Perović

Ontology – the study of the most fundamental categories of

being – lies at the very heart of metaphysics. The reason why
it appears to be so central is because it takes on the following
questions: What sorts of entities are there? What features
do those entities have? How do they relate to one another?
And so on. Section 1 of this Element presents a fast-paced Metaphysics
historical overview of some of the notable approaches to
these questions. Section 2 tells the story of how one of the
oldest, most disputed, but also most developed ontological
categories – universals – got introduced. Section 3 builds on
the discussion of universals as it considers the desiderata for
a promising system of ontological categories. And Section 4

Ontological Categories
Ontological
looks at ways in which philosophers might break with tradition
and explore some new ontological categories.

Categories
About the Series Series Editor
This highly accessible series of Elements Tuomas E. Tahko
provides brief but comprehensive University of Bristol
introductions to the most central topics
in metaphysics. Many of the Elements

Katarina Perović
also go into considerable depth, so the
series will appeal to both students and

https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

academics. Some Elements bridge the
gaps between metaphysics, philosophy
of science, and epistemology.

Cover image: peterkai / iStock /

Getty Images Plus ISSN 2633-9862 (online)
ISSN 2633-9854 (print)
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press
 Elements in Metaphysics
edited by
Tuomas E. Tahko
University of Bristol

ONTOLOGICAL
CATEGORIES

A Methodological Guide

Katarina Perović
University of Iowa
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press
 Shaftesbury Road, Cambridge CB2 8EA, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467

Cambridge University Press is part of Cambridge University Press & Assessment,

a department of the University of Cambridge.
We share the University’s mission to contribute to society through the pursuit of
education, learning and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781009475679
DOI: 10.1017/9781108973861
© Katarina Perović 2024
This publication is in copyright. Subject to statutory exception and to the provisions
of relevant collective licensing agreements, no reproduction of any part may take
place without the written permission of Cambridge University Press & Assessment.
First published 2024
A catalogue record for this publication is available from the British Library
ISBN 978-1-009-47567-9 Hardback
ISBN 978-1-108-97825-5 Paperback
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

ISSN 2633-9862 (online)

ISSN 2633-9854 (print)
Cambridge University Press & Assessment has no responsibility for the persistence
or accuracy of URLs for external or third-party internet websites referred to in this
publication and does not guarantee that any content on such websites is, or will
remain, accurate or appropriate.
 Ontological Categories

A Methodological Guide

Elements in Metaphysics

DOI: 10.1017/9781108973861
First published online: January 2024

Katarina Perović
University of Iowa
Author for correspondence: Katarina Perović, [email protected]

Abstract: Ontology – the study of the most fundamental categories of

being – lies at the very heart of metaphysics. The reason why it appears
to be so central is because it takes on the following questions: What
sorts of entities are there? What features do those entities have? How
do they relate to one another? And so on. Section 1 of this Element
presents a fast-paced historical overview of some of the notable
approaches to these questions. Section 2 tells the story of how one of
the oldest, most disputed, but also most developed ontological
categories – universals – got introduced. Section 3 builds on the
discussion of universals as it considers the desiderata for a promising
system of ontological categories. And Section 4 looks at ways in which
philosophers might break with tradition and explore some new
ontological categories.

This Element also has a video abstract: www.cambridge.org/Perović

https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Keywords: ontology, categories, entities, metaphysics, universals

© Katarina Perović 2024

ISBNs: 9781009475679 (HB), 9781108978255 (PB), 9781108973861 (OC)
ISSNs: 2633-9862 (online), 2633-9854 (print)
 Contents

Introduction 1

1 A Brief Historical Sketch 4

2 Introducing an Ontological Category: A Case Study

in Universals 10

3 Some Desiderata for a Promising System of Ontological

Categories 28

4 On the Value of Exploring New Types of Ontological

Categories 53

References 68
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press
 Ontological Categories 1

Introduction
Ontology – the study of the most fundamental categories of being – lies at the
very heart of metaphysics. The reason why it appears to be so central is because
it takes on the following questions: What sorts of entities are there? What
features do those entities have? How do they relate to one another? And so
on. The first question is both general and assuming. It is general in a sense that is
not to be answered by listing all the things that one can think of, for it asks about
the sorts of entities that there are, not just the sheer number of them. It is also an
assuming question insofar as it takes for granted that the world is not empty but
filled with entities and that these entities can in fact be categorized by appeal to
their distinctive features and roles.
So how might one go about addressing such vast questions concerning the
categories of being? In Section 1, I present a fast-paced historical overview of
some of the notable approaches. I start with the naïve linguistic realism of Plato
and Aristotle, who took natural language to be a reliable guide to the fundamental
categories of being. Then I briefly discuss the three main skeptical attitudes to this
form of ontological realism – Kantian skepticism, which questions whether the
categories of being are indeed part of the structure of the world or merely part of
our conceptual framework; skepticism about the types and numbers of categories;
and skepticism about natural language as a guide to what there is. These forms of
skepticism are still very much alive in contemporary approaches to metaphysics,
as is the ontological realism that resists them.
Section 2 tells the story of how one of the oldest, most disputed, but also most
developed ontological categories – universals – got introduced. I present the
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

famous “one-over-many” argument for universals, as the argument was presented

by Plato and then picked up and amplified by Aristotle. The goal of the section is
to illustrate, through a case study of universals, what making a case for a category
looks like, as well as to show the points at which other philosophers might
disagree. One of the most prominent contemporary realists about universals –
David Armstrong –believed that his version of “one over many” only offered
a “preliminary case” for universals, not a decisive one. Building on this prelimin-
ary case, he added further arguments – the argument from laws of nature and the
so-called negative arguments from Russell against the denial of relational univer-
sals; and from Pap and Jackson, from the lack of adequate paraphrases of apparent
references to universals. In this way, Armstrong believed himself to be making
a more convincing, cumulative, case for universals.
In Section 3, I draw on the example of the category of universals from
Section 2 as I discuss the methodological desiderata for building a promising
system of ontological categories. The first desideratum is to make sure that the
 2 Metaphysics

metaphysical question or problem that the ontological entity is introduced to

address is well-stated and motivated. This requires taking special care to avoid
engaging with metaphysical pseudo-problems. I understand a metaphysical
pseudo-problem in a particular way: it is the sort of problem that in its statement
commits the fallacy of a complex question. This type of fallacy is committed
when, in stating the question or problem, one tacitly takes for granted contro-
versial assumptions that are not shared by all sides of the debate, or assumptions
that are simply not true, or that indeed presuppose a false dichotomy. The result
of engaging with a question so posed is that answers/solutions are either false or
severely restricted by those same tacit assumptions implied by the question.
When evaluating a given ontology and the categories it introduces, philo-
sophers tend to consider ontological and ideological economy. I discuss Quine’s
famous distinction between ontology (the entities that a theory claims exist) and
ideology (the concepts and predicates that theoretical explanations rely upon)
and the challenges one faces when attempting to measure them against one
another. I also discuss the important distinction between simple and more
complex ontological entities, and the widely held but questionable assumption
that fundamentality and simplicity ought to go hand in hand in a metaphysical
explanation.
I then move on to the meta-ontological discussion of the relationship between
ontological entities, categories, and ontological roles. Following Oliver (1996),
I suggest that the term “ontological category” should be reserved to pick out
a particular ontological role that a philosopher finds in need of being filled in
their theory, as they address a particular philosophical problem. Ontological
entities, on the other hand, can fill one or multiple such roles, and thus can fall
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

under one or more distinct ontological categories. This, I argue, tracks nicely
what metaphysicians already tend to do – that is, it tracks how they assign
distinct roles to distinct entities in their system and how they then employ such
entities in different places in their overall metaphysical system. Of course, not
everyone approaches ontological categories in this way; I conclude the section
by contrasting this kind of bottom-up approach to ontological categories with
descriptive and prescriptive top-down approaches.
Finally, in Section 4, I look at ways in which philosophers might break with
some traditional category divisions and explore new ones. The search for the
“correct” ontological categories, and the redrawing of different categorical
divisions, carries a certain weight only if one assumes some form of realism
about ontological categories. This is why I start the section by examining the
circumscribed way that the discussion between an ontological realist and an
ontological anti-realist about categories frequently plays out, and I find specific
aspects of the dialectical exchange wanting. The anti-realist attacks tend to
 Ontological Categories 3

trivialize and misrepresent the endeavor that ontological realists are engaged in,
while the realists often end up alienating more scientifically minded philo-
sophers with their insistence that they are describing ultimate, immutable, and
metaphysically necessary features of reality. I suggest in place of such a form of
ontological realism, a more modest, cautious, revisionary form which might
accommodate some of the anti-realist’s complaints, without, however, giving
up on the realist’s overarching goals. I discuss the importance of trying to
become aware of our metaphysical “blind spots” and the value of exploring
new, nontraditional ontological categories and category systems. I argue that
part of what constitutes progress in ontological inquiry has to do with opening
ourselves up to different ways of conceiving of the familiar. Exploring new
ontological entities and systems of categories is integral to that goal. To this end,
I consider the example of Whitehead’s original temporal ontology of events and
objects, and the more recent approaches to gender categories as discussed by
Barnes (2019).
From the description given so far, the reader will see that the approach of this
Element is primarily methodological – it is a sort of “how to” guide to introdu-
cing ontological categories. My goal has not been to propose a single preferred
system of ontological categories; such a project requires a more wide-ranging
and in-depth discussion of a number of metaphysical problems, as well as
a comparative analysis of different ontological systems. My goal has also not
been to systematically describe the most commonly invoked ontological cat-
egories in metaphysical discussions, though a number of such categories do
make an appearance. And my goal has certainly not been to engage in a meta-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

ontological study of the various ways that philosophers have employed, or

ought to have employed, the concept of an “ontological category.”
Instead, I have tried to illustrate, in some detail, by reference to the case study
of universals, a number of concerns that are pertinent to metaphysical theory-
building as one makes a case for a particular system of ontological categories.
I have deliberately chosen to move between examining an ontological discus-
sion at base level, and then looking at the parameters of the debate on a meta-
level, because I think it is instructive. I firmly believe that taking the time to pay
close attention to the main parameters of any given debate is crucial; and I find
this is to be particularly important (though quite challenging) in fundamental
ontology. By revealing and examining the underlying assumptions and param-
eters that are often taken for granted we start to become aware of the possibil-
ities we weren’t necessarily aware of before. There is also a value in providing
illustrative examples from ground-level debates when discussing meta-
theoretical issues. Too often in ontological discussions, we trail off into abstract
discussion in which examples are hard to come by and the theoretical claims
 4 Metaphysics

become esoteric and increasingly difficult to illustrate. Taking the time to

explain, motivate, and provide examples is one way that we can keep our
debates healthy and our metaphysical discussions intellectually honest. I have
tried to do this here, but I assume that I too have failed in some places.
Of course, this methodological exploration of ontological categories is itself
not neutral; it is carried out from a particular point of view, which makes it an
opinionated one. I firmly believe that ontology is central to philosophy; that
ontological categories are not merely features of our conceptual framework; and
that universals are a genuine ontological category. I am thus a realist in more
than one sense.

1 A Brief Historical Sketch

A historically informed approach to contemporary ontological debates is
important for several reasons: it helps us understand better at the outset the
historical context in which certain arguments and theories were developed and
the assumptions that such a historical context might have left unexamined; it
helps us understand those arguments in a contemporary setting, with new
improvements, distinctions, and clarifications; in many circumstances, it helps
us to build on the achievements of our predecessors rather than having to
reinvent the wheel; and finally, one would hope, it helps guard us against
repeating the mistakes of the past. In this section, due to space constraints,
I can only provide a rough, bird’s-eye view of the traditional realist approach to
ontological categories and the types of skepticism that it gave rise to. It is
noteworthy that any proposed list of categories immediately faces questions
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

about method (the way that such a list is arrived at) and accuracy (whether it
adequately captures the true divisions in the world).

1.1 Naïve Realism about Ontological Categories

It seems fair to say that the “traditional approach” to ontological categories origin-
ated with Plato and Aristotle, and that it carried a decidedly realist stamp. That is,
these philosophers thought of fundamental categories of being as real mind-
independent entities. Ancient philosophers were inclined to rely on natural language
as a reliable guide to what there is. According to this view, there is a straightforward
matchup between sentences such as “Xanthippe is wise” and what such sentences
are about, namely, the particular woman – Xanthippe – and the property of wisdom
she happens to have. The assumption of a straightforward matchup between our
linguistic descriptions of reality and the ontological structure of reality is crucial for
this outlook. Plato believed that an understanding of the world around us involves
providing definitions of the true nature of things, and that this requires us to place
 Ontological Categories 5

those things in the correct categories, that is, under the correct Forms. Getting this
right is essential, for it is the only way to find the real divisions in nature, it is the
only way that one can, as Plato famously put it, “carve nature at its joints.”
Aristotle shared this general outlook with Plato, though he disagreed with the
Platonic understanding of universals as independently existing Forms.
Universals, for Aristotle, were not to be conceived in a Platonic way, as existing
independently from particulars in a realm of their own; rather, he thought of the
two types of entities as mutually dependent on one another. Aristotle’s grouping
of entities also involves a much more diverse lot than just particulars and Forms;
he lists the following ten as categories: substance, quantity, quality, relation,
place, date, posture, state, action, and passion. This list was meant to be
exhaustive and of the highest level of generality – for Aristotle, there was no
further higher category that all of these categories could be subsumed under. But
regardless of whether or not we find Aristotelian categories compelling today,
what is important to recognize is that both Aristotle and Plato assumed a certain
naïve form of realism about the categories: that is, they assumed that our
categorizations of entities picked out genuine features of the outside world.
We can immediately see one difficulty that naïve realist assumptions give rise
to: even a surface level comparison between Plato’s and Aristotle’s categoriza-
tions makes clear the discrepancy between what are arguably the most funda-
mental mind-independent features of reality. The two-category ontology of
particulars and Forms is clearly quite different from the ten-category ontology
described by Aristotle. What are we to make of such clashes between the
category systems within the realist framework? One possibility is to argue
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

that one philosopher’s category system is the correct one, while the other
philosopher is simply wrong. Another possibility is to try reconciling the two
systems of categories by showing that the more numerous category system is
just an elaboration on the two-category one, and that it might be subsumed
under it, with some modifications. A further possibility still is to claim that they
are both wrong and that some other category system is the correct one. But the
central problem underlying the disagreement is one concerning whether they
have even employed the same method in arriving at their categorizations.
Indeed, how does one even go about deciding which categories there are?
Interestingly, there has been significant controversy about what guided Aristotle
in making the particular categorizations that he did, since he did not reveal his
reasoning. Studtmann (2021) highlights four dominant interpretative
approaches to Aristotelian categories – the question approach; the grammatical
approach; the modal approach; and the medieval derivational approach. I want
to draw attention only to the first two, since they seem to me to be the most likely
interpretations and the most relevant to this discussion. The grammatical
 6 Metaphysics

approach, as advocated by Baumer (1993), focuses on Aristotle’s reliance on

natural language and its inherent grammatical structure as a guide to categoriz-
ing entities in the world. The question approach, on the other hand, is promoted
by Ackrill (1963) and it focuses on the questions one might ask about an entity.
Take, for instance, a clever nine-year-old named Lulu; one can ask many
questions about her: “What is Lulu?” could be answered with “Lulu is a nine-
year-old human.” Then one might ask, “What is a nine-year-old human?,” to
which one might answer, “it is an animal with particular sorts of features.”
Further questioning could then lead us to higher and higher categorizations, and
ultimately to substance, as it was understood within Aristotle’s framework. If
we ask such questions about different sorts of beings in the world, we will,
according to Ackrill, get the ten Aristotelian most general categories. The
alternative line of questioning described by Ackrill does not press on with the
same “What is X?” sort of question; rather it asks different specificatory
questions about the X. So, for example, these questions could take the following
shape: “What is Lulu?,” “Where is Lulu?,” “How old is Lulu?,” “What is Lulu
doing?,” “How tall is Lulu?,” and so on, until – again – presumably one arrives
at the ten most fundamental Aristotelian categories.
I am not interested here in whether or not either of these approaches actually
leads neatly to the Aristotelian list of categories. I merely want to point out that
though at first glance it might seem correct to say that a philosopher might arrive at
a list of categories by asking a number of questions about various entities, such
a method is too haphazard. Take a nine-year-old from the example just given and
imagine her performing a sorting exercise in a room surrounded by a variety of
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

objects. How might she sort things? She might decide to sort them by color, shape,
function, or she might decide to sort them by material that they are made out of, or
she might sort them by whether or not they make a sound, or she might sort them
by how pretty she finds them, or whether she finds them interesting or not, and so
on and so forth. Clearly, there are a great many different ways that humans can sort
things into groups and there are a great many different questions one can ask about
any given thing and any given assortment of things. The search for ontological
categories cannot be guided by such an unconstrained process of inquiry.
Thus, it should be clear that not just any type of question regarding entities
will be conducive to finding the most fundamental categories of being. Even if
one were to replace a child in the example discussed with Aristotle himself, the
category system that the latter came up with would not be a better one just
because it is devised by a famous philosopher rather than a child.1 Rather, what
1
A quick application of a version of the Euthyphro dilemma demonstrates this. If we ask, “What
makes Aristotle’s system of categories the correct one?,” it would not be good to answer, “It is the
correct one because it is chosen by Aristotle.” Hopefully, it is a better system of categories
 Ontological Categories 7

will make one system of categories better than another will have to do, at least in
part, with the types of questions being asked about the structural features of the
world. The questions need to be well-defined and the problems that they raise
need to be well-articulated and motivated. It is only when metaphysicians
postulate entities with specified explanatory roles that it becomes possible to
compare the different types of entities and categories that philosophers employ
in their different explanations. Without a certain amount of agreement on the
philosophical problems that need to be addressed, and the type of methodology
for solving such problems that should be employed, fundamental ontology risks
looking like a solitary and idle endeavor, just one person’s way of sorting things
against another.

1.2 Three Forms of Skepticism about Categories

From the previous subsection, we can see that it is no surprise that a naïve realist
approach to categories has historically suffered attacks from different direc-
tions. We can roughly distinguish three: (i) skepticism about whether the
categories being described are indeed part of the structure of the world or are
merely part of our conceptual framework with which we approach it;
(ii) concern that the categorizations being taken for granted are wrong (perhaps
it is not two types of entities that we should admit, but only one, or three, or four,
or more); and (iii) questioning of our reliance on natural language as a guide to
what there is, and efforts to provide better logical analyses of language as a way
of getting to the correct ontological structure of reality.
Skepticism of the first kind goes back to Kant’s Critique of Pure Reason, and it
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

has had significant philosophical repercussions: on the one hand, we see that the
worry that there is no way of arriving at genuine mind-independent categories has
led philosophers to set their goals to more modestly defined descriptivist
approaches to metaphysical categories, where the aim is to describe the concepts
we use to understand the world, not the categories of the mind-independent world
itself. Husserl’s phenomenology and Strawson’s descriptivist approach to meta-
physics are two notable examples of this sort of attitude. On the other hand, the
same Kantian worry has led some philosophers toward an idealism which takes
the external world to be unknowable and the categories used to describe such
a world as completely misleading and even contradictory. In this camp, we find
British idealists such as F. H. Bradley developing vicious regress arguments
against the possibility of external relations in an attempt to show that ontological
pluralism is false and that ontological monism – a commitment to the existence of

because it describes the structure of reality more accurately, where the accuracy conditions are
specified independently.
 8 Metaphysics

only one Absolute whole – must be the only true view of reality. Here too, we find
McTaggart’s notorious dismissal of the mind-independent existence of temporal
series, based on his quick reductio arguments against the A-series and the B-series
of time. Also going along with the idealist worldview we find Joachim’s defense
of the coherentist theory of truth with its obscure and mystifying reference to the
indivisible Truth that only applies to the Absolute whole.
Skepticism of the second type mentioned – the one that is aimed at the number
and types of categories introduced – is present in ancient Greek philosophy, and is
even more visible in the Middle Ages with respect to the ancients’ particular–
universal distinction. For instance, Aquinas, Duns Scotus, Ockham, and others
are often seen as early supporters of the one-category trope ontology. Building on
disagreements concerning the type and the number of categories introduced, one
can quickly arrive at skepticism concerning the method – whether there can be
a right way at arriving at the fundamental categories at all.
One can then see how the two types of skepticism just described influenced
the third type, which takes aim at natural language as a guide to ontological
categories. We see this very clearly in early Russell, with his break from
idealists and commitment to the careful logical analysis of language as
a guide to the correct logical structure of the world. Note that this third type
of skepticism is quite limited – it does not renounce the project of discovering
the correct ontological categories altogether. Rather, it merely corrects the
route; it requires that the project be carried out through careful logical analysis.
Early Wittgenstein in Tractatus, at least in parts, seemed to be on this track too.
In other parts, of course, he famously denounced the very possibility of meta-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

physics and our ability to make any sense of it.

Subsequently, the logical positivists’ insistence on naturalism and a posteriori
method of inquiry led to an even stronger anti-metaphysical and anti-
ontological sentiment that dominated the decades that followed. Insistence on
a posteriori inquiry and continuity with science seemingly left no room for
questions about ontological categories. Whatever was left of the “old philoso-
phy” got translated into merely linguistic approaches to what were once con-
sidered substantive philosophical problems.
But despite the logical positivists’ best efforts to leave metaphysics behind,
a number of them were not able to practice what they preached, and they found
themselves unable to resist ontological discussions. This irony was not lost on
Gustav Bergmann, who joined the Vienna Circle in the late 1920s and early
1930s, but found himself increasingly disagreeing with some of the main tenets
of the Circle. Indeed, by the early 1950s, Bergmann’s philosophy had already
taken an “ontological turn.” He thought that metaphysics was, in fact, unavoid-
able and that many of the logical positivists were tacitly committing themselves
 Ontological Categories 9

to various ontological and metaphysical theses. He made this critique clear in

his suggestively titled collection of papers The Metaphysics of Logical
Positivism (1953). In the years that followed, Bergmann revived many of the
traditional ontological questions, including questions about the types of onto-
logical entities that should be admitted and how those entities relate to one
another. His own approach to these questions was heavily influenced by a post-
Russellian search for an “ideal language” whose ontological commitments
would be transparent. But perhaps most importantly for our purposes,
Bergmann marks a forceful return to realism about ontological categories. He
was the first to use the phrase “ontological ground” and to use it ubiquitously in
making arguments for why one might need to admit a certain category of entity.
Quine’s criticism of the logical empiricists’ analytic–synthetic distinction
refocused attention on ontology. With the breaking down of the boundary between
the two domains, Quine hoped to show that philosophical inquiry was continuous
with science and could, in fact, be carried out with the same scientific rigor and in
a naturalistic spirit. To help make metaphysical discussion more regimented, Quine
tried to impose certain meta-ontological constraints. He proposed an explicit
criterion of ontological commitment (“to be is to be a value of a bound variable”
in statements of first-order logic); he urged metaphysicians to provide clear identity
criteria for entities they wanted to postulate (“no entity without identity”); and he
offered a cost–benefit approach to evaluating opposing metaphysical theories (by
appeal to a trade-off between ideology and ontology, and appeals to simplicity and
explanatory power). This had a double effect: on the one hand, it freed up ontology
in a way that allowed many different theories with different ontological commit-
ments to flourish and compete with one another; on the other hand, the proliferation
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

of different ontologies, together with Quine’s impactful instrumentalist approach to

philosophical theories, gave momentum to ontological relativism.
This brief and inevitably broad-strokes sketch of the way that the fundamen-
tal ontological question concerning categories has been approached is a project
that deserves further attention and careful historical development in its own
right. My goal here is not to take on such a project in any real depth, but just to
highlight some ideas that shaped the debate about the very possibility of the
ontological project. It helps to know, for instance, that the heavily naturalist
approaches exemplified by many metaphysicians today, as well as the popular-
ity of different nominalist approaches to ontology, can be traced back to Quine
and his taste “for desert landscapes.” Similarly, it’s good to keep in mind that the
scientific realist approach to ontology has its roots in Bergmann and his
students’ reckoning with logical positivism, on the one hand, as well as in
John Anderson’s influence on David Armstrong and the Australian school of
metaphysics, on the other.
 10 Metaphysics

It is thanks to figures like David Armstrong, David Lewis, Keith Campbell,

D. C. Williams, E. J. Lowe, and others who engaged in substantial ontological
debates during the last few decades of the twentieth century, that ontology is
a much livelier field today than it was 100 years ago.
This ontological revival is in no small measure due to what might be termed
“the truthmaker turn” in metaphysics, advocated for by Armstrong and
C. B. Martin. The truthmaker principle that they advocated marked a strong
return to the correspondence theory of truth and its main idea that the truth of
a sentence consists in its correspondence to reality. The truthmaker principle thus
requires that all truths need to be made true by something in reality – a
truth-maker. The principle itself does not legislate as to which truthmakers need
to be admitted in order to make various truthbearers true, nor does it require
a simple one-to-one correspondence between true sentences and truthmakers.
(The same truthmaker can make multiple sentences true and one truth can have
multiple truthmakers.) Nor does the truthmaker principle require us to naïvely
read ontology off the surface structure of a sentence. But what the truthmaker-led
debate has done is to refocus attention back on ontological entities and categories
that need to be admitted to anchor the truths of our statements about the world.

2 Introducing an Ontological Category: A Case Study

in Universals
This section tells a story of how universals – one of the oldest, most disputed,
but also most worked-out ontological categories – got introduced. Universals
are particularly relevant to our discussion because right from their inception,
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

they were thought of as filling a dual role: they were thought of as an ontological
category as well as an entity that helped categorize other entities. I discuss the
“one-over-many” argument for universals as it was presented by Plato, as well
as the way that the argument was developed by a prominent contemporary
realist about universals – David Armstrong. Armstrong believed that his version
of “one over many” only offered a “preliminary case” for universals, not
a decisive one. His case for universals includes the argument from laws of
nature and the so-called negative arguments, from Russell, against the denial of
relational universals; and from Pap and Jackson, from the lack of adequate
paraphrases of apparent references to universals.

2.1 The Original One-over-Many Argument for Universals

Plato introduced universals in the context of searching for an adequate Socratic
definition. His dialogues frequently open with Socrates asking a question of the
type, “What is X?” about relevant ethical or aesthetic qualities such as virtue,
 Ontological Categories 11

justice, beauty, and so on. The right answer to this kind of question, according to
Plato, had to take the form of a definition of X. And yet not just any definition
would do. The definition that Plato was after was not supposed to merely list the
sort of things that were considered to be X; nor would it have been sufficient to
explain just the meaning of the word “X.” What Plato wanted the correct
Socratic definition to do was to explain the very nature of the quality X: it had
to give a universally true description of its essence. It had to be a definition that
would guide us in deciding whether various new candidates were indeed X or
not.
Thus, right at the outset, Plato was making an ontological distinction between
many things that exemplify X, on the one hand, and the single entity X itself, on
the other. He contrasts the plurality of just actions with justice itself, the
plurality of beautiful things with beauty itself, and so on. In the same breath,
though, Plato stresses that what makes various things beautiful and just is their
partaking in the “absolute” beauty and justice:

It seems to me that whatever else is beautiful apart from absolute beauty is

beautiful because it partakes of that absolute beauty, and for no other
reason. . . . The one thing that makes that object beautiful is the presence in
it or association with it, in whatever way the relation comes about, of absolute
beauty. I do not go so far as to insist upon the precise details – only upon the
fact that it is by beauty that beautiful things are beautiful. (Phaedo 100 c–d,
Plato 1963, pp. 81–82)

Beauty itself, justice itself, the X itself are entities Plato called Forms and in the
passage quoted we already see at work the famous “one-over-many” argument
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

for the existence of Forms.2 The argument proceeds from the assumption that
many different particular things – say, a, b, and c – are all beautiful, to the
conclusion that there must be a single Form of beauty that is responsible for
conferring beauty to the particular things. This movement, from the plurality of
things that have something in common to a conclusion that asserts the existence
of the entity that the plurality of entities shares in, is even more explicit in the
following passage from the Republic 596a–b:

We are in the habit, I take it, of positing a single idea or form in the case of the
various multiplicities to which we give the same name. . . . Let us take any
multiplicity you please; for example, there are many couches and tables. . . .
But these utensils imply, I suppose, only two ideas or forms, one of a couch
and one of a table. (Plato 1963, p. 820)

2
The name of the argument comes from Aristotle who referred to an argument that Platonists used
for the existence of Forms. Plato himself never actually used the phrase but he did sometimes refer
to a Form as being the “one over many.”
 12 Metaphysics

We can see here that the distinction between Forms and particulars that partake
in those Forms marks for Plato a difference between two different ontological
categories that occupy two separate realms of being, and that have contrasting
features. Forms are the more perfect, single, independent, changeless entities,
whereas particulars are less perfect, multiple entities that have their natures in
virtue of their relations to Forms, and they are subject to constant change.
Why Plato held that Forms exist independently of particulars is not entirely
clear. His arguments for this seem to be based on the assumption that there are
perfect properties that particular things never fully exemplify, but only approxi-
mate to. Mathematics, and geometry in particular, provided Plato with plenty of
examples of this sort: for instance, one can speak of a perfectly circular shape
but whichever circle we draw it’s always going to have some irregularity and
imperfection. This led Plato to believe that things in the natural world can only
approximate the perfection described by a mathematician. Generalizing to all
particulars, he thought that they could only approximate to the perfection of
Forms; that is, even though perfect justice may never be encountered in any just
action that any one person performs, perfect justice, that is, the Form of justice,
must exist, as a perfect paradigm.
Plato described the relation between the particulars and Forms in two differ-
ent ways: sometimes he described it as participation or sharing in, and some-
times as imitation. In the first case, particulars participate in the Form or the
Form is somehow shared by particulars; the second case is more mysterious:
particulars are related to the Forms as imperfect copies are to the originals.
Since Forms possess the ultimate perfection and particulars are only poor copies
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

of it, Forms were thought of by Plato as having greater reality.

In Parmenides, Plato addressed the issue of how inclusive his realm of Forms
should be. Moral and aesthetic qualities such as justice, beauty, and goodness
seem to deserve honorary membership in the realm, but what about Forms of
man, fire, or water? In Parmenides, the young Socrates discloses his puzzlement
and inability to decide such cases as well as a reluctance to admit less dignified
Forms of things like hair, or mud, or dirt, and so on. Parmenides’ response to
this is that when he gets older he will learn not to be so fastidious and
discriminatory when it comes to Forms.
Plato’s discussion of “lesser” Forms is analogous to the distinction that
contemporary metaphysicians make between sparse properties and relations,
the sharing of which makes for genuine resemblance between particulars, and
abundant properties and relations the sharing of which does not make for
resemblance. A contemporary realist about universals tends to draw this dis-
tinction by saying that the former properties correspond to universals while the
latter don’t; similarly, Plato allowed some properties to be Forms and not others.
 Ontological Categories 13

Although it might seem that Plato’s Socrates is basing his decisions on ad hoc
judgments about which properties and relations appear worthy of a place in the
realm of Forms, the grounds for the distinction run deeper. This can be espe-
cially appreciated in Plato’s later writings, in which his earlier conception of
Forms as paradigms gives way to the conception of Forms as categorizers.
Therein, Plato makes it very clear that not just any collection of particulars
corresponds to a Form; only the collections that respect the “objective articula-
tions” in nature are eligible. This view is confirmed by the following passages
from the Statesman 262b–e, where Plato cannot stress enough the importance of
following the true classifications of things in pursuit of the correct Socratic
definition:

We must only divide where there is a real cleavage between specific forms.
The section must always possess a specific form. . . . [I]t is dangerous,
Socrates, to chop reality up into small portions. It is always safer to go
down the middle to make our cuts. The real cleavages among the Forms are
more likely to be found thus, and the whole art of these definitions consists in
finding these cleavages. (Plato 1963, pp. 1025–1026)

“Real cleavages” in the world are cleavages between the classes of genuinely
resembling particulars. To each of those classes corresponds a Form which
unites it. To avoid mistakes of misclassification one has to be careful and try not
to give in to hasty and arbitrary divisions dictated by one’s own subjective
interests. The example of such a mistake is:

The kind of mistake a man would make who, seeking to divide the class of
human beings in two, divided them into Greeks and barbarians. This is
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

a division most people in this part of the world make. They separate the
Greeks from all other nations making them a class apart; thus they group all
other nations together as a class, ignoring the fact that it is an indeterminate
class made up of peoples who have no intercourse with each other and speak
different languages. Lumping all this non-Greek residue together, they think it
must constitute one real class because they have a common name “barbar-
ian” to attach to it. . . . A division setting Lydians or Phrygians or any other
peoples in contradistinction to all the rest can only be made when a man fails
to arrive at a true division into two groups each of which after separation is
not only a portion of the whole class to be divided but also a real subdivision
of it. (Plato 1963, p. 1026, italics mine)

This passage is an important one, as it clearly illustrates how concerned Plato

was about making “real” classifications as opposed to hasty and confused
groupings that only on the surface appear to belong together. The key to correct
classifications is finding genuine resemblances between the members of a given
class – where there is such resemblance, there is a corresponding Form; where
 14 Metaphysics

there isn’t, there is no corresponding Form. Thus Forms turn out to be not just
one of the most fundamental ontological categories introduced, alongside
particulars, but are themselves are entities that sort and categorize.
This dual role that Forms are meant to play leads to a problem that is brought
out by Plato’s infamous Third Man Argument, an infinite regress argument
which exhibits the following structure:

(1) a, b, and c are all large because there is a Form of largeness L that they all
participate in.
(2) a, b, c, and L are all large because there is a Form of largeness L1 that they
all participate in.
(3) a, b, c, L, and L1 are all large because all of them participate in a Form of
largeness L2 – and so on ad infinitum.

Three main assumptions are at work in this argument. First is “one over many”
which states that many things being L must be explained by the existence of the
Form L that all of them participate in. The second premise is implicit and it has
been called the premise of “non-self explanation”; it simply assumes, that in
explaining what a, b, c, and L have in common one cannot appeal to the L itself,
but needs to appeal to a different Form of largeness, L1. The third premise which
is implied in this argument is the “self-partaking” premise; it assumes that the
Form L is itself L, or, in this particular case, that the Form of largeness is itself
large.
The correct interpretation of this argument has been debated at length by
commentators on Plato. My aim here is not to engage in detail with that debate,
but to merely point out the tension that it brings out very clearly – namely,
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Forms conceived as entities that confer to particulars their nature (in this
instance largeness to a, b, and c), then need some other Form to confer that
nature to them (since largeness is itself assumed to be large). But if there is
a different, higher-order Form of largeness that confers largeness L1 to L, then
we are off on a regress of higher and higher-order Forms.
It might seem that the easiest way to escape this difficulty would be to claim
that Forms have a dual role: the role of conferring qualities to particulars as well
as the role of conferring that very same quality to itself – the form of largeness
would itself be large due to this self-conferring feature. But the outcome of this
approach is a very real puzzle about what it means for a Form itself to be large or
beautiful or just, and so on (since these appear to be qualities that particular
things have, not Forms themselves). To avoid this puzzle, Plato would have to
embrace the alternative and claim that Forms confer qualities to particulars but
do not themselves exemplify such qualities – so the Form of largeness is not
itself large and the Form of beauty is not itself beautiful, and so on. This would
 Ontological Categories 15

nicely avoid the Third Man regress altogether, leaving Plato with the task of
explaining of how the Forms confer beauty, justice, virtue, and so on to
particulars, if they are themselves distinctly not those things.
Although Aristotle disagreed with Plato on the nature of Forms and their
interaction with particulars, he did believe in the existence of some such entities.
He called them ta katholou, that is, universals. The main point of disagreement
was over Plato’s view of Forms as entities completely independent of, and
separately existing from, particulars. Plato saw Forms as entities in a realm of
their own, outside of space and time, and yet more real than particulars.
Aristotle, on the other hand, wanted to restore the balance in favor of particulars.
Particulars and universals are seen as mutually dependent entities; in fact, it is
an essential part of the nature of Aristotelian universals to be related to particu-
lars. The reason why Plato could not see this, according to Aristotle, was
because he understood Forms to be substances. In his earlier works, such as
Categories, Aristotle uses the term “substance” to refer to things that are neither
in anything nor said of anything; they are things like a particular horse or
a particular man. Thus, Plato’s mistake, in Aristotle’s view, was that his
Forms were more like particular substances than universals.
Although he harshly criticized the outcome of the argument – the existence of
Forms as such – Aristotle fully appropriated and relied upon the “one-over-
many” argument as proof of the existence of his own universals. Indeed, he
took the validity of the ontological version of the argument for granted and
employed different versions of “one over many” for different purposes. For
instance, Aristotle mentions a version of the argument when arguing for the
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

existence of universals as objects/contents of thought and knowledge. In

Posterior Analytics, he describes the inductive process that leads from the
perception and memory of numerous objects to the universal “that is one apart
from the many” and “one and the same in all those things” as a “principle skill of
understanding” (100a4–9). In Metaphysics, he states that “if there is nothing apart
from individuals, there will be no object of thought” and as a consequence of this,
“all things will be objects of sense” (999b1–2). Thus, for Aristotle, if there are no
universals there is nothing that we can properly think about; and without thought
there is no knowledge, “for the knowledge of anything is universal” (1003a14).
To summarize, Plato relied upon the “one-over-many” argument to provide
him with the entities he needed for definitions of the natures of things – to
explain what made different things beautiful, just, wise, large, and so on. These
entities were Forms, the self-sufficient, independent, immutable, perfect, and
eternal entities existing in a realm of their own. In his earlier writings, Plato
thought of Forms as perfect paradigms, but later came to think of them as
entities that categorize other entities and through such categorizations reflect the
 16 Metaphysics

objective, mind-independent classifications of reality. The exact way that Forms

were meant to fulfill this role remains unclear.
The way that Plato went about introducing universals left a lot of his
assumptions implicit. For instance, he took it for granted that the phenom-
enon that he was describing (different particulars seemingly having
a common nature) was an obvious and indisputable one. He also took it for
granted that this phenomenon was inherently problematic and in need of an
explanation. He then assumed that introducing an entity – a Form – would
provide an adequate explanation of the phenomenon in question. Finally,
Plato did not question the existence and the role that multiple distinct
particulars play. The assumption was that the universals help explain the
natures and the proper groupings of many different particulars, but that the
particulars themselves were in no need of special argument or explanation.
As we will see in the next section, more recent arguments for the existence of
universals attempt to make some of these assumptions more explicit and
motivated.

2.2 A Contemporary Approach to Universals: Armstrong’s Use

of the One-over-Many Argument
The contemporary realists’ reliance on the “one-over-many” argument is more
guarded than Plato’s and Aristotle’s, but the argument still presents the main
motivating force for the introduction of universals as a sui generis ontological
category. David Armstrong, one of the most prominent contemporary realists,
puts it this way: “[t]he main argument for the existence of universals is Plato’s
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

‘One over Many’. I do not think that it proves straight off that there are
universals. But I think that it shows that there is a strong preliminary case for
accepting universals” (Armstrong 1997a, p. 101).
This “strong preliminary case” is built on the fundamental assumption that
facts of sameness of type are objective “Moorean” facts:

G.E. Moore thought . . . that there are many facts which even philosophers
should not deny, whatever philosophical account or analysis they gave of
these facts. He gave as an example the existence of his hands. We can argue
about the philosophical account which ought to be given of material objects,
such as Moore’s hands. But we should not deny that there are such
things. . . . I suggest that the fact of sameness of type is a Moorean fact.
(Armstrong 1997a, p.102)

Calling the sameness of type a “Moorean fact” makes it very clear what
Armstrong takes to be the starting point of the ontological debate. The apparent
sameness of type is something that all sides of the debate should agree upon,
 Ontological Categories 17

according to him. The explanation of this fact, on the other hand, is the issue
about which philosophers disagree. Armstrong himself believes that apparent
sameness of type is due to genuine resemblance between particular things, and
that this resemblance is due to partial identity. The structure of Armstrong’s
version of the “one-over-many” argument is thus roughly as follows:

(1) Facts of sameness of type are Moorean facts, that is, objective facts that no
philosopher should deny.
(2) Apparent sameness of type should be explained.
(3) Genuine resemblance among different particular things is what gives rise to
the apparent sameness of type among distinct particulars.
(4) Genuine resemblance among different particulars is due to their having
something in common, literally sharing something.
(5) Universals are what different resembling particulars have in common and
what makes them resemble each other.

Plato’s version of “one over many” took for granted (1), (2), and (3) and directly
inferred (5) from (4).
Armstrong’s structure of the argument, on the other hand, is more careful and
allows opponents to disagree at various points. Contra Armstrong, one might try
to deny (1). This denial could be developed in different ways. Theoretically, one
might deny that there is even an appearance of sameness of type. This is a difficult
claim to pull off, since it goes against what most people report as their experience
of reality; thus, one would have to appeal to a very different phenomenological
experience of reality – one which does not recognize similarities and patterns in
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

nature at all. Communication with someone who makes such a claim would be
impossible, since the very structure of our language and our thought relies on
a shared understanding of commonalities and patterns; thus, even the denial of the
appearance of such patterns would, in its formulation and communication,
presuppose the very things which are being denied. A more plausible denial of
(1) would acknowledge that there is indeed an appearance of sameness of type,
but deny that such an appearance captures a real phenomenon. In reality, the
claim would go, there is no genuine resemblance between particulars, no
Moorean fact of sameness of type; there is only an apparent fact. The interesting
question then becomes analogous to those that other skeptical attacks give rise
to – namely, how come there is a need to postulate the sameness of type when in
reality there isn’t one? What gives rise to such a pervasive illusion and what, if
anything, is to be found in its stead? An error theory of some sort is called for.
Armstrong was correct in his assessment that most philosophers would not
outright deny (1), and would indeed agree that there are many objective resem-
blances to be found in reality. But such philosophers may disagree about (2); that
 18 Metaphysics

is, they may disagree on whether an explanation of such facts is necessary at all.
For instance, nominalists tend to reject outright the demand for an analysis of the
facts of resemblance and insist on taking them to be primitive. In Quine’s words:
“That the houses and roses and sunsets are all of them red may be taken as
ultimate and irreducible” (Quine 1997, p. 81) Along similar lines, David Lewis
remarks:

An effort at systematic philosophy must indeed give an account of any

purported fact. There are three ways to give an account. (1) “I deny it” –
this earns a failing mark if the fact is really Moorean. (2) “I analyse it thus” –
this is Armstrong’s response to the fact of apparent sameness of type. Or (3) “I
accept it as primitive.” Not every account is an analysis! A system that takes
certain Moorean facts as primitive, as unanalysed, cannot be accused of
failing to make a place for them. (Lewis 1983, p. 20)

This is a forceful response to Armstrong, and, alongside him, to any realist who
insists on the need for an explanation of facts of resemblance. Lewis believes
that “making a place” for a phenomenon in one’s philosophical system ought to
count as a good enough response. He writes:

An adequate Nominalism . . . is a theory that takes Moorean facts of apparent

sameness of type as primitive. It predicates mutual resemblance of the things
which are apparently of the same type; or it predicates naturalness of some
property that they all share, i.e. that has them all as members; and it declines
to analyse these predications any further. (Lewis 1983, p. 21)

One can debate whether taking a phenomenon as primitive and unanalyzable

amounts to a sufficient engagement, and I will have more to say about in
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Section 3.2.3. What is worth highlighting in this Armstrong–Lewis exchange

is the value of an open debate concerning whether a certain phenomenon should
receive further analysis at all, especially when such an analysis quickly leads to
the introduction of a new ontological category. Indeed, if we embark on
providing a reductive analysis of facts of resemblance, we are already halfway
to introducing the category of universals. The subsequent steps, as we will see
shortly, are to analyze resemblance in terms of partial identity, and then to admit
universals as the entities that ground that partial identity between particulars.

2.2.1 Universals as Entities That Ground Resemblance

As Armstrong himself openly acknowledged, there is nothing in the “one-over-
many” argument that compels us to draw the conclusion that universals exist.
The ontological inflation that takes place is not necessitated by the argument.
But very few things in metaphysics ever are. What we are offered is an
explanation of resemblance in terms of universals.
 Ontological Categories 19

The explanation runs as follows: genuine resemblances among different

particulars are nothing else but identity in a certain respect. For example, two
apples resembling in being red and round is due to their being strictly identical
in those respects; the two apples have parts which are numerically identical in
both and which make for their partial identity – these parts are the universal of
redness and the universal of roundness. Universals are thus the sorts of entities
that can multiply occur and that are identical in each of their occurrences; our
intuition that things that resemble each other literally have the same thing in
common is thus given an ontological ground.
With universals, a realist is able to explain a few other features of resem-
blance. Two things that share all their universals are exact duplicates – they
could be said to perfectly resemble one another. Perfect resemblance understood
as sharing all the respects of resemblance is transitive: if a perfectly resembles
b, and b perfectly resembles c, then a will perfectly resemble c (since all three
will be perfect duplicates). This is obviously not the case when resemblance is
less than perfect, since the universals that a and b share may not all be the same
universals that b and c share.
One can also account for degrees of resemblance among particulars in terms
of universals: one way of doing this would be to say that particulars resemble to
a larger degree if they share more universals, and to a lesser degree if they share
fewer universals. But this is perhaps a somewhat crude criterion, as it is easy to
imagine two particulars, a and b, sharing one universal F, say being human, but
resembling each other more than a and c, which share two universals G and H,
say having black hair and having brown eyes (where, for example, a and b are
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

two children and c is an adult poodle). In this case we would be judging

resemblances according to which universals “carry more weight” by capturing
respects of resemblance better.
In any case, genuine, deep, resemblances between particulars are few.
A scientific realist like Armstrong believes that which universals there are is indeed
an a posteriori matter: it is up to scientific theories to discover the real cleavages in
nature and thus provide the candidates for universals.3 According to this sort of
scientific realist, most of our everyday talk about the sharing of properties and
relations is actually not a talk about genuine resemblances in nature.

2.2.2 Genuine Resemblances without Universals

Metaphysicians can agree with Armstrong’s setup of the “one over many” and
accept (1), (2), and (3) – that there is indeed a Moorean fact of the sameness of

3
Of different scientific inquires, Armstrong favors physics, and believes that it will give us our
ultimate, most fundamental universals.
 20 Metaphysics

type and that it is indeed due to genuine resemblance among different particulars –
but refuse to analyze this fact any further, and thus refuse to postulate the category
of universals. This is the move that nominalists such as David Lewis make when
they deny the existence of universals and take resemblance as a primitive.4

Nominalism
More specifically, Lewis defines properties as sets of all of their instances:
actual and possible. For him, redness is to be identified with the set of all red
things – across actual and possible worlds. Thus, the only kind of entity
admitted in Lewis’s ontology are particulars and classes of particulars. The
need to admit particulars in other possible worlds alongside this-worldly par-
ticulars was prompted by the coextension problem. If properties are nothing
more than sets of their instances, then properties that have the same extensions
(such as, say, the property of having a heart and the property of having a kidney)
amount to being the same property. By appeal to possibilia, Lewis was able to
address this difficulty: namely, although the set of people with hearts is ordin-
arily the same as the set of people with kidneys, this is considered to be merely
a contingent feature of people in this world; when one considers possible worlds
and their inhabitants, the two sets of instances come apart.
With the coextension problem addressed, Lewis still worried that his
account of properties was too undiscriminating, that it did not capture real
resemblance among its members. He wrote: Because properties are so abun-
dant, they are undiscriminating. Any two things share infinitely many proper-
ties, and fail to share infinitely many others. That is so whether the two things
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

are perfect duplicates or utterly dissimilar. Thus properties do nothing to

capture facts of resemblance. That is work more suited to the sparse universals
(Lewis 1983, p. 13).
Lewis then went on to introduce “an elite minority of special properties” – the
natural properties –which were meant to capture genuine resemblance classes.
These were envisioned, essentially, as a class nominalist’s response to univer-
sals: “we could call a property perfectly natural if its members are all and only
those things that share some one universal” (Lewis 1983, p. 13). But Lewis’s
ontology did not accept universals and his appeal to them here is merely
illustrative. What he wanted was that his own ontological categories would

4
I am here using the term “nominalism” to refer to views that reject universals and construe
properties out of ordinary particulars. Trope theories are not nominalist in this sense, for they tend
to construe particulars out of property-instances. If we were to adopt the view that any denial of
universals is a nominalist view, then Lewis’s view as well as all the varieties of trope theoretic
views would count as nominalist. I am not using the terminology in this way, however, and am
following Armstrong (1978, 1980, 1983, 1997) in my usage.
 Ontological Categories 21

somehow be able to do the work that universals do in the realist’s ontology. He

grappled with this issue and considered two avenues: the first was to take
“naturalness” as a primitive predicate (see Lewis 1983, p. 14) which applies
to all and only natural classes (resemblance would then be explained as
a relation holding among the members of natural classes picked out by such
primitive predicates); the second was to take genuine resemblance as a primitive
relation and then provide a definition of natural properties in its terms (as
properties whose members are all and only the members that resemble each
other). He did not adjudicate between these two paths and evaluating them
further need not concern us here. What should be mentioned is that the same
difficulties also assail the more recent version of nominalism –Rodriguez-
Pereyra’s (2002) brand of resemblance nominalism and its commitment to
primitive predicates of resemblance.
Despite the above-mentioned difficulties, both the class nominalism of David
Lewis and the resemblance nominalism of Rodriguez-Pereyra explicitly
acknowledge that facts of resemblance are real mind-independent features of
reality. This is not the case for predicate and concept nominalists who take
resemblance between two particulars to consist in sharing the same predicate or
falling under the same concept. A predicate nominalist claims that different
particulars resemble each other in being red due to the fact that they fall under
the predicate “is red,” while the concept nominalist appeals to the concept of
redness. But predicates and concepts are created by us, and it appears wrong-
headed to claim that properties of things depend on the existence of the
predicates we make to describe them or the concepts we create for them to
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

fall under. To put it slightly differently: if there were no humankind, and thus no
spoken language, there would still seemingly be things that were red, square,
had mass or charge, and so on, and resembled other particulars in some of these
respects. Moreover, it seems plausible to assume that there are already resem-
blances and properties in the actual world that we do not yet have predicates or
concepts for. But a predicate or a concept nominalist would have to deny this –
the properties would only come into existence once we have found predicates or
concepts to describe them. Or, perhaps they would have to appeal to such things
as possible predicates or concepts that await our discovery of them. Either way,
this kind of nominalist story seems rather implausible.

Trope Theory
Trope theorists have argued for a midway position between the realist and
the nominalist approach, introducing a sui generis entity – a nonrepeatable,
singly occurring sort of property they call a “trope.” Trope theorists such as
 22 Metaphysics

Keith Campbell tend to think of their theory as primarily particularist: “The

trope theory is Particularist. It accepts Locke’s thesis that all things are
always only particular. It denies that there are any literally common elem-
ents present in all members of a group of resembling particulars” (Campbell
1990, p. 27).
But although it accepts only particulars in its ontology and denies any
universals whatsoever, trope theory should not be confused with a nominalist
position:

Some writers use the label “nominalist” for every denial of universals, but
this blurs a crucial distinction: ordinary nominalisms, in denying universals,
deny the existence of properties, except perhaps as shadows of predicates or
classifications. They recognize only concrete particulars and sets. . . . But the
trope philosophy emphatically affirms the existence of properties (qualities
and relations). Indeed, it holds that there is nothing but properties (or nothing
but properties and space–time). However, it insists that these properties are
not universals but, on the contrary, particulars with a single, circumscribed
occurrence. (Campbell 1990, p. 27)

In this paragraph, Campbell changes the terms of discussion – the division is

no longer between realists who believe in the existence of universals and
nominalists who deny it. Properties have now become central and, since
nominalists construe properties out of particulars (plus sets, resemblance
relations, etc.), their properties appear to Campbell not to have enough
ontological weight; he says they are mere “shadows of predicates or classifi-
cations.” The trope theorist, on the other hand, “emphatically affirms” the
existence of properties, which for them means that there are ontological
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

entities filling this particular property role. In fact, for Campbell, “there is
nothing but properties.”
When it comes to the problem of sameness of type, Campbell is very much in
agreement with Armstrong that the problem is a serious one that a “responsible
ontology” ought to tackle: “The world is not a chaos, with every aspect, at
every minute, unique in character. Nor is it an undifferentiated blancmange. It is
a diverse and orderly cosmos displaying patterns of recurrence. No responsible
ontology can evade this very general fact; and no responsible ontology can
avoid offering its assay of this situation” (Campbell 1990, p. 28).
Thus, just like the realist and the class and resemblance nominalists men-
tioned, the trope theorist also recognizes the importance of the Moorean facts of
resemblance. Campbell clearly agrees with (1), (2), and (3) in Armstrong’s
version of the “one over many” outlined in Section 2.2. Unlike Lewis, and more
along the realist’s lines, Campbell recognizes the importance of providing an
“assay of this situation.” At the same time, Campbell believes that the realist’s
 Ontological Categories 23

stating of the problem should be broken down into two related but separate
questions – the questions he refers to as “A” and “B.” The “A question” is about
the single particular object and its properties; it is a question that takes “one
single red object” and asks of it, “what is it about this thing in virtue of which it
is red?” The “B question,” on the other hand, is about two or more objects; it is
a question that takes “any two red things” and asks, “what is it about these two
things in virtue of which they are both red?”
Campbell points out that discussions of the problem of universals mostly
assume that the two questions are to be given parallel answers. This can be seen
from the fact that the realist gives the same response to A and B – that the single
thing is red in virtue of the universal redness present in it, and that two things are
both red in virtue of the universal of redness present in each. The conclusion that
Campbell draws from this is that: “The conflation of the A and the B questions is
responsible for making the realist position seem much more inevitable than it
really is” (Campbell 1990, p. 29). By disentangling the two questions he hopes
to show that the realist’s solution is not inevitable.
For Campbell, once the two questions are separated, universals start to seem
somewhat ad hoc as an answer to the A question. In his view, it amounts to
saying, without any further justification, that “a nature or a character of an item
can never be particular.” Or, if one is in search of a more “scientific” answer to
the question of what it is in virtue of which some object has a certain property,
the quest is delayed a little. The property in question will be explained away by
some other property (or properties) of the underlying structure and those by,
perhaps, other properties. However, Campbell insists, there will be a moment in
which the only available answer to a question such as “what is it about charge in
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

virtue of which it is charge?” would be “its being what it is” (Campbell 1990,
p. 30). This sort of tautological response is inevitable at a certain level of every
system:

It is critical to the trope vision of the world that particulars can be natures, that
something can just be a case of charge, or colour, or whatever. Philosophers
are rightly suspicious of tautological-seeming answers to questions . . . but it
is important to remember that such answers arise at some point in every
system. The realist about universals has a substantial seeming answer to our
A question, even in the case of basic properties. But the rock bottom is not far
away. What is it about electric charge in virtue of which the presence of this
universal is necessary and sufficient for something’s having charge? Its being
what it is. (Campbell 1990, p. 30)

In other words, something’s having electric charge may come down to its
having the right sort of universal, but the next question about that universal –
“what is it about that universal that enables the particular to have electric
 24 Metaphysics

charge?” – would then have no informative answer; it would probably have to

be something like “it being that particular type of universal.” The point of these
observations, for Campbell, is to show that introducing universals in order to
give a slightly more substantial answer to the A question is an ontological
redundancy, unless, of course, universals are required elsewhere in one’s
ontology.
Campbell might be right about the fact that an answer to an A question alone
is not a good enough reason for introducing universals. But the realist’s main
argument for the existence of universals comes from trying to answer the
B question – in virtue of what do two or more objects resemble each other in
being red, round, human, and so on? It is the intuition about there being certain
shared natures among different particulars that motivates the introduction of
universals. With universals, realists take themselves to have provided an
explanatory answer to the B question, along with a “more substantial answer”
to the A question.
In contrast, Campbell’s own answer to the B question is that the apparent
sameness of type is simply due to the presence in them of distinct resembling
tropes:

What is it about two objects in virtue of which they are both red? Each
includes a red trope. What is it about those tropes in virtue of which they are
both red tropes? Their likeness to one another is what makes them tropes of
the same kind. Their natures make this the red, rather than the blue, or oblong,
kind. . . . What is being offered here, of course, is a Resemblance theory of
resemblance and recurrence, an assay in terms of resembling particulars, in
place of one involving identical universals. It takes resemblance to be ultim-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

ately, in basic cases, unanalysable. (Campbell 1990, p. 31)

Thus, resemblance of tropes themselves, just as in the case of class and resem-
blance nominalism, is to remain a given and not a further analyzable feature of
trope theory.5

2.3 Additional Arguments for Universals

I have thus far presented “one over many” as the realist’s “strong preliminary
case” for universals. But realists tend to support their case with additional
arguments. For instance, Armstrong also wants his theory of universals to
sustain his non-Humean account of laws of nature. According to this account,
roughly, “All Fs are Gs” is a law of nature only if there is a second-order
law-making relation of necessitation, N, holding between the universals’ F-ness

5
For a more thorough discussion of trope theory, good starting points are Simons (1994) and
Maurin (2018).
 Ontological Categories 25

and G-ness: N (F, G). Armstrong (1983, 1997b) singles out functional laws as
laws that have the best claim to be fundamental, and determinable universals as
the ones that ought to feature in them. One example of a functional law is
Newton’s law of gravitation: it correlates the determinables force, mass, and
distance, and under each of these determinables there is a class with possibly an
infinite many determinate universals that the law applies to.
Additionally, there have also been some prominent “negative arguments” for
universals. Such arguments are notable for making an indirect case for the
existence of universals by showing the extent of the difficulties that arise
without them.

2.3.1 Russell’s Regress Argument against the Denial of Universals

In The Problems of Philosophy, Russell (1912) presents a famous negative

argument for the existence of universals by claiming that avoidance of univer-
sals leads to a vicious infinite regress. He describes the regress as follows:

If we wish to avoid the universals whiteness and triangularity, we shall

choose some particular patch of white or some particular triangle, and say
that anything is white or a triangle if it has the right sort of resemblance to our
chosen particular. But then the resemblance required will have to be
a universal. Since there are many white things, the resemblance must hold
between many pairs of particular white things; and this is the characteristic of
a universal. It will be useless to say that there is a different resemblance for
each pair, for then we shall have to say that these resemblances resemble each
other, and thus at last we shall be forced to admit resemblance as a universal.
The relation of resemblance, therefore, must be a true universal. And having
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

been forced to admit this universal, we find that it is no longer worth while to
invent difficult and unplausible theories to avoid the admission of such
universals as whiteness and triangularity. (Russell 1912, p. 55)

Thus, if a nominalist wants to avoid postulating universals such as whiteness

and triangularity, they need to find alternative ways of accounting for proper-
ties. One way of doing this is to pick out some particular, d – a particular patch
of white or a particular triangle – and then take the properties of whiteness and
triangularity to consist in the “right sort of resemblance” of particulars to the
chosen paradigm, d. This position has come to be known as resemblance
nominalism, or, more specifically, as paradigm resemblance nominalism.
Now, as there are many white and triangular things, there will be many pairs
of things resembling each other in the relevant way. Multiple recurrence is
a characteristic of universals and if one wants to avoid admitting resemblance as
a universal, a way to do this is by saying that there is a different resemblance for
each pair (a,d), (b,d), (c,d), and so on – r(a,d), r(b,d), r(c,d), and so on. But if
 26 Metaphysics

a nominalist says this, they find themselves again having to explain what makes
all the resemblances, r(a,d), r(b,d), r(c,d), and so forth, resemble each other.
They may say that all of them resemble each other because each of them
resembles some arbitrarily picked resemblance r(x,d). In this way we get new
pairs of resembling resemblances: (r[a,d], r[x,d]); (r[b,d], r[x,d]); (r[c,d], r[x,d]).
And again, if in each case we ascribe the same resemblance, there arises a risk
for the nominalist of admitting a universal into their system. Therefore, they
have to say that each of the new pairs of resemblances are different particular
resemblances, and so forth. The conclusion that Russell draws from this is that
since the resistance to admit universals leads to an infinite regress of resem-
blance relations, a nominalist might as well accept the resemblance relation as
a universal in the first place. And once one universal is admitted, why not let
all of them in?
Immediately after his presentation of the regress argument for universals,
Russell notes that Berkeley and Hume failed to appreciate the force of this
argument because they thought of qualities rather than relations as exemplars of
universals. For Russell, at this time, the opposite is the case: it is relations that
are the main candidates for universals. Qualities can be treated, as the regress
argument itself suggests, in terms of resemblance of one particular to another;
relations, however, are irreducible. An attempt to rid oneself of relations leads to
an infinite regress of further relations of resemblance or likeness. The argument
thus proves very simply, according to Russell, that an ontology that admits of
only particulars is flawed – it overlooks the fact that relations are ineliminable,
and thus that universals are too.6
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

2.3.2 Against the Possibility of Paraphrasing Away the Reference

to Universals

Another form of negative argument for universals that Armstrong presents

originates from Pap (1959) and was reworked by Jackson (1977). It challenges
the nominalist to translate the following two true statements about colors: (1)
“Red resembles orange more than it resembles blue,” and (2) “Red is a color.”
A common nominalist paraphrase of the first of these statements would run
something like: (1*) “For all particulars, x, y, and z, if x is red and y is orange and
z is blue, then x resembles y more than x resembles z.” But Pap points out that the
translation proposed by the nominalist is not equivalent to the original statement
since “x may resemble z more than y in other respects though x is red and y orange

6
There is a lot more that can be said about the origin and impact of Russell’s regress. For a further
discussion, see Cargile (2003), Rodriguez-Pereyra (2004), and Perović (2015). There is also an
important discussion about what makes for a vicious regress in Nolan (2001).
 Ontological Categories 27

and z blue” (Pap 1959, p. 334). In other words, ordinary particulars x, y, and z will
have other characteristics besides color. For instance, let x be a red cube, y an
orange sphere, and z a blue sphere; in this case, the truth of statement (1) will
stand unaffected whereas the translation (1*) turns out to be false.
At this point a nominalist can try rephrasing statement (1) in a way that
specifies the respects of resemblance involved; Pap suggests something like:
(1**) “For all particulars, x, y, and z, if x is red and y is orange and z is blue, then
x resembles y in respect of color more than x resembles z in respect of color.”
The problem with this translation, Armstrong argues, is that “resemblance in
respect of color” seems to refer to a four-term relation holding between x, y, z,
and color – an option hardly attractive to nominalists, since they do not wish to
be committed to relational universals any more than they want to be committed
to property universals. Moreover, this paraphrase seems to introduce a reference
to an additional universal – color. A better paraphrase would probably be:
(1***) “For all particulars, x, y and z, if x is red and y is orange and z is blue,
then x colour-resembles y more than x colour-resembles z” (Armstrong 1978,
p. 59). Here, in place of “resemblance in respect of color,” which seems to refer
potentially to universal color, we have a predicate “color-resembles” which
might avoid such commitment if it is taken to be a primitive predicate. But then
concerns arise as to whether such a predicate is indeed sufficiently conceptually
simple to be treated as a primitive and how it is to be distinguished from other
resemblance predicates such as, say, shape-resemblance, temperature-
resemblance, and so on. All of these resemblances are distinct – but in virtue
of what?, challenges the realist.
Statement (2) “Red is a color” poses difficulties as well. One suggestion is to
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

try and translate it along the lines of (1*), that is, as (2*): “For all particulars, x, if
x is red, then x is colored.” It is clear that (2) entails (2*) but in order for the
translation to be successful the reverse has to be true as well. To test whether
(2*) does entail (2), Jackson (1977) proposed consideration of an analogous
case, the case of statement (3*): “For all particulars, x, if x is red, then x is
extended.” (3*) is a true statement; in fact, it appears to be a necessary truth like
(2*). It seems, then, that if (2*) entails (2), then analogously, (3*) should entail
(3): “Red is an extension.” But (3) is clearly false. This, according to Jackson,
and Armstrong who follows him, proves that (2) says something more than (2*)
and that the nominalist hasn’t managed to capture in their translation that extra
something which (2) seems to express. In Jackson’s words:

If red’s being a color were nothing more than a matter of every red thing
necessarily being colored, then red’s being a shape and an extension would be
nothing more than the fact that necessarily every red thing is shaped and
 28 Metaphysics

extended. And red is not a shape and not an extension. It seems that “Red is
a color” says, as realists maintain, something about red not reducible to
something about red things. (Jackson 1977, p. 427)

Of course, one need not take these arguments from the lack of adequate
paraphrases of apparent references to universals as demonstrating anything
more than a quirk of language. Namely, one may very well acknowledge that
our ordinary language contains apparent references to universals that cannot be
easily eliminated without loss of meaning, while at the same time maintaining
that such a feature of language should not be seen as ontologically significant.
It’s important to bear in mind that not one of the arguments discussed in this
section is put forward as a decisive argument for universals. These arguments
are presented here in order to illustrate what a cumulative case for one onto-
logical category – the category of universals – looks like. It’s inevitably a partial
story, as some arguments can be replaced with others, perhaps more compelling
ones. The “one-over-many” argument plays an important role in the overall
story being told. Plato thought of it as an obviously decisive argument for
universals, but contemporary metaphysicians today rarely think of it that way.
The reconstruction of Armstrong’s version of the argument was meant to show
the different places where other philosophers might disagree and dispute the
need to introduce the category of universals, and appeal to other categories and
explanations in their stead.

3 Some Desiderata for a Promising System of Ontological

Categories
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Drawing on the case study of universals from Section 2, this section discusses
some theoretical desiderata for building a promising system of ontological
categories. The first desideratum is to make sure that the metaphysical question
or problem that the ontological entity is introduced to address is well-stated and
motivated. This, as we will see, requires taking special care to avoid engaging
with metaphysical pseudo-problems. I then describe considerations of ontological
and ideological economy that tend to guide philosophers in their theory-building.
I also discuss the distinction between simple and more complex ontological
entities, and the widely held assumption that fundamentality and simplicity go
hand in hand in a metaphysical explanation. Sections 3.2.5 and 3.2.6 engage with
the meta-theoretical considerations about the relationship between ontological
entities, categories, and roles. Drawing on Oliver (1996), I make a case for
characterizing ontological categories by appeal to ontological roles they play in
solving certain metaphysical problems. I believe that this characterization best
accommodates the piecemeal, revisionary, bottom-up approach that I advocate.
 Ontological Categories 29

3.1 Motivating the Question and Steering Clear

of Pseudo-Problems
In Section 2, we saw the category of universals introduced primarily through the
“one-over-many” argument. We saw that Plato arrived at the conclusion of the
argument – his introduction of Forms – very quickly, whereas Armstrong’s
development of the argument was a bit more gradual and careful. The crucial
steps in Armstrong’s version of the argument are to assert that there is
a Moorean fact of sameness of type and then to require that this apparent fact
be further explained.
Realists and trope theorists take this task on – the former attempting to provide an
explanation by appeal to universals, while the latter appeal to tropes and their
mutual resemblance. The question “what makes for genuine resemblance between
this piece of paper and this table?” would thus, for a realist, be that they both share
one and the same universal of rectangularity, whereas trope theorists would say that
they each have a distinct trope of rectangularity, r1 and r2, and that the resemblance
between these two tropes requires no further explanation or grounding.
Class nominalists resist the demand for an explanation of resemblance and
take it to be a primitive and unanalyzable predicate which applies to natural
classes of particulars. As Lewis forcefully puts it, “not every account is an
analysis!” Taking certain facts as primitive and unanalyzed, according to him, is
still a way of making room for them in one’s ontology.
I have taken the debate surrounding universals as itself a paradigmatic case of
the way that the introduction of a new and distinct ontological category often
proceeds. Of course, each debate has its own individual features, parameters,
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

distinctions, and so on. But there are definite similarities among them, too.
Consider, for example, the debate about whether or not we ought to be commit-
ted to causal relations, or bare particulars, or states of affairs, or temporal parts,
and so on. In each of these cases the starting point is a certain target phenom-
enon, a certain undeniable “Moorean fact.” It seems that when I push this table
I cause it to move – what is this due to? Is there something corresponding to my
sense that there is a causal relation between the two events? It looks as if exact
duplicates such as, say, these two qualitatively identical erasers, can share all of
the same qualities but still be two. What makes for distinctness between exactly
similar objects? All sorts of things seem to remain the same, even through
changes in properties: How can something change and yet remain the same?
In all of these cases, the starting point is a demand for an analysis of a certain
undeniable “Moorean fact.” A refusal to provide an analysis is, of course, an
option, but certain dialectical circumstances might make such a choice harder or
easier. If the majority view in a given ontological debate is that a problem is an
 30 Metaphysics

important one and in definite need of being addressed, then it becomes a problem
which is hard to ignore. Not engaging becomes in itself a controversial stance, one
which needs to be explained and justified. In addition, if a number of prominent
philosophers have recognized the problem and tackled it, this adds to the pressure
and makes it difficult to disengage with the problem as stated. At the same time, if
a certain philosophical problem is considered to be a serious one only by
a handful of philosophers, it is easier not to be drawn into a dispute whose
significance many do not recognize. Thus, which question is considered to be
a compulsory one in a given ontological debate is often, to a greater degree than
metaphysicians would like to admit, a circumstantial matter.
Regardless of circumstances and philosophical trends and fashions, what
helps make an ontological question or problem a compulsory and an enduring
one is the way it is stated. The statement of it should be clear and well-
motivated. The theses that give rise to the problem should be made explicit,
as well as any controversial underlying assumptions presupposed by those
theses. Formulating the problem as a puzzle that needs solving can also be
very helpful, as it often sheds light on assumptions that cannot all be held
together.7 The risks of not stating and motivating the problem properly are
great – for we might find ourselves addressing a metaphysical pseudo-problem
and postulating entities aimed at addressing it. By a metaphysical pseudo-
problem, I understand the sort of problem that in its statement commits the
fallacy of a complex question. Such a fallacy is committed when, in posing the
question or problem, one tacitly takes for granted controversial assumptions
which others would not grant if made explicit; or assumptions which are simply
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

untrue; or assumptions which assume a false dichotomy. As a result, in response

to the question so posed, answers, too, are severely restricted by those tacit
assumptions – in other words, if one wants to engage a certain problem or
question, one seems to have no option but to accept those very same tacit and
unsubstantiated assumptions which were presupposed by the question.
A nonphilosophical example of a complex question is someone asking, for
instance: “Did you poison your husband on Tuesday or on Wednesday?” This
question makes three assumptions that need to be established rather than merely
assumed: the first one is that the person addressed has a husband; the second one
is that the person has poisoned their husband; and the third one is that they did it
on one of the two days mentioned. If this person replied, “But I did not poison
my husband!” or “I don’t have a husband” or “I have a husband and he is alive
and well,” one wouldn’t be addressing the question as posed.

7
Sider (Conee and Sider 2007) provides a nice example of this in his articulation of the puzzle of
coinciding objects.
 Ontological Categories 31

I take one of the paradigmatic metaphysical pseudo-problems to be: “Why is

there something rather than nothing?” This question is often taken to be one of the
most fundamental metaphysical problems; it captures the wonder one feels about
there being a universe at all. One way of arriving at the question is via a subtraction
argument. The thinking is as follows: it seems as if there need not have been any
contingent existents – this table, this chair, this office, me, you, and so on. It seems
that any number of contingent beings might not have existed. An entire universe,
which also seems to be a contingent being, might not have existed. In fact, it seems
quite possible that there might not have been anything at all. So how come – what
is the reason – for there being something rather than nothing?
Metaphysicians are usually quick to dismiss as plausible any causal answer to
this question. As Conee (Conee and Sider 2007) points out, the question is not
concerned with physical causes of the origin of the universe; that is, it is not
after the causal story that explains the big bang. Rather, it is after something
more general and more fundamental. The idea is that even if there were
a satisfactory answer regarding what caused the big bang, the metaphysical
question would still remain: What caused the big bang? And why was there such
a cause rather than nothing? The answer to the metaphysical version of “Why is
there something rather than nothing?,” we are told, is not meant to take anything
for granted – no physical laws of any kind, no objects, nothing at all.
Historically, the most popular answers to this question have appealed to necessi-
tarianism of some kind. The answer to the question would then take the form: “the
reason why there is something rather than nothing is because there are beings that
have to be, that exist out of necessity.” But arguments in favor of necessary beings –
whether it be a god, or a number, or a proposition – are problematic in many ways.
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Further still, it is often pointed out that even if we were to find compelling
arguments for the existence of certain necessary beings, the existence of contin-
gents would be open to the same type of question as the initial one, namely: “Why
are there these contingent beings rather than none (or some others)?”
It is not my aim here to engage in any depth with the various answers that
have been given to this question. I just want to pinpoint aspects of the question
which, I believe, make it a poorly motivated “pseudo-problem” and thus one of
those questions that is not compulsory.
First, the phrasing of the question is such that it seems to privilege nothing
over something. To ask “why is there something rather than nothing?” pre-
supposes that “nothing” is a default natural state, as opposed to a something,
which is taken to be surprising and in need of explanation. But, for the
question to be a fruitful one, the naturalness of the nothing state needs to be
established, not merely assumed. Grünbaum (2009) has argued that the
assumption of “the ontological spontaneity of nothingness”(SoN) finds its
 32 Metaphysics

historical origin in the Christian doctrine. Considering nothing as a natural

state seems to be presupposed by the doctrine that mandates that an all-
powerful being acts as a creative (and maintaining) cause of the universe.
He also notes that SoN has been defended by Leibniz and Swinburne as an
ontologically and conceptually simpler alternative. But it is not at all clear that
ontological simplicity is more spontaneously realized in the absence of the
overriding cause. Thus, proponents of the question “why is there something
rather than nothing?” need to provide further arguments for SoN if the
question is to be posed in this way.
There is an additional difficulty with the question when it comes to “nothing”;
it is very unclear how we ought to characterize it, so that it actually presents
a genuine ontological possibility. One suggestion might be to think of it as
a negative fact; but this only leads to more questions: Is it a negative fact of
totality or a totality of infinitely many negative facts? How are we to think of
negative facts – as absences of facts, or negative instantiations of properties by
particulars, or as instantiations of negative universals, or in some other way
entirely? And so on and so forth. It might be hard to address these questions, but
if we are engaging with a genuine problem, rather than a pseudo-problem, some
clarity about these needs to be established.
Finally, the “why is there something” part of the question seems to be
relying on a version of the Principle of Sufficient Reason (PSR). According to
PSR, “for each thing that exists or obtains, there is an explanation of its
existence, a reason that it exists” (Della Rocca 2010, p. 1). Now, for PSR to
present support and motivation for the “why is there something” part of the
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

question, the sort of explanation it is after needs to be made clearer.

Explanations of existence are difficult to evaluate. On the one hand, propon-
ents of PSR are usually quick to reject the idea that they are after a causal
story of some kind. On the other hand, it is not quite clear what sort of
explanatory story is sought in this context. Is it an explanation of how come
there is a specific something? If so, explanations will vary depending on the
existent in question. (There will be different – possibly again, causal – stories
for dinosaurs, persons, artifacts, and yet a different one again for abstract
particulars or universals etc.). Perhaps the PSR in this case is after some kind
of a global explanation of an entire universe; but it is not clear how global
explanations are meant to work. (Are they made up of many partial explan-
ations of individual existents or are they supposed to be an explanation of
a whole universe? If it’s the latter, then one needs to establish first the
existence of such an “entity” as an “entire universe” as something over and
above all of its parts, and as something in need of a different sort of explanation
than all of its parts.)
 Ontological Categories 33

To briefly summarize, the question “why is there something rather than

nothing?,” without further clarification, is an example of a pseudo-problem.
That is, it uses terms such as “nothing” which need an ontological account and
clarification. It also takes the form of a complex question, insofar as it seems to
privilege “nothing” as a default state, and this sort of assumption needs signifi-
cant defense in its own right. And, finally, the question seems to tacitly rely on
a certain form of PSR, a principle which – as I have suggested – itself needs to
be defended, especially in its application in this context.
To be sure, complex questions might be useful and revelatory in some
contexts – such as when used as an interrogation technique by investigators,
or by prosecutors in courtrooms. Such questions may even prove to be rhetoric-
ally effective when employed by psychologists; for instance, a therapist might
ask: “I wonder what prevents you from acting upon your decision?” In fact, such
questions might act as a useful shortcut whenever it is assumed that there is
a shared background context of discussion. But in a philosophical context,
where the context of discussion has not been established and where the tacit
assumptions are not shared and properly argued for, complex questions are
extremely problematic and give rise to pseudo-problems. In such cases, it is
paramount that the assumptions that are being made are brought to light and
examined, as well as whether they are properly supported and whether there is
still a good question to be asked once this work has been done.
This sort of work is particularly important in the context of ontological
discussions. These kinds of discussion are foundational, and if one ends up
introducing ontological categories hastily and on shaky and unchecked founda-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

tions, it will reverberate throughout one’s metaphysical system.

Before moving on to the next section, I will offer a couple of clarifications.
First, I wouldn’t want someone to take away from discussion this that I am
advocating that each and every assumption in one’s system needs to be thor-
oughly supported and that I do not allow for so-called primitives in one’s theory.
On the contrary, as we will in Section 3.2, I acknowledge that such primitives
are necessary; the stress here is only on making such primitives explicit,
especially in the formulation of what one wants to pose as a “compulsory”
metaphysical problem. Second, I do not want to imply that in order to avoid
stating pseudo-problems, philosophers must at all cost state philosophical
problems in some perfectly “neutral” way. Here too, I suggest that transparency
is the best policy. One might, just like Armstrong in advocating for universals,
provide opinionated introductions to a favored philosophical problem in order
to draw readers in and make them “feel” the pull of the problem for themselves.
But in doing so, one must also be careful to make one’s philosophical opinions
and preferences explicit.
 34 Metaphysics

3.2 Ontological Categories and Theory-Building

3.2.1 Ontology versus Ideology

Once one has followed the suggestions in Section 3.1, and hopefully managed to
motivate and state the metaphysical problem reasonably well – perhaps as in the
Armstrongian version of the “one over many” – there is a great deal more to do.
Section 2 showed how one metaphysical problem can be addressed in a number
of ways, and introduced different types of entities. A useful way of thinking about
this is in terms of ontological roles and role-fillers. An entity that gets introduced
into one’s ontology is meant to fill a certain role in the overall theory. As we saw,
universals were introduced to ground the sharing and having of genuine proper-
ties; tropes were introduced to ground the having of properties, but the resem-
blance between distinct tropes was considered not to require a further common
ground. Thus, the initial decision about which categories are introduced depends
on what sort of phenomenon one wants to provide an explanation for, and what
sort of entity the philosopher deems to be best equipped for that job.
Then, when evaluating a metaphysical theory, philosophers often appeal to
Quine’s (1951) distinction between ontology and ideology. Ontology refers to the
entities that the theory claims exist; ideology refers, somewhat vaguely, to ideas,
concepts, and explanations provided by a theory. According to this distinction,
discussion about virtues of a given ontology, descriptions of different ontological
roles, characterizations of different types of entities, the sorting of different
entities into different categories, and so on all fall within the realm of ideology.
When comparing metaphysical theories, philosophers consider how they fare
with respect to both ontological and ideological economy. A theory is onto-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

logically economical if it introduces only a small number of fundamental

entities. This idea captures nicely the gist of Occam’s razor, that is, the famous
medieval dictum that prescribes that “we should not postulate entities beyond
necessity.” On the other hand, a theory is ideologically economical if it has very
few unanalyzable primitive predicates. Philosophers tend to think of both of
these types of economy as desirable and as enhancing a theory’s explanatory
power: an explanation that is ideologically simple and elegant and that postu-
lates only a small number of entities is an ideal one. But there is often a certain
amount of trade-off between ideological economy and ontological economy;
that is, explanations that appeal to only a small number of primitive predicates
often end up postulating a greater number of entities, and vice versa – explan-
ations which postulate only a few fundamental entities will often end up with
a greater number of complicated predicates.
The trade-off between ideological and ontological economy could be seen in
Section 2 in the discussion on different approaches to properties and relations.
 Ontological Categories 35

Postulating universals alongside particulars inflates the realist’s ontology more

than, say, a one-category ontology of tropes or Lewis’s ontology of particulars,
and classes of actual and possible particulars. At the same time, universals were
introduced as entities whose role was to account for the having and sharing of
genuine properties between particulars; thus a boost in ontology came with
a prima facie gain in ideological economy. Trope theorists and class and
resemblance nominalists, on the other hand, found themselves admitting more
of the complex primitive predicates than realists, such as a primitive predicate
of resemblance or, in the case of class nominalists, a contrastive and variably
polyadic predicate of naturalness.
Lewis (1973, p. 87) has made a distinction between two types of ontological
economy: the qualitative economy and the quantitative economy. Qualitative
economy is measured by the number of kinds of postulated entities; quantitative
economy refers to the sheer number of entities, of any kind, postulated by
a theory. Lewis has argued that only qualitative economy matters – a position
that allows him to claim that his ontology is more economical than the realist’s.
Rodriguez-Pereyra (2002), on the other hand, argues that both sorts of economy
matter but that qualitative economy “takes precedence” over the quantitative one.
Thus, for example, a boost in ontology brought about by admitting concrete
possibilia into one’s ontology would reflect a quantitative gain, but not
a qualitative one (since one would continue to have just particulars in one’s
ontology). The worry then might be whether such an application of the distinction
is rather ad hoc and the commitment to infinite concrete possible individuals such
as talking donkeys, flying pigs, unicorns, gods, and so on is not just more of the
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

same plain-old particulars. A challenger could perhaps argue that concrete pos-
sibilia are in fact a different kind of entity altogether. I won’t try to settle this
dispute here; it is just important to keep in mind that there is less consensus about
this aspect of metaphysical methodology than one might assume.
Assessing ideological economy is similarly challenging – it is both important
but also controversial. It is not just complex primitive predicates that are found
to be undesirable but, more generally, any kind of convoluted metaphysical
explanation and theory. What philosophers are after, ideally, are theories that are
clear, elegant, and fruitful. Such theories should only take as unanalyzable
concepts that are clear and self-evident and proceed in a systematic and gradual
way to explain the more complex components of the system in terms of the more
simple ones. The fruitfulness of the theory is also very important and is
measured by how well it explains the phenomenon it was intended to explain,
and whether it can be reapplied successfully to other areas of inquiry.
Let me illustrate this last point with the following example. When a realist
introduces universals to ground resemblance between distinct particulars, it
 36 Metaphysics

may not look good for the fruitfulness of the theory if resemblance between
universals cannot receive the same type of treatment. Examples of resemblance
between universals are as follows: crimson resembles vermilion in being red,
triangularity is more like quadrilaterality than it is like circularity, and both
triangularity and circularity resemble in being shapes. But analyzing these
resemblances among universals by appeal to universals is highly problematic.
If resembling universals are themselves taken to instantiate second-order uni-
versals, it would seem to imply that they would themselves have to be instances
of those universals. And yet this can’t be, since universal properties of triangu-
larity and circularity are not themselves instances of shapes; that is, they are not
themselves shaped. Similarly, crimson and vermilion are not themselves red.
Take, again, triangularity and circularity; these two property universals resem-
ble in that they are both shapes. But they also differ as shapes. Armstrong
himself recognized that this was not a tenable view, since “things cannot differ
in the respect in which they are identical” (Armstrong 1978, p. 106). Thus,
rather than have one and the same thing serve as the ontological ground of
sameness and difference between two other entities, what is needed are distinct
entities one (or more) of which will serve as a resemblance-maker, and the other
which will serve as a difference-maker. Being a shape fails on both accounts: it
is neither specific enough to be able to capture a resemblance-order that exists
between different shapes, nor specific enough to account for what makes
different shapes distinct. Resemblance between universals thus has to receive
a different account entirely from that which applies to resemblance between
particulars and this, arguably, does not look good for the fruitfulness of the
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

realist’s ideology.

3.2.2 Complex and Simple Entities

In metaphysical theory-building, there are entities that are taken to belong to the
most fundamental ontological categories and that serve as the building blocks of
one’s ontology; and then there are entities that are more complex, composed of
the more fundamental entities. For example, particulars and universals can be
taken as the most fundamental ontological building blocks, while states of
affairs can be seen as the more complex entities that are made up of these two.
The more complex entities should not be introduced ad hoc; they, too, require
a well-motivated argument in favor of their postulation. Armstrong, for
example, produces a truthmaker argument for states of affairs. This argument
rests on the assumption that all truths require an ontological ground, that is,
something in the world that makes them true. He then goes on to make a case
that, truthmakers for truths about particular things having properties, and
 Ontological Categories 37

standing in relations to other things, need to appeal to states of affairs. He thinks

the truth of the sentence “this chalk is white,” or the form “a is F,” cannot be
made true by a all by itself construed as what he calls a “thin particular” (since
a does not include all its properties); nor does it suffice to appeal to a universal,
F, by itself. A pair (a, F) won’t do either, he claims, since both a and F could
exist without it being the case that a is F. Thus, according to Armstrong, the
correct truthmaker and ontological ground for such truths has to be the state of
affairs a’s being F; such a state of affairs is made up of the particular a and the
universal F, but is also seen as something over and above the two.
Arguments for complex entities face most of the same challenges as argu-
ments for the simpler ones. Philosophers might disagree with the assumptions
of the argument presented (in the example just given, one might dispute the
truthmaker approach itself based on, say, an alternative theory of truth, or how
the truthmaker argument applies in this context). Or, they may disagree about
whether the introduced complex entity is in fact needed, that is, whether the
explanatory role it is introduced to fill warrants the ontological inflation that
takes place (e.g. one might dispute that the truthmaker argument requires states
of affairs to fill the particular truthmaking role). They can further disagree about
which ontological category the new entity belongs to, that is, whether it falls
within the existing categories or whether it is in a new sui generis category
(when it comes to states of affairs, one might wonder if they themselves are
particulars, universals, or something else). But the new aspect that complex
entities bring into view are questions about the nature of the relationship that the
complex has to the simpler entities that constitute it (in the case of states of
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

affairs this has been a particularly fraught debate, since these entities do not
abide by the model of mereological constitution).

3.2.3 Fundamentality and Primitives

As noted in the previous section, metaphysicians commonly take complexity of an

entity to imply that such an entity has constituents, while simple entities, by
contrast, have no further constituents. States of affairs are thus assumed to be
constituted of particulars and universals, and in this sense they are indeed complex
entities. For many, this also means that such entities cannot be fundamental.8 The
background assumption here is that fundamentality and simplicity go hand in hand.
But this assumption itself must be treated with caution, since there are different
ways in which an entity can be taken to be fundamental.9

8
Bradley (1910) famously puzzled over Russell’s suggestion that an entity can be fundamental and
have constituents. Many more philosophers after him have had a similar thought.
9
For an in-depth treatment of the concept of fundamentality, see Tahko (2018).
 38 Metaphysics

A great deal of the recent literature in metaphysics has been dedicated to the
concepts of ontological dependence, metaphysical explanation, and
grounding.10 All too often it is assumed that metaphysical explanation and
ontological dependence go hand in hand; the thought is that the metaphysical
explanation will follow the chains of ontological dependence, where the latter is
frequently characterized in terms of existential dependence of entities and in
terms of relationship of constitution. In its turn, then, ontological dependence is
characterized as an ontological analogue of metaphysical explanation – thus, the
“ontologically fundamental” entities are considered to be the ones that are
fundamental for explanatory purposes.
Where the two notions become completely indistinguishable is in loose
metaphysical talk of entities explaining a certain phenomenon. But I believe
this to be a misguided way of talking. For how might an entity do any explain-
ing? It seems to me more correct to say that we explain phenomena by appeal to
certain ontological categories or entities; entities themselves don’t explain
anything. Of course, an appeal to certain entities and their features might
make our explanations simpler and easier to follow, while other explanations
might seem more complex and difficult. But regardless of how the explanations
might go, the entities are doing no more explaining than, say, a virus or
a bacterium itself explains the disease that it causes.11
Once we dissociate the simplicity/complexity of an entity from the simpli-
city/complexity of the explanation that appeals to such an entity, we can see that
one need not associate the notion of fundamentality with an entity’s simplicity.
Indeed, it seems useful to distinguish between at least the following three senses
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

in which an entity can be taken to be fundamental: constitutive, explanatory, and

existential. Although some of these types of fundamentality end up picking out
the same entities, they need not do so, for they are quite different. I have
elsewhere suggested the following characterizations:12

Constitutive fundamentality. An entity e is constitutively more fundamental than

entity e* iff e is a constituent of e*, where constitution is construed broadly to
include mereological and nonmereological forms of constitution.

Explanatory fundamentality. An entity e is explanatorily more fundamental

than entity e* iff the definition or a characterization of e**’s ontological role

10
See, for instance, Cameron (2008), Correia (2008), Schaffer (2012), Tahko and Lowe (2020).
11
An anonymous reviewer has pointed out that some philosophers such as Lowe (2018) do indeed
find that essences explain metaphysically. I suspect that what is meant by this claim is simply that
an appeal to the essential properties of an entity helps us understand its nature.
12
See Perović (2016).
 Ontological Categories 39

cannot be made without reference to e, where e is either taken as an explanatory

primitive or can be characterized independently from e*.

Existential fundamentality. An entity e is existentially more fundamental than

entity e* iff e* cannot exist without e, whereas e can exist without e*.

These distinctions help characterize the relationships between simple and

complex entities in more nuanced ways. Applied to our case study of
universals, particulars, and states of affairs, they give us the following
conceptual possibilities. First, particulars and universals appear to be con-
stitutively more fundamental than Armstrongian states of affairs; tropes are
constitutively more fundamental than, say, bundles of tropes, and so on. But,
at the same time, a realist like Armstrong could make a case that facts or
states of affairs are explanatorily more fundamental than particulars and the
universals that make them up; states of affairs can be characterized via
the truthmaking argument – they are the entities that make sentences about
the having and sharing of properties true. One might think that what makes
up such entities is a further question that may have different answers depend-
ing on the type of ontology one embraces. From this we can start to see how
constitutive and explanatory fundamentality might come apart and pick out
different entities.
Existential fundamentality also need not match up with constitutive fundamen-
tality. Some entities might be constituents of other more constitutively complex
ones, but both simple and complex ones might be mutually existentially dependent
upon one another. For instance, in our case study, all three entities – particulars,
universals, and states of affairs – seem to be on a par when it comes to existential
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

dependence. None can exist without the others. Particulars cannot exist but in states
of affairs; universals cannot exist uninstantiated and thus in states of affairs; and, for
an immanent realist, states of affairs cannot exist without particulars and universals.
Existential fundamentality can be further refined in a number of ways. It can
be useful to distinguish, for instance, between the type of existential dependence
that holds between a given circle, c, and its essential property of extendedness,
and a circle, c, and its nonessential property of redness. It seems as if c cannot
exist without it being extended and circular, but it can exist without it being red
(it might be some other color). Various more fine-grained types of existential
fundamentality are discussed in detail by Tahko and Lowe (2020).
To sum up, these distinctions help us see how some entities might indeed be
constitutively complex but also, in a different sense, fundamental and unana-
lyzable. For instance, states of affairs appear to be existentially as fundamental
as particulars and universals; they are constitutively complex (and thus not
 40 Metaphysics

fundamental in this sense); and they may even be seen as explanatorily more
fundamental than both particulars and universals.
Relating to concepts of simplicity and fundamentality is the concept of
a “primitive” in one’s theory. Metaphysical literature frequently loosely appeals
to the use of “primitives” whenever a certain type of entity or concept cannot be
further explained. But here, too, just as in the case of fundamentality, we need to
be careful. Postulated entities, strictly speaking, are not “primitive”; they are
simple or fundamental in one of the above senses. Predicates and concepts, on
the other hand, when they cannot be further explained and analyzed, are taken to
be primitives of the theory. For instance, in one’s theory of time, one might treat
the passage of time as a primitive; that is, it is taken as something that cannot
receive further analysis or a noncircular explanation (i.e., an explanation that
does not refer to the very concept it sets out to explain). This does not mean,
however, that there is nothing more to be said about the primitive. In fact, one
would hope that different descriptions and characterizations can be afforded.
The very idea is to take something intuitively compelling as a primitive notion
that one would hope to be able to describe further; but because it is explanatorily
fundamental, it cannot be further defined or explained away in terms of some-
thing more explanatorily simple.
There are, of course, gray areas, and even strong philosophical disagreements
about whether or not a certain concept or predicate makes for a convincing
primitive. For example, we have seen Lewis argue for the notion of a natural
property, or his notion of a contrastive and polyadic predicate of resemblance to
be taken as a primitive. In such contexts, philosophers appeal to primitives as
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

a way of signaling that explanation does not go any further, that one has reached
the bottom line of a theorist’s ideology.
And yet philosophical inquiry and the need for analysis cannot be hushed so
easily. Philosophers have been known to put pressure on concepts that other
philosophers have taken to be primitive and unanalyzable. Take, for example,
relations. External relations, such as spatial and temporal relations (e.g. “two feet
apart,” “to the left of,” “preceding”) have been taken by many philosophers to be
unproblematic. At one point, Russell (1911) thought that relations might just be
the ultimate constituents of reality and that everything might indeed be reducible
to bundles of universals and relations among them. Other philosophers have been
suspicious of relations, finding them to be extremely puzzling. F. H. Bradley
(1893), for instance, famously argued that relations conceived as external lead to
an infinite regress, while relations conceived as internal, that is, as grounded in
their relata (e.g. “taller than,” “greater than”) also lead to contradictions and an
infinite regress of relations. Regress arguments were a commonly used philo-
sophical tool at the turn of the twentieth century. They were often generated in
 Ontological Categories 41

order to show that using a certain concept or entity to explain a certain phenom-
enon was somehow insufficient; more entities or concepts of the same kind would
need to be introduced, ad infinitum, never arriving at an ultimate reductive
explanation. Bradley’s regress arguments against relations attempted to make
just that point – he wanted to demonstrate that relations could not fill the relating
task in a satisfactory way and that more and more relations would need to keep
being introduced to do the job. The assumption that generated the relational
regresses was that relations could not relate, that they were simply not up to the
task. Thus, the deeper dispute about relations had to do with the question whether
they can actually fulfill the ontological role assigned to them. Once it became
clear that Bradley’s deeper concern was that relations could not relate, while
Russell thought that they could, there was not much of a dispute left. The two
philosophers found themselves at an impasse. Bradley (1911) expressed his final
concern to Russell by asking him to explain further “how relations relate.” This
sort of question has been echoed by contemporary metaphysicians suspicious of
relations’ ability to relate. Some philosophers respond to this with the claim that
this is just what relations do; others find such an answer deeply frustrating and
unsatisfactory. The former then insist that they have an entity in their arsenal that
fulfills a certain role, and that the way that it fulfills that role is not something up
for dispute; how an entity fulfills its role is a primitive of a theory. Opponents, on
the other hand, stress that the “how” question is a compulsory one, and that
without addressing it adequately the entire matter is shrouded in mystery.
As this example illustrates, what makes for a good primitive concept in one’s
theory is often a controversial matter. Intuitiveness is a notoriously unreliable guide
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

in philosophy. Among philosophers, intuitions are highly variable and change over
time. In addition, one needs to be careful about how far to push the “why” and
“how” questions in philosophy. What seems to power questions of the type “how
does a relation relate?” is reasoning similar to that in the PSR. This principle states,
roughly, that everything must have “a sufficient reason.” But this, of course, is too
vague. The “sufficient reason” sometimes refers to an explanation of a certain
phenomenon, at other times to a physical cause, and at yet other times to onto-
logical ground. These are all clearly very different ways of articulating a “reason”
for something. What they all have in common, and what I assume the PSR tries to
capture, is the sense that there ought to be some sort of explanation for every thing/
occurrence/phenomenon. This kind of search for an explanation well encapsulates
what lies at the core of our attempts to understand ourselves, the world around us,
and our place in such a world. It is the sort of wonder that has led to various
scientific discoveries, both large and small. But if the PSR is only meant to capture
something as vague as “human sense of wonder” and “search for explanations,”
then it is indeed too broad and vague to do any real philosophical work.
 42 Metaphysics

We can be in favor of searching for explanations of a variety of phenom-

ena and also recognize that some concepts should be taken as primitive. We
can appreciate the importance of wonder and curiosity, while recognizing
that some questions are better formulated and more fruitful than others; and,
indeed, that some questions might just be confused, misplaced, or even give
rise to pseudo-problems. In doing fundamental ontology, metaphysicians
sometimes have a difficult time walking the thin line between rightful
questioning of certain fundamental assumptions, on the one hand, and
risking sounding like a small child who won’t stop asking “why” of any
explanation they are provided with, on the other. Unfortunately, PSR is of no
help in distinguishing between the two sides of the line. If it is to capture
a particular kind of request for an explanation, it should be made more
specific. If it is to remain broad and vague, it confers no particular weight or
legitimacy to its demands for reasons and explanations.

3.2.4 Naturalistic Considerations

In providing reasons for postulating a certain kind of entity, some philosophers

have appealed to naturalistic considerations. Armstrong has famously appealed
to the Eleatic principle which dictates that one should only believe in entities
that are causally efficacious. But causal efficacy can be understood in at least
two ways. One way is to include entities that play an important role within
a causal explanation; another way is to postulate entities that themselves must
causally partake in the natural world. On the first interpretation, naturalism
would come to encompass entities that need not themselves be causally effica-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

cious and need not be part of the spatio-temporal physical world, but that
nonetheless feature in explanations of such a world. Such entities might be
numbers, propositions, universals, thoughts, and so on. This understanding of
naturalness in ontology is probably too permissive and not what most philo-
sophers are after when they talk about naturalistic ontology. But if interpret an
appeal to naturalism as a way to include only entities that are themselves
causally efficacious, this can seem both too restrictive and too vague. It can
be seen as too restrictive in the sense that it might leave out too many entities
that play an important explanatory role in a metaphysical theory – such as
possibilia, substances, selves, and so forth – while at the same time not being
causally efficacious in a straightforward way. The vagueness is also a problem –
causal efficacy of an entity would need to be spelled out further; for instance,
is causal efficacy to apply only to things or also their properties and, if so, which
account of properties is presupposed? Armstrong, for example, believed that
only scientific universals – those that feature in fundamental physics – should be
 Ontological Categories 43

allowed. But this, of course, potentially leaves many other perfectly good
candidates out: universals from different sciences that are not fundamental
physics as well as the category of universals itself seem to be left out of the
picture. In any case, if appeal to naturalism is to play an important role in
deciding which sorts of ontological entities one should admit, then at the very
least one should make sure that the sense of “naturalism” is properly delineated,
as well, of course, as the particular sense in which causal efficacy is to be
understood in that philosopher’s theory.

3.2.5 Ontological Categories, Entities, and Roles

The phrase “ontological categories” is commonly used to describe the most

general groupings of everything that exists; “ontological entities,” on the other
hand, ordinarily refers to the entities that are taken to fall under those general
groupings. When it comes to universals, the two terms are often used inter-
changeably. The reason, as we have seen, is that realists tend to introduce
universals as the type of entity that itself sorts other entities into groups, as
well as the entity that provides the ontological ground of resemblance between
the distinct particulars within those groupings. These two roles that universals
play has been present from their inception, and we have seen it on display in
Section 2 in the way that Plato introduced his Forms.
But strictly speaking, the two terms can and do come apart, and should be kept
separate. The term “ontological category” is mostly used to pick out a type of
entity which has certain distinctive features, though those features need not be
universals. This is something that nominalists and trope theorists are more
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

accustomed to, since they do not admit universals to provide the ontological
ground of genuine resemblances in nature. Instead, they tend to think of resem-
blance between such entities as holding due to those entities’ natures, where the
latter can be understood by appeal to, say, nonrepeatable property tropes.
Distinction between ontological categories and entities that fall under such
categories is important to maintain for realists, as well. This is most clearly seen
when examining the issue of the most general categories. To help explain the
issue, I will utilize the determinate/determinable distinction introduced by
Johnson (1921). Johnson outlined five main features of the relationship between
determinate and determinable:

(1) if a particular has a determinable property, then it also has some determinate
property that falls under that determinable (for instance, if a is red, it may
not be crimson nor scarlet but it must be some specific shade of red; or if
something has a shape, it must have a specific shape such as
rectangularity);
 44 Metaphysics

(2) if a particular has a determinate property, then it also has the corresponding
determinable property (if a is crimson, then it is automatically red);
(3) a particular cannot have more than one determinate property of a specific
determinable at the same time (a cannot at once be crimson and scarlet all
over, weigh 4 kg and 5 kg, be triangular and quadrilateral);
(4) the relationship between determinable and determinate is not to be confused
with the relationship between species and genus (the definition of species
involves differentia which is an independent property from genus, whereas in
defining determinates there is no such independent property); and
(5) within a class of determinates that fall under the same determinable there
are resemblances and these resemblances can be ordered (scarlet resembles
vermilion more than it resembles crimson, for instance); this, however, is
not the case with “highest” determinables.

As these features illustrate, determinate and determinable are introduced as

fundamentally relational categories, defined in terms of one another. They are
notions that do not seem to unequivocally fix their referents, since what is
a determinate with respect to one determinable can itself often be a determinable
with respect to another, “lower,” determinate.
Now, when it comes to the most general groupings of entities – the most
general categories – it might be tempting for a realist to appeal to determinable
universals such as having shape, being extended, or perhaps even being
a universal or being an entity. But the challenges for taking such high determin-
able properties to be universals are serious: one challenge has to do with what
one should take the bearer of the higher-order determinable universal to be
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

(should it be a particular or a first-order universal, and how should either of the

two accounts be spelled out?); but the more pressing challenge in this context
has to do with how undiscriminating such highly general determinables actually
are. Being extended, being a universal, and being an entity appear to be too
general for the role of resemblance-making; indeed, from that perspective, it
doesn’t seem right to think of them as universals at all.
I thus suggest that the best way to describe categories for realists, trope
theorists, and nominalists alike, is by appeal to ontological roles. This approach
was captured well by Oliver (1996), and I believe that such an approach most
clearly acknowledges that coming up with categories of entities is not done
outside of a specific metaphysical context. In metaphysics, we find many prob-
lems/puzzles that need to be solved, and it is in addressing such problems that the
appeal to ontological categories plays an important part. When ontological
categories and entities within them are charged with specific explanatory roles,
their introduction can be properly motivated and critically assessed.
 Ontological Categories 45

The assumptions present in the search for ontological categories, as well as

the assessment of different category systems, become more transparent within
such a framework. We can question, for instance, whether there is a need to
introduce the category of universals or not, and whether the ontological role that
universals are meant to fulfill needs to be filled, and if so, with what kind of
entity.
The way in which one ontological role is characterized will in turn determine
other ontological roles that need to be filled. For example, when universals are
characterized as resemblance-makers, particulars tend to be characterized as
distinctness-makers: they provide the ground of the multiplicity of numerically
distinct entities. The need for a role of this kind can be gleaned by considering
the problem of individuation. Namely, if the realist about universals wanted to
just have universals in their ontology, they would have to analyze particulars
such as, say, an apple, by appeal to property universals like juiciness, redness,
sweetness, and so on. Thus, a particular red and ripe apple would just be
a bundle of property universals. But it quickly becomes apparent that there
can be two or more exact duplicates that share all of the same property
universals – two or more apples that have exactly the same juiciness, redness,
and sweetness. What then makes it so that there are two (or more) distinct apples
rather than just one? If one is committed only to universals, and universals are
seen as repeatable entities that are wholly present in each of their instances,
what metaphysically distinguishes one bundle of universals from another
exactly resembling one? What prevents all of the apples that exactly resemble
one another in the universals they instantiate from collapsing into just one
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

bundle of universals? These questions motivate many realists to explicitly

postulate two categories of entities – particulars and universals – rather than
just universals. Particulars’ role is thus to help solve the problem of individu-
ation by providing an ontological ground of distinctness of entities. Particulars
can, of course, play additional roles and be used to solve further problems,
similar to how we saw universals supported by further arguments in Section 2.
For instance, one might appeal to them as what underlies change in properties
through time, what bears properties and stands in relations to other particulars,
and so on.
We can see that the main ontological roles that particulars and universals are
assigned to fill seem well-defined and exclusive; when it comes to resemblance-
making and difference-making, entities typically fall into one of the two
categories – they are either particulars or universals, not both at once. This is
not to say that a more complex entity – such as a state of affairs – cannot contain
both of these entities. This is indeed how states of affairs are understood: they
are entities made up of a particular having a certain universal (Fa), or two or
 46 Metaphysics

more particulars standing in a certain relation (aRb). But a state of affairs is

ordinarily itself conceived as a nonrepeatable entity – a particular. Armstrong
referred to this feature of states of affairs as “victory of particularity.”
The same entity can be cross-categorized – indeed, it often will belong to
a number of different ontological categories. Take, for example, an abstract
universal being a prime number; this entity belongs to a category of universals
as well as to a category of abstract entities. Or take the event of Lulu’s 9th
birthday party – this event can be seen as a nonrepeatable entity that falls under
the category of “particular,” but is also a concrete spatio-temporal event with
a certain duration. And so on and so forth. Different metaphysical systems will
make different choices about what they take to be the most fundamental
ontological categories as well as what they take to be the ontological entities
within those categories, and what their features are. For instance, a universal
will be seen as an abstract entity by some, while it will be understood as
a concrete spatio-temporal entity by others. An event might be seen as a sum
of static temporal parts by proponents of the four-dimensional approach to
persistence, while it will be understood as an intrinsically dynamic temporally
extended state of affairs by some proponents of the three-dimensional approach
to persistence.
Not only will there be differences among the entities admitted and their
characterizations, but the way that distinctions between entities are drawn up
will vary as well. Most philosophers believe that ontological roles need to be
defined neatly and exclusively, so that there is no confusion about which entities
are introduced to fill which roles. Property universals ground resemblance
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

between what appears to be sameness of qualities between particulars; particu-

lars make for numerical distinctness; relations relate; and so on. But these neat
divisions of roles, though desirable, are not always carried out. Sometimes the
roles are found to be not defined well enough and/or are not exclusive enough.13
Take, for instance, Russell’s (1911, 1992) characterization of relations as
entities that can fill two roles, depending on the situation they are in: in some
situations they relate, in others they are merely terms of the other relating
relation that relates into a complex. Still, even on this view, it is a unique feature
of relations that they can fill both of these roles, which is what distinguishes
them from just particulars that can only be terms. Now, I don’t want to suggest
that such multiple-role characterizations are optimal. In fact, Russell faced
objections to this characterization of relations and found it difficult to explain
what spurred the exhibition of such different roles within relations, as he
13
For instance, Frank Ramsey (1925) famously argued against a number of different ways of
drawing up the particular–universal distinction; in recent literature, this challenge has been
elaborated upon and amplified by Fraser MacBride (2005).
 Ontological Categories 47

understood them. But it does seem as if it is possible to assign to the same entity
multiple ontological roles, although not opposing roles at the same time.
Once the ontological role(s) are specified, philosophers can start testing the
adequacy of the different candidate entities for the assigned roles. If the
ontological role to be played is a property role, then we might have different
candidates such as classes of possibilia, universals, tropes, and so on. If the role
is “ontological ground of distinctness of entities,” possible role-fillers might be:
bare particulars, substrata, states of affairs, tropes, ordinary particulars, and so
forth. One might wonder if very different entities might end up playing the same
ontological roles equally well, thus giving rise to equally good but incompatible
rival ontological systems.
Armstrong and Lewis were remarkably intellectually honest in their debate
about fundamental ontology. Given the balance of trade-offs among different
theories of properties, for instance, they didn’t always think that there was
a straightforward winner among them. Armstrong frequently talked about
how such matters could be decided “in the end game”; that is, which theory
will be deemed a better one would have more to do with its overall explanatory
power in a number of fields of metaphysics, and not just in the debate about
properties.
I believe that this seems right. Given the unlikely situation in which two
rival ontologies explain a certain phenomenon or address a certain metaphys-
ical problem equally well, there is certainly more work to focus on in other
areas of metaphysics. This is why the debate between nominalism, realism,
and trope theory doesn’t just end with the account of properties and relations;
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

it also has to do with accounts one might offer, by appeal to these entities, of
modality, laws of nature, propositions (if there are such), and so on.
Systematic metaphysics is a difficult, gradual, and patient endeavor; it takes
a great deal of thought to develop an ontology that can be employed in various
areas of philosophy, and that can be used to address a number of philosophical
problems.
This last point brings us back to the initial considerations of this section:
namely, the importance of making sure that metaphysical problems are properly
motivated and clearly stated. For if we don’t have a certain amount of philo-
sophical consensus on what the main problems are that need to be addressed, it
is difficult to have agreement on which ontological roles need to be filled, since
it is the problems that determine the ontological roles and the ontological roles
that determine the possible categories. Absent agreement on the problems, and
absent some form of general agreement on the methods, it is very difficult to
make progress in ontological system-building and the evaluation of rival
ontologies.
 48 Metaphysics

3.2.6 Some Other Ways of Approaching Ontological Categories

My preferred way of approaching ontological categories in this Element has
been by appeal to ontological roles that different kinds of entities play in solving
certain metaphysical problems. I believe that this bottom-up approach high-
lights the tight connection that exists between first-order ontological inquiry
and second-order methodological inquiry. The move from the first-order
domain to the meta-domain is thus not an inadvertent confusion; rather, it is
in itself a methodological choice to anchor the meta-debate about categories to
concrete examples and explicit goals of philosophical problem-solving.
Grossman seemed to have something similar in mind. In his The Categorial
Structure of the World, he states that the task of ontology is to try to answer the
following two questions: (1) What are the categories of the world?; and
(2) What are the laws that govern these categories? (Grossman 1983, p. 3).
This is not too dissimilar from the goal of chemistry, he says, with its search for
the fundamental chemical elements and the laws they obey. At the same time,
Grossman notes that ontology is not a science among other sciences; it has
a larger scope. One starts to do ontology, he says, when one realizes not just
“that there are not only different kinds of individual thing, but also different
kinds of entity” (Grossman 1983, p. 3). In The Categorial Structure of the
World, Grossman painstakingly elaborates on different categories – individuals,
properties, relations, classes, structures, multitudes, and so on. But when faced
with the question “what is a category?” he only says this:

It is a kind of entity. What kind of kind? In answer to this question, we can

only give examples. It is that sort of kind, as we have seen, that distinguishes
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

between individuals, on the one hand, and properties on the other. It is that
sort of kind . . . which obeys a certain kind of law, namely, categorial laws.
But this reply does not really help much either. We must therefore rest
content, as on so many other occasions, with examples rather than definitions.
In these most fundamental matters of metaphysics, definitions are impossible.
(Grossman 1983, p. 5).

Simply put, categories are kinds of entities, not entities themselves. Which
kinds of entity? This depends on the ontology we develop. No general theory
of kinds of entities is to be had, according to Grossman. In fact, he believes that
“disputes about the true nature of ontology, just like similar disputes about
philosophy as a whole, are singularly barren and tedious. The proof of the
pudding, as the saying goes, is in the eating” (Grossman 1983, p. 18).
Some philosophers strongly disagree with this bottom-up approach to onto-
logical categories. What motivates a top-down approach is pretty straightforward.
If we take the main job of ontology to be to determine the correct list of
 Ontological Categories 49

ontological categories, then it seems central to this task to first articulate what it is
we are looking for – that is, what ontological categories are in the first place. Take,
for instance, Westerhoff’s (2005) exploration of the issue. Right at the outset he
finds it important to keep ontological questions separate from meta-ontological
ones. The two central meta-ontological questions, according to Westerhoff, are:
“1. What kind of things are ontological categories? What makes the categories
incorporated in the seven systems discussed ontological categories?” and
“2. How can these categories be related?” (Westerhoff 2005, p. 20). For him,
these two questions need to be distinguished from object-level ontological
questions: “1’. Which ontological categories should appear in an ontological
theory?” and “2’. How are these categories in fact related, i.e. which are included
in one another, which are coextensional etc.?” (Westerhoff 2005, p. 20).
Whereas Westerhoff is right that the two sets of questions are distinct, it
doesn’t seem to follow from this that the considerations involved in the second
set should have little or no bearing on the first. Indeed, I take it that the most
relevant meta-ontological questions are brought about by the first-order ques-
tions, and how one goes about answering the former will be very much influ-
enced by how one goes about answering the latter.
A philosopher can choose to stay at a meta-ontological level, list different
category systems provided by different philosophers, and then proceed to ask,
following Westerhoff: What, if anything, do all these category systems, articu-
lated by different philosophers, have in common? What are all these philo-
sophers engaged in, when they offer a list of categories? Take, for example,
Aristotle’s list of categories, mentioned in Section 1.1, which appeals to sub-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

stance, quantity, quality, relation, place, date, posture, state, action, and passion;
Grossman’s (1983) categories of simple entities (individuals, numbers, proper-
ties, relations) and complex entities (sets, structures, facts); Lowe’s (2006)
categories of particulars (objects, modes) and universals (kinds, attributes);
Chisholm’s (1996) categories of contingent entities (individuals, states) and
necessary entities (states, nonstates); and so on. To ask “what makes different
categories in these systems ontological categories?” risks, I worry, being
a pseudo-question. To ask about the commonalities between different category
systems assumes, without demonstrating, that there are indeed such common-
alities, and that an appeal to them can explain what makes something an
ontological category.
But the trouble here is that different philosophers often mean very different
things by each of the listed categories, and the lists themselves can serve very
different categorizing purposes. Now, even if we found that there are in fact certain
commonalities between the distinct category systems proposed – say, their appeal
to particulars, properties, relations, and so on – this by itself does not establish any
 50 Metaphysics

deep meta-ontological fact. The commonalities might only be apparent – philo-

sophers might use the same terms but articulate what they mean by them in very
different ways – or they might mean similar things by their use of the term
“ontological category” as well as what they mean by invoking their preferred
categories such as “particulars,” “universals,” “relations,” and so on, but what
such common usage demonstrates will itself be up for interpretation and discussion.
An apparent commonality in the use of terms might be due to similar patterns of
thinking about categories, or it might be due to shared understanding of the implicit
application conditions of the terms. In other words, it seems to me that without
engaging with bottom-level ontological inquiries that undergird the introduction of
these particular categories rather than some other ones, the meta-inquiry can at most
provide us with a descriptivist account of what the term “ontological category” has
come to mean in certain philosophical contexts.
In contrast with a descriptivist approach, the prescriptivist top-down project
tries to determine what metaphysicians “ought” to mean by an “ontological
category.” I take this to be a more promising approach and it is one that
McDaniel (2017) is favorable to in his recent work The Fragmentation of
Being, where he argues that ontological categories should be seen as ways of
being. He finds this approach present already in Aristotle and Aquinas, though
the reference to them does not exactly help clarify the ambiguities inherent in
the phrase “way of being.”
It seems to me that one way of understanding the phrase is “a way of
existence.” This then leaves the door open for the view that various things can
be ascribed different types of existence, and perhaps even different degrees of
existence – for example, one might think that an abstract object such as
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

a number has a different degree of existence than a concrete object such as

a chair. This reading faces familiar difficulties of ontological overpopulation
raised by treating existence as a predicate or a property. Another way of
understanding the phrase “way of being” is as referring to one and only one
being, a monistic whole of some sort, and then to think of this whole as the
British idealists did, as appearing to us under different guises. Another way still
is to just think of the phrase as a thinly veiled reference to most general
properties, the highest determinable properties – properties, particulars,
abstracta, and so on. Or, perhaps all that McDaniel has in mind is that there
are some essential properties that determine a kind that ontological entities
belong to, and that these kinds are the highest ontological categories. He writes:

The view defended here respects the intuition that the fact that an object
belongs to a particular ontological category is a deeper fact than any fact
concerning the properties had by the object. The category that an object
 Ontological Categories 51

belongs to is not just another property among many had by the object, but
rather is ontologically prior to any property had by the object. Properties
partition the beings in the world. Ontological categories partition being itself.
(McDaniel 2017, p. 124)

The language of partitioning being itself is not particularly helpful. But, setting
that aside, McDaniel’s broader intent is certainly to allow for different ontolo-
gies to fill out the details as they see fit. Some ontologies might be very coarse-
grained in their appeal to only one category of entity (be it an individual or
a trope), others might be very fine-grained (with many categorical subdivi-
sions), and then there is everything else in between. This view also leaves open
how one might analyze properties themselves (by appeal to natural classes,
tropes, or universals).
In E. J. Lowe’s (2006) The Four Category Ontology we find a more robust
prescriptivist top-down approach to categories. Lowe describes ontology as
being concerned, at its heart, with “what kinds of things can exist and co-
exist” (Lowe 2006, p. 5). He immediately clarifies that by “kinds of things”
he means categories, and that by “things” he means entities. Central to this
project is thus determining which categories should be admitted and how
they should be individuated, that is, identified and distinguished from other
categories. Such categories are organized hierarchically – at the very top is
the category of “entity” and right beneath it are particulars and universals,
according to Lowe. The most fundamental level, however, is the third level,
at which we find Lowe’s four categories: kinds, properties (and relations),
objects, and modes. He then goes on to define each of the four categories by
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

appeal to two formal relations of characterization and instantiation. These

relations are “formal” in the sense that, for Lowe, they are not genuine
external relations and thus do not themselves fall into one of the categories;
they are not themselves entities, though they do help explain how entities
depend on one another. They are akin to internal relations, understood as
capturing a stronger type of dependence than is usually found in internal
relations.14 Lowe grants a similar status to categories themselves. He
explores whether categories could somehow fit into his system – as perhaps
kinds or particulars – and quickly concludes that, given how each of his
categories is defined, category itself is just not an entity that belongs in
either of his fundamental categories. He concludes this discussion
as follows: “[T]he only acceptable thing to say . . . is that the ontological

14
Internal relations such as taller than supervene on the natures of the relata, in this case height.
Lowe’s formal relations capture internal dependence relationships which are much stronger. The
example he gives is of Fido and doghood: Fido could not have existed in the absence of
doghood – the two are “made for each other,” so to speak (Lowe 2006, pp. 46–47).
 52 Metaphysics

categories are not themselves entities and are thus not to be included in an
exhaustive inventory of what there is. There are, quite literally, no such
things as ontological categories” (Lowe 2006, p. 43).
This fact should not, however, be taken to threaten metaphysical realism in
any way. Lowe explains:

Does this mean that no ontological system can have a realist foundation?
Does it mean that ontological categorization is all just a matter of how we
choose to classify and describe things – of how we choose to “carve up”
reality, to use the rebarbative metaphor so often favored by anti-realists? Not
at all. The difference between, say, an object and a property, or between an
object and a mode, is as fundamental, objective and real as any difference
could possibly be. (Lowe 2006, p. 43)

In other words, categories need not be entities for us to be realist about them.
What they capture are real distinctions between types of entities – such as
Armstrong’s particulars and universals, or Lowe’s modes and objects – and
such distinctions are due to the intrinsic natures of such entities; they are not due
to us and our ways of describing them. To categorize correctly is simply to
categorize according to the existence and identity conditions of things that are
before us. This exercise is done purely on a priori grounds, which makes it very
different from the taxonomies we find in sciences such as biology or chemistry,
Lowe notes.
Thus, for Lowe, ontology is a science of the most general and abstract
kind – it is a science of being. In doing ontology, we are trying to discover
correct ontological categories which are based on the metaphysical natures of
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

entities.

Why do we need a “science of being”, and how is such a science

possible? Why cannot each special science, be it empirical or a priori,
address its own ontological questions on its own behalf, without recourse
to any overarching “science of being”? The short answer to this question
is that reality is one and truth indivisible. Each special science aims at
truth, seeking to portray accurately some part of reality. But the various
portrayals of different parts of reality must, if they are all to be true, fit
together to make a portrait which can be true of reality as a whole. No
special science can arrogate to itself the task of rendering mutually
consistent the various partial portraits: the task can alone belong to one
overarching science of being, that is, to ontology. (Lowe 2006, p. 4)

This passage captures nicely the traditional realist approach to ontology, with
its bold overarching goal of discovering the most general mind-independent
categories of being. This approach is an admirable and inspiring one; but, as
I will argue in Section 4, it can seem alienating to anti-realists, on the one hand,
 Ontological Categories 53

and prohibiting to more modest, revisionary, and piecemeal approaches to

ontological categories, on the other.

4 On the Value of Exploring New Types of Ontological

Categories
This section is more speculative and tentative in spirit. It takes a look at ways in
which philosophers might break with tradition and explore new ontological
categories. I start by examining one common way that the discussion between
an ontological realist and an ontological anti-realist about categories plays out,
and I point out aspects of the dialectical exchange that I find wanting. The anti-
realist attacks tend to trivialize and misrepresent the endeavor that ontological
realists are engaged in, while realists often end up alienating the more scientif-
ically minded philosophers with their insistence that they are describing ultim-
ate, immutable, and metaphysically necessary features of reality. I suggest
a more modest, cautious, revisionary form of realism which might accommo-
date some of the anti-realist’s concerns, without giving up on the metaphysical
realist’s overarching goals. I discuss the importance of trying to break out of
“traditional” systems of categories and explore “new” categories.

4.1 Realism versus Anti-Realism about Ontology and Its Categories

The prevalent assumption when philosophers are devising systems of onto-
logical categories is that they are getting at “reality as it truly is.” The kinds of
entities that exist and their features are assumed to be independent of us; if we
succeed in reasoning well about reality and its categories, we are discovering
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

the true features of mind-independent reality. This, in a nutshell, is realism.

Anti-realism disagrees; it claims that the ontological categories that philo-
sophers come up with are not getting at mind-independent reality, but are rather
describing our own conceptual apparatus. According to anti-realists, it’s not the
case that there is a mind-independent reality, but that we cannot properly get at
it; this would still be a form of realism, just a pessimistic kind. For anti-realists,
there simply is no mind-independent system of categories to be had.
We can more easily see how one might be an anti-realist about aesthetic
qualities such as beauty – these are often taken to lie in the eye of the beholder;
but it might seem more difficult to argue for anti-realism about things such as
planets, dogs, trees, water molecules, and such. The latter do not appear to
depend upon a judging mind for their existence. Anti-realists can counter such
observations in a couple of ways: one is to argue that genuine agreement on the
things that realists assume must be uncontroversial does not exist (simply, there
are no clear-cut cases to be had; dogs, trees, and planets are no different than
 54 Metaphysics

mind-dependent qualities such as color or taste). Another way for an anti-realist

to proceed is to accept that there are genuine commonalities among people’s
judgments, but argue that this form of agreement is merely a contingent feature
of human perception, cognition, and psychology.
It is not my aim here to engage in depth with the different types and facets of
the realist–anti-realist debate as it applies to ontology. What I do want to bring
out, however, are the different values that a realist and an anti-realist end up
placing on ontological categories and ontological commitments. It is not
uncommon to hear realists worry that ontology, and metaphysics more broadly,
would lose its value within an anti-realist framework. “If we were merely
engaged in describing how the world appears to us, and if we were merely
describing our own conceptual schemas, then ontology would no longer be
a study of the most fundamental categories of being and it would lose its value,”
the realist’s reasoning goes. “This is why we are engaged in finding out the truth
about how reality is. We are after the godly perspective, the sub specie aeterni-
tatis view of reality,” they add. Anti-realists, however, often seem completely
unfazed by this: “In that case, so much worse for ontology,” they say. Realists’
insistence on devising ontological categories strikes them as an idle endeavor,
with no real consequences for anyone except those interested in a very narrow
metaphysical debate. This is, of course, a bit of a caricature, but it does illustrate
well what is at stake. Now let’s look at a concrete example.
In their recent article, Bueno et al. (2015) advocate for an ontology without
categories; in doing so, the authors reenact a version of the exchange just
described. They specifically take aim at Lowe’s (2006) four-category ontology
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

and his arguments for the realist stance about ontological categories. The
authors highlight Lowe’s repeated appeal to indispensability arguments for
the categories that he proposes:

The indispensability argument is central to Jonathan Lowe’s general defense

of the fairly traditional project of metaphysics, but also to his articulation of
the need for a metaphysical type of possibility and for the grounds for
embracing distinctively metaphysical objects of the sorts covered by his
four primary ontological categories: kinds, attributes, modes, and objects.
Each plays a metaphysical role that it and only it can play, not only within the
philosopher’s domain of metaphysics, but also in the various domains of the
specialized sciences which, according to him, could not be as they are, were
there none of the objects with which he populates his four-category ontology.
(Bueno et al. 2015, pp. 234–235)

Bueno et al. (2015) find such pervasive appeal to indispensability unconvincing.

They don’t much engage with the specifics of the case that Lowe makes for the
introduction of each of the categories; their main focus is on Lowe’s claim that
 Ontological Categories 55

his proposed categories are indispensable for scientific enterprise. For example,
Lowe criticizes regularity accounts of laws of nature as unsatisfactory and
believes that conceiving of laws of nature in terms of relations between univer-
sals provides a better foundation for their generality and predictive power. The
authors reply to this is that Lowe’s proposal is “unlikely to be recognizable to
practitioners of science as a proper solution to the problem” (Bueno et al. 2015,
p. 242). Lowe also believes that his ontological categories are routinely presup-
posed by scientific descriptions of the world, as these make pervasive appeals to
kinds, objects, properties, and so on. But the authors find this claim to be
extremely problematic. They say that Lowe’s categories are indeed dispensable,
since they are of no concern to a practicing scientist.
I find that there is something strange about these arguments. It is as if Bueno
et al. (2015) are purposely missing the point that a realist is trying to make:
namely, Lowe is not claiming that all scientists need to do metaphysics first, by
reading, for instance, his account of ontological categories as a prerequisite for
their scientific research. Nor is he claiming that all scientists need to learn about
universals and his own account of laws of nature. The uncharitableness of
interpretation that anti-realists are engaged in is striking. What Lowe is claim-
ing, however, is that scientific reasoning presupposes the very categories that he
has outlined; he claims that if scientists were to ask themselves foundational
questions about features of laws of nature, they could use some metaphysical
background on universals. The types of criticism of ontological realism that
appeal to what scientists do in their practice are thus entirely beside the point.
Perhaps the upshot of these anti-realist arguments is just that Lowe, and
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

realist approaches to ontology more generally, are overplaying their hand. There
are no genuinely indispensable categories, claim anti-realists. We need categor-
izations, since we need to sort and categorize things, but we should not assume –
the claim goes – that the entities subsumed under such categorizations pick out
mind-independent entities. “Categorizations without reification” is their motto;
Bueno et al. (2015) believe that abstract concepts rather than categories can do
the categorizing job well enough. Concepts are flexible, easily swapped around,
eliminated, and introduced as needed. Categories, in contrast, are not so flex-
ible, according to Bueno et al.; they seem rigid and permanent.
Now, what should be noted is that these anti-realist concerns seem to rely on
a particular understanding of the nature of truth, which is antithetical to the
realist conception. The backdrop to the realist’s commitment to ontological
categories is some form of correspondence theory of truth. In truthmaker terms,
the assumption can be put simply as follows: there are truthbearers, assertions
that are either true or false, and truthmakers, worldly facts or states of affairs
that make those sentences true. For example, “this piece of paper is white” is
 56 Metaphysics

made true by the way things are, in this case, by this piece of paper indeed being
white. Whereas we can express the meaning of a given sentence in different
ways, truthmakers are what they are, they exist independently of our minds, as
does the truthmaking relationship between truthbearers and the facts that make
them true. One can provide different ontological accounts of truthmakers
(nominalists might appeal to particulars and their class membership; trope
theorists will appeal to tropes and trope-bundles; immanent realists will appeal
to particulars, universals, and possibly states of affairs), but the assumption is
that there are some such entities. One can also take different attitudes toward the
scope of the truthmaker approach –that is, whether one ought to be a truthmaker
maximalist and take it that all truths must have truthmakers, or argue more
modestly that perhaps some truths, such as negative truths or general truths, do
not have a truthmaker – but the assumption is nonetheless that truthmaking
applies widely.
In any case, anti-realists must reject all of this. They must reject correspond-
ence and assume some other theory of truth – perhaps a coherence theory, or the
pragmatist theory, or the deflationary theory. But such theories are notoriously
incapable of explaining what it is that makes sentences true and what might
guide one in evaluating the truth of a given sentence. Such theories are also quite
incapable of explaining what it is that makes certain concepts better than others
for categorizing purposes. If a certain concept is more useful, or serves a certain
purpose better than other concepts, how should such usage or purpose be
characterized and what makes one concept better than another? Why apply
the concept “rectangularity” to all of the things we perceive to be rectangular?
Or what do all of our appeals to “objects” have in common? If there is no appeal
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

to what the concepts pick out, and no specified criteria of applicability of

concepts, then it becomes very difficult to have a meaningful conversation
with an anti-realist. If, on the other hand, anti-realists have already tacitly
assumed that certain concepts are better than others because they pick out
certain features of reality better than others, then they seem to have relied
upon some form of correspondence theory after all. In that case, the dispute
with the realist is no longer about whether or not we ought to admit ontological
categories, but rather about which ontological categories should be admitted.
Thus, rather than locate their disagreement at the general level of concepts
versus categories, many anti-realists might actually just be disagreeing about
particular types of reifications – they might not believe in universals and might
find the entire debate about their features, multiple occurrences, instantiations,
and so forth pointless and off-putting. Similarly with relations, and perhaps
some other types of ontological postulates. Or, maybe they simply do not
believe in the existence of the most general determinables – such as being
 Ontological Categories 57

a property, being a relation, being an object, being a kind, being an abstract

entity, being a concrete entity, and so on. But this too is an intratheoretical
dispute about which ontological categories should be admitted, not whether they
should be admitted. Indeed, whenever these theorists appeal to certain objects
rather than others, whenever they describe properties of these objects and expect
to be understood and agreed or disagreed with, their talk of mind-dependent
concepts ends up looking like a thinly veiled implicit commitment to certain
fundamental ontological categories.
Thus, although anti-realists might find it quite tempting to eliminate reference
to ontological categories altogether, it might be wiser to proceed with caution.
A revisionist approach might be a better way forward than wholesale elimin-
ation. For, as I have tried to show, it is often not clear how deep the disagreement
between the two sides actually goes, especially in light of the fact that concept-
talk often ends up actually presupposing at least some of the categories it is
trying to rid itself of.

4.2 Benefits of the Ontological-Role Approach to Categories

Another factor that drives the anti-realist pushback of the kind described by Bueno
et al. (2015) has to do with the perceived rigidity, inflexibility, and absoluteness of
categories. Recall how Lowe, in the quote at the end of Section 3.2.6, answers the
following two questions: “Why do we need a ‘science of being’, and how is such
a science possible? Why cannot each special science, be it empirical or a priori,
address its own ontological questions on its own behalf, without recourse to any
overarching ‘science of being’?” (Lowe 2006, p. 4). He says that “reality is one and
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

truth indivisible” and that whereas “special sciences” aim to portray some part of
reality correctly, these different portrayals must all fit together to make an accurate
picture of reality as a whole. It is thus the job of ontology, according to Lowe, to
paint this accurate picture of such a holistic reality.
I suspect that this underlying commitment to oneness of reality and the
supposed indivisibility of truth alienates some scientifically minded philo-
sophers. The appeal to this oneness can appear to leave no room for partial
explorations and piecemeal pursuits of truth to count as fully fledged truths in
their own right. If all of these different partial explorations must add up to one
reality described properly only by the one correct ontology in order to count as
descriptions of reality, and if various truths must all add up to one indivisible
truth, then only those engaged in the “science of everything” will stand a chance
of getting at the truth. And that seems like a tall order indeed.
I am not sure, however, that Lowe intended to embrace this sort of monistic
picture, despite sometimes sounding as if he did. For instance, he writes:
 58 Metaphysics

The relativist must hold that reality itself is many, not one – that we do not all
inhabit the same world. He must say that the sum total of existence for me is
not necessarily the same as the sum total of existence for you. . . . Against
[pluralist ontologies] is posed a monistic ontology which holds that reality is
fundamentally one: that there is just one sum total of existence – one world –
which is the same for all thinkers, places and times. And my suggestion is
that, to the extent that we are committed to the unity of truth, at least inasmuch
as this amounts to an unconditional acceptance of the principle of non-
contradiction, we are committed to the oneness of reality and to its mind-
independence. We are committed, in short, to a fully realist metaphysics.
Fortunately, this still leaves plenty of scope for many forms of pluralism. In
accepting that reality is one, we need not accept that there is only one truth, or
only one truthmaker, or only one kind of truthmaker. Our ontology will admit
of multitudes within The One. (Lowe 2006, p. 191)

In this passage we can see that what drives Lowe to think in apparent monistic
terms about reality is the concern about truth. The unity or indivisibility of truth
mentioned here simply has to do with what he calls “an unconditional acceptance of
the principle of non-contradiction.” There cannot be a truth according to you and
a truth according to me, for this would inevitably lead to a reality of contradictions,
and this simply cannot be, for Lowe. For there to be no violation of the principle of
non-contradiction, the truth must be mind-independent and the reality that our true
descriptions attempt to get at, must be one. It is one not in the almost mystical sense
of the British idealists; rather, it is one in the sense that it is a shared reality.
Seen in this way, ambitious metaphysical realist projects like Lowe’s in fact
leave plenty of room for scientific and metaphysical explorations of all kinds.
They leave room for the metaphysical problem-driven, bottom-up approach to
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

ontology that I favor, as well.

Admittedly, the piecemeal approach is less ambitious, but in spite of it (or
because of it), it might just be more fruitful. In Section 3, I suggested that
approaching the search for fundamental ontological categories is best done
within the framework of ontological roles. Metaphysical problems determine
the ontological roles that need to be played (such as property roles, particular
roles, relation roles, etc.) and the roles, in turn, characterize the sorts of entities –
ontological categories – that can fill them (universals for both property and
relation roles, states of affair or trope-bundles for roles of particulars, etc.).
The hope is that such a framing of the issues will highlight the fact that
determining the categories of entities is not done outside of a specific problem-
solving context. It allows for a more transparent debate about the ways that
metaphysical problems get stated; it stresses the importance of guarding against
engaging in pseudo-problems; and it helps us set up criteria of comparison
between different candidate entities and systems.
 Ontological Categories 59

Saying that ontological roles characterize ontological categories does not

mean that all ontological roles need to be filled by a category of entity.
Ontological roles that need to be filled can be large and small, and some such
roles might need to be filled by entities, while others might be filled with
conceptual distinctions. Thus, whereas categories should be characterized by
appeal to ontological roles, not all roles need to be played by categories.
Ontological categories need not merely be of the highest level. It seems wise
to leave it open how fine-grained one might like the ontological categories to be.
As I noted in Section 3, the highest determinables for a realist about universals
will not themselves be universals, but the lowest, most specific determinates
certainly will be. Will the lowest determinates such as being crimson or being
rectangular themselves be categories? I believe that it will depend on the
explanatory roles assigned to them, within a given theory. Although the highest
categories will likely be set in stone within a particular theory, the lower
categories might not be set at all; and that is a good thing, for it allows for all
kinds of revisionary projects within metaphysical systems.
I will here illustrate just two such cases of category revisions. The first
example comes to us from A. N. Whitehead (1920), with his ontology of events
as the highest category of entity. The second example comes from a relatively
recent reframing of gender categories.

4.2.1 Whitehead’s Temporal Ontology of Events

In various category systems, one usually finds place and time listed as the
highest categories that cut across those of, say, facts, or particulars, or proper-
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

ties. But A.N. Whitehead (1920) offered a radically new and interesting tem-
poral ontology that made no such separation. He explicitly argued against what
he called “a bifurcation of nature” into two systems of reality. One reality was
the given physical reality, say, the reality of microphysics with entities such as
electrons, protons, and other microparticles; the other reality he referred to was
the mind-dependent one, or what is given to us in our sense-awareness. What
Whitehead was getting at is, essentially, the distinction between Lockean
primary and secondary qualities, a distinction that he took to be wrong and
pernicious. The so-called bifurcation of nature, according to Whitehead, created
a schism that induced us to arrive at wrong conclusions about both realms. The
way forward, then, was to close the schism, by assuming that “everything
perceived is in nature. We may not pick and choose. For us the red glow of
the sunset should be as much part of nature as are the molecules and electric
waves by which men of science would explain the phenomenon. It is for natural
philosophy to analyze how these various elements of nature are connected”
 60 Metaphysics

(Whitehead 1920, p. 29). Whitehead’s way of closing the gap between the two
realms was thus not to pick one realm over the other, but rather, to integrate
them.
The result of this approach was a highly original temporal ontology, with
irreducibly dynamic events and processes at its core. Whitehead (1920) con-
ceived of events as chunks of passage of nature, and thought of objects as what
remains the same through time, through all the change and flux. Objects, for
him, are always to be found within events, and could only be isolated through
the process of abstraction. The traditional particular–universal distinction was
scrapped and replaced by an event–object distinction in Whitehead’s temporal
ontology.
As we saw in Section 2, the particular–universal distinction emerged from
the recognition of genuine resemblance in the world, and from the need to
explain and ontologically ground this phenomenon in some way. Universals
were introduced as entities that ground such resemblances by being present in
each particular that has an apparent property. Particulars, in this framework,
provided the ground of distinctness of entities. In such a context, change was
often seen as a state of dissimilarity, described in static terms such as:
a particular, a, has a certain property, F, at a certain time, t1, and then
a different property, G, at time t2. But Whitehead’s event–object distinction
turned this picture on its head. The background assumption in his ontology is
the constant change and transitoriness in nature. Nature appears to us as
passing and changing, and events are the basic currency of that temporal
passage. Objects are then derived as what seems to us to be constant in that
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

change; what we can grasp and abstract from it.

Interestingly, what helped lead Whitehead to this change in perspective was
the idea of seeing ourselves and our perceptions of reality as an integral part of
it: “sense-awareness and thought are themselves processes”; in addition to what
ordinarily gets considered as passage in nature, we need to consider that “there
is a passage of sense-awareness and a passage of thought” (Whitehead 1920,
p. 66). He even adds: “We may speculate, if we like, that this alliance of the
passage of mind with the passage of nature arises from their both sharing in
some ultimate character of passage which dominates all being” (Whitehead
1920, p. 69). It is a matter for a different investigation to consider how
successful Whitehead’s unified approach to ontology really is. Here, I just
wanted to point out that there is merit in exploring attempts similar to his, as
they reveal different ontological categorizations that are worth taking seriously.
In their piece on blind spots in physics and scientific inquiry more broadly,
physicists Frank and Gleiser, and philosopher Thompson (Frank et al. 2019),
apply Whitehead’s insight to the study of time and the study of consciousness.
 Ontological Categories 61

They argue that we mustn’t lose sight of how experientially charged the objects
of scientific inquiry really are; overlooking this fact leads to an unwarranted
faith in science as providing absolute knowledge:

When we look at the objects of scientific knowledge, we don’t tend to see

the experiences that underpin them. We do not see how experience makes
their presence to us possible. Because we lose sight of the necessity of
experience, we erect a false idol of science as something that bestows
absolute knowledge of reality, independent of how it shows up and how
we interact with it. (Frank et al. 2019)

This belief in absolute knowledge leads to blind spots: “The Blind Spot arises
when we start to believe that this [scientific] method gives us access to unvar-
nished reality. But experience is present at every step. Scientific models must be
pulled out from observations, often mediated by our complex scientific equip-
ment. They are idealizations, not actual things in the world” (Frank et al. 2019).
They conclude: “[T]o finally ‘see’ the Blind Spot is to wake up from a delusion
of absolute knowledge. It’s also to embrace the hope that we can create a new
scientific culture, in which we see ourselves both as an expression of nature and
as a source of nature’s self-understanding” (Frank et al. 2019).
All this is very broad and it is not entirely clear how such an insight is to be
applied on a case-by-case basis. Nonetheless, there is something valuable here that
we can apply to metaphysical investigations as well as the scientific ones. It does
seem to be the case that certain blind spots in metaphysics arise in a similar way to
the ones described – that is, they arise when we start to take our models of reality,
our rough generalizations and categorizations, for the reality itself. Whitehead was
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

particularly worried about this when discussing the notion of an instant of time,
a construct he took to be useful but ultimately very misleading when it comes to the
true nature of physical reality. He puts this concern as follows:

Instantaneousness is the concept of all nature at an instant, where an instant is

conceived as deprived of all temporal extension. For example we conceive of
the distribution of matter in space at an instant. This is a very useful concept
in science especially in applied mathematics; but it is a very complex idea so
far as concerns its connexions with the immediate facts of sense-awareness.
There is no such thing as nature at an instant posited by sense-awareness.
What sense-awareness delivers over for knowledge is nature through
a period. (Whitehead 1920, p. 57, italics mine).

This passage nicely illustrates Whitehead’s motivation for disposing of

instants as genuine features of physical reality. For him, we need to begin
our study of nature from our experience of it, and what experience offers are
only temporally extended events, not instantaneous entities. Instants, he says
 62 Metaphysics

at one point, are “a metaphysical fairytale.” This does not mean, though, that
we are to dispose of instants altogether. Just because instants are not given in
our sense-awareness does not mean that they are not useful abstractions in
mathematics and science. What we must not lose track of is that they are just
that – useful abstractions. And, for Whitehead, abstractions should not be
substituted for reality, which they are introduced to describe and explain,.
Thus, we should take Whitehead seriously when he warns: “[W]e are apt to
fall into the error of thinking that the facts are simple because simplicity is the
goal of our quest. The guiding motto in the life of every natural philosopher
should be ‘Seek simplicity and distrust it’.” (Whitehead 1920, p. 163).
Even if one disagrees with the particulars of Whitehead’s ontological pro-
posal, the motto he offers is instructive. The change of perspective that his
denial of the bifurcation of nature offers is illuminating. By integrating our
experience of change with the passage of nature itself, he fused the two sets of
categories that are often kept separate – space and time, on the one hand, and
atemporal categories such as particulars and universals, on the other. The result
is surprising and makes us reconsider whether passage and change should
indeed be derived from a static picture of reality, or whether passage and flux
of nature should come first with static entities derived as abstractions from it.

4.2.2 Reframing of Gender Categories

In social ontology, discussions of ontological categories are particularly conse-

quential for our practical concerns. Take, for instance, the question of whether
gender is a genuine mid-range ontological category, and whether it should be
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

used to categorize people into only two distinct groups of men and women.
Many have thought that how one answers this metaphysical question ought to
guide one’s use of gender terms, but Barnes (2020) has recently argued against
this straightforward alignment.
Indeed, “in virtue of what, if anything, do people have genders?” seems to be
a distinctly metaphysical question. In her recent paper Barnes (2020) nicely
summarizes two broad types of answers that this question has received in recent
literature: (1) social-position accounts, which explain gender by appeal to
external factors; and (2) identity-based accounts, which explain gender by
appeal to internal factors.
Social-position accounts tend to explain gender by appeal to external social
factors such as how people are perceived, or what social roles they are expected
to play in society, and so on. For Witt (2011), for example, gender seems to play
a pretty fundamental role in structuring and ordering other social properties and
roles. She takes human beings to be comprised of the human organism, the
 Ontological Categories 63

person, and the social individual. Gender is then a social property that structures
and unifies other social roles which come to be part of the social individual. For
example, I am not just a parent, I am a mother; I am not just a spouse, I am
a wife; I am not just a sibling, I am a sister; and so on. Gender thus has a central
social role that is imposed on persons based on various assumptions about the
reproductive role that they play. Haslanger (2012), too, is a proponent of the
social-position account of gender, but in her definitions of women and men, she
makes salient the structural features of society which consistently place women
in subordinate roles and positions based on their perceived or imagined bodily
features, and which place men in privileged roles and positions based on their
perceived or imagined bodily features.
Barnes notes that the trouble with social-position accounts is that they face
what Jenkins (2016) has called an “exclusion problem.” That is, such accounts
tend to leave out of their classifications trans women and trans men if they are
not perceived by others as falling into these social categories. By relying merely
on how others in society perceive and engage with a person, and not on how
people themselves identify, social-position accounts fail to capture an important
aspect of gender – the relevant first-person reports of people themselves.
Identity-based accounts avoid the exclusion problem just described by put-
ting weight, in their explanations of gender, on the internal features of persons.
What matters is how one feels about oneself, how one identifies, which groups
one sees oneself as belonging to, and so on. This kind of view is endorsed by
McKitrick (2015) in her appeal to behavioral dispositions. If a person has
a certain number of dispositions identified as feminine (such as how they tend
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

to express themselves in various contexts, how they dress, how they wear their
hair and makeup, what name they go by, which bathroom they choose, and so
on), then that person is a woman. But the trouble with these kinds of accounts is
that though they avoid the previous exclusion problem, they face a problem of
their own. The worry here is that if too much weight is put on a person’s own
self-identification and expression of gender, it might leave out those who are not
actively engaged, or not able to engage in this kind of self-sorting activity. The
example that Barnes (2020) discusses here is of cognitively disabled women
and men who, it seems, should not be excluded from being identified as women
or men just because they lack the capacities for explicit self-identification.
Barnes herself ends up endorsing a modified social-position account of
gender. She finds this account to be important insofar as it tracks social reality
well. At the same time, she admits that such an account far from offers
a complete story; gender, according to her, also encompasses gender identity,
gender expression, and so on. In the light of the discussion of the messiness and
difficulties in characterizing gender categories, Barnes insists that gender terms
 64 Metaphysics

should not be expected to neatly track such categories. She makes a strong case
for keeping the application of gender terms as permissive and flexible as
possible, while at the same time recognizing the social reality captured by the
social-position accounts. In this way, the flexible use of these terms is meant to
help the process of overcoming the still very real and grim social reality of
gender inequality.
These sorts of discussions are clearly important and relevant although they
do not engage with traditional metaphysics of the highest ontological categor-
ies of substance, mode, and property. But exploring the status of mid- to
lower-level categories such as gender or race is important and revealing. We
find important truths when we ask whether such categories are genuine; in
what way they should be defined; what role they play in current social and
political reality; and whether or not we should worry if our terms pick these
categories neatly or not. These are all questions that a revisionist ontological
project makes room for.

4.3 Concluding Remarks: Ontological Progress and the Revisionist

Project
I conclude with a few words to the skeptic who, after reading my discussion of
ontological categories in this Element, is left wondering whether the whole
project is utterly hopeless. They might think that if, after more than two and
a half millennia since Plato’s discussion of Forms, philosophers have not yet even
remotely settled on fundamental ontological categories, what hope can there be
for progress in ontology, or metaphysics, or philosophy more broadly?
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

I believe that such a pessimism is too hasty and that a lot depends on how
we interpret progress in philosophy. In his recent paper, Daly (2016) equates
the presumed lack of progress with persistent philosophical disagreement,
that is, the fact that no philosophical problem appears to have been perman-
ently solved. He then goes on to explore three different explanations for this
lack of progress: (1) Russell’s explanation, which contrasts science and
philosophy; (2) MacBride’s explanation, which invokes “the epistemic end
of days”; and (3) the explanation that appeals to the notion of cognitive closure.
Russell’s explanation is succinct: he claims that science is “what you more or
less know” whereas philosophy is “what you do not know,” thus contrasting the
presumed certainty of science with the speculative nature of philosophy. Russell
also pointed out that science has been successful at solving some of the problems
that initially originated as philosophical ones. But, as Daly (2016) rightly points
out, this Russellian characterization of both science’s certainty and philosophy’s
uncertainty is not accurate. There is a fair amount of speculation in science itself;
 Ontological Categories 65

we just need to turn our attention to theoretical physics. So, Russell’s account is
not helpful in diagnosing the reason for the presumed lack of philosophical
progress.
MacBride’s (2014) explanation attributes the lack of philosophical progress
to the fact that philosophical problems have a broad range – they are both
general and require understanding of several areas of knowledge. So, until the
supposed “epistemic end of days” we cannot expect progress to happen because
of the deep interconnectedness of philosophical problems with each other, and
with problems in other adjacent disciplines. Daly (2016) finds this account
objectionable as well, because, according to him, it exaggerates the degree to
which philosophical problems are indeed interconnected. Many problems in
science are deeply interconnected too, notes Daly, but that by itself does not
translate into a lack of progress.
I am particularly intrigued by the idea of the “epistemic end of days” and
wonder if a similar hypothesis can help us think about progress in philosophy
and the nature of our discipline more broadly. Imagine an omniscient being:
Does such a being philosophize? And, more to the point, does such a being do
ontology and engage in the search for ontological categories? If we are inclined
to say “no, the omniscient being does not engage in ontological inquiry,” such
an answer would likely be based on the following thought: when one does
ontology, one is engaged in the discovery of ontological facts; such facts are
perhaps, as MacBride (2014) suggests, extremely difficult to get right. But
a higher omniscient being would surely have no trouble grasping them. Such
a being would not need to engage in philosophical pursuit any more than it
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

would need to engage in scientific pursuits, experiments, and so on. If, however,
we thought that an omniscient being would indeed do philosophy, such an
answer might indicate several things. It could reveal a view of philosophy as
riddled with puzzles that no amount of factual knowledge can solve; or it could
reveal a view of philosophy as a process of reasoning of a particular type, with
different possible outcomes, due to different possible combinations of views as
well as patterns of reasoning. I am sure there are more possibilities still; the goal
of the question “does an omniscient being philosophize?” is simply to help us
see that the question of philosophical progress perhaps has less to do with the
disagreements we have with one another and more to do with our understanding
of the very nature of philosophy and philosophical pursuit.
Related to this discussion is the third explanation of the lack of philosophical
progress with its appeal to our presumed cognitive closure. Daly attributes this
view to Van Inwagen (2008), and the thought is simply that we cannot make
progress in philosophy because our minds are cognitively closed to solutions of
philosophical problems; our minds are simply not up to the task. I find this kind
 66 Metaphysics

of explanation to be the most hopeless of all, as it amounts to suggesting that we

might as well just give up – our philosophical pursuits are the vacuous exercise
of an incompetent mind. It is quite ironic that such a suggestion comes from one
of the most prominent contemporary metaphysicians.
This approach can be rebuffed, of course, by questioning the claim that we are
cognitively closed to solutions of philosophical problems. For what is it,
exactly, about our cognitive capacities that makes us particularly incapable of
finding solutions to philosophical problems, but allows us to find solutions to
scientific or mathematical problems? This last question shows us the way out –
that is, it shows us that we have been focusing on the wrong side of the issue.
Rather than worry about the presumed lack of solutions to philosophical
problems, we should be focusing on the formulations of our problems.
Daly draws a similar conclusion about methodology. He believes that “the
fault doesn’t lie with the questions we ask – that they are somehow defective –
or with our minds – that they are ill-suited or too limited – but with the methods
and ambitions involved in our inquiry” (Daly 2016, pp. 35–36). Daly takes
philosophical methods to be at once too weak and too strong. They are too weak,
because one can always question the justification and the evidence for the
premises, even in deductive arguments. And they are too strong in the sense
that no conclusion is safe – it is subject to the same scrutiny that the premises of
the argument are subjected to. Philosophical ambitions are also at fault, accord-
ing to Daly, since philosophy “seeks to identify the most fundamental level of
epistemic justification for claims, and it aspires to an especially high degree of
clarity and understanding” (Daly 2016, p. 38).
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

I tend to agree with the spirit of Daly’s diagnosis, though not with the details.
First, it is not clear to me that a presumed lack of progress in philosophy should
be deduced from the persistence of disagreement. Persistence of disagreement
in philosophy can be seen as its feature, perhaps even a philosophical virtue,
a sign of healthy skepticism within a discipline that is highly speculative but
also uniquely self-reflective and self-questioning. Moreover, a case can be made
that there has been a tremendous amount of progress in philosophy across the
board, as well as in ontology more specifically. Often times it seems as if there
are very few stones that philosophers have left unturned, very few arguments or
avenues that have been left unexplored.
But there is room for improvement, of course, and I believe that philosophical
methods need to be further scrutinized and better regimented. I have argued that
part of this task is to engage in a more careful analysis of the formulations of the
problems we set out to address, by being more transparent about the underlying
motivations and implicit assumptions. I also agree with Daly’s concern about
philosophical ambitions, though I see the concern as having more to do with the
 Ontological Categories 67

level of generality than with the aims themselves that philosophers set their
sights on.
I find that as admirable as systematic metaphysical top-down theory-building
is, there should also be room for more piecemeal and humble approaches to
ontology. Trying to carry out both projects is probably the best way forward,
since if we only take the global approach, we might miss catching the blind
spots and/or force the categorizations top-down in a way that may not fit the
different philosophical phenomena in need of explanation. If, on the other hand,
we only take the piecemeal approach, we might miss the ways that different
areas of inquiry hang together, and miss out on the big picture that drives the
commitment to the most fundamental ontological categories. The most fruitful
way forward is probably the one that puts forward an initial system, but in
various areas keeps amending and refining the categories, as needed.
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press
 References
Ackrill, J. L. (1963). Aristotle: Categories and De Interpretatione. Oxford:
Clarendon Press.
Armstrong, D. M. (1978). Nominalism and Realism: Universals and Scientific
Realism, Volume I. Cambridge: Cambridge University Press.
(1983). What Is a Law of Nature? New York: Columbia University Press.
(1997a). “Against ‘Ostrich’ Nominalism: A Reply to Michael Devitt.” In
D. H. Mellor and A. Oliver eds., Properties. Oxford: Oxford University
Press, 1997, 101–111. [Original date of publication 1980.]
(1997b). A World of States of Affairs. Cambridge: Cambridge University Press.
Barnes, E. (2019). “Gender and Gender Terms.” Noûs, 54/3, 704–730.
Baumer, M. (1993). “Chasing Aristotle’s Categories Down the Tree of
Grammar.” Journal of Philosophical Research, 18, 341–449.
Bradley, F. H. (1893). Appearance and Reality. Oxford: Clarendon Press.
(1910). “On Appearance, Error and Contradiction.” Mind, 19/74, 153–185.
(1911). “Reply to Mr. Russell’s Explanations.” Mind, 20/77, 74–76.
Bueno, O., Busch, J., and Shalkowski, S. A. (2015). “The No-Category
Ontology.” The Monist, 98/3, 233–245.
Cameron, R. P. (2008). “Turtles All the Way Down: Regress, Priority and
Fundamentality.” Philosophical Quarterly, 58/230, 1–14.
Campbell, K. (1990). Abstract Particulars. Oxford: Basil Blackwell.
Cargile, J. (2003). “On Russell’s Argument against Resemblance Nominalism.”
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Australasian Journal of Philosophy, 81/4, 549–560.

Chisholm, R. M. (1996). A Realistic Theory of Categories: An Essay on
Ontology. Cambridge: Cambridge University Press.
Conee, E., and Sider, T. (2007). Riddles of Existence: A Guided Tour of
Metaphysics. Oxford: Oxford University Press.
Correia, F. (2008). “Ontological Dependence.” Philosophy Compass, 3/5,
1013–1032.
Daly, C. (2016). “Persistent Philosophical Disagreement.” Proceedings of the
Aristotelian Society, 1, 3–25.
Della Rocca, M. (2010). “PSR.” Philosopher’s Imprint, 10/7, 1–13.
Frank, A., Gleiser, M., and Thompson, E. (2019). “The Blind Spot.” Aeon,
January 8. https://aeon.co/essays/the-blind-spot-of-science-is-the-neglect-
of-lived-experience.
Grossman, R. (1983). The Categorial Structure of the World. Bloomington:
Indiana University Press.
 References 69

Grünbaum, A. (2009). “Why Is There a World At All, Rather Than Just

Nothing?” Ontology Studies, 9, 7–19.
Haslanger, S. (2012). Resisting Reality: Social Construction and Social
Critique. New York: Oxford University Press.
Jackson, F. C. (1977). “Statements about Universals.” Mind, 86/343,
427–429.
Jenkins, K. (2016). “Amelioration and Inclusion: Gender Identity and the
Concept of Woman.” Ethics, 126/2, 394–421.
Johnson, W. E. (1921). Logic. Cambridge: Cambridge University Press.
Lewis, D. (1973). “Counterfactuals and Comparative Possibility.” Journal of
Philosophical Logic, 2/4, 418–446.
(1983). “New Work for a Theory of Universals.” Australasian Journal of
Philosophy, 61, 343–377.
Lowe, E. J. (2006). The Four-Category Ontology: A Metaphysical Foundation
for Natural Science. Oxford: Clarendon Press.
(2018). “Metaphysics as the Science of Essence.” In A. Carruth, S. Gibb, and
J. Heil eds., Ontology, Modality, and Mind: Themes from the Metaphysics
of E.J. Lowe. Oxford: Oxford University Press, 14–34.
MacBride, F. (2005). “The Particular–Universal Distinction: A Dogma of
Metaphysics?” Mind, 114, 565–614.
(2014). “Analytic Philosophy and Its Synoptic Commission: Towards the
Epistemic End of Days.” Royal Institute of Philosophy Supplement, 74,
221–236.
Maurin, A.-S. (2018). “Tropes.” In Edward N. Zalta ed., The Stanford
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Encyclopedia of Philosophy. https://plato.stanford.edu/entries/tropes/.

McDaniel, K. (2017). The Fragmentation of Being. Oxford: Oxford University
Press.
McKitrick, J. (2015). “A Dispositional Account of Gender.” Philosophical
Studies, 172/10, 2575–2589.
Nolan, D. (2001). “What’s Wrong with Infinite Regresses?” Metaphilosophy,
32/5, 523–538.
Oliver, A. (1996). “The Metaphysics of Properties.” Mind, 105, 1–80.
Pap, A. (1959). “Nominalism, Empiricism and Universals: I.” Philosophical
Quarterly, 9/37, 330–340.
Perović, K. (2015). “The Importance of Russell’s Regress Argument for
Universals.” In D. Wishon and B. Linsky eds., Acquaintance,
Knowledge, and Logic: New Essays on Bertrand Russell’s The Problems
of Philosophy, 171–187. Stanford, CA: CSLI Publications.
(2016). “A Neo-Armstrongian Defense of States of Affairs: A Reply to
Vallicella.” Metaphysica, 17/2, 143–161.
 70 References

Plato. (1963). The Collected Dialogues, ed. E. Hamilton and H. Cairns. Princeton,
NJ: Princeton University Press.
Quine, W. V. (1951). “Ontology and ideology.” Philosophical Studies: An
International Journal for Philosophy in the Analytic Tradition, 2/1, 11–15.
(1997). “On What There Is.” In D. H. Mellor and A. Oliver eds., Properties.
Oxford: Oxford University Press, 74–88. [Original date of publication
1948.]
Ramsey, F. P. (1925). “Universals.” Mind, 34/136, 401–417.
Rodriguez-Pereyra, G. (2002). Resemblance Nominalism. Oxford: Oxford
University Press.
(2004). “Paradigms and Russell’s Resemblance Regress.” Australasian
Journal of Philosophy, 82/4, 644–651.
Russell, B. (1911). “On the Relations of Universals and Particulars.”
Proceedings of the Aristotelian Society, 12, 1–24.
(1912). The Problems of Philosophy. London: William & Norgate.
(1992). Theory of Knowledge: The 1913 Manuscript. New York: Routledge.
Schaffer, J. (2012). “Grounding, Transitivity, and Contrastivity.” In F. Correia
and B. Schneider eds., Metaphysical Grounding: Understanding the
Structure of Reality. Cambridge: Cambridge University Press, 122–138.
Simons, P. (1994). “Particulars in Particular Clothing: Three Trope Theories of
Substance.” Philosophy and Phenomenological Research, 54/3, 553–575.
Studtmann, P. (2021). “Aristotle’s Categories.” In Edward N. Zalta ed., The
Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/
aristotle-categories/.
Tahko, T. E. (2018). “Fundamentality.” In Edward N. Zalta ed., The Stanford
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Encyclopedia of Philosophy. https://plato.stanford.edu/entries/fundamen

tality/.
Tahko, T. E., and Lowe, E. J. (2020). “Ontological Dependence.” In Edward
N. Zalta ed., The Stanford Encyclopedia of Philosophy. https://plato.stan
ford.edu/entries/dependence-ontological/.
Van Inwagen, P. (2008). Metaphysics. 3rd ed. Boulder, Co: Westview Press.
Westerhoff, J. (2005). Ontological Categories. Oxford: Clarendon Press.
Whitehead, A. N. (1919). An Enquiry concerning the Principles of Natural
Knowledge. Cambridge: Cambridge University Press.
(1920). The Concept of Nature. Cambridge: Cambridge University Press.
Witt, C. (2011). The Metaphysics of Gender. Oxford: Oxford University Press.
 Metaphysics

Tuomas E. Tahko
University of Bristol
Tuomas E. Tahko is Professor of Metaphysics of Science at the University of Bristol, UK.
Tahko specializes in contemporary analytic metaphysics, with an emphasis on
methodological and epistemic issues: ‘meta-metaphysics’. He also works at the interface of
metaphysics and philosophy of science: ‘metaphysics of science’. Tahko is the author of
Unity of Science (Cambridge University Press, 2021, Elements in Philosophy of Science),
An Introduction to Metametaphysics (Cambridge University Press, 2015) and editor of
Contemporary Aristotelian Metaphysics (Cambridge University Press, 2012).

About the series

This highly accessible series of Elements provides brief but comprehensive
introductions to the most central topics in metaphysics. Many of the Elements also go
into considerable depth, so the series will appeal to both students and academics.
Some Elements bridge the gaps between metaphysics, philosophy
of science, and epistemology.
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press
 Metaphysics

Elements in the series

Substance
Donnchadh O’Conaill
Essence
Martin Glazier
Truthmaking
Jamin Asay
Laws of Nature
Tyler Hildebrand
Dispositions and Powers
Toby Friend and Samuel Kimpton-Nye
Modality
Sònia Roca Royes
Parts and Wholes: Spatial to Modal
Meg Wallace
Indeterminacy in the World
Alessandro Torza
Parts and Wholes
Meg Wallace
Formal Ontology
Jani Hakkarainen, Markku Keinänen
https://doi.org/10.1017/9781108973861 Published online by Cambridge University Press

Chemistry’s Metaphysics
Vanessa A. Seifert
Ontological Categories: A Methodological Guide
Katarina Perović

A full series listing is available at: www.cambridge.org/EMPH