspect we can antedate this usage. In contrast, the
OED has no uses of ‘semantic’ prior to the nineteenth century. However, I shall talk of denota-
tion, etc., as the semantic properties of signs.
7. One might think that the next point, about the need for an account of generality in thought,
arises here, for how can we make a sign general without having the intention to do so, and that
Berkeley on Abstraction, Universals, and Universal Knowledge  •  273

However, treating this as a problem in the philosophy of language8 appears to


leave unanswered the more fundamental question which was also addressed by
the appeal to abstract ideas: How do we think general thoughts, which are pre-
sumably what we are trying to express with these general terms? We can see the
need for such an account in the rather rhetorical conclusion to the Introduction:

And he that knows names do not always stand for ideas, will spare himself
the labour of looking for ideas, where there are none to be had. It were
therefore to be wished that every one would use his utmost endeavours, to
obtain a clear view of the ideas he would consider, separating from them
all that dress and encumbrance of words which so much contribute to
blind the judgment and divide the attention. (PHK Intro 24)

The first sentence tells us not to go looking for abstract ideas as the meanings
of general terms and the second tells us to focus our attention not on words but
on the ideas we have in mind. But if all Berkeley had said about universality was
to give an account of the semantics of general words, then when we discard the
dress of words, we would be left with nothing but particular ideas in our minds: it
would seem that there is no space for general thoughts in Berkeley’s system.
And yet section 15 seems to confidently assert that he can allow for universal
knowledge:

It is I know a point much insisted on, that all knowledge and demonstra-
tion are about universal notions, to which I fully agree: but then it doth
not appear to me that those notions are formed by abstraction in the man-
ner premised; universality, so far as I can comprehend, not consisting in
the absolute, positive nature or conception of any thing, but in the relation
it bears to the particulars signified or represented by it. (PHK Intro 15)

If generality were an entirely linguistic phenomenon, and Berkeley thinks that we


can and should turn our attention from words to the contents of our minds when
doing philosophy, it would be puzzling how he can be so confident that there is

intention must have a general content. However, it is an oversimplification to think that all
semantic intentions are simply intentions to match up signs with preexisting concepts.
8. It seems that in the Manuscript Introduction Berkeley did see generality as an entirely lin-
guistic phenomenon, but by 1710 he realized that if he wanted to avoid the “embarrass and
delusion of words” and still allow for universal knowledge, he would need to allow general
ideas in some sense. See Pitcher 1977, 82–​83.
2 7 4   •  The Problem of Univer sals in Early Modern philosophy

universal knowledge which is “about universal notions” rather than the particular
(for all ideas are particular) ideas adverted to in section 24.
There is, however, no real puzzle because in section 12 he already talks about
“how ideas become general,” and he immediately follows it with an example of a
geometric proof. Thus it seems that he is prepared to attribute semantic proper-
ties not only to words and diagrams but also to the particular ideas we have before
the mind. Crucially, he is prepared to attribute to particular ideas the semantic
property of indifferently denoting all things of a certain sort. And they acquire
this semantic property in the same way that words do, namely by a deliberate act
of ours to make them “indifferently denote.” The crucial move here is to allow
that ideas are the sorts of thing that can be signs.9
That particular ideas can signify other ideas is a crucial feature of the New
Theory of Vision (NTV, 1709). That work is primarily addressed to the question
of how we see depth; that is, how do we see objects as being distant from us? His
premise is that nothing strictly visible could ever carry information about depth
because

distance being a line directed end-​wise to the eye, it projects only one
point in the fund of the eye, which point remains invariably the same,
whether the distance be longer or shorter. (NTV 2)

He concludes that distance is only ever immediately perceived by touch, how-


ever, there are regular, though contingent and arbitrary, connections between
certain visual appearances and tangible distances. These contingent connections,
once learnt, allow us to know that certain visual appearances represent distance.
Berkeley thinks that not merely do these visual objects—​light, colors, shapes—​
possess semantic properties when combined in various ways, but also that they
meet the other conditions for being a language and that we should treat the
objects of vision as words in a Language of Nature giving us information about
the unseen which we need in order to survive (on the claim that they literally
form a language, see Stoneham 2013). Because these semantic properties derive
from natural connections which we experience all our lives, we do not notice our-
selves learning them. Furthermore, once we have learned them, the visual ideas
suggest to the mind the ideas they signify without any conscious intervention on
our part, much like the way we hear the meaning of words spoken in a familiar
language without noticing the step from sound to meaning, and often without
even noticing the sounds at all. Hence we think we see distance, for when we see

9. There is good reason to think this rules out an adverbial account of ideas.
Berkeley on Abstraction, Universals, and Universal Knowledge  •  275

certain visual cues, strictly tangible ideas of distance are straightaway presented
to the mind, just as if they were part of the visual experience. And in a sense they
are: because the objects of visual experience—​light, colors, shapes—​have seman-
tic properties, they signify other ideas; to one who knows those signification rela-
tions the visual experience has a double object, the sign and what it signifies, in
precisely the way our experience of human languages has a double object, namely
the words and what they signify.
We can now see Berkeley’s account of generality as extending the range of
semantic properties possessed by ideas. In NTV, the semantic properties of visual
ideas were natural, they were created by the contingent, arbitrary, but systematic
connections between those ideas and ideas of touch (as well as other ideas of
sight). The semantic properties of human languages are human creations, be they
singular or general. Given the strict parallelism in the Introduction between the
generality of words and of ideas, one can infer that this semantic property of our
ideas is also a human creation: it is we who make particular ideas stand indiffer-
ently for all ideas of the same sort. Hence the emphasis on words and ideas “being
made to represent.” Thus, when thinking the general thought I might choose to
express as “Man is mortal,” I might have before my mind the idea of some par-
ticular human, it matters not whom but let us say Xanthippe, and I think about
humans,10 rather than just Xanthippe, by giving that particular idea of Xanthippe
the semantic property of standing for all humans whatsoever.
If this is Berkeley’s account of generality, why is he so confident that it solves
the epistemic problem of universals, that it explains the cognitive content of uni-
versal claims? If those claims are not about Platonic universals, and they are not
about pluralities of particulars, and nor are they about abstract ideas, what are
they about? The closest Berkeley comes to answering this question directly is in a
much later work, where he is defending the possibility of belief in the Christian
mysteries by drawing parallels with scientific knowledge:

If I mistake not, all sciences, so far as they are universal and demonstrable
by human reason, will be found conversant about signs as their immediate
object, though these in the application are referred to things. (A 7.13, 305;
see also DHP1, 173, DM 7, A 7.11, for similar passages).

Notice how this echoes the PHK Introd. 15 comment that “all knowledge and
demonstration are about universal notions.” So universal sciences are actually

10. Here we can see that the usage of ‘man’ and ‘men’ for humans rather than just males becomes
impossibly strained when we use an example of a woman. Which is why it is appropriate that
that usage should be actively discouraged.
2 7 6   •  The Problem of Univer sals in Early Modern philosophy

about signs—​words, diagrams or particular ideas—​and these are their immedi-


ate content; but being signs, these objects have signification and thus by being
about signs the sciences are also about what those signs represent. A sign is not
a formal object, but something that essentially has a semantic value. Even in the
case of arithmetic, where we have the option to study the signs independently of
what they signify because they are “capable to represent aptly, whatever particular
things men had need to compute” (PHK 122), if we do study them “for their own
sake,” this is as cognitively pointless as “controversies purely verbal” (PHK 122).
Which is to say that in all cases of universal knowledge, there are two
objects: the signs and what they represent or signify. The former gives no knowl-
edge without the latter, but the latter is, in the case of generality, unthinkable
without the former. So that on any given occasion we consider the proposition
that man is mortal, the immediate cognitive content is a particular idea of a par-
ticular human, say Socrates, or even a particular word, ‘men’. Thus, for example,
when I consider that proposition, I might be thinking of Socrates and Socrates’s
famous demise. But if Socrates and his particular death are signs and have been
“rendered universal,” then they signify all men and all deaths, and in virtue of
thinking about the particulars I can also think and know the universal proposi-
tion they represent.
Thus, Berkeley’s solution to the problem of universal knowledge is that, unlike
knowledge of particulars that is concrete and direct, it is essentially semiotic. The
problem was generated in such a way that we seemed to need a special kind of
object to be the kind of thing a universal proposition is about, but all such objects
are found wanting. Berkeley denies we need a special kind of object and instead
finds them to be about ordinary objects which have special properties, namely
they have been “rendered universal” in the sense of being made—​by us—​to repre-
sent all things of a particular sort.
This is a striking and original solution to an age-​old problem, but one might
think that all it does is sweep the problem under the carpet. Surely the prob-
lem recurs when Berkeley talks of an idea or other sign being made to signify
all things “of a particular sort” (PHK Introd 12)?11 For then we should ask what
makes it the case that two particulars are of the same sort, for example, both are
men and thus, since we have claimed that man is mortal, that both are mortal.
If the reason both are men is that they are each one of the men, where ‘the men’
refers to a specific collection of particular men, then it seems that we have failed
to make a universal claim at all and our general term ‘man’ is just another name,

11. This is a familiar criticism of Berkeley on abstraction which can be found in several places,
including Aaron 1967, 65; Bolton 1987, 65–​66; and Pitcher 1977, 89–​90.
Berkeley on Abstraction, Universals, and Universal Knowledge  •  277

but for a plurality rather than a singularity. However, if we say that both are men
because both have the property of being human, or both partake of humanity,
then we have appealed to something nonparticular, a universal of some kind.
It seems that the Platonist and the abstractionist both have accounts of what
makes several distinct particulars belong to the same sort, what make Socrates,
Xanthippe, and Plato all humans; but, by doing away with anything except the
particulars themselves, Berkeley appears to leave himself without the possibil-
ity of such an account. When particulars are rendered universal they are given
a semantic property which determines that they signify some things and not
others, and this property creates a partition not merely among the experienced
objects, or even the potentially experienced objects, but among all objects what-
soever. The possibility of having such a semantic property is precisely the philo-
sophical problem we are dealing with, because it is the problem of determining
the cognitive content of a universal claim.
Berkeley is well aware that Platonists and abstractionists are giving an account
of what makes several things all of one sort, but seems to think such an account
is unnecessary:

From which it must necessarily follow, that one word be made the sign of a
great number of particular ideas, between which there is some likeness, &
which are said to be of the same sort. But these sorts are not determin’d &
set out by Nature, as was thought by most philosophers. Nor yet are they
limited by any precise, abstract ideas settled in the mind, with the general
name annexed to them as is the opinion of the author of the Essay, nor
do they in truth, seem to me to have any precise bounds or limits at all.
(Berkeley 1948–​57, 2:128)12

However, his optimism that such an account is unnecessary seems unfounded


and he is left with nothing to say about an important philosophical problem.
Those who feel the force of this objection have underestimated the extent
of Berkeley’s pragmatism. Remember that he has said the immediate objects of
universal sciences are in fact signs. The philosophical theory that mathematics is
really just about signs and symbols and the rules for manipulating them, known
as formalism, is a form of antirealism, so we should be struck by the antirealist

12. I have here quoted from Jessop and Luce, in Berkeley 1948–​57, despite the liberties they
take with the text. This is because the end of this passage involves much crossings out and
rephrasings and I only want to illustrate Berkeley’s awareness of the point rather than his con-
sidered opinion upon it. The published introduction contains even less on the issue, support-
ing my reading that he takes it to be unnecessary to give an account of what makes for sorts of
thing. See Appendix.
2 7 8   •  The Problem of Univer sals in Early Modern philosophy

tone of this view of Berkeley’s (though, as PHK 122 makes clear, he is no for-
malist). By saying that the immediate object of universal knowledge is the signs,
Berkeley has moved away from a straightforward realist position, which holds
that the universal proposition that man is mortal is straightforwardly about men
in general, for it is necessarily about some particular object, be it a man or an idea
or a word. Yet those particular objects are also signs which are given significance,
and significance of a special sort. Thus, mathematics and other universal sciences
and branches of knowledge are not only about signs, just immediately about them.
What those signs signify are concrete particulars.
The objection being raised asks what determines that a given general sign
indifferently signifies all and only particulars of a given sort, what makes it that
‘men’ signifies all and only men. And the first step in Berkeley’s answer is the
antirealist move of claiming that, independently of human activity, specifically
human representational activity, there are no facts about what sorts of things
there are.13 It is because we have a sign which signifies indifferently Socrates and
Xanthippe and Plato and all other humans that they are of the same sort. The
semantic properties of general signs do not track the sorts of things there are in
the world, for a thorough-​going nominalist no more believes in objective sorts
than in Platonic forms. Rather, when we “render the sign universal,” when we
make it represent in a certain way, we also create the sameness of sort possessed
by all humans.
Given that we create these semantic properties, Berkeley not unreason-
ably concludes that they are a function of our natures and interests. But now a
new version of the problem occurs, for if the distinction between a human and
another animal on the basis of which the term ‘man’ applies to the former and
not the latter is a function of our interests and nature, then it is far from obvious
that the distinction will in fact project determinately over an indefinitely large
and varied set. Consider Lucy, the famous australopithecus afarensis, whose com-
plete fossilized skeleton was discovered by Johanson and Gray in 1974. Was she
human? Does our commitment to the universal claim that humans are mortal
include Lucy? Well, some paleoanthropologists talk of “early humans” and others
are careful to avoid that and only talk of hominins. Is there a definitive answer as
to whether Lucy is a human or not, whether she and I are both this sort of thing?
Berkeley’s answer appears to be that the answer is relative to our interests in mak-
ing the classification, and ultimately our interests boil down to the “never enough

13. Of course, there is nothing to prevent God from rendering some sign universal, but then
either he is doing it by reference to his own interests or ours. If the latter, then we can grasp the
signification of the sign, but the sameness of sort is no different from that possessed by signs we
make general. If the former, we cannot grasp the signification.
Berkeley on Abstraction, Universals, and Universal Knowledge  •  279

admired laws of pain and pleasure” (PHK 146). Thus, if a question of classifica-
tion is not useful, that is, it has no impact however indirect on human14 pleasure
and pain, then it is arbitrary. Thus, if we find it useful to treat Lucy as a human,
then we should; if we find it hinders us, perhaps impeding our understanding of
evolution, then we shouldn’t; if it seems to be indifferent to us, then perhaps there
is no determinate answer. As he put it in the Manuscript Introduction (admit-
tedly here thinking exclusively about language):

nor do they [sorts] in truth, seem to me to have any precise bounds or lim-
its at all. For if they had I do not see, how there could be those doubts &
scruples, about the sorting of particular beings, which are observ’d some-
times to have happened. Neither do I  think it necessary the kinds or
species of things should be so very accurately bounded & marked out.
(Berkeley 1948–​57, 2:128)15

As such, Berkeley’s approach to universal knowledge may in fact only secure that
our universal knowledge is humanly universal, that universal propositions apply
determinately only within the range of actual and possible human experience, and
even then they will only be as determinate as we have reason to want them to
be.16 Since the sameness of sort that we recognize is a human construction, from
God’s perspective—​at least from God’s perspective on the world of ideas—​there
are no facts of the matter whether two particulars belong to the same sort or not
(though there will be facts about whether humans take them to so belong, a fact
which God can make use of ). It follows that, at least from God’s perspective, our
universal knowledge is not truly universal. But if true universality is possible, it is

14. Berkeley is quite liberal about which species can feel pleasure and pain, and would thus
allow that a classification may be nonarbitrary in virtue of its impact on, say, avian or even
apian pleasure and pain. But we can only make our words general in that way by reference to
those other species hedonic states, so we would have to know about them and choose to use
our terms that way.
15. Again I quote from the inaccurate Jessop and Luce edition (in Berkeley 1948–​57) rather
than the Belfrage diplomatic edition (Belfrage 1987), this time because the whole passage is
struck through in the manuscript. While this passage lends some support to my interpretation,
I do not offer it as evidence of Berkeley’s considered views, merely his earlier openness to the
kind of pragmatist position I am articulating here. See Appendix.
16. We can, and should, ask whether there is space here for incorrect classification. The answer
seems to be that an individual can mistake what the cognitive community has determined the
most useful classification to be, and a community can mistake which classification best serves
its interests. But it seems that the community, though not the individual who is trying to make
her classification conform to the community’s, cannot misapply its own classification. Clearly,
at this point the epistemic problem of universals connects with the skeptical problem about
meaning raised by Kripke’s Wittgenstein (Kripke 1982).
2 8 0   •  The Problem of Univer sals in Early Modern philosophy

only accessible to an infinite being, for only such a being could render one of its
signs to be truly universal. Hence, insofar as our universal sciences fall short of true
universality and express only the humanly universal, that does not matter, for there
is nothing they are missing which is intelligible to us finite beings. Furthermore,
even if there is true universality, even if God does render signs universal, that is not
absolute or objective universality, for it is still relative to His (infinite) interests.
Thus, Berkeley’s solution to the problem of universal knowledge is that, unlike
knowledge of particulars, which is concrete and direct, it is essentially semiotic and
indirect. This is a nominalist solution because the semantic properties of our general
signs are merely arbitrary connections between particulars that are created by finite
minds. But we should admit that Berkeley has not really given an adequate solution
to the original problem, if we are to take the condition of an adequate solution to
be one which achieves all that the question presupposes an answer would achieve.
For the epistemic problem of universals was the problem of how we make claims
with universal content, claims which extend beyond our knowledge of particulars
and are true of an indefinitely large number of things with which we may have no
acquaintance: how can we make claims about all beds or all humans if we, and the
people we talk to, have only ever come across some finite subset of beds or humans?
Berkeley’s answer reveals that our universal claims are not really as universal as the
Platonist, and perhaps the abstractionist, took them to be. They do extend beyond
our individual experience and the collective experience of any specific group of us,
but not indefinitely, for their scope is determined by human interests and those are
finite. Were an infinite being to create general signs, they might enable him to make
fully universal claims, but we could not grasp their signification. Instead, we must
settle with what is humanly universal.

12.4. Language and Formalism


The account I  have given here of Berkeley’s views on universal knowledge has
certain similarities to the account of his views on scientific knowledge given in
(Peterschmitt 2009). However, as is so often the case in these matters, the differ-
ences are more significant than the similarities. I shall discuss two.
The first significant difference is scholarly. Peterschmitt suggests that the
development of Berkeley’s thought on these matters between 1710 and 1732 is so
substantial that “on peut dire qu’elles sont parfaitement contraires” (2009, 413).
The crucial change is that in 1710 Berkeley takes our knowledge to be restricted
to our ideas, but in 1732 he recognizes how the formal character of the languages
of science allows us to extend our knowledge beyond our ideas. On my view, in
contrast, the epistemic problem of universals, which is precisely a problem of
how our knowledge can extend beyond our particular ideas, had been a matter
of concern since 1708 and a solution was offered in 1710. Admittedly, some parts
Berkeley on Abstraction, Universals, and Universal Knowledge  •  281

of the solution are not made fully explicit until Alciphron in 1732, but this does
not represent a change of view but an addition of important detail. Furthermore,
in the discussion of arithmetic at PHK 122, where we are told that “we regard
not the things but the signs” in recognition of its formal (though not formalist—​
see above, section 12.3) character, Berkeley explicitly notes the connection with
“what we have before observed, of words in general [PHK Introd. 19].” And in a
letter to Molyneux (December 19, 1709), Berkeley writes that “to me it appears
that all grammar & every part logic contain little else than rules for discourse &
ratiocination by words” (Hight 2012, 31).
The more interesting disagreement with Peterschmitt is over the formal character
of languages which extend our knowledge beyond our ideas. It is to a large extent this
emphasis on the formal in Alciphron which makes Peterschmitt think there is such a
large departure from the Principles. But ‘formal’ is not one of Berkeley’s words and it is
far from obvious to me that formal properties are really doing much work at all here.
One clear sense in which a language can be formal is that it contains syntactic
inference rules. That is, rules which allow one to perform valid inferences without
knowing the meaning of the terms in question. Clearly algebra is a formal language
in this sense, but in fact all natural languages have some formal elements. Any given
language can be more or less formally complete; that is, its syntactic inference rules
can allow one to perform a greater or lesser proportion of the valid inferences stat-
able in that language. Mathematics and the artificial languages of formal logic tend
to have a high degree of completeness—​possibly 100 percent—​whereas natural lan-
guages contain many valid inferences which are not captured by syntactic inference
rules (famously: if the book is red it is not green).
Now, it seems that any given science is distinctively formal in this sense—​
that is, is formal compared with nonscientific or vulgar discourse—​exactly to the
extent it is mathematical in the broad sense, which includes mathematical logic.
And it may well be true that all sciences are more mathematical than nonscientific
discourse, but it looks like there is a continuum here, with the vulgar using a fair
amount of basic geometry and arithmetic, and theoretical physics being almost
entirely mathematical. Being formal in this sense certainly enables a language to
extend our knowledge beyond our immediate ideas and in ways that have practi-
cal consequences. A simple example which has nothing to do with physics and
mechanics is when someone analyzes a series of trades on a commodities market,
working out how to maximize profit, while having no idea at all about what is
being traded.17 In contrast, sign systems indicating toilets and exits, while they

17. When applied in scientific areas where we couldn’t have ideas, such as Newtonian kinemat-
ics or atomic physics, there is a question about whether the knowledge generated is knowledge
of insensible things or merely of the structure of reality. On this I disagree with Peterschmitt
(see Stoneham and Cei 2009; Stoneham forthcoming), but that is another issue.
2 8 2   •  The Problem of Univer sals in Early Modern philosophy

have a superficial appearance of syntactic rigor, in fact lack any useful formal
properties.
However, with respect to the epistemic problem of universals Berkeley is
addressing, it is not the case that the formal character of a language is relevant.
This is because the crucial point is the nondenotational semantics of the signs, not
their syntactic properties. One might think that the claim that the subject matter
of universal claims is the signs themselves makes them in some sense “formal.” But
notice that this is not the sense of “formal” which Peterschmitt is using to get his
result about knowledge extending beyond our ideas: there is nothing about signs
themselves which entails the existence of syntactic inference rules. And in fact the
situation is worse than that. Syntactic inference rules require that syntactically
specified terms are unambiguous. Consider the simple, formal, inference:

Fa
a = b
——​
Fb

Setting aside the identity sign, there are here three terms (a, b, F), each of which
occurs twice. If the semantic value of the two occurrences of any of those terms
were different, then the inference would be invalid. But notice that in natural
languages, many words, formally defined as sequences of letters, are ambiguous and
thus they can have occurrences that differ in semantic value (and if the form of a
spoken word is the sound, things are even worse). So in fact, no formally speci-
fied instance of that inference in a natural language is guaranteed to be valid.
Consider, for example,

George is a novelist.
George is Eric.
——​–––––––––––––––
Eric is a novelist.

If ‘Eric’ in the second premise refers to Eric Blair (making it true), but ‘Eric’ in
the conclusion refers to Eric Bloodaxe (making it false), the inference is invalid.
Similar problems arise with predicates. Without semantic knowledge, we cannot
tell whether this is a good inference or a fallacy.
So the more interested we are in formal inferences, the more work we need
to do to remove ambiguities from our language. But Berkeley’s sign system does
exactly the opposite, for it actively encourages radical polysemy by insisting (as
he has to, given his nominalism) that the signs which possess semantic value are
Berkeley on Abstraction, Universals, and Universal Knowledge  •  283

particulars (e.g., “so the name line which taken absolutely is particular, by being
a sign is made general”; PHK Intro 12, my emphasis), be they marks on paper,
sounds, or ideas. So if we both write the word ‘triangle’ there are two signs, and
in the inference above there are four names and two predicates: to say there are
two names and one predicate, each with two occurrences, as I did above, is not
for Berkeley to speak of signs but of what they signify, for it is to type the signs
by their significations.
Even worse, one particular sign can signify different things on different occa-
sions or in different contexts. This is particularly obvious when the signs are
ideas: my idea of Peter might be made to signify man in one context and animal
in another (PHK Intro 16, but most clear at NTV 72–​73).
Of course, there are plenty of things we can do to make a specific sign system
more formal, eliminating context sensitivity and polysemy, but these are hard-​
won achievements in math and science, not essential features of sign systems
themselves. On the contrary, given Berkeley’s theory of signs, formal properties
have no essential role at all and semantic knowledge is fundamental.18

Appendix
I quoted two consecutive passages from the Jessop and Luce edition of the
Manuscript Introduction, noting that the actual Manuscript contains many
strikings and variations that Jessop and Luce simplify. For completeness, this is
the diplomatic edition of the same passages, from the bottom of folio 10 (Figure
12.1) and the top of folio 11 (Figure 12.2):

Figure 12.1  Diplomatic transcription of Berkeley’s “Manuscript Introduction,” folio 10.

18.  Earlier versions of this chapter were presented at the Iranian Research Institute in
Philosophy, the International Berkeley Conference in Zurich, the Scuola Normale Superiore in
Pisa, and the Institute for Foreign Philosophy at Peking University. Audiences at these events
provided excellent questions and discussion that have helped improve the chapter considerably.
2 8 4   •  The Problem of Univer sals in Early Modern philosophy

Figure  12.2   “Manuscript Introduction,” folio 11.

Abbreviations
A = Alciphron, Berkeley 1948–​57, 3:31–​329; cited by dialogue, section,
and page.
DM = De Motu, Berkeley 1949–​57, 4:31–​52 cited by section.
PHK = Principles of Human Knowledge, Berkeley 1948–​57, 2:41–​113; cited
by section.
DHP = Three Dialogues between Hylas and Philonous, Berkeley 1948–​57,
2:163–​263; cited by dialogue and page.
NTV = Essay towards a New Theory of Vision, Berkeley 1948–​57, 1:171–​239; cited
by section.
13 HUME ON ABSTRACTION AND IDENTITY

Donald L. M. Baxter

Hume’s critique of traditional abstraction entails a result that under-


cuts his account of the idea of identity. To save his account of identity,
Hume would have to accept abstraction as well. What links these two
discussions is (1)  Hume’s widely shared assumption that traditional
abstraction is separating in the mind what are inseparable in reality,
(2)  his principle that what are different are mentally separable, and
(3)  his principle that we cannot conceive of the impossible. Given
these, it will turn out that abstraction is mentally separating some-
thing from itself, which will entail that abstraction is conceiving of
something as distinct from itself. But it is impossible for something to
be distinct from itself, and so it is inconceivable. Therefore abstraction
is impossible.1 Yet consider Hume’s account of the idea of identity. On
that account, to conceive of an identity is to conceive of something
as one single thing viewed one way and as two distinct things viewed
another. How we can take these opposing views of the same thing is
a problem that I’ve termed Hume’s Difficulty Concerning Identity. It
will turn out that we can take the opposing views only if we can con-
ceive of the single thing viewed one way as somehow distinct from
itself when viewed the other way. That is, we must be able to conceive
of something as distinct from itself. However, if we cannot conceive
of something as distinct from itself when abstracting, then we cannot
do it when conceiving of an identity. So traditional abstraction and
Hume’s account of the idea of identity stand or fall together.
I first give a characterization of traditional abstraction, then
explain the problem that Hume finds with it, then show that the same
problem affects his own account of identity. Along the way I  argue

1.  For predecessors of this approach see Baxter 1997, Pitcher 1977, Weinberg
1965, and Winkler 1989.
2 8 6   •  The Problem of Univer sals in Early Modern philosophy

that Locke’s version of abstraction is the traditional one rather than the partial
consideration attributed to him by some commentators.

13.1. Traditional Abstraction
Berkeley distinguishes three senses of “abstraction”:  (1)  separating in thought
parts that are actually united but can exist apart, (2) separating in thought quali-
ties that cannot exist apart, (3) separating in thought qualities that make some-
thing a particular individual from qualities that make it of a given kind.2 Hume,
like Berkeley, would object only to abstraction in the second and third senses.
His express arguments mainly concern abstraction in the third sense and on that
I will focus. It shares with the second the assumption that abstraction is separat-
ing in thought what are inseparable in reality.
Hume took himself to be confirming Berkeley’s arguments against the
“receiv’d opinion” concerning abstraction. Berkeley argues against Locke’s view.
So it is safe to assume that Hume also had Locke foremost in his mind as an
expositor of the view to be opposed. Locke is explicit that abstraction is separat-
ing in thought what are not separate in reality. Both Berkeley and Hume assume
that he is committed to their being inseparable in reality, as well. Berkeley makes
the assumption clear in his three-​way disambiguation of “abstraction.” Hume
too, as witnessed by his claim “that ’tis utterly absurd to suppose a triangle really
existent, which has no precise proportion of sides and angles” (T 1.1.7.6, 19). To
show that their assumption is correct will take a little argument. I will first discuss
mental separation, then discuss inseparability in reality.
Locke gives a version of the traditional account of abstraction.

Ideas become general, by separating from them the circumstances of


Time, and Place, and any other Ideas, that may determine them to this or
that particular Existence. (Essay 3.3.6)

And also,

Afterwards, when time and a larger Acquaintance has made [children]


observe, that there are a great many other Things in the World, that in
some common agreements of Shape, and several other Qualities, resem-
ble their Father and Mother, and those Persons they have been used to,
they frame an Idea, which they find those many Particulars do partake in;

2. Berkeley 1982, Introduction, paragraph 10.


Hume on Abstraction and Identity  •  287

and to that they give, with others, the name Man, for Example. And thus
they come to have a general Name, and a general Idea. Wherein they make
nothing new, but only leave out of the complex Idea they had of Peter and
James, Mary and Jane, that which is peculiar to each, and retain only what
is common to them all. (Essay 3.3.7)

And again,

And he that thinks general Natures or Notions, are any thing else but such
abstract and partial Ideas of more complex ones, taken at first from partic-
ular Existences, will, I fear, be at a loss where to find them. For let any one
reflect, and then tell me, wherein does his Idea of Man differ from that of
Peter, and Paul; or his Idea of Horse, from that of Bucephalus, but in the
leaving out something, that is peculiar to each Individual; and retaining so
much of those particular complex Ideas, of several particular Existences, as
they are found to agree in? (Essay 3.3.9)

So for Locke, one starts with ideas of various particulars that have been observed
to resemble in some ways. The ideas are complex ones because they are of particu-
lars with a variety of qualities. One then presumably takes one of these ideas and
leaves out parts that represent features not shared by all of the particulars. One
especially leaves out features that are unique to a given particular such as “existing
at any determin’d time and place” (Essay 2.27.1). The resulting general idea is able
to represent all the individuals that resemble in the ways first observed, because
it conforms to all of them. That is, each has all the features represented by the
general idea (Essay 3.3.6).
The leaving out of parts of the original idea is mental separation in two senses.
In the first sense it is a case of thinking of two things and then continuing to
think of one while ceasing to think of the other. In the second sense it is a case
of having two ideas (especially two parts of a complex idea) present to mind and
then continuing to have just one of them present to mind while ceasing to have
the other present to mind. On Hume’s reading of Locke, these senses can be used
interchangeably, where the relevant separation is removal. The mental separation
of objects of thought by mentally removing one of them while leaving only the
other, just is the separation in the course of thinking of the vehicles of thought—​
the ideas of those objects—​by the removal from the mind of one of the ideas
while leaving only the other.
There is an additional kind of mental separation that Locke writes of. To men-
tally separate something extended is “to make in the Mind two Superficies, where
before there was a Continuity, and consider them as removed [“disjoined”] one
2 8 8   •  The Problem of Univer sals in Early Modern philosophy

from the other” (Essay 2.13.13). In such a case, both the disjoined parts continue
to be present to mind. In contrast, in the case of abstraction some parts cease to
be present to mind and some parts continue to be present. Of these parts retained
the mind makes “a new distinct complex Idea” (Essay 3.3.9).
There is an influential way to read Locke that is an alternative to Hume’s and
Berkeley’s way.3 While I could still make my points about Hume without defend-
ing his interpretation, it increases the interest in them if Hume’s interpretation is
correct. And I think it is.
The alternate reading sees “thinks of x” as ambiguous between “has an idea
of x present to mind” and “has an idea of x present to mind and attends to x.”
This ambiguity adds a way to mentally separate objects of thought in addition
to separating ideas in the mind. The motivation for the alternate reading is that
sometimes the same idea can be both an idea of x and an idea of y. For instance,
the same idea might be both an idea of triangularity and trilaterality. One might
begin by having the idea present to mind and attending to both x and y, and then
continue to attend to x while withdrawing attention from y. On the alternate
reading, this would be the other way to mentally separate objects. There is no
removal of any idea, there is just restriction of attention to less than everything
represented by an idea that remains present to mind all along. That is, there is a
shift from full consideration to partial consideration of the object of thought.
In my view, this alternate interpretation is motivated more by charity than by
the text. It is an attempt to save Locke from Berkeley’s and Hume’s criticisms by
undercutting the interpretation that the criticisms are based on. However, there
are textual reasons to refuse the charity.
First, Locke explicitly distinguishes partial consideration from separation.
In his discussion of space he says that the parts of space cannot be separated
nor even mentally separated. He admits that one part can be considered with-
out attending to the rest, then baldly states, “But a partial consideration is not
separating” (Essay 2.13.13). Since he distinguishes these and uses the language
of separation in explaining the formation of abstract ideas, it is unlikely that
there he means mere partial consideration. This point is not conclusive, how-
ever, since the kind of separation under discussion at 2.13.13 is different from
that involved in abstraction.
Secondly, in the 3.3.9 passage quoted above, Locke uses the phrase “abstract
and partial Ideas.” A partial idea is a part of an idea, not an idea used in a partial
consideration. That would be a complete idea along with restricted attention.

3. Ayers 1975, xx, and 1986, 12–​13; Mackie 1976, 107–​12; Taylor 1978, 97–​115; Urmson 1982,
26–​27; Winkler 1989, 39–​41.
Hume on Abstraction and Identity  •  289

Thirdly, Locke makes clear that the separation in abstraction is the removal of
distinct ideas in the following passage.

The Acts of the Mind wherein it exerts its Power over its simple Ideas are
chiefly these three, 1. Combining several simple Ideas into one compound
one, and thus all Complex Ideas are made. 2. The 2d. is bringing two Ideas,
whether simple or complex, together; and setting them by one another, so
as to take a view of them at once, without uniting them into one; by which
way it gets all its Ideas of Relations. 3. The 3d. is separating them from
all other Ideas that accompany them in their real existence; this is called
Abstraction:  And thus all its General Ideas are made. This shews Man’s
Power and its way of Operation to be muchwhat the same in the Material
and Intellectual World. For the Materials in both being such as he has no
power over, either to make or destroy, all that Man can do is either to unite
them together, or to set them by one another, or wholly separate them.
(Essay 2.12.1)

In this passage Locke speaks of three operations on two distinct ideas: (1) uniting


them into a compound one, (2)  setting them beside each other without unit-
ing them, (3) wholly separating them. In the last sentence of the paragraph the
“them” clearly refers to the same distinct ideas each time. The relevant separation
is the removal of one idea from a distinct one, not the mere shifting of attention
while keeping the same idea present to mind.
Thus, for Locke, the relevant mental separation, of one object of thought from
the other in abstraction is ceasing to think of one object of thought while con-
tinuing to think just of the other. That is accomplished by removing from the
mind the idea of one while retaining in mind only the idea of the other.
I have said that, in abstraction, the relevant objects of thought are inseparable
from each other in reality. Locke does not make this inseparability explicit. The
above passage from Essay 2.2.1 is characteristic, where he describes abstraction
as mentally removing qualities from ones that do accompany them in their real
existence, without saying that they must accompany them.4 In another context
he says,

Though the Qualities that affect our Senses, are, in the things themselves,
so united and blended, that there is no separation, no distance between

4. Locke uses the word “Ideas,” but he has cautioned us by saying, “which Ideas, if I speak of
sometimes, as in the things themselves, I would be understood to mean those Qualities in the
Objects which produce them in us” (Essay 2.8.8).
2 9 0   •  The Problem of Univer sals in Early Modern philosophy

them; yet ’tis plain, the Ideas they produce in the Mind, enter by the
Senses simple and unmixed. (Essay 2.2.1)

Again, Locke only says that such qualities are not separate, not that they are
inseparable.
Nonetheless, Hume and Berkeley are right to assume that for Locke the rel-
evant qualities are inseparable in reality. To avoid some complication, I will con-
fine attention to primary qualities.5
The basis for Hume’s and Berkeley’s assumption is Locke’s saying “All Things,
that exist, being Particulars …” and “all things that exist are only particulars”
(Essay 3.3.1 and 3.3.6). Even here, Locke does not say that necessarily all things
that exist are particulars, but he is likely committed to that. The traditional basis
for this conclusion, going back to Aristotle and Plato’s Parmenides, was that it was
impossible for one and the same thing as a whole to be common to many items
at one time. Boethius gives a succinct presentation of the arguments that it is not
possible for universals to be common to distinct particulars, and he concludes
that they do not exist but are grasped only by thought.6 Since Locke’s view of
abstraction echoes that presented by Boethius, it is safe to assume that Locke is
simply following the ancient tradition. From here, the argument to inseparability
would presumably go as follows.

(1) If the features of an object that make it a particular were separable from it,
then the object could exist without any such feature.
(2) An object existing without any of the features that make it a particular
would not be a particular.
(3) Only particulars can exist.
(4) So, the features of an object that make it a particular are inseparable from it.

One might object that the first premise is false. After all, it is possible that a given
child be separated from his current precise height by growing, even while it is
impossible that he exist without some height or other. However, such cases as a
child’s growth are not relevant to the kind of separation at issue. The separation
in thought at issue in abstraction is the ceasing to think of the qualities that make
something a particular while continuing to think only of the other qualities that
do not. The corresponding separation in reality would then be the ceasing to exist

5. It is interesting to consider abstract ideas of secondary qualities or of the features of acts of an
immaterial mind, for instance, but it is simpler not to worry about those here.
6. Boethius, From His Second Commentary on Porphyry’s Isagoge, in Spade 1994, 21–​22.
Hume on Abstraction and Identity  •  291

of the qualities that make something a particular while only those continue to
exist that do not make it a particular. The so-​called separation envisioned in the
child-​growing objection is not separation in the relevant sense.
Thus Hume and Berkeley are correct to construe abstraction in Locke’s sense
as separating in the mind what are inseparable in reality.

13.2. Inseparability and Identity


Hume takes as a principle that things inseparable in reality are identical. Given
this principle, abstraction is mentally separating something from itself. Let me
call the principle at issue the Real Separability Principle. Strictly speaking, it is
not a principle for Hume, since he derives it from two others: what I will call the
Mental Separability Principle and the Conceivability Principle.
Hume states the Mental Separability Principle early on:

First, We have observ’d that whatever objects are different are distinguish-
able, and that whatever objects are distinguishable are separable by the
thought and imagination. And we may here add, that these propositions
are equally true in the inverse, and that whatever objects are separable are
also distinguishable, and that whatever objects are distinguishable are also
different. (T 1.1.7.3, 18).

I assume from context that by ‘in the inverse’ Hume means what we mean by
‘conversely’.7
When Hume says ‘different’ he means numerically distinct. For instance, in
discussing time, he argues that the impression of time that one has in hearing
five musical notes is not “different” from the impressions of the notes themselves,
which is to say it is “not a sixth impression” (T 1.2.3.10, 36). Were it different, it
would be a sixth impression, that is, numerically distinct. Likewise at 1.3.1.1, 69
he explicitly uses the phrase ‘numerically different’.
When Hume says that things are “distinguishable,” he means that we can
think of them as numerically distinct. As he argues in his passage on distinctions
of reason, where we cannot think of things as numerically distinct, they are “in

7. Garrett (1997, 58) calls this simply the Separability Principle and says it and its converse are
new with Hume. See also Bricke 1980, 68. However, see Descartes’s discussion of distinctions
in Principles, Part I, Principles 60–​62 (Descartes 1988, 180–​182). Suarez (1947, 40) finds the
root of such principles in Aristotle. See Laird against the mental separability principle (1931,
82–​83).
2 9 2   •  The Problem of Univer sals in Early Modern philosophy

effect the same and undistinguishable.” Any apparent distinguishing of “them” is


really a distinguishing of things related to “them” (T 1.1.7.18, 25).
This move from what is true to what we can think calls for some explanation.
Hume must mean that for any different things we can in principle distinguish
them. The fact that we might be unacquainted with them, or the fact that we
might not be practiced in telling them apart cannot be serious counterexamples
to Hume’s principle. However, might it not be possible that there be distinct
things that we in principle cannot be acquainted with, or in principle could never
tell apart? I suspect that Hume’s answer would be that such things could not be
objects of thought.

Now since nothing is ever present to the mind but perceptions, and
since all ideas are deriv’d from something antecedently present to the
mind; it follows, that ’tis impossible for us so much as to conceive or
form an idea of any thing specifically different from ideas and impres-
sions. (T 1.2.6.8, 67)

Things that we in principle could not think of, or in principle could not tell
apart would be too different from perceptions to be something we could think
of at all.
Hume’s appeal to the possibility of a “relative idea” does not rebut this claim.
Relative ideas do not allow us to think of objects specifically different from per-
ceptions. The notion of such an object is absurd. “For as to the notion of external
existence, when taken for something specifically different from our perceptions,
we have already shown its absurdity” (T 1.4.2.2, 188). Relative ideas only allow
us to approach somewhat toward thinking of such an object. “The farthest we
can go towards a conception of external objects, when suppos’d specifically differ-
ent from our perceptions, is to form a relative idea of them, without pretending
to comprehend the related objects” (T 1.2.6.9, 68). However, “’tis impossible to
found a relation but on some common quality” (T 1.4.5.11, 236). So even the
unknown relative must be thought to have some resemblance to some percep-
tion even to be thought of as a relative. Hume does allow that “we may suppose,
but never can conceive a specific difference betwixt an object and impression”
(T 1.4.5.20, 241), but this does not count as thinking of such a thing. We may
use the phrase “object specifically different from perceptions” in our supposition,
but without an idea the phrase has no meaning (see 1.3.14.14, 162; 1.4.3.10, 224;
1.4.5.6, 234).
So in the Mental Separability Principle, there is an implicit restriction to
possible objects of thought. That hardly seems a restriction, however. It is like
Hume on Abstraction and Identity  •  293

pointing out that in a universal generalization, there is an implicit restriction to


objects that we can quantify over.8
When Hume says objects are “separable by the thought and imagination” he
means that it is possible to cease thinking of one while continuing to think of
the other. Presumably we would do this by ceasing to have an idea of the one
while continuing to have an idea of the other. When things are inseparable by the
thought and imagination it is not possible to think of one while ceasing to think
of the other. For instance, “A person, who desires us to consider the figure of a
globe of white marble without thinking on its colour, desires an impossibility”
(T 1.1.7.18, 25).
Hume takes having an idea of something to be the same as conceiving it to
exist (T 1.2.6.4, 66–​67). Therefore, for Hume, to mentally separate two things
is to conceive of one continuing to exist on its own in the absence of the other.
In sum, Hume assumes the Mental Separability Principle that objects are
numerically distinct only if we can think of them as distinct, and we can think of
them as distinct only if we can continue thinking of one while ceasing to think of the
other (in other words, only if we can conceive of one continuing to exist on its own in
the absence of the other).
I have suggested that we mentally separate the objects of thought by separat-
ing during the course of thinking the vehicles of thought—​the ideas. For exam-
ple, we might mentally separate a cause and an effect by separating during the
course of thinking the idea of the cause from the idea of the effect. Assuming that
this is correct, Hume must be presupposing another proposition about separabil-
ity, one undergirding the Mental Separability Principle: if ideas are distinct then
they are separable in the course of thought. Call it the Idea Separability Principle.
As a shortcut in reasoning, he will sometimes use the Idea Separability Principle
interchangeably with the Mental Separability Principle, but strictly speaking they
should be distinguished.
The other main principle is the Conceivability Principle. Hume says that
“nothing of which we can form a clear and distinct idea is absurd and impos-
sible” (T 1.1.7.6, 19–​20), and “’Tis an establish’d maxim in metaphysics, that
whatever the mind clearly conceives includes the idea of possible existence, or in
other words, that nothing we imagine is absolutely impossible” (T 1.2.2.8, 32),
“Whatever is clearly conceiv’d may exist; and whatever is clearly conceiv’d, after
any manner, may exist after the same manner” (T 1.4.5.5, 233), and “whatever
we conceive is possible” (T 1.4.5.10, 236). Here, Hume uses a principle tracing

8. This appeal to possible objects of thought answers Laird’s challenge to Hume’s contention
that “all differents are distinguishable” (1931, 82–​83).
2 9 4   •  The Problem of Univer sals in Early Modern philosophy

back to Bishop Tempier’s condemnations in 1277, that the clearly conceivable is


possible.9
Using the two main principles, Hume reasons as follows:

(1) If things are distinct, then we can think of them as distinct, so can clearly
conceive of one continuing to exist without the other.
(2) The clearly conceivable is possible.
(3) So the one can continue to exist without the other.

In other words, he uses the Conceivability Principle to infer from the Mental
Separability Principle a result about real separability—​viz., that any numerically
distinct things are really separable. I call this result the Real Separability Principle
(despite its being derived).10
These principles raise a problem for traditional abstraction.

(1) Whatever are numerically distinct are separable in reality (by the Real
Separability Principle).
(2) Whatever are inseparable in reality are numerically identical (by

contraposition).
(3) Abstraction is mentally separating what are inseparable in reality.
(4) Thus abstraction is mentally separating something from itself.
(5) Things that are mentally separable can be conceived to be distinct (by
Hume’s converse of the Mental Separability Principle).
(6) So if abstraction is possible, then something can be conceived to be distinct
from itself (by [4]‌and [5])
(7) If something is impossible, it is inconceivable (by contraposition on the
Conceivability Principle).
(8) It is impossible for something to be distinct from itself.
(9) So nothing can be conceived to be distinct from itself.
(10) So abstraction is not possible.

Hume certainly held (8). It is a contradiction that something be distinct from


itself. Were something distinct from itself, it would have to be both one single
thing and yet two distinct things, which Hume explicitly says is impossible
(T 1.4.2.28, 200).

9. See Bosley and Tweedale 1997, xx–​xxi, 440–​41, editors’ introductions. See also Grant 1982.
10. Garrett 1997 does not distinguish the mental separability principle from the real separabil-
ity principle. I think he is concerned with both. By the way, despite the use of the words ‘real’
and ‘really’, I am on Hume’s behalf still talking about the world of appearance.
Hume on Abstraction and Identity  •  295

Hume did not give the precise argument that I’ve given. He gives three related
ones. His general example of abstraction is of forming an idea of a quantity or
quality without forming a precise notion of its degree: for example, forming an
idea of a line without forming an idea of its precise length. His first argument
is that “it is evident at first sight” that the quantity or quality is identical (“not
different nor distinguishable”) with the degree of it. Therefore, they are not men-
tally separable, by the converse of the Mental Separability Principle. Hume’s sec-
ond argument is that no impression can be of a quality lacking a precise degree,
all ideas are copied from impressions, so no idea can be of such a quality. Hume’s
third argument is that it is impossible in reality for a quality or quantity to exist
without having a particular degree, so by modus tollens on the Conceivability
Principle, it is impossible to conceive the one without the other. So it is impos-
sible to mentally separate them (T 1.1.7.3–​6, 18–​20).
The argument I have given borrows the principles and approaches of Hume’s
arguments, plus the traditional assumption about abstraction, to draw the con-
sequence that abstraction would entail conceiving of something as distinct from
itself. This result, however, raises a problem for his theory of identity.

13.3. Identity and Hume’s Difficulty


Given Hume’s principles and the traditional view of abstraction, abstraction
entails thinking of something as distinct from itself. As I  will argue, Hume’s
account of identity entails the same thing.
I’ve been talking so far as if by 'identity' Hume meant the same thing as we
do—​numerical identity. Officially, however, by 'identity' Hume means numeri-
cal identity through time (T 1.1.5.4, 14; 1.4.2.29, 201; see also 1.4.3.4, 220).11 For
identity at a time he apparently uses ‘simple’ or ‘inseparable’ as well as ‘same’ (T
1.1.1.2, 2; 1.1.7.7 n. 5 App, 637; 1.1.7.18, 25; 1.4.6.22, 263). Nonetheless, I will con-
tinue to use the word ‘identity’ as we do. As I will show, Hume’s account of the
idea of identity through time is designed to solve a problem about identity in our
more general sense. So I will often talk as if he is giving an account of the idea
of identity in our sense. Furthermore, the feature of his account that comes into
conflict with his critique of abstraction is best expressed using our current sense
of ‘identity’.
Hume begins his account with a problem I’ve called Hume’s Difficulty
Concerning Identity. The difficulty begins with the fact that sometimes we don’t

11. Nonetheless, as Annemarie Butler pointed out to me, Hume says that “difference of num-
ber” is “oppos’d” to identity, which suggests that he there means numerical identity in general
(T 1.1.5.10, 15).
2 9 6   •  The Problem of Univer sals in Early Modern philosophy

know whether or not two things are identical. When I say “two things” here, I am
speaking colloquially and not prejudicing the case in favor of them being distinct.
When we don’t know whether or not two things are identical, we can alternately
imagine them to be two distinct things and imagine them to be one and the same
thing. Hume thinks that it is essential to the concept of identity that it include
the possibility of imagining both ways. Whether two things are identical is not
something we can know a priori; it is not a relation of ideas, in the terminology
of the first Enquiry (EHU 4.18). For things knowable only a posteriori—​matters
of fact—​both alternatives are imaginable. Hume makes this clear about identity.
In a case of identity, the alternative is conceivable.12 “Two objects, tho’ perfectly
resembling each other, and even appearing in the same place at different times,
may be numerically different” (T 1.3.1.1, 69). Hume is here using “two objects”
colloquially as well. It is clear from context that he is not stating a necessary truth.
Rather he is saying that two objects which are in fact identical, may for all we
know be numerically distinct.
It seems obvious that something appearing at different times can sometimes
be thought of as either distinct or identical. If I show you a coin, conceal it, then
show it to you again, you can’t be sure whether I have shown you the same coin
or two exactly resembling coins in succession. What is not so obvious is that the
coin can be thought of either way even when there is no interruption in view,
when it seems manifest that it is the same thing. For consider that you can place
a coin in front of a very young child and after a bit quickly grab it away, and the
child will not see your motion. To him it will appear as if the coin just disap-
peared. Likewise, you could quickly substitute another in its place without detec-
tion. We can imagine that a sleight-​of-​hand artist could repeat the experiment
with us adults. So even when we are watching without apparent interruption, we
can nonetheless alternately imagine something at one time and the same thing
at another time to be one and the same thing or two distinct things. We will
believe that there is a single coin, but can still suspend the belief and imagine the
alternatives.
Hume takes it to be essential to the concept of identity that it allow for these
alternate imaginable possibilities. An analysis of the concept of identity that does
not explain the possibility of thinking of the identical things alternately as two
distinct things and as one and the same thing, is not a full analysis of identity,
he thinks. This requirement on the concept of identity is what he explores at
the beginning of his extremely compressed discussion of the origin of the idea of
identity.

12. For the interchangeability of imagining and conceiving, see T 1.2.2.8, 32.


Hume on Abstraction and Identity  •  297

One might protest that Hume’s requirement on the concept of identity is ask-
ing for more than is necessary. All he needs to explain is why identity claims are
not knowable a priori. For that, he needs to explain is how it is possible that we
can fail to know that something is one single thing. And, reminiscent of Frege
(1980), the explanation is merely that we fail to know that something is one thing
when we have distinct representations of it and do not know if they are represen-
tations of one thing or two things.
To the contrary, there is more to explain. That more is how, for the things
taken to be identical, we might have thought of them as distinct—​them, without
thought of their representations. For example, suppose you meet a new colleague
at a meeting and another day meet a new neighbor, and suppose, like me, you
are someone who has trouble recognizing people in different contexts. You may
wonder whether or not the person who is the new colleague and the person who
is the new neighbor are the same person. You can imagine that they are and can
imagine that they are not. You are thinking of them, and can imagine that they
are two distinct persons. You are thinking of them even if they are in fact one and
the same person. You are certainly not thinking of your representations of them
and wondering whether they co-​refer.
The sort of representational item, whether impression or idea or description
or sense or whatever, does not matter. That we can imagine two such represen-
tational items alternately picking out one thing or two things does not help in
explaining the alternate imaginable possibilities essential to the concept of iden-
tity. We are able alternately to imagine that two things are one and the same
thing, on the one hand, or two distinct things, on the other hand. How we can
do this is Hume’s concern.
Hume’s talk of needing distinct ideas of the objects involved in order to have
a meaningful identity proposition should not mislead us (T 1.4.2.26, 200). The
point of seeming to need two ideas is to be able to think of the identical things
as two distinct things. We think about the identical things via ideas; we do not
think about the ideas and their relations to the identical things.
Before turning to Hume’s text, let me note that where I talk about his analysis
of the concept of identity, Hume talks about the origin of the idea of identity. For
our purposes these are interchangeable. Hume’s Copy Principle links the parts
and structure of an idea to its origin. Ideas are what Hume thinks concepts are. So
in explaining the origin of ideas he is giving the parts and structure of concepts.
Hume begins by saying that the idea of identity is not copied from the impres-
sion given by a single object. There must be more to the idea than that. It is not
possible to think of a single object, recognized as such, as two distinct things.
That would be to think of it as distinct from itself. So the idea copied from such
an impression does not allow for the alternate possibilities essential to the concept
2 9 8   •  The Problem of Univer sals in Early Modern philosophy

of identity. “One single object conveys the idea of unity, not that of identity”
(T 1.4.2.26, 200).
For example, suppose you have a single impression of a man. According to
Hume, to think that he is identical with himself, you would have to be able to
imagine that the man before you is distinct from the very man before you. But
that would be to imagine that someone you are seeing is distinct from himself.
Everything will depend on understanding this point that to think of a single
object, recognized as such, as two distinct things is to think of it as distinct from
itself. Two natural ways to attempt to evade it both involve equivocation.
First attempt: It would seem possible to think of a single thing with two parts
as two distinct things. But that would be to think of it as composed of distinct
things that it is distinct from. What is needed for the idea of identity is the pos-
sibility of thinking of it as being two distinct things. The first attempt involves
equivocating between identity and the composition relation.
One might persist by saying that a single thing with two parts is numerically
identical with the two parts collectively; that’s what composition is.13 So that
allows for the possibility of thinking of the whole as being two distinct things.
However, the goal is to be able to think of what are identical with each other as
distinct from each other. So the goal would be accomplished only if one could
somehow think of the whole as identical to the distinct parts individually, not
just collectively. So, the first attempt still equivocates between identity and com-
position, even given this special version of the composition relation.
Second attempt: It would seem possible to think of a single thing as two dis-
tinct things by unknowingly representing it with the two different ideas copied
from two different impressions. However, the possibility of representing a single
thing with two distinct ideas is not yet to think of it as two distinct things. It
is at best a failure to know that it is a single thing. This attempt to evade the
point I am making depends on an equivocation between ideas and what they are
of. Furthermore, it overlooks that the concern is with a single thing, recognized
as such.
So Hume says that the idea of identity is not copied from the impression given
by a single object. It is not possible to think of a single object, recognized as such,
as two distinct things. That would be to think of something as distinct from itself.
Likewise, Hume says that the idea of identity is not copied from the impres-
sions given by a multiplicity of objects. Suppose there are two objects. It is not
possible to think of two objects, recognized as such, as one and the same thing.
If we can imagine, for two things recognized as such, that they are one and the

13. See Baxter 1988b, as well as Yi 1999, and Sider 2007.


Hume on Abstraction and Identity  •  299

same thing, then we can imagine the situation in the reverse. We can imagine,
for one thing recognized as such, that it is two distinct things, and so distinct
from itself. However, as we have seen, we cannot imagine something being dis-
tinct from itself. When we perceive a multiplicity recognized as such, we cannot
think of those very objects we are perceiving as anything but many. “The mind
always pronounces the one not to be the other, and considers them as forming
two, three, or any determinate number of objects, whose existences are entirely
distinct and independent” (T 1.4.2.27, 200). So multiple objects would convey
the idea of number, not that of identity.
But what is the alternative to copying the idea of identity from an impression
of a single thing, recognized as such, and from impressions of more than one
thing recognized as such? “Betwixt unity and number there can be no medium;
no more than betwixt existence and non-​existence” (T 1.4.2.28, 200). One is per-
ceiving there to be either one single thing or more than one.
My saying “recognized as such” adds a wrinkle that at first glance does not
appear in Hume’s text. But it is clear from the text that Hume is not so much
concerned with how many things are actually being perceived as with how
many things one perceives there as being. This distinction between what there
really is that is being represented versus what one is representing there as being
is important for understanding Hume’s account of identity. If what are really
many things are, through some confusion, perceived as being a single thing,
then that impression cannot be the source of the idea of identity, for the reasons
Hume has given concerning the perception of a single thing. Likewise, if what
is really a single thing is, through some trick, perceived as being many things,
then those impressions likewise cannot be the source of the idea of identity. So
when I speak of an impression of a single thing recognized as such, I am con-
cerned only with the fact that the impression represents there as being a single
thing. Likewise, impressions of many things recognized as such represent there
as being many things.
Recognizing the wrinkle makes Hume’s question: What could be in between
representing there as being one single thing and representing there as being many
distinct things? Can we represent there as being one and a half things—​one
and a half apples, say? But half an apple is still a single thing. So, one and a half
apples are two distinct things. So we would be representing there as being two
distinct things. Alternatively, to find a medium, could we say that there is one
thing that exists and a second thing that only half exists? But what would this
mean? Something that half exists is something, and therefore it exists. So, again,
we would be representing there as being two distinct things.
The upshot is Hume’s Difficulty Concerning Identity:  How can we represent
there as being things that are perhaps identical and perhaps distinct? If we represent
3 0 0   •  The Problem of Univer sals in Early Modern philosophy

some things as being one, then we cannot alternatively represent it as being many
because we can’t alternatively represent it as being distinct from itself. If we rep-
resent some things as being many, then we cannot alternatively represent them
as being one single thing. If we could, then we could represent it as being them,
which leads to the same absurdity.
Take an example. Suppose we meet Cicero one day and Tully another and later
wonder whether they are the same person. Having met them, we are equipped to
wonder about them. Current philosophers may segue into wondering about their
names, and whether or not these names are co-​referential, but that is not what is
going on here. We are wondering about Cicero and Tully, not about their names.
Suppose we first imagine that Cicero and Tully are the same person. Then we
must be able to imagine as well that this person we imagine there to be is the same
Cicero as Cicero and the same Tully as Tully, even when we imagine Cicero and
Tully to be distinct. For we are imagining, of Cicero and Tully themselves, alter-
nately that they be identical and that they be distinct. So, it is the same Cicero
and the same Tully in either imagining. So, again, we must be able to imagine
of the Cicero/​Tully of the imagining that there is one person, that it is the same
Cicero and the same Tully as the distinct Cicero and Tully of the other imagin-
ing. So we must be able to imagine there being someone of whom we can imagine
that he is distinct from himself (Figure 13.1).
The idea of identity apparently needs to be a medium betwixt the idea of unity
and the idea of number that is somehow alternately both. But such a medium
seems impossible. Further, there seem to be no impressions from which such an
idea could be copied.

Same Cicero

Distinct men

Same Tully

Imagining them to be Imagining them to be


one single thing two distinct things
Figure  13.1  Diagram of the alternate imaginings.
Hume on Abstraction and Identity  •  301

Hume’s answer to this difficulty is that the idea of identity is cobbled together
by means of an irresistible fiction (where “fiction” entails “falsehood”).14 Here is
where Hume brings in time, so that in answer to a problem concerning identity in
general, he proposes an account of identity through time in particular.
The fiction that Hume appeals to is the fiction of a steadfast object with
duration. Hume holds that the idea of duration is copied from the impressions
received when perceiving a succession recognized as such. Consequently the
idea of duration is the idea of successiveness. He holds further that no idea can
be applied without fiction to something it could not be derived from. He holds
lastly that sometimes we perceive steadfast objects, where steadfast objects are
nonsuccessions that coexist with successions (T 1.2.3.6–​11, 35–​37). It follows
that applying the idea of duration to a steadfast object is a fiction. After all, it is
false that a nonsuccession has successiveness. Nonetheless, Hume argues that we
naturally come to apply the idea of duration to steadfast objects in addition to
successions. One first gets this fiction by having a succession of ideas that are in
fact of the same steadfast object, though one does not pay attention to that fact.
Such application becomes a habit we cannot resist (T 1.2.5.28–​29, 64–​65).

I have already observ’d, that time, in a strict sense, implies succession, and
that when we apply its idea to any unchangeable object, ’tis only by a fic-
tion of the imagination, by which the unchangeable object is suppos’d to
participate of the changes of the co-​existent objects, and in particular of
that of our perceptions. This fiction of the imagination almost universally
takes place; and ’tis by means of it, that a single object, plac’d before us,
and survey’d for any time without our discovering in it any interruption
or variation, is able to give us a notion of identity. (T 1.4.2.29, 200–​01)

One day in our development, we contemplate a steadfast object and are struck
by its nonsuccessiveness. We never paid attention to that before. And yet, despite

14. Hume’s concept of fiction is ambiguous. There are two senses, identified by Saul Traiger, in
which a fiction is not a falsehood. In his first sense a fiction is an idea of a characteristic that is
applied to an object from which the idea cannot be derived. The idea cannot be derived because
the object does not have the characteristic. For instance, and paradigmatically, when an idea
of duration is applied to a steadfast object, the idea of duration is a fiction. In Traiger’s second
sense a fiction is a feigning, a regular “process of the imagination” by which something that
lacks a characteristic is treated as if it has it. In both cases there is an opinion about something
lacking a characteristic that it has that characteristic. This third sense of "fiction" does entail
falsehood, and this is the sense I use here. Hume uses it when he characterizes the “fiction of
a continu’d existence” as “really false,” just like the “false opinion that any of our objects, or
perceptions, are identically the same after an interruption” (T 1.4.2.43, 209). See Traiger 1987,
385–​86, and Traiger 2010, 52.
3 0 2   •  The Problem of Univer sals in Early Modern philosophy

this striking experience, we cannot resist the habit of regarding everything that is
not minimally brief as having duration. So we are faced with a palpable nonsuc-
cession that we, nonetheless, cannot resist regarding as a succession. The mind is
unable to give up either of these incompatible views of the steadfast object, so the
mind alternates between them. The alternation hides this incompatibility from
the mind. It is in this alternation that the idea of identity comes into being.

Here then is an idea, which is a medium betwixt unity and number; or


more properly speaking, is either of them, according to the view, in which
we take it: And this idea we call that of identity. (T 1.4.2.29, 201)

A succession is many things in succession. A steadfast object is a single thing that


coexists with successions but is not a succession itself. So, to regard a steadfast
object alternately as a succession and as a nonsuccession is to regard it alternately
as many distinct things and as a single thing. This is Hume’s closest approach to
representing there as being a medium betwixt unity and number.
So the idea of identity is the result of alternating between an accurate idea of
the steadfast object as a single thing and a fictitious idea of the steadfast object as
many distinct things.
Unfortunately, Hume’s idea of identity is at best a way of evading Hume’s
Difficulty. It certainly is not a solution and, it turns out, is not even a successful eva-
sion. Even a successful evasion in this way would involve the very problem entailed
by his critique of abstraction: thinking of something as distinct from itself. The
idea of identity was supposed to be a way around this problem, but it isn’t.
In order to explain in detail why Hume’s attempt to simulate an idea of a
medium betwixt unity and number cannot work, let me return to a distinction
made earlier, that between what there is that an idea represents and what an idea
represents there as being. For example, suppose in the dark you mistake a mailbox
for a boy scout. What your idea is of is a mailbox. What your idea represents there
as being is a boy scout.
The fiction included in Hume’s idea of identity is that the steadfast object
has duration. We think of it via a succession of ideas that represent it as being
many distinct things. Thus, what there is that the succession of ideas represents
is in fact a single steadfast object. What the succession of ideas represents there
as being, however, is something with duration, that is, a succession. Because they
are successive, they represent what they are of as being a succession.15 Next, we

15. There is a later fiction that occurs for Hume after one has the idea of identity. One takes
successive objects in close relation to be the same thing (see T 1.4.2.34–​35, 203–​04; 1.4.6.6,
Hume on Abstraction and Identity  •  303

think of the steadfast object as one single thing via a single idea copied from an
uninterrupted contemplation of the object. Because it is a single idea, the idea
represents the steadfast object as being a single thing. To have the idea of identity
is to alternate between these views. Hume intends that this makeshift idea will
simulate a solution to Hume’s Difficulty, since there is no genuine solution. How
do we represent there as being things perhaps identical and perhaps distinct? We
alternate between an idea representing it as a single thing and some successive
ideas representing it as being many distinct things.
However, the makeshift idea cannot even simulate a solution. To alternate
between a single idea and a succession of ideas is just to have a longer succession
of more ideas. Being successive, they will represent what they are of as being suc-
cessive. So, the longer succession would still only represent the steadfast object as
being a succession. So, even if the mind is in fact alternating between views of the
object as one and the object as many, there is no way on Hume’s account for the
mind to recognize that it is doing so. For the makeshift idea even to simulate a
solution to Hume’s Difficulty something extra must be added that Hume cannot
account for. Somehow the viewer must recognize that it is the same object in the
alternating views. Its merely being the same object is not enough.
In other words, Hume’s account cannot distinguish (1)  representing some-
thing simply as many things in succession, some of which are briefer and some
of which are longer, from (2) representing it alternately as one thing and as many
things in succession. His account could only supply the same sequence of ideas
for each of these crucially different cases. What it cannot supply is a recognition
of the sameness of what the alternate views are of, as the second case requires.
Let me give an illustration. Suppose a drunk’s eyes were crossing then uncross-
ing as the same bug flies around his head. He alternately sees one bug then a
doubled bug, even though it is in fact the same bug each time. For all the drunk
knows, he is being plagued by a succession of bugs, some of which fly in pairs.
To think that it is the same bug viewed in different ways, he has to think that
each bug in the apparent pairs of bugs is the same bug as the apparent single bug.
Hume’s account gives no way to think of this sameness. If it did, it would solve
Hume’s Difficulty.
Here, then, is the problem with Hume’s account. For Hume’s makeshift idea
of identity even to simulate a solution to Hume’s Difficulty, it would have to be
a genuine solution. The idea would have to represent there as being the same
thing whether viewed as one single thing or viewed as many distinct things. The
idea would have to represent a single thing as a single thing and then switch to

253–​55). One, as it were, “runs together” these distinct objects. However, this subsequent
more famous fiction is not part of Hume’s discussion of first forming the idea of identity.
3 0 4   •  The Problem of Univer sals in Early Modern philosophy

representing it as the same “it” and yet as distinct things. But as we have seen, it
could do this only if it could represent there as being something distinct from
itself.16
Thus, in his account of identity Hume faces the same problem that he raises
for the proponents of traditional abstraction. The problem of abstract ideas is a
version of Hume’s Difficulty Concerning Identity. In order to abstract we must
be able to think of the same thing alternately as one and the same thing and as
two distinct things. So we must be able to think of something as distinct from
itself. If we cannot conceive this, Hume’s account of the idea of identity falls. If
we can, then his critique of traditional abstraction falls. For Hume, traditional
abstraction and his account of identity thus stand or fall together.17

Abbreviations
EHU = Hume 2000; cited by section and paragraph.
Essay = Locke 1975; cited by book, chapter, section.
T = Hume 2007, 1978; cited by book, part, section, paragraph, page in Hume 1978.

16.  I  myself think that Hume is exactly right about what the concept of identity requires.
The fact that Hume’s Difficulty follows means that an account of identity needs to be able to
address the difficulty. In the terms of my own peculiar account of Many-​One Identity, sense
can be made of someone’s being distinct from himself. I say that numerical identity is relative
to “counts.” And I say that “aspects” of a person can be numerically identical in one count and
numerically distinct in another. “Cross-​count identity” is secured by the “aspectival identity” of
these aspects. Thus to say that Cicero/​Tully is distinct from himself, in a sense that is not con-
tradictory, is to refer to his Cicero aspect and to his Tully aspect in the count in which they are
identical and then to predicate distinctness of them in a count in which they are numerically
distinct. In addition to Baxter 1988b, see Baxter 1988a, 1989, 1999, and 2001.
17. I’m grateful for comments from audiences at these conferences: “The Problem of Universals
in Modern Philosophy,” Scuola Normale Superiore, Pisa; “Hume’s Metaphysics and Humean
Metaphysics,” University of Tampere, Finland; and the 39th International Hume Conference,
University of Calgary, Alberta. I’m especially grateful for discussion with Stefano Di Bella,
Jani Hakkarainen, Don Garrett, Philipp Keller Blum, Roberta Ballarin, Todd Ryan, Tad
Schmaltz, Samuel Newlands, Antonia Lolordo, Tom Stoneham, Lawrence Nolan, Saul Traiger,
Annemarie Butler, Lewis Powell, Jonny Cottrell, Dario Perinetti, and Donald Ainslie. Thanks
also to Toby Napoletano for research assistance.
14 KANT AND ABSTRACTIONISM ABOUT
C O N C E P T F O R M AT I O N

Alberto Vanzo

Unlike several of the authors discussed in this volume, Kant does not
provide any extended discussion of the metaphysical status of univer-
sals, understood as the sort of items that many particulars may share
(properties and relations).1 Kant is much more interested in our
capacity to represent properties and relations as the sort of items that
many particulars share. According to Kant, we do this by employing
concepts. For instance, we represent the color of a drape as a feature
that other objects share by subsuming it under the concept red.
This chapter focuses on Kant’s abstractionist account of the for-
mation of empirical concepts such as red, an account that, along
with abstractionist theories as such, has been the object of numerous
criticisms. Sections 14.1 and 14.2 provide a reconstruction of Kant’s
account. Sections 14.3 to 14.5, which focus on color concepts, dis-
cuss two criticisms that have been advanced against not only Kant’s
account but also Locke’s account and abstractionist theories as such.
As we will see, neither of the objections is convincing as it stands. Kant
can offer replies to both objections that are consistent with his views
and with the empirical evidence concerning the perception and repre-
sentation of colors and sensory properties.

1. For a discussion of Kant’s brief remarks on this issue, see Oberst


2015. References to the Critique of Pure Reason are to the first-​and sec-
ond-​edition pagination (A and B). Otherwise, the pagination to which
I refer in Kant’s texts is from the Akademie-​ Ausgabe (Kant 1900–​ ),
except for L. Bauch, L. Hechsel and Warschauer L., which are cited from Kant 1998, and
A. Dohna, which is cited from Kowalewski 1924. The abbreviations used are
listed at the end of the chapter. Translations, where available, are from the
Cambridge Edition of the Works of Immanuel Kant.
3 0 6   •  The Problem of Univer sals in Early Modern philosophy

14.1. Background
Kant distinguishes two ways of representing properties:  by means of thoughts
and by means of nonconceptual intuitions. We can either “think” or only “intuit”
the “roundness” of a plate (A137/​B176). In Kant’s vocabulary, thoughts are con-
ceptual representations and thinking is having a mental representation informed
by concepts (Fort. 20:325; Pr. Anthr. 7:196). Intuitions, too, can be informed by
concepts. However, at least in principle, intuitions need not be conceptualized.
An intuition is “that which, as representation, can precede any act of thinking
something” (B67, see B132).
According to Kant, what differentiates thoughts from nonconceptual intu-
itions is not their “matter,” that is, the items they are about.2 Rather, a house can
be represented by both a “mere intuition” and an “intuition and [a]‌concept at
the same time.”3 What differentiates thoughts from nonconceptual intuitions is
their “form,” that is, the way in which they represent the items they are about.
Intuitions are “singular representations,”4 whereas the form of conceptual repre-
sentations is “its generality.”5 This means that only conceptual intuitions repre-
sent individual objects and their features as instantiating universals. Conceptual
and nonconceptual intuitions can represent the red color of cinnabar, a carmine
drape, or a poppy, but only conceptual intuitions represent their redness as a
feature that “can be encountered in anything” (B133n) and that is “common to
many objects” (Wiener L. 24:905). For Kant, representing properties as features
that many particulars share is representing them by means of concepts.
What is the origin of the concepts that enable us to represent properties as fea-
tures that many particulars share? Some concepts, like the concept of unmarried

2. “In every cognition, one must distinguish matter, i.e., the object, and form, i.e., the way in
which we cognize the object” (L. Pölitz 24:510 = Jäsche-​L. 9:33; see Wiener L. 24:805; Refl.
1628 [1780–​89] 16:45).
3. Jäsche-​L. 9:33. Other passages in the Kantian corpus make the same point: L. Pölitz 24:510;
Wiener L. 24:909 = L. Hechsel 397; L. Bauch 47–​48. See also the passages on nonconceptual
intuitions of the parts of objects (Entd. 8:217n; Refl. 220 [ca. 1776–​1783] 15:84; L. Bauch
46). I  leave the question undecided as to whether Kant regards these as merely theoretical
possibilities or whether he holds that, as a matter of fact, there are nonconceptual representa-
tions of objects. However things may be, Kant states that, when we conceptualize a “horse” as
a “four-​footed animal,” we represent “something that was already apprehended in the sensory
intuition,” prior to that conceptualization (Fort. 20:273–​74).
4. B136n; L. Pölitz 24:565; M. Dohna 28:651.
5.  Fort. 20:273–​74. Several passages identify another difference between intuitions and
thoughts: intuitions are immediate representations, whereas thoughts are mediate (e.g., A68/​
B93, A320/​B377). I do not take a stance on whether the singularity of intuitions and general-
ity of concepts are more basic than their immediacy or mediateness.
Kant and Abstractionism about Concept Formation  •  307

adult male, can be formed by combining other concepts, whereas others may be
formed by analyzing complex concepts.6 These processes allow us to derive new
concepts from other concepts, but they cannot account for the origins of our first
concepts. According to Kant, no concepts are preformed in our minds from the
beginning of our lives. Not even the categories, “pure a priori concepts” that “con-
tain nothing empirical” (A95), are preformed.7 They are generated by “reflection”
and “abstraction” on mental acts that we carry out “on occasion of experience”
(Refl. 409 [1772–​1779?] 15:155; M. L1 28:233–​34). Regrettably, however, Kant
never provided more than scant remarks on the formation of the categories. His
statements on the formation of empirical concepts are more explicit.
Empirical concepts are, by definition, those concepts whose representational
content depends on the stimuli that are given to our senses. Nevertheless, their
property of representing features as “common to many objects”8 is never given
through the senses, but always “made,”9 contributed by our mind through acts of
comparison, abstraction, and reflection. These acts are performed on “empirical
intuition[s]‌,” on which empirical concepts are “grounded.”10 They enable us to
represent specific features, like the carmine of a drape and the ruby of a stone, as
instances of universals like the color red.

14.2. Comparison, Reflection, and Abstraction


The Kantian texts on the formation of empirical concepts have been criticized
for being excessively concise, fragmentary,11 “cryptic and obscure,”12 so much
so that it is allegedly “problematic” to reconstruct a unitary account on their

6. Warschauer L. 613 = L. Pölitz 24:570; see Jäsche-​L. 9:99. As I argue in Vanzo 2012, 94n52,
these passages are best read as statements of Kant’s views, not just as explanations of a doctrine
found in the textbook used for his lectures. If this is correct, the phrases “every concept what-
soever” in Jäsche-​L. 9:94 and “no concept” in Wiener L. 24:909 are too strong. I expand on
other claims of this chapter in Vanzo 2012.
7. Entd. 8:221–​23; Lett. (1789) 11:82; M. K3 29:949, 951–​52.
8. Wiener L. 24:905.
9. Jäsche-​L. 9:93 = Refl. 2855 (1772–​1778?) 16:547.
10. A47/​B64. Not by chance, Kant’s lecture notes paraphrase the question “how do representa-
tions become concepts?” with “how does a concept derive from intuition?” (L. Busolt 24:654;
see Wiener L. 24:907). I leave the question open as to whether, according to Kant, we form
nonempirical concepts like the categories and mathematical concepts through comparison,
reflection, and abstraction.
11. Carpenter 1995, 227.
12. Allison 2004, 80.
3 0 8   •  The Problem of Univer sals in Early Modern philosophy

basis (Vásquez Lobeiras 1998, 141). This section surveys Kantian texts from the
Critical period (ca. 1780–​1804) and outlines the unitary account of empirical
concept formation that can in fact be found in them.13
Kant’s account applies to sortal concepts (cup, tree) as well as to character-
izing concepts (red, tall). In order to form both kinds of concepts, “I must
have distinctly cognized many individual objects and I must represent distinctly
what is common to them.”14 “I compare things and attend to that which they have
in common, and I abstract from all other things; thus this is a concept, through
which all these things can be thought.”15

I see, e.g., a spruce, a willow, and a linden. By first comparing these objects
with one another I note that they are different from one another in regard
to the trunk, the branches, the leaves, etc.; but next I reflect on that which
they have in common among themselves, trunk, branches, and leaves
themselves, and I abstract from the quantity, the figure, etc., of these; thus
I acquire a concept of tree. (Jäsche-​L. 9:94–​95)

Kant’s Critical texts provide two analyses of this process. The first set of texts
identifies three phases:

(1) comparison [die Comparation, d.  i. die Vergleichung] of representations


among one another in relation to the unity of consciousness;
(2) reflection as to how various representations can be conceived in one con-
sciousness; and finally
(3) abstraction of everything else in which the given representations differ
(Jäsche-​L. 9:94).

The kind of comparison that takes place in concept formation is carried out on
“many representations,”16 namely intuitions, as they are made the objects of a

13. Since most of these texts are Kant’s personal notes (the so-​called Reflexionen) and lecture
transcripts, they must be used with caution. On the necessary cautions and the dating of the
lecture transcripts, see Capozzi 2001, 145–​82; Conrad 1994, 43–​65; Naragon 2006.
14.  L. Bauch 44, cf. 43. The context makes clear that the distinct cognitions of individual
objects mentioned in this passage are intuitions. Along similar lines, A195–​96/​B240–​41 states
that, if the concept of cause were an empirical concept, it would derive from the perception of
many events.
15. Wiener L. 24:907. See Refl. 2854 (1773–​1779?) 16:547; Refl. 2876 (1776–​1779?) 16:555;
Wiener L. 24:907–​909 = L. Hechsel 393–​97; Jäsche-​L. 9:93–​95.
16. Jäsche-​L. 9:94; L. Pölitz 24:566.
Kant and Abstractionism about Concept Formation  •  309

single mental act, “one consciousness.”17 The Jäsche-​Logik (9:94) associates com-
parison with the search for differences among objects. Most logic texts, however,
associate it with the search for differences as well as shared properties.18 The texts
on concept acquisition use “reflection” to refer to the identification of shared
properties: “from reflection, one cognizes that which many things have in com-
mon.”19 Abstraction is the act of diverting one’s attention from the features with
respect to which compared objects differ. Kant stresses that we do not abstract
the shared properties of objects, but rather abstract from the features for which
they differ.20 Abstraction is a necessary21 but “negative” phase of concept forma-
tion because it only excludes certain features from conceptual content (L. Dohna
24:754; Wiener L. 24:907, 909; L. Busolt 24:654). Comparison and reflection
play a positive role because they enable us to identify the features that will consti-
tute the content of concepts.
The first set of texts does not make explicit claims regarding the temporal
order of comparison, reflection, and abstraction. However, at least from a logical
point of view, reflection and abstraction presuppose comparison. We can only
identify shared features and divert our attention from nonshared features if we
compare them with one another.
The second set of passages inverts the order of the three phases and gives a new
meaning to “reflection.”22 According to these texts, reflecting is the mental act of

17. Warschauer L. 610; Wiener L. 24:909; see Jäsche-​L. 9:94. Longuenesse (1998, 113–​15) iden-
tifies other meanings of “comparison” in Kant’s texts.
18.  Two Reflexionen suggest that one can acquire concepts without carrying out compari-
sons: Refl. 2876, 2878 (1769–​1789) 16:555–​57. As far as I am aware, no other texts corroborate
this suggestion. On comparison, see also EE 20:213.
19. Wiener L. 24:909 = L. Hechsel 396; see Warschauer L. 610; Jäsche-​L. 9:94. Unlike the logic
transcripts, the Anthropologie Dohna (147) associates reflection to the identification of differ-
ences as well as shared properties.
20. See, e.g., Pr. Anthr. 7:131, against Meier (1752) 1924, 16:551. Kant’s use of “abstraction”
is similar to Locke’s. It does not map on any the meanings of “abstraction” singled out by
Berkeley. See Baxter, chapter 13, section 13.1, in the present volume.
21. Warschauer L. 610; Refl. 2871 (1760–​1777?) 16:554; M. L1 28:328. According to La Rocca
(2004, 281n65), three Reflexionen deny that we form concepts through abstraction (Refl.
2851, 2865 [1769–​1775, before 1766?], 16:546, 552; Refl. 2878 [ca. 1776–​1789], 16:557–​58).
None of those passages denies that concept formation involves abstraction. Refl. 2851 employs
“Reflexion” in a broad sense, which includes abstraction besides comparison and reflection.
Refl. 2865 denies that abstraction is sufficient for concept formation, but not that it is nec-
essary. Refl. 2878 suggests that we may form concepts without comparison, but not without
abstraction.
22. Refl. 2860 (ca. 1776–​1789) 16:549; L. Pölitz 24:566; L. Busolt 24:654. Two passages men-
tion only two phases of concept formation, but they do not accurately express Kant’s views.
They are located near passages that, like most texts, identify three phases. Compare Wiener
3 1 0   •  The Problem of Univer sals in Early Modern philosophy

becoming conscious of, and paying attention to, the features of objects (L. Pölitz
24:566; L. Busolt 24:654). It precedes comparison, through which we identify
shared features, and abstraction.
The two sets of texts on empirical concept formation are compatible with one
another. They outline the same process, although they single out different phases.
They can be integrated into a unitary account by distinguishing between two
senses of “reflection” (reflection1 and reflection2) and identifying four phases of
concept formation, as a few Kantian passages do.23 Let us assume that we lack a
concept of tree and we are comparing the intuitions of a spruce, a willow, and a
linden. They may be perceptively present to us, or the vision of a tree or a leaf could
bring nonperceptual mental images of trees to our mind.24 The act of turning our
attention to their features is reflection1. As we consider them, our mind identi-
fies and records the features that they share (reflection2) and those with respect
to which they differ (comparison). Although the texts provide scant details, we
can think of this process in sequential terms, as the identification of a feature in
the first tree, followed by a search for that feature in the other trees.25 We may be
doing this by ourselves, or we may be guided by an instructor’s verbal feedback.26
Environmental feedback too may act as an instructor: “if, for example, primroses
were edible, and all other flowers toxic … , feedback from the consequences of
the sensorimotor interactions would be supervision enough” (Harnard 2005,
39). If the search is successful, the feature is recorded in a mental list of shared
features. Otherwise, it is recorded in a mental list of differences. Then we turn to
another feature and repeat the process, until we have identified a sufficient num-
ber of shared features and reflection2 stops. At that point, we divert our attention

L. 24:907, lines 22–​23 with 909, lines 19–​20, 24–​26; Warschauer L. 608, line 525 with 609,
line 571.
23. M. K2 28:740; R. Pölitz 28:1052–​53; Danz. RT 28:1269. These passages call reflection1
“attention.” For instance, according to the Logik Busolt (24:654), reflection1 is “attention to the
manifold that is thought in an intuition.” Other texts use “attention” in different ways, relating
it to comparison (e.g., Refl. 2976 [1776–​1789?] 16:555) or reflection2 (e.g., L. Pölitz 24:567).
La Rocca (2004, 281) identifies four phases in Kant’s account of empirical concept formation.
24. Wolff ’s account of concept formation, that Kant knew, mentions the comparison between
perceptual and non-​perceptual mental images (Wolff 1751, §§273, 832–​33).
25. Kant’s passages refer to features that we can detect through ocular inspection. However,
the same process can be applied to a wide array of features, including those that trees have only
under certain conditions, like losing leaves if it is autumn (Longuenesse 1998, 145); disjunctive
features; or features that we can only detect by employing instruments and certain concepts,
like the features sought by genetic taxonomists.
26. Pace Lyssy (2007, 162), the fact that concepts, for Kant, are mental entities does not rule out
that “linguistic social interaction” may be involved in their formation.
Kant and Abstractionism about Concept Formation  •  311

from the differences and focus on the shared features (abstraction), by which the
content of the concept is formed.27
What counts as a sufficient number of shared features may vary depending on
the quantity and level of detail of our intuitions, the strength of our memory, our
level of attention, and the amount of time devoted to comparison and reflection.
These factors depend in turn on other factors, such as our aims and activities. The
biological concepts that we form when we are trying to develop a new taxonomic
theory are likely to rely on more intuitions and more careful comparisons than
those that we form when we casually note a new breed of dog during leisurely
walks in the neighborhood.

14.3. Two Objections
The broad lines of Kant’s account of empirical concept formation are not deeply
original. They recall, among others, Locke’s comments on how we form abstract
ideas28 and Christian Wolff ’s explanation of how we come to have distinct uni-
versal concepts.29 However, several specific features of Kant’s account are original.
For example, the triad comparison-​reflection-​abstraction and the claim that we
must carry out all three mental operations to generate concepts from intuitions
cannot be found in Baumgarten, Berkeley, Crusius, Hobbes, Hume, Leibniz,
Locke, Meier, or Wolff. Among them, Locke holds that we acquire some ideas by
comparison, others by reflection, and others by abstraction.30 He does not state
that we must carry out all three acts to acquire any kind of ideas, although his
account of the acquisition of abstract ideas recalls Kant’s account of empirical
concept formation (Locke, Essay 2.11.9, 3.3.7). Georg Friedrich Meier, the author

27. A728/​B756. The search for shared features and differences may resume at any point in the
future if conceptual revision is required (Kitcher 1990, 210–​11). According to Kant, we do not
juxtapose the features that form the content of a concept in a “simple list,” in “no particular
order” (Frege [1884] 1988, §88), but we organize them in a structure of co-​ordinated and sub-
ordinate marks. See Hanna 2001, 125–​54.
28. Locke, Essay 2.11.9, 3.3.7.
29. See, e.g., Wolff 1751, §§273, 832, and Wolff 1738, §283 on the formation of universal con-
cepts, that is, concepts of genera and species, as opposed to concepts of individuals. Given
Wolff ’s endorsement of innatism (Wolff 1740, 508; 1751, §819), these passages may be taken
to outline the process whereby we become conscious of the content of our innate concepts of
genera and species. According to Wolff, this happens when we render a concept distinct. Wolff
calls a mental content distinct if we can distinguish it from other mental contents and we can
state what the difference between them is (Wolff 1751, §206).
30. They are respectively the ideas of relations (Locke, Essay 2.11.4), reflection (2.2.8, 24), and
abstraction (2.11.9). Other ideas are acquired passively, through sensation or perception (2.2.3,
9, 23).
3 1 2   •  The Problem of Univer sals in Early Modern philosophy

of the textbook that Kant used in his logic lectures, mentions a sequence of three
mental acts corresponding to Kant’s reflection1, comparison, and abstraction, but
only when he formulates the rules to render our cognitions clear, not when he
discusses the origin of concepts.31
Despite the relative originality of Kant’s account of empirical concept for-
mation, several objections that have been directed to his account have also been
raised against other authors, such as Locke, and against abstractionist theo-
ries as such. In what follows, I discuss two objections that, if successful, would
undermine views held not only by Kant but also by several of his early modern
predecessors. I argue that the objections are not successful as they stand.
I focus on the acquisition of color concepts for two reasons. First, Kant holds
that we acquire color concepts from empirical intuitions through comparison, reflec-
tion, and abstraction. However, color concepts are often mentioned as an example
of concepts that we cannot acquire in this way. The same has been said of other
concepts, like democracy (Prinz 2005) and the concepts of modern cosmology
(Gaukroger 1978, 107). Kant could reply that we acquire them by combining previ-
ously acquired concepts. This reply is less plausible for color concepts, because their
content has a particularly strong relation to sensory experience. If we cannot acquire
even those concepts that are straightforwardly related to visual experience through
acts of comparison, reflection, and abstraction upon empirical intuitions, there is
reason to doubt that we can acquire any other concepts in this way.
Second, although color concepts may be as straightforwardly related to sen-
sory information as auditory or tactile concepts, psychologists, cognitive scien-
tists, and cultural anthropologists have studied color concepts much more than
other sensory concepts. Kant was not aware of this research. However, I do not
claim that Kant rejected the objections in the ways that I illustrate below. I only
claim that, given Kant’s views and the empirical results presented below, the
objections are not successful as they stand.
According to Kant, although color concepts have empirical origin,32 they are
not given in experience: “the intuition of red does not yet give any concept of
the understanding” (L. Dohna 24:752). They are generated from visual intuitions
through acts of comparison, reflection, and abstraction:

He who wished to have a representation of the color red first had to see the
color red. When he compared the color red in the red of cinnabar, carmoisin

31. Meier (1752) 1997, §162; Meier (1752) 1924, §131. For Meier’s definition of clarity, see his
(1752) 1924, §124.
32. See, e.g., L. Pölitz 24:566; L. Busolt 24:654.
Kant and Abstractionism about Concept Formation  •  313

[carmine], and ponceau, however, he became aware that there is something


general in the color red, that is contained along with other things in other
representations of the color red, and he thought by red that which was com-
mon to many objects, and this was a concept.33

Other passages add that, when we focus on the red that is “common to many
objects,” we abstract from the differences between them. For instance, when I focus
on the redness of a scarlet cloth, “I abstract from the cloth” (L. Pölitz 24:567). The
color concept that I form is “thought of as common to several” intuitions of red
objects, “that in addition” to being red “also have something different in themselves”
(B133–​34n).

14.4. First Objection: Color Shades and Shared Features


Against Kant, Locke, and abstractionist theories of concept formation as such, it is
often claimed that we do not form color concepts by abstraction because “it is false
that all instances of a given color share some common features” (Carruthers 1992,
59). “[T]‌he different basic shades of red do not have anything in common, which
can be singled out in attention, and thus give rise to the more general concept ‘red’ ”
(Newman 1992, 104). “[R]edness consists in a continuous range of shades, each of
which is only just distinguishable from its neighbors. Acquiring the concept red is a
matter of learning the extent of the range” (Carruthers 1992, 59). The location of the
boundaries between ranges is “set by the ordinary meaning of the word ‘red’ ” (Ayers
1991, 1:259).34
Kant does not share this view. He holds that the boundaries of at least some
colors, like red, are not conventional because there is “something general,” a fea-
ture that is “common to” all and only their instances, although he does not state
what this feature is.35 Current studies on categorical perception provide support
for Kant’s view and furnish him with an answer as to what that feature might be.

33. Wiener L. 24:904–​05 = L. Hechsel 390. This passage implies that blind people lack color
concepts. It is unclear whether, according to Kant, those who have seen some colors can form
concepts of the colors that they have not seen. Compare Pr. Anthr. 7:167–​68 with M. L1
28:233–​34. Kant also seems undecided on whether Euler’s theory of color perception is correct
(compare KU 5:224 with 5:324) and on the role of imagination in color perception (Berger
2009, 41–​45).
34. A similar argument for the claim that we cannot form the concept color by comparison,
reflection, and abstraction has been put forward by Geach (1957, 37–​38) and endorsed by Hark
(2008, 103).
35. Wiener L. 24:904–​05 = L. Hechsel 390; see B133–​34n. Among the authors known to Kant,
Baumeister (1747, §36n) acknowledged not to know what this feature is.
3 1 4   •  The Problem of Univer sals in Early Modern philosophy

Let us imagine that we are observing a white square on a black background.


The quantity of light that reaches the square is reduced gradually and uniformly,
until we are unable to distinguish the square from the background. As the quan-
tity of light is reduced, the color of the square turns gradually from white to black.
It would be hard to tell when the square stopped being white and started being
grey, or when it turned from grey to black. Lighter and darker shades morph
gradually and continually into one another. If, however, we project red light on
the square, and we lower its wavelength gradually and uniformly until the square
becomes violet, we will witness a rather different phenomenon. The square will
first become orange. It will remain orange for some time. It will then become yel-
low, green, light blue, and dark blue. It will remain dark blue for longer than any
other color, before becoming violet. From time to time, when the square turns
from one color to another, it will briefly have an indistinct color.
The phenomenon that we witness if we look at the changing hues of a square
is the same as we can observe when we look at a rainbow or at light rays refracted
through a prism. We do not see a series of shades that morph gradually and uni-
formly into one another. We see colored stripes, each of which is rather uniform
and distinct from the adjacent colors, with narrow stripes of an indistinct color
between them.36

This is due to the fact that our perceptual system compresses certain
frequency ranges, which we see as just varying shades of the same quali-
tative color. These compressed ranges are then separated from adjacent
qualitative regions, also compressed, by small, boundary regions that look
like indefinite mixtures, which are neutral between the two adjacent cat-
egories. And just as there is compression within each color range, there is
expansion between them. Equal-​sized differences look much smaller and
are harder to detect when they are within one color category than when
they cross the boundary from one category to the other.37

We are able to discriminate stimuli belonging to different color categories (such


as a shade of red and a shade of green) more quickly and more accurately than
stimuli that are equally distant on the spectrum, but belong to the same cate-
gory (such as two shades of red). This phenomenon is the categorical perception
of color.

36. For a list of transition points from one color to another, see Knoblauch 2002, 66.
37. Harnard 2005, 26. For experiments that prove the existence of categorical perception of
color, see Uchikawa and Shinoda 1996.
Kant and Abstractionism about Concept Formation  •  315

As a result of categorical perception, the location of the boundaries between


red and adjacent colors is not purely conventional. The fact that we designate
a certain area of the spectrum with the term “red” instead of “blue” is conven-
tional, but the boundaries of each area are set by the processes of compression and
expansion that are involved in color perception. All shades of red have a common
property. This is the property of falling in a specific area of the spectrum, an area
that appears to us as homogeneous and rather well distinct from the adjacent
areas.38
Against this view, one could claim that categorical perception depends on
the possession of color concepts and color terms. The categorical perception
of shades of red cannot provide the basis for the formation of the concept red
because we perceive shades of red categorically only if we possess the concept
red. The best sources of support for this view are not anthropological studies
of color naming,39 but studies of the cerebral activity associated with color per-
ception. In adults, the categorical perception of color is associated with activity
in the left hemisphere, which encodes linguistic information (Franklin et  al.
2008). This provides strong evidence for the claim that the categorical percep-
tion of adults is influenced by their color terms, and hence—​at least for those
who, like Kant, posit a strong link between concepts and language40—​by their
concepts.
Interestingly, however, the study that proved that the categorical color per-
ception of adults is lateralized to the left hemisphere identified a second type of
categorical color perception, which is not influenced by color terms. It is found
in infants of four to six months who have not mastered verbal language. It is asso-
ciated with activity in the right hemisphere, which does not encode linguistic
information, but metric information.41 Kant can appeal to this nonlinguistic
categorical perception to explain the acquisition of color concepts from non-
conceptual intuitions. He can claim that, in virtue of nonlinguistic categorical
perception, the shades of the same color share a feature that can be perceived in
absence of color concepts.

38. I identify a second common property in Vanzo 2012, 176.


39. The majority view among cultural anthropologists is that there is a “universal pattern” of
color naming across languages and cultures (Boster 2005, 109). For evidence, see Cook, Kay,
and Regier 2005. The opposite view is defended by Davidoff, Davies, and Roberson 1999,
Roberson, Davies, and Davidoff 2000. A survey of the debate that followed these two studies
can be found in Regier and Kay 2009.
40. See, e.g., Pr. Anthr. 7:155, 192; Capozzi 1987.
41. Franklin et al. 2008; Regier and Kay 2009, 439–​42.
3 1 6   •  The Problem of Univer sals in Early Modern philosophy

14.5. Second Objection: Shared Features and


Comparisons
Besides holding that all shades of red share a feature, Kant holds that, in order to
identify it, we must compare those shades with one another. Similarly, in order
to notice that a spruce, a willow, and a linden all have a trunk, branches, and
leaves, we must compare them with one another. What is required for us to do
this? The critics of Kant,42 Locke,43 and abstractionist theories as such44 claim
that we must “recognize” the objects that we are going to compare as “associable
objects.”45 However, “(1) this kind of comparison seems to presuppose awareness
of what is presented to us as having the feature corresponding to the concept to
be made explicit, and (2) that in turn seems to presuppose a prior synthesis of the
manifold according to that concept.”46
To address this objection, Kant can deny (1). Kant would agree that, in order
to represent a feature as shared by several objects, we employ concepts. He can
also grant that, sometimes, we compare objects with one another because we
hold that they share certain features. For instance, taxonomists may notice that
certain plants share specific genetic features, which are the basis for the con-
cept of a new taxon. However, Kant need not claim that we always compare
objects because we hold that they share certain features. We can compare them
because, without employing concepts, we have noted a similarity among them.
For instance, Kant can claim that we did not compare shades of red with one
another, so as to form the concept red, because we became aware that they all
are red. Instead, we compared them because, without employing concepts, we
detected that they are similar to one another, more so than they are to shades of
green, blue, or other colors.
Consider Leibniz’s and Kant’s example of a dog which, having been beaten
in the past, sees its owner raise a stick and cries (in Kant’s version) or runs away
(in Leibniz’s version).47 In this case, the perception of an event brings to mind
the memory of a similar event. Kant can explain this by noting that dogs, like
humans, have a natural disposition to compare perceptual mental images with

42. E.g., Allison 2001, 22; Carpenter 1995, 233; Ginsborg 2006, 39; Kalar 2006, 48.
43. Carruthers 1992, 55; Ginsborg 2006, 43.
44. E.g., Atkinson 1982, 49; N. Bolton 1977, 14–​17; Sigwart 1904, 328–​29.
45. Pippin 1982, 113.
46. Ginsborg 2006, 41, numbers added; see Cassirer 1923, 16–​17.
47. Leibniz (1714) 1890, §5; Refl. 377 (1753–​1756? 1762–​1763?) 15:151.
Kant and Abstractionism about Concept Formation  •  317

nonperceptual, memorized mental images.48 When a similarity is detected, the


representations of the relevant objects, features, or perceptual scenes are brought
to our consciousness.
In order to detect similarities, it is necessary to represent them. Kant can claim
that dogs can detect similarities, even though they lack concepts,49 because he
holds that concept possession is not necessary to represent and detect similarities.
One can detect similarities between particulars (including objects, their features,
and entire perceptual scenes) by performing the kind of operations on imagistic,
nonconceptual representations that Kant ascribes to the faculty of imagination.50
An example of such an operation, mentioned by Kant, is the superimposition
of the mental images of different particulars for the purpose of comparing their
shapes (KU 5:231–​36). This is a simple procedure since it concerns only shapes,
and it can fall prey to the vanishing intersection problem (Harnard 2005, 28–​29).
However, we can conceive of more complex ways of detecting similarities among
objects, color shades, or perceptual scenes if we think of the Kantian imagination
as operating on similarity spaces. Similarity spaces are used in accounts of concep-
tual and nonconceptual mental content.51 In what follows, I understand similar-
ity spaces as nonconceptual representations of sensory features of particulars and
similarities between them. Kant can reject the second objection, with regard to
color shades, by claiming that we represent their similarities nonconceptually by
means of similarity spaces.
The idea underlying the notion of similarity space is that it is possible to repre-
sent sensory properties by employing geometrical structures. Consider pitch per-
ception. Humans and other animals can memorize the pitches of three sounds a,
b, and c, order them from the lowest to the highest, and tell whether the pitch of b
is more similar to the pitch of a or c. One can represent them as points on a line, as
shown in Figure 14.1, where A, B, and C represent, respectively, the pitches of a,
b, and c. If, and to the extent that, the pitch of b is perceived as being more similar
to the pitch of a than to c, B will be closer to A than to C.

48. Kant ascribes to nonhuman animals the capacity to compare representations (EE 20:211;
A. Dohna 145), identify identities and differences among them (Jäsche-​L. 9:65; Wiener L.
24:845–​4 6), and associate them with one another (Lett. [1789] 11:52).
49. On nonhuman animals’ lack of concepts, see, e.g., Wiener L. 24:845–​4 6, an early statement
in Spitzf. 2:59, and Jäsche-​L. 9:65.
50.  Kant distinguishes the imagination from the understanding, which is the “faculty of
concepts” (KU 5:287). In 1781, Kant distinguishes the imagination from the sensibility and
the understanding alike (A115). In 1787, he states that the imagination belongs to sensibility
(B151).
51. See respectively Gärdenfors 2000, Gauker 2011.
3 1 8   •  The Problem of Univer sals in Early Modern philosophy

A B C

Figure  14.1   Representation of sound pitches on a line.

The claim that sound pitches can be represented through spatial structures
finds support in the neurophysiology of pitch perception (Gärdenfors 2000, 13).
Each sound frequency stimulates a specific area of the cochlea. The area that a
sound stimulates is directly proportional to its frequency, with higher frequen-
cies stimulating areas closer to the base and lower frequencies stimulating areas
closer to the apex. This linear organization of auditory stimuli is reproduced in
the primary auditory cortex, where there are groups of neurons, called cochleo-
topic maps, that reproduce the spatial organization of the cochlea. As a result,
“the orderly mapping of neurons with sound frequencies is preserved from the
cochlea to the auditory cortex.”52
Unlike sound pitches, other sensory features must be represented through
more complex geometrical structures than points on a line. Consider, for instance,
the widely held view that there are four basic tastes (sweet, sour, salty, and bitter)
and that each other taste derives from the combination of three of them.53 Given
these assumptions, each taste can be represented as a point on a face of the regular
tetrahedron represented in Figure 14.2, called Henning’s tetrahedron, with the
four basic tastes at the vertices.54
Among the sensory features that can be represented by means of similarity
spaces are color hues. For the sake of simplicity, I assume that they can be repre-
sented as points on a segment, whose extremes border respectively infrared and
ultraviolet radiations. In fact, more complex geometric structures than segments
are required to represent them (Knoblauch 2002, 51–​54).
Let us assume that we represent color hues a, b, and c by means of points A, B,
and C on a segment that mirrors the visible spectrum. The more hue a is perceived
as being similar to hue b than to hue c, the closer the point representing a on the

52.  Gärdenfors 2000, 50. The cortex contains other topographic maps, such as those that
reproduce the spatial organization of stimuli on the retina and the localization of tactile sen-
sations on the body. Gallistel (1990, 477)  reviews “neurophysiological data supporting the
hypothesis that the nervous system does in fact quite generally employ vectors to represent
properties of both proximal and distal stimuli. The values of these representational vectors are
physically expressed by the locations of neural activity in anatomical spaces whose dimensions
correspond to descriptive dimensions of the stimulus.”
53. For an overview and critique of this position, see Erickson 2008.
54. This model was first put forward in Henning 1916. Those who deny that there are basic
tastes might prefer, as an example, the representation of the heaviness of an object as a point in
a three-​dimensional space (Shockley, Carello, and Turvey 2004).
Kant and Abstractionism about Concept Formation  •  319

Bitter

Salty
Sweet

Sour
Figure  14.2   Henning’s tetrahedron.

segment will be to the point representing b than to that representing c. In order to


generate a visual representation of the segment and the points, it is not necessary
to employ concepts or formulate judgments. It is sufficient to generate mental
images. Employing concepts or formulating judgments is also not necessary to
detect if a is more similar to b or to c. It is sufficient to manipulate segments in
the imagination. For instance, given Figure 14.1, one could either translate AB, or
rotate it 180 degrees around B, so as to superimpose AB over BC. Once AB and
BC are superimposed, it is visually apparent which segment is shorter and, hence,
whether b is more similar to a or to c.
Even if one grants that we can represent similarities between particulars
nonconceptually by means of similarity spaces, one could claim that we can use
similarity spaces only to formulate judgments. As Ernst Cassirer puts it, “[t]‌he
similarity of things” can “only be effective and fruitful, if it is understood and
judged as such.”55 Since, for Kant, the act of judging requires concept possession,56
only beings that possess concepts can use similarity spaces.
However, it is not the case that similarity spaces can be used only to formu-
late judgments. They can also be used to sort objects, as in the following exam-
ple. There is a bag full of colored chips, some of which are red or have colors
close to red, such as dark orange, whereas others are green or have colors close
to green, like some shades of blue. Several animals are able to sort the chips into

55.  Cassirer 1923, 15. More recently, Carsten Held (2001, 104)  claimed that the “compari-
son of different objects” is “unintelligible without the thinker already possessing general con-
cepts.” Held assumes that, in order to compare different objects, it is necessary to formulate
judgements.
56. This can be gathered from Kant’s claim that “there is a concept” in every judgment, under-
stood as the mental content associated with an act of judging (A68/​B93).
3 2 0   •  The Problem of Univer sals in Early Modern philosophy

two groups, those whose color is red or more similar to red than green and those
whose color is green or more similar to green than red.57 In order to carry out this
task, it is necessary to detect whether the color of each chip is closer to the color
of the chips placed in one group or the other. This task requires the detection of
similarities, but it gives rise to the sorting of objects, rather than the formulation
of judgments.
One could think that, although we can use similarity spaces to sort objects,
we can use them only if we possess concepts because sorting objects requires con-
cept possession. Placing chip a in the group of red chips in virtue of its color
requires the formulation of a judgment such as “a is more similar to these items
than to those items.” Formulating this judgment requires the possession of some
concepts, including at least a concept of similarity.
Kant would reject this view because he holds that nonhuman animals are
able to identify similarities among particulars, even though they lack con-
cepts.58 We may be inclined to think that if Kant had access to current-​day
ethological knowledge, he would have ascribed concepts to some nonhuman
animals. However, not only primates, but also pigeons can carry out sorting
tasks like the one in the colored chips example (Harnard 2005, 24–​25). The
claim that Kant would have ascribed concepts to pigeons is not very attrac-
tive, given the range of capacities that Kant closely links to concept posses-
sion. They include the capacities to carry out rule-​based categorizations,59 to
employ verbal language,60 to justify beliefs,61 and to possess mini-​theories of
the world.62
Since Kant holds that nonhuman animals can carry out sorting tasks even
though they cannot judge, he ought to explain what mental representations

57.  Categorical perception reduces the number of chips that one would not know how to
categorize.
58. See notes 48 and 49.
59. On the difference between rule-​based and similarity-​based categorization, see Smith and
Sloman 1994. Kant holds that concepts enable us to “think the particular as contained in the
universal” (KU 5:179), that is, to categorize particulars. These are rule-​based categorizations
for two reasons. First, Kant associates concepts to rules (e.g., A106, A722/​B750, A724/​B752;
Enzikl. 29:16, 17; L. Hechsel 396 = Wiener L. 24:909; M. Dohna 28:672). Second, Kant is
aware that nonhuman animals, which he takes to lack concepts, can detect similarities and,
hence, carry out similarity-​based categorizations, but he does not ascribe the capacity to “think
the particular as contained in the universal” to animals.
60. See note 40.
61. Kant relates conceptual content to the justification of the beliefs that are expressed by ana-
lytic judgments. See, e.g., Entd. 8:198.
62. According to Kant, all but simple concepts have a structure of subordinated and co-​ordi-
nated marks that encodes a set of beliefs regarding the items falling under those concepts.
Kant and Abstractionism about Concept Formation  •  321

underpin their sorting behavior. He could claim that this behavior is guided by
the mental act of including a particular, represented by means of nonconceptual
intuitions, in a similarity class, represented as a set of particulars whose represen-
tations in a similarity space are close to one another.63 This involves ascribing to
Kant the view that nonhuman animals can represent particulars by means of non-
conceptual intuitions. As for humans, Kant could claim that they too represent
particulars by means of nonconceptual intuitions. Alternatively, he could claim
that, prior to the formation of empirical concepts, humans represent particulars
by means of intuitions subsumed under the categories, but not under empirical
concepts. However things may be with regard to humans, this view entails that
Kant admits the existence of nonconceptual intuitions, which nonhuman ani-
mals possess and which represent particulars. Although I take these claims to be
supported by Kant’s texts, they are not uncontroversial.64

14.6. Conclusion
This chapter has reconstructed Kant’s abstractionist account of empirical con-
cept formation and discussed whether it falls victim to two objections. According
to the first objection, we cannot acquire even color concepts, which are straight-
forwardly related to visual experience, through acts of comparison, reflection,
and abstraction upon empirical intuitions, because “it is false that all instances
of a given color share some common features” (Carruthers 1992, 59), and the
boundary between colors is conventional. According to the second objection,
even if all instances of a given color share certain features, we can identify them
only if we possess a concept of that color. Kant can rebut the first objection by
arguing that all instances of some colors share a common feature. He can rebut
the second objection by arguing that, in order to compare the instances of a given
color with one another, it is not necessary to possess a concept of that color. We
can compare them because we have detected a similarity among them. We can
represent that similarity by means of nonconceptual similarity spaces, which can
be employed to group particulars in absence of empirical concepts. Once those
particulars have been grouped, it is possible to carry out the acts of comparison,
reflection, and abstraction that issue in the formation of an empirical concept.

63.  Christopher Gauker (2011, 145–​83) explains how animals lacking concepts can group
objects and carry out other tasks by manipulating mental images. Kant could adopt a similar
account.
64. David Landy (2009, 240, 243) and others hold that Kantian intuitions, as such, are con-
ceptual representations. Stefanie Grüne (2014, §2) denies that there is strong textual evidence
for the view that Kant ascribes intuitions to nonhuman animals.
3 2 2   •  The Problem of Univer sals in Early Modern philosophy

In addition to the objections discussed in this chapter, other criticisms have


been leveled against Kant’s abstractionist account of empirical concept forma-
tion. They aim to show that Kant’s account cannot explain, without circular-
ity, the formation of our first empirical concepts, but only, at most, the process
whereby we become conscious of their content (see, e.g., Ginsborg 2006, 40).
Although these criticisms take several forms, they often revolve around three
claims. The first is the claim that since, whatever may be the case for nonhuman
animals, humans’ “intuitions without concepts are blind” (A51/​B75), they can
represent the particulars which will be compared with one another only if they
are informed by concepts.65 The second is the claim that since, for Kant, all acts
of the understanding are acts of judging (A69/​B94) and acts of judging employ
concepts,66 the acts of comparison, reflection, and abstraction whereby we form
empirical concepts require the employment of other concepts.67 The third is the
claim that those other concepts must be empirical because we can apply the cat-
egories to intuitions only if we already possess empirical concepts.68
Addressing these objections would involve discussing three issues on which
Kant’s texts are less than clear, and on which there is little consensus among schol-
ars. The first is the meaning of the claim that intuitions without concepts are blind
and the thorny issue of whether Kant was a conceptualist or a nonconceptualist
about perception.69 The second is Kant’s view on the origin of the categories and
their role in the formation of empirical concepts. The third is whether Kant can
allow for a judgmental or proto-​judgmental activity that does not employ con-
cepts and can lead to the formation of concepts.70
The objections on which this chapter focused do not revolve on distinctive
Kantian claims but have also been advanced against Locke’s account of concept
formation and against abstractionist theories as such. I argued that none of the
objections are successful as they stand. The abstractionist accounts of concept
formation put forward by Kant and several of his early modern predecessors can-
not be dispensed with as quickly as has often been suggested.

65. See, e.g., Heller 1993, 82–​83; Vásquez Lobeiras 1998, 151.


66. See note 56.
67. Several scholars hold that, according to Kant, we form concepts by means of acts of judg-
ment. See, e.g., Allison 1973, 61–​65; Longuenesse 1998, 112, 164–​65.
68.  See, e.g., Allison 2001, 24; Kalar 2006, 48; Stern 1977, 20. By contrast, according to
Claudio La Rocca (2004), empirical concept formation operates on intuitions informed only
by the categories.
69. See, e.g., Bauer 2012, Hanna 2005.
70. Zuckert 2007, among others, employs the notion of proto-​judgment in relation to Kant.
Kant and Abstractionism about Concept Formation  •  323

Abbreviations
Unless otherwise indicated, from Kant 1900–​; cited by volume and page.

A. Dohna = Anthropologie Dohna-​Wundlacken; in Kowalewski 1924; cited


by page.
Danz. RT = Danziger Rationaltheologie
EE = Erste Einleitung in die Kritik der Urteilskraft
Entd. = Über eine Entdeckung
Enzikl. = Vorlesung Philosophische Enziklopädie
Fort. = Fortschritte der Metaphysik
Jäsche-​L. = Jäsche-​Logik
KU = Kritik der Urteilskraft
L. Bauch = Logik Bauch; in Kant 1998; cited by page.
L. Busolt = Logik Busolt
L. Dohna = Logik Dohna-​Wundlacken
Lett. = Briefwechsel
L. Hechsel = Logik Hechsel; in Kant 1998; cited by page.
L. Pölitz = Logik Pölitz
M. Dohna = Metaphysik Dohna-​Wundlacken
M. K2 = Metaphysik K2
M. K3 = Metaphysik K3
M. L1 = Metaphysik L1
Pr. Anthr. = Pragmatische Anthropologie
Refl. = Reflexionen from Kant’s handschriftlicher Nachlass
R. Pölitz = Philosophische Religionslehre Pölitz
Spitzf. = Die Falsche Spitzfindigkeit der vier syllogistischen Figuren
Warschauer L. = Warschauer Logik; in Kant 1998; cited by page.
Wiener L. = Wiener Logik
WORKS CITED

P r i m a r y S o u r c es

Aquinas, Thomas. 1947. Summa theologiae. Latin text of the Leonine edition. Translated
by Fathers of the English Dominican Province. 35  vols. New  York:  Benziger
Brothers.
—​—​—​. 1955–​57. Summa contra Gentiles. Translated by Anton C. Pegis. New  York,
Hanover House.
Arnauld, Antoine. 1967. Œuvres de Messire Antoine Arnauld Docteur de la maison et
société de Sorbonne. 43 vols. Brussels: Culture et Civilisation.
—​—​—​. 2001. Antoine Arnauld:  Textes philosophiques. Edited by Denis Moreau.
Paris: Presses Universitaires de France.
Baumeister, Friedrich Christian. 1747. Elementa philosophiae recentioris usibus iuven-
tutis scholasticae accommodata. Leipzig: Gleditsch. Reprinted in Wolff 1962–​, series
3, vol. 83.
Berkeley, George. 1948–​57. The Works of George Berkeley, Bishop of Cloyne. 9  vols.
Edited by T. Jessop and A. Luce. London: Thomas Nelson and Sons.
—​—​—​​. 1982. A Treatise concerning the Principles of Knowledge. Edited by Kenneth
Winkler. Indianapolis and Cambridge: Hackett.
Boethius, Anicius Manlius Severinus. 1968. The Theological Tractates. With an English
translation by H. F. Stewart and E. K. Rand. The Consolation of Philosophy. With
an English translation by “I. T.” (1609), revised by H. F. Stewart. London:  W.
Heinemann; Cambridge, MA: Harvard University Press.
Cudworth, Ralph. 1678. The True Intellectual System of the Universe. London: printed
for R. Royston.
—​—​—​​. 1690–​91. Bibliotheca Cudworthiana, sive Catalogus variorum librorum plurimis
facultatibus: insignium bibliothecae instructissimae Rev. Doct. Dr. Cudworth, S. T. P.,
Coll. Christi apud Cantab. nuperrimè praefecti . . . London: E. Millington.
3 2 6   •  Works Cited

—​—​—​​. 1731. A Treatise Concerning Eternal and Immutable Morality. London:  J. &
J. Knapton.
—​—​—​​. 1995. Traité de morale et Traité du libre arbitre. Introduction, translated and
edited by Jean-​Louis Breteau. Paris: Presses Universitaires de France.
Descartes, René, 1964–​74. Œuvres de Descartes. 11 vols. Edited by Charles Adam and
Paul Tannery. Paris: Vrin.
—​—​—​​. 1976. Descartes’ Conversation with Burman. Edited and translated by John
Cottingham. Oxford: Clarendon Press.
—​—​—​​. 1984–​85. The Philosophical Writings of Descartes. Vols. 1–​2. Translated by John
Cottingham, Robert Stoothof, and Dugald Murdoch. Cambridge:  Cambridge
University Press.
—​—​—​​. 1988. Descartes:  Selected Philosophical Writings. Translated by John
Cottingham, Robert Stoothoff, and Dugald Murdoch. Cambridge:  Cambridge
University Press.
—​—​—​. 1991. The Philosophical Writings of Descartes. Vol. 3. Translated by John
Cottingham, Robert Stoothof, Dugald Murdoch, and Anthony Kenny.
Cambridge: Cambridge University Press.
[ Desgabets, Robert]. 1675. Critique de la Critique de la recherche de la vérité,
ou l’on découvre le chemin qui conduit aux connoissances solides. Pour servir de reponse
à la letter d’un academicien. Paris: du Puis.
Desgabets, Robert. 1983–​85. Dom Robert Desgabets:  Œuvres philosophique inédites.
Edited by Joseph Beaude. 7 vols. Amsterdam: Quadratures.
Foucher, Simon. 1969. Critique de la Recherche de la vérité. Edited by Richard A.
Watson. New York: Johnson.
Frege, Gottlob. 1980. Translations from the Philosophical Writings of Gottlob Frege.
Edited by Max Black and Peter Geach. Oxford: Blackwell.
—​—​—​​. 1988. Die Grundlagen der Arithmetik: Eine logisch mathematische Untersuchung
über den Begriff der Zahl. Edited by Christian Thiel. Hamburg: Meiner. First pub-
lished in 1884.
Gassendi, Pierre. 1658. Opera Omnia in sex tomos divisa. Lyon: Laurent Anisson and
Jean-​Baptiste Devenet. Reprinted in facsimile with an introduction by Tullio
Gregory, Stuttgart-​Bad Canstatt: Friedrich Frohmann, 1964.
—​—​—​​. 1959. Exercitationes paradoxicae adversus Aristoteleos. Edited and translated
by Bernard Rochot, Dissertations en forme de paradoxes contre les Aristotéliciens.
Paris: Vrin.
—​—​—​​. 1962. Disquisitio metaphysica; seu, Dubitationes et instantiae adversus Renati
Cartesii Metaphysicam et responsa. Recherches métaphysiques; ou, Doutes et instances
contre la Métaphysique de R.  Descartes et ses réponses. Edited and translated by
Bernard Rochot. Paris: Vrin.
—​—​—​​. 1972. The Selected Works of Pierre Gassendi. Translated by Craig B. Brush.
New York, London: Johnson Reprint Corporation.
Works Cited  •  327

—​—​—​​. 1981. Pierre Gassendi’s Institutio Logica (1658). Edited and translated by
Howard Jones. Assen, Netherlands: Van Gorcum.
Hobbes, Thomas. 1961. Thomae Hobbes Malmesburienis Opera Philosophica quae
Latinae Scripsit, Studio et labore Guliemie Molesworth. 5 vols., London, 1839–​45.
Reprint Aalen: Scientia.
—​—​—​​. 1962. The English Works of Thomas Hobbes of Malmesbury; now first collected
and edited by sir William Molesworth, 11 vols, London 1839–​45. Reprint Aalen:
Scientia.
—​—​—​​. 1966. Thomae Hobbes Malmesburienis Opera Philosophica quae Latinae Scripsit,
Studio et labore Guliemie Molesworth. 5 vols. Reprint. Aalen: Scientia Verlag.
—​—​—​​. 1973. Critique du De Mundo de Thomas White. Edited by Jean Jacquot and
Harold Whitmore Jones. Paris: Vrin.
—​—​—​​. 1981. Part I of De Corpore. Translated by Aloysius P. Martinich. New York: Abaris.
—​—​—​​. 1990. The Elements of Law, Natural and Politic. Edited by J. C.  A. Gaskin.
Oxford: Oxford University Press.
—​—​—​​. 1994a. The Correspondence of Thomas Hobbes. Edited by Noel Malcolm.
Oxford: Clarendon Press.
—​—​—​​. 1994b. Leviathan. Edited by Edwin Curley. Indianapolis: Hackett.
—​—​—​​.1999. De Corpore. Edited by Karl Schuhmann. Paris: J. Vrin.
Hume, David. 1978. A Treatise of Human Nature. Edited by L. A. Selby-​Bigge and P. H.
Nidditch. 2nd edition. Oxford: Oxford University Press.
—​—​—​​. 2000a. An Enquiry concerning Human Understanding. Edited by Tom L.
Beauchamp. Oxford: Clarendon Press.
—​—​—​​. 2007. A Treatise of Human Nature. Edited by David Fate Norton and Mary J.
Norton. Oxford: Clarendon Press.
Kant, Immanuel. 1900–​. Gesammelte Schriften. Edited by the Königlich Preussische
(Deutsche) Akademie der Wissenschaften. 29 vols. Berlin: Reimer; de Gruyter.
—​—​—​​. 1998. Logik-​Vorlesung:  Unveröffentlichte Nachschriften. Edited by Tillmann
Pinder. 2 vols. Hamburg: Meiner.
Leibniz, Gottfried Wilhelm. 1875–​90. Die philosophische Schriften. 7  vols. Edited
by Carl Immanuel Gerhardt. Berlin:  Weidemann. Reprint Hildesheim:  George
Olms, 2008.
—​—​—​. (1714) 1890. “Principes de la Nature et de la Grace, fondés en raison.” In Die
philosophischen Schriften von Gottfried Wilhelm Leibniz, edited by C. I. Gerhardt,
vol. 6, 598–​606. Berlin:  Weidmann; translated in Philosophical Essays (1989),
edited by Roger Ariew and Daniel Garber, 207–​13. Indianapolis: Hackett.
—​—​—​. 1923–​. Gottfried Wilhelm Leibniz:  Sämtliche Briefe und Schriften. Edited by
[Preussischen] Deutsche Akademie der Wissenschaften zu Berlin. Darmstadt and
Berlin: Akademie-​Verlag.
—​—​—​. 1966. Logical Papers. Edited and translated by G. H.  R. Parkinson.
Oxford: Clarendon Press.
3 2 8   •  Works Cited

—​—​—​. 1969. Philosophical Papers and Letters. Translated by Leroy E. Loemker. 2nd
edition. Dordrecht: Kluwer.
—​—​—​. 1989. Leibniz: Philosophical Essays. Translated and edited by Roger Ariew and
Daniel Garber. Indianapolis: Hackett.
—​—​—​. 1996. New Essays on Human Understanding. Translated and edited by Peter
Remnant and Jonathan Bennett. Cambridge: Cambridge University Press.
Lelevel, Henri de. 1694. La Vraye et la fausse métaphysique [. . .] où l’on refute les senti-
mens de M. Régis. 2 vols. Rotterdam: Leers.
Locke, John. 1823. “An Examination of P. Malebranche’s Opinion of Seeing All Things
in God.” In The Works of John Locke in Nine Volumes, vol. 8, 211–​255. 12th ed.,
London: Rivington.
—​—​—​. 1975. An Essay Concerning Human Understanding. Edited by Peter H.
Nidditch. Oxford: Clarendon Press.
Malebranche, Nicolas. 1700. Treatise Concerning the Search after Truth. Translated by
Thomas Taylor. London: W. Bowyer, T. Leigh, and W. Midwinter.
—​—​—​. 1871. Œuvres, tome 1, Entretiens sur la métaphysique. Edited by J. Simon. Paris:
Charpentier.
—​—​—​. 1958–​67. Œuvres complètes. Edited by André Robinet. 20 volumes.
Paris: Vrin CNRS.
—​—​—​. 1962–​6 4. De la recherche de la vérité, où l’on traite de la nature de l’esprit de
l’homme et de l’usage qu’il en doit faire pour éviter l’erreur dans les sciences. Edited by
G. Rodis-​Lewis, t. III, Eclaircissement X, 131. Paris: Vrin.
—​—​—​. 1997. The Search after Truth. Edited and Translated by Thomas M. Lennon and
Paul J. Olscamp. Cambridge: Cambridge University Press.
—​—​—. 1999. Dialogues on Metaphysics and on Religion. Edited by Nicholas Jolley and
David Scott. Cambridge: Cambridge University Press
Meier, Georg Friedrich. (1752) 1924. Auszug aus der Vernunftlehre. In Kant 1900–​,
vol. 16.
—​—​—​. (1752) 1997. Vernunftlehre. Edited by Günter Schenk. 3  vols. Halle/​
Saale: Hallescher Verlag.
More, Henry. 1659. The Immortality of the Soul. London.
—​—​—​. 1662. A Collection of several Philosophical Writings. London: J. Flesher.
—​—​—​. 1675–​79. Opera Omnia, tum quae latine, tum quae anglice scripta sunt; nunc
vero latinitate donata. London: J. Macock for J. Martyn and W. Kettilby. Reprinted
1966. Opera Omnia. II. Opera philosophica. 2 vols. Hildesheim: Georg Olms.
—​—​—​. 1679. Enchiridium Metaphysicum, sive, de rebus incorporeis succincta et luculenta
Dissertatio. In More 1675–​79, vol. 2, 133–​334. Reprinted 1966, Opera Omnia. II.
Opera philosophica, vol. 1.
—​—​—​. 1682. “Annotations upon the Discourse of Truth.” In Two choice and useful trea-
tises the one, Lux orientalis … the other, A discourse of truth. London: J. Collins and
S. Lownds.
Works Cited  •  329

—​—​—​.1995. Henry More’s Manual of Metaphysics. A Translation of the “Enchiridium


Metaphysicum” [1679], with An Introduction and Notes. [Part I, Chapters 1–​10 and
27–​28]. Translated by Alexander Jacob. Hildesheim: Olms.
—​—​—.1998. A Platonick Song of the Soul. Edited with an introductory study by
Alexander Jacob. Lewisburg:  Bucknell University Press; London:  Associated
University Press.
Nizolio, Mario. 1956. De veris principiis et vera ratione philosophandi contra pseudophi-
losophos. Book 4. Edited by Quirinus Breen. Rome: Bocca.
Norris, John. 1687. A Collection of Miscellanies. Oxford: John Crosley.
—​—​—. 1689. Reason and Religion. London: S. Manship.
—​—​—​. 1722. An Essay towards the Theory of the Ideal or Intelligible World. 2  vols.
London: E. Parker.
Ockham, William. 1998. Ockham’s Theory of Terms:  Part I  of the “Summa Logicae”.
Translated by Michael J. Loux. South Bend, IN: St. Augustine’s.
Regis, Pierre-​Sylvain. 1690. Système de philosophie, contenant la logique, la metaphy-
sique, la physique, et la morale. 2 vols. in 4. Paris: Thierry.
—​—​—​. 1970. Cours entire de philosophie; ou, Système generale selon les principes de
M. Descartes, contenant la logique, la metaphysique, la physique, et la morale. Edited
by R. A. Watson. 3 vols. New York: Johnson. [Retitled edition of Regis 1690.]
—​—​—​. 1996. L’Usage de la raison et de la foy, ou l’accord de la foy et de la raison. Edited
by Jean-​Robert Armogathe. Paris: Fayard.
Spinoza, Benedict. 1925. Opera. Edited by Carl Gebhardt. 4  vols. Heidelberg:  Carl
Winters.
—​—​—​. 1985. The Collected Works of Spinoza. Vol. 1. Translated and edited by Edwin
Curley. Princeton, NJ: Princeton University Press.
—​—​—​. 2002. Complete Works. Translated by Samuel Shirley. Indianapolis: Hackett.
Steuchus Augustinus Eugubinus. 1542. De Perenni Philosophia. Basileae: per Nicolaum
Bryling et Sebastianum Francken.
Suárez, Francisco. 1947. On the Various Kinds of Distinction. Edited and translated by
Cyril Vollert. Milwaukee, WI: Marquette University Press; reprinted 2007.
—​—​—​. 1964. On Formal and Universal Unity. Edited and translated by James F. Ross.
Milwaukee, WI: Marquette University Press.
—​—​—​. 1965. Disputationes Metaphysicae. 2  vols. Hildesheim:  Georg Olms;
reprinted 2009.
—​—​—​. 1983. On the Essence of Finite Being As Such, On the Existence of That Essence
and Their Distinction (from Disputationes Metaphysicae, vol. 31). Translated by
Norman Wells. Milwaukee, WI: Marquette University Press.
Wolff, Christian. 1738. Psychologia empirica. Revised ed. Frankfurt: Renger. Reprinted
in Wolff 1962–​, series 1, vol. 5.
—​—​—​. 1740. Der vernünfftigen Gedancken von Gott, der Welt und der Seele des
Menschen, auch allen Dingen überhaupt, anderer Teil, bestehend in ausfürlichen
3 3 0   •  Works Cited

Anmerckungen. 4th ed. Frankfurt:  Renger. Reprinted in Wolff 1962–​, series 1,


vol. 3.
—​—​—​. 1751. Vernünftige Gedancken von GOTT, der Welt und der Seele des Menschen.
Enlarged ed. Halle: Renger. Reprinted in Wolff 1962–​, series 1, vol. 5.
—​—​—​. 1962–​. Gesammelte Werke. Hildesheim: Georg Olms.

S eco n d a r y L i t er at u r e

Aaron, Richard. 1967. The Theory of Universals. Oxford: Oxford University Press.


Acworth, Richard. 1975. La philosophie de John Norris, 1657–​1712. Paris: H. Champion.
Adams, Marilyn McCord. 1987. William of Ockham. 2 vols. Notre Dame, IN: Notre
Dame University Press.
Allison, Henry E. 1973. “A Historical-​Critical Introduction.” In The Kant-​Eberhard
Controversy, 1–​104. Baltimore, MD: Johns Hopkins University Press.
—​—​—​. 2001. Kant’s Theory of Taste: A Reading of the Critique of Aesthetic Judgment.
Cambridge: Cambridge University Press.
—​—​—​. 2004. Kant’s Transcendental Idealism: An Interpretation and Defense. 2nd edi-
tion, revised and enlarged. New Haven, CT: Yale University Press.
Angelelli, Ignacio. 1965. “Leibniz’s Misunderstanding of Nizolius’ Notion of
‘Multitudo.’” Notre Dame Journal of Formal Logic 6: 319–​22.
Ariew, Roger. 2001. “Leibniz’s Metaphysical Disputation on the Principle of
Individuation: A Scholastic Exercise.” In Nihil sine ratione. Akten des VII Internationalen
Leibniz-​Kongresses, edited by Herbert Breger, 33–​40. Berlin: Leibniz-​Gesellschaft.
Atherton, Margaret. 1987. “Berkeley’s Anti-​Abstractionism.” In Essays on the Philosophy
of George Berkeley, edited by Ernest Sosa, 45–​60. Dordrecht: Reidel.
Atkinson, Martin. 1982. Explanations in the Study of Child Language Development.
Cambridge: Cambridge University Press.
Ayers, Michael. 1975. Editor’s Introduction to Philosophical Works, by George Berkeley,
vi–​xxiv. London: Dent.
—​—​ —​ . 1981. “Locke vs. Aristotle on Natural Kinds.” Journal of Philosophy
78: 247–​72.
—​—​—​. 1986. “Are Locke’s ‘Ideas’ Images, Intentional Objects, or Natural Signs?” Locke
Newsletter 17: 3–​36.
—​—​—​. 1991. Locke: Epistemology and Ontology. 2 vols. London: Routledge.
—​—​—​. 2008. “Locke’s Account of Abstract Ideas—​Again.” In Studies on Locke: Sources,
Contemporaries, and Legacy, edited by Sarah Hutton and Paul Schuman, 59–​74.
Dordrecht: Springer.
Bardout, Jean-​Christophe. 1998. “Toute-​puissance et singularités.” In La légèreté de
l’être: Études sur Malebranche, edited by Bruno Pinchard, 95–​119. Paris: Vrin.
Baxter, Donald L. M. 1988a. “Many-​One Identity.” Philosophical Papers 17: 193–​216.
—​—​—​. 1988b. “Identity in the Loose and Popular Sense.” Mind 97: 575–​82.
Works Cited  •  331

—​—​—​. 1989. “Identity through Time and the Discernibility of Identicals.” Analysis
49: 125–​31.
—​—​—​. 1997. “Abstraction, Inseparability, and Identity.” Philosophy and Phenomenological
Research 57: 307–​30.
—​—​—​. 1999. “The Discernibility of Identicals.” Journal of Philosophical Research
24: 37–​55.
—​—​—​. 2001. “Instantiation as Partial Identity.” Australasian Journal of Philosophy
79: 449–​6 4.
Bauer, Nathan. 2012. “A Peculiar Intuition:  Kant’s Conceptualist Account of
Perception.” Inquiry 55: 215–​37.
Belfrage, Bertil. 1987. George Berkeley’s Manuscript Introduction. Oxford: Doxa.
Bennett, Jonathan. 1994. “Descartes’s Theory of Modality.” Philosophical Review
103: 639–​67.
—​—​—​. 2001. Learning from Six Philosophers:  Descartes, Spinoza, Leibniz, Locke,
Berkeley, Hume. 2 vols. Oxford: Clarendon Press.
Berger, David. 2009. Kant’s Aesthetic Theory:  The Beautiful and Agreeable.
London: Continuum.
Bernhardt, Jean. 1985. “Nominalisme et mécanisme dans la pensée de Hobbes.” Archives
de Philosophie 48: 235–​49.
Berti, Enrico. 2009. In principio era la meraviglia:  Le grandi questioni della filosofia
antica. Rome and Bari: Laterza.
Bloch, Olivier. 1971. La philosophie de Gassendi:  Nominalisme, Matérialisme et
Métaphysique. The Hague: Martinus Nijhoff.
Bolton, Martha Brandt. 1977. “Leibniz and Hobbes on Arbitrary Truth.” Philosophy
Research Archives 3: 242–​73.
—​—​—​. 1987. “Berkeley’s Objection to Abstract Ideas and Unconceived Objects.”
In Essays on the Philosophy of George Berkeley, edited by Ernest Sosa, 61–​83.
Dordrecht: Reidel.
—​—​—​. 1990. “Locke’s Taxonomy of Ideas in Locke’s Essay.” In The Cambridge
Companion to Locke’s “Essay concerning Human Understanding”, edited by Lex
Newman, 67–​100. Cambridge: Cambridge University Press.
—​—​—​. 1998. “Universals, Essences, and Abstract Entities.” In The Cambridge History of
Seventeenth-​Century Philosophy, edited by Daniel Garber and Michael Ayers, vol. 1,
178–​211. Cambridge: Cambridge University Press.
—​—​ —​. 2006. “Leibniz and the Limits of Perception.” In Einheit in der
Vielheit:  Proceedings of VIII. Internationaler Leibniz-​Kongress, edited by Herbert
Breger, Jügen Herbert, and Sven Erdner, 10–​19. Hannover: Hartmann GmbH.
—​—​—​. 2007. “Leibniz’s Nouveaux Essais: A Contest by Dialogue.” In Leibniz and the
English Speaking World, edited by Pauline Phemister and Stuart Brown, 111–​32.
Berlin: Springer Verlag.
Bolton, Neil. 1977. Concept Formation. Oxford: Pergamon.
3 3 2   •  Works Cited

Bosley, Richard N., and Martin Tweedale. 1997. Basic Issues in Medieval Philosophy.
Peterborough, ON: Broadview.
Boster, James S. 2005. “Categories and Cognitive Anthropology.” In Handbook of
Categorization in Cognitive Science, edited by Henri Cohen and Claire Lefebvre,
91–​118. Amsterdam: Elsevier.
Bracken, Harry M. 1974. Berkeley. London: Macmillan.
Breger, Herbert. 2006. “The Proof of the Axioms.” In Leibniz selon les Nouveaux Essais,
edited by François Duchesneau, 49–​58. Québec: Éditions Bellarmine.
Breteau, Jean Louis. 1995. “Introduction to R. Cudworth.” In Traité de morale et Traité
du libre arbitre, 1–33. Paris: Presses Universitaires de France.
Bricke, John. 1980. Hume’s Philosophy of Mind. Princeton, NJ:  Princeton
University Press.
Burgess, John P., and Gideon Rosen. 1997. A Subject with No Object:  Strategies for
Nominalistic Interpretation of Mathematics. Oxford: Oxford University Press.
Callaghan, G. K. 2001. “Nominalism, Abstraction, and Generality in Hobbes.” History
of Philosophy Quarterly 18: 37–​55.
Capozzi, Mirella. 1987. “Kant on Logic, Language and Thought.” In Speculative
Grammar, Universal Grammar and Philosophical Analysis of Language, edited by
Dino Buzzetti and Maurizio Ferriani, 97–​147. Amsterdam: Benjamins.
—​—​—​. 2001. Kant e la logica, vol. 1. Naples: Bibliopolis.
Carpenter, Andrew. 1995. “Kant’s (Problematic) Account of Empirical Concepts.” In
Proceedings of the Eighth International Kant Congress:  Memphis, 1995, edited by
Hoke Robinson, vol. 2, 227–​33. Milwaukee: Marquette University Press.
Carriero, John. 1995. “On the Relationship between Mode and Substance in Spinoza’s
Metaphysics.” Journal of the History of Philosophy 33: 245–​73.
Carruthers, Peter. 1992. Human Knowledge and Human Nature: A New Introduction to
an Ancient Debate. New York: Oxford University Press.
Cassirer, Ernst. 1922. Das Erkenntnisproblem in der Philosophie und Wissenschaft der
Neueren Zeit. Berlin: Bruno Cassirer.
—​—​—​. 1923. Substance and Function and Einstein’s Theory of Relativity. Translated by
William Curtis Swabey and Marie Collins Swabey. Chicago:  Open Court. First
published in 1910.
Ceyssens, Lucien. 1974. Sources relatives à l’histoire du jansénisme et de l’anti-​jansénisme
des années 1677–​1679. Louvain: Presses Universitaires de Louvain.
Chappell, Vere. 1997. “Descartes’s Ontology,” Topoi 16: 111–​27.
Cleary, John J. (ed.). 1997. The Perennial Tradition of Neoplatonism. Leuven: Leuven
University Press.
Cohen, Henri, and Claire Lefebvre, eds. 2005. Handbook of Categorization in Cognitive
Science. Amsterdam: Elsevier.
Conrad, Elfriede. 1994. Kants Logikvorlesungen als neuer Schlüssel zur Architektonik der
Kritik der reinen Vernunft: Die Ausarbeitung der Gliederungsentwürfe in den Logikvorles­
ungen als Auseinandersetzung mit der Tradition. Stuttgart: Frommann-​Holzboog.
Works Cited  •  333

Connell, Desmond. 1967. The Vision in God:  Malebranche’s Scholastic Sources.


New York: Humanities Press.
Cook, Richard S., Paul Kay, and Terry Regier. 2005. “The World Color Survey
Database.” In Handbook of Categorization in Cognitive Science, edited by Henri
Cohen and Claire Lefebvre, 223–​41. Amsterdam: Elsevier.
Crocker, Robert. 2003. Henry More, 1614–​1687. A Biography of the Cambridge Platonist.
Dordrecht: Kluwer.
Cunning, David. 2003. “True and Immutable Natures and Epistemic Progress in
Descartes’s Meditations.” British Journal for the History of Philosophy 11: 235–​48.
—​—​—​. 2010. Argument and Persuasion in Descartes’ “Meditations.” New York: Oxford
University Press.
Dascal, Marcelo. 1987. Leibniz:  Language, Signs and Thought. Amsterdam and
Philadelphia, PA: Benjamins.
Davidoff, Jules, Ian Davies, and Debi Roberson. 1999. “Colour Categories in a Stone-​
Age Tribe.” Nature 398: 203–​04.
De Boer, John. 1931. The Theory of Knowledge of the Cambridge Platonists.
Madras: Methodist Publishing House.
Della Rocca, Michael. 2002. “René Descartes.” In A Companion to Early Modern
Philosophy, edited by Steven Nadler, 60–​79. Oxford: Blackwell.
De Rosa, Raffaella. 2011. “Rethinking the Ontology of Cartesian Essences.” British
Journal for the History of Philosophy 19: 605–​22.
Di Bella, Stefano. 1998. “L’astratto e il concreto. Hobbes, Leibniz e la riforma
dell’ontologia.” Rivista di storia della filosofia 53: 235–​66.
—​—​—​. 2014. “Leibniz et l’inhérence: ‘Diu multumque investigavi generalem notionem
tou inesse.’” In Leibniz. Monades e mondes. Archives de Philosophie, 77: 17–​42.
Doney, Willis 1982. “Is Berkeley’s a Cartesian mind?” In Berkeley: Critical and Interpretive
Essays, edited by Colin Turbayne, 273–​82. Minneapolis: University of Minnesota Press.
Doolan, Gregory. 2008. Aquinas on the Divine Ideas as Exemplar Causes. Washington,
DC: Catholic University of America Press.
Duncan, Stewart. 2010. “Leibniz on Hobbes’s Materialism.” Studies in History and
Philosophy of Science 41: 11–​18.
—​—​—​. 2011. “Hobbes, Signification, and Insignificant Names.” Hobbes Studies 24: 158–​78.
Dupré, John. 1993. The Disorder of Things: Metaphysical Foundations of the Disunity of
Science. Cambridge, MA: Harvard University Press.
Easton, Patricia. 2009. “What is at Stake in the Cartesian Debates on the Eternal
Truths?” Philosophy Compass 4: 348–​62.
Emilsson, Eyjólfur Kjalar. 2007. Plotinus on Intellect. Oxford: Clarendon Press.
Erickson, Robert P. 2008. “A Study of the Science of Taste:  On the Origins and
Influence of the Core Ideas.” Behavioral and Brain Sciences 31: 59–​105.
Favaretti Camposampiero, Matteo. 2010. “Pure Intellect, Brain Traces, and
Language: Leibniz and the Foucher-​Malebranche Debate.” Oxford Studies in Early
Modern Philosophy 5: 115–​4 4.
3 3 4   •  Works Cited

Faye, Emmanuel. 2005. “The Cartesianism of Desgabets and Arnauld and the Problem
of the Eternal Truths.” Oxford Studies in Early Modern Philosophy 2: 193–​209.
Field, Hartry. 1980. Science without Numbers:  A  Defense of Nominalism. Princeton,
NJ: Princeton University Press.
Flew, Antony, ed. 1984. A Dictionary of Philosophy. London: Pan Books in ass. with The
Macmillan Press.
Franklin, Anna, Gilda V. Drivonikou, Laura Bevis, Ian R.  L. Davies, Paul Kay, and
Terry Regier. 2008. “Categorical Perception of Color Is Lateralized to the Right
Hemisphere in Infants, but to the Left Hemisphere in Adults.” Proceedings of the
National Academy of Sciences of the United States of America 105: 3221–​25.
Friedman, Russell L. 1999. “Divergent Traditions in Later-​ Medieval Trinitarian
Theology: Relations, Emanations, and the Use of Philosophical Psychology, 1250–​
1325.” Studia Theologica: Nordic Journal of Theology 53: 13–​25.
Gallistel, Charles R. 1990. The Organization of Learning. Cambridge, MA: MIT Press.
Galluzzo, Gabriele. 2011. Breve storia dell’ontologia. Rome: Carocci.
Garber, Daniel, and Michael Ayers, eds. 1998. The Cambridge History of Seventeenth-​
Century Philosophy. Cambridge: Cambridge University Press.
Gärdenfors, Peter. 2000. Conceptual Spaces:  The Geometry of Thought. Cambridge,
MA: MIT Press.
Garrett, Don. 1997. Cognition and Commitment in Hume’s “Treatise.” Oxford: Oxford
University Press.
Gauker, Christopher. 2011. Words and Images:  An Essay on the Origin of Ideas.
Oxford: Oxford University Press.
Gaukroger, Stephen. 1978. Explanatory Structures: A Study of Concepts of Explanation
in Physics and Philosophy. Hassocks, UK: Harvester.
Geach, Peter. 1957. Mental Acts: Their Content and Their Objects. London: Routledge
& Kegan.
Gewirth, Alan. 1970. “The Cartesian Circle Reconsidered.” Journal of Philosophy
67: 668–​85.
—​—​—​. 1971. “Two Disputed Questions.” Journal of Philosophy 68: 288–​96.
Gilson, Étienne. 1979. Index scolastico-​cartésien. 2nd revised ed. Paris: Vrin. First pub-
lished in 1913.
Ginsborg, Hannah. 2006. “Thinking the Particular as Contained under the Universal.”
In Aesthetics and Cognition in Kant’s Critical Philosophy, edited by Rebecca Kukla,
35–​60. Cambridge: Cambridge University Press.
Gouhier, Henri. 1976. “La première polemique de Malebranche.” In Études d’histoire de
la philosophie française, 64–​113. Hildesheim: Georg Olms.
Grant, Edward. 1982. “The Effect of the Condemnation of 1277.” In The Cambridge
History of Later Medieval Philosophy, edited by Norman Kretzmann, Anthony
Kenny, and Jan Pinborg, 537–​39. Cambridge: Cambridge University Press.
Works Cited  •  335

Grüne, Stefanie. 2014. “Reply to Colin McLear.” Critique. https://​virtualcritique.word-


press.com/​2014/​08/​20/​reply-​to-​colin-​mclear/​; archived at http://​www.webcita-
tion.org/​6VPDP3r58.
Gueroult, Martial. 1955. Malebranche. 3 vols. Paris: Aubier-​Montaigne.
—​—​—​. 1968–​74. Spinoza. 2 vols. Hildesheim: Georg Olms.
—​—​—​. 1984. Descartes’ Philosophy Interpreted According to the Order of Reasons.
Translated by Roger Ariew. 2 vols. Minneapolis: University of Minnesota Press.
Hacking, Ian.1975. Why Does Language Matter to Philosophy? Cambridge: Cambridge
University Press.
Hanna, Robert. 2001. Kant and the Foundations of Analytic Philosophy. Oxford: Oxford
University Press.
—​—​—​. 2005. “Kant and Nonconceptual Content.” European Journal of Philosophy
13: 247–​90.
Hark, Michel ter. 2008. “Wittgenstein, the Secondary Use of Words, and Child
Psychology.” In Wittgenstein’s Enduring Arguments, edited by David K. Levy and
Edoardo Zamuner, 87–​107. London: Routledge.
Harnard, Stevan. 2005. “To Cognize is to Categorize: Cognition is Categorization.” In
Handbook of Categorization in Cognitive Science, edited by Henri Cohen and Claire
Lefebvre, 19–​43. Amsterdam: Elsevier.
Haserot, Francis S. 1950. “Spinoza and the Status of Universals.” Philosophical Review
59: 469–​92.
Hazard, Paul. 1995. La crise de la conscience européenne. Paris: Livre de Poche. First pub-
lished in 1935.
Held, Carsten. 2001. “Analyticity and the Semantics of Predicates.” In Kant’s
Legacy: Essays in Honor of Lewis White Beck, edited by Predrag Cicovacki, 93–​116.
Rochester, NY: University of Rochester Press.
Heller, Edmund. 1993. “Kant und J.  S. Beck über Anschauung und Begriff.”
Philosophisches Jahrbuch 100: 72–​95.
Henning, Hans. 1916. “Die Qualitätenreihe des Geschmacks.” Zeitschrift für Psychologie
und Physiologie der Sinnesorgane 74: 203–​19.
Hight, Marc A. ed. 2012. The Correspondence of George Berkeley. Cambridge
University Press.
Hübener, Wolfgang. 1977. “Ist Thomas Hobbes Ultranominalist gewesen?” Studia
Leibnitiana 9:77–​100.
Jaffro, Laurent, Geneviève Brykman, and Claire Schwartz, eds. 2009. Berkeley’s
“Alciphron”: English Text and Essays in Interpretation. Hildesheim: Georg Olms.
Jesseph, Douglas. 1999. Squaring the Circle. Chicago: University of Chicago Press.
Jolley, Nicholas. 1982. “Leibniz and Locke on Essences.” In Leibniz:  Critical and
Interpretative Essays, edited by Michael Hooker, 196–​208. Minneapolis: University
of Minnesota Press.
3 3 6   •  Works Cited

—​—​—​. 1990. The Light of the Soul:  Theories of Ideas in Leibniz, Malebranche, and
Descartes. Oxford: Oxford University Press.
Kalar, Brent. 2006. The Demands of Taste in Kant’s Aesthetics. London: Continuum.
Kenny, Anthony. 1968. Descartes: A Study of His Philosophy. New York: Random House.
—​—​—​. 1970. “The Cartesian Circle and the Eternal Truths.” Journal of Philosophy
67: 685–​700.
King, Peter. 2003. “Scotus on Metaphysics.” In The Cambridge Companion to Duns
Scotus, edited by Thomas Williams, 15–​68. Cambridge: Cambridge University Press.
Kitcher, Patricia. 1990. Kant’s Transcendental Psychology. New  York:  Oxford
University Press.
Knoblauch, Kenneth. 2002. “Color Vision.” In Sensation and Perception, edited by
Steven Yantis, 41–​75. Vol. 1 of Stevens’ Handbook of Experimental Psychology, 3rd
ed. New York: Wiley.
Kowalewski, Arnold, ed. 1924. Die philosophischen Hauptvorlesungen Immanuel
Kants:  Nach den aufgefundenen Kollegheften des Grafen Heinrich zu Dohna-​
Wundlacken. Munich: Rösl.
Kripke, Saul A. 1980. Naming and Necessity. Oxford: Blackwell.
—​—​—​. 1982. Wittgenstein on Rules and Private Language. Oxford: Blackwell.
Kristeller, Paul Oskar. 2005. Il pensiero filosofico di Marsilio Ficino. Florence: Le Lettere.
Laird, John. 1931. Hume’s Philosophy of Human Nature. New York: Dutton.
—​—​—​. 1934. Hobbes. London: Ernest Benn.
Landy, David. 2009. “Inferentialism and the Transcendental Deduction.” Kantian
Review 14: 1–​30.
LaPorte, Joseph. 2004. Natural Kinds and Conceptual Change. Cambridge: Cambridge
University Press.
La Rocca, Claudio. 2004. “How Are Synthetic a priori Judgments Possible? The
Conditions and Process of Empirical Knowledge in Kant.” Quaestio 4:265–​93.
Leduc, Christian. 2006. “Définition et substance dans les Nouveaux Essais.” In Leibniz
selon les Nouveaux Essais sur l’entendement humain, edited by François Duschesneau
and Jérémie Griard, 179–​88. Montréal: Bellermin.
Lennon Thomas. 1993. The Battle of the Gods and the Giants: The Legacy of Descartes
and Gassendi, 1655–​1715. Princeton, NJ: Princeton University Press.
—​—​—​. 1994. “The Problem of Individuation among the Cartesians.” In Individuation
and Identity in Early Modern Philosophy: Descartes to Kant, edited by Kenneth F.
Barker and Jorge J. E. Gracia, 13–​39. Albany: State University of New York Press.
Levy, Samuel, 2002. “Leibniz and the Sorites.” Leibniz Review 12: 25–​52.
Longuenesse, Béatrice. 1998. Kant and the Capacity to Judge: Sensibility and Discursivity
in the Transcendental Analytic of the “Critique of Pure Reason.” Princeton,
NJ: Princeton University Press.
Lotti, Brunello. 2004. Ralph Cudworth e l’idea di natura plastica. Udine: Campanotto.
Works Cited  •  337

Lovejoy, Arthur O. 1908. “Kant and the English Platonists.” In Essays, Philosophical and
Psychological. London: Longmans, Green.
—​—​—​. 1936. The Great Chain of Being: A Study of the History of an Idea. Cambridge,
MA: Harvard University Press.
Lowe, E. J. 1998. The Possibility of Metaphysics:  Substance, Identity, and Time.
Oxford: Clarendon Press.
—​—​—​. 2006. The Four-​Category Ontology:  A  Metaphysical Foundation for Natural
Science. Oxford: Clarendon Press.
—​—​—​. 2008. “Two Notions of Being:  Entity and Essence.” In Being:  Developments
in Contemporary Metaphysics, edited by Robin Le Poidevin, 23–​48. Cambridge:
Cambridge University Press.
Lyssy, Ansgar. 2007. “Subjektiv, Intersubjektiv, Objektiv:  Über die Struktur der
Erfahrungsurteile bei Kant.” In Transzendentalphilosophie und Person: Leiblichkeit,
Interpersonalität, Anerkennung, edited by Christoph Asmuth, 147–​62. Bielefeld:
Transcript.
Mackie, J. L. 1976. Problems from Locke. Oxford: Clarendon Press.
MacKinnon, Flora Isabel. 1910. The Philosophy of John Norris of Bemerton. Baltimore,
MD: The Review Publishing Company.
Mander, William J. 2008. The Philosophy of John Norris. Oxford: Oxford University Press.
Margolis, Joseph. 1982. “Berkeley and Others on the Problem of Universals.” In
Berkeley:  Critical and Interpretive Essays, edited by Colin Turbayne, 207–​27.
Minneapolis: University of Minnesota Press.
MacBride, Fraser. 2004. Review of Rodriguez-​Pereyra (2002). Notre Dame Philosophical
Reviews. 2004.02.07. http://​ndpr.nd.edu/​news/​23684-​resemblance-​nominalism-​
a-​solution-​to-​the-​problem-​of-​universals/​.
Malcolm, Noel. 2002. Aspects of Hobbes. Oxford: Clarendon.
Martinich, Aloysius P. 1981. Translator’s Commentary to Part I  of De Corpore, by
Thomas Hobbes, 333–440. Translated by Aloysius P. Martinich. New York: Abaris.
Mates, Benson. 1978. “The Lingua Philosophica.” Studia Leibnitiana Sonderheft 8:
59–​66.
—​—​—​. 1980. “Nominalism and Evander’s Sword”. In Akten des III. Internazionalen
Leibniz-​Kongresses. Studia Leibnitiana, Supplement 21: 213–​25.
—​—​—​. 1986. The Philosophy of Leibniz:  Metaphysics and Philosophy of Language.
New York and Oxford: Oxford University Press.
Michael, Emily, and Frederick S. Michael. 1989. “Two Early Modern Concepts of
Mind:  Reflecting Substance vs. Thinking Substance”. Journal of the History of
Philosophy 27: 29–​48.
Mondadori, Fabrizio. 1990. “Modalities, Representations and Exemplars: The ‘Region
of Ideas.’” In Mathesis Rationis: Festschrift für Heinrich Schepers, edited by Albert
Heinekamp, Wolfgang Lenzen, and Martin Schneider, 153–​67. Münster: Nodus.
3 3 8   •  Works Cited

Moreau, Denis. 1999. Deux cartésiens: La polémique Arnauld Malebranche. Paris: Vrin.


Mugnai, Massimo. 1976. Astrazione e realtà. Saggio su Leibniz. Milan: Feltrinelli.
—​—​—​. 1990. “Leibniz’s Nominalism and the Reality of Ideas in the Mind of God.” In
Mathesis Rationis: Festschrift für Heinrich Schepers, edited by Albert Heinekamp,
Wolfgang Lenzen, and Martin Schneider, 153–​67. Münster: Nodus.
—​—​—​. 1992. Leibniz’s Theory of Relations. Stuttgart: Steiner.
—​—​—​. 2005. “Leibniz on Substance and Changing Properties.” Dialectica 59: 503–​16.
Muirhead, John Henry. 1931. The Platonic Tradition in Anglo-​Saxon Philosophy: Studies
in the History of Idealism in England and America. London: G. Allen and Unwin.
Nadler, Steven. 1989. Arnauld and the Cartesian Philosophy of Ideas. Princeton,
NJ: Princeton University Press.
—​—​—​. 1992. Malebranche and Ideas. New York: Oxford University Press.
—​—​—​. 2006. Spinoza’s Ethics:  An Introduction. Cambridge:  Cambridge University
Press.
—​—​—​. 2008. “Arnauld’s God.” Journal of the History of Philosophy 46: 518–​38.
—​—​—​. 2011. Occasionalism:  Causation among the Cartesians. Oxford:  Oxford
University Press, 2011.
Naragon, Steve. 2006. Kant in the Classroom. http://​www.manchester.edu/​kant/​
Home/​index.htm.
Nelson, Alan. 1993. “Cartesian Actualism in the Leibniz-​Arnauld Correspondence.”
Canadian Journal of Philosophy 23: 675–​94.
Nelson, Alan, and David Cunning. 1999 . “Cognition and Modality in Descartes.” Acta
Philosophica Fennica 64: 137–​53.
Newlands, Samuel. 2015. “Spinoza’s Early Anti-​ Abstractionism.” In The Young
Spinoza:  A  Metaphysician in the Making, edited by Yitzhak Melamed, 255–​71.
Oxford: Oxford University Press.
—​—​—​. 2012. “Thinking, Conceiving, and Idealism in Spinoza.” Archiv für Geschichte
der Philosophie 94, no. 1: 31–​52.
Newman, Andrew. 1992. The Physical Basis of Predication. Cambridge:  Cambridge
University Press.
Nolan, Lawrence. 1997a. “The Ontological Status of Cartesian Natures.” Pacific
Philosophical Quarterly 78: 169–​94.
—​—​—​. 1997b. “Reductionism and Nominalism in Descartes’s Theory of Attributes.”
Topoi 16: 129–​4 0.
—​—​—​. 1998. “Descartes’s Theory of Universals.” Philosophical Studies 89: 161–​80
—​—​—​. 2015. “Cartesian Trialism on Trial: The Conceptualist Account of Descartes’
Human Being.” In Gods and Giants: Papers Presented to Thomas M. Lennon, edited
by Patricia Easton and Kurt Smith, 137–​74. Leiden and Boston: Brill.
Oberst, Michael. 2015. “Kant on Universals.” History of Philosophy Quarterly 32: 335–​52.
Osler, Margaret. 1994. Divine Will and the Mechanical Philosophy. New York: Cambridge
University Press, 1994.
Works Cited  •  339

Owen, David. 1991. “Locke on Real Essence.” History of Philosophy Quarterly 8: 105–​18.
Panaccio, Claude. 2000. “Semantics and Mental Language”. In The Cambridge
Companion to Ockham, edited by Paul Vincent Spade, 53–​75. Cambridge: Cambridge
University Press.
Pappas, George S. 2000. Berkeley’s Thought. Ithaca, NY: Cornell University Press.
Pécharman, Martine. 1992. “Le discours mental selon Hobbes.” Archives de Philosophie
55: 553–​73.
Peters, Richard. 1956. Hobbes. Harmondsworth: Penguin.
Peterschmitt, L. 2009. “L’aspect formel du langage de la physique dans l’Alciphron.”
In, Berkeley’s “Alciphron”:  English Text and Essays in Interpretation, edited
by Laurent Jaffro, Geneviève Brykman, and Claire Schwartz, 414–22.
Hildesheim: Georg Olms.
Pitcher, George. 1977. Berkeley. Arguments of the Philosophers. London: Routledge.
Pippin, Robert B. 1982. Kant’s Theory of Form: An Essay on the “Critique of Pure Reason.”
New Haven, CT: Yale University Press.
Priarolo, Mariangela. 2004. Visioni divine. La teoria della conoscenza di Malebranche tra
Agostino e Descartes. Pisa: ETS.
Price, H. H. 1953. Thinking and Experience. Cambridge, MA: Harvard University Press.
Prinz, Jesse J. 2005. “The Return of Concept Empiricism.” In Handbook of
Categorization in Cognitive Science, edited by Henri Cohen and Claire Lefebvre,
679–​95. Amsterdam: Elsevier.
Putnam, Hilary. 1975. “The Meaning of ‘Meaning.’” In Mind, Language and
Reality: Philosophical Papers, vol. 2, 215–​71. Cambridge: Cambridge University Press.
—​—​—​. 1990. “Is Water Necessarily H2O?” In Realism with a Human Face, edited by
James Conant, 54–​79. Cambridge, MA: Harvard University Press.
Quinton, Anthony. 1973. The Nature of Things. London: Routledge and Kegan Paul.
Rauzy, Jean-​Baptiste. 1993. “Leibniz et les termes abstraits: Un nominalisme par provi-
sion.” Philosophie 39: 108–​28.
—​—​—​. 2001. La doctrine leibnizienne de la vérité. Aspects logiques et ontologiques.
Paris: Vrin.
Regier, Terry, and Paul Kay. 2009. “Language, Thought, and Color: Whorf was Half
Right.” Trends in Cognitive Sciences 13:439–​4 6.
Reid, Jasper. 2003. “Malebranche on Intelligible Extension.” British Journal for the
History of Philosophy 11: 581–​608.
—​—​—​. 2012. The Metaphysics of Henry More. Dordrecht: Springer.
Roberson, Debi, Ian Davies, and Jules Davidoff. 2000. “Colour Categories are not
Universal: Replications and New Evidence from a Stone-​Age Culture.” Journal of
Experimental Psychology: General 129: 369–​98.
Robinet, André. 1965. Système et existence dans l’oeuvre de Malebranche. Paris: Vrin.
Rodis-​Lewis, Geneviève. 1981. “Polémiques sur la création des possibles et sur
l’impossible dans l’école cartésienne.” Studia cartésiana 2: 105–​23.
3 4 0   •  Works Cited

Rodriguez-​ Pereyra, Gonzalo. 2002. Resemblance Nominalism:  A  Solution to the


Problem of Universals. Oxford: Oxford University Press.
Rozemond, Marleen. 2008. “Descartes’s Ontology of the Eternal Truths.” In
Contemporary Perspectives on Early Modern Philosophy:  Essays in Honor of Vere
Chappell, edited by Paul Hoffman, David Owen, and Gideon Yaffe, 41–​ 63.
London: Broadview.
Russell, Bertrand. 1990. Problems of Philosophy. Indianapolis: Hackett. First published
in 1912.
Rutherford, Donald. 1988. “Truth, Predication and the Complete Concept of an
Individual Substance.” In Leibniz: Questions de logique, 130–​144. Studia Leibnitiana
Sonderheft 15. Stuttgart: Steiner.
Schepers, Heinrich. 2014. “Leibniz’ Konzeptualismus. Ein Programm.” In Leibniz: Wege
zu seiner reifen Metaphysik, 227–​231. Berlin: Akademie Verlag.
Schmaltz, Tad M. 1991. “Platonism and Descartes’ View of Immutable Essences.” Archiv
für Geschichte der Philosophie 73: 129–​70.
—​—​—​. 2000. “Malebranche on Ideas and the Vision in God.” In The Cambridge
Companion to Malebranche, edited by Steven Nadler, 59–​86. Cambridge: Cambridge
University Press.
—​—​—​. 2002. Radical Cartesianism:  The French Reception of Descartes. New  York:
Cambridge University Press.
—​—​—​. 2008. Descartes on Causation. New York: Oxford University Press.
—​—​—​. 2011. “Causa Sui and Created Truth in Descartes.” In The Ultimate Why
Question:  Why Is there Anything at All Rather than Nothing Whatsoever?,
edited by John F. Wippel, 109–​24. Washington, DC:  Catholic University of
America Press.
—​—​—​. 2014. “The Fifth Meditation:  Descartes’ Doctrine of True and Immutable
Natures.” In The Cambridge Companion to Descartes’ “Meditations”, edited by David
Cunning, 205–​22. Cambridge: Cambridge University Press.
Scribano, Emanuela. 1996. “Malebranche: Visione di Dio e visione in Dio.” Rivista di
storia della filosofia 3: 519–​54.
—​—​—​. 2003. “Foucher and the Dilemma of Representation: A ‘Modern’ Problem?”
In The Return of Scepticism from Hobbes and Descartes to Bayle, edited by Gianni
Paganini, 197–​212. Dordrecht: Kluwer Academic.
—​—​—​. 2006. Angeli e beati: Modelli di conoscenza da Tommaso a Spinoza. Rome and
Bari: Laterza.
Shockley, Kevin, Claudia Carello, and M. T. Turvey. 2004. “Metamers in the Haptic
Perception of Heaviness and Moveableness.” Attention, Perception, and Psychophysics
66:731–​42.
Shoemaker, Sydney. 1998. “Causal and Metaphysical Necessity.” Pacific Philosophical
Quarterly 79: 59–​77.
Sigwart, Wilhelm. 1904. Logik. 3rd revised ed. 2 vols. Tübingen: Mohr.
Works Cited  •  341

Sider, Theodore. 2007. “Parthood.” Philosophical Review 116: 51–​91.


Smith, Edward E., and Steven A. Sloman. 1994. “Similarity-​versus Rule-​Based
Categorization.” Memory and Cognition 22: 377–​86.
Sosa, Ernest, ed. 1987. Essays on the Philosophy of George Berkeley. Dordrecht: Reidel.
Spade, Paul Vincent, ed. 1994. Five Texts on the Mediaeval Problem of Universals.
Indianapolis and Cambridge: Hackett.
Spruit, Leen. 1995. Species Intelligibilis: From Perception to Knowledge. Vol. 2, Renaissance
Controversies, Later Scholasticism and the Elimination of the Intelligibilis Species in
Modern Philosophy. Leiden: Brill.
Steel, Carlos. 1997. “Breathing Thought:  Proclus on the Innate Knowledge of the
Soul.” In The Perennial Tradition of Neoplatonism, edited by J. J. Cleary, 293–​309.
Leuven: Leuven University Press.
Stern, Carl. 1977. “Kant’s Theory of Empirical Concept Formation.” Southwestern
Journal of Philosophy 8, no. 2: 17–​23.
Stoneham, Tom. 2002. Berkeley’s World:  An Examination of the “Three Dialogues.”
Oxford: Oxford University Press.
—​—​—​. 2005. “Berkeley’s Principles of Human Knowledge.” In Central Works of
Philosophy, edited by John Shand, 2:137–​65. Chesham: Acumen.
—​—​—​. 2013. “Response to Atherton:  No Atheism without Scepticism.” In Debates
in Modern Philosophy: Essential Readings and Contemporary Responses, edited by
Stewart Duncan and Antonia LoLordo, 216–​26. New York: Routledge.
—​—​—​. Forthcoming. “On Being Unreasonable: The Character of Lysicles in Berkeley’s
Alciphron.”
Stoneham, Tom, and Angelo Cei. 2009. “‘Let the Occult Quality Go’:  Interpreting
Berkeley’s Metaphysics of Science.” European Journal of Analytic Philosophy
5: 73–​91.
Suarez, Francis. 1947. On the Various Kinds of Distinctions. Translated by Cyril Vollert.
Milwaukee, WI: Marquette University Press.
Swoyer, Chris. 1995. “Leibnizian Expression.” Journal of the History of Philosophy
33: 65–​99.
Taylor, C. C. W. 1978. “Berkeley’s Theory of Abstract Ideas.” Philosophical Quarterly
28: 97–​115.
Thiel, Udo. 1998. “Individuation.” In The Cambridge History of Seventeenth-​
Century Philosophy, edited by Daniel Garber and Michael Ayers, vol. 1, 212–​62.
Cambridge: Cambridge University Press.
Tipton, I. C. 1974. Berkeley: The Philosophy of Immaterialism. London: Methuen.
Traiger, Saul. 1987. “Impressions, Ideas, and Fictions.” Hume Studies 13:381–​99.
—​—​—​. 2010. “Experience and Testimony in Hume’s Philosophy.” Episteme 7:42–​57.
Trottmann, Christian. 1998. “La vision en Dieu chez Malebranche, Thomas d’Aquin et
Suárez.” In La légèreté de l’être: Études sur Malebranche, edited by Bruno Pinchard,
73–​92. Paris: Vrin.
3 4 2   •  Works Cited

Turbayne, Colin, ed. 1982. Berkeley:  Critical and Interpretive Essays. Minneapolis:
University of Minneapolis Press.
Uchikawa, Keiji, and Hiroyuki Shinoda. 1996. “Influence of Basic Color Categories on
Color Memory Discrimination.” Color Research and Application 21:430–​39.
Urmson, James O. 1982. Berkeley. Oxford: Oxford University Press.
Vanzo, Alberto. 2012. Kant e la formazione dei concetti. Trento: Verifiche.
Vásquez Lobeiras, Maria Jesús. 1998. Die Logik und ihr Spiegelbild:  Das Verhältnis
von formaler und transzendentaler Logik in Kants philosophischer Entwicklung.
Frankfurt: Lang.
Walski, Gregory. 2001. “Descartes’s Doctrine of the Creation of the Eternal Truths.”
PhD dissertation, University of Michigan.
—​—​—​. 2003. “The Cartesian God and the Eternal Truths.” Oxford Studies in Early
Modern Philosophy 1: 23–​4 4.
Warnock, Geoffrey. 1953. Berkeley. London: Pelican.
Watkins, John W. N. 1965. Hobbes’s System of Ideas. London: Hutchinson.
Watson, Richard. 1966. The Downfall of Cartesianism, 1673–​ 1712:  A  Study of
Epistemological Issues in Late 17th Century Cartesianism. The Hague:  Martinus
Nijhoff.
Weinberg, Julius R. 1965. “The Nominalism of Berkeley and Hume.” In Abstraction,
Relation, and Induction:  Three Essays in the History of Thought, 3–60.
Madison: University of Wisconsin Press.
Whitaker, C. W. A. 2002. Aristotle’s “De Interpretatione”: Contradiction and Dialectic.
Oxford: Oxford University Press.
Williams, Donald C. 1953. “The Elements of Being.” Review of Metaphysics 7, no. 2:
3–​18 and 171–​92.
Wilson, Margaret Dauler. 1978. Descartes. London: Routledge and Kegan Paul.
—​—​—​. 1999. “Leibniz and Locke on ‘First Truths.’” In Ideas and Mechanism, 353–​72.
Princeton, NJ: Princeton University Press.
Winkler, Kenneth. 1989. Berkeley: An Interpretation. Oxford: Clarendon Press.
—​—​—​. 2003. “Lockean Logic.” In The Philosophy of John Locke: New Perspectives, edited
by Peter Anstey, 154–​75. London: Routledge.
Wippel, John. 1982a. “The Relationship between Essence and Existence in Late-​
Thirteenth Century Thought:  Giles of Rome, Henry of Ghent, Godfrey of
Fontaines, and James of Viterbo.” In Philosophies of Existence: Ancient and Medieval,
edited by Parviz Morewedge, 131–​6 4. New York: Fordham University Press.
—​—​—​. 1982b. “Essence and Existence.” In The Cambridge History of Later Medieval
Philosophy, Edited by Norman Kretzmann, Anthony Kenny, and Jan Pinborg,
385–​410. Cambridge University Press.
—​—​—​. 2000. The Metaphysical Thought of Thomas Aquinas:  from Finite Being to
Uncreated Being. Washington, DC: Catholic University of America Press.
Works Cited  •  343

Wolfson, Harry. 1934. The Philosophy of Spinoza: Unfolding the Latent Process of His
Reasoning. 2 vols. Cambridge, MA: Harvard University Press.
Yi, Byeong-​Uk. 1999. “Is Mereology Ontologically Innocent?” Philosophical Studies
93: 141–​60.
Zarka, Yves-​Charles. 1985. “Empirisme, nominalisme, et materialisme chez Hobbes.
Archives de philosophie 48:177–​233.
Zehnder, Jean. 1944. Dom François Lamy … essai d’introduction à sa vie et son œuvre.
Freiborg: n.p.
Zuckert, Rachel. 2007. Kant on Beauty and Biology: An Interpretation of the “Critique of
Judgment.” Cambridge: Cambridge University Press.
INDEX

Aaron, Richard  276 n. Atherton, Margaret  270 n.


abstraction  75, 286–​290, 308 Atkinson, Martin  316 n.
abstract terms  210–​215 attention, selective  38
accident  57, 59, 62, 208, 216–​218 attributes  29, 78, 79, 80, 82, 91
Acworth, Richard  188 n., 191 n. Augustine of Hyppo (St.)  2, 5, 92, 98,
Adams, Marylin McCord  98, 127, 129, 131 and n., 134, 144 and
100, 208 n. n., 184, 188, 194 n.
Ainslie, Donald  304 n. Ayers, Michael  222 n., 239 n., 249 n.,
Alciphron 271, 281 288 n., 313
Allison, Henry E.  307 n., 316 n., 322 n.
Angelelli, Ignacio  202 n. Bacon, Francis  154 n.
Angelini, Elisa  165 n. Ballarin, Roberta  304 n.
Anselm of Canterbury  98 Bardout, Jean-​Christophe  155 and n.,
Aquinas, Thomas (St.)  2, 63 n., 156 n., 157 and n.
68 and n., 69, 81 n., 90, 107, 129, Bauer, Nathan  322 n.
130, 131, 139, 157, 158, 159, 160, 161, Baumeister, Friedrich Christian  313 n.
162, 163, 193 n., 194 n., 195 n. Baumgarten, Alexander  311
Ariew, Roger  199 n. Baxter, Donald  5 n., 11, 116 n., 285 n.,
Ariste, 153 298 n., 304 n., 309 n.
Aristotelianism  14, 15, 16, 23 n., Belfrage, Bertil  268, 279 n.
63 and n. Bennett, Jonathan  91 n., 112
Aristotle  2, 19, 20, 21, 25, 26 n., 87, 139, Berger, David  313 n.
154 n., 166, 172 n., 183, 188, 194 Berkeley, George  5 n., 7, 11, 34, 186 n.,
n., 201 and n., 203 n., 208, 211 n., 222, 267 and n., 268 and n., 269,
254 and n., 262, 269 n., 290, 291 n. 270, 271, 272, 273 and n., 274, 275,
Arnauld, Antoine  118, 123, 128, 129 and n., 276 and n., 277 and n., 278, 279 and
130 and n., 131 and n., 132, 133, 136, n., 280, 281, 282, 283 and n., 286
140 and n., 141, 144, 146 n., 148, 150 and n., 288, 290, 291, 309 n., 311
3 4 6   •  Index

Bernhardt, Jean  41, 42 n. concept 13


Berti, Enrico  194 n. c. of color  312–​318
Blair, Eric (true name of empirical c. 307
G. Orwell) 282 formation of c.  307–​311
Bloch, Olivier  14 n., 18 n. conceptualism  4, 7, 10, 51, 63 and n., 64,
Bloodaxe, Eric  282 89, 117, 128, 131, 132, 63 and n.
Blum, Philipp Keller  304 n. Connell, Desmond  157 n.
Boethius  2, 170 and n., 290 and n. Connolly, Patrick  116 n.
Bolton, Martha Brandt  1 n., 5 n., 11, Conrad, Elfriede  308 n.
53 n., 55 n., 116 n., 166 n., 172 n., conventionalism  205–​206, 207, 241
223 n., 231 n., 270 n., 276 n. Cook, Richard S.  315 n.
Bolton, Neil  316 n. Cottingham, John  14 n.
Bosley, Richard N.  294 n. Cottrell, Jonny  304 n.
Boster, James  315 n. Crocker, Robert  174 n.
Bracken, Harry  270 n. Cross, Richard  85 n.
Breger, Herbert  246 n. Crusius, Christian August  311
Bréhier, Émile  88 Cudworth, Ralph  10, 166 and n., 167,
Bretau, Jean-​Louis  184 n. 168 n., 177 and n., 178 and n., 179
Bricke, John  291 n. and n., 180, 181, 182 and n., 183, 184,
Brush, Craig G.  14 n. 185 and n., 186 and n., 187, 188 and
Bucephalus 287 n., 191 n., 192 and n., 195 and n., 196
Burgess, John P.  61 n. Cunning, David  111, 121 n., 142 n.
Burman, Frans  94 n., 101, 109 Curley, Edwin Moses  78 n.
Butler, Annemarie  295 n., 304 n.
Dascal, Marcelo  205 n.
Callaghan, Gerry K.  41, 61, 62 and n. Davidoff, Jules  315 n.
Capozzi, Mirella  308 n., 315 n. Davies, Ian  315 n.
Carello, Claudia  318 n. De Boer, John  177 n., 178 n.
Carpenter, Andrew  307 n., 316 n. definition 230
Carraud, Vincent  128 n. Della Rocca, Michael  188 n.
Carriero, John  66 Democritus 31
Carruthers, Peter  313, 316 n., 321 De Rosa, Raffaella  88 n., 116 n.
Cassirer, Ernst  190 n., 316 n., 319 and n. Descartes, René  4, 5, 7, 8, 9, 10, 18, 22, 23,
Cei, Angelo  281 n. 45 n., 52, 53, 54, 55, 62 and n., 64, 66
Ceyssens, Lucien  130 n. n., 68 and n., 74 and n., 75, 77, 79 and
Chamberlain, Colin  85 n. n., 80 n., 82 n., 87 n., 87, 88 and n.,
Chappell, Vere  88, 112, 119, 89, 90, 91 and n., 92, 93 and n., 94,
123 and n., 142 n. 95, 96, 97 and n., 98, 99, 101 and n.,
Cicero, M. Tully  300 102, 103, 104, 105, 106, 107, 108, 109,
common nature  19 110 and n., 111, 112, 113, 115, 116, 117,
common notions  8, 10, 82, 83, 84 118, 119, 120 and n., 121, 122 and n.,
conceivability 293 123 and n., 124 and n., 125, 126, 128,
Index  •  347

129 and n., 130, 131 and n., 132, 134 Ficino, Marsilio (Ficinus)  31 n., 194 n.
n., 135, 137, 138, 139, 140 and n., 141, Field, Hartry  61 n.
142 and n., 143 n., 144, 146, 148, 150, Flew, Anthony  166 n.
152, 163, 164 n., 168 and n., 170 n., formalism  281–​283
171 and n., 172 and n., 179 and n., Foucher, Simon  125, 126 and n.,
185 and n., 186 and n., 188 and n., 134, 157 n.
205, 211 and n., 291 n. Franklin, Anna  315 n.
Desgabets, Robert  10, 134 and n., 135, Frege, Gottlob  297, 311 n.
136 and n., 137, 138, 139 n. Friedman, Russell L.  197 n.
Di Bella, Stefano  10, 85 n., 116 n.,
165 n., 185 n., 210 n. Gallistel, Charles R.  318 n.
Dionysius (Pseudo-​), Areopagite  158 n., Galluzzo, Gabriele  172 n
159, 195 n. Gärdenfors, Peter  317 n., 318
distinctions (theory of, rational, Garrett, Don  291 n., 294 n., 304 n.
formal) 79, 81, 96 Gassendi, Pierre  8 and n., 9, 13 and n.,
Doney, Willis  270 n. 14, 15 and n., 16 and n., 17, 18, 19,
Doolan, Gregory  161 20, 21, 22, 23, 24, 26, 27, 28, 29, 30,
Dortous de Mairan, Jean-​Jacques  140 n. 32, 33, 34, 35, 36, 37, 38, 39 and n.,
Duncan, Stewart  9, 43 n., 45 n. 68 and n., 185 n., 199 n., 204
Dupré, John  258 n. Gauker, Christopher  317 n.
Gaukroger, Stephen  312
Easton, Patricia  113 Geach, Peter  313 n.
emanation 197 genus  2, 29, 41, 242–​243
Emilsson, Eyjólfur Kjalar  184 n. Gewirth, Alan  88
ens rationis  63, 71, 72, 74, 81 Giles of Rom  107
Epicurus 31 Gilson, Etienne  109
Erickson, Robert P.  318 n. Ginsborg, Hannah  316 n.
essence  7, 23, 88, 114–​115, 118, 135, 138, God
140, 168 God’s knowledge  103–​109
mathematical  112, 118–​119, 138, 186 simplicity of God  91, 98–​103, 193
nominal  8, 229–​230, 232–​233, 239 Goodnick, Liz  85 n.
real  8, 228, 232, 254, 255 Gouhier, Henri  126 n.
eternal truth  22, 23, 24, 180, 240 Grant, Edward  294 n.
creation of  92–​93, 113, 118, 123–​124, Gray, Tom  278
130, 134–​136, 187–​188 Grüne, Stefanie  321 n.
Euclid 217 Gueroult, Martial  78 n., 79 n., 84
Euler, Leonhard  313 n. n., 119, 150 n., 155 and n., 157,
extension, intelligible  136, 150–​152 163 and n., 164

Favaretti Camposampiero, Hacking, Ian  43 n.


Matteo 157 n. Hakkarainen, Jani  304 n.
Faye, Emmanuel  130 n. Hanna, Robert  311 n., 322 n.
3 4 8   •  Index

Hark, Michel ter  313 n. individual 155, 170


Harnard, Stevan  310, 317, 320 individuation  169, 172, 199–​200
Haserot, Francis S.  78 n. Istvan, Michael  66 n., 85 n.
Hazard, Paul  143 n.
Held, Carsten  319 n. Jesseph, Douglas  41, 56 n., 60 n., 61
Heller, Edmund  322 n. Jessop, Thomas Edmund  277 n.,
Henning, Hans  318 and n., 319 279 n., 283
Hobbes, Thomas  6, 7, 8, 9, 10, 41, Johanson, Donald  278
42 and n., 43 and n., 44 and n., 45, Jolley, Nicholas,  87 n., 116 n., 127 n.,
46 and n., 47, 48, 49, 50, 51, 52, 54, 235 n., 239 n.
55 and n., 56 and n., 57, 58, 59 n., Jones, Howard  14 n.
61, 62 and n., 64 and n., 66 n., 68
and n., 70 n., 77, 172, 173, 174 and Kalar, Brent  316 n., 322 n.
n., 179 and n., 183, 184 n., 200 n., Kant, Immanuel  9, 305, 306, 307,
204 and n., 205, 206, 208, 209, 210, 308 and n., 309 and n., 310 n., 311
211 and n., 212, 214, 268 n., 311 and n., 312, 313 and n., 315, 316, 317
Hooke, Robert  254 and n., 319 and n., 320 and n., 321
Hübener, Kurt  53 n. and n., 322 and n.
Huet, Pierre-​Daniel  132 Kay, Paul  315 n.
Hume, David  7, 11, 142, 268, 285, 286, Keckermann, Bartholomäus  170 n.
287, 288, 290, 291, 292, 293 and n., Kenny, Anthony  88 and n., 119, 142 n.
294 and n., 295 and n., 296, 297, kind, natural  13, 76, 207, 225–​226, 232,
298, 299, 301 and n., 302 and n., 235, 257–​266
303 and n., 304 and n. King, Peter  100
Huygens, Gommaire  130 and n. Kitcher, Patricia  311 n.
Hylas 267 Knoblauch, Kenneth  314 n., 318
hylomorphism 4, 107 Kripke, Saul  11, 253, 259 and n., 260,
261, 262, 264, 265, 266, 279 n.
idea Kristeller, Paul Oskar  194 n.
abstract  18, 25, 28, 34, 35, 51 n., 148,
222, 268, 270–​271 Laird, John  41, 45 n., 51 n.,
aggregate  18, 25, 28, 34, 35 291 n., 293 n.
divine  103–​109, 118, 159–​162, 180, Lamy, François  131 and n.
188–​191, 196 Landy, David  321 n.
general  13, 17, 153–​154, 220, 222, LaPorte, Joseph  260 n.
224–​225, 269–​270 La Rocca, Claudio  309 n., 310 n., 322 n.
innate  5, 110, 120, 172–​173 Lascano, Marcy  85 n.
particular  13, 152, 155, 164 LeBuffe, Michael  84 n.
identity  285, 295–​301 Leduc, Christian  239 n.
image, mental  27, 35, 44, 45, 49, 51 Leibniz, Gottfried Wilhelm  6, 8, 10, 11,
immaterialism 270 45 n., 52, 53 and n., 56, 64 and n.,
Index  •  349

198 and n., 199 and n., 200 and n., Malebranche, Nicolas  6, 10, 90, 118, 124,
201 and n., 202 and n., 203 and 125, 126 and n., 127 and n., 128,
n., 204 and n., 205, 206 and n., 129, 130, 131, 132, 133, 134 and n.,
207 and n., 208, 209, 210, 212, 213 136, 139, 140 and n., 141, 142,
and n., 214, 215, 216, 217, 218 and n., 143 and n., 144, 145 and n., 146
223, 224, 225, 229, 232, 235, 236, and n., 147 and n., 148, 149, 145
237, 238, 239, 240, 241, 242, 243, and n., 146, 149, 150, 151, 152,
244, 245, 246, 247, 249, 250, 153, 154 and n., 156, 157 and n.,
311, 316 n. 158, 162, 163 and n., 164 and n.,
Lelevel, Henri de  127 and n., 128, 140 190, 191, 192 and n., 194 and n.,
Lennon, Thomas  15 and n., 138 n. 195 and n., 197, 268 n.
Levey, Samuel  241 n. Mander, William  188 n., 189 n., 190 n.,
Liu, Cathay  116 n. 192 n., 194 n., 195 n.
Locke, John  4, 5 and n., 8, 9, 10, 11, Margolis, Joseph Z.  270 n.
51 n., 171 n., 190, 191, 207 and n., Martinich, Aloysius,  43 n.
220, 221, 222, 224, 225, 226, materialism 45 n., 73
227, 228, 229 and n., 230, 231, Mates, Benson  201 n.
232, 234, 235, 236, 237, 238, 239, Meier, Georg Friedrich  309 n., 311, 312
240, 243, 244, 245, 246, 247, 248, Mersenne, Marin  45 n., 46, 92, 93,
249, 250, 252, 253, 254, 255, 105 and n., 106, 110, 188 and n.
256 and n., 257, 258 and n., 259, Meyer, Lodewijk  74 n.
260, 261, 262, 264, 265, 266, mode  96, 136–​137, 248
268 and n., 269 and n., 270, moderate realism  16, 82, 210
272 and n., 286, 287, 288, 290 and Mondadori, Fabrizio  201 n.
n., 290, 291, 309 n., 311 and n., 312, monster  233, 237, 256
313, 316 More, Henry  10, 53, 55, 56, 166 and n.,
LoLordo, Antonia  9, 11, 13, 167, 168 and n., 169, 170 and n., 171
116 n., 304 n. and n., 172, 173 and n., 174 and n.,
Longuenesse, Béatrice  309 n., 322 n. 175 and n., 176, 189 n.
Lotti, Brunello  10, 31 n., 185 n. Moreau, Denis  128 n., 130 n., 144 n.
Lovejoy, Arthur O.  166 n., 256 n. Mugnai, Massimo  201 n., 217 n., 223 n.
Lowe, Jonathan  11, 254 n., 258 n. Muirhead, John Henry  166 n., 185 n.
Luce, Arthur Aston  277 n., 279 n., 283 Murdoch, Dugald  14 n.
Lucy 278
Lyssy, Ansgar  310 n. Nadler, Steven  78, 127 n., 128 n., 144 n.,
145 n., 148 n.
MacBride, Frazer  62 n. name, general  44, 47, 48, 49, 50, 54,
Mackie, John  224, 288 n. 58, 221
MacKinnon, Flora Isabel  188 n., Napoletano, Toby  304 n.
191 n., 195 n. Naragon, Steve  308 n.
Malcolm, Noel  46 nature. See essence
3 5 0   •  Index

Nelson, Alan  93 n., 111, 116 n. Philonous 267


Newlands, Samuel  62 n., 77 n., 80 n., Philo of Alexandria  184, 194 n.
81 n., 116 n., 304 n. Pippin, Robert B.  316 n.
Newman, Andrew  313 Pitcher, George  273 n., 276 n., 285 n.
Newman, Lex  116 n. Plato  1, 17, 19, 20, 25, 26, 30 and n.,
Newton, Isaac  266 67 n., 70, 72, 87, 92, 127, 131, 161,
Nicole, Pierre  131, 148 166 n., 183 and n., 188, 191 n., 194
Nizolius, Marius (Mario n., 201, 244, 277, 290
Nizolio)  53 and n., 198 and n., 199, Platonism  2, 3, 5, 6, 10, 63 and n., 73, 87,
200, 201, 202 and n., 203 and n., 88 and n., 117, 124, 131, 139, 176, 200
204, 207, 208, 209 and n., 210 English (also: Cambridge)
noéma  183–​186, 195 Platonism  6, 10, 166 and n.
Nolan, Lawrence  10, 88, 93 n., 95 n., 97 Theological (also: Christian)
n., 101 n., 105, 110, 115, 119, 120, 121 Platonism  89, 94 n., 99, 101, 102,
n., 122 n., 123 n., 124 n., 304 n. 105, 111, 112, 180, 182, 184, 201
nominalism  3, 6, 13, 14, 15, 16, 17, 41, Plotinus  168 n., 184 n., 194 n.
42, 45 n., 52, 60 and n., 63 and n., Porphyry  2, 43 n., 46, 194 n., 290 n.
178, 198, 200 possibilia 203
resemblance  218 n., 252 Powell, Lewis  304 n.
Norris, John  10, 166 and n., 167, predication  213–​215
178 and n., 188, 189 and n., Priarolo, Mariangela  10, 116 n., 127 n.,
190 and n., 191 and n., 192 and n., 143 n., 157 n.
193 and n., 194 and n., 195 and n., Price, Henry H.,  61
196 and n., 197 Putnam, Hillary  11, 253, 259 and n.,
260, 261, 262, 263, 264, 265, 266
Ockham, William  3, 7, 42 and n.,
43 and n., 51 n., 53, 61, 63 n., 89, 98, Quinton, Anthony  270 n.
100, 198 n., 199 n., 200, 207 n., 208
n., 216, 268 n. Raho, Joseph  165 n.
Osler, Margaret  16 and n. Rauzy, Jean-​Baptiste  202 n., 210 n.
Owen, David  116 n., 228 Regier, Terry  315 n.
Régis, Pierre-​Sylvain  10, 113, 118 and n.,
Panaccio, Claude,  42 124, 127 n., 132 and n., 133, 134 and
Pappas, George  270 n. n., 137, 138 and n., 139 and n.
Payne, Robert  46 Reid, Jasper  170 n.
Pearce, Kenneth  269 n. Reid, Thomas  3
Pécharman, Martine  43 n. representation 306
Peters, Richard  41 resemblance.  See similarity
Peterschmitt, Luc  280, 281 n., 282 Roberson, Debi  315 n.
phantasm  6, 37, 52, 57, 58 and n., 59, 68. Robinet, André  144, 150 n.,
See also image, mental 155 and n., 157
Index  •  351

Rochot, Bernard  14 n. Spangler, Al  116, n.


Rodis-​Lewis, Geneviève  118 n., species  42, 233–​234, 235–​238
134 n., 154 n. Spinoza, Baruch  8, 9, 62 and n., 63,
Rodriguez-​Pereyra, Gonzalo  61 n., 62 n. 64 and n., 65 and n., 66 and n., 67
Roscelin 198 n. and n., 68 and n., 69 and n., 70, 71,
Rosen, Gideon  61 n. 73, 74 and n., 75, 76, 77 and n., 78,
Rozemond, Marleen  88, 89, 99, 100, 79 and n., 80 and n., 81 and n., 82
101, 102, 103, 105, 106, 107, 109, and n., 83 and n., 84, 85, 138 n.
123, 124 n. Spruit, Leen  146 n.
Russell, Bertrand  62 n., 94 n., 114 n. Steel, Carlos  172 n.
Rust, George  175, 189 n., 195 n. Stencil, Eric  85 n.
Rutherford, Donald  210 n. Stern, Carl  322 n.
Ryan, Todd  304 n. Steuchus (Steuco), Agostino  191 n.
Stoneham, Thomas  5 n., 11, 116 n., 271,
Schepers, Heinrich  207 n. 272, 274, 281 n., 304 n.
Schmaltz, Tad  10, 79 n., 85 n., 88 and n., Stoothoff, Robert  14 n.
105, 114 and n., 116 n., 120 n., 123, Suárez, Francisco  14, 63 n., 64, 65 and n.,
127 n., 136 n., 139 n..165 n., 304 n. 66 and n., 79 n., 80, 81 and n., 107,
Scotus, Duns  63 n., 88, 99, 100, 101, 107 108 and n., 109, 172 and n., 196 and
Scribano, Emanuela  143 n., 146 n., n., 199, 204 and n., 291 n.
148 n., 157 n., 165 n. substance 79
semiotic theory of general secondary  171–​172
terms  271–​280 simplicity of  98, 137
Seneca 215 Swoyer, Chris  232
separability 222
mental 291, 293 Taylor, C. C. W.  288 n.
real 291, 294 Tempier, Etienne  294
Shinoda, Hiroyuki  314 n. Theodore 153
Shirley, Samuel  68 n. Theseus’ ship  217
Shockley, Kevin  318 n. Thiel, Udo  172 n.
Shoemaker, Sydney  262 n. Thomasius, Jakob  199 and n., 201
Sider, Theodore  298 n. Tipton, I. C.  270 n.
Sigwart, Wilhelm  316 n. Traiger, Saul  301 n., 304 n.
similarity  17, 19, 20, 28, 30, 57, 60, 65, transcendentals (transcendental
68, 69, 207, 229, 252 concepts) 76
similarity spaces  317–​318 tropes  66, 67, 215–​216
Sloman, Steven A.  320 n. Trottmann, Christian  157 n.
Smith, Edward E.  320 n. Turvey, Michael T.  318 n.
Socrates  17, 25, 26, 30 and n., 46, 58, Tweedale, Martin  294 n.
214, 276, 277 Twin Earth, thought experiment  260,
Spade, Paul Vincent  290 n. 263–​264
3 5 2   •  Index

Uchikawa, Keiji  314 n. Wells, Aaron  85 n.


ultranominalism Wells, Norman  108 and n.
(or: supernominalism)  9, 41, 52, Whitaker, C. W.  269 n.
204–​206 Williams, Donald Cary  66 and n.
universal knowledge  203–​204, 246–​249 Wilson, Margaret  123, 245 n.
Urmson, James O.  288 n. Winkler, Kenneth  245 n., 285 n., 288 n.
Wippel, John  107, 160 n.
van den Sanden, Henri  130 Wittgenstein, Ludwig  279 n.
Vanzo, Alberto  11, 307 n., 315 n. Wolff, Christian  310 n., 311 and n.
Vásquez Lobeiras, Maria Wolfson, Harry  78 n., 80, 84 n.
Jesús 308, 322 n. Xanthippe 275, 277
vision in God  127, 130, 143, 149, 159
Yi, Byeong-​Uk  298 n.
Walski, Gregory  93 n.
Warnock, Geoffrey  270 n. Zarka, Yves-​Charles  45 n.
Watkins, John W. N.  41, 59 n. Zehnder, Jean  131 n.
Watson, Richard  126 n. Zeno (of Elea)  73
Weinberg, Julius  285 n. Zuckert, Rachel  322 n.

You might also like