Untitled

OUP CORRECTED PROOF – FINAL, 6/4/2018, SPi
Reality and its Structure

Reality and its

Structure
Essays in Fundamentality
edited by
Ricki Bliss
and Graham Priest
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© the several contributors 2018
The moral rights of the authors have been asserted
First Edition published in 2018
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2017959735
ISBN 978–0–19–875563–0
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
List of Contributors vii
0. The Geography of Fundamentality: An Overview 1

Ricki Bliss and Graham Priest
Part I. The Hierarchy Thesis

1. Grounding Orthodoxy and the Layered Conception 37
Gabriel Oak Rabin
2. Symmetric Dependence 50
Elizabeth Barnes
3. Grounding and Reflexivity 70
Ricki Bliss
4. Cosmic Loops 91
Daniel Nolan
5. Metaphysical Interdependence, Epistemic Coherentism,
and Holistic Explanation 107
Naomi Thompson
6. Buddhist Dependence 126
Graham Priest
7. Bicollective Ground: Towards a (Hyper)Graphic Account 140
Jon Erling Litland
Part II. The Fundamentality Thesis

8. Indefinitely Descending Ground 167
Einar Duenger Bohn
9. Inheritance Arguments for Fundamentality 182
Kelly Trogdon
10. From Nature to Grounding 199
Mark Jago
11. Grounding in Mathematical Structuralism 217
John Wigglesworth
12. Fundamentality and Ontological Minimality 237
Tuomas E. Tahko
vi contents
13. The Structure of Physical Reality: Beyond Foundationalism 254

Matteo Morganti
Part III. The Contingency and Consistency Theses

14. On Shaky Ground? Exploring the Contingent Fundamentality Thesis 275
Nathan Wildman
15. Heidegger’s Grund: (Para-)Foundationalism 291
Filippo Casati
Index of Names 313

General Index 316
List of Contributors
Elizabeth Barnes University of Virginia

Ricki Bliss Lehigh University
Filippo Casati Kyoto University
Einar Duenger Bohn University of Agder
Mark Jago University of Nottingham
Jon Erling Litland University of Texas at Austin
Matteo Morganti University of Rome Tre
Daniel Nolan University of Notre Dame
Graham Priest The Graduate Center of the City University of New York, and the
University of Melbourne
Gabriel Oak Rabin New York University Abu Dhabi
Tuomas E. Tahko University of Helsinki
Naomi Thompson University of Southampton and University of Gothenburg
Kelly Trogdon Virginia Tech
John Wigglesworth University of Vienna
Nathan Wildman University of Glasgow and Tilburg University
0
The Geography of Fundamentality
An Overview
Ricki Bliss and Graham Priest
Reality is a rather large place. It contains protons, flamingos, economies, headaches,

sentences, smiles, asteroids, crimes, and numbers, amongst very many other things.
Much of the content of our reality appears to depend on other of its content.
Economies, for example, appear to depend upon people and the way they behave,
amongst other things. Some of the content of our reality also appears to be, in some
significant sense, more important than other of its content. Whilst none of us would
wish to deny the very important role that economies play in our lives, most of us would
agree that without matter arranged certain ways in space, for example, there could be
no economies in the first place.
The reality that we happen to occupy is, in some important sense, a physical one.
Accordingly, matter is afforded a special place in our story about it. Indeed, not only is
matter accorded a special place in our ontology, but some from amongst its elements
are also thought to be particularly important. Chairs and flamingos and people are
made from parts, and those parts from further parts and so on—with most folks being
of the view that at some point these dependence chains must terminate in absolutely
basic, or simple, parts which themselves have no further parts. It is these basic parts,
so the story goes, that give rise to everything else.
The content of reality to which these parts give rise is arranged relatively neatly
into layers: facts about economies and crimes reside at a higher level than facts
about biological systems, which reside at a higher level than facts about chemical
systems and so on. Or perhaps we might prefer to say that economic systems are
further up the Great Chain of Being than ecosystems, which are further up the
chain than carbon compounds.1 This picture, or something very much like it, looms
large over contemporary analytic metaphysics: a picture according to which reality is
hierarchically arranged with chains of entities ordered by relations of ground and/or
ontological dependence terminating in something fundamental.
1 The Great Chain is normally taken as running downwards, with the ground at the top; we upend it here.
 the geography of fundamentality: an overview
The historical literature is also littered with what appear to be variations on this
kind of view. Consider both Plato and Aristotle, for example. The former believed
that everything was grounded in the Forms, with all of the Forms being ultimately
grounded in the Form of the Good. The latter distinguished between primary and
secondary substances, with a priority ordering amongst them—along, arguably, with
making appeal to prime matter, without which there would be nothing whatsoever.
Just as very many of the Medievals (Aquinas, for example) and Early Moderns
(Descartes, Spinoza, Leibniz) thought that everything depended on God, the need
to establish a fundamental ground breaks out in certain of the Continental thinkers,
such as Heidegger, in the form of The Problem of Being: there must be something
(fundamental), Being, if we are to account for the fact that anything has being at all.
Turning also to non-Western traditions, we see that the idea that reality is structured
by metaphysical dependence relations, where there is something fundamental, is by
no means an unfamiliar one.2 Various of the Indian, Chinese, and Japanese traditions
rely heavily on notions of metaphysical dependence and fundamentality. In fact,
whole schools were formed based on disagreements over the fundamental structure of
reality. According to the Indian Abhidharmika tradition, for example, there must be
dharmas—simples—as there are aggregates which are built from them. And according
to Kyoto School thinker Nishida, the ultimate ground of everything is consciousness,
which is also absolute nothingness. The idea that reality is structured, and that there
must be something fundamental, is by no means the monopoly of contemporary
Western analytic thought.
The kind of view, or cluster of views, that appear to dominate the contemporary
analytic debate can be thought of broadly as, or as species of, metaphysical foun-
dationalism. As will become clearer in due course, there are, in fact, a variety of
ways in which one can be a metaphysical foundationalist; with different species of
foundationalism involving different core commitments. Although this list is by no
means exhaustive, we assume the following to be amongst the core commitments of
metaphysical foundationalism as commonly endorsed in the contemporary literature.
1. The hierarchy thesis: Reality is hierarchically structured by metaphysical de-
pendence relations that are anti-symmetric, transitive, and anti-reflexive.
2. The fundamentality thesis: There is some thing(s) which is fundamental.
3. The contingency thesis: Whatever is fundamental is merely contingently
existent.
4. The consistency thesis: The dependence structure has consistent structural
properties.
Strictly speaking, in order to be considered a species of foundationalism, a view
needs only commit to the the fundamentality thesis: 2., then, is both necessary and
2
See Bliss and Priest 2017.
ricki bliss and graham priest 
sufficient for a view to count as a kind of foundationalism. For proponents of what we

can think of as the standard view, however, all four theses are necessary, with no one
of them being sufficient.3
Is this the only view of the fundamental, or basic, structure of reality that is available
to us, though? Of course it isn’t. To be sure, deviations from the standard view exist
in the literature.4 But the full spread of possible views has, so far as we can tell, been
both grossly underestimated and grossly underexplored.
It is important and interesting to note that in foundational epistemology—where
the structuring relations are strikingly similar to those invoked in talk of foundational
metaphysics—one can be an epistemic foundationalist (of various sorts), an epistemic
infinitist, or an epistemic coherentist. Is a similar spread of possible views available to
us in foundational metaphysics? We are inclined to think that it is, as do Morganti and
Thompson (this volume). Just as an epistemic infinitist thinks that chains of beliefs
ordered by an anti-symmetric, anti-reflexive, transitive relation orders beliefs without
termination, a metaphysical infinitist thinks that chains of entities ordered by an anti-
symmetric, anti-reflexive, transitive relation orders entities without termination. So
too for coherentism. Just as an epistemic coherentist thinks that beliefs are organized
into a highly integrated web, with justification emerging from it, the metaphysical
coherentist thinks that entities are organized into a highly integrated web with
something like being or reality emerging from it. As one might expect, there will also
be various possible shades between.
The papers contained within this volume can be thought of as contributing to a
broader discussion of the reasons for which we are supposed to believe aspects of
the standard view, the reasons we might have for embracing one or other of the
alternatives, and what those alternatives might be like. Not all of the papers in this
volume endorse types of anti-foundationalism, but each of them speaks to, and chal-
lenges, in some way or other, one or other of the core commitments of metaphysical
foundationalism as noted above. In some cases, our authors even support one or other
of the assumptions, with the aim of their contribution being to highlight weaknesses
in the arguments commonly offered in their defence. The papers in this volume are
arranged, then, according to the core assumption that they primarily address.
3 The idea that the world is ontologically ‘flat’, with everything being fundamental—a rejection of 1—
has been described by Bennett 2011 as ‘crazy pants’, for example. Just as many philosophers baulk at the
suggestion that the fundamentalia are necessary beings.
4 It is worth noting that it does not follow from the appearance of a smattering of papers challenging the
standard view that the standard view is not still just that, the standard view. A handful of dissenting papers
does not a heterodoxy make. Although some authors have challenged aspects of the foundationalist picture,
the dominant paradigm that drives many contemporary analytic research programmes is one according to
which reality has a layered structure and a fundamental level. Even though a small number of philosophers
have challenged aspects of the standard view, to the best of our knowledge, these challenges have not
resulted in research programmes of their own, nor have they impacted upon the way much research is
conducted.
 the geography of fundamentality: an overview
In what remains of this introduction, we take up the mantle of introducing and

engaging with some of the most important issues that we believe need to be dealt
with if foundationalism is to be a view that we actually have good reasons to endorse;
and if the alternatives are to be considered not just logically, but also metaphysically,
possible.
1 The Lie of the Land

Many philosophers accept a view according to which the world has an overarching
causal structure. Thunderstorms cause trees to fall down, and water is caused to
boil by the application of heat. This volume takes as one of its starting assumptions
that the world (also) has an overarching metaphysical structure. Of course, causal
structure is a kind of metaphysical structure; however, what philosophers tend to
mean nowadays when they speak of metaphysical structure is that this structure is
induced by relations of ground and/or ontological dependence.5 We refer to these as
metaphysical dependence relations, and they are the relations around which the ideas
presented in the following essays are centred.
There is a lot that has been, and continues to be, written on metaphysical depend-
ence relations. And there is an enormous amount of disagreement over even the
most basic of concepts in operation in the relevant literature.6 Is grounding to be
understood on the operator view or the sentential connective view? Is grounding just
explanation? How are grounding and ontological dependence related? Is grounding
unitary? These are amongst some of the many issues that those working on issues
pertaining to the structure of reality are concerned with. This volume is not primarily
concerned with most of those disagreements, however. We leave it to our contributors
to assume what they will regarding how they define their terms and the conceptual
connections that they take to be in operation; and we leave it to our readers to find
appropriate reading material if what they are interested in are those debates. For the
sake of clarity in this introduction, however, we think it wise to say something about
how we shall be understanding things.
It is not uncommon to see a distinction drawn in the literature between relations
of ground and ontological dependence. Relations of ground, say many, obtain between
facts, where relations of ontological dependence obtain between entities of any and
all categories.7 So, where one would say that the fact that the weather is miserable
today is grounded in the fact that it is pouring, one would say that the shadow
ontologically depends on the object that casts it. And where one would say that the
fact that the sky is blue or we are in Australia, is grounded in the fact that the sky
5 See Schaffer 2016 for a discussion of the relationship between grounding and causation, and a view
according to which grounding is a kind of causing.

6 See Bliss 2014 for an overview of some of the major sources of disagreement.
7 See Schaffer 2009 for the development of a view according to which grounding obtains between entities
of any and all categories and cross-categorically.

ricki bliss and graham priest 
is blue, one would also say that the fact that the sky is blue ontologically depends
on its constituents—the sky and blueness. Again, when we talk about relations of
metaphysical dependence, we mean this term to act as a covering term for both
grounding and ontological dependence. Where, in this introduction, we think it
necessary to discriminate between the two, we say as much. We also don’t think much
as regards the reasons to endorse one fundamental view of reality over another is going
to turn on whether grounding obtains between facts alone, for example. What bears
consideration when settling the kinds of matters that this volume is concerned with
will be the same, we believe, whether it turns out that ontological dependence just is
a kind of grounding or not.
It is a plank of the grounding literature that grounding is somehow involved with
metaphysical explanation. It is an open question, however, whether the relations are
merely associated with metaphysical explanation or whether they are identical with it.
Thompson (this volume) offers us some compelling reasons to think that grounding
is better thought of as being an explanatory relation. She argues that were grounding
relations to be relations that underwrite our explanations, we would still need to
account for how the relations and the explanations they back are related to one
another. If the way they are related to one another is via grounding, then we are
really in trouble, says Thompson, because the notion of a metaphysical explanation
is typically invoked to shed light on how we are supposed to understand grounding in
the first place. Trogdon (this volume), on the other hand, thinks it natural to assume
that grounding relations back metaphysical explanations. So far as we can tell, not
much turns on resolving this particular issue for what we have to say here in this
introduction. It is enough for us to point out that we assume that grounding is most
certainly involved with metaphysical explanation, however that turns out to be, and
move on.
It has been suggested that the connection between ontological dependence and
explanation is weaker than the connection between ground and explanation. Tahko
and Lowe suggest, for example, that the existence of hydrogen and oxygen—upon
which water depends—do not, alone, explain the existence of water.8 Whilst we agree
that the mere existence of hydrogen and oxygen does not fully explain the existence
of water, we struggle to understand how the existence of the two could fail to be
appealed to in an explanation of the other. Perhaps Tahko and Lowe are correct that
the connection is weaker, but we here feel confident proceeding on the assumption
that ontological dependence is sufficiently strongly tied to metaphysical explanation
nonetheless.
Let us turn now to the notion of fundamentality itself. We assume that the categories
of fundamental and derivative are exclusive and exhaustive. Some entity is either
fundamental or derivative but never both.9 The category of derivative things is just
8 9
See Tahko 2015. See Barnes 2012 for arguments against the exclusivity assumption.
 the geography of fundamentality: an overview
the category of metaphysically dependent things; which is just to say it is the category
of grounded and ontologically dependent entities. It is true by definition that a
derivative entity is dependent and, thus, that it has a metaphysical explanation. The
fundamentalia, on the other hand, by definition, depend upon nothing else (except
perhaps themselves) and are, thus, without metaphysical explanation (except perhaps
in terms of themselves). This is not to say, however, that being independently existent
is a sufficient condition for being fundamental (on some accounts, it’s not even
necessary). There may well be a plethora of independent entities that, nonetheless,
do not serve as candidate fundamentalia.10 Although there are alternative ways of
understanding fundamentality, such as discussed by Takho and Barnes (this volume),
Fine, and Sider, we are happy to proceed on the independence understanding.11
It is open, and indeed the case on many accounts, that the fundamental facts be
fundamental qua grounding structure and yet dependent qua ontological dependence
structure. This is because for any account according to which a fact is dependent
upon its constituents, a fundamental fact will be ungrounded and yet, nonetheless,
dependent. The term ‘fundamentalia’ can then be taken to refer to either fundamental
facts or fundamental things depending upon which ordering one wishes to fore-
ground.
We recognize that there are also subtly different ways in which the notion of
being fundamental can be formally cashed out. One distinction that we think it
particularly important to mention is that between the relation being well-founded and
it having a lower bound.12 To say that dependence relations are well-founded is to say
that (i) chains ordered by the relation downwardly terminate in a fundamentalium,
and (ii) that there is a finite number of steps between any member of a chain and
the fundamentalium that it terminates in. Although it’s not uncommon to hear
philosophers speak in the language of well-foundedness, what they often mean is
that any chain of entities ordered by that relation has a lower bound. Importantly,
where a relation is bounded from below, there need not be a finite number of steps
between any member of that set and the fundamentalium that grounds it. To better
understand this, consider the relationship between God and the contents of reality;
although there may be an infinite number of steps between, say, the number 7 and
God, the number 7, along with everything else, depends on him nonetheless. In
order to remain neutral on an understanding of fundamentality as well-foundedness
and fundamentality as lower boundedness, we choose to capture this aspect of
foundationalism formally in terms of the notion of extendability (E) and its negation;
more of which anon.
10 Facts about numbers, for example, may be independent, without that entailing that they are therewith
fundamental.
11 Fine 2001 and Sider 2011. See Raven 2016 for another alternate account of fundamentality.
12 See Dixon 2016, and Rabin and Rabern 2016, for formal treatments and discussions of different
possible ways of understanding fundamentality.

ricki bliss and graham priest 
2 Taxonomy
The hierarchy thesis says that the dependence relation is anti-symmetric, transitive,
and anti-reflexive. The fundamentality thesis says that there must be something
fundamental. Although it is common to assume that the relevant dependence rela-
tions have some combination of the aforementioned properties, a variety of different
combinations are at least logically possible. To see this, let us first introduce some
notation.13

We write ‘x depends on y’ as x → y.14 (We may write x → x as x .) Next, four
structural properties:
Anti-reflexivity, AR.
• ∀x¬ x → x [Nothing depends on itself.]
• So ¬AR: ∃x x → x [Something depends on itself.]
Anti-symmetry, AS.
• ∀x∀y(x → y ⊃ ¬ y → x) [No things depend on each other.]
• So ¬AS: ∃x∃y(x → y ∧ y → x) [Some things depend on each other.]
Transitivity, T.
• ∀x∀y∀z((x → y ∧ y → z) ⊃ x → z) [Everything depends on anything a
dependent depends on.]
• So ¬T: ∃x∃y∃z(x → y ∧ y → z ∧ ¬x → z) [Something does not depend on
what some dependent depends on.]
Extendability, E.
• ∀x∃y(y = x ∧ x → y) [Everything depends on something else.]
• So ¬E: ∃x∀y(x → y ⊃ y = x) [Something does not depend on anything else.]
We can now give a taxonomy, which is as follows. After the enumeration column,
the next four columns list the 16 possibilities of our four conditions.
AR AS T E Comments Special Cases

1 Y Y Y Y Infinite partial order I
2 Y Y Y N Partial order A, F, G
3 Y Y N Y Loops I
4 Y Y N N Loops F, G
13 The contents of this section are reproduced from Bliss and Priest 2017.
14 One may distinguish between full dependence and partial dependence. (See e.g. Dixon 2016, sec. 1.)
Just to be clear: the notion of dependence we are concerned with here is partial dependence.
 the geography of fundamentality: an overview
5 Y N Y Y ×
6 Y N Y N ×
7 Y N N Y Loops of length >0 I
8 Y N N N Loops of length >0 F, G
9 N Y Y Y ×
10 N Y Y N ×
11 N Y N Y ×
12 N Y N N ×
13 N N Y Y Preorder C, I
14 N N Y N Preorder C, F, F , G
15 N N N Y Loops of any length I
16 N N N N Loops of any length F, F , G
Consider, next, the Comments column. Here’s what it means.

• There is nothing in categories 5, 6, since if there are x, y, such that x y, then

by T, x y , contradicting AR. (¬AS and T imply ¬AR.)
• There is nothing in categories 9–12, since if for some x, x → x, then for some
x and y, x y, contradicting AS. (¬AR implies ¬AS.)
• All the other categories are possible, as simple examples (left to the reader) will
demonstrate.
• In cases 13–16, since ¬AR implies ¬AS, the second column (AS) is redundant.
• In categories 1 and 2, → is a (strict) partial order; and in category 1, the objects
involved must be infinite because of E.
• In categories 13 and 14 → is a (strict) preorder, so loops are possible. (A loop
is a collection of elements, x1 , x2 , . . . , xn−1 , xn , for some n 1, such that
x1 → x2 → . . . → xn−1 → xn → x1 .)
• In cases 3, 4, 7, 8, 15, 16, transitivity fails, and there can also be loops. In cases 7, 8,

there are no loops of length zero, x , since AR holds.
Turning to the final column, this records some important special cases.
• The discrete case is when nothing relates to anything. Call this atomism, A. In this
case, we have AR, AS, T, ¬E. So we are in case 2 (though this is not the only thing
in case 2).
• If → is an equivalence relation (reflexive, symmetric, transitive), we have ¬AR,
¬AS, T, so we are in cases 13 or 14 (though this is not the only thing in these
two cases). In case 13, there must be more than one thing in each equivalence
class, because of E. A limit case of this is when all things relate to each other:
∀x∀y x → y. Call this coherentism, C.
• Call x a foundational element (FEx) if there is no y on which x depends, except
perhaps itself: ∀y(x → y ⊃ x = y). Foundationalism, F, is the view that everything
grounds out in foundational elements. One way to cash out the idea is as
ricki bliss and graham priest 
follows.15 Let X0 = {x : FEx}, and for any natural number, n ∈ ω: x ∈ Xn+1

iff x ∈ Xn or ∀y(x → y ⊃ y ∈ Xn ). X = Xn . F is the view that everything
n∈ω
is in X, ∀x x ∈ X.16 Intuitively, this means that everything is a foundational
element, or depends on just the foundational elements, or depends on just those
and the foundational elements, and so on. E entails that there are no foundational
elements. Hence, this is incompatible with F. So, given F, we must be in an even
numbered case—except those that are already ruled out by other considerations.
(All are possible. Merely consider x → y → z. z is foundational; add in arrows
as required to deliver the other conditions.)
• A special case of foundationalism is when the foundational objects, and only
those, depend on themselves: ∀x(FEx ≡ x → x). Call this view F . Since AR
must fail in this case, we must be in cases 14 or 16 of the taxonomy.
• Another special case of foundationalism is when there is a unique foundational
object on which everything else depends: ∃x(FEx ∧ ∀y(y = x ⊃ y → x) [Some-
thing is a foundational element, and everything else depends on it.] The x in
question does not depend on anything, except perhaps itself, and it must be
unique, or it would depend on something else. Call this case G (since the x could
be a God which depends on nothing, or only itself). This is a special case of F,
and could be in any of the cases in which F holds.
∗
• Write x → y to mean that y is in the transitive closure of → from x. That is, one
can get from x to y by going down a finite sequence of arrows. An element, x,
is ultimately ungrounded, UGx, if, going down a sequence of arrows, one never
∗
comes to a foundational element: ∀y(x → y ⊃ ¬FEy). Infinitism, I, is the view
that every element is ultimately ungrounded: ∀x UGx.17 We note that Infinitism
allows for the possibility of loops, that is, repetitions in the regress. Thus, we have
the following possibility: x → y → z → x → y → z →. . . . However, if →
is transitive and anti-symmetric (T and AS), such loops are ruled out. Infinitism
entails Extendability, E. So if I holds we must be in an odd numbered category
of our taxonomy (which is not ruled out by other considerations). All such are
possible, as simple examples demonstrate. (Merely consider x0 → x1 → x2 →
x3 →..., where these are all distinct. Add in other arrows as required.) Note that
if there are at least two elements, then C is a special case of I.
15 We note that, how, exactly, to cash out the idea of foundationalism is contentious. For some discussion
of the matter, see Dixon 2016. We suspect that the notion may be vague, to a certain extent, and so
susceptible to different precisifications. The definition we give here is strong, simple, and very natural.
16 One may, if one wishes, iterate the construction into the transfinite, collecting up at limit ordinals in
the obvious way.

17 We note that Infinitism, also, is certainly susceptible to various precisifications. For example, one
might require that only some element is ungrounded. Again, the definition we give here is strong, simple,
and natural.
 the geography of fundamentality: an overview
• A final special case. Let x y iff x → y ∨ y → x. Then x and y are connected

along the dependence relation, xCy, iff for some n 1:
x y ∨ ∃z1 z2 . . . zn (x z1 ∧ z1 z2 ∧ . . . ∧ zn y)
[Everything relates to everything else along some sequence of dependence rela-
tions.] → itself is connected iff ∀x∀y xCy. In all of the ten possible cases, →
may be connected or not connected. G is a special case of connectedness; C is
an extreme case of connectedness; and A is an extreme case of disconnectedness.
Let us finish this section with an informal summary. The taxonomy is built on four
conditions. (i) Anti-reflexivity, AR: nothing depends on itself. (ii) Anti-symmetry, AS:
no things depend on each other. (iii) Transitivity, T: everything depends on whatever
a dependent depends on. (iv) Extendability, E: everything depends on something else.
This gives us 16 (= 24 ) possibilities. Six of these are ruled out by logical considerations,
leaving ten live possibilities. Within these, some special cases may be noted. Atomism,
A: nothing depends on anything. Foundationalism, F: everything is a fundamental
element or depends, ultimately, on such. F : Foundationalism, where the fundamental
elements and only those depend on themselves. G: Foundationalism where the
fundamental element is unique. Infinitism, I: there are no fundamental elements.
Coherentism, C: everything depends on everything else.
3 On the Metaphysical Possibility of the Alternatives

So far, we have seen that alternatives to metaphysical foundationalism in general, and
the standard view in particular, are logically possible: lines 1–4, 7, 8, and 13–16. One
might wonder, however, if they are metaphysically possible. In this section, we will
argue that they are. But before turning to a discussion of the viability of the alternatives
to the standard view, let us first address one particular issue that we will face time
and again.
It is quite common to hear friends of the standard view defend their commitments
to various aspects of the view by appeal to their intuitions. These philosophers will
claim to have intuitions that there is something fundamental, that nothing can ground
itself, and so on. Moreover, these philosophers appear to take their intuitions to serve
as something like arguments in defence of the view: these philosophers will not only
claim to have said intuitions, but also that nothing more needs to be said on the
matter. We simply do not share these intuitions. In fact, neither of us has any intuitions
whatsoever regarding a subject matter as abstract and recherché as the fundamental
structure of reality. But, more importantly, we also firmly believe that intuitions are
no replacement for actual arguments. That intuitions have been allowed to play the
role they have in the dependence/fundamentality debates thus far is, in our view, why
alternative views have been so poorly explored, and why actual arguments in defence
of the view have been allowed to be so bad.
ricki bliss and graham priest 
In what follows, although appeal to intuition is often made in defence of one

commitment or another, we will not respond to them further. Our response in
each case is as stated here. Let us continue our investigation, then, by turning to a
consideration of actual arguments, beginning with the hierarchy thesis.
3.1 The Hierarchy Thesis
According to the proponent of the standard view, reality is hierarchically arranged.
That reality is like this, we are told, is intuitive and somehow obvious.18 It has been
suggested that to challenge the idea that reality has such a shape, by questioning
whether dependence relations are transitive, irreflexive, and anti-symmetric, is pre-
posterous for the reason that metaphysical dependence relations are introduced into
the philosophical vernacular exactly to capture this aspect of reality. A reason often
cited in favour of abandoning talk of supervenience—a symmetric and reflexive
relation—in favour of, say, grounding talk, is that we need a relation that can capture
reality’s hierarchical structure. We agree that if metaphysical dependence relations
are introduced exactly to allow us to capture the idea that reality has a hierarchical
structure, then it makes little sense to call into question the properties that are securing
that structure. But the important question, we think, is why we ought to believe reality
has such a structure in the first place. And it is when we focus on this question that
reasons so often offered to commit to the hierarchy thesis look less compelling. Let us
now consider them.
.. anti-reflexivity
In defence of the claim that dependence relations are necessarily anti-reflexive,
philosophers have tended to argue that it would be absurd to assume that something
can ground itself, or that, given the tight connection between grounding and expla-
nation, as it is a principle of explanation that nothing explains itself, it ought to also
be a feature of dependence relations.19
Let us first consider why one might think it absurd to assume that metaphysical
dependence relations can be reflexive. As dependence talk is about reality, it is
reasonable to wonder if self-dependence is absurd because there is some way that
the world would have to be, such that things can depend on themselves, which is
unacceptable. But what might this be?
A first worry about self-dependence is that anything that depends upon itself would
have to bootstrap itself into being. But why think this is a problem? In the case
of causation, the problem is apparent: something that is self-caused would have to
exist prior to itself in time in order to bring itself into existence. But metaphysical
dependence relations are typically thought of as being synchronic, so what goes
18 See Raven 2013 for a well-articulated defence of the hierarchy thesis.

19
See Jenkins 2011for a somewhat different discussion of dependence and irreflexivity.
 the geography of fundamentality: an overview
for causation here does not (necessarily) go for metaphysical dependence.20 As

metaphysical dependence relations are thought of as inducing a priority ordering,
perhaps the problem, then, is that where the relations are reflexive, the very idea of a
priority ordering goes out the window. This may well be the case, but of course this is
no reason to think that dependence relations cannot be reflexive, for it is just to assert
that the relation must be anti-reflexive in the first place. Exactly what is required in
order to have a priority ordering is that the ordering relation is anti-symmetric and
anti-reflexive.
Anyone with even a passing familiarity with the historical literature would be aware
that there is, in fact, precedent for a view according to which there is at least one thing
that is self-dependent, namely, Leibniz’s account of God. According to Leibniz, God
exists, indeed, exists necessarily. He does so because existence is part of his essence;
but to say this means, inter alia, that God necessarily exists. So God necessarily exists
because he necessarily exits. One might wonder, then, if a good reason to reject
the possibility of reflexive instances of ground is that anything that is self-grounded
would be a necessary being. Now, of course this is only going to be a problem if the
wrong things, or kinds of things, turn out to be self-grounded; take, for example,
the fundamentalia. A potential serious worry, then, is that if the fundamentalia are
necessary beings, and they ground the being of everything else, then there is only one
way the world can be, which is exactly how the world actually is.21
Are we compelled, though, to accept this story—the story according to which
self-grounded entities are necessary beings? Bliss (this volume) suggests that we are
not. But if this is the case, we seem no closer to understanding (i) what reflexive
dependence amounts to and (ii) why it is unacceptable. Failing all else, one might
simply worry that the idea that anything can depend upon itself is absurd just because
it is plain weird. Maybe it is weird (the judgement of which would seem to require
knowing what self-dependence actually amounts in), but we struggle to see how self-
dependence is any weirder than the commonly held belief that there are some entities
that pop into being from nowhere and for no reason at all—which is exactly what the
fundamentalia are like by most people’s lights. Metaphysically speaking, it is not so
clear what is so bad about something’s being self-dependent.
More compelling, we think, are explanatory reasons for thinking that reflexive
instances of dependence are unacceptable. It is a plank in much of the literature on
explanation that reflexive explanations are trivial, uninformative, and explanatorily
useless. A reflexive explanation, so the thought goes, is as good as no explanation
at all. We are inclined to think, though, that whilst there may be something to this,
matters here are thornier and more subtle than they appear.22 For a start, not all
20 There are reasons to believe that there are cases of non-synchronic grounding, just as there are cases
of synchronic causation. Obviously the intricacies of these issues cannot be covered here.
21 See Dasgupta 2016 for a discussion of this view according to which it would not be a problem.
22 See Keefe 2002 for a most illuminating discussion of issues relevant to this debate.
ricki bliss and graham priest 
circular explanations are trivial—we have already seen this in the case of God and
his explanatory relationship to himself—nor are they necessarily uninformative or
useless. After all, coming to understand that something has no further explanation is
coming to understand something more about that thing. In the worst case, what we
may be dealing with is a problem with explanatory superfluity: something’s explaining
itself is as good as it having no explanation whatsoever, so why bother permitting self-
dependence in the first place.
As things stand, the reasons to disavow self-dependence appear to be fairly thin on
the ground. Metaphysically speaking, it’s not clear how a world would have to be such
that things depend on themselves, leaving us with explanatory considerations. But
if this is the conclusion it is hardly welcome. Suddenly the problems with reflexivity
appear to be epistemic rather than metaphysical which would seem to fly in the face of
how the friends of foundationalism understand the overarching structure of reality.
.. anti-symmetry
Let us now turn our attention to anti-symmetry. Advocates of the standard view rely
on (some combination of) arguments from intuition, arguments from the data, and
arguments from structural similarities with explanation. Appeal is also made to what
we might call arguments from relative fundamentality. The argument from relative
fundamentality is just a variation on the kind of argument in terms of structure that
we mentioned in the introduction to this section. We consider these first.
According to the argument from relative fundamentality ‘dependence is intimately
connected to (and perhaps even explains or is one and the same things as) relevant
notions of fundamentality, priority, grounding, etc. Dependence is the kind of rela-
tion that explains the connection between the fundamental and the derivative (the
dependent) to the fundamental (the independent). Any relation that plays this role
must be asymmetric’ (Barnes, this volume). The idea that reality is ordered into a
hierarchical structure is a very old one that can be traced back to the Ancient Greeks.
Indeed, right the way through the history of the Western tradition, many philosophers
have been engaged in some way or other with filling in the details of this picture.23
That some folks claim to have intuitions regarding the structure of reality is hardly
surprising given the pervasiveness of this view (and imagery) in the history of Western
thought.24
As Barnes points out, if moving us from the fundamental to the derivative is the role
that dependence is supposed to play, then it seems right to suppose that dependence
must be anti-symmetric. Indeed, as already mentioned, it just follows from the idea
23 See Lovejoy 1934 for an informative and charming discussion of the notion of the Great Chain of
Being and its centrality to the development of Western metaphysics.

24 The idea that reality is hierarchically structured has not only been the purview of the metaphysician,
but was also commonplace in the sciences, art, and theology up until the end of the nineteenth century.
This view went out of vogue with the momentous changes to our understanding of the world precipitated
by scientific developments.
 the geography of fundamentality: an overview
that reality is hierarchically structured that the structuring relation is anti-symmetric.

But exactly what the argument from relative fundamentality does not provide us with
is a reason to suppose that the relation is anti-symmetric—it simply assumes it. One
way to respond to the relative fundamentality argument, then, is to challenge the idea
that we have reasons to suppose that reality is hierarchically structured in the first
place. To put the point more finely, we can challenge the idea that reality exhibits a
robust hierarchical structure by arguing that metaphysical dependence relations are
either symmetric (which might generate a species of metaphysical coherentism) or
that they are non-symmetric (a weaker claim that may yet allow for a hierarchy to
emerge nonetheless).
Whilst we agree that the world appears to present us with cases of anti-symmetric
dependence, that dependence relations are necessarily anti-symmetric is not obvious
to us at all. As Barnes and Thompson (this volume) argue, some of our most beloved
metaphysical theories appear to posit symmetric instances of dependence; or at least
make more sense if they do. Consider, for example, Armstrong’s account of states
of affairs. Armstrong’s picture is one according to which atomic states of affairs are
ontological rock-bottom with their constituents as abstractions from those states of
affairs. The problem with this picture is that the states of affairs really seem to depend
on their constituents, with those constituents explaining the nature and existence of
those states of affairs. Barnes suggests that if Armstrong were to allow symmetric
instances of dependence, then he could have his cake, as it were, and eat it too: atomic
states of affairs depending on their constituents, but the constituents depending on
their state of affairs.
Theoretical cases aside, consider also the relationship between the north and south
poles of a magnet: without the north pole, the south pole would not exist and without
the south pole, the north pole would not exist. The list, it would seem, goes on. We
appear to have compelling reasons to temper our commitment to anti-symmetry and
endorse the more modest suggestion that the relation(s) is non-symmetric.
What about the much stronger claim that dependence is, in fact, symmetric?
Can the case be made for such a strong view? Well, it can because it has been. As
Priest (this volume) discusses, the Chinese Huayan Buddhist tradition endorses a
species of full-blown coherentism with everything depending symmetrically upon
everything else.25
It has been suggested that as metaphysical dependence is intimately involved
with explanation, we can infer from the structural properties attributed to (good)
explanations on some models that metaphysical dependence relations also share such
properties. As explanations are anti-symmetric, so the objection goes, so too are
dependence relations. But as both Barnes and Thompson (this volume) point out,
there are alternative (very good) explanatory models on which explanations are not
25 See also Priest 2014, esp. chapters 11–13, for a contemporary presentation of a coherentist picture
inspired, in part, by Huayan.

ricki bliss and graham priest 
necessarily anti-symmetric. Indeed, according to Barnes and Thompson, explanation

as understood wholistically, may well do a better job of capturing certain aspects of
our everyday and theoretical explanatory praxis.
That metaphysical dependence relations are introduced to capture reality as hierar-
chically structured does not provide us with a reason to think reality has that structure
in the first place. Although ‘the data’ suggests that some instances of dependence
relations are anti-symmetric, this is also no reason to suppose that the relation is in
general. Indeed, the cherry-picking of instances of dependence relations that appear
to be anti-symmetric to use as our paradigmatic cases of dependence ought not blind
us to the presence of other instantiations of the relation that are plausibly thought to
be symmetric. All told, there seems to be good reasons to suppose that metaphysical
dependence relations are at least non-symmetric.
.. transitivity
There is something natural-seeming about the idea that metaphysical dependence
relations are transitive. Where a person depends on their vital organs, it also seems
true that they depend upon the cells that compose those vital organs. However,
a number of authors, including Nolan (this volume), have pointed out that at the
very least, we could well allow that some instances of dependence relations fail to
be transitive, and hold a view according to which metaphysical dependence is non-
transitive.
Why question the transitivity assumption? Well, one good reason is that reality
appears to present with actual cases of failures of transitivity. Schaffer asks us to
consider the following propositions: (1) the fact that o has a dent, d, grounds the fact
that o has shape S, (2) the fact that o has shape S grounds the fact that o is more or less
spherical, and (3) therefore, the fact that o has a dent grounds the fact that o is more or
less spherical. If grounding were transitive, then we would expect this argument to go
through but, Schaffer argues, it does not because o ‘is more-or-less spherical despite
the dent, not because of it’.26 As far as Schaffer is concerned, the fact that o has a dent
does not ground the fact that o is more or less spherical, in which case grounding is
not necessarily transitive.27 Or consider other problematic cases: singleton Obama is
dependent upon its member Obama, and Obama is dependent upon his parts, and yet
we might well not want to say that the existence of Obama’s heart (partially) explains
the existence of singleton Obama.
One way to respond to these sorts of cases is to point out that dependence is not
univocal. What one might think is going on in these cases is the chaining together of
instances of dependence relations that don’t, in fact, properly belong together. One
might try and argue, for example, that the way in which a singleton depends upon
26 Schaffer 2012, p. 127.

27 This is not the only purported failure of transitivity that Schaffer presents us with. See also Raven 2013
for a defence of the thought that grounding is transitive.
 the geography of fundamentality: an overview
its member is different to the way in which the member depends upon its parts.28
Were one to pursue such an approach, however, one must remain mindful of the costs
such an approach might incur: do we really want or need a proliferation of species of
dependence relations, for example?29
Another approach might be to distinguish between relations of mediate and imme-
diate dependence, where the former is transitive and the latter is not. Indeed, in the
literature, philosophers have suggested that we should take seriously a distinction
between immediate and mediate dependence.30 Purported failures of transitivity can
then be understood as involving the transitive closure of an intransitive relation. So
what appears to be a failure of transitivity, in fact, involves a case of mistaken identity.
There are advantages to admitting a distinction between a transitive and a non-
transitive species of the relation. On the one hand, it allows us to avoid a proliferation
of relation-types in response to the purported problem: where part/whole relations are
a species of dependence relation, truthmaking another and so on. And, on the other
hand, it allows for certain possibilities. Nolan (this volume) suggests, for example,
that some species of dependence, or instances of the relation, may fail to be transitive
allowing the possibility of giant cosmological loops. And more generally, where there
is a species of the relation that is intransitive, loops of various sizes could be admitted
without being forced to sacrifice anti-symmetry and anti-reflexivity. All told, there
are reasons to doubt that metaphysical dependence relations are necessarily transitive.
Not only do we appear to be in possession of counterexamples to the transitivity thesis,
but we have reasons to suppose that admitting an intransitive species of the relation
to our repertoire would be to our advantage.
Of course there is so much more to be considered regarding the widespread
commitment to the hierarchy thesis, and the possible alternatives to it. Rabin (this
volume) believes that unorthodox accounts of grounding allow us to better capture the
layered conception of reality. Looking to other traditions, as Priest (this volume) does,
we can see that a number of accounts from the Asian Buddhist traditions, for example,
reject the idea that reality is hierarchically structured. Anyone seriously interested in
non-standard conceptions of the structure of reality would do well to look beyond the
Western canon. And Litland (this volume) argues that, what he calls a bi-collective
account of ground, may have interesting applications for certain types of coherentist
structures.
28 Consider what happens when we say that Harry banks on Sally and Sally banks on Tuesday. No one
would claim that, therefore, Harry banks on Tuesday. Nor would anyone claim that the relation banking on,
as demonstrated by this example, is not, therefore, transitive. What we would be inclined to say is that the
expression ‘banks on’ picks out different relations in the two cases.
29 See Wilson 2014 for a defence of the claim that all we need are the many different kinds of small-
g grounding relations with which we are familiar—supervenience, parthood, etc.—rather than one big-G
grounding relation.
30 See, for example, Fine 1994, 1995, and 2013 for discussions of such possibilities as regards both
ontological dependence and grounding.

ricki bliss and graham priest 
3.2 The Fundamentality Thesis

One might well have the impression that nary a paper is produced in analytic meta-
physics these days that does not make reference to the notion of fundamentality.31
Somewhat surprising, then, is the dearth of good arguments available in the literature
in defence of the fundamentality thesis. Broadly construed, there appears to be at
least three types of argument on offer. The first of these, as might be expected, are
arguments from intuition; the second of these are arguments from vicious infinite
regress; and the third, arguments from theoretical virtue. In keeping with our promise
above, we desist from discussing arguments from intuition and turn immediately to
regress arguments.
.. regress arguments

What explains the fact that we exist? A good place to start will surely involve appeal to
facts about the existence of our parents and the genetic material they have bequeathed
to us, our vitals organs, and so on. Of course, we are causally dependent upon on our
parents, but we are also metaphysically dependent upon them: it’s not simply that our
parents cause us to exist, but they also ground our existence as well. Although the
story of the existence of any one of us is metaphysically complex, most of us would
feel confident in assuming that we have some rough idea of how to tell it. Suppose,
now, that we also wish to explain the existence of our parents and our vital organs.
Again, a complex matter, but surely one that will involve appeal to their parents—our
grandparents—and the cellular structure of the organs and so on. At each stage, it
would appear as though we have explained something about the entities for which
we are seeking an explanation, and that this process could go on successfully without
termination. But exactly what the fundamentality thesis tells us is that it doesn’t (or
can’t) go on forever, and a justification for this position is going to have to tell us why
this is the case.32
One obvious seeming thought is that where we have limitless descending depend-
ence chains, although we have explained something (probably even a lot), we haven’t
yet explained everything that we need an explanation for.33 Or another thought might
be that where we have limitless descending dependence chains, although we have
explained something, we haven’t yet arrived at an explanation that is complete, or
at least completely satisfactory. And, of course, there is a way of understanding these
two explanatory concerns that is intimately related, for an explanation will surely be
31 One needs not only be reading from the dependence/fundamentality literature to notice this. Appeal
to fundamentality is made in the literature ranging from topics as diverse as the philosophy of mind to
aesthetics and ethics, to offer but a few examples.
32 See Bliss (forthcoming), from which much of the following discussion in this section is borrowed, for
a more sustained elaboration of these thoughts.

33 See Bliss 2013.
 the geography of fundamentality: an overview
unsatisfactory exactly when we have failed to explain everything that we need an

explanation for.
Both of these kinds of concerns are echoed by various authors in the literature.
Schaffer, for example, claims that where there is nothing fundamental ‘being is
infinitely deferred and never achieved’.34 Dasgupta suggests that it is at least plausible
to think that we might justify our commitment to fundamentality as ‘the desire for
this special kind of explanation . . . in which one looks at the surrounding mountains
and oceans and thinks “good grief, how come it all turned out like this?”’35 Where the
‘special kind of explanation’ he refers to is exactly the kind of explanation we don’t
have when we point out that mountains depend upon arrangements of matter in
space, and so on. Although concluding that the best reason we have for supposing
that there is something fundamental is that it would be better to have a unified
explanation of everything that needs explaining, Cameron also states that ‘for if there
is an infinitely descending chain of ontological dependence, then while everything
that needs a metaphysical explanation (a grounding for its existence) has one, there
is no explanation for everything that needs explaining. That is, it is true for every
dependent x that the existence of x is explained by the existence of some prior object
(or set of prior objects), but there is no collection of objects that explains the existence
of every dependent x.’36 And finally, concerned with satisfaction, Fine suggests that
‘ . . . given a truth that stands in need of explanation, one naturally supposes that it
should have a “completely satisfactory” explanation, one that does not involve cycles
and terminates in truths that do not stand in need of explanation’.37
An unfortunate consequence of the alleged obviousness of the fundamentality
thesis is that remarks such as these are seldom presented in the form of arguments
in the literature. It is not uncommon, nor unreasonable, to suppose that comments
such as these can be reconstructed in the form of arguments from vicious infinite
regress. One might suppose, for example, that where there is nothing fundamental, a
regress is generated, and it is vicious because it leaves us without an explanation for
something that we think needs explaining, or that we are left without an explanation
that is completely satisfactory.
We think, however, that there is a simpler way of reconstructing arguments in
defence of fundamentality of this stripe that does not make direct appeal to arguments
from vicious infinite regress.38 Reconstructing the arguments after such a fashion has
the added advantage of allowing us to bring to the fore an assumption crucial to the
foundationalist view that appears to have gone largely unnoticed in the literature.
One way of reconstructing the kinds of claims mentioned above as arguments
requires two assumptions. The first of these is an assumption that stipulates an
34 35 36
Schaffer 2010, p. 62. Dasgupta 2016, p. 4. Cameron 2008, p. 12.
37 Fine 2010, p. 105.
38 This approach also fits with our view—which we have argued for independently—that the infinite
regress is, in most cases, never the disease but, rather, a symptom. See Bliss 2013 and Priest 2014, 1.4.
ricki bliss and graham priest 
explanatory target. Such a stipulation might make appeal to something that needs to be
explained; or it might make appeal to a type of explanation. In light of the discussion
above, our explanatory targets might include (i) why anything has being whatsoever,
(ii) why things turned out this way rather than any other, or (iii) that we need
completely satisfactory explanations of everything that we think needs explaining.
But note that having stipulated what our explanatory target is, or could be, we do
not yet have an argument in defence of fundamentality. For it is not enough that
we know that there is something that needs explaining, or some particular kind of
explanation that we are after, but we also need an assumption that tells us that no
dependent entity is up to the task to hand. Arguments in defence of fundamentality
rely, crucially, on a second assumption which tells us that no dependent entity can do
the kind of explanatory work that we are after.
For the sake of economy let us reconstruct two possible arguments in defence
of fundamentality; arguments that are congruous with suggestions made in the
literature.39 Assuming that the world divides exclusively and exhaustively into the
fundamental and the derivative:
Argument I
1. There is an explanation for why anything has being whatsoever.
2. No dependent entity can explain why anything has being whatsoever.
3. Therefore, there must be something fundamental.
Argument II
1. There is a complete metaphysical explanation for things that have metaphysical
explanations.
2. No dependent entity can generate a complete explanation for things that have
metaphysical explanations.
3. Therefore, there must be something fundamental.
What are we to make of these arguments? In particular, are these good argu-
ments in defence of the fundamentality thesis? Let us begin by considering the first
assumption of our first argument. It seems obvious that what is at issue on this
kind of reconstruction is a variation on an old theme: the cosmological argument.
Understood in this way, the foundationalist is concerned to answer some version or
other of a cosmological question. Indeed, many historically important figures have
been engaged with such explanatory projects, including, as Casati (this volume) points
out, Heidegger. Foundationalism, so understood, is of course not motivated by a
concern to establish an ultimate cause of reality, but, rather, by a concern to establish
an ultimate ontological ground.
39 We are of the view that many of the philosophers who worry about the grounds of being, or
explaining the existence of everything, and so on, are, in fact, circulating roughly in the same waters. These
philosophers are concerned with age-old questions such as why are there any beings whatsoever.
 the geography of fundamentality: an overview
Before assessing the merits of these arguments, it is interesting to note that their
very appearance would appear to be in tension with what is a common view amongst
contemporary analytic thinkers. Inspired by Hume, it is not uncommon for philoso-
phers to suppose that having explained the existence of this thing here, and the
existence of that thing there, everything that needs a (causal) explanation has one. This
is just to say that, following Hume, many folks are of the view that there is nothing left
over that needs to be explained and therewith, no blazing cosmological questions that
demand an answer. Indeed, some philosophers have even gone so far as to claim that
cosmological questions are ill-formed and non-sensicle.40 It is an item of curiosity
why it is, then, that in the causal case, cosmological arguments (and the kinds of
questions they are offered in response to) are passé and, yet, in the metaphysical case
they are not. This is not to say that there is not a principled reason for the difference,
but that it would be nice to know what it is.
Sociological observations aside, there is what we believe to be a considerable
concern with the use of cosmological questions to motivate metaphysical founda-
tionalism: they appear to rely on an application of the principle of sufficient reason
(PSR). Although there may be a suitably constrained version of the principle in the
vicinity, the employment of the full-blown principle—according to which every thing
has an explanation for its existence—to motivate foundationalism would be a disaster
for the view: exactly what the foundationalist believes is that not everything has an
explanation. Metaphysical foundationalism, so motivated, runs the risk of pulling the
rug out from beneath itself.
Let us now turn to the second argument and consider the thought that there is a
complete metaphysical explanation for things that have metaphysical explanations.
We do not wish to be distracted by how we have formulated the assumption here.
Whether we formulate the target as all or only some things that have metaphysical
explanations have complete explanations, what we are concerned with is why we
should think anything that has a metaphysical explanation has a complete one in the
first place. So what can we say about this assumption? One might suppose, as Fine
does, that it is a plausible demand on explanations that they be completely satisfactory.
Alternatively, one might be of the view that, independently of any general explanatory
considerations, it is a plausible demand on metaphysical explanations in particular
that they be completely satisfactory.
But there appears to be a lot to be concerned about with the first assumption in
our second argument as well. First and foremost, there is a way of understanding
the assumption that looks as though it simply begs the question. We assume that
no argument in defence of fundamentality can contain an assumption from which
it follows that there is something fundamental. But the demand that some (or all) of
our metaphysical explanations be complete just seems to be the demand that those
40
See Maitzen 2012 and 2013.
ricki bliss and graham priest 
explanatory chains terminate, which, of course, is just to say that there must be
something fundamental.
A good reason to think that our metaphysical explanations ought to be complete
is that there is something wrong with explanations (in general) that are incomplete.
But explanations are not typically rendered defective by dint of being incomplete.
If someone wants to know why their window is broken, a story that makes appeal
to the storm the previous night would be adequate. It is simply not the case that an
explanation for a broken window is rendered defective in virtue of its failing to make
appeal to the origins of the universe. Of course, what goes for causal explanations
needs not go for metaphysical explanations, and the foundationalist may well be better
off making recourse to the idea that there is something special about metaphysical
explanations in particular which means they must be complete.
We think it is worth pointing out at this juncture that there is something of
an odd tension between the demand, on the one hand, for completely satisfactory
explanations that can only be achieved by terminating our dependence chains and, on
the other hand, the notion of a full ground. Let us suppose that singleton Socrates—
{Socrates}—is fully grounded in Socrates. The way we are often encouraged to
understand what full grounding amounts to is that, in this case, the existence of
{Socrates} is fully explained by the existence of Socrates. If, however, where the non-
terminating dependence chain of which these two are members leaves us with an
incomplete or a not completely satisfactory explanation of {Socrates}, this would
indicate that Socrates doesn’t completely explain {Socrates} in the first place. If, on the
other hand, Socrates does completely explain the existence of {Socrates}, then there
must, in fact, be something else at issue such that our explanations are not satisfactory
unless there is something fundamental.
Returning to broader explanatory considerations, one thought might be that whilst
causal explanations may be incomplete, metaphysical explanations cannot be, for it
is the purview of metaphysical explanations to afford us a complete explanation of
reality. We can’t help but think that something a bit slippery has gone on here, though.
First, where there is something fundamental, exactly what we don’t have is a complete
explanation of reality, for we have the fundamentalia that are unexplained. Second,
this proposal looks a lot like a cloaked version of the question-begging insistence that
there is something fundamental mentioned above.
Let us turn now to a consideration of the second assumptions in our arguments. As
pointed out above, to note that there is something that has not yet been explained, is
not yet to have an argument in defence of fundamentality. What is further required
is an assumption that stipulates that no dependent entity is up to the task to hand.
Without such an assumption, we have no need to move beyond the collection of
dependent entities and, thus, no need to posit the existence of something fundamen-
tal. If our foundationalism, however, is to be well motivated, we need to know why
this is the case. We need an answer to the question, why can’t any dependent entity
explain where, say, being comes from?
 the geography of fundamentality: an overview
We think there are at least five prima facie plausible reasons to suppose that no
dependent entity is able to be invoked to explain that for which it is being invoked.
We list these as follows: (a) the reflexivity objection; (b) the never-ending questions
objection; (c) the same questions objection; (d) the predicate-satisfaction objection;
(e) the same kind objection. We discuss each in their turn.
Some versions of cosmological arguments to the existence of God arrive at their
conclusion by pointing out that no contingent thing can explain why there are any
contingent things at all on pain of violating an anti-reflexivity assumption. They
claim that as any contingent thing would be amongst the collection of things to be
explained, were something contingent to explain why there are any contingent things
at all, then the collection would be self-explanatory. Or put slightly more formally,
let [A] be the state of affairs described by A. Suppose that there only two states of
affairs, [A] and [B], and that [A] causally explains [[A] and [B]], then [A] causally
explains [A] (and [B]). One might think that an analogous worry is what motivates the
metaphysical foundationalist. The worry in this case would be that where [A] grounds
[[A] and [B]], [A] grounds [A] (and [B]).
We think there are at least two reasons to reject this concern as a reason for
accepting the second assumption of the proposed foundationalist argument. The
first of these pertains to understanding the explanatory target as a conjunctive fact.
Cashing out the foundationalist concern over the ground of being in terms of a
giant conjunctive fact doesn’t seem to really respect the concern that is driving the
view in the first place. Moreover, the logic of ground, as it is commonly understood,
is such that conjunctions are grounded in their conjuncts—exactly what explains
the super conjunction are its conjuncts. Our second reason for rejecting this way
of understanding the foundationalist concern also relates to the logic of ground.
Although [[A] and [B]] necessitates [A] and [B] it does not metaphysically explain
them. Quite on the contrary, as we have seen. The reflexivity objection, we would like
to suggest, provides us with no good reason to suppose that no dependent entity can
explain why anything has being whatsoever.
Perhaps the reason we ought to endorse the second assumption, then, is that were
our chains not to terminate we would be forced to ask a never-ending series of
questions: dependent entities, by their very nature, have explanations, and for every
new dependent entity that we invoke, we can ask of it ‘why does this thing exist?’ (or
something of the like) ad infinitum. It’s not hard to see how some of the concerns
extant in the literature can be understood in these terms. Exactly what a never-
ending series of questions would seem to leave us without is a completely satisfactory
explanation, for example.
Once again, we find this line of reasoning—the never-ending questions’ objection—
wanting. Why? In short because it appears to us to beg the question. When do we cease
to ask questions? When we arrive at the existence of something that does not demand
that we ask of it certain (relevant) questions. And when do we arrive at the existence
of such a thing? When we arrive at something fundamental, of course. To insist that
ricki bliss and graham priest 
our explanatory chains terminate is just to insist that there is something fundamental.
Or put another way, to insist that our explanations be completely satisfactory is just
to insist that there is something fundamental.
What would not be a question-begging motivation for endorsing the never-ending
questions’ objection would be if we had an independent reason for endorsing it: a
reason over and above the mere stipulation that explanatory chains need to terminate.
An independent reason to endorse the never-ending questions objection might just be
that where we are forced to keep asking questions this must be because we haven’t
answered the question we are seeking an answer to in the first place. This is one way
we might interpret Schaffer’s concerns over the grounds of being, for example. Where
of each new thing we are compelled to ask ‘and what explains the being of this thing?’
one might suppose we have not really answered the question that we were seeking an
answer to in the first place.
Be that as it may, this reason to endorse the second assumption of the argument is
peculiar. Our reason for thinking so is that it seems to trade on a confusion. Where
we are forced to ask a never-ending series of questions, the problem may not be that
the chain does not terminate, but that one may be going about answering the question
in the wrong way. Put differently, the never-ending questions are not themselves the
disease, but, instead, a symptom of a deeper problem.
Interesting as this may be, this is not a good reason to suppose that no dependent
entity can explain where being comes from. Why? Let us grant that the series of never-
ending questions and answers is generated because we are going about answering the
questions in the wrong way. But if this is the case, what good will terminating the
chain do? How does terminating the chain at some, likely arbitrary, point solve our
problem if the problem is generated by a mismatch between question and answer in
the first place? It is hard to see how it could. Moreover, what reason could we have
for supposing that our answers are incorrect? If this reason for endorsing the second
crucial assumption is to play the justificatory role that we need it to, it cannot be
because we are lumbered with a never-ending series of questions and answers because
no dependent entity can explain why anything has being whatsoever. Exactly what we
appear to be left without is a motivation for the assumption.
Let us turn, then, to the same questions objection. Suppose one of us were to ask
you why there are any flamingos whatsoever. Suppose that you responded that there
are flamingos because there are an enormous number of them living in the Rann of
Kutch. Dissatisfied with your response, we might press you and say, ‘Ok, fine. So why
are there those flamingos?’ Were you to respond by pointing out that those particular
flamingos exist because their parents existed, we would be forced to suggest that you
seem to be missing the point. Whilst it is surely true that the particular flamingos
presently inhabiting the Rann of Kutch exist because their parents existed, no number
of flamingos can help us explain why there are any flamingos whatsoever. By parity
of reasoning, no dependent entity—entity with being—can help us explain why there
are any beings whatsoever. What is going wrong in both of these cases is that we are
 the geography of fundamentality: an overview
invoking the very thing for which we are seeking an explanation in our explanans.41
The problem is not that we have an infinitude of explanations, but rather that things
go badly out of the gate. We are forced to keep asking the same question because we
simply never receive an answer to it.
Whilst we find this line of reasoning compelling, what it seems to supply us
with—as with one interpretation of the never ending questions’ objection—is more
a restatement of the principle for which we are seeking a justification and less a
justification itself. The same questions objection seems to presuppose the idea that
no dependent entity can explain where being comes from rather than justify it.
But perhaps there is some principle lurking in the background here according
to which where F is any predicate that applies to dependent entities only, you can’t
explain why there are any F things at all by invoking only those things that are F, even
if your explanations go on forever.42 Let us call this principle the predicate satisfaction
principle. According to the predicate satisfaction objection, no dependent entity can
be invoked to explain why anything has being whatsoever because this would violate
the predicate satisfaction principle.
Although plausible seeming, we don’t think this is the right reason to endorse our
second assumption. The reason for this is that we do seem to allow explanations where
the G things explain the F things, but all the Gs happen to be Fs as well: all that is
required to explain why there are any F things at all is the G things that happen to
be the F things as well.43 Consider explanations of pain in terms of C-fibre firings.
Anything that satisfies the predicate ‘being in pain’ will also satisfy the predicate ‘has
C-fibre firings’, according to an appropriate version of physicalism, for example. As
much as we can explain why there are any pains at all, some theories do so in terms
of C-fibre firings, even though what satisfies the former predicate will also satisfy the
latter. Or how about the predicate ‘is the auditory threshold for the normal human
ear’? Let this predicate be denoted by F. The instantiation of this predicate is explained
by the G things—‘sounds falling within a range of 16 to 32 hertz’—where everything
that is a G is also an F. Other, non-scientific, examples also come to mind. Let F
be ‘is money’ and G be ‘is used as money’, for example. At first blush, the predicate
satisfaction principle appears intuitive and plausible, but it seems to be a principle
stronger than one that we ought to accept. We frequently explain why there are any
F things by making appeal to things that are, in fact, F things. So let us set this
principle aside.
Finally we come to what we call the same kind objection. According to this
objection, no member of a kind can explain why that kind exists at all. A reason to
endorse our second assumption would be that no dependent thing can explain, say,
41 See Bliss 2013 and Passmore 1970.

42 See Maitzen 2013, p. 263 for the formulation of the principle from which the one here was borrowed.
43 See Keefe 2002 for a discussion of ways in which explanations that fit this structure can be unprob-
lematic, and Maitzen 2013, p. 264 for an elaboration of the same point.
ricki bliss and graham priest 
why anything has being whatsoever because dependent things form a kind and no
member of a kind can explain why that kind exists at all. Again, at least one of us finds
this argument compelling (which is not to say it is well motivated!). And allusion to
the idea that no member of a kind can explain why that kind exists at all can be found
at various places in the literature.44
The argument also appears to be in keeping with at least one of the aforementioned
motivations for foundationalism. Let us return to the idea that metaphysical expla-
nations must be complete because it is the job of a metaphysical theory to give us
a complete story of reality. In previous remarks, we suggested that there is at least
one problem with this understanding of foundationalism cum metaphysical theory of
everything: it leaves something out, namely, the fundamentalia. What appears to be
implicit in this line of reasoning, however, is the idea that what we need an explanation
for is all the dependent entities. It at least accords with foundationalism understood in
this way that the world carves into two fundamental kinds—the derivative and the
fundamental—and that whatever is of the same kind as the derivative cannot explain
why there are any derivative things in the first place.
Understanding what motivates foundationalism in these terms, and as ultimately
being motivated by the same kind of objection, whilst plausible, brings with it its
own problems. There are going to be difficult issues associated with the thought that
‘dependent entity’ and ‘fundamental entity’ are kind-terms. Where ‘dog’ seems like
a good example of a kind term, it is less clear that ‘dependent entity’ is. Secondly,
foundationalism, so motivated, seems to land us in the awkward position whereby
the fundamentalia are invoked to explain the being of the dependent entities, but the
being of the dependent entities also explains the fundamentalia!
.. arguments from theoretical virtue

To the best of our knowledge, only one philosopher, Cameron, has explicitly endorsed
an argument from theoretical virtue in defence of the fundamentality thesis.45
Cameron argues that a theory of reality on which we have a unified explanation
of everything that needs an explanation is more virtuous than one on which we have
no such unity. And metaphysical foundationalism, unlike its rivals, is just such a
theory, according to Cameron.
In addition to the argument from theoretical virtue that is available in the literature,
one can imagine other possible arguments in the same spirit in defence of the funda-
mentality thesis. One might argue, for example, that metaphysical foundationalism
has the virtue of being parsimonious where its rival, metaphysical infinitism, does
not. Just as one might argue that foundationalism is simpler or more elegant than
coherentism.
44 45
See Lowe 2003, p. 91. Cameron 2008.
 the geography of fundamentality: an overview
Arguments from theoretical virtue are not designed to determine whether a theory
is impossible. Rather, their role is to help us adjudicate between theories that we
already believe to be possible. No argument from theoretical virtue, then, can lead
us to conclude that any one from amongst our theories is to be stricken from our list
of possibilities. But of course, what these arguments can do is help us make choices as
to which of our theories are better than the others.
That said, arguments from theoretical virtue are tricky, it seems, at least twice over.
On the one hand, how we are to understand the virtues is a matter of contention.
And, on the other hand, how the virtues interact with one another can make it hard
to determine, in some cases, when a theory is, in fact, better than another.
Consider the thought that foundationalism is more parsimonious than infinitism.
Are we to understand this as a claim regarding quantitative or qualitative parsimony?
If it’s the former, it is not entirely clear why we should believe this to be the case.
Moreover, it is not clear why we should believe that any foundationalist could, in fact,
run such an argument. Even though the infinitist denies that there is a fundamental
level, and is, therewith, committed to infinitely descending chains, foundationalism
says nothing about the number of entities that reside at the fundamental level; or any
other for that matter. It is not at all obvious, then, that infinitism is more ontologically
splashy than foundationalism after all, if what we are concerned with is the number
of things. Things may look differently, however, if what we are counting are the levels
themselves. Foundationalism does seem to do better as it does not commit us to
ever deeper layers or levels. But again, things here aren’t as straightforward as they
might appear. Were the world to be open at the top—with infinitely ascending layers,
then whether or not there is something fundamental makes little difference to the
parsimony of either view.46 Of course, it is not unreasonable to suppose that the world
is closed at the top, but how both infinitism and foundationalism fare in terms of
quantitative parsimony will be both complex and intimately involved with additional
commitments.
More often that not, what philosophers claim to be concerned with is qualitative
rather than quantitative parsimony. But here, again, matters do not appear to be
straightforward. Which view is more parsimonious than the other will depend upon
which kinds we think are there to be counted. On one way of carving up the space,
metaphysical infinitism is, in fact, more parsimonious than foundationalism; where
foundationalism has two fundamental kinds (the derivative and the fundamental)
infinitism only has one (the derivative). Suppose one were to argue, instead, that
qualitative parsimony pertains to kinds and not categories, and that terms such as ‘fun-
damental thing’ are category terms. What we ought to count, so this argument goes,
is all the cats, protons, and wave functions (rather than derivative and fundamental
things), and that where there is nothing fundamental there is surely an obnoxious
46
See Bohn 2009 and Schaffer 2010 for contrasting discussions of this possibility.
ricki bliss and graham priest 
number of kinds instantiated in the world. But even here, if we want to push such
an argument through, we require some additional assumptions. We would need to
assume, for example, that for the foundationalist there is only a finite number of kinds
of things that reside at the fundamental level. Alternatively, it might be the case, as has
been suggested by Tahko (this volume) that below a certain level for the infinitist there
are repetitions. It is possible, then, that in spite of not committing to a fundamental
level, the infinitist is still not committed to there being an infinite number of kinds in
the world.
Matters are more complex still when we consider virtues such as simplicity or
explanatory power. One might suppose that a reality with a hierarchical structure
and a fundamental level is simpler than one on which, say, everything depends on
everything else. But why this is the case is not altogether clear. It certainly seems
simpler, but that could just be because it is the picture in terms of which most of us
are accustomed to thinking. Arguably, a picture of reality on which everything is at
the same level is simpler than one that contains multiple levels.
Just as it is not clear which one of our theories wins the prize regarding explanatory
power. On the one hand, foundationalism looks to do well as the presence of a
fundamental level allows us to explain the existence of everything else. On the other
hand, anti-foundationalisms look to do better as there is nothing that is posited that
does not have an explanation. The balance could tip here, however, if it turns out
that where there is nothing fundamental there is something that is unexplained. As
we have seen in the discussion above, what anti-foundationalisms might leave us
without an explanation for is, say, why anything has being at all. But of course, this
is its whole own additional commitment that, as we have seen, brings with it its own
potential strife.
Much work remains to be done on the virtues of metaphysical foundationalism and
its alternatives. What we think the outcome of such work will be is that it is far from
clear that metaphysical foundationalism is obviously the most virtuous of the theories
available to us.
Of course there is much more to be said regarding the fundamentality thesis and
the kinds of arguments offered in defence of it. Bohn (this volume) argues that
we do not have good reasons to support the fundamentality thesis. Moreover, he
argues in addition to this that we even have good reasons to think that it is false
once we consider arguments involving gunk, junk, and hunk, and what he calls
the metaphysical principle of sufficient reason. Trogdon (this volume) suggests that
we can better understand Schaffer’s concern over the grounds of being in terms of
the notion of reality inheritance and that the argument so understood doesn’t work.
This is not to say, thinks Trogdon, that we, therewith, have no argument(s) in defence
of the fundamentality thesis, but that we need to understand fundamentality (as
motivated by the inheritance principle) as a kind of causal foundationalism or concrete
foundationalism. Jago (this volume), on the other hand, proposes an account of a
thing’s nature or essence that can allow us to provide grounding conditions for that
 the geography of fundamentality: an overview
thing. Essences, so understood, vindicate the hierarchy thesis as endorsed by the

proponent of the standard view, but allow that the relation may be non-well-founded.
It is somewhat surprising that the literature on metaphysical dependence and
different kinds of structuralisms are not brought more often into dialogue with one
another. Wigglesworth (this volume) argues that there are species of mathematical
structuralism that can plausibly be understood (i) to involve metaphysical depend-
ence relations and (ii) to challenge almost all of the structural features typically
attributed to those relations. In particular, he argues, there are species of structuralism
that involve both infinitely descending grounding chains and something fundamental.
Tahko (this volume) argues that standard accounts of fundamentality are generally
framed in terms of a kind of atomism. He argues that the fundamentality thesis, so
understood, has problems accounting for the picture of reality that emerges from
certain kinds of structuralisms. In place of this he proposes an account in terms
of ontological minimality which, interestingly, can accommodate both species of
fundamentality and infinitism. Morganti (this volume) undertakes a more general
investigation of alternative conceptions of physical reality. In particular, he defends the
idea that physics may well be able to be interpreted as supporting both infinitist and
coherentist structures, supporting a kind of pluralism about metaphysical structure.
3.3 The Contingency and Consistency Theses
We come now to a consideration of the contingency and consistency theses. We
discuss each in their turn. As Wildman (this volume) correctly points out, how
fundamentality intersects with modality is a spectacularly underexplored topic in
the current grounding literature. It is safe to assume, however, that the standard
view is one on which the fundamentalia are contingently existent. Of course, it is
not necessary for a foundationalist to believe that the fundamentalia are merely
contingently existent. Indeed, paradigmatic accounts of fundamentality have it that
the fundamentalia are necessary beings: consider God or Plato’s forms, for example.
The problem for such views, however, is how we are to preserve contingency in the
world, for where the fundamentalia are necessary beings, and beings that necessitate
the existence of everything else, there is only one way that the world can be, namely
exactly how it actually is.47 Not everyone agrees that this problem is as serious as it
sounds. Dasgupta, for example, has argued that a sufficiently constrained picture of
reality on which the fundamentalia are necessary existents (facts about essences in his
case) is plausible and appealing in certain ways.48
In his contribution to this volume, Wildman argues that there is a further issue
related to the contingency thesis that has, thus far, not been treated in the literature.
Whatever the modal status of the fundamentalia, is being fundamental itself a
47 See Skiles 2014 for a defence of the thought that grounding does not involve necessitation and
Trogdon 2013 for a defence of the thought that it does.

48 Dasgupta 2016.
ricki bliss and graham priest 
necessary or merely contingent property of the fundamentalia? Whilst a number

of combinations of views are possible—where the fundamentalia are, say, neces-
sary beings and necessarily fundamental—Wildman argues that several prominent
accounts appear to assume that the property of being fundamental is a merely contin-
gent property of the fundamentalia. Wildman aims to explore how one might go about
thinking about the intersection between modality and fundamentality, and argues that
the contingency of fundamentality is not as problematic as one might suppose.
Let us turn to the consistency thesis. Although one of us has developed and
defended vigorously both logics and metaphysical accounts according to which
contradictions are tolerable or even actual,49 there is no denying that the idea that
contradictions are insufferable is a stalwart in the Western tradition. As noted above,
some philosophers have been willing to question the first three of the foundationalists’
core commitments, but to the best of our knowledge, no one, to date, has challenged
the thought that whatever properties grounding structures have, they have them
consistently.
As we have seen, in the case of the first three of the foundationalist’s commitments,
philosophers have either offered, or it is at least possible to see, what the reasons might
be for defending or rejecting any of these commitments. In the case of the consistency
thesis, were a philosopher to defend their commitment, they would likely make appeal
to the host of arguments commonly levelled against inconsistencies already available
in the literature. We have no desire to rehearse or discuss the relevant arguments
here.50 In the final paper in this collection, Casati develops an account of the ‘second
Heidegger’ according to which we can understand him as espousing a kind of para-
foundationalism; where the grounding structure both is and is not anti-symmetric,
anti-reflexive, and extendable. One way of making sense of the later Heidegger, argues
Casati, is to bite the bullet, accept that he endorses contradictions, and with it, a view
according to which the grounding structure has inconsistent properties.
Let us return momentarily to the taxonomy presented in §2. Recall that there
were a number of lines—combinations of formal properties—that we dismissed as
impossible. We said, for example, that the dependence structure cannot be symmetric,
transitive, and anti-reflexive. Our taxonomy, along with every paper with one excep-
tion in this volume, has assumed the consistency thesis. In our taxonomy, we rule
out multiple views as impossible on the assumptions that for each of our four formal
properties a grounding structure either does or does not have (but not both) that
property. Our taxonomy assumes consistent axioms and rules out as impossible any
combination of views that, in spite of this, yields an inconsistency. But the kind of
view presented by Casati does not appear in our taxonomy for the reason that, unlike
us, the axiomatic system he proposes is itself inconsistent—a view so radical that it
does not appear on our taxonomy in order to be ruled out in the first place. Rather
49 50
See e.g. Priest 1987, 2006. Discussion can be found in Priest 2006.
 the geography of fundamentality: an overview
than return to the same old arguments in defence of consistency, and the arguments
against them, we prefer to say a few words about why one might go to the trouble of
challenging the consistency thesis to begin with.
The history of philosophy (East and West) is a history littered with accounts that
are plausibly construed as harbouring contradictions. This is not to suggest that the
history of philosophy is a history of dialethism, for, indeed, many of The Greats found
themselves deeply troubled by the appearance of contradictions in their systems.
What it is to suggest is that many interesting and important philosophical accounts
have invariably involved contradictions, and that one way of dealing with these
contradictions is just to accept them. Of course, contradictions can crop up all over
the place, but what we are particularly interested in are accounts of the structure of
reality—accounts that are plausibly construed as being couched in the language of
metaphysical dependence relations—that involve contradictions.
Consider the picture of reality espoused by twentieth-century Japanese thinker
Nishida.51 What emerges from his writings in influential texts such as his Basho is
the idea that to be an object just is to be enplaced—what it is for an object to be a
cat is to lie in the place ‘being a cat’. In the same way, a cat lies in the place ‘being a
mammal’, and a mammal lies in the place of ‘being an animal’, and so on and so forth.
This cannot go on forever, thinks Nishida, and there is the ultimate place—the place of
all places—which for Nishida is absolute nothingness (which also happens to be pure
consciousness). Importantly, if the place of all places is to do the work required of it, it
must not, itself, lie in a place; which is just to say it cannot be an object. However, this
is where the trouble begins. Indeed, as we have stated above, we know that, according
to Nishida, absolute nothingness does not lie in any place. But it turns out that what
this means is that absolute nothingness lies in at least one place, which is the place of
not lying in a place! So it turns out that for Nishida, the ultimate ground both is and
isn’t an object, which means it both is and isn’t fundamental.
Faced with this seeming contradiction at the bottom of his world, one might
suppose that Nishida was confused and that his system ultimately failed. There is
textual evidence to support the thought, though, that pure consciousness as a dialethia
was, in fact, exactly how Nishida intended it to be. Supposing that the enplacement
relation is a metaphysical dependence relation, we appear to have an historical
example of an inconsistent grounding theory.52
It is not simply that inconsistent grounding theories might be a useful tool for
engaging with certain historical figures. They may well have other interesting appli-
cations. Let us suppose that the membership and parthood relations are kinds of
metaphysical dependence relations. If this is the case, then it would seem that
inconsistent set theory, inconsistent mathematics, and inconsistent mereologies all
51 See Maraldo 2015.

52
Nishida’s view is closely related to the view of nothingness discussed in Priest 2014, ch. 13.
ricki bliss and graham priest 
entail, or at least have need of, inconsistent grounding theories. Of course, this will
not convince anyone who is not on board with these particular research programmes
in the first place, but the connections between the two, given a small number of very
plausible assumptions, are wide-ranging and interesting.
4 Taking the Alternatives Seriously

The taxonomy we presented in §2 makes certain matters clearer. We can now see, for
example, that there are, in fact, many more logically possible views about the structure
of reality than commonly supposed. But the taxonomy, and our classifications of
certain views, have their limitations. As is clear, our taxonomy rules out as impossible
very many combinations of properties of the dependence relations. Were we to employ
certain types of non-classical logics, however, these views might become worthy
of further consideration. We have also seen that our taxonomy fails to include the
type of para-grounding account Casati (this volume) believes to be attributable to
Heidegger.
Moreover, the way in which we have distinguished between foundationalism,
infinitism, and coherentism is overly simplistic. Had we tried to accommodate all
the possible ways in which species of these views could be, the taxonomy would
have become unwieldy and enormous. Much will turn on matters of definition, but
it certainly seems that mixed worlds might be possible. One can imagine a world in
which some dependence chains terminate in fundamentalia, where others do not;
what we call such a view will depend upon how foundationalism and infinitism
are defined. Understanding coherentism as a view according to which everything
depends upon everything else is particularly strong. It is possible to understand
coherentism as a kind of view, where various species of it may be possible. One might
think that views that allow ontological loops of any kind ought to be considered
weaker species of coherentism. If this is the case, it is worth pointing out that the
mere presence of loops does not entail a denial of the fundamentality thesis, as
defined. One can imagine a world in which there is something fundamental—even
a world in which every grounding chain terminates in the fundamental—but some
grounding chains contain loops. What we call such a view will depend upon how
foundationalism and coherentism are defined. Indeed, we can even imagine a world
in which some grounding chains are infinitely descending, some chains terminate
in something fundamental, and some chains contain loops of various sizes. Our
taxonomy, unfortunately, does not cater for such nuances.
The metaphor of the Great Chain of Being has wielded very significant influence—
both overt and covert—on the history of Western philosophy. It is about time to think
outside that particular box. We believe that the contents of this volume provide ample
evidence for this claim. Reality may well not have the metaphysical structure of a well-
founded chain, but a much more complex and fascinating one.
 the geography of fundamentality: an overview
References
Aikin, S.F. (2005), ‘Who is Afraid of Epistemology’s Regress Problem’, Philosophical Studies,
vol. 126, pp. 191–217.
Armstrong, D.M. (1997), A World of States of Affairs, Cambridge University Press.
Audi, P. (2013), ‘A Clarification and Defence of the Notion of Grounding’, in Metaphysical
Grounding: Understanding the Structure of Reality, Fabrice Correia and Benjamin Schnieder
(eds), Cambridge University Press.
Barnes, E. (2012), ‘Emergence and Fundamentality’, Mind, vol. 121, pp. 873–901.
Bennett, K. (2011), ‘By our Bootstraps’, Philosophical Papers, vol. 25, pp. 27–41.
Bliss, R.L. (2013), ‘Viciousness and the Structure of Reality’, Philosophical Studies, vol. 166,
pp. 399–418.
Bliss, R. (2014), ‘Viciousness and Circles of Ground’, Metaphilosophy, vol. 45, pp. 245–56.
Bliss, R. (forthcoming), ‘What Work the Fundamental?’, Erkenntnis.
Bliss, R. and Priest, G. (2017), ‘Metaphysical Dependence and Reality: East and West’, in
Buddhist Philosophy: A Comparative Survey, Steven Emmanuel (ed.). Basil Blackwell.
Bliss, R.L. and Trogdon, K. (2014), ‘Metaphysical Grounding’, in E. Zalta (ed.), Stanford
Encyclopedia of Philosophy, http://plato.stanford.edu/entries/grounding/.
Bohn, E.D. (2009), ‘Must There Be a Top Level?’, The Philosophical Quarterly, vol. 59, no. 235,
pp. 193–201.
Cameron, R. (2008), ‘Turtles All the Way Down’, Philosophical Quarterly, vol. 58, pp. 1–14.
Corkum, P. (2013), ‘Substance and Independence in Aristotle’, in Varieties of Dependence:
Ontological Dependence, Supervenience, and Response-Dependence, Benjamin Schnieder,
Alex Steinberg, and Miguel Hoeltje (eds), Basic Philosophical Concepts Series, Philosophia
Verlag, pp. 36–67.
Corkum, P. (2016), ‘Ontological Dependence and Grounding in Aristotle’, Oxford Handbooks
Online in Philosophy, Oxford University Press.
Dasgupta, S. (2016), ‘Metaphysical Rationalism’, Noûs, vol. 50, no. 2, pp. 379–418.
Dixon, T.S. (2016), ‘What is the Well-Foundedness of Grounding?’, Mind, vol. 125, pp. 439–68.
Fine, K. (1994), ‘Essence and Modality’, Philosophical Perspectives, vol. 8, pp. 1–16.
Fine, K. (1995), ‘Ontological Dependence’, Proceedings of the Aristotelian Society, vol. 95,
pp. 269–90.
Fine, K. (2001), ‘The Question of Realism’, Philosopher’s Imprint, vol. 1, pp. 1–30.
Fine, K. (2010), ‘Some Puzzles of Ground’, Notre Dame Journal of Formal Logic, vol. 51,
pp. 97–118.
Fine, K. (2013), ‘Guide to Ground’, in Metaphysical Grounding: Understanding the Structure of
Reality, Fabrice Correia and Benjamin Schnieder (eds), Cambridge University Press.
Jenkins, C.S. (2011), ‘Is Metaphysical Dependence Irreflexive?’, The Monist, vol. 94, no. 2,
pp. 267–76.
Keefe, R. (2002), ‘When Does Circularity Matter?’, Proceedings of the Aristotelian Society, New
Series, vol. 102, pp. 275–92.
Lovejoy, A. (1934), The Great Chain of Being, Harvard University Press.
Lowe, E.J. (2003), ‘Individuation’, in The Oxford Handbook of Metaphysics, Michael J. Loux and
Dean W. Zimmerman (eds), pp. 75–99.
Lowe, E.J. (2009), More Kinds of Being: A Further Study of Individuation, Identity and the Logic
of Sortal Terms, Wiley-Blackwell.
ricki bliss and graham priest 
Lowe, E.J. (2013), ‘Asymmetrical Dependence in Individuation’, in Metaphysical Grounding:

Understanding the Structure of Reality, Fabrice Correia and Benjamin Schnieder (eds),
Cambridge University Press.
Maitzen, S. (2012), ‘Stop Asking Why there’s Anything’, Erkenntnis, vol. 77, pp. 51–63.
Maitzen, S. (2013), ‘Questioning the Question’, in The Puzzle of Existence: Why is there
Something Rather than Nothing?, Tyron Goldschmidt (ed.), Routledge.
Maraldo, J.C. (2015), ‘Nishida Kitarō’, The Stanford Encyclopedia of Philosophy, Edward N. Zalta
(ed.), http://plato.stanford.edu/entries/nishida-kitaro/.
Passmore, J. (1970), Philosophical Reasoning, Duckworth.
Priest, G. (1987), In Contradiction, Martinus Nijhoff. Second edition, Oxford University Press
(2006).
Priest, G. (2006), ‘Doubt Truth to be a Liar’, Oxford University Press.
Priest, G. (2014), One, Oxford University Press.
Rabin, G.O. and Rabern, B. (2016), ‘Well Founding Grounding Grounding’, Journal of Philo-
sophical Logic, vol. 45, no. 4, pp. 349–79.
Raven, M. (2013), ‘Is Ground a Strict Partial Order?’, American Philosophical Quarterly, vol. 50,
pp. 191–9.
Raven, M. (2018), ‘Fundamentality Without Foundations’, Philosophy and Phenomenological
Research, vol. 93, no. 3, pp. 607–26.
Schaffer, J. (2009), ‘On What Grounds What’, in Metametaphysics: New Essays on the Founda-
tions of Ontology, David Manley, David J. Chalmers, and Ryan Wasserman (eds), Oxford
University Press, pp. 347–83.
Schaffer, J. (2010), ‘Monism: The Priority of the Whole’, Philosophical Review, vol. 119,
pp. 31–76.
Schaffer, J. (2012), ‘Grounding, Transitivity and Contrastivity’, in Metaphysical Grounding:
Understanding the Structure of Reality, Fabrice Correia and Benjamin Schnieder (eds),
Cambridge University Press, pp. 122–38.
Schaffer, J. (2016), ‘Grounding in the Image of Causation’, Philosophical Studies, vol. 173,
pp. 49–100.
Sider, T. (2011), Writing the Book of the World, Oxford University Press.
Skiles, A. (2014), ‘Against Grounding Necessitarianism’, Erkenntnis, vol. 80, pp. 717–51.
Thomasson, A. (2007), Ordinary Objects, Oxford University Press.
Trogdon, K. (2013), ‘Grounding: Necessary or Contingent?’, Pacific Philosophical Quarterly,
vol. 94, pp. 465–85.
Tahko, T.E. and Lowe, E.J. (2015), ‘Ontological Dependence’, The Stanford Encyclopedia of Phil-
osophy, Edward N. Zalta (ed.), http://plato.stanford.edu/entries/dependence-ontological/.
Wilson, J. (2014), ‘No Work for a Theory of Ground’, Inquiry, vol. 57, pp. 535–79.
PA R T I
The Hierarchy Thesis
1
Grounding Orthodoxy and the
Layered Conception
Gabriel Oak Rabin
1 Introduction
Our world contains a shocking variety of stuff, from the very large (planets, quasars,
galaxies) to the very small (quarks, leptons, bosons), with lots in between (koalas,
canyons, coins). Here’s a common thought: All this stuff can be organized into a
hierarchy of levels. The galaxies and quasars are “on top”, the canyons and koalas lie
in the middle, below them come molecular compounds, and at the very bottom are
the tiny particles and other phenomena (nuclear forces, electromagnetism) discussed
in fundamental physics. The idea of “levels” in the special sciences reflects this
hierarchical conception of the world. In the layering of special sciences, physics
occupies the bottom, with chemistry, then biology, then psychology, then economics,
lying on top.
What makes one phenomenon “higher” than another? One answer is that a relation
of dependence creates the hierarchical structure. Psychology depends on biology,
which depends on chemistry, which depends on physics. Of course, it’s not the
sciences themselves that depend on each other (psychology predates chemistry), but
rather the phenomena the sciences study. Which psychological states I have depends
on which biological states I have, but not vice versa. Which biological states I have
depends on which chemical states I have, but not vice versa. Et cetera. Let’s use the
phrase ‘the layered conception of reality’ (‘the layered conception’ for short) as a label
for the general idea that reality is layered in a hierarchy structured by relations of
dependence. We can add a claim about fundamentality to the layered conception: the
lower tiers of the layering are more fundamental than the higher tiers. I will make this
further assumption in what follows.
Much philosophical ink has been recently spilled inquiring into the nature of
ground. Ground is alleged to be a/the relation of metaphysical dependence, expla-
nation, and/or priority. It is that relation the physicalist alleges to hold between the
mental and the physical, that the utilitarian claims holds between moral facts and
the facts about pleasure and pain, and that many claim to hold between the fact that
P and the fact that P or Q. In each case, the ground makes the grounded obtain.
 grounding orthodoxy and the layered conception
The grounded metaphysically depends on is metaphysically explained by, and/or is

ontologically posterior to, the ground. Ground should be distinguished from causal
dependence. Ground often (and perhaps always) holds synchronically, between two
relata at the same time. For example, the physicalist claims that my current pain
is grounded in my current brain state. In contrast, causal dependence relates items
across time. The dualist can admit that my past brain state caused my current pain,
while denying that pain is grounded in the brain.1
Once we have a notion of ground on board, it seems natural to slot that notion into
the layered conception. After all, relations of dependence generate the layering, and
ground is metaphysical dependence. Voila! Let’s plug in everything we’ve learned in
all the literature on ground to generate the layered picture of the world. Theorists of
ground have had exactly this idea (deRosset [2013]). In fact, much of the appeal of
the notion of ground, and its recent rise to prominency in metaphysics, comes from
the intuitive appeal of the layered conception. Using ground to generate a hierarchy
of dependence, and thereby vindicate the layered conception, is a nice thought, but it
faces serious obstacles.
Only a relation with certain formal features is capable of delivering the layered
conception of the world. For example, a layered hierarchy generated by a relation
that loops will contain X above Y, above Z, but X will appear again down below Z!
Loops aren’t amenable to creating the type of structure characteristic of the layered
conception. Thankfully, the orthodox views on ground hold that ground has several
features that ensure that ground will be able to provide the structure characteristic
of the layered conception. Let’s label the conjunction of the following four theses
‘the orthodoxy’. (All of these claims should be interpreted as preceded by universal
quantifiers ∀X, ∀Y, ∀Z.):
(TS) Transitivity: If X grounds Y and Y grounds Z, then X grounds Z.
(AS) Antisymmetry: If X grounds Y, then Y does not ground X.
(IR) Irreflexivity: X does not ground X.
(FD) Foundationalism: Everything is ultimately grounded in a bottom layer with
no further ground.2
A relation that is transitive and antisymmetric cannot contain loops. This takes care
of the worry that ground might generate loops, and thereby be unable to vindicate
the layered conception. Or does it? The problem here is that every component of
the orthodoxy has been challenged. Schaffer [2012] denies transitivity. Barnes [2018]
denies antisymmetry. Jenkins [2011] questions irreflexivity. Bliss [2014] even argues
that ground might generate loops!
1 I leave open the possibility that causation might, in the end, turn out to be a form of ground. Or vice
versa. But prima facie, they look different, despite sharing some similarities.
2 This constraint sometimes goes under the banner that ground must be “well-founded”
(Schaffer [2010]: 37). This is an unfortunate choice of terminology: a relation of ground that is not
well-founded in the set-theoretic sense can still have a foundation. For clarification of these issues and of
what “well-founded” amounts to when it comes to ground, cf. Rabin & Rabern [2015].
gabriel oak rabin 
For the most part, theorists have either ignored the alleged counterexamples
and continued to insist on the orthodoxy, or fought against the counterexamples
outright (e.g. Litland [2013]). A major reason for maintaining the orthodoxy in the
face of alleged counterexamples is the worry that without the formal features the
orthodoxy provides, ground will prove unable to vindicate the layered conception.
In the rest of this paper, my goal will be to alleviate this worry. I will argue that, even
without any of the formal features listed above—transitivity, asymmetry, irreflexivity,
or foundationalism—ground can still provide the dependence structure the layered
conception requires. In fact, I will argue that relaxing the assumptions in the ortho-
doxy actually makes ground better able to generate the structure characteristic of the
layered conception.
Here’s a roadmap for the remainder of the paper. In the next section (2: “Ground
as the Generator as Layers”), we put some flesh on the bones of the idea of the layered
conception and how ground interacts with it. Each of Sections 3–6 explores how
ground fares in its ability to vindicate the layered conception under the relaxation
of some element of the orthodoxy. We consider abandoning foundationalism, anti-
symmetry, irreflexivity, and transitivity (in that order). The conclusory Section 7 steps
back to consider the resulting overall picture.
2 Ground as the Generator of Layers

The layered conception is admittedly vague. In this section, we examine ways to put
flesh on the bones of the bare idea and how we might utilize ground to elucidate the
structure the layered conception mandates.
The layered conception, at first pass, looks something like this:
economics
psychology
biology
chemistry
physics
As I mentioned before, the claim is not that the sciences themselves, considered
as fields of inquiry, depend on each other. Economists can and should go about their
business without asking chemists for instructions. Instead, the phenomena studied by
one field of inquiry depend on, and are determined by, phenomena studied by another
 grounding orthodoxy and the layered conception
field of inquiry. But that is not quite right. Biology depends on chemistry, but the
camouflage in Arabian cuttlefish (a biological process) has absolutely nothing to do
with the oxidization of steel beams (a chemical process) in a shipyard in New Orleans.
Most particular concrete biological happenings have nothing to do with, and certainly
don’t depend on, most particular concrete chemical happenings. The same is true at
the level of types. It’s likely that the biological phenomenon of cuttlefish camouflage
has nothing to with the chemical process of oxidization. (The marine biologists could
prove me wrong here, but I feel like I’m on safe ground.)
However, the camouflage patterns of a particular cuttlefish do depend on some
chemical facts about that particular cuttlefish. And the camouflage of a different
cuttlefish depends on chemical facts about that cuttlefish. Furthermore, the two
instances of cuttlefish camouflage might depend on the very same type (not token)
of chemical property—call it ‘C’. If the pattern is widespread, then we might claim
a dependence of cuttlefish camouflage on chemical property C. This yields a lesson.
We infer dependencies between types of properties from patterns in dependencies of
particular tokens of those properties.
We now come to ground. Ground is typically understood as a dependence relation
between particular facts, states of affairs, particulars, or properties. The mass of this
table is grounded in the mass of these four legs and this tabletop. Ground gives us the
particular instances of dependency. From these particular token-dependencies we can
infer the type-dependencies characteristic of the layered conception.
Sometimes, the type-dependencies are specific, such as when the firing of neurons
is grounded in an electrical imbalance between positively charged potassium ions and
negatively charged sodium. But these cases are rare. More often, the dependency is not
specific, and a higher-level type, such as cuttlefish camouflage, does not depend on
only one lower-level type, such as potassium/sodium interaction. In each particular
case of cuttlefish camouflage, there is some chemical processes underlying it. But
it needn’t be the same type of chemical process in each case. These points are
familiar from research on multiple realizability. Most phenomena are realizable, or
groundable, in a wide variety of underlying lower-level phenomena. The various
lower-level phenomena that all give rise to a single type of higher-level phenomenon
might have little in common, other than the fact that they ground, or give rise to,
the same type of higher-level happenings. Of course, these lower-level happenings,
despite their dissimilarities, remain, for example, chemical. So at the very least, we can
say that cuttlefish camouflage, even if it does not depend on any particular chemical
type, depends on “chemistry”, or “the chemical level”.
Call a complete story of the world’s grounding relations between particular facts
a grounding graph (so called because it can be represented by a graph in the mathe-
matical sense: a set of nodes with directed relations between them). The grounding
graph gives us both more and less than we want from the layered conception. It gives
us more because it gives us thousands of cuttlefish camouflage dependences—one for
each cuttlefish. That’s more than we need. But the grounding graph also gives us less.
gabriel oak rabin 
psychology geology
chemistry
Figure 1.1. Option (i): geology and psychology on the same level, equally fundamental.
psychology geology
chemistry
Figure 1.2. Option (ii): geology and psychology incommensurable, neither more, nor less, nor
equally fundamental.
The layered conception says that biology is above chemistry. This entails that cuttlefish
x’s biological camouflage is above cuttlefish y’s chemical properties. But grounding
relations don’t deliver this verdict. There are no grounding relations between the two.
In mathematical terms, the layering conception seems to demand a total order, in
which every pair of items is related by either the “higher than”, “lower than”, or “at
the same level as” relation. In contrast, ground is a (very) partial order. A randomly
chosen pair of items is unlikely to be related by ground at all. There’s no easy recipe
for generating a total order from a partial order.
However, there are reasons to be optimistic that the ordering characteristic of
the layered conception can be gleaned from the grounding graph. First, as dis-
cussed above, we can look for patterns in the particular grounding claims. There
are many such patterns. Sometimes the patterns are specific (neural firing depends
on potassium–sodium ion imbalance). Other times they are not (each instance of
cuttlefish camouflage depends on some chemical property). But the patterns are there.
If they weren’t, the layered conception wouldn’t be so appealing in the first place.
Second, we may not want the layered conception to deliver a total ordering. Both
geology and psychology are above chemistry. Neither lies above the other. Two options
remain: (i) they are at the same level or (ii) they are incommensurable.
If the layered conception demands a total ordering, then (i) is the only option.
A total ordering does not permit cases in which two items are incommensurable.
However, I think that option (ii) is preferable, and that we should give up the idea
that the layered conception requires a total ordering. Here’s why. It remains open
to discover some other range of phenomena, below psychology, but which contains
no grounding relations to geology. Computation provides a potential example. If
all psychological phenomena are ultimately grounded in computational phenomena
(a not implausible hypothesis), then psychology will lie above computation. Suppose
 grounding orthodoxy and the layered conception
we choose option (i), placing geology on the same level with psychology. Ground tells
us to place computation below psychology, which we’ve placed on the same level as
geology. We’re now forced to put computation below geology. This seems wrong. The
relation between computational phenomena and geological phenomena is exactly the
same as the relation between psychological phenomena and geological phenomena:
nil. Whatever the reasons in favor of placing psychology and geology on the same
level were, exactly the same reasons apply to placing computation and geology on the
same level. It would be arbitrary to place geology and psychology on the same level
with computation below, rather than, say, geology and computation on the same level,
with psychology above.
The desire to place neither geology nor psychology above the other can be satisfied
without placing them at the same level in the layered conception. Instead, we should
give up the idea that the layered conception mandates a total ordering. Once we do
so, ground, with its very partial order, looks better as a guide to reality’s layers (as
conceived by the layered conception). Admittedly, the layered conception demands
an ordering that is closer to total than the ordering provided by ground. But patterns
among ground’s partial ordering can bridge the gap between ground’s very partial
order and the layered conception’s less partial order.
3 Foundationalism and the Layered Conception

Foundationalism is the easiest bit of the orthodoxy for the fan of the layered con-
ception to reject. Simply put, the layered conception does not require a foundation.
The Greek philosopher Xenophanes was an early proponent of the layered conception
(Patzia [n.d.], Lesher [1992]). Arguably, he also believed foundationalism to be false,
and that the world consisted of alternating layers of earth and water.3
Of course, one could build foundationalism into the layered conception, forming
the-layered-conception-with-a-bottom. In so doing, one would make the layered
conception developed via ground incompatible with rejection of foundationalism
about ground. But one certainly need not insist on a bottom layer. The basic idea of a
reality structured by relations of dependence does not require a foundation.
4 Anti-Symmetry and the Layered Conception

The basic idea of using ground to generate the layered conception comes from the
following principle:
(The Simple Principle) If x grounds y, then x is at a lower level / more fundamental than y.
3 Xenophanes believed in an infinite temporal descent of watery and earthy stages (Hippolytus of
Rome [2015] attributes this view to Xenophanes in his Refutation of All Heresies: 1.14). Whether this entails
anti-foundationalism of ground will turn on whether temporal, or causal, dependence can be parlayed into
metaphysical dependence.
gabriel oak rabin 
The simple principle gets us from claims about grounding relations between
particulars (facts, objects, or properties) to claims about where those particulars fit
into reality’s layers. To generate the full layered conception, we still need to discern
patterns concerning where certain types of things occur in reality’s layers. But, via the
simple principle, ground gives us a good start.
The simple principle does not work so well, however, if ground fails to be anti-
symmetric. According to the simple principle, if x grounds y, then x is lower than y. If
y grounds x (violating anti-symmetry), then y is lower than x. And that doesn’t make
sense, at least in so far as I understand the layered conception. Biology can’t be both
above and below chemistry.
There are decent prima facie considerations in favor of rejecting the anti-symmetry
of ground. Barnes [2018] argues that we should accept symmetric dependence in a
wide variety of cases, from immanent universals to states of affairs to mathematical
ontology. In one example, she argues that it is essential to the evacuation at Dunkirk
that it is part of World War II. And it is essential to World War II that it contain the
evacuation at Dunkirk. If this is correct, it is plausible to maintain that each of World
War II and the evacuation at Dunkirk depend on the other. Voila: symmetric ground!
This is neither the time nor the place to have the fight over whether ground is or
is not anti-symmetric. Barnes presents some plausible cases. At the least, proponents
of the theory-combinations Barnes discusses might want to take advantage of a non
anti-symmetric (i.e. sometimes symmetric) notion of ground. For their sake, it’s
worth exploring how rejecting the orthodoxy regarding the anti-symmetry of ground
interacts with the layered conception.
I believe that, ultimately, rejection of anti-symmetry for ground does not impugn
ground’s ability to vindicate the layered conception. In fact, cases of symmetric ground
might help us better understand how reality is layered. I argue for these claims in the
remainder of this section.
The simple principle, above, is one way to infer layering from relations of ground.
But once we recognize the possibility of symmetric ground, we can opt for the
following slightly less simple principle.
(The Slightly Less Simple Principle) If x grounds y, and y does not ground x, then x is more
fundamental/at a lower level than y.
The Slightly Less Simple Principle is a clear improvement over the Simple Principle.
If ground is anti-symmetric, then the ‘y does not ground x’ clause in the Less Simple
Principle is vacuous, and the Less Simple Principle reduces to the Simple Principle.
But if symmetric ground does occur, the Slightly Less Simple Principle avoids the
problematic result above, where x is both above and below y in reality’s layering.
In cases of symmetric ground, what should we say about the layering relations of
the items that ground each other? We should not place either above the other. This
leaves two options, which we’ve already seen: (i) they are at the same level or (ii) they
are incommensurable. I believe that (i) is the better option here. x and y are related
 grounding orthodoxy and the layered conception
by ground. It seems odd to say that they bear no relation to each other in reality’s
layering. The layering is still a layering based on dependence. And x and y depend on
each other. I propose we place x and y on the same level.
Considerations involving the transitivity of ground further support placing x and y
on the same level. The transitivity of ground will guarantee that, in cases of symmetric
ground, the symmetric grounders will be at the pseudo-same level. For any x and y, x
and y are at the pseudo-same level in reality’s layering if and only if for any z, if z is
above x, then z is above y, and if z is below x, then z is below y. In simple terms, two
items at the pseudo-same level are both above, and below, all the same stuff. This does
not quite guarantee sameness of level. x and y might still be incommensurable.
It is worth noting that this case is slightly different than the geology–biology
case discussed earlier, in which I argued for incommensurability of level. In that
case, computation lay below biology, but remained incommensurable with geology.
This would not be possible if geology and biology were incommensurable but at
the pseudo-same level. Their pseudo-sameness would guarantee that if biology were
higher than computation, geology would be too.
I admit that my arguments leave some space for claiming that symmetric grounders
are incommensurable in level. But given that (a) they are related by dependence and
(b) they are at the pseudo-same level, I believe we should say that they lie at the same
level in reality’s layering.
5 Irreflexivity and the Layered Conception

Grounding orthodoxy holds that ground is irreflexive: nothing grounds itself.
Jenkins [2011] has challenged the orthodoxy, claiming that it’s better to leave open
the possibility that something could ground itself. For example, an identity theorist
in philosophy of mind might simultaneously claim that (a) consciousness is identical
to electrical flow in the brain’s dorsal stream and (b) mental phenomena, including
consciousness, are grounded in brain phenomena, such as electrical flow in the dorsal
stream. If the orthodoxy is correct, this position is incoherent: ground is irreflexive.
Consciousness can’t be grounded in electrical flow in the dorsal stream, to which it is
identical.
Understandably, Jenkins argues that our conception of ground should not rule out
by fiat the combination of metaphysical views espoused by the envisioned identity
theorist. We want ground to provide a useful philosophical tool for conceptualizing
various debates in metaphysics. In so far as an irreflexive conception of ground
makes unintelligible a plausible and popular view in the philosophy mind, it fails to
accomplish this goal. The best solution, argues Jenkins, is to give up the irreflexivity of
ground. The result is not that ground is reflexive (i.e. everything grounds itself), but
that sometimes, things do ground themselves.
There are ways to resist this line of thought. But acceptance, in certain cases, of
reflexive ground, seems desirable, particularly so in light of specific philosophical
gabriel oak rabin 
views, like the identity theory in philosophy of mind. How much does giving up the
irreflexivity of ground affect ground’s ability to vindicate the layered conception? The
answer is, “Not much.”
In combination with the simple principle, cases of irreflexive ground cause
problems. Continuing with the identity claim as our example, the two will entail
that conscious experience is above itself (and below itself) in the layered conception.
That’s weird. Like with symmetric ground, a shift from the simple principle to
the slightly less simple principle saves the day. The slightly less simple principle
avoids the result that conscious experience is above (and below) itself in reality’s
layering.
The choice between an irreflexive conception of ground and a reflexive conception
might be partly terminological. In the semantics of Fine [2012], weak ground, in
which everything grounds itself, is taken as the primitive notion. Fine does this partly
for reasons of simplicity. But we might think that formal simplicity provides some
reason for taking the reflexive conception of ground to be more fundamental, even
if talk of an entity’s grounding itself rubs against thought of ground as a form of
metaphysical explanation and/or determination.
One important difference between giving up irreflexive ground and giving up anti-
symmetric ground is worth noting. If ground is reflexive, that is, if everything grounds
itself, there is no serious challenge to the layered conception. We need simply shift
from the unreflective simple principle to the slightly less simple principle. Such a
move will avoid the unsavory implications of cases of reflexive ground (e.g. that
conscious experience is both above and below itself), but still allow ground to play
its intended role in generating the remainder of reality’s layering. On the other hand,
if ground is symmetric, that is, if every time x grounds y, y grounds x, the goal of using
ground to generate reality’s layering falls into serious jeopardy. There’s no simple fix for
symmetric ground. (Thankfully, to my knowledge no one has suggested that ground
is symmetric.)
The basic thought behind using ground to generate the layered conception is that if
x grounds y, x is lower than y in the layered conception. In the first instance, ground
relates tokens, or particular facts. The layered conception relates types (as well as
tokens of those types). Some theorizing is required to get from the tokens to the
types. Adoption of a reflexive conception of ground, in which everything grounds
itself, requires only minimal modification of the basic idea. A shift from the basic
idea, expressed in the simple principle, to a more nuanced version of the same idea
via the slightly less simple principle, does the trick and rescues a reflexive conception
of ground’s ability to generate reality’s layering.
In contrast, symmetric ground, in which every time x grounds y, y also grounds x,
completely voids the basic idea. Ground will never give us the result that x is above (or
below) y in reality’s layering. In Section 4, I argued that in cases of symmetric ground
we should maintain that the symmetric groundees should be placed at the same level
in reality’s layering. If this is correct, then ground will provide some, but not much,
 grounding orthodoxy and the layered conception
guide to reality’s layers. Ground will be sufficient for sameness of level. But some other
relation will be required to do the heavy lifting in the generation of reality’s vertical
hierarchy.
In sum, I claim that neither acceptance of particular cases of reflexive ground nor
acceptance of a fully reflexive conception of ground seriously challenges the ability of
ground to vindicate the layered conception. Particular individual cases of symmetric
ground can be easily handled. But a full-blown symmetric conception of ground will
void ground’s ability to provide reality’s layering.
6 Intransitivity and the Layered Conception

Lastly, we come to the transitivity of ground, which says that if x grounds y, and
y grounds z, then x grounds z. Schaffer [2012] has challenged this principle. One
of his arguments revolves around a dented sphere. Schaffer claims that while it’s
plausible that (a) the fact that the dented sphere has a dent grounds that fact that
it has determinate shape S and (b) the fact that the dented sphere has determinate
shape S grounds the fact that it is more-or-less spherical, it is implausible that (c) the
fact that the dented sphere has a dent grounds the fact that it is more-or-less spherical.
After all, writes Schaffer, “the thing is more-or-less spherical despite the minor dent,
not because of it” (127).
There are ways to resist the argument, but I do not wish to weigh in on the
issue here. Schaffer’s example is prima facie plausible, and he provides other alleged
counterexamples to transitivity. At the least, some will want to deny the transitivity
of ground. For their sake, it’s worth exploring how such a denial will affect ground’s
interaction with the layered conception.
The layered conception’s hierarchical structure is transitive. If biological phenom-
ena lie above chemical phenomena, and psychological phenomena lie above the
biological, then psychological phenomena lie above chemical phenomena. We need
some transitivity in the layered conception.
Consider a graphical representation of the world’s grounding relations, in which
nodes represent the relata of grounding relations and arrows between nodes represent
relations of ground. (Arrows point from the ground to the grounded.) From the
graph, we can observe the beginnings of reality’s layering. The fact that my brain
contains proton p lies below the fact that my brain contains potassium molecule
m, which lies below the fact that my brain contains neuron n. This layering of
particular facts proceeds from the physical to the chemical to the biological. The
generation of this layering does not require an arrow, or a relation of ground,
between proton p and neuron n or their associated facts. A failure of transitivity,
say, between the proton and the neuron, will not interfere with the generation of this
layering.
From a formal standpoint, this should be no surprise. For any non-transitive
relation R one can always take the transitive closure of R to generate a transitive
gabriel oak rabin 
neuron n
molecule m
proton p electron e
relation R* that will contain R as a subset, in the sense that if Rab, then R*ab. Even if
ground is not transitive, we can take ground’s transitive closure to generate ground*.
But we need not resort to such formal tricks.
The layered conception involves a layering of fundamentality. The chemical is
more fundamental than the biological. “More fundamental than” is transitive, as
are “higher than” and “lower than” in reality’s layering. Ground and fundamentality
are linked by the simple and/or slightly less principle we’ve discussed. Grounding
relations have implications for relations of relative fundamentality and for reality’s
layering. But ground can fail to be transitive, and even be anti-transitive, yet still
have these implications for the transitive relations for “more/less fundamental than”
and “lower/higher in reality’s layering than.” Assuming this transitivity, a double
application of the simple (and/or slightly less simple) principle yields the result that if
x grounds y, and y grounds z, then x is more fundamental than z, and x is lower than
z in reality’s layering. This is so even if x does not ground z.
One final worry goes as follows. If ground is not transitive, but the hierarchical
structure of the layered conception is, what is the layered conception a hierarchy
of? The preceding discussion should alleviate the worry. The layered conception’s
hierarchical structure captures relations of relative fundamentality, which have an
intimate relation to ground, despite the fact that they remain transitive even when
ground is not.
We can get the transitive structure constitutive of the layered conception even if
ground fails to be transitive. The transitivity can come in later, with the relations
(“more/less fundamental than,” “lower/higher than”) that properly constitute reality’s
layering, and to which ground is a guide.
 grounding orthodoxy and the layered conception
7 Conclusion
The key to making unorthodox views about the formal properties of ground com-
patible with the layered conception is to recognize that there is a gap between what
grounds what and the layered conception. One can’t just “read off ” reality’s layering
from the facts about ground. The move from what grounds what to reality’s layering
is substantive. I believe we should be optimistic about gleaning from the facts about
ground a useful and informative structure that roughly matches our pre-theoretic
conception of how the features of reality are layered.
First, principles linking ground and layering, or fundamentality, such as the simple
and/or slightly less simple principle, give us a healthy start in generating a layering
from ground. But the task of evaluating the patterns in the grounding relations
between particulars, and gleaning from those patterns a layering of the various
properties, and types of properties (geological, biological), remains. Second, we may
have to abandon some of our pre-theoretic ideas about reality’s layering. I argued that
we should abandon the claim that reality’s layering generates a total order. Geology
and biology are incommensurable; neither lies above or below the other. The layering’s
order is closer to total than ground’s order. But both are partial.
The gap between ground and layering both helps and harms. It harms because it
makes the task of discerning reality’s layering more difficult. Even after we possess
a complete story of what grounds what, we must still do philosophical work to
determine what is more fundamental than what. It helps because it permits the
layering to be well-behaved even when ground is not. For example, symmetric cases
of ground don’t force us to claim that the symmetric groundees each lie above (or
below) the other in reality’s layering.
The grounding orthodoxy ensures that ground behaves nicely. It will be a good
little transitive, anti-symmetric, irreflexive, foundationalist relation. This obedient
behavior ensures the absence of problematic grounding structures, such as loops, that
create problems when we move from ground to reality’s layering. But the heretics are
out there. Not all theorists of ground believe in the orthodoxy. I’ve covered a variety
of reasons to doubt various parts of that orthodoxy. These theorists will probably be
willing to give up some nice behavior in order to have a theoretical tool that can do
the metaphysical work they want done. For this reason alone, it’s worth exploring how
reality’s layering might go if we accept an unorthodox view about ground and want to
maintain an intimate link between ground and the layered conception.
There are good reasons for the orthodoxy. The principles seem prima facie correct.
It’s convenient to have a formally well-behaved relation. But there are good reasons
to doubt the orthodoxy. Cases like Jenkins’ reflexive dependence of pains on brains,
or Barnes’ symmetric dependence of World War II and the evacuation at Dunkirk,
should force us to seriously reconsider. There is something to the idea of mutual
dependence in those cases. This dependence should at least be taken into consid-
eration when we move to generate reality’s layering. A non-orthodox conception of
gabriel oak rabin 
ground will better reflect whatever it was about dependence that Barnes and Jenkins
latched on to, and which we want reflected in the reality’s layering. Even staunchly
orthodox views, when they move from the grounding graph to reality’s layering, might
decide to reflect that symmetric relation in reality’s layering, even if they do not choose
to call it ‘ground.’ In this way a non-orthodox conception of ground better reflects
reality’s relations of dependence, and enables the generation of a more, rather than
less, accurate, picture of reality’s layering.
In the end, we might reject the arguments of Barnes, Bliss, Jenkins, and Schaffer,
and maintain that the orthodoxy about ground is correct. But knowing that the
layered conception is perfectly compatible with the heretical views that challenge the
orthodoxy should grease the wheels for rejecting that orthodoxy (a move with which
I have considerable sympathy). A non-orthodox view of ground can not only have a
nice layering of reality, but the non-orthodox view is, in various ways, better suited to
that layering. The grounding heretics can have their (layered) cake and eat it too.4
References
Barnes, Elizabeth. Symmetric Dependence. 2018. In Bliss, Ricki and Priest, Graham (eds),
Reality and its Structure: Essays in Fundamentality. Oxford: Oxford University Press.
Bliss, Ricki. 2014. Viciousnesss and Circles of Ground. Metaphilosophy, 45, 245–56.
deRosset, Louis. 2013. Grounding Explanations. Philosophers’ Imprint, 13(7), 1–26.
Fine, Kit. 2012. A Guide to Ground. Pages 37–80 in Correia, Fabrice and Schnieder, Benjamin
(eds), Metaphysical Grounding: Understanding the Structure of Reality. Cambridge: Cam-
bridge University Press.
Hippolytus of Rome. 2015. Refutation of All Heresies. CreateSpace Independent Publishing
Platform.
Jenkins, Carrie. 2011. Is Metaphysical Dependence Irreflexive? The Monist, 94, 267–76.
Lesher, J.H. 1992. Xenophanes of Colophon: Fragments: A Text and Translation with Commen-
tary. University of Toronto Press.
Litland, Jon. 2013. On Some Counterexamples to the Transitivity of Grounding. Philosophical
Essays, 14, 19–32.
Patzia, Michael. Xenophanes. The Internet Encyclopedia of Philosophy, http://www.iep.utm.
edu/xenoph/.
Rabin, Gabriel and Rabern, Brian. 2016. Well Founding Grounding Grounding. Journal of
Philosophical Logic, 45(4), 349–79.
Schaffer, Jonathan. 2010. Monism: The Priority of the Whole. The Philosophical Review,
119(1), 31.
Schaffer, Jonathan. 2012. Grounding, Transitivity, and Contrastivity. Pages 122–38 in Correia,
Fabrice and Schnieder, Benjamin (eds), Metaphysical Grounding: Understanding the Structure
of Reality. Cambridge: Cambridge University Press.
4 Thanks are due to two anonymous referees for extremely helpful comments and suggestions, and to
Ricki Bliss and Graham Priest for inviting me to think about these topics and contribute to this volume.
2
Symmetric Dependence
Elizabeth Barnes
Metaphysical orthodoxy maintains that the relation of ontological dependence is

irreflexive, asymmetric, and transitive. The goal of this paper is to challenge that
orthodoxy by arguing that ontological dependence should be understood as non-
symmetric, rather than asymmetric. If we give up the asymmetry of dependence,
interesting things follow for what we can say about metaphysical explanation—
particularly for the prospects of explanatory holism.
1 Background: Ontological Dependence

The term ‘dependence’ is employed in different ways across different sub-literatures.
So I first need to be clear about what I mean by ‘dependence’, and what specific
literature I’m focusing on. To begin with, I’m concerned with ontological dependence.
There are no doubt other forms of dependence—causal, conceptual, logical, and so
on—but such relations aren’t my target here.
What is ontological dependence? That’s a vexed question. Moreover, it’s not a
question I’m going to attempt to answer in full here—not the least because many
contemporary metaphysicians take it to be primitive. Rather, I’m going to highlight
some key features of the relation, which will hopefully be enough for my purposes.
1.1 Paradigm cases
Talk of ontological dependence is typically introduced via paradigm cases or exam-
ples. The whole ontologically depends on its parts. The mental ontologically depends
on the physical. Secondary qualities ontologically depend on primary qualities.
Esthetic ontology depends on non—esthetic ontology. And so on.
One thing to note about these paradigm cases is that—fitting with the orthodoxy—
dependence holds asymmetrically in each of them. The whole depends on the parts,
but the parts don’t depend on the whole.1 The mental depends on the physical, but the
1 Or, at least, there is a dependence relation between part and whole. Most people think wholes depend
on parts, but not everyone does—see especially Schaffer (2010b).
elizabeth barnes 
physical doesn’t depend on the mental. And so on. From this, it is sometimes reasoned
that we have justification for thinking that the relation of dependence is asymmetric.
For example, Kathrin Koslicki remarks, after introducing a list of paradigm cases of
dependence, that if in fact ‘[these cases] do constitute examples of pairs of entities
related by an ontological dependence relation of some sort, the dependence relation
in question may plausibly be taken to be asymmetric.’2 Yet it’s a mistake to reason as
follows: ‘Paradigm cases of F are , therefore all cases of F are .’ All the paradigm
cases of redness are determinately red. But you can’t conclude from that that all cases
of redness are determinately red.
1.2 Hyperintensionality
So what do these paradigm cases of dependence—the mental on the physical, a whole
on its parts, and so on—have in common with one another? What is the relation
of dependence? It’s been, in recent times, very common to try to appeal to modal
concepts to answer this question—to try to give some sort of modal definition or
analysis of dependence. The usual thought is that the salient modal notion is ‘can’t
exist without’. The xs depend on the ys just in case the xs can’t exist without the ys, or
duplicates of the xs can’t exist without duplicates of the ys, and so on. Yet these modal
analyses look too coarse, for a variety of reasons.3
To begin with, there is the problem of necessary co-existents. Kit Fine (1995) gives,
as an example, the famous case of Socrates and {Socrates}, which exist in all the
same worlds and yet while {Socrates} depends on Socrates, the dependence does
not hold in the other direction. A further problem is created by necessary existents.
Suppose, for example, that there are, necessarily, numbers. It shouldn’t follow from
this that everything is dependent on numbers, simply because nothing can exist
without numbers. Likewise, the theist believes in a necessary existent (God). Yet, while
some theists might be interested in defending the claim that everything depends on
God, it doesn’t look like this dependence claim should simply follow from the idea
that God exists necessarily.
These concerns have led many contemporary metaphysicians to argue that we need
a hyperintensional account of dependence. Nothing modal is going to be fine-grained
enough to do the work we want dependence to do, for example, to allow us to say
that sets are dependent on their members but not vice versa, or that numbers exist
necessarily but nothing non-numerical depends on them, and so on.
2 Koslicki (2013), p. 32.

3 The counterexamples I give are phrased as counterexamples to the modal analysis of dependence
as ‘can’t exist without’. But given some plausible assumptions, they’re also counterexamples to the modal
analysis in terms of duplicates. So, for example, if there’s a necessary existent, x, that has all of its intrinsic
properties essentially (which is plausible in the case of numbers, and perhaps also for the theistic God), then
not only can nothing exist without x existing, but nothing can exist without a duplicate of x exists. Likewise,
if we assume that the intrinsic nature of sets supervenes on the intrinsic natures of their members, you can’t
have a duplicate of Socrates without a duplicate of {Socrates}.
 symmetric dependence
Opting for hyperintensionality—and thereby divorcing dependence from modal

notions like ‘can’t exist without’—opens up some interesting options for dependence
claims in the presence of contingency. For example, it’s common to say that the
whole depends on the parts. And yet unless we adopt a strong form of mereological
essentialism, we don’t want to say that the whole can’t exist without its parts—we
want to allow that the whole could have been composed of different parts. What does
this do to our dependence claim? Those attracted to modal definitions need to do
some fancy footwork here—they need to argue, for example, that there’s a difference
between de re and de dicto dependence (or between rigid and generic dependence, or
the like). The whole depends on having some parts or other, but not on the parts it
in fact has. But why should we think that the whole is necessarily a complex object,
even if it is actually so? Perhaps that there are possible worlds in which this thing
which is in fact a complex object is instead an extended simple, for example. And
so we can introduce a further complication—talk of duplicates. Yes, the whole could
exist without having any parts at all. But a duplicate of the whole can’t exist without
having some parts or other. And so we continue, the modal definitions getting more
and more intricate. But once dependence is divorced from the modal notion of ‘can’t
exist without’, it’s not clear that any such complication is needed—or, indeed, that
we need a distinction between de re and de dicto dependence at all. Once we give
up on modal analyses of dependence, we might consider the option that necessary
connections aren’t even necessary, let alone sufficient, for dependence. We could then
say simply that the whole depends on its parts—on the parts it in fact has in the actual
world. Yes, there’s a possible world in which the whole has different parts. Yes, there’s a
possible world in which the whole has no proper parts at all. But none of that precludes
us from saying that, in the actual world, the whole depends on the parts it actually has.
Not all accounts of dependence will want to embrace this option, certainly—more on
this in §4.3—but its availability is an interesting upshot of separating dependence and
modality.
Saying that dependence is hyperintensional doesn’t preclude trying to give a
definition or analysis of dependence—it just precludes giving that analysis in modal
terms. Kit Fine (1995), for example, characterizes dependence via essence—x depends
on y just in case part of what it is to be x involves y—y is a constituent of some essential
property of x. In a similar vein, Benjamin Schnieder (2006) defines dependence via
metaphysical explanation—x depends on y just in case there exists some F such that
x exists because y is F. In recent work, Karen Bennett (2017) defines dependence via
her notion of a building relation. Something is independent just in case it is unbuilt,
otherwise it is dependent.
Others take dependence as primitive. Schaffer (2010b), for example, argues that
we can say many informative things about dependence, but that we shouldn’t
attempt to define or analyze it. Rosen (2010) likewise eschews attempts at defining
dependence in favor of giving examples of it and then showing what work it
can do.
elizabeth barnes 
In what follows, I’ll remain neutral on this issue. I don’t have any particular
definition of dependence in mind, nor am I assuming dependence cannot be defined.
My arguments should be applicable no matter which of these competing accounts of
dependence you favor.4 But it is important for my purposes—as will be clear—that
dependence is understood hyperintensionally.
1.3 Unification
Finally, there is the question of whether there are lots of different varieties of ontolog-
ical dependence, or whether there is just a single relation of ontological dependence.
There’s been somewhat of a cottage industry devoted to identifying different types
of ontological dependence—distinguishing between, say, rigid existential necessary
dependence and generic existential necessary dependence and identity dependence.5
Discussions of these varying types of dependence, and how we can define and
distinguish them, has generated a complex literature with lots of epicycles.
But, perhaps as a backlash to this increasing complexity, it’s become prevalent in
recent discussions in metaphysics to assume that there is a single, unified relation of
ontological dependence. This is the strategy employed in, inter alia, Cameron (2008a),
Rosen (2010), Schaffer (2010a), and Schnieder (2006). In what follows, I’ll proceed
along similar lines and speak of ontological dependence simpliciter. But I’ll argue in
§4.3 that nothing much hangs on this choice.
2 The Orthodoxy
Orthodoxy about dependence includes the claim that dependence is asymmetric. But
a striking feature of this orthodoxy is that little in the way of argument is given to
support it.6 The asymmetry of dependence is very often simply assumed without
further comment,7 and is perhaps something we’re meant to find intuitive or obvious.
Perhaps the most prevalent argument for the asymmetry of dependence has less
to do with dependence itself, and more to do with other relations or concepts that
dependence is often assumed to be connected to: in virtue of, grounding, priority,
fundamentality, and so on. Dependence is often mentioned in the same breath with
4 An exception here is Bennett (2017)’s definition of dependence. Bennett defines the independent as the
‘unbuilt’ (in her terminology). But in cases I’ll give below, there are things which are plausibly ‘unbuilt’ in
Bennett’s sense, but which I’m arguing are dependent. So if you accept Bennett’s definition of independence,
you won’t find these cases persuasive. But I’m hoping that the cases will give you reason to reconsider
Bennett’s definition of dependence.
5 See especially Lowe (2009) for an overview.
6 See especially Bliss (2012) for a very helpful overview of the relative paucity of argument for many
of the key assumptions in discussions of dependence and cognate notions. E.J. Lowe (1994) gives a
brief suggestion at an argument for asymmetry (p. 39), saying that our objection to symmetric cases of
dependence is analogous to our objection to circular arguments. I’m not exactly sure what to make of this
argument, other than to say that there’s a difference between circular arguments and holistic explanations.
7 As in, inter alia, Bennett (2017), Cameron (2008a), Schaffer (2010a), and Rosen (2010).
 symmetric dependence
these other (equally fashionable) notions. More significantly, even, as perhaps the least
esoteric of this cluster, dependence is often used as something which can help explain
or get traction on the somewhat more slippery notions of priority, grounding, and in
virtue of.8
So, for example, Karen Bennett (2017) remarks: ‘I do not think there is any
question that independence is a–the–central aspect of our notion of fundamentality.’
Similarly, Schaffer (2010b) takes a rejection of limitless or circular dependence to be
a consequence of the claim that some things are fundamental and that ‘all being must
originate in basic being’ (p. 37). And Koslicki (2013) proposes (although acknow-
ledging it to be controversial) the ability to illuminate disputes about fundamentality
where there is not a dispute about what exists as a criteria of success for accounts of
ontological dependence.
Relations of priority and relative fundamentality are, insofar as I have any grip
on them, plausibly asymmetric. And that is because they need to be asymmetric in
order to do the work we want them to do. These are relations that are introduced
in an attempt to take us from the derivative (the constructed, the grounded, the
nonfundamental) down toward the bedrock (the ultimate grounds, the fundamental,
the basic). It’s not a constraint of such relations that they ultimately bottom out.9 But it
does seem to be a constraint that they’re headed in a single direction. Their asymmetry
is built into the work we want them to do—it’s part of what they are for.10
The case is somewhat less clear for in virtue of and grounding.11 But certainly, if
you want to treat these as relations that take you from something you should treat
with less ontological seriousness (or even, something that is ‘less real’; see Fine (2001)
or McDaniel (2013)) to something that you should treat with more ontological
seriousness, then you need them to be asymmetric. The basic point, then, is this:
relations which purport to take us from the derivate to the fundamental are plausibly
viewed as asymmetric.
So here is an argument that dependence must be asymmetric. Dependence is
intimately connected to (and perhaps even explains or is one and the same thing
as) relevant notions of fundamentality, priority, grounding, and so on. Dependence
is the kind of relation that explains the connection between the fundamental and
8 So, for example, Schaffer (2010a), (2010b) and Cameron (2008a) both explain priority partly in terms
of dependence (and Schaffer especially often uses dependence-talk and priority-talk interchangeably);
Rosen (2010) explains ‘in virtue of ’ in terms of dependence; Bennett (2017) explains relative fundamen-
tality in terms of dependence; and Wilson (2014) identifies the relation of grounding as the target of ‘the
idioms of dependence’.
9 See Cameron (2008a) for discussion.
10 This is evidenced by the way we use them. We say ‘prior to’ and ‘more fundamental than’. I genuinely
cannot make sense of what it would mean to say ‘x is prior to y and y is prior to x’ or ‘x is more fundamental
than y and y is more fundamental than x’, nor do I know what locutions we might replace these with that
would render such claims coherent. So, at least as they are commonly used, I simply cannot make sense of
symmetrical cases of priority or relative fundamentality.
11 Wilson (2014) makes a case for the non-symmetry of grounding, for example.
elizabeth barnes 
the derivative—it takes us from the derivative (the dependent) to the fundamental
(the independent). Any relation that plays this role must be asymmetric. And so
dependence must be asymmetric.
I think it’s correct that if dependence is to play this role, then dependence must be
asymmetric. But what I’m going to argue is that it’s far too quick to simply assume that
this is the kind of role dependence ought to play. And a big part of the reason it is far
too quick is that there’s good reason to think that dependence isn’t asymmetric.
The idea that dependence and fundamentality come apart is one that we might
find plausible regardless of whether we think dependence is asymmetric, and it’s
an idea that can be put to useful work. For example, in previous work I argue that
dependence and fundamentality come apart in both directions—that there can be
fundamental dependent entities and derivative independent entities.12 Distinguishing
the two notions lets us make sense of a range of interesting (and independently
motivated) positions in metaphysics, including Agustin Rayo’s (2013) trivialism about
mathematical ontology (according to which numbers are plausibly construed as inde-
pendent but not fundamental13 ) and ontological emergence, which can be plausibly
understood as the idea that there are fundamental dependent entities.14 In what
follows, I’ll give a further reason for thinking that dependence and fundamentality
come apart: dependence should be understood as a non-symmetric relation.
3 The Case for Non-Symmetry

To make the case that dependence should be understood as non-symmetric, rather
than asymmetric, I’m going to make the case that dependence can sometimes hold
symmetrically. And to make the case that dependence can sometimes hold symmet-
rically, I’m going to proceed by a series of examples. Of course, any of the particular
cases I offer can be resisted. But when viewed as a whole, the range of cases is striking.
Examples of apparently symmetrical dependence are not hard to come by—they can
be found across a wide range of metaphysical theories, and in wide variety. The upshot
of this, I’ll argue, is that we can’t maintain that dependence is asymmetric without
ruling out wide swathes of the metaphysical landscape. And that quite simply isn’t the
job of a notion of dependence—which is, after all, meant to be neutral across various
ontologies—especially in the absence of independent argument that dependence must
be asymmetric.
In discussions of ontological dependence, there are at least two (potentially distinct)
ways of characterizing dependence: via essence and via explanation. The Finean
12 See Barnes (2012).

13 See especially chapter 3.
14 Indeed, it’s for precisely this reason, i.e. that emergence is the idea that there are fundamental things
which are also dependent things, that Bennett (2017) argues that emergence is deeply mysterious, and
possibly nonsensical. But absent further argument that dependence and fundamentality must go together,
such skepticism is unmotivated.
 symmetric dependence
account of dependence says that x depends on y just in case what it is to be x involves y

(y is a constituent of some essential property of x). Whereas an explanatory approach
(like the one endorsed by Schnieder) says that x depends on y just in case x exists, or
is the way it is, because y is F. I’m not endorsing either of these characterizations of
dependence. But in what follows, I take these two criteria—the essentialist claim and
the explanatory claim—as indicators of dependence, and I provide cases that motivate
symmetrical dependence for each.15
3.1 Immanent universals and essentialism
The first case I’ll offer is inspired by neo-Aristotelian metaphysics. Here are two claims
that are broadly Aristotelian in spirit: universals are immanent and membership in
natural kinds is had essentially. If universals are immanent, then universals require
the existence of their instantiations. An uninstatiated universal is impossible, perhaps
even incoherent. Universals don’t exist in some Platonic heaven and then get stapled
on to their instances (or not, if they’re uninstantiated). Rather, universals are inti-
mately bound to their instances, and, more generally, to being instantiated.16 If kinds
are had essentially, then for any x that’s a member of kind K, part of what it is to be x
is to be a member of K.
Both these claims are quite naturally understood as dependence claims. Immanent
universals depend on their instances. Part of what it is to be a universal, on this picture,
is to have instances. And individuals depend on their kinds—part of what it is to be
those particular individuals is to instantiate those kinds. If being F is essential to x,
then anything that fails to instantiate F isn’t x. Part of what it is to be x is to be F. And
so, plausibly, we can say that x depends on being F.
But the combination of these two doctrines straightforwardly yields symmetrical
cases of dependence. Suppose that being an electron is a universal, the instantiations
of which make up a natural kind (the electrons). If universals are immanent, then
the universal of being an electron depends on its instances. But, likewise, if natural
kinds are essential then its instances depend on the universal—all the things that
are electrons wouldn’t be the very things they are without the universal of being
an electron. And so, on this sort of neo-Aristotelian picture, we get cases where
dependence holds symmetrically. For those universals which correspond to natural
kinds—and, in general, to essential properties—the universal depends on instances,
and the instances depend on the universal.
15 I am being explicitly neutral about whether modal connections such as ‘can’t exist without’ are a
necessary condition for dependence, but I am taking it that Fine’s essentialist criterion for dependence
is at least an indicator of dependence (though perhaps not necessary for dependence). In what follows, I’ll
assume that Schnieder’s explanatory criterion—which I’m assuming doesn’t appear to have the same modal
consequences as Fine’s essentialist criterion (e.g. a complex object, x, could exist or be the way it is because
it’s parts, the ys, are arranged in a certain way F, even if there are possible worlds where that very thing x is
mereologically simple)—is also an indicator of dependence.
16 See especially Armstrong (1978b) for an overview and defense of immanent universals.
elizabeth barnes 
3.2 Armstrongian states of affairs

Consider the states of affairs metaphysic popularized by David Armstrong. There’s
a deep puzzle in Armstrong’s metaphysics regarding the relationship between states
of affairs and their constituents (particulars and universals). Consider the state of
affairs of Jane being human.17 That state of affairs binds together two constituents:
the particular individual Jane and the universal of being human. But the puzzling
question for Armstrong is what the relationship is between states of affairs and their
constituents. Do states of affairs depend on their constituents? Or do constituents
depend on states of affairs?
The trouble is that embracing either horn of this dilemma is problematic for
Armstrong. If we say that states of affairs depend on their (independent) constituents,
we get a picture in which the explanatory bedrock is particulars and universals. But
if what’s ultimately independent are the constituents of states of affairs—rather than
the states of affairs themselves—then Armstrong’s metaphysics loses its Tractarian
ambitions. Armstrong wants an ontology of facts—a ‘world of states of affairs’—
in which facts are fundamentally explanatory. For Armstrong, what explains the
existence of the particular Jane and the universal being human ought to be the
existence of the state of affairs—the worldly fact—of Jane’s being human. The reason
there are particulars and properties, on a Tractarian metaphysics, is because there are
things having properties—that is, because there are states of affairs. This picture is
undermined, however, if Armstrong takes particulars and universals as independent
and understands states of affairs as asymmetrically dependent on—and thus asym-
metrically explained by—particulars and universals.
But Armstrong encounters a different problem if he takes states of affairs to be
independent, and constituents to be dependent on states of affairs. If this horn of the
dilemma is embraced, then the metaphysic becomes explanatorily impoverished. For
example, we want to be able to say that the states of affairs of Jane’s being human
and Tom’s being human have something in common. But if the ultimate explanatory
bedrock is just the states of affairs, and not their constituents, then it’s hard to see how
we could explain this commonality. We want to be able to say that the constituents of a
state of affairs explain why that state of affairs is the way it is. Jane’s being human is the
state of affairs it is because of the constituents Jane and being human, and it is more
similar to Tom’s being human than to Rex’s being a dog because of the constituents
involved in each state of affairs.
The most stable position for Armstrong, I contend, is that states of affairs are a
case of symmetrical dependence. States of affairs depend on—and are thus explained
by—their constituents. But likewise individual constituents depend on—and thus are
17 Armstrong wouldn’t allow that things like Jane and being human are constituents of fundamental
ontology. So replace Jane and humanness with more scientifically respectable terms, if you’re worried about
that.
 symmetric dependence
explained by—states of affairs. That is, the most stable position for Armstrong is a
type of explanatory holism (discussed in more detail in §5.1). But if we separate
dependence from fundamentality, this doesn’t preclude Armstrong from saying that
both states of affairs and their constituents are fundamental. They are fundamental, but
they each depend on the other. This allows Armstrong to respond to the resemblance
problem, and it allows him to have his world of facts. The cost of this picture is, of
course, a cost to parsimony—we end up with a fundamental ontology of both states
of affairs and their constituents. But the claim here is that this is the most stable way
of making sense of the fact-based ontology that Armstrong wants to defend.
3.3 Tropes and the problem of ‘bare mass’
According to trope metaphysics, properties are individual ‘particular thisnesses’.18
A traditional property metaphysics says that if the rose and the carnation are both red,
then they both have the same property—they both instantiate redness. But the trope
theorist says that properties are particulars. The rose’s redness and the carnation’s
redness are two different (non-repeatable) things. What the rose and the carnation
have in common is that the rose’s redness and the carnation’s redness are similar
(perhaps exactly similar).
Trope theory is commonly combined with bundle theory—the view that objects
are nothing more than collections of properties.19 According to trope bundle theory,
objects just are collections of particular thisnesses (there is not an underlying sub-
stance which instantiates or is the bearer of properties). The combination of tropes
and bundle theory gives rise to an explanatory puzzle sometimes called the problem
of ‘bare mass’ or ‘free mass’.20 If properties are particulars, and objects are nothing
more than collections of properties, could you have an object that was nothing but an
individual mass trope? Nothing about trope bundle theory rules this out, and yet it
seems incoherent. So much the worse for trope bundle theory.
But allowing that dependence can hold symmetrically gives the trope bundle
theorist an easy line of response to this objection. The problem for the bundle theorist
is that she cannot appeal to an underlying object on which properties depend—
objects just are collections of properties. But if dependence can hold symmetrically,
what she can say instead is that there are tropes which mutually depend upon each
other.21 You cannot have a mass trope without a size trope and a shape trope, for
example. And so on, mutatis mutandis, for shape tropes and size tropes. The picture
here is one of ‘dependence clusters’—mass depends on shape and size, size depends
on mass and shape, and so on. Part of what it is to have mass is to have shape and
18 The contemporary discussion of tropes goes back to at least Williams (1953). See Maurin (2013) for
an excellent overview and discussion.

19 See Paul (2013) for a good introduction and overview.
20 See, inter alia, Armstrong (1997) and Schaffer (2003) for discussion.
21 This sort of interdependence between tropes is posited as a solution to the free mass problem in both
Denkel (1996, 1997) and Simons (1994).

elizabeth barnes 
to have size, for example. And part of what it is to have shape is to have mass and to
have size. And so on. These properties are all interdependent. And so the resulting
ontology that trope bundle theory can offer includes clusters of interdependence—
properties are particular ‘thisnesses’, but that doesn’t mean that ‘particular thisnesses’
are independent. Trope bundle theory needn’t encounter the problem of bare mass if
dependence can hold symmetrically between tropes.
3.4 Mathematical ontology
Thinking that there are numbers might also give you good reason to accept symmet-
rical cases of dependence. This is particularly evident if your mathematical ontology is
that of non-eliminativist structuralism. That is, you think there are numbers, and you
think that what numbers are are nodes or positions in a mathematical structure.22
Non-eliminativist structuralists often say that each node of the structure depends
on all the others nodes—and perhaps even on the structure itself as well. And, as
Linnebo (2008) persuasively argues, it’s easy to see why such dependence claims
are needed. What it is to be a particular node in the structure is bound up in the
other nodes being what they are. Consider the number six.23 The non-eliminativist
structuralist is a realist about mathematical ontology. She thinks that the number six
exists. Moreover, she thinks that what the number six is is a particular node in a
complex mathematical structure. But that particular node is the number six in virtue
of the relations it stands in to the other nodes in the structure. Likewise, the fact
that the particular node is the number six is explained by the relations it stands in
to the other nodes in the structure. And so for the non-eliminativist structuralist,
the number six is dependent on the other numbers (which are, mutatis mutandis,
themselves dependent on the other numbers). The non-eliminativist structuralist is
(plausibly) committed to symmetrical cases of dependence in order to explain her
ontology.24
While the case for symmetrical dependence is most vivid for the structuralist, other
versions of realism about mathematical ontology might have similar explanatory need
for such inter-dependencies. On a Finean conception of dependence, for example,
x depends on y if what it is to be x involves y—that is, if y is a constituent of
some essential property of x. On such an understanding of dependence, numbers
are plausibly interdependent—that is, they depend on each other. The mere fact
of their necessary co-existence doesn’t entail interdependence, but the explanatory
22 See especially Shapiro (1997) for explication and discussion of this view.
23 Linnebo argues that the structural realist shouldn’t think this about all mathematical ontology—it
is implausible for sets, for example—but maintains that it’s a central part of the structuralist picture in
many cases. My use of natural numbers here is no doubt not the most compelling instance of symmetry—
Linnebo (2008) provides much more sophisticated examples in his paper.
24 It’s worth noting that structuralisms in general—whether mathematical or not—are likely to give rise
to symmetric dependencies, simply because of the holistic style of explanation they favor. Structural realism
about the ontology of physics might similarly be interpreted as involving claims of symmetric dependence
between individuals and structures, for example. See e.g. French (2014).
 symmetric dependence
connections will. What it is to be the number six is bound up in what it is to be the

number five and the number seven, and so on. The number six would not be the thing
it is were it not the successor of five, but it also would not be the thing it is were seven
not its successor, so the number six is dependent on the numbers five and seven; but
seven would not be the thing it is were it not the successor of six, and so seven is
dependent on six; and so on.
3.5 Events
Many people who endorse an inflationary metaphysics of events are attracted to both
the idea that at least some events contain/are constituted by smaller events and to
the idea that at least some events have some of the smaller events they contain/are
constituted by essentially.25 The event WWII contains many smaller events—some
insignificant (such as a particular lighting of a cigar by Winston Churchill) some much
more significant (such as the evacuation of Dunkirk). And while WWII might have
been the same event without that particular lighting of Churchill’s cigar,26 it’s plausible
that WWII just wouldn’t have been the same event without the evacuation at Dunkirk.
Without the evacuation at Dunkirk, it literally would have been a different war—the
evacuation is an essential part of the war.
But, similarly, we might think that being a part of WWII is essential to the
evacuation of Dunkirk. Sure, you could have a duplicate of that event that doesn’t
take place in the wider context of WWII. But that duplicate isn’t the evacuation at
Dunkirk—part of what it is to be the evacuation at Dunkirk is to be a part of WWII.
It’s part of the character of the event that it had the goals it had, that it was part of a
wider mission, that it took place within the particular geopolitical context that it did,
and so on.
But if the events-ontologist accepts both these claims, she accepts a symmetric
case of dependence. The event of the evacuation depends on the event of WWII.
A qualitatively similar event that isn’t a part of WWII isn’t the same event. But
likewise the event of WWII depends on the event of the evacuation. An event that
doesn’t contain the evacuation at Dunkirk isn’t WWII. The two events—WWII and
Dunkirk—each depend on each other to be what they are.
3.6 Summing up
I have presented a range of cases—across a variety of topics and debates in
metaphysics—which might motivate the claim that dependence can hold sym-
metrically. None of these cases are, by themselves, knock-down reasons to reject
25 See Hornsby (1997), chapter 3, for an excellent articulation of the former claim. Hornsby also seems
in many places to endorse the latter, although this is less explicit.
26 Although see Lombard (1986) for an argument that we should embrace a radical form of essentialism
about events.
elizabeth barnes 
asymmetric dependence. But their dialectical force when taken together is, I’ll argue,
greater than the sum of their parts.
Orthodoxy assumes that dependence is asymmetric. But, as already noted, there’s
very little in the way of argument to support this tenet of orthodoxy. It is, more often
than not, assumed rather than argued for. And it’s against this backdrop that I give
this range of cases in which dependence is better understood as non-symmetric,
rather than asymmetric. These cases are, taken collectively, quite striking. Cases
where dependence holds symmetrically were not hard to find—there are plenty of
them, including some very popular and well-known theories in metaphysics (and
if the goal of this paper had simply been to list potential examples then the list
could have continued for some pages). Nor is it a single niche area or type of view
that’s giving rise to such cases—rather, the examples come from across a wide range
of theories in metaphysics, and from a variety of different traditions. This makes
a default, undefended assumption of asymmetry in dependence look odd—to say
the least.
Suppose we take on board the default assumption that dependence is asymmetric. If
I’m right that the above cases should plausibly read as ones in which dependence holds
symmetrically, then to take this assumption on board is to rule out these cases. That
is, to assume that dependence is asymmetric is to rule out vast swaths of interesting,
historically grounded metaphysics—or at least to force on them unpalatable inter-
pretations. That—I contend—isn’t dialectically appropriate. Absent some compelling
argument that dependence must be understood as asymmetric, it isn’t the role of a
notion of dependence to simply rule out (or even severely constrain) diverse and
promising metaphysics. That’s not what a notion of dependence is for—if we can
rule out all such views simply by pointing out that they run afoul of the asymmetry
criteria of metaphysical dependence (which, again, there isn’t much argument for)
then dependence is doing too much work.
4 Objections
4.1 These cases are all impossible
I’ve presented the above examples as more or less argument by cases. But a clear
objection is simply this. Most metaphysicians don’t think the views described in the
cases above are true. Most metaphysicians think that whatever ultimate metaphysical
theory is true is necessarily true. Therefore most metaphysicians will think that the
views I’ve described are necessarily false. Why think you can convince people that
dependence can sometimes hold symmetrically by giving a bunch of impossible cases?
In reply, let me clarify an important point. I’m not arguing that there are in fact
cases of symmetrical dependence. Here’s what I’m arguing, in a nutshell:
• People assume that dependence is asymmetric. They shouldn’t.
• People assume that asymmetry is built into the concept of dependence. It isn’t.
 symmetric dependence
• Insofar as we want a relation of dependence that is neutral across varying

ontologies (as it’s often assumed to be in the literature on dependence),
dependence needs to be non-symmetric, not asymmetric.
• Insofar as we’re warranted in talking about a relation of dependence, a non-
symmetric relation seems to fit our purposes better than an asymmetric relation.
• We shouldn’t rule out ontologies just because they allow for symmetric cases of
dependence.
All of that is compatible with it being the case that dependence only ever in fact
holds asymmetrically. Suppose Jonathan Schaffer (2010b) is right, and monism is
the true theory of fundamental metaphysics. If monism is true, then everything
asymmetrically depends on the world. Does that mean that Schaffer should think
dependence is in fact asymmetric? Well, it depends on what role dependence is
playing in the dialectic. If dependence is something that’s supposed to be neu-
tral across different ontologies—something that we can use to explain common
structures between ontologies, or some kind of generic explanatory principle—then
the fact that it only ever occurs asymmetrically doesn’t mean the relation itself is
asymmetric.
Schaffer seems to use dependence in this specific-ontology-neutral sense. (As do
Bennett, Koslicki, Fine, Rosen, etc.) Indeed, Schaffer (2010b) starts out with general
reflections about dependence (not tied to any particular ontology) that include the
claim that dependence is asymmetric, and then uses these reflections about depend-
ence as part of his motivation for monism. It would be an odd sort of bootstrapping, to
say the least, to then point to the asymmetry of dependence in monism as an argument
for the general claim that dependence is asymmetric.
Dependence might, prior to commitment to a particular ontology, turn out to be
non-symmetric, and so we shouldn’t simply assume that it’s asymmetric, and shouldn’t
use the assumption that it’s asymmetric to rule out particular ontologies.
4.2 These cases aren’t symmetric dependence, they’re joint dependence

Another way of objecting to what I’ve said above is that in the cases I describe, the
reason it looks like you get two things which depend on each other is that they each
depend on the same further thing. These are cases of joint dependence, not symmetric
dependence.
But this move doesn’t look like it’s available for all the cases given above. If we
combine Aristotelian universals with essentialism, it doesn’t look like there’s anything
further we can point to that both universals and their instances depend on. (Aristotle
himself may have thought that everything ultimately depends on the Aristotelian god,
but I doubt this move will be particularly popular.) Likewise, for trope bundle theory,
it doesn’t seem plausible that there’s any one supertrope on which all the other tropes
in the bundle depend. So as an across-the-board response, appeal to joint dependence
doesn’t work.
elizabeth barnes 
In other cases, there might be candidates for joint dependence. Maybe what we
ought to say about structuralist realism in mathematics is that all the nodes depend
on the structure itself, but that the structure doesn’t depend on the nodes. But this
option looks ad hoc and forced. Why would the structure be independent of the
nodes? How could it be? The ontology we’re forced into if we want something we
can say is a source of joint dependence begins to look mysterious and bizarre. Why go
in for such an ontology, rather than just allowing that this is a case where dependence
holds symmetrically? The answer had better not be an assertion that dependence just
has to be asymmetric.
4.3 These cases only arise because you failed to distinguish
different kinds of dependence
As outlined in §1, I’m following the tradition in the literature that treats dependence
as a single, unified relation. But it could be objected that it’s precisely the refusal to
distinguish between different varieties of dependence that’s leading to apparent cases
of symmetric dependence. After all, it’s prima facie cases of symmetry like the de-
pendence of universals on their instances and instances on universals that, for
example, motivates Jonathan Lowe (1994) to distinguish between generic existential
necessary dependence and rigid existential necessary dependence, and between exis-
tential and identity dependence.27 And similar worries have motivated the distinction
between de re and de dicto dependence. So we don’t, if we’re careful, really have a case
where we’ve got a single relation that’s holding symmetrically—we’ve just got two (or
more) different forms of dependence.
The key thing to say about this objection, once again, is that it doesn’t look like a
response that can be leveled at all the cases I give above. Even if we allow for different
forms of dependence, whatever sense in which a trope bundle’s shape trope depends
on its size trope is the same sense in which its size trope depends on its shape trope.
And likewise for individual nodes in a mathematical structure, and for Dunkirk and
WWII. So even if we granted that there are different kinds of dependence—a view
which, as discussed in §1.2, has its problems—that wouldn’t eliminate all apparent
cases of symmetrical dependence.
But let’s consider the case of de re and de dicto dependence. Someone might object
that de re and de dicto dependence are very different things—in the case given in
§3.1, for example, the particulars depend on that very universal, but the universal
merely depends on having some particulars or other that instantiate it. This, it might
be protested, is not the same relation of dependence in both cases. And so the case is
not, in fact, a case of symmetric dependence.
27 Interestingly, though, some of Lowe’s arguments for accepting multiple kinds of dependence seem
to rest on the assumption that there cannot be symmetrical cases of (strong) ontological dependence. The
above discussion can be taken as a reason to apply modus tollens where Lowe applies modus ponens.
 symmetric dependence
As discussed in §1.2, divorcing dependence from modality might give us reason to

push back against this thought. The motivation for a deep distinction between de re
and de dicto dependence seems like an after-effect of modal accounts of dependence.
The universal could exist without any of the things that in fact instantiate it—it just has
to be instantiated by something, not by the particular things that in fact instantiate it.
Whereas the particular things that instantiate it couldn’t exist without the universal.
But if ‘couldn’t exist without’ doesn’t give us a grip on dependence, why think this is
relevant to the question of whether the universal depends on the things that in fact
instantiate it, just as the things that in fact instantiate it depend on the universal?28
Not all accounts of dependence can accept this. Someone attracted to Fine (1995)’s
essentialist account of dependence, for example, will want to maintain something like
the de re/de dicto distinction for this case—part of what it is to be the particulars is that
they instantiate that very universal, but it’s not part of what it is to be that universal
that it is instantiated by those particulars (part of what it is to be that universal
is simply that it’s instantiated by some particular or other). For these accounts of
dependence, there are two points to make. The first is again simply that the cases given
in §3.3, §3.4, and §3.5 (and perhaps §3.2 as well) look to be cases of symmetrical de re
dependence.29 The second is that even if you think that, in a case like §3.1, it’s not the
same relation of dependence going in both directions, the resulting picture—where
both particulars and universals are dependent, even if they are dependent in different
ways—yields an interesting explanatory structure that further undercuts that idea that
fundamentality and independence always go together.
A more complicated case is that of the distinction between full and partial depen-
dence. For those inclined to think this distinction is important, the cases given above
might seem like they only give support to the idea that partial dependence can
sometimes hold symmetrically. The Aristotelian universal is partially dependent on
each of its instances, but not wholly dependent on any of them. The mass trope is
partially dependent on the size trope and the shape trope, but not wholly dependent
on either. And so on. Perhaps full dependence is asymmetric, regardless of whether
partial dependence might be non-symmetric. And perhaps full dependence is the
more important, bedrock notion.
28 This point is particularly salient if we separate dependence from relations like priority—which, once
the prospect of the non-symmetry of dependence is raised, I think we should. It’s plausible to think that
certain kinds of necessitation claims go along with priority. If the xs are prior to y, then necessarily if you
have the xs you have y. That’s one way of interpreting the idea that, if the xs are prior to the y, then in some
sense having the xs gets you y ‘for free’.
29 And it’s also possible to motivate symmetric cases of de dicto dependence. In some of the medieval
discussion of ‘substantial form’, the relationship between matter and form seems to suggest such depen-
dence. The account of substantial form given by Suarez, for example, seems to suggest a reading in which
matter depends on having some form or other (but not on having any particular form), and likewise form
depends on being realized in some matter or other (but not on any particular matter). See Pasnau (2011),
pp. 561–3.
elizabeth barnes 
But I’m skeptical that there’s an important difference between full and partial
dependence—or at least I’m skeptical that if there is an important difference, the cases
I’ve given only address the latter. Full dependence seems simply like the limit case of
partial dependence, rather than something different in kind. Consider again the case
of Aristotelian universals. Universals depend on their instances, but they don’t wholly
depend on any particular instance—they partially depend on each instance, and
collectively depend on all their instances. But if universals depend on their instances,
it looks like it’s possible for a universal to depend on a single instance. Suppose that
all natural kinds correspond to universals, and suppose further that the elements of
the periodic table each represent natural kinds. Many of the elements of the periodic
table are plentiful and naturally occurring, but some can only be made in specialized
laboratory conditions, and have only ever been made a few times. An element like
Einsteinium, for example, has only had a few instances. Now consider the possible
world in which Einsteinium is only made once. That’s a world in which Einsteinium
only has one instance. In that world, the universal Einsteinium wholly depends on the
single instance, and the single instance wholly depends on the universal Einsteinium.
5 Morals of the Story

5.1 Holistic explanation
If dependence is non-symmteric, why does it matter? Another objection to symmetric
cases of dependence that will doubtless crop up is that symmetric cases of dependence
license unacceptable circular explanations. Dependence—whatever else it may be—
is inextricably linked to metaphysical explanation. If x depends on y, then x is at
least in some sense explained by y. But suppose that we have a case of symmetric
dependence—x depends on y and y depends on x. In that case, the explanation
suggested is that x explains y and y explains x. Surely that’s unacceptable circularity.
However, I think that allowing such forms of explanation is a feature, not a bug,
of understanding dependence as non-symmetric. What non-symmetric dependence
allows is a certain kind of explanatory holism. The existence of the state of affairs
explains the existence of its constituents. The existence of the constituents explains
the existence of states of affairs. The existence of the mass trope explains the existence
of the shape and size tropes. The existence of the shape and size tropes explains the
existence of the mass trope. And so on. The most common models of explanation in
metaphysics are analogous to foundationalism in epistemology. Chains of explanation
ultimately ground out in primitive, unexplained explainers. But that needn’t be the
only way that explanation in metaphysics can work. We could have alternative pictures
that are more like coherentism: the overall explanatory structure can be holistic, and
there are no unexplained explainers.
Certainly, the availability of this kind of explanation is one reason why people
have in fact objected to non-symmetric dependence. E.J. Lowe (1994) and (2009), for
 symmetric dependence
example, remarks that our objection to symmetric cases of dependence is analogous

to our objection to circular arguments. But it’s hard to know what to make of this
claim. Circular arguments are valid. The reason that they’re bad arguments is an
epistemic reason—they don’t provide any new information, or any further warrant,
for thinking that the conclusion is true. And so they don’t play the justificatory role
we want arguments to play. But metaphysical explanation is explicitly non-epistemic.
When we say that x explains y and y explains x, we’re not saying that the way we come
to have knowledge of x is via y, and the way we come to have knowledge of y is via x, or
something along those lines. So it’s not clear what the analogous objection to circular
arguments would be in the case of metaphysical explanation.30
Another way of motivating a circularity objection can be found in Fraser
MacBride (2006). MacBride objects to the symmetry of dependence in mathematical
structuralism (or at least explores the following objection without fully endorsing it).
In order for a relation to obtain between two objects, MacBride argues, those objects
need to be:
independently constituted as numerically diverse. Speaking figuratively, they must be numer-
ically diverse ‘before’ the relation can obtain; if they are not constituted independently of the
obtaining of …[the] relation then there are simply no items available for the relation to obtain
between. (p. 67)
But symmetric cases of dependence seem to be cheating in this regard—it’s the

very obtaining of the relation that allows for the existence of the objects (because
they depend on each other). On MacBride’s picture—at least as I understand it—
objects are like pins in a bulletin board and relations are like bits of string that
you can hang between pins. You’ve got to have the pins there before you can hang
the string. It can’t be the hanging of the string that somehow magically gives you
the pins. But on this very flat-footed reading, MacBride’s objection carries just as
much weight against dependence-as-grounding or dependence-as-priority as it does
against symmetric cases of dependence (like the mathematical structuralist). On
such understandings of dependence, if x depends on y, then x isn’t ‘independently
constituted as numerically diverse’ in a way that’s explanatorily prior to the obtaining
of the dependence relation. It’s precisely because x depends on y—that is, precisely
because the relation of dependence obtains between x and y—that x exists. Symmetric
cases of dependence aren’t asking us to countenance anything radically different.
They’re just positing a case in which both relata—rather than one relatum—require
the obtaining of the dependence relation for their existence.
30 Lowe (2009) further asserts that metaphysical explanation is asymmetric, and therefore dependence—
because it tracks or is intimately bound up with metaphysical explanation—must likewise be asymmetric.

But he gives no argument for the claim that metaphysical explanation must be asymmetric. Schaffer (2010b)
appears to endorse this claim of Lowe’s, but likewise does not say why. Schnieder (2006) makes a similar
claim, but again does not argue for the asymmetry of explanation.
elizabeth barnes 
But a MacBride-esque worry that cuts directly against symmetric cases of depend-
ence would be this: we need to have the existence of at least some of the relata of a
relation independently of the obtaining of that relation. We need, as MacBride puts
it, a relatum ‘before’ (figuratively speaking) we can have a relation. In symmetric
cases of dependence, though, the very existence of the relata requires the obtaining
of the relation—the relata depend on each other. But put this way, the objection just
sounds like a denial of (rather than an argument against) the sort of explanatory
holism that non-symmetric dependence allows. It’s important not to be misled here
by the temporal metaphor. The objection is not whether we need objects temporally
before the obtaining of relations between those objects. Rather, the point—at least
as I understand it—is that we need at least some objects to be explanatorily prior to
the obtaining of any dependence relations. Or, to put it more simply, we need some
objects to be independent. That’s not an argument against holism—that’s just a denial
of holism.
Holistic explanations have a long and rich history in philosophy. They are, it’s
safe to say, out of fashion in much of contemporary metaphysics.31 But it isn’t
clear that they should be, especially considering the interesting work that holistic
explanation can do, and the interesting explanatory models it provides. And more
importantly, holistic explanations don’t look like the kind of thing that we should
dismiss without argument, simply by asserting that dependence is asymmetric. A non-
symmetric dependence relation allows for holistic as well as foundationalist models
of metaphysical explanation, and that’s one major reason why non-symmetry, rather
than asymmetry, should be the default assumption for dependence.
5.2 Dependence and grounding relations
The other main moral of the story is this: if dependence can be non-symmetric, then
dependence needs to be separated from talk of grounding, priority, in virtue of, and
so on. These relations are relations that aim to take us from the derivative to the
fundamental. They take us from things we treat with less ontological seriousness, or
‘get for free’, down to the ultimate ontological bedrock. But if dependence is non-
symmetric, it can’t play this role, and it can’t be jumbled together with these other
relations.
Suppose that the symmetric dependence interpretation of Armstrong really is the
best interpretation. If that’s the case, then for Armstrong nothing is independent. His
basic ontology is states of affairs and their constituents. But both are dependent (each
depend on the other). That doesn’t mean that Armstrong should think nothing is
31 Although if you look a little outside of ’mainstream’ metaphysics—especially to feminist
metaphysics—you will find plenty of champions of holistic explanation. See especially Haslanger (1995).
Some of the most salient examples can be found in feminist discussions of social construction and social
kinds. See, for example, Haslanger (2016) and Witt (2011). Much of the discussion of holism in feminist
philosophy is more directed toward epistemology, but often has striking consequences for metaphysical
holism—see especially Haraway (1991) and Harding (1993).
 symmetric dependence
fundamental. It just means that dependence isn’t a good guide in all cases to getting
at fundamentality.
Dependence is something distinct from theoretical gizmos—like grounding, prior-
ity, and in virtue of—tailored specifically to take us from the less fundamental to the
more fundamental. Dependence can do a lot of interesting work in our theories, but
it can’t do that. Nor can dependence be used to explain priority, grounding or the like.
Whatever sense (if any) we can make of those other relations and whatever work they
can do (if any) in our theories, they need to be clearly separated from dependence.32
References
Armstrong, David M. (1978a). Nominalism and Realism, volume 1 of Universals and Scientific
Realism. Cambridge: Cambridge University Press.
Armstrong, David M. (1978b). A Theory of Universals, volume 2 of Universals and Scientific
Realism. Cambridge: Cambridge University Press.
Armstrong, David M. (1997). A World of States of Affairs. Cambridge: Cambridge University
Press.
Barnes, Elizabeth (2012). ‘Emergence and Fundamentality’. Mind 121(484), pp. 873–901.
Bennett, Karen (2017). Making Things Up. Oxford: Oxford University Press.
Bliss, Ricki (2012). ‘Viciousness and the Structure of Reality’. Philosophical Studies (online first):
http://link.springer.com/article/10.1007%2Fs11098-012-0043-0.
Cameron, Ross (2008a). ‘Turtles all the Way Down: Regress, Priority and Fundamentality’. The
Philosophical Quarterly 58(230), pp. 1–14.
Cameron, Ross (2008b). ‘Truthmakers and Necessary Connections’. Synthese 161(1), pp. 27–45.
Denkel, Arda (1996). Object and Property. Cambridge: Cambridge University Press.
Denkel, Arda (1997). ‘On the Compresence of Tropes’. Philosophy and Phenomenological
Research 57, pp. 599–606.
Hornsby, Jennifer (1997). Simple Mindedness: In Defense of Naive Naturalism in Philosophy of
Mind. Cambridge, MA: Harvard University Press.
Fine, Kit (1995). ‘Ontological Dependence’. Proceedings of the Aristotelian Society 95,
pp. 269–90.
Fine, Kit (2001). ‘The Question of Realism’. Philosophers’ Imprint 1(2), pp. 1–30.
French, Steven (2014). The Structure of the World: Metaphysics and Representation. Oxford:
Oxford University Press.
Haraway, Donna (1991). ‘Situated Knowledges’. In Donna Haraway, Simians, Cyborgs, and
Women: The Reinvention of Nature. New York: Routledge.
Harding, Sandra (1993). ‘Rethinking Standpoint Epistemology: “What is Strong Objectivity?”’
In Linda Alcoff and Elizabeth Potter (eds), Feminist Epistemologies. New York: Routledge,
pp. 49–82.
32 Many thanks for helpful feedback and discussion to Ross Cameron, Sally Haslanger, Kris McDaniel,
Trenton Merricks, Daniel Nolan, Jason Turner, Robbie Williams, two anonymous referees, and audiences
at the Pacific APA, Birmingham University, Harvard University, Leeds University, and the University of
Virginia.
elizabeth barnes 
Haslanger, Sally (1995). ‘Ontology and Social Construction’. Philosophical Topics 23, pp. 95–125.
Haslanger, Sally (2016). ‘What is a (Social) Structural Explanation?’ Philosophical Studies
173(1), pp. 113–30.
Koslicki, Kathrin (2013). ‘Ontological Dependence: An Opinionated Survey’. In B. Schnieder,
M. Hoeltje, and A. Steinberg (eds), Varieties of Dependence: Ontological Dependence,
Grounding, Supervenience, Response-Dependence (Basic Philosophical Concepts). Munich:
Philosophia Verlag, pp. 31–64.
Linnebo, Øystein (2008). ‘Structuralism and the Notion of Dependence’. Philosophical Quarterly
58, pp. 59–79.
Lombard, Lawrence (1986). Events: A Metaphysical Study. London: Routledge.
Lowe, E.J. (1994). ‘Ontological Dependency’. Philosophical Papers 23(1), pp. 31–48.
Lowe, E.J. (2009). ‘Ontological Dependence’. The Stanford Encyclopedia of Philosophy (Spring
2010 Edition), ed. Edward N. Zalta: http://plato.stanford.edu/archives/spr2010/entries/
dependence-ontological/.
MacBride, Fraser (2006). ‘What Constitutes the Numerical Diversity of Mathematical Objects?’
Analysis 66(1), pp. 63–9.
McDaniel, Kristopher (2013). ‘Degrees of Being’. Philosophers’ Imprint 13(19), pp. 1–18.
Maudlin, Tim (2007). The Metaphysics Within Physics. Oxford: Oxford University Press.
Maurin, Anna-Sofia (2013). ‘Tropes’. The Stanford Encyclopedia of Philosophy (Fall 2013
Edition), ed. Edward N. Zalta: http://plato.stanford.edu/archives/fall2013/entries/tropes/.
Pasnau, Robert (2011). Metaphysical Themes 1274–1671. Oxford: Oxford University Press.
Paul, L.A. (2013). ‘Mereological Bundle Theory’. In Hans Burkhardt , Johanna Seibt, and Guido
Imaguire (eds), The Handbook of Mereology. Munich: Philosophia Verlag.
Rayo, Agustin (2013). The Construction of Logical Space. Oxford: Oxford University Press.
Rosen, Gideon (2010). ‘Metaphysical Dependence: Grounding and Reduction’. In Bob Hale
and Aviv Hoffmann (eds), Modality: Metaphysics, Logic, and Epistemology. Oxford: Oxford
University Press.
Schaffer, Jonathan (2003). ‘The Problem of Free Mass: Must Properties Cluster?’ Philosophy and
Phenomenological Research 66(1), pp. 125–38.
Schaffer, Jonathan (2010a). ‘The Internal Relatedness of All Things’. Mind 119(474), pp. 341–76.
Schaffer, Jonathan (2010b). ‘Monism: The Priority of the Whole’. Philosophical Review 119(1),
pp. 31–76.
Schnieder, Benjamin (2006). ‘A Certain Kind of Trinity: Dependence, Substance, Explanation’.
Philosophical Studies 129, pp. 393–419.
Shapiro, S. (1997). Philosophy of Mathematics: Structure and Ontology. Oxford: Oxford Univer-
sity Press.
Simons, Peter (1994). ‘Particulars in Particular Clothing: Three Trope Theories of Substance’.
Philosophy and Phenomenological Research 54, pp. 553–75.
Williams, D.C. (1953). ‘On the Elements of Being I’. The Review of Metaphysics 7(1): pp. 3–18.
Wilson, Jessica (2010). ‘What is Hume’s Dictum and Why Believe It?’ Philosophy and Phe-
nomenological Research 80, pp. 595–637.
Wilson, Jessica (2014). ‘No Work for a Theory of Grounding’, Inquiry 57, pp. 535–79.
Witt, Charlotte (2011). The Metaphysics of Gender. Oxford: Oxford University Press.
3
Grounding and Reflexivity
Ricki Bliss
Philosophers interested in the notion of ground are commonly of the view that the
grounding, or metaphysical dependence, relation is necessarily asymmetric, irreflex-
ive and transitive: they deny that anything can be self-grounded or self-dependent.
What the reasons are for this commitment, however, are less than clear. One might
suppose that this attitude is born from a commitment to the unacceptability of
circularity more generally. The coherence theory of truth, for example, was eschewed
by many for reasons of circularity. Just as the appearance of a seemingly vicious circle
in the foundations of naive set-theory leads philosophers either to abandon the reality
of sets, or to set about the formidable task of refining its axioms.
The kinds of circles with which this paper is concerned are not those generated
when we try, directly or indirectly, to explain grounding in terms of itself. Nor are
the kinds of circles with which we are interested those generated where we have
circular arguments, or become tangled in circular reasoning. The reason for this is
that metaphysical explanations are not arguments, and the issue is, presumably, not
one of trying to convince anyone of anything. Nor is the matter to hand to be confused
with what it is to suggest that some fact is self-evident. Suppose that I happen upon
a crime scene where the victim is laid out with a dagger in his chest. Although we
might suggest that it’s self-evident how the man met his end, we do not infer from
this that the poor fellow drove a knife into his own chest; or that what explains the
man’s death is the fact that the man is dead. Self-evident facts may bear no explanatory
relationships to themselves whatsoever. And self-grounded facts may not be self-
evident by anybody’s lights.
The kinds of cases this paper addresses are cases in which we say that some fact is
either fully or partially grounded in itself. Which is just to say that the kinds of cases
we are interested in are cases in which some fact bears the right kind of real, or perhaps
merely explanatory, connection to itself. But what does this even mean? And why, as
we are so often told, is it absurd or unacceptable?
If real relations of ground are a feature of objective reality, we might suppose that
there are metaphysically laden reasons for denying that there can be reflexive instances
of dependence: perhaps the consequences for the reality that the relation structures
ricki bliss 
would be unseemly—on some, as yet undefined, sense of unseemly. Or perhaps the

way facts would have to be, such that they can be self-grounded, is metaphysically
incoherent or impossible. Alternatively, one might be of the view that even though
there are real relations of ground that structure reality, instances of self-dependence
are completely innocuous. It has been suggested, for example, that the identity relation
is a relation that things bear harmlessly to themselves. One might wish to suggest,
then, that grounding—or some sub-species of it—behaves in just this way as well.
Why, exactly, instances of self-dependence would then be considered to be a problem
is a little bit difficult to see. Perhaps it is simply that there is no real need to posit them
in the first place for they add little to nothing of value to our theory.
Bearing in mind that ground is associated with explanation, however, we may,
instead, discover that there are powerful explanatory reasons for supposing there can
be no circles of ground. Indeed, for anyone who refrains from positing the existence
of real relations of ground, explanatory concerns may well be the most likely locus
of justifications for the commitment to the no-circularity assumption. Suppose that
reflexive instances of ground give rise to explanations of the form ‘x because x’; these
explanations might be unacceptable simply because they are trivial, uninformative,
and explanatorily useless.1 Or, alternatively, the connection between reflexivity and
non-well-foundedness might itself ensure the place for a particular kind of explana-
tory failure on our catalogue of reasons to eschew the possibility of circles of ground:
where the step between non-well-foundedness and explanatory failure would be
mediated by an argument from vicious infinite regress.
This paper aims to focus the reasons for which we might find reflexive instances of
dependence unacceptable: a task that necessitates an investigation into what it even
means for a fact to ground itself. In §1, I introduce the notion of ground along with
the kinds of circles of ground I will be considering. In §2, I present several different
reasons to motivate the need to think about circles of ground more seriously. In §3,
I discuss possible metaphysically substantive reasons to deny that anything can be
self-dependent. Both historically and contemporarily, philosophers have expressed
worries over the ontological priority ordering, bootstrapping, and the connection
between self-dependence and the necessary and the divine. In §4, I turn to a con-
sideration of explanatory reasons to avoid circles of ground. I discuss connections
between circularity, non-well-foundedness, and viciousness, along with the thought
that circles of ground are unacceptable for the more (deceptively) humdrum reason
that they give rise to trivial and uninformative explanations. I conclude that the most
salient reasons we have for supposing grounding is irreflexive are explanatory rather
than metaphysical, and that reasons to reject or accept instances of reflexivity need to
be assessed with a greater eye to other of our commitments.
1
See Keefe 2002 for a discussion of the problems with circular accounts.
 grounding and reflexivity
1 Grounding and Circularity

Broadly, we can distinguish two different approaches to the notion of ground. The first
approach, amongst whose proponents are philosophers such as Kit Fine and Fabrice
Correia, hold what we can call the sentential connective view.2 According to this view
locutions such as ‘because’ (on certain appropriate uses), ‘makes it the case that’ (on
certain appropriate uses) and ‘in virtue of ’ (on certain appropriate uses), connect
sentences. Importantly, although it can be convenient to talk in terms of facts, we
need not take the view as committing us to the existence of facts. Take the sentence
‘the table exists because its legs exist’. We could well recast this sentence in the language
of facts, but we need not. And even if we did, we need not take the sentence to commit
us to their (the facts’) existence.
The second approach, as endorsed by philosophers such as Trogdon and Schaffer
does involve a commitment to a realm of facts.3 This approach, which we can refer
to as the relational view, holds that real relations of ground obtain between facts. In
understanding grounding in this way, we are committing both to the existence of facts,
and to the existence of relations of ground that hold between them.
Here I understand the relational view to align with what I will refer to as the property
transfer view: relations of ground transmit a property.4 There are a variety of kinds of
properties that the grounding relation might transmit: modal properties are amongst
them. It is at least prima facie plausible to suppose that some fact will inherit its
modal status from the facts upon which it depends. The property-transfer view, as I
understand it, though, involves a commitment to the idea that there is some property
that is the property that the grounding relation transmits. Candidates for the kind of
property at issue are the properties of existence (or being), reality, and truth.5
Grounding is involved with metaphysical explanation. We say that where the fact
that it is raining grounds the fact that it is raining or it is not raining, it also, in a
distinctively metaphysical sense, explains it. Similarly, we might think that where the
physical facts ground the mental facts, or the natural facts ground the moral facts,
they also explain them.
One reason advocates of the relational view cite in defense of their commitment to
real relations is that there are good reasons to suppose that successful explanations are
always backed by real relations.6 For advocates of this approach it is because grounding
2 Fine 2012 and Correia 2010.

3 Schaffer 2009 and Trogdon 2013a. Schaffer believes that the relata of grounding relations can be drawn
from all ontological categories, and is committed to the existence of whatever it is that we take real relations
of ground to obtain between.
4 Not all proponents of the relational view explicitly endorse the property transfer view. One could
advance the relational view without holding the property transfer view; although it seems natural to suppose
that there is some property or other at issue where relations are instantiated. See also Trogdon (Chapter 9,
this volume) for a further elaboration of this view.
5 Bliss 2013, pp. 406–8, mentions this way of understanding the grounding relation. Morganti 2014 also
discusses grounding in these terms.

6 Kim 1994.
ricki bliss 
is involved with explanation that we ought to assume it to be a real relation.7 Advocates

of the sentential connective view, on the other hand, tend to think that grounding just
is metaphysical explanation. To say that some fact grounds another just is to say that
it metaphysically explains it. All parties agree, however, that grounding is intimately
involved with a distinctively metaphysical kind of explanation.
We can distinguish between cases of full ground and cases of partial ground. The
fact that I am happy partially grounds the fact that I am happy and rich, whereas the
fact that I am happy fully grounds the fact that I am happy or rich. The fact that I
am happy, and the fact that I am rich, together fully ground the fact that I am happy
and rich.
We come now to the matter of circles of ground. Circles of ground can be achieved
in a number of different ways. Suppose we were to introduce a further distinction
between mediate and immediate ground. Intuitively, whilst the fact that I exist is
immediately grounded in facts about the existence of my vital organs, it is only
mediately grounded in facts about the existence of fundamental properties, for
example. It is to this relation of mediate ground that philosophers commonly refer
when they talk of grounding. This relation, as we already know, is transitive. The
relation of immediate ground, however, is not.
With these two conceptions of ground in mind, we can imagine a situation in which
grounding loops are formed by chaining together instances of the immediate relation,
where these chains double back on themselves. As the relation is not transitive, these
chains do not collapse into themselves, avoiding reflexivity. Alternatively, circles of
ground can be formed where the relevant notion of ground is symmetric. Consider
the relationship between the north and south poles of a magnet.8 The fact that the
north pole exists is grounded in the fact that the south pole exists, and the fact that
the south pole exists is grounded in the fact that the north pole exists. The two poles
are symmetrically dependent upon one another.9 Where the relation is transitive,
such cases will also yield instances of reflexive dependence. Where the relation is not
transitive, they will not.
It is with the possibility of reflexive instances of ground that I shall be concerned
in this paper. Although there is much that can be said on circles of ground more
generally, I restrict myself here to a discussion of circles yielded by reflexive instances
of the relation.
2 Why Circles of Ground Matter

One might wonder why we should bother thinking about circles of ground at all.
Perhaps more pertinently, one might worry that the line of inquiry pursued in this
7 See Thompson (Chapter 5, this volume) for a reason to doubt that grounding involves real relations.
8 Priest, 2014, p. 178.
9 Note, this dependence is not merely nominal. The poles of a magnet are, by their natures, symmetrically
dependent upon one another. Note also that the fact that the poles are symmetrically dependent may well
be grounded in further facts about the existence and/or nature of magnetic fields.
 grounding and reflexivity
paper is simply illegitimate. After all, isn’t it just true by definition that grounding is
asymmetric, transitive, and irreflexive?10
Let’s consider briefly the notion of well-foundedness. Grounding is commonly
described as a relation that is asymmetric, transitive, irreflexive, and well-founded.
But although we can stipulate that the relation is well-founded, we are still under
obligation to offer not only an elaboration of what we think that amounts to—an
underexplored topic in the literature—but also the reasons for which we believe that to
be the case.11 And we can see from exploring some of the contemporary literature that
philosophers have, indeed, felt the need to conduct just this kind of investigation.12
To the best of my understanding, then, no one infers from the claim that grounding
is, by definition, well-founded, that it is illegitimate to ask whether or not, or why, we
should believe it to be the case, that reality comports with our definition. As should
also be the case, I venture to suggest, with the property of irreflexivity. We can stipulate
that grounding is irreflexive and yet still wonder whether or not reality is like this.
And, arguably, where philosophers are emphatic that a relation cannot have a certain
property, we have all the more reason to suppose that there is a principled reason for
the view; and all the more reason to want to know what that is.
Matters here might seem to become complicated by the fact that some philosophers,
at least, are willing to grant the existence of a second, reflexive notion of ground. Fine,
for example, distinguishes between strict and weak ground, where the latter is not
only possibly, but necessarily, reflexive.13 For proponents of the distinction, it will not
be the case, then, that there is anything particularly bad about circles of ground, but
rather, that where there are such circles we must realize that we are dealing with a
distinctive flavour of the relation.14
But this distinction, for those who grant that it is there to be drawn, is itself
informative, for it appears to reveal assumptions both about the relation that can be
reflexive and about the one that we are told cannot be. We can wonder, for example,
if the weak relation is something like a degenerate cousin of the strict—irreflexive—
relation. If this is the case, what reasons have we to suppose that the strict relation is
the central notion of ground? Surely whatever these reasons are, they are informed by
the reasons we have to suppose that the strict relation cannot be reflexive. Does every
fact weakly ground itself? If so, are we to infer from this that the kind of grounding
relation that a fact bears to itself is somehow harmless? Assuming the relational view,
is self-grounding harmless because the relation fails to deliver some property to the
relevant fact? Or is it harmless because the relation does deliver a property to the fact,
10 I have met with this response on a surprising number of occasions.

11 See both Dixon 2016 and Rabin and Rabern 2016 for excellent discussions of what we might mean
when we talk about the well-roundedness of grounding.
12 See Bliss 2013, Morganti 2009 and 2014, Tahko 2014.
13 Fine 2012, pp. 51–3.
14 I have borrowed this expression from Fine 2012.
ricki bliss 
but does so unproblematically? Are we then to distinguish between a metaphysically

substantive relation of strict ground—say, a property delivering relation—and a soft,
or inert—non-property delivering—relation of weak ground? If so, is this because
there is something wrong with circles of ground after all? Drawing distinctions
between different flavors of the relation only piques, rather than satisfies, the curiosity.
In what follows, however, when attempting to ascertain what is so bad about circles
of ground, I mean to ask this of whatever relation we are told is not, or cannot be,
reflexive. Coming to understand why some particular relation of ground is necessarily
irreflexive will surely help us to understand grounding, the reality that it orders, and
why it is, if indeed it is, that some flavors of ground can be reflexive. Before turning
to an exploration of self-grounded facts, though, I first consider some further reasons
that might motivate us to think more carefully about circles of ground.
The first of these is that reality appears to be populated with example instances
of them.15 Fine provides us with a number of cases.16 Take the fact that everything
exists—[everything exists]. This fact, itself an existent, helps ground [everything
exists]. [everything exists] partially grounds [everything exists]. Or consider the
proposition <every proposition is true or false>. This proposition is, itself, either true
or false. If it is true it helps ground its own truth, so too if it is false. Similarly with
the sentence ‘every sentence is true or false’. Fine considers these cases to present us
with a kind of puzzle exactly because it seems undeniable that these are instances of
partial self-grounding, and yet, this situation is, according to fine, ‘an absurdity, given
that nothing can hold in virtue of, or partly in virtue of, itself ’.17
Jenkins, on the other hand, suggests that mind/brain identity theories may present
us with a host of instances of circles of ground.18 According to some versions of
mind/brain identity theories, token mental states are identical to token brain states.
Someone’s being in pain state x just is them being in brain state y. These theories may
also assume there to be an explanatory connection between facts about tokens of the
one and facts about tokens of the other. Understanding these theories in terms of a
notion of ground, we appear to have cases in which pain states are both explained by,
and identical to, brain states, thus giving us instances of self-dependence.
Paseau points out that it is conceptually possible that at least some things are self-
dependent.19 For this reason, he is critical of the tendency to build irreflexivity
(and asymmetry) into our definitions of dependence and ultimate ontological basis.
Moreover, he cites the existence of non-well-founded sets—self-membered sets—as
15 See Raven 2013, for a treatment of some purported cases of reflexivity.

16 17 18
Fine 2010. Fine 2010, p. 98. Jenkins 2011.
19 Paseau 2010, p. 171. Although Paseau talks in the idiom of ontological dependence, it is acceptable to
assume this point to also hold true in the case of grounding. This is because Paseau’s paper is a response to
Ross Cameron on fundamentality. In that paper Cameron also talks in the idiom of ontological dependence,
but he now considers the discussion to be one that subsumes the notions of ontological dependence and
ground. Cameron’s original paper was written during a time in which it was not commonplace in the
literature to distinguish between relations of ontological dependence and relations of ground.
 grounding and reflexivity
examples of actual cases of self-dependence. And Graham Priest not only suggests that
it is possible that everything is partially self-grounded, but he argues that everything
is actually partially self-grounded.20
Changing tack now, a further compelling reason to take self-dependence seriously
is that a theory of reality on which some things are self-dependent may be, all
things considered, better than one on which they are not. Consider metaphysical
foundationalism: there is a derivative metaphysical superstructure that is grounded
in a fundamental ontological ground. This is far and away the most commonly
held view amongst metaphysicians who embrace a notion of ground. Whether that
which is fundamental is singular—as in some kind of monism—or plural—as in
pluralism—the fundamentalia/um grounds/gives rise to/explains everything else. But
why suppose there is something fundamental? Why not be metaphysical infinitists, for
example?
Ross Cameron argues that although we have no good reason to suppose that
metaphysical foundationalism is necessarily true, we do have good reason to suppose
that it is contingently true of the actual world.21 All things considered, claims
Cameron, a theory that affords a unified explanation of the phenomenon under
consideration is more virtuous than one that does not. Metaphysical foundationalism,
claims Cameron, is just such a theory: positing the existence of fundamentalia allows
us a unified explanation of the contents of reality.
Putting aside questions of what it means for a theory to be unified, amongst
other issues, what this line of argumentation brings to our attention is the role
considerations of theoretical virtue can play in our theorizing about the fundamental
structure of reality. We might then wonder whether a theory that leaves a large number
of things out (i.e. the fundamentalia) is ceteris paribus better than one that does not.22
Is a theory that posits unexplained fundamentalia obviously more virtuous than a
theory that posits nothing fundamental, but in which everything is explained? Maybe
it is. At least maybe it is if denying the existence of something fundamental ushers
in infinitely long grounding chains. But maybe it isn’t. Perhaps a theory that posits
the existence of quasi-fundamental phenomena—fundamental in the sense of being
metaphysically ‘rock bottom’—but then allows that they are self-explanatory, or self-
dependent does better. This way, everything has an explanation—nothing is left out—
but we are not forced to violate considerations of quantitative parsimony, for example.
I’m not suggesting that this view is correct, but just that it is not prima facie incorrect.
And whether or not the theory is apt to have its virtues considered would seem to
turn, at least partly, on whether or not anything can be self-dependent.
20 Priest 2014, p. 179. Although he describes everything as interpenetrating with itself, given other of
Priest’s assumptions, this claim is tantamount to stating that everything is (partially) grounded in itself.
21 Cameron 2008.
22 Foundationalism does not purport to explain the fundamentalia and one might worry that it is unfair to
criticize a theory for failing to explain that which it never attempted to explain in the first place. Nonetheless,
we can still weigh theories against one another and evaluate them on their virtues.
ricki bliss 
A final reason to think we ought to think more carefully about the possibility of cir-
cles of ground comes from noting the parallels between issues in foundational episte-
mology and issues in foundational metaphysics.23 In foundational epistemology, one
can be an epistemic foundationalist—the dominant view—an epistemic coherentist,
or an epistemic infinitist. All three views, although varying in degree of popularity,
are established, developed views. It is a striking feature of foundational metaphysics
that metaphysical foundationalism is the only view that is well-developed. Some
philosophers have defended the possibility of metaphysical infinitism, but metaphys-
ical coherentism—species of which would admit circles of ground—remains wholly
unexplored.24 Owing to the similarities between the relation of ground as it orders
reality—on some views at least—and the relation of justification as it orders our
beliefs, we may have good reason to think that where epistemic coherentism gets
going, metaphysical coherentism might just get going as well.
3 The Problems
Jenkins suggests that the term ‘dependence’ is quasi-irreflexive.25 What she means by
this is that ‘ . . . it always sounds bad to say “x metaphysically depends on x” or “x
grounds itself ”’.26 One reason that reflexive statements of ground might sound bad
is if they are just plain false. Suppose I claim that I exist because I exist. I would seem
to have uttered a falsehood. I exist because, amongst other things, a certain sperm met
with a certain ovum, for example. But exactly whether or not it is necessarily false—
indeed false at all—to utter statements of this form is exactly what is under issue. We
cannot assume that such statements are necessarily false as a premise in an argument
to the conclusion that reflexive grounding claims are necessarily false. Moreover, we
cannot infer from a single case—or class of cases—that no one may be suggesting are
instances of self-dependence in the first place, that nothing can be self-dependent at
all. This would be like contemplating the number seven, believing it not to have any
causes and inferring from this that nothing whatsoever is ever caused.
Let’s assume that the truth of statements of ground place demands on reality—
this is just what I mean when I claim that grounding talk is metaphysically laden.
Broadly construed, this just means that for any statement of ground to be true, the
world must be thus and such a way. On some views, as we have seen, this will also
involve an additional commitment to a binary relation of ground and the facts that
it holds between.27 Reflexive statements of ground will come out as false where the
entity under consideration does not happen to ground itself; and necessarily false if it
is unacceptable, or impossible, that anything ground itself.
23 See Bliss 2012, and Morganti 2014 and (this volume), for discussions that draw out some of these
similarities.
24 See Morganti 2009 and Schaffer 2003 for two different discussions of the possibility of infinitism.
25 26
Jenkins 2011, p. 1. Jenkins 2011, p. 2.
27 For the suggestion that relation needs not be binary, see Jenkins 2011 and Schaffer 2012.
 grounding and reflexivity
So why might it be unacceptable or impossible that anything ground itself? One

reason could be that if something’s being self-grounded commits us to a further
metaphysically substantive thesis regarding the nature of the self-grounding fact, or
the reality that it structures; and whatever this thesis commits us to is impossible or
unacceptable. What I have in mind here, in the case of the facts themselves, is that
where some fact grounds itself, how that fact is such that it is the kind of fact that can
ground itself is (i) non-trivial, and (ii) problematic.
The metaphysically laden view could then be either what we might think of as
substantive or shallow. On the shallow view, the truth of ‘x because x’ demands nothing
more of the world other than x grounding itself. We might think of something’s being
self-identical in just this same way: it is metaphysically laden—it is about the world—
but shallow—there is no further interesting story to be told about why things are such
that they are self-identical. This is just how things are.
Owing to the relationship between grounding and explanation, perhaps there will
be additional explanatory reasons for denying that grounding can be reflexive. Such
reasons may turn out to be the main locus of problems for the shallow approach to
self-dependence. In particular, we might worry that where a grounding tree contains
a loop, that tree will be non-well-founded, issuing in a vicious infinite regress. This
generates an explanatory concern for the simple reason that where a grounding tree is
non-well-founded, we fail to explain anything within that tree at all. Alternatively, we
might wonder if the appearance of tight grounding loops is unacceptable just because
the explanations that they issue in are trivial, uninformative, and explanatorily useless.
On this last approach, we require nothing metaphysically untoward in order to have
good, albeit epistemic, reasons to think grounding is irreflexive.
At this stage, I would like to address a worry that some of these suggestions might
raise. I can imagine an interlocutor objecting that the metaphysical explanations
associated with grounding are objective and thus devoid of any epistemic or cognitive
dimensions. According to this objection, it would be a mistake to entertain anything
that looks like an epistemic reason to reject reflexivity.
I think this objection is wrong-headed. First, how, exactly, we are to understand
the notion of a metaphysical explanation is far from clear. In the grounding liter-
ature to date, there is a tendency to point to the connection between grounding
and metaphysical explanation without elaborating upon how we are to understand
the relevant notion of explanation. Of course, one can stipulate that metaphysical
explanations are purely objective, but without good reasons to think this is the case,
such a stipulation does not help us very much. Importantly, though, the idea that the
objective nature of an explanation entails that it is altogether devoid of any epistemic
dimension is a mistake: a mistake that likely turns on conflating objectivity with
mind-independence.28 To claim that an explanation is objective is not to claim that
28 Consider money. Money is not a feature of mind-independent reality, and yet there are plenty of
objective facts about it: it is important to our lives, the source of much evil, and so on. Mind-independence
and objectivity are not one and the same thing.
ricki bliss 
it is mind-independent. The suggestion that our cognitive lives play a role in our
explanatory praxis does not necessarily threaten the objectivity of explanations.29
In order for objectivity to be threatened, we would need to have our cognitive lives
playing very particular kinds of roles in our accounts of explanation, and not just any
old role whatsoever.
Second, in light of our epistemic innocence regarding the reasons for which we are
to reject the possibility of reflexive instances of ground, it is incumbent upon us to
explore the field of options before us. If it turns out that the best reasons we have to
reject reflexivity are epistemic rather than metaphysical, and we are wedded to the idea
that there is no place for our cognitive lives in our theory of metaphysical explanation,
then we need to revise our commitment to irreflexivity. Alternatively, if it turns out
that the best reasons we have to reject reflexivity are epistemic, we may need to revise
our understanding of metaphysical explanation.
Grounding, we are so often told, is about reality. In fact, grounding is assumed by
many to be the relation that structures reality. Even on the sentential connective view,
we are still urged to understand the (relevant) sentences, and the connectives that join
them, as expressing truths about reality. If this is correct then we might expect that
reality itself provides us with reasons to suppose there is something wrong with circles
of ground.
3.1 Metaphysical worries

At various places in the literature, both historical and contemporary, philosophers
have gestured at what purport to be metaphysically laden reasons for supposing that
nothing can be self-dependent.
In the second of his Five Ways, Aquinas states the following:
The second way is from the nature of efficient cause. In the world of sense we find there is an
order of efficient causes. There is no case known (neither is it, indeed, possible) in which a thing
is found to be the efficient cause of itself; for so it would be prior to itself, which is impossible.30
The notion of efficient causation in operation here is to be understood in terms of

the medieval notion of causation per se. Causes per se exist contemporaneously with
their effects and sustain them. In this respect at least, this notion of efficient causation
is similar to our notion of ground. Aquinas is suggesting that it is impossible that
anything is the efficient cause of itself, for it would then be prior to itself. But why
is this? Were the notion of causation at issue here one on which causes temporally
preceded their effects, this objection might make sense, for it seems correct that
nothing can exist prior to itself in time. But the dependence relation we are concerned
with is not like this: grounds, and that which they ground, exist contemporaneously.
29
See Trogdon 2013b, p. 473 for a similar observation.
30
Aquinas 1964, Part 1, Question 2, Section 3.
 grounding and reflexivity
So the worry cannot be that things would have to exist before they exist in order to
cause themselves to be.
The notion of priority also plays a role in contemporary discussions of the structure
of reality. Putting aside Aquinas exegesis, the contemporary worry is best understood
against the backdrop of a commitment to the idea of reality as hierarchically struc-
tured, or composed of levels. Regimenting the worry in the language of facts, it might
be that any fact that is self-grounded would exist prior to itself within this structure:
it would reside at two different levels of the structure at the one time.
If levels are individuated by their occupants, however, two levels that contain all
of the same occupants would be indistinguishable. The real problem, then, seems to
be that reflexive instances of ground conflict with the very idea of a priority ordering
itself.31 Where grounding is reflexive, there is no ontological hierarchy and with it, no
priority. But as a reason to suppose that grounding cannot be reflexive, this objection
will not do. We cannot argue that grounding cannot be reflexive because it conflicts
with the notion of priority, where that priority is achieved by way of a relation that is
asymmetric, transitive, and irreflexive. That is as good as saying that grounding cannot
be reflexive because it cannot be reflexive. Of course, one could argue this, were one
to be in possession of independent reasons to believe this to be the case. And what
some such independent reasons are is exactly what the present exploration is hoping
to uncover.
A related worry, or perhaps just another way of presenting the same worry, is
in the language of relative fundamentality.32 One might worry that exactly what an
irreflexive notion of ground is introduced to capture is the idea that some things
are more or less fundamental than others: where grounding is reflexive—or non-
reflexive—this notion is no longer captured by our notion of ground. One might think
that just as nothing can be larger than itself, nothing can be more fundamental than
itself. It seems right that this is true—nothing can, presumably, be more fundamental
than itself. But such a line of reasoning goes no way towards justifying why we are
supposed to think anything is more or less fundamental than anything else in the
first place. Stipulating that some things are more or less fundamental relative to
other things does not provide us with reasons to suppose that reality exhibits such
a structure in the first place. As exactly what we are after is a reason to suppose that
grounding is necessarily irreflexive, the appeal to relative fundamentality goes no way
towards helping us uncover such a reason.
There is an issue here that I would like to pause and address. Some philosophers
claim that the very reason that we need a notion of ground is that we need a relation
that captures the hierarchical structure of reality (a notion that captures relative
31 Paseau 2010 objects to Cameron’s (2008) use of the term ‘ontological priority’ as the converse of
‘ontological dependence’ for he claims that it ‘…encourages the assumption, ex vi termini, that both relations
are irreflexive’ (footnote 5, p. 171). The notion of priority carries with it that of irreflexivity.
32 Thank you to an anonymous referee for pointing out the need to address this issue.
ricki bliss 
fundamentality). These philosophers will claim that we already have relations at our
disposal that are symmetric and reflexive, such as supervenience and identity. What
work is grounding to do at all, if not to capture this hierarchical structure, thinks
the philosopher of this ilk. The relationship between grounding and supervenience
is complicated, and I make no attempt to discuss this here.33 The most oft-cited
advantage that ground has over supervenience is that it allows us to capture the
asymmetry of dependence: where supervenience picks out modal companionship,
ground allows us to say what depends on what. But the advantage a notion of
ground has over that of, say, supervenience, isn’t simply that it captures asymmetric
dependencies. Ground is just, I take it, a richer notion. When wielding a notion of
ground, we are not just suggesting that two facts are yoked to one another across
possible worlds, we are pointing to the fact that two facts are related in a particularly
rich and important way. Grounds metaphysically explain that which they ground, in a
way that the supervenience base does not explain what supervenes upon it.34 The
point is just that ground plus reflexivity does not necessarily yield supervenience.
Of course, it remains to be seen whether anything can explain itself in any kind
of interesting way, but ground remains an, in principle, richer notion than that of
supervenience nonetheless. Admitting that some things can ground themselves is not
necessarily just to reintroduce supervenience in ground’s clothing.
Returning to the central discussion, the question I am interested in addressing is
why is it that some instances of ground, in all its richness, cannot be reflexive? One
very good reason to think that grounding is necessarily irreflexive is that reflexive
instances of ground are unacceptable or impossible. But absent such independent
reasons, the kinds of objections we have seen so far are useless.
Fortunately, an example of an independent reason to think that no fact can ground
itself has been alluded to in recent discussions: nothing can be self-grounded because
anything that exists in this way would have to bootstrap itself into being.35 This
objection sounds like it is in keeping with the spirit of the worry conveyed by Aquinas,
but it is also as enigmatic. Again, the relevant relation is non-temporal, so self-
dependent things would not need to exist before themselves in time in order to
bring themselves into existence. Indeed, grounding doesn’t seem to be involved with
bringings of things into being or existence at all.
Perhaps what the bootstrapping objection is getting at is that it is just plain weird
to think that something could be its own ground. I confess that I find this concern
a little bit difficult to appreciate. A pervasive view amongst metaphysicians is that
there is a fundamental level populated by brute, independent fundamentalia: they are
ungrounded and without explanation.36 But suppose, now, for argument’s sake, that
33 See Leuenberger 2014 for a recent and interesting discussion on grounding and supervenience.
34 See McLaughlin and Bennett 2014.
35 I am not aware of anyone who has stated this in print. I have, however, heard this concern expressed
on a number of separate occassions.

36 Cameron 2008, Schaffer 2009.
 grounding and reflexivity
our fundamentalia turn out to be self-dependent. It is far from clear how something
bootstrapping itself into being (whatever that turns out to mean) is any weirder, or any
more or less explanatory, than something that gets flung into existence from nowhere
and for no reason whatsoever. In the particular case of the contrast with brute facts
or existents, it is not clear why one view is more acceptable, or less weird, than the
other. And this is all the more so the case where we understand our fundamental
facts to be contingent. Whilst necessary facts, one might think, do not stand in need
of explanation, contingent facts seem to be exactly the kinds of facts that do need
explanations! Weirdness to which we have become accustomed—as in the case of
brute, contingent facts—is still weirdness after all. And if it’s weirdness that we wish
to avoid, then the proponent of brute, contingent facts also has some work to do.
One way that we might try and make something of the bootstrapping worry is in
light of the property-transfer view: where a fact grounds itself, the relation transfers
some property from that fact to itself, and this is unacceptable. But, again, why? The
success of this objection would seem to turn on the idea that properties—even if
only of a certain kind—have to originate from somewhere else. But why is this? The
contrast with brute, contingent facts here is, again, pertinent—why is the idea that
something lends a property to itself any more bizarre than the idea that something
plucks a property out of thin air? After all, we seem perfectly willing to countenance
that there are some things that don’t borrow certain, important, properties from
anything else—the fundamentalia—so why can’t some facts garner certain properties
from themselves?37 At the very least, we would seem to have a problem of superfluity.
If something already possesses a property, such that it is available to lend it to itself, we
can wonder why we need to go to the trouble of invoking that thing and its property
possession to explain where that property came from in the first place. If some fact
already possesses a particular property, there is no need to invoke the grounding
relation it bears to itself to explain where that property comes from. But is there
anything more serious or troubling going on than a potential explanatory breakdown?
Perhaps the real problem for this view is if there is some way that facts need to be
such that they are the kinds of facts that can ground themselves, and however this is is
unacceptable. That is to say, a good reason to deny the possibility of reflexive instances
of ground is if it commits us to a further, metaphysically substantive, yet unacceptable,
thesis regarding the natures of the entities that are involved.
There is historical precedence for the positing of self-explanatory facts. According
to Leibniz, if God exists, He exists because He exists. God’s existence, or the fact
that God exists, is self-explanatory. But far from being trivial, God’s existence is self-
explanatory because it is in His essence to exist. There is, then, a further, metaphys-
ically substantive, story to be told regarding God’s self-explanatory existence. And
importantly, this is not the end of the story. It follows from God’s essential existence
37 In claiming that the fundamentalia are ungrounded, no one wishes to deny that they have being, for
example.
ricki bliss 
that He is also a necessary being, as essential properties are necessary properties; and
that self-explanatory facts about God’s existence are necessary facts. There is a route
from some fact’s being self-explanatory to its being necessary and it is via the necessary
existence of its constituents: the kinds of facts that could be self-explanatory facts,
then, would be existence facts. Moreover, once we are in the business of presenting
explanations for how some fact is such that it can ground itself, we might think that
the explanatory demand that got us this far also obliges us to present an explanation
for why it is that God is the kind of being that can exist essentially.38 After all, I have
lots of my properties essentially—such as being human—but I am excluded from
having the property of existence in this way. Leibniz proposed that the appropriate
explanation here is in terms of God’s divinity: God has the property of existence
essentially because He is perfect.39
Putting the pieces together, we might wonder if the problem with self-grounded
facts is that they would be (i) necessary facts, where (ii) the subjects of those facts are
essential existents, which are (iii) in possession of divine properties. Of course, the
idea of a necessary fact is not particularly controversial. What would be controversial,
however, is a picture of reality on which the fundamental facts are necessary. The
reason for this is that where the fundamentalia are necessary, and everything else
follows from them by necessity—assuming necessitarianism about grounding—then
the only world that exists is the actual world. All contingency drops out of the picture
and we are left with full-blown necessitarianism.40
First (iii) and the connection between self-groundedness and the divine. It was
wielding a Principle of Sufficient Reason (PSR) that Leibniz was driven to invoking
God’s divinity as an explanation for why it is that God exists essentially. Without
such a principle, however, we are not compelled to make the additional step from
essential existence to the possession of divine properties. The proponent of the notion
of ground is not normally wont to make appeal to a Principle of Sufficient Reason, and
without it, I see no reason to suppose that one would be forced to account for why it
is that some things exist in a certain way, namely, essentially, whilst others do not. For
anyone who did make use of a PSR, however, things may look a little bit different.
Nonetheless, even where some explanation or other of why it is that some thing exists
essentially is warranted, are we really forced to frame such an explanation in terms
of divine attributes? Perhaps there are other kinds of explanations available to us that
allow us to explain why it is that some things exist essentially and others not?
38 As Schopenhauer 1974 wittily remarks on the Principle of Sufficient Reason, ‘the causal law therefore
is not so accommodating as to let itself be used like a hired cab, which we dismiss when we have reached
our destination. . . . ’
39 Dasgupta 2016 also notes the potential to draw this connection between necessary facts and the divine,
but he dismisses it in the context of his broader project. The reason for this, so far as I understand it, is that
there are examples of necessary beings, such as numbers, that we have no reason to suppose are divine (see
esp. p. 26).
40 See Dasgupta 2016 for a riveting discussion of how things would be at such a world.
 grounding and reflexivity
Let us turn, now, to (i) and the purported connection between self-groundedness
and the necessary. A move that could be made at this stage would be to question the
very idea that any individual can have the property of existence essentially. If existence
is not a property (of individuals), as has been suggested by the likes of Kant and Frege,
then no individual can have it essentially. If this is the case, however things need to
be, such that facts about their existence are self-grounded, it cannot be because the
subjects of our existence facts have existence as an essential property. But for anyone
who does accept that existence is a property, surely this is not an option. Another
possibility would be to take issue with the notion of an essence: nothing can exist
essentially because nothing has an essence whatsoever. Again, however facts need to
be such that they are self-grounded, it cannot be because the subjects of those facts
exist essentially.
More promising, I believe, would be to question the connection between self-
dependence and essential existence altogether (ii). Recall that the reason we were
discussing the connection between self-grounded facts and essential existence was
because we were exploring an extant account of how the world needs to be such that
some facts ground themselves. But does citing God’s essential existence really yield a
situation in which the fact that God exists explains itself? Does it not yield a situation
in which the fact that God exists is explained by the fact that God has it in His essence
to exist? We appear to have two separate facts here and not, actually, a case of one
grounding itself.
Absent the connection between self-dependence and essential existence, have we
any other reason to suppose that self-grounded facts force us to commit to a metaphys-
ically substantive thesis that is otherwise undesirable? In particular, is there any other
route from self-dependence to necessity that is not mediated by a thesis regarding the
essential existence of the subjects of our self-grounded facts? It is not clear to me that
there is. Even if we wished to argue that necessary facts are self-explanatory, there is no
obvious entailment from a fact’s being self-grounded to its being necessary. Moreover,
if David Lewis is right, there may even be examples of self-explanatory facts that are
contingent. Loops formed where a time-traveler visits himself with instructions on
how to build a time-machine involve a series of contingent, self-explanatory facts.41
3.2 Explanatory worries
Far more salient, in my view, are the explanatory reasons for finding the prospect of
circles of ground troubling. I will canvas two clusters of explanatory worries, along
with arguing that the issues are both more complex and less disturbing than they
may initially appear. The first set of worries involve infinite regresses, and the second
explanatory failure.
Grounding chains, or trees, that contain loops may well be non-well-founded.
A reason to reject the possibility of circles of ground, then, is that they would commit
41
Lewis 1976.
ricki bliss 
us to the possible existence of (downwardly) non-terminating grounding structures.42

Whether or not this is a compelling reason to reject the possibility of reflexive
instances of ground will turn upon how powerful the reasons are for supposing that
grounding chains must (downwardly) terminate.
Chief amongst the arguments often offered in defense of fundamentality are
arguments from vicious infinite regress.43 Philosophers will claim that where our
grounding chains do not (downwardly) terminate we are off on an infinite regress and
the regress is vicious. But infinite grounding regresses are not, however, necessarily
vicious; sometimes they are benign.44 If this is correct, then the pertinent questions
to ask in the current context are, in fact, two: do reflexive instances of ground generate
infinite regresses, and if they do, are those regresses vicious or benign?
Reflexive instances of ground generate loops of the form [A] grounds [A]. These
loops could be expressed by sentences of the form ‘A because A’. Whether or not a
regress is generated at all, I venture to suggest, will depend upon what we think is
required a regress to make. Consider first that such a tight loop contains only one
relatum, [A] and a single instance of a relation, G. How are we supposed to generate an
infinite regress from this? Of course, where [A] grounds itself, [A] follows from itself,
but it does not follow from this that there are, say, an infinite number of instances
of [A], and an infinite number of instances of the grounding relation that obtain
between [A] and itself.45 From what materials, exactly, are we supposed to construct
our regress?
From a single relatum and a single instance of the grounding relation we can,
however, draw an infinite number of explanatory inferences should we choose to; each
at a later moment in time. In such a case, we could generate an infinite sequence
of sentences ‘A because A. A because A. A because A . . . . ’ Why one would wish
to continue on in this way, however, is somewhat difficult to imagine. But that one
could continue on in this way seems to be reason to suppose that an infinite regress of
sorts can be generated. Suppose, though, that we are willing to grant that an infinite
explanatory regress can be generated, we are then compelled to establish whether it
would be vicious or benign.
Consider the fact that X—[X]. Suppose also that this fact is a dependent fact.
As dependent facts have grounds [X] must also have a ground. Suppose that what
grounds [X] is [Y]. In citing [Y], we have, in some sense of explains, explained [X].
Where [Y] is itself a dependent fact, we might worry that although [Y] explains [X], it
cannot completely explain [X]. This is because [Y] also stands in need of explanation.
Supposing that our grounding chain is infinitely descending, at each stage of the
42 Dasgupta 2016, p. 6 also notes this possibility.

43 The role regress arguments play in defenses of fundamentality are, in my view, often underestimated.
I suspect this might be because of their cryptic and commonly terse nature. For examples of regress
arguments in defense of fundamentality, see Fine 2010 and Schaffer 2012, for example. For a discussion
of how to understand regress arguments in the context of the notion of ground, see Bliss 2013.
44 45
Bliss 2013. Bliss 2014, §3 and Gratton 2010, esp. ch. 4.
 grounding and reflexivity
regress, although we have said something of the facts at the level above, we may not
have explained everything for which we think we need an explanation.46
Although there is no hope of adequately covering the thicket of issues associated
with arguments from vicious infinite regress here, enough can be said, I hope, to
satisfy the reader. Arguments from vicious infinite regress require the inclusion of
an assumption that stipulates an explanatory target; which is exactly what we fail
to reach where a regress is generated. And the very same regress can be considered
vicious or benign depending on the explanatory target we have in mind.47 Where that
for which we are seeking an explanation is some dependent fact or other, that fact
is, presumably, explained by the facts upon which it depends. Regresses generated
after this fashion are benign. Regresses are not benign, however, where what we wish
to explain is something that no member of the regress, or the entire regress taken
together, can explain. Good candidates for such explanatory targets are: why there are
any dependent entities whatsoever; why everything turned out this way rather than
that way;48 or the expectation that dependent entities have complete metaphysical
explanations. There is not the space to address all of the relevant issues here, but it is
enough to recognize that in order to establish whether a regress is vicious or benign,
we need to be clear on what our explanatory target is in the first place.
In cases in which some fact purportedly explains itself, though, it might appear
that, in citing that fact as its own explanation, we have explained nothing about that
fact whatsoever. Let’s, for argument’s sake, assume as much. Importantly, we need
not be embarked upon an infinite number of moves through our very tight loop to
make such an observation. The explanatory failure that may be noted at infinity is an
explanatory failure that presents itself having moved through the loop but once. The
infinite explanatory regress that could be generated is, therewith, superfluous, for the
problem presents itself at the very first stage: explaining some fact in terms of itself
is as good as offering no explanation of that fact at all. What does this mean, though,
for cases in which the entire structure has a self-grounded fact at the tip of one of its
branches? Can something that has no explanation serve as the terminus point of such
a structure? The standard view is that it certainly can. Exactly what the metaphysical
foundationalist thinks is that there is a fundamental level of ungrounded facts that
explain everything else. Once again, at the very least, it is not clear that self-grounded
facts are any worse off than ungrounded, fundamental facts.
46 If grounds do not explain (or ground) that which they ground, I have no idea why anybody bothers
talking about grounding in the first place. To deny that grounds do any work at all seems preposterous
(for anyone who cares to employ a notion of ground, that is. There may be good reasons to wish to deny
grounding talk altogether. My target with this comment is the person who (i) embraces a notion of ground,
but (ii) tries to claim that grounds don’t do any work).
47 See Aikin 2005 and Bliss 2013 for discussions of the regress problem in these terms. See also
Passmore 1970.
48 See Dasgupta 2016 for an account of fundamentality in these terms.
ricki bliss 
This brings us to the second, and final, cluster of explanatory worries that I
will consider. The most conspicuous form that a reflexive explanation can take is
‘A because A’. This explanation of A in terms of itself would seem to be trivial,
uninformative, and explanatorily useless. We might suppose that it is for the reason of
explanatory failure that philosophers express an aversion to the possibility of circles
of ground.
At first blush, these explanatory worries seem convincing. It does, indeed, seem
to be the case that ‘A because A’ is trivial, uninformative, and explanatorily useless.
Trivial explanations are bad, we might think, because they are uninformative and
explanatorily useless. But are explanations of this form necessarily trivial? Earlier in
the discussion, we saw that although on some accounts it is true to say of God ‘He
exists because He exists’ it is not the case that God’s existence is trivial. Although I
argued that the way God is, such that we can say of Him that He is self-explanatory,
seems problematic, that is not to say that there may not be some other way that
self-dependent things are, such that they are self-explanatory. Far from being trivial,
what the appearance of a reflexive metaphysical explanation may alerts us to is that
something is such that it is able to explain itself.
Alternatively, what the appearance of a reflexive metaphysical explanation could
alert us to is simply that we have reached an explanatory dead end. Reflexive meta-
physical explanations let us know that we have arrived at a breakdown in our
explanatory progress. One reason for this breakdown might be that we have arrived at
some fact that simply does not stand in need of further explanation. Good candidates
for such facts may be facts about essences.49 I would feel confused if someone asked
me to explain why it is in my essence to be a human. And this is not because I think
there is some further story to be told, but happen not to know what it is. Rather, it is
because there isn’t anything else to say. Arguably, all necessities might involve us in
explanatory dead ends of this kind. Or perhaps a reflexive explanation could just be
taken to indicate that we have arrived at something that may well have an explanation
but we don’t, given our current state of knowledge, happen to know what it is. There
are a variety of different reasons for which our explanatory progress can be halted.
Perhaps we do not achieve the kind of explanation we are looking for, but this does not
mean that we have not explained anything whatsoever. In this sense, then, reflexive
instances of ground could be metaphysically laden—they are about the world—but
shallow—there is no special way that self-grounded facts have to be in order to ground
themselves.
There is a further issue, or set of issues, that complicate these matters though.
Assume the sentence ‘S is in pain state x because S is in brain state y’. This sentence, in
fact, involves two sentences joined together with the sentential connective ‘because’.
It suggests an explanation of a token pain state in terms of a token brain state.
49 These kinds of facts are facts that Dasgupta 2016 describes as autonomous—they stand in no need of
further explanation.
 grounding and reflexivity
We can represent this explanation more simply as ‘A because B’. Sentences express
propositions, but it would be a mistake, however, to suppose that the foregoing
sentential version of our explanation can necessarily be rendered into its propositional
form as <B> explains <A>. According to identity-based reductionism, token pain
states and token brain states are (contingently) identical. As names denote referents,
‘pain state x’ and ‘brain state y’ share a referent.
On a broadly Russellian view of propositions, propositions are built up out of
worldly entities. The propositions <Sally loves Louis> is built up from Sally and Louis
and the loving relation. Sally and Louis enter into the proposition directly. As the
expressions ‘brain state x’ and ‘pain state y’ are co-referring terms the two sentences
in our explanatory statement above express one and the same proposition. This means
that the explanans and explanandum of our explanation are identical. On a Russellian
view of propositions, token identity theories that are also explanatory will yield
technically circular explanations.50 On a Fregean view of structured propositions they
will not. According to broadly Fregean views, we can distinguish between senses and
referents, where structured propositions are composed of senses. Two terms that share
a referent may diverge in their senses. What this means in the case of identity theories
that involve reductive explanations is that, as the terms ‘pain state x’ and ‘brain
state y’ differ in their senses, they express different propositions. On a Fregean view
of structured propositions, our explanations in these cases are not even technically
circular.
Identity theories are controversial, and it is a matter of debate whether they
can afford us any kind of explanation whatsoever. The problem generalizes beyond
mind/brain identity theories, however. Many scientific explanations are also thought
to involve identities: Water = H2O and genes = DNA molecules, and yet we use
the latter to explain the former all the time. Whether or not we even have reflexive
explanations in these cases will depend, in part, on how we fine grain what we assume
to be our explanatory relata.51 Whether or not we are dealing with genuine cases of
explanatory circularity, and the reasons for which they may be trivial, will be sensitive
to very many of our theoretical commitments.52
50 In her discussion of how we might avoid reflexive metaphysical explanations in cases where we say
that pain states are grounded in brain states, Jenkins 2011 suggests that we go hyperintensional. How
successful this route is, however, will depend upon what we understand the relata of explanations to be
and what, where our relata are propositions, the metaphysics of those propositions are. Suggesting that we
‘go hyperintensional’ is not yet sufficient to deal with the problems of circularity.
51 For the proponent of the view that real relations of ground obtain between facts, establishing where
we have genuine instances of reflexivity will involve clarifying both the relationship between propositions
and facts, and the particular metaphysic of facts or propositions employed.
52 Ruben also notes that certain facts, on certain accounts, will yield problems with reflexivity. Recog-
nizing this trouble he states, ‘…explanation is not just a relation between facts as constituted by worldly
particulars and their properties, apart from how they are conceptualized. If P=Q, the fact that x is P and the
fact that x is Q introduce the same feature. What matters in explanation isn’t only property introduction,
but the way in which we conceptualize the property, viz. whether the property P is introduced as property
P or as property Q’ (Ruben 1990, p. 176).
ricki bliss 
4 Concluding Remarks
Philosophers are generally adamant that grounding is necessarily irreflexive. What
the reasons are for this commitment, however, are not often clear. One might suppose
that there are a host of metaphysical reasons for supposing that grounding cannot
be reflexive. I have explored variations on what I understand to be the most salient
metaphysical reasons to reject the possibility of tight grounding loops and found them
wanting. I have also explored some explanatory reasons to suppose that grounding is
necessarily irreflexive. Although these reasons are more compelling, this result has
the upshot that our cognitive lives play a far greater role in what counts as a good or a
bad metaphysical explanation than friends of the notion of ground would commonly
like to acknowledge. Reflexive explanations are bad because, in some cases, they are
trivial and uninformative. That said, we need to be tremendously careful because the
charge of triviality will be sensitive to other of our theoretical commitments; and our
explanatory demands will play an important role in fixing what counts as an adequate
metaphysical explanation.
References
Aikin, S.F. (2005), ‘Who is Afraid of Epistemology’s Regress Problem’, Philosophical Studies,
vol. 126, no. 2, pp. 191–217.
Aquinas, T. (1964), ‘The First Three Ways’, Summa Theologica, Blackfriars, pp. 13–15.
Bliss, R.L. (2012), ‘Against Metaphysical Foundationalism’ (unpublished PhD dissertation,
University of Melbourne).
Bliss, R.L. (2013), ‘Viciousness and the Structure of Reality’, Philosophical Studies, vol. 166,
no. 2, pp. 399–418.
Bliss, R.L. (2014), ‘Viciousness and Circles of Ground’, Metaphilosophy, vol. 45, no. 2,
pp. 245–56.
Cameron, R. (2008), ‘Turtles All the Way Down’, Philosophical Quarterly, vol. 58, no. 230,
pp. 1–14.
Correia, F. (2010), ‘Grounding and Truth-Functions’, Logique et Analyse, vol. 53, no. 211,
pp. 1–29.
Dasgupta, S. (2016), ‘Metaphysical Rationalism’, Noûs, vol. 48, pp. 1–40.
Dixon, T. Scott (2016), ‘What is the Well-Foundedness of Grounding?’, Mind.
Fine, K. (2010), ‘Some Puzzles of Ground’, Notre Dame Journal of Formal Logic, vol. 51, no. 1,
pp. 97–118.
Fine, K. (2012), ‘Guide to Ground’, in Fabrice Correia and Benjamin Schnieder (eds), Meta-
physical Grounding: Understanding the Structure of Reality, Cambridge University Press,
pp. 37–80.
Gratton, C. (2010), Infinite Regress Arguments, Springer.
Jenkins, C.S. (2011), ‘Is Metaphysical Dependence Irreflexive?’, The Monist, vol. 94, no. 2,
pp. 267–76.
Keefe, R. (2002), ‘When Does Circularity Matter?’, Proceedings of the Aristotelian Society, New
Series, vol. 102, pp. 275–92.
 grounding and reflexivity
Kim, J. (1994), ‘Explanatory Knowledge and Metaphysical Dependence’, Philosophical Issues,

vol. 5, pp. 51–69.
Leuenberger, S. (2014), ‘From Grounding to Supervenience’, Erkenntnis, vol. 79, pp. 227–40.
Lewis, D. (1976), ‘The Paradoxes of Time Travel’, American Philosophical Quarterly, vol. 13,
no. 2, pp. 145–52.
McLaughlin, B. and Bennett, K. (2014), ‘Supervenience’, The Stanford Encyclopedia of Philoso-
phy, ed. Edward N. Zalta: https://plato.stanford.edu/entries/supervenience/.
Morganti, M. (2009), ‘Ontological Priority, Fundamentality and Monism’, Dialectica, vol. 63,
no. 3, pp. 271–88.
Morganti, M. (2014), ‘Dependence, Justification and Explanation: Must Reality be Well-
Founded?’, Erkenntnis, vol. 60, no. 3, pp. 1–18.
Paseau, A. (2010), ‘Defining Ultimate Ontological Basis and the Fundamental Layer’, The
Philosophical Quarterly, vol. 60, no. 328, pp. 169–75.
Passmore, J. (1970), Philosophical Reasoning, Duckworth.
Priest, G. (2014), One: Being an Investigation into the Unity of Reality and of its Parts, Including
the Singular Object which is Nothingness, Oxford University Press.
Rabin, G.O. and Rabern, B. (2016), ‘Well Founding Grounding Grounding’, Journal of Philo-
sophical Logic, vol. 45, no. 4, pp. 349–79.
Raven, M. (2013), ‘Is Ground a Strict Partial Order?’, The American Philosophical Quarterly,
vol. 50, pp. 191–9.
Ruben, D. (1990), Explaining Explanation, Routledge.
Schaffer, J. (2003), ‘Is there a Fundamental Level?’, Noûs, vol. 37, no. 3, pp. 498–517.
Schaffer, J. (2009), ‘On What Grounds What’, in David Manley, David J. Chalmers, and
Ryan Wasserman (eds), Metametaphysics: New Essays on the Foundations of Ontology, Oxford
Schaffer, J. (2012), ‘Grounding, Transitivity and Contrastivity’, in Fabrice Correia and Benjamin
Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality, Cambridge
Schopenhauer, A. (1974), The Four-fold Root of the Principle of Sufficient Reason, trans.
E.F.J. Payne, Hackett.
Tahko, T. (2014), ‘Boring Infinite Descent’, Metaphilosophy, vol. 45, no. 2, pp. 257–69.
Trogdon, K. (2013a), ‘An Introduction to Grounding’, in Varieties of Dependence: Ontological
Dependence, Grounding, Supervenience, Response-Dependence, Basic Philosophical Con-
cepts, Philosophia Verlag.
Trogdon, K. (2013b), ‘Grounding: Necessary or Contingent’, Pacific Philosophical Quarterly,
vol. 94, no. 4, pp. 465–84.
4
Cosmic Loops
Daniel Nolan
1 Introduction
Those of us interested in thinking about outré possibilities will be familiar with
scenarios where there are large temporal and causal loops—for example, scenarios
where time goes in a loop, so that, for example, a big crunch is immediately followed
by a big bang. (I intend here a “one time around” loop, as opposed to the kind of eternal
recurrence where there are infinitely many bang-to-crunch stretches, laid end to end.)
In these scenarios, there are temporal loops and causal loops, but only ones that go
all the way around the history of the universe. One example of these, of more than
just metaphysical interest, is the closed temporal loop universe described by Gödel
1949, which appears to show that such temporal loops are allowed by Einstein’s general
theory of relativity.
Scenarios that are less familiar are ones where there are cosmic grounding loops:
where the whole structure of grounding ensures that if you follow the chain around
from any point, after enough steps you can arrive back where you started. In this paper
I want to distinguish several interesting ways of thinking about such grounding loops,
argue for the coherence of such models of grounding, consider whether they are meta-
physically possible, and discuss how we might embed grounding structures which are
locally irreflexive, anti-symmetric, and transitive in worlds with such cosmic loops.
Any loop of grounding, of course, enables one in principle to trace it around and
get back to the start. What is distinctive about cosmic loops is that they would require
going around “the whole way”, in a way that is analogous to the way that a cosmic
temporal loop would require going through every other time to arrive back at the
original time. The nice thing about times is that, when they are well behaved, they
come with a complete ordering, but this is not true in general for objects that stand in
grounding relationships. So it is a bit harder to say what “going around the whole way”
would amount to for a grounding loop. It would be convenient if everything came
with a grounding “level”, as is supposed by some simple versions of the “layer cake”
model of the special sciences: chemistry on top of physics, biology on top of chemistry,
psychology on top of biology, and so on. Then we could insist that a cosmic loop of
 cosmic loops
ground pass through all of the levels before coming back to the original one. Other
patterns in the world come with convenient layers that are less all-encompassing: the
relation of part-to-whole can be used to order my fingernail as part of my finger, my
finger as part of my hand, my hand as part of me. On its own, it will not serve as a
convenient way of ordering everything, since there are distinct hierarchies of parts:
my table leg is not part of my leg, nor vice versa. We would have a cosmic loop of
part-to-whole if we started with one world (call it world 1) which had many atoms at
one end of the part–whole hierarchy, and at the other end of the part–whole hierarchy
a Universe that contained everything as parts, and considered another world, world 2,
with the same pattern of part-to-whole except that the thing which was the Universe
of world 1 was part of all the things which were atoms of world 1. In world 2, you
could follow the chain of “part of ” relations starting at the object which is world 1’s
Universe, right around to that very object again. World 2 would plausibly contain a
cosmic grounding loop too, given the common assumption that wholes are grounded
in their parts. (Perhaps world 2 would only be an impossible world, rather than a
possible one: more on this question in Section 3 below.)
While I have hopefully said enough to get the idea of cosmic loops across, I have
not yet provided a general definition. Rather than bogging down in a specification that
avoids various tricky corner cases, I will present some exemplars which we may use as
paradigms: especially since the issues that arise for my exemplars don’t really depend
on whether we have pinned down a unique concept of cosmic loops. One thing I do
want to leave open, at least as far as the definition of “cosmic loop” goes, is that cosmic
loops of ground might co-exist with shorter loops of ground. Again, time provides a
useful analogy: even if the entire universe is a great temporal loop, say with a big bang
at the “start” also serving as a big crunch at the “end”, there may also be shorter loops
created by time-travel machines or unusual spatio-temporal wormholes. Likewise,
even if there are cosmic loops of ground that go “all the way around”, there may also
be short loops (e.g. the fact that there are some facts may ground itself1 ). I also want
to allow that a loop can be cosmic without bringing everything in a universe into its
scope: a layer-cake universe might have several cosmic loops that contain a member
from each layer, but do not share any members.
When we are considering cosmic loop scenarios, which loops will be grounding
loops will depend on what kinds of relationships go along with relationships of
grounding in those scenarios. I suppose that we could brutally stipulate grounding
connections between different entities or facts, but it will be more natural, and more
familiar, to think of grounding as going along with other relationships, such as the
part–whole relation or the determinate–determinable relation. (Though there are of
1 See Fine 2010, though of course Fine himself is not tempted to allow that this fact grounds itself. It
is instructive to see how difficult it is to avoid allowing it to be a ground of itself, if we make some other
standard assumptions about grounding.
daniel nolan 
course debates to be had about which direction grounding goes even in these cases:
part-to-whole, or whole-to-part, or sometimes one and sometimes the other, for
example.)
Rival theories of grounding differ on whether grounding claims are most perspic-
uously to be expressed using a sentential or propositional connective, or a relational
predicate. That is, if we wish to express a particular grounding connection to do with
being scarlet and being red, whether G(Apple A is scarlet, Apple A is red), or G(A’s
scarletness, A’s redness) best gets to the heart of the matter, assuming determinates
ground determinables. I will talk as if grounding is a relation between objects in this
paper, but this for convenience rather than to take a stand on this question. I will also
not be making much of the distinction, often drawn, between full and partial ground:
some cases I will discuss below are best seen as loops of full grounding and others
only of partial grounding, but little relevant will hang on which are which. Finally,
I will restrict my discussion to talk of singular grounding, instead of also talking about
cases where some things collectively ground another (or some things are collectively
grounded by a thing, or when some things collectively ground some others): this is
not to take a stand on whether there is any irreducibly plural grounding, but again
only because that distinction is not important for current purposes.
Warning: well-brought up readers of this paper are likely to have been taught that
no sense can be made of talk of loops of grounding, cosmic or otherwise, so may find
the cases to be discussed below repugnant to their grounding sensibilities. I would
encourage those readers to do their best to get their heads around the cases, perhaps
in the spirit of intellectual exploration of foreign conceptual landscapes. I will turn to
discussing whether any of these examples are possible, coherent, or even conceivable
in Section 3.
2 Examples of Cosmic Loops

One of my favourite thought-experimental curiosities is a universe described by Rudy
Rucker in his Infinity and the Mind (Rucker 1982, pp. 33–4).
2.1 The Rucker Loop
What appears to be our entire universe is just a sub-atomic particle in a larger universe,
which is but a sub-atomic particle in a yet larger “universe”, and so on ad infinitum.
This is also true in the other direction: what seem to us now to be our smallest sub-
atomic particles have the internal structure of an entire “universe”, the sub-atomic
particles of which are entire “universes” themselves, and so on ad infinitum. What is
distinctive about Rucker’s thought is that this world also loops: go up through enough
stages and you will arrive back at one of our sub-atomic particles, or go down through
enough stages and you will reach our entire universe.
 cosmic loops
Rucker focuses on aspects of this imagined world like it having no absolute scale
from smallest to largest (nothing is once-and-for-all the smallest or the largest, for
example), and the prospect that it could nevertheless contain finitely many objects,
despite, for example, everything being divisible without end. But the Rucker Loop
suggests an interesting pattern of grounding, as well. It is often thought that a whole
is grounded in its parts: and when there is a loop like this, that suggests that there is a
loop in grounding. Even if we reversed this grounding connection, so that the parts of
our cosmos are all grounded in the cosmos, we would get a loop of grounding—our
cosmos grounded in the one “above”, grounded in the one “above” that . . . grounded
in our cosmos. Furthermore, we can suppose the loop (or the many loops) are all-
encompassing—that no cosmos lacks a step in the loop, and that we have to go through
a cosmos of each other level before arriving back at the cosmos we began with. Let us
focus on one of the loops in this world, that begins and ends with our familiar cosmos.
This loop is cosmic in the sense I have in mind for this paper.
Another cosmic loop of the part-to-whole relationship that has been discussed in
the literature is one suggested by a story of Borges (Borges 2000, originally published
1949). In Borges’s story, he describes an object, “the Aleph”, which, on one reading,
has everything in the universe as a proper part, even though it itself is a small globe
found in a cellar in Buenos Aires. (On another reading, the Aleph merely provides
a viewpoint on everything. Borges notes this reading within the story, suggesting
that the true object that contains everything else in the universe may be a pillar in
Cairo. I suppose it could be contested whether the part-to-whole loop goes all the
way around mereological levels in this case, but Borges seems to describe at least a
near-cosmic-sized loop.) Sanford 1993 and Parsons unpublished both discuss Borges’s
Aleph, on the interpretation where the Aleph does contain everything as a part (and so
looking into the Aleph, one even sees the Aleph itself within its basement, containing
within itself the whole universe . . .). They both find it worthwhile to try to make
coherent sense of it as a possibility. Parsons further seems to suggest that if the Aleph
is genuinely possible then the part–whole relationship is not anti-symmetric and
transitive. I am not sure of Parsons’s reasoning here, but perhaps he is using “anti-
symmetric” in a way that a relation is anti-symmetric only if necessarily the relation
does not relate an object to a distinct object and also vice versa.
Once the option of cosmic loops is noticed, it is easy to come up with other
examples. Here are two examples that may be of use as thought experiments, or as
pieces of speculative theology for those who are so inclined.
2.2 The last shall be first
In this scenario, there is a god—let me label Her TLSBF (for “the last shall be first”).
TLSBF is both immanent and transcendent in the following way. (Leave aside any
quibbles for now about whether this characterization is strictly “immanence” or
“transcendence” in the senses used in, e.g. Christian theology.) TLSBF is within in the
smallest places: let Her be a proper part of each space–time point, or if you prefer let
daniel nolan 
physical space–time be gunky with no atomic physical parts, with Her being a proper
part of every region. We could also directly specify that She is located within every
point (or every gunky region), or perhaps her being parts of those points and regions
will be enough, on some conceptions of location, to already guarantee this. Thus She
is immanent in her world. (We may add that She is also part of every physical object
too, if you wish.)
TLSBF is at the “bottom”. But She is also at the “top”. There is a region which has
all other regions as sub-regions (the “universal region”), and TLBSF is located at that
region. There is an entity which has everything in the universe as parts (as is standard
in most theories of parts and wholes), and that universal entity is identical to TLSBF
Herself. Let us explicitly include all the space–time regions among her parts. Finally,
let us stipulate that in this scenario, entities are grounded in their proper parts: so
TLSBF is grounded in her parts, and there is a chain of grounds that lead from TLSBF
to Herself.
Let us restrict our attention to concrete objects, and leave aside questions about
the grounding of abstract objects (if any) in our scenario. The TLSBF scenario is
incompatible with classical extensional mereology, which does not allow an object
to be a proper part of itself (or indeed to stand in the ancestral of “proper part” to
itself—since classical extensional mereology insists that “proper part” is transitive, in
ruling out one it rules out the other). Indeed, even much weaker mereologies may
rule out this scenario unless we can find some other part of space–time points to be
co-parts of those points with TLSBF.
2.3 The One and the cosmos: emanating and constituting

Another class of cosmic loop scenarios come into view if we pay attention to the option
of saying that there are several kinds of grounding (whether this is because grounding
is a genus which admits of various species, or because grounding, though unified,
holds in different kinds of cases).
Consider a world which is in one respect rather neo-Platonist. The One is the
ultimate source of emanation, and this relationship passes through Soul, Wisdom, and
other such luminaries, down through Forms, through the Intelligences that are to be
found throughout the world, and finally to brute material entities, furthest from the
One. Emanation goes with grounding, so that, for example, Wisdom is immediately
grounded in the One, the Soul is immediately grounded in Wisdom . . . and the small
material parts of intelligences are grounded in the intelligences they emanate from.
(I don’t suppose that this specific emanation structure matches that posited by any
particular neo-Platonist, but you should be able to tinker with the emanation structure
somewhat without affecting the point of the example.)
On the other hand, the smallest material parts make up the objects they belong to;
those larger material objects constitute the Intelligences, the natures or activities of the
Intelligences constitute the Forms, the Forms compose the Soul, which constitutes
 cosmic loops
Wisdom, which is the constitution of the One itself. Wholes are grounded in their
parts, and the constituted by what constitutes it (at least in the scenario being
described), so in this respect, grounding runs from the lowest to the highest.
This is not the same kind of loop as in the previous two cases. In the other two
cases, the cosmic loop followed a circle: while we could pick our universe as the place
to start and end in the Rucker Loop, it occupied no particularly privileged place, and
while TLSBF served both as a proper part of space–time points and as the Universe,
we could follow the entire loop around by going from part to whole at each step. In
this case, however, we have two grounding arrows facing in opposite directions: part-
to-whole and constitution going in one direction, and emanating coming from the
other. To follow grounding around to get a loop requires going all the way up and
then all the way down again.
A variant of this case can be imagined that would have a hybrid kind of loop. In this
variant, the meanest of the material particles, furthest along the path of emanation
from the One, each directly constitute the One and so directly and fully ground it.
(Perhaps they do this by being simple, and therefore they are the ones that give rise
to the One. Perhaps our hypothetical neo-Platonist constructing the account of this
scenario has been meditating on the second half of Plato’s Parmenides.) This loop con-
nects most saliently at the One, which directly grounds Wisdom through emanation,
and is directly grounded by each of the ones through constitution. While grounding
goes in a circle in this variant, none of the relations that go along with grounding do:
emanation and part–whole are one-way only, as is the “direct constitution” link from
the material simples to the One.
Imaginative readers will probably be able to think of other interesting scenarios
containing cosmic loops, but the three examples above should be enough to illustrate
the idea and give some idea of the range of cosmic loop scenarios. The three cases
presented are all cases taking entities to be grounded by other entities: those interested
in expressing grounding using a sentential connective in the manner of Fine 2001
should be able to construct further scenarios where there are cosmic loops of such
grounding without involving any cosmic loops of grounding between entities, but I
have stretched our theoretical imaginations enough for one paper, so I will refrain
from exploring any options of that sort here.
3 Cosmic Loops and Principles of Ground

Cosmic loops, on the face of it, conflict with some standard constraints on a theory of
ground put forward in the literature. Ground is normally defined so that it is transitive,
asymmetric, and irreflexive, which would rule out loops: any loop would result, by
transitivity, in something grounding itself. Furthermore, the correct principles of
ground, whatever they are, are often thought to be necessary. (At least metaphysically
necessary, though sometimes these principles are discussed as if they are partially
daniel nolan 
definitional of ground, so may be intended to be analyticities as well.) Can cosmic

loops be dismissed as impossibilities?
I would be tempted to argue that such loops are possible in some generous sense,
since descriptions of them are coherent.2 But even if they are not possible at all (in
any worldly or alethic sense, as opposed to being, e.g. doxastic possibilities), I do not
think this would justify an immediate dismissal of any discussion of cosmic loops. One
main reason for this is that we are interested in deciding what to think about which
principles of ground are correct, so even if alternatives to the true theory of ground are
all metaphysical impossibilities, working out which theory of ground is correct may
well require us to judge between alternatives to select the best one. In metaphysical
inquiry, as elsewhere, dogmatic rejection of alternative theories as even being fit for
discussion would be a terrible methodology, since it is often only by appreciating what
alternative theoretical options there are to one’s preferred views that we can work out
whether our current opinions are better than alternatives, and so whether they are
worthy of our continued belief.
Those uncongenial to these cosmic loop scenarios might doubt that they are even
coherent. Anti-symmetry and transitivity are often put forward as if they are axiomatic
of grounding, so some may suspect that it is a conceptual truth (or perhaps an
analytic truth) that there are no loops of ground.3 Perhaps there is some concept
of a grounding-like relation that, by conceptual stipulation, is anti-symmetric and
transitive. But I doubt that such a concept is a very useful tool for investigating the
world. One of the central aims of a theory of grounding, I would have thought, would
be to discover what sorts of fundamental (and less-than-fundamental) metaphysical
relationships obtain between entities (and/or what connections, more broadly speak-
ing, hold between facts). If the important relationships in our world display the sort
of loop structure suggested, a theory of ground should reflect that: it would be far less
fruitful to declare that we have discovered there is no grounding, but there is merely
grounding*, a relation that is found where we thought grounding might be, with all of
the features of grounding except (e.g.) transitivity. Substantial metaphysical progress
is not to be made by analytic stipulation, so we should select our conceptual tools with
an eye to what can be used to illuminate our target of inquiry, rather than to try to bake
in some of our favoured conclusions about that target in advance. Those who insist
that according to their concept of grounding we can rule out cosmic loops of ground
as incoherent are invited to deploy a concept better suited for substantive inquiry.
2 They are possible in at least some of the ways I distinguish as candidates to be “metaphysical possibility”
in Nolan 2011, though perhaps not in all.

3 A similar concern can be raised about whether it is a conceptual falsehood that the part–whole relation
allows of loops: van Inwagen 1993, in response to Sanford’s Aleph example (see p. 94, this chapter), takes the
line that the Aleph case involves a conceptual falsehood. I am tempted by a similar response in the case of
part–whole as I am in the case of grounding: I would argue against the conceptual truth of, e.g. asymmetry
and transitivity of the part–whole relation, just as I argue against elevating principles of grounding to
conceptual truths in the main text.
 cosmic loops
Some readers may find cases of cosmic loops so bizarre or outrageous that they
may doubt that cosmic loops are even conceivable. (I intend to use “conceivable” in its
ordinary sense, or something like its ordinary sense, and not in any of the stipulative
senses introduced by philosophers such as Chalmers 2002.) As against this, it is hard
to know what to offer in response beyond the plain fact that I and others do conceive
of such scenarios—Rucker seems to have conceived of one of the scenarios above,
I came up with two of the scenarios above myself, and I hopefully described them in
enough detail to get across what is going on in them, at least to those not antecedently
committed to finding such scenarios unintelligible. Perhaps some familiarity with
neo-Platonist emanation will help for scenario 3. No doubt there will be those who
suspect neo-Platonist emanation is unintelligible on its own: those people face an
interesting psycho-historical challenge in explaining how hundreds of people over
hundreds of years seemed to communicate and debate about emanation, without any
of them conceiving of it.
Why would there be resistance to the claims that these scenarios are intelligible or
imaginable or conceivable? One source of such resistance will be from people who
think that conceivability is a good guide to possibility (Yablo 1993), and who also
judge the scenarios I described to be impossible. Once one thinks that one’s grasp of
possibility is usually mediated by conceiving, it will be easy to pass from the thought
that something seems impossible to the thought that it must be inconceivable. While
being able to conceive of something often goes along with its possibility, trying to
insist on too tight a connection either leaves one at the mercy of counterexamples
to be found everywhere from philosophical theorizing to Escher to the far reaches
of speculative fiction, or encourages a dangerous attitude that the thing to do with
an alternative one takes to be impossible is to try to convince oneself that one does
not understand it. That would be an unhelpfully dogmatic move in many areas of
science—imagine if opponents of the general theory of relativity had all reacted that
way—and it seems no less dogmatic in philosophy.4
Another source of resistance will be less motivated by theory: I expect some people
will find it difficult to understand the scenarios described, and not (necessarily) due to
any defect in my presentation. One the face of it, one might have thought that people
would take their own inability to conceive something as very weak evidence that it
is inconceivable, especially when there are others who apparently conceive of the
scenario under discussion. In some areas, this does seem to be people’s response: those
who find they cannot conceive of relativistic space–time, for example, are often willing
to defer to experts who claim to conceive of it, and so count relativistic space–time as
conceivable. But it is a curious fact that philosophers who have trouble conceiving of a
4 A third option would be to allow that many more things are possible than one might have thought,
just because we can form some conception of them: see Mortensen 1989. But why engage in a large revision
of our views of what is possible rather than a minor revision of a theory of the connection between
conceivability and possibility?
daniel nolan 
scenario proposed by other philosophers are often keen to pronounce such scenarios
inconceivable. (This often happens to me in conversation with philosophers when I
claim to conceive things others claim they cannot, at least.)
To those inclined to take these cases to be inconceivable for this reason, let me
remind them that familiarity with unusual scenarios can be mind-expanding, and to
play with cosmic loop scenarios for a while before being confident that the scenarios
are inconceivable, and not merely ruled out by principles that they endorse.
4 Recovering “Local” Irreflexivity, Symmetry,

and Transitivity in Cosmic Loops
A scenario can be a cosmic loop scenario even if grounding is closed under transitivity
in it: these are cases where everything in a circle of ground grounds everything in that
circle, including itself. But there is a more natural way to understand many of these
circles of ground as being intransitive: while A grounds B which grounds C which
grounds D which grounds E which . . . grounds A, these are not scenarios where A
grounds itself or is somehow a causa sui. Or at the very least, this seems plausible
for many of the entities in these loops: maybe TFSBL or the One are most naturally
thought of as self-grounders, but entities in the “middle” of each loop, a given human
hand, for example, are not naturally thought of as self-grounders.
Even aside from this natural thought, it will be interesting to explore what the
options are here for recovering local irreflexivity, asymmetry, and transitivity in
cosmic loop scenarios. That is, to what extent can we “save the appearances” and allow
that even if, on some cosmic scale, there is a loop of grounding, we need not change
our attitudes to the grounding relationships that hold, for example, between the cells
and other components of my hand and my hand itself, or between the distribution
of rain, clouds, and lightning, on the one hand, and a thunderstorm, on the other?
Can things as we ordinarily take them to be be embedded in a cosmos containing one
or more cosmic loops at scales we are unfamiliar with? (Compare: in a universe with
a unique big crunch that is immediately before its unique big bang, the direction of
time might still be locally one-way, with no small loops letting people live through
2014 before 2013.)
What would “local” mean in this context? One stab at characterizing it would be to
say that grounding is locally irreflexive, asymmetric, and transitive iff when we restrict
the domain of entities quantified over to some domain D, then for all x in D, x does not
ground x, for all x and y in D, if x grounds y then y does not ground x, and for all x, y,
and z in D, if x grounds y and y grounds z, then x grounds z. Then we should insist on
some restrictions on the appropriate D so that it is appropriately “local”. We would be
aiming to capture the idea that with a certain “distance”, grounding behaves as if it is
irreflexive, asymmetric, and transitive, and cases where there are loops of ground only
show up when we look at “long distances”. The challenge then is to specify the relevant
 cosmic loops
domains D that are “local” to each other, or alternatively to specify a “distance” so that
any entities within that distance of a given object O count as belonging to the same
domain D as O.
One way to pursue the former strategy would be to find some independent way
of specifying domains and which objects share a common domain. Perhaps each of
Rucker’s “universes” could be its own domain, for example. One way to pursue the
latter strategy would be to say that a domain D is local if there are no more than n steps
of immediate grounding between any two members of D, for some suitably low n.
This would require that we rely on a notion of “immediate ground”, and find a way to
apply it to the grounding chains we are concerned with. Sometimes it is easy to see
what immediate ground would be: intuitively, the singleton of Socrates is immediately
grounded in Socrates, but the singleton of the singleton of Socrates is plausibly
immediately grounded only in the singleton of Socrates, and its grounding in Socrates
is only mediate. In other cases, though, it is harder to draw the distinction. Am I
immediately grounded in my cells, or only immediately grounded in objects such as
my brain and liver, and only mediately (partially) grounded in my liver cells? Or am I
immediately grounded in all my parts, down to the quarks and leptons? If we were to
apply notions of immediate and mediate grounding in one’s parts in the Rucker world,
for example, we would at least want it to turn out that I was not immediately grounded
in any of the galaxies that are parts of one of my electrons: though we may have to add
stipulations to the original thought experiment if we wanted to guarantee this.
While an account of locality that appeals to immediate ground might capture a
sufficient condition for a domain D to be of objects “local” to each other, it is probably
too restrictive, in several ways. One is that there may well be cases where there
is grounding, but no immediate grounding. This could be because some forms of
grounding do not lend themselves to an immediate/mediate distinction, and it could
also be because there may well be cases where a kind of grounding is in general
amenable to that distinction, but unusual cases defy categorization. Consider this
sort of structure: suppose that we have an infinite sequence where, for each finite
stage, each stage after the first is immediately grounded in the stage below. Suppose
now that this sequence has a first “infinite” member: if we were ordering the stages
by ordinals, we would assign that stage the ordinal ω. That stage may be plausibly
grounded in the stages that came before, but not plausibly immediately grounded in
any of them: there is no stage “immediately before” it in the series. One might even
think the ordinals themselves are like this. It is more usual to think that ordinals
are immediately grounded in all the ordinals that precede them (if “member of ”
corresponds to immediate grounding, and we accept the von Neumann definition
of ordinals), but orthodoxy here is not compulsory.
Another challenge the particular account of “locality” offered here faces, even
apart from any concerns about its relying on a notion of immediate ground, is that
it does not yet rule out gerrymandered “neighbourhoods”. A selection of a handful
of things that do not stand in any chains of grounding to each other will count as
daniel nolan 
a “neighbourhood”: one of my electrons, the singleton of Socrates, and Abraham

Lincoln’s last thought are an example of a three-membered domain that we may want
to rule out as one of the relevant domains D that we are defining locality over. On
the other hand, we do not want to insist that every member of D either grounds, or
is grounded by, every other: if we want these domains to include ones we typically
think about, we might want to include me, and all of my quarks and leptons, as well
as intermediate parts, in a single D, without insisting that each of my electrons either
grounds, or is grounded by, each of the others.
I will resist the temptation to go down the rabbit hole of developing and critiquing
different criteria we might have for selecting a domain D, and ensuring that each
domain shares a “locality” in an intuitive sense. Instead, I will turn to a different
challenge. To ensure that grounding can be “locally” irreflexive and asymmetric,
transitivity must fail somewhere in the cosmic loop—otherwise everything in the
loop will ground itself, for example, since we will be able to go from a thing back
to itself by steps of grounding. (A failure of irreflexivity is automatically also a failure
of asymmetry). The challenge then is to say how grounding could fail to be transitive
around the whole loop while being locally transitive, especially if we desire that it
is locally transitive everywhere: otherwise enough applications of local transitivity
might take us around the whole loop, provided the “locations” overlap.
There are a few ways to try to satisfy the demand for local transitivity in the face
of this need for a counterexample to transitivity somewhere in the loop. The most
conservative way would be to abandon the demand for local transitivity everywhere:
perhaps there are no counterexamples to transitivity in parts of the cosmos we are
familiar with, but the counterexamples occur somewhere else. A version of this
strategy that would work with case 3 would be to insist that grounding per se is
not transitive, but only the species of grounding are (in case 3, emanation and
constitution). In the second variant of case 3, for example, the obvious point where
the counterexample to transitivity of grounding would occur is from the ones to the
One and then to Wisdom, since the ones constitute the One but Wisdom only stands
in the emanation relation from the One.
Suppose we wanted to get closer to the idea that grounding was always locally
transitive. We could tinker with our account of what entities are “local” to which, so
counterexamples to transitivity only occur when entities do not share a locality (e.g.
if there were clear borders between cosmoi in the Rucker loop, perhaps two entities
would need to share a cosmos to be local to each other). Or we could wheel out more
high-powered philosophical resources. One traditional area where philosophers have
struggled with the conflicting desires to have a local inheritance principle that fails
over longer distances is in dealing with the paradoxes of vagueness: in the sorites
paradox, for example, we would like to hold onto the idea that subtracting one grain of
sand from a heap always leaves a heap, but also to the idea that subtraction of enough
grains of sand turns a heap to a non-heap. Likewise, if we want local transitivity
without full-strength transitivity, we would like the ground of a ground to always
 cosmic loops
be a ground, but there to be some number of iterations of the immediate grounding

relation that takes us from a ground to a non-ground, with it being vague where the
series breaks down. Perhaps resources developed to help with the sorites could be
employed to help with the marriage of local transitivity to a failure of full-strength
transitivity?
The literature on ways to resist the sorites paradox is vast, and so I will not try to list
all the available options here. Options include taking the transitivity principle to have
no false instances, but some instances that fail to be true (as with supervaluationist
approaches, for example); or to think that some instances of the transitivity principle
are almost fully true, and perhaps all steps involving immediate grounding are true
enough to assert, though the slow leakage of truth from antecedent to consequent in
each instance allows us to have a series of steps of x1 immediately grounding x2 , x2
immediately grounding x3 , and so on, without it being at all true that x1 grounds x1000
(as in fuzzy-logical treatments of vagueness). Both of these approaches compromise
the (full) truth of the general transitivity principle, while salvaging the absence of
some kinds of counterexamples—there will be no particular “break” in the chain to
be identified.
Other, slightly more exotic, options, would be to retain the full truth of the
transitivity principle but weaken our logical resources so that we cannot validly apply
it multiple times: just as we cannot validly reach the conclusion that a single grain
of sand is a heap, we will not be able to validly reach the conclusion that x1 grounds
x1000 , even if x1 grounds x2 , and x2 grounds x3 , and grounding is transitive. Ways
of doing something similar in the case of the sorites include the contraction-free
approach explored by Slaney 1988 and Restall 1994 ch. 8, and the intransitivity-
of-validity approach explored by Cobreros et al. 2015, among others. Yet another
option would be to adopt an approach that rendered at least the material version of
transitivity true while making it unsuitable for use in inferring the consequent from
the antecedent, by analogy with the subvaluational approach to vagueness explored in
Hyde 1997. I am inclined to think that any commitment to transitivity of grounding
would not be strong enough to motivate these sort of logical modifications to preserve
local transitivity of grounding: but those already keen on these resources, perhaps to
preserve tolerance principles in vague cases, may find it appealing to treat apparent
failure of transitivity in cosmic loop cases with similar resources.
Armed with an understanding of how to ensure local transitivity and asymmetry
without global transitivity in cases of loops of grounding (whichever understanding
we might adopt), we can apply the same resources to other relations of interest, includ-
ing those that appear to underpin grounding relationships. One thing that makes
Rucker loops so mind-bending, for example, is that they challenge our assumptions
about the part–whole relationship: that I, for example, could be a proper-part of a
proper-part of a proper-part of . . . myself. That would be impossible were the proper-
part relation both asymmetric and transitive. However, one thing that makes the
case less mind-bending than short loops of the proper–parthood relation (e.g. just
daniel nolan 
stipulating that A is a proper part of B and B is a proper part of A) is that in a

Rucker world with local transitivity and asymmetry the relation of proper-part to
whole would behave just as it actually does in cases we are familiar with: I am part
of the Milky Way galaxy, but the Milky Way galaxy is not part of me. This is not the
only way to conceive of a Rucker loop, of course: another way would be to conceive of
the Milky Way as being one of my parts, but just much further down a natural chain of
part-to-whole than one might have thought. But at least the option of retaining local
transitivity and asymmetry gives us one way to think of the Rucker loop scenario as
being one in which our ordinary particular judgements about what is part of what do
not need to be revised.
Similar devices could also be deployed if we wished to claim there was local near-
transitivity, near-asymmetry and near-irreflexivity of ground and of other notions.
After all, a number of authors have wanted to motivate exceptions to each of these
principles independently of anything to do with very long chains of ground. See,
for example, Jenkins 2011 on reflexivity, Barnes 2018, on symmetry, Schaffer 2012
on transitivity, and Bliss 2011 on all three, as well as many of the other papers in
this volume. One, perhaps inelegant, way to modify local transitivity to local near-
transitivity, for example, would be to say that except for such-and-such cases grounding
is locally transitive within a domain D. A more elegant way to specify local near-
transitivity would be to have a positive story about when, for entities among a given D,
it is the case that when x grounds y and y grounds z, x also grounds z. Even more
elegant would be such a principle that applies to all “local” domains D at once, rather
than separately specified principles about each D individually.
Suppose we do secure local transitivity (or something close to it) without requiring
transitivity tout court. What advantages could that offer a theory that postulated a
cosmic loop? One advantage is that grounding relationships would be more selective.
A grounding loop which is transitive requires everything in the loop to ground
everything else in that loop, which might sometimes seem counterintuitive: even if
both the Milky Way and an electron in it are part of the one Rucker loop, intuitively
the electron partly grounds the Milky Way and not the other way around. Perhaps we
should think that cosmic loops where grounding is transitive, and so everything in a
loop grounds everything else in that loop, are also conceivable and maybe possible:
but it is natural to think that not all grounding loops are like this, and perhaps not the
ones that most naturally come to mind when presented with cases like those given
in Section 2.
Another advantage follows if we antecedently thought that instances of grounding
we are familiar with appear to never relate entities to themselves, relate in an asym-
metric “direction”, and at least appear transitive. While there are many papers in this
volume that will argue that even grounding we are familiar with need not always be
like this, we retain the option of leaving our theory of the grounding relationships
between familiar entities as being traditional, while accepting (or leaving open the
possibility) that there are cosmic grounding loops outside our familiar domain. The
 cosmic loops
options for preserving local transitivity will also be valuable, apart from any doc-
trines about grounding, when dealing with other metaphysical relationships we are
tempted to think are asymmetric and transitive, such as the relation of part-to-whole.
The Rucker loop, for example promises to shed light on the conceivable, and perhaps
possible, options for mereology as well as for grounding.
Those suggesting philosophical innovations, or even scepticism about received wis-
dom, are often under pressure to “save the phenomena”: to explain why it seemed that
the old orthodoxy was right, or why we can often rely on generalizations or inferences
supported by the old orthodoxy. For example, the nihilist about tables and chairs owes
us a story of our apparent success in home decoration and lunch preparation, or the
dialetheist logician owes us a story about why classical mathematics seems to have
been such a success in the twentieth century while apparently relying on classical logic.
One way to “save the phenomena” is to corral exceptions to previous orthodoxies to
cases that are relatively unusual: classical physics can be used to build bridges or aim
cannons, because, for example, moving objects do not get appreciably more massive as
they speed up until they are close to the speed of light. Recovering “local” transitivity,
asymmetry, and irreflexivity for grounding is one way to show how the exceptions to
those principles do not show up in the cases we were most familiar with.
Grounding loops will appear exotic to many, but if a theory postulating a grounding
loop only offends our intuitions in cases far removed from those with which we are
familiar, then we may not wish to trust our intuitions very far about those cases. The
comparison with theorizing about causation may be instructive: while we are, in my
view, properly reluctant to abandon the view that rock throwings sometimes cause
window breakings or that stock market crashes cause unemployment, we are much
less certain about our ordinary causal generalizations and intuitions when considering
cases like quantum mechanical phenomena or the Big Bang. And rightly so: exotic
phenomena might behave exotically. To work out whether there are cosmic loops of
ground, or of part–whole, or other such relations, we would do well not to just trust
our off-the-cuff generalizations but to carefully investigate cases outside familiar ones.
5 Conclusion
Cosmic loops are of intrinsic interest: thinking about them can satisfy the same urges
to grapple with the unfamiliar which are satisfied by various sorts of speculative
fiction, from science fiction to the stories of Borges. Metaphysical fiction is a genre
in its infancy, but a promising one for all that.
I have argued that thinking about cosmic loops serves several more academic
purposes, however. They demonstrate, that we can make sense of loops of ground
in a different way from the usual examples of loops achieved through only a few
steps, and the conceivability and perhaps possibility of them are supported in ways
different from other arguments I know of to support failures of asymmetry and
transitivity. This should give us additional reason, were additional reason needed, to
daniel nolan 
admit the conceivability and consider seriously the possibility, that grounding need
not be transitive (and to a lesser extent, reason to take non-symmetry seriously, if
we think that some cosmic loops of ground are not counterexamples to transitivity).
Finally, through exploring options for recovering local transitivity (and so local
asymmetry and irreflexivity, should we want them), we can see that confidence about
grounding relations between familiar items should not lead us to overconfidence
about general principles of ground, no more than experiencing the local asymmetry
of the direction of time should lead us to assert dogmatically that cosmic temporal
loops are impossible.
Those who want to reject the possibility of cosmic loops, let alone those who reject
the coherence of them, would be well served to defend the principles they think rule
out such loops, rather than just taking those principles to be obvious or analytic. And
this applies just as much to cosmic loops of part-to-whole, or cosmic loops of any
other relation, as it does to cosmic loops of grounding. Metaphysics, with its hope to
be a completely general investigation of what there is, should be particularly wary of
the perils of overgeneralization.5
References
Barnes, E. 2018. “Symmetric Dependence”, in R. Bliss and G. Priest (eds), Reality and its
Structure: Essays in Fundamentality. Oxford: Oxford University Press, pp. 50–69.
Bliss, R. 2011. “Against Metaphysical Foundationalism”. PhD thesis, University of Melbourne.
Borges, J.L. 2000 [1949]. The Aleph and Other Stories. London: Penguin.
Chalmers, D. 2002. “Does Conceivability Entail Possibility?” in T. Gendler and J. Hawthorne,
Conceivability and Possibility. Oxford: Oxford University Press, pp. 145–200.
Cobreros, P., Egré, P., Ripley, D., and van Rooij, R. 2015. “Vagueness, Truth and Permissive
Consequence”, in T. Achourioti, H. Galinon, K. Fujimoto, and J. Martinez-Fernandez (eds),
Unifying the Philosophy of Truth. Dordrect: Springer.
Fine, K. 2001. “The Question of Realism”. Philosopher’s Imprint 1: 1–30.
Fine, K. 2010. “Some Puzzles of Ground”. Notre Dame Journal of Formal Logic 51.1: 97–118.
Gödel, K. 1949. “A Remark About the Relationship Between Relativity Theory and Idealistic
Philosophy”, in P.A. Schlipp (ed.) Albert Einstein: Philosophical Scientist. Library of the Living
Philosophers, La Salle, IL: Open Court Press, pp. 555–62.
Hyde, D. 1997. “From Heaps of Gaps to Heaps of Gluts”. Mind 106.424: 641–60.
Jenkins, C.S. 2011. “Is Metaphysical Grounding Irreflexive?” The Monist 94.2: 267–76.
Mortensen, C. 1989. “Anything is Possible”. Erkenntnis 30.3: 319–37.
Nolan, D. 2011. “The Extent of Metaphysical Necessity”. Philosophical Perspectives 25.1: 313–39.
Parsons, J. Unpublished. “The Earth and the Aleph”. http://www.joshparsons.net/draft/aleph/.
Accessed 16 January 2016.
Restall, G. 1994. “On Logics Without Contraction”. PhD Thesis, University of Queensland.
Rucker, R. 1982. Infinity and the Mind. London: Paladin Books.
5 Thanks to Sara Bernstein, Ross Cameron, Alex Sandgren, and two anonymous referees for feedback.
 cosmic loops
Sanford, D. 1993. “The Problem of the Many, Many Composition Questions, and Naive
Mereology”. Noûs 27.2: 219–28.
Schaffer, J. 2012. “Grounding, Transitivity and Constrastivity”, in F. Correia and B. Schnieder
(eds), Grounding and Explanation. Cambridge: Cambridge University Press, pp. 122–38.
Slaney, J.K. 1988. “Vagueness Revisited”. Australian National University, Automated Reasoning
Project Technical Report, TR-ARP-15/88.
van Inwagen, P. 1993. “Naive Mereology, Admissible Valuations, and Other Matters”. Noûs 27.2:
229–34.
Yablo, S. 1993. “Is Conceivability a Guide to Possibility?” Philosophy and Phenomenological
Research 53: 1–42.
5
Metaphysical Interdependence,
Epistemic Coherentism, and
Holistic Explanation
Naomi Thompson
1 Introduction
This paper develops an argument for metaphysical interdependence; an alternative
to orthodox foundationalist accounts of metaphysical structure as characterized by
grounding relations. Friends of metaphysical interdependence take facts to be related
in networks of grounding such that there might be no foundational facts, and that
a given fact can appear in its own grounding ancestry. Grounding is an explanatory
relation, and the need to recognize holistic explanations (and in particular, holistic
metaphysical explanations) generates a requirement for an account of grounding with
a holistic structure. Metaphysical interdependence is such an account.
After briefly introducing the notion of ground in §2, §3 outlines both the core of
the foundationalist approach, and that of metaphysical interdependence. §4 develops
an analogy between metaphysical interdependence and coherentism in epistemology.
§5 argues that grounding is to be thought of as an explanatory relation. In §6, the
view that grounding is an explanatory relation is considered against the backdrop
of different approaches to explanatory structure. In §7 I respond to some perceived
objections to holistic explanation. §8 concludes this chapter.
2 Grounding
I take grounding to be a relation of metaphysical dependence, which obtains between
facts.1 Grounding is said to be an explanatory relation, such that when some fact [A]
1 This point is contentious amongst friends of grounding, and I make no attempt to defend it here.
I talk of grounding as obtaining between facts merely in order to simplify the discussion, and not because
I think there are conclusive arguments for thinking that grounding relations do not relate entities of other
ontological categories. I take it that what I say here could also be applied to cases of grounding between
entities of other ontological categories with only minor adjustments.
 interdependence, coherentism, and holistic explanation
grounds a further fact [B], [A] explains [B]. Like explanation, grounding is a one-
many relation, such that some fact may have a number of grounds. Following Fine
(e.g. 2012: 50) I take [A] to be a partial ground for [C] if [A], on its own or with some
other grounds, is a full ground of [C]. A full ground for [C] is sufficient, by itself, to
ground [C]. It is generally assumed that grounding claims can be expressed using a
variety of different locutions. They can be signified with a sentential operator such as
‘because’, or with a relational formulation such as ‘[A] grounds [B]’ or ‘[B] depends on
[A]’. For ease of expression I’ll generally use the relational formulation, but it shouldn’t
be assumed that anything follows from this choice.
Grounding claims seem each to involve an explanatory element—it is the way that
the painting is received that explains its being beautiful, it is the non-moral features of
an action that explain an action’s being morally permissible, and so on. Grounding is
usually considered a primitive relation; it can’t be analysed in other terms. Along with
citing putative examples of grounding, a favourite recourse for the friend of grounding
when attempting to clue us in to what the notion is supposed to be is to point to
the explanatory character of ground (see e.g. Fine, 2001; 2012). This tactic is taken
to be particularly useful when defending the notion against sceptical attacks on its
intelligibility.2
A quick look to the relevant literature reveals the ubiquitousness of the idea that
grounding and explanation are closely connected. Dreier (2004: 35) says that the
ground for some fact is the ‘most illuminating explanation’ of that fact; Raven (2012:
689) says ‘a fact’s grounds explain it by its holding in virtue of them’, and Trogdon
(2013: 97) says ‘in causal explanations the explanans and explanandum are connected
through a causal mechanism, while in metaphysical explanations they’re connected
through a constitutive form of determination, that of grounding’.
Orthodoxy has it that grounding is transitive (if [A] grounds [B] and [B] grounds
[C], then [A] grounds [C]); irreflexive (nothing grounds itself); hyperintensional
(necessarily co-referring terms cannot be substituted salva veritate); non-monotonic
(if [A] grounds [C], it doesn’t follow that [A] and [B] ground [C]); and asymmetric (if
[A] grounds [B], [B] doesn’t also ground [A]). Note that these properties of grounding
are also generally taken to be properties of explanation. In fact, it is standard to claim
that grounding inherits these properties from the properties of explanation, since
grounding is an explanatory relation (see e.g. Raven, 2015: 327).
Grounding relations are taken to describe the structure of reality, in the sense that
ontological dependence is to be cashed out in terms of grounding. In the literature,
one conception of the structure of reality dominates: metaphysical foundationalism.
Orthodoxy has it that grounding relations form a well-founded partial order. In
§3, I briefly outline the foundationalist conception of metaphysical structure, and
introduce my preferred alternative: metaphysical interdependence.
2 Those mounting such attacks include Daly (2012) and Wilson (2014). Defences which appeal to the
explanatory nature of grounding include Audi (2012a; 2012b); Barnes (2013); Trogdon (2013).
naomi thompson 
3 Foundationalism and Interdependence

Metaphysical foundationalists hold that facts are related in non-repeating chains of
grounding, terminating in a collection of fundamental or foundational facts, which
are not themselves grounded in anything further. We can characterize metaphysical
foundationalism in terms of its commitment to two theses (where x and y are facts):
Well-foundedness: For all x, x is either grounded by some foundational fact or
facts, or is itself a foundational fact.
Asymmetry: For all x and all y (where x = y) if x grounds y then y does not
ground x.3
Well-foundedness guarantees that each fact is ultimately grounded in some founda-
tional fact or facts, and asymmetry guarantees that grounding hierarchies run only in
one direction; from the more fundamental to the less fundamental. The higher up the
chain of grounding a fact appears, the further it is from the foundational facts that
ultimately ground it. (I assume here that all facts are part of the grounding hierarchy.
Anybody who wishes to resist that assumption might take the universal quantifiers in
the definitions above to be restricted to range over facts which they think are part of
that hierarchy.)
Metaphysical foundationalism is roughly analogous to a foundationalist approach
to epistemic justification, where justification is inferred along linear chains of beliefs
from basic beliefs at the end of the chain which are the source of justification. The most
prominent alternative picture of how beliefs are justified is a coherentist approach,
whereby justification emerges when beliefs form a coherent network.
At the heart of the coherentist’s approach is the idea that the structure of justifica-
tion is one of mutual support. A set of facts or beliefs is coherent just in case every
element in the set is supported by all the other elements taken together (Lewis, 1946).
Support might be understood in a weak probabilistic sense, such that P is supported
by Q if the probability of P is raised if we assume that Q is true. Alternatively, we might
define coherence in terms of logical consequence; a coherent set must be consistent,
and every member of the set must follow by logical deduction from the rest (see
Olsson, 2017).
An analogous picture is plausible in the metaphysical case. Grounding between
facts might be such that each fact is supported by all the other facts taken together,
rather than (as in foundationalism) being supported by a set of foundational facts.
If grounding relations are taken to hold with metaphysical necessity (as is stan-
dard), then probable relations will have little role to play in either the metaphysical
foundationalist or the interdependence picture. Instead, we can focus on something
3 That x and y are distinct facts is built into my definition of asymmetry here so as not to make any
assumptions about the reflexivity of grounding; so that those (e.g. Jenkins, 2011) who do not wish to rule
out the possibility that ground might be a reflexive relation are not thereby forced to deny that ground is
asymmetric. This is sometimes called antisymmetry.
 interdependence, coherentism, and holistic explanation
closer to logical deduction. It is common for foundationalists to claim that once the
foundations are settled, the rest of reality follows—a complete story can be told in
terms of the foundations alone. Friends of metaphysical interdependence might claim
that any element of the system can be deduced from all the rest taken together, and
that we can’t tell a complete story without considering the system as a whole.
In addition, support has an explanatory dimension. Just as the metaphysical foun-
dationalist takes the fundamental facts to explain or to account for all the rest,
the friend of metaphysical interdependence can take each fact in the system to be
supported, in this explanatory sense, by all the other facts taken together. Facts
are related in web-like networks of ground such that each fact in the network is
partially grounded by other facts in the network. Metaphysical interdependence is
thus most similar to versions of coherentism that focus on inferential connections
between component beliefs. These beliefs often form a linear order within the system,
but a given belief might appear in its own reason ancestry. By the transitivity of
partial ground, each fact in the interdependent network is partially grounded by
every other fact in the network, and itself. The grounds for some fact [A] might
be the further facts [B], [C], and [D], but [D] might itself be grounded by [E],
[F], and also by [A]. There might therefore be no fundamental facts, but the facts
in the network enter into mutually supporting grounding relations. The friend of
metaphysical interdependence thus denies both well-foundedness and asymmetry.
Note that a weak version of interdependence requires only that there be at least one
counterexample to both well-foundedness and asymmetry. In that case, grounding
structures might be hybrids featuring chains of ground that bottom out in foun-
dational facts alongside small pockets of interdependence. Such a case would be
much less closely analogous to the coherentist position sketched above than a strong
version of interdependence where there are no foundational facts. Though I think
a weak version of interdependence deserves attention, my discussion here concerns
the stronger version of the view. For our purposes then, we can assume that the
friend of interdependence endorses non-well-foundedness; the view that there are
no foundational facts. Non-well-foundedness is a commitment usually shared with
metaphysical infinitists (see e.g. Cameron, 2008; Morganti, 2009), who hold that linear
chains of ground extend infinitely in some direction.4
Metaphysical interdependence requires that asymmetry also be rejected. If [A]
grounds [B], [B] might (either fully or partially) also ground [A]. In order to motivate
my claim that asymmetry should be rejected, I’ll mention two examples which purport
to demonstrate that grounding is not asymmetric, the first involving full grounding,
and the second partial grounding (see Rodriguez-Pereyra, 2015 and Thompson,
2016 for more detailed examples). First, consider the following true propositions:
A = <B is true> and B = <A is true>. Assume that both propositions are true. It is very
4
Unfortunately I do not have the space here to discuss versions of infinitism in any detail.
naomi thompson 
plausible to suppose that propositions are true in virtue of the relationship between
their constituents and the world (thus, e.g. the fact that <snow is white> is true is
grounded in the fact that snow is white). In this case, the fact that A is true is grounded
in the fact that B is true, and the fact that B is true is grounded in the fact that A is true.
If we accept that this is an example of grounding, we must accept that full grounding
is not asymmetric.
Second, consider the relationship between the qualities of mass, density, and
volume in a homogeneous fluid. A natural way to describe that relationship is in
terms of grounding; the volume of the fluid seems to have the value it does in virtue
of the values of the mass of the fluid and the density of the fluid. But each of the
parameters seems to have the value it does in virtue of the other two parameters;
the three parameters are interrelated. If grounding is transitive, then facts about the
value of each of the parameters partially ground facts about the value of the other
two, and itself. If we accept this example, we must accept that partial grounding is not
asymmetric.
Once we have at least cast doubt on the orthodox view that grounding must be
asymmetric, we can consider arguments for metaphysical interdependence. Three
such arguments are given in Thompson (2016). I’ll briefly outline two of them. First,
metaphysical interdependence is the only theory capable of reconciling competing
intuitions about certain cases of grounding. Take, for example, the grounding between
facts about an organism and its facts about its organs. It seems, for example, that
the fact that the heart pumps blood around the body depends on the fact that the
organism exists and has a properly functioning circulatory system. But the fact that
the organism exists and has a properly functioning circulatory system depends on the
fact that the heart pumps blood around the body. More generally, in the case of what
Schaffer (2010: 47) calls integrated wholes (which are to be distinguished from mere
aggregates, such as heaps), there are good reasons to think that the parts are grounded
in the whole. Consider, for example, a circle; any divisions (e.g. its semi-circles) are
‘arbitrary partition[s] on the circle’ (Schaffer, 2010: 47). But there are also good reasons
for thinking that wholes are always grounded in their parts. Just as atoms might
compose a table, semi-circles might compose a circle, and so the fact that the circle
exists can be explained by the fact that the semi-circles exist; the existence of the circle
seems to depend on the existence of the semi-circles. Metaphysical interdependence
can simultaneously account for both of these seemingly competing intuitions about
dependence.
Second, we cannot rule out the possibility that the world is gunky—that everything
has proper parts. It is also possible that the world is junky—that everything is a proper
part of something. Assuming that parthood relations entail grounding relations,5 we
cannot rule out the possibility of an infinitely extending grounding chain in either
5
See e.g. Cameron (2008); Schaffer (2010).
 interdependence, coherentism, and holistic explanation
direction. Metaphysical foundationalists require that grounding chains terminate,

whether they terminate in multiple foundational facts as in priority pluralism (which
is incompatible with gunky worlds) or a single totality fact as in priority monism
(which is incompatible with junky worlds). The foundationalist must therefore argue
(if grounding relations are taken to hold with metaphysical necessity) that one or other
of gunk and junk is metaphysically impossible.6 Metaphysical interdependence does
not require that there be any foundational facts, and so is compatible with gunk, junk,
both, or neither (this also makes interdependence more attractive than infinitism,
which requires that there be infinite extension of grounding chains in at least one
direction).
With our account of metaphysical interdependence on the table, I motivate meta-
physical interdependence via an analogy with coherentism about epistemic justifica-
tion in §4. The rejection of asymmetry and of well-foundedness is held in common by
friends of metaphysical interdependence, advocates of coherentism, and defenders of
the holistic approaches to explanation in general which I discuss in the final sections
of the paper.
4 Epistemic Coherentism
In §3, I highlighted an analogy between metaphysical interdependence and epistemic
coherentism. Both structures are non-well-founded and involve relations of mutual
self-support. Coherentists think that no belief is the source of its own justification,
but rather that justification is an emergent feature of coherent sets of beliefs. Friends
of metaphysical interdependence think that no fact is ungrounded. The analogy
with coherentism isn’t perfect, but an understanding of how beliefs might work
together to support one another via inferential connections might help make clear
how facts support one another in a system of ground characterized by metaphysical
interdependence. The example I describe below focuses on explanations for beliefs.
Justificatory explanations are one type of explanation, and so the example might also
help to motivate the more general claim I make in this paper, that holistic explanations
can be good explanations.
Here’s the example: Aimee notices that her neighbour, Bob, seems a little upset and
withdrawn. She wonders what could explain his behaviour, and realizes that she hasn’t
seen Bob’s partner, Chris, in a week or so. She then remembers hearing raised voices
coming from Bob’s house one evening, about a week ago. She hypothesizes that Bob’s
behaviour is due to him and Chris having split up. With that explanation in mind, she
reasons that the raised voices she heard were Bob and Chris, and that the reason she
6 Even if the foundationalist denies that grounding is necessary in this way, the inability to rule out the
possibilities of gunk or junk in the actual world poses the same problem.
naomi thompson 
hasn’t seen Chris lately is that Chris and Bob have split up. On this basis she reasons
that Bob is indeed upset, and that he’s upset because he and Chris have split up.
In the above example, Aimee’s belief that Bob and Chris have split up is explained by
her observations that she hasn’t seen Chris lately, that she heard raised voices a week
or so ago, and that Bob seems upset. In turn, it is Aimee’s belief that Bob and Chris
have split up that lends support to her belief that Bob really is upset, and explains why
Chris hasn’t been around lately, and why voices were raised. This is a holistic system
of explanation, characterized by mutual support relations between explanans and
explanandum. It’s important to note that without taking each belief in this system to
be supported by all the other beliefs in the system, this explanation of Bob’s behaviour
wouldn’t seem like a good explanation.
Without further argument one might maintain that holistic explanations can be
plausible in epistemic cases such as that of the justification of our beliefs, but are
implausible when it comes to metaphysics, and to grounding. I think explanation
should be understood in general terms, such that holism about explanation in one
area should strengthen the case for holism in others. In §5, I argue that grounding is
an explanatory relation, and in §6 and §7 that grounding explanations can be holistic
explanations.
5 Explanation and Ground

What precisely is the relationship between ground and explanation? There seem to be
two options. Either the relationship is one of identity; grounding just is a relation of
metaphysical explanation, or it is the weaker relationship of tracking; explanations
track grounding relations. Support for both views can be found in the literature.7
In this section I offer two considerations in favour of the former view, the first
based on a problem for the tracking view, and the second based on the epistemology
of ground.
Assume that explanations track grounding relations. This idea has some precedent
with respect to causation; we generally think of causal explanations as ‘tracking’
causal relations in the world (see e.g. Salmon, 1984; Lewis, 1986; Woodward, 2003).
A question then arises as to the nature and the mechanism of this tracking relation.
Kim (1994: 26) argues that explanations track dependence relations. More precisely,
Kim says that ‘the relation that “grounds” the relation between an explanans, G, and
its explanatory conclusion, E, is that of dependence; namely G is an explanans of E
just in case e, the event to be explained depends on g, the event evoked as explaining
it’. Kim thus takes dependence relations to ground explanations.
7 Defenders of the former view include Dasgupta (2014); Fine (2012); Litland (2013); Raven (2012); and
Rosen (2010). Defenders of the latter view include Audi (2012a); Schaffer (2012); and Trogdon (2013).
 interdependence, coherentism, and holistic explanation
Let’s suppose then that the relationship between a grounding relation and an
explanation that tracks it is that of grounding. Kim distinguishes between the relata
of the dependence relation (which in the above quote are events) and the relata of
the explanatory relation. Since we are assuming that the relata of the grounding
relation are facts, we won’t make a corresponding distinction; explanatory relations
and grounding relations both have facts as their relata.
We can suppose then that when [A] grounds [B], there is a corresponding explana-
tory relationship between [A] and [B], and that the relationship between the facts
that [A] grounds [B] and that [A] explains [B] is one of grounding; [[A] grounds [B]]
grounds [[A] explains [B]]. Here’s the worry. That the relationship between ground
and explanation is itself one of grounding is not viciously circular, but it is problematic
because it renders the connection between explanation and ground explanatorily
redundant.
Friends of grounding rely on the close connection between explanation and
grounding in order to elucidate the relatively opaque notion of primitive metaphysical
grounding with the far more familiar notion of explanation (see e.g. Audi, 2012a; Fine,
2012; Raven, 2015; Trogdon, 2013). If an appeal to explanation is to shed light on the
notion of ground, part of what must be understood is how ground and explanation
are related. Here we are told that the relationship between ground and explanation
is in fact one of ground, but ground was what we were seeking elucidation of in the
first place! If the connection between ground and explanation is to illuminate the
notion of ground, we need an account of that connection not itself cashed out in
terms of ground.
So if not ground, then what? Merely modal notions like supervenience and entail-
ment are too coarse-grained to respect the sense in which it is the grounding relations
that back the explanations, and not the other way around. Making sense of the
dependence of the explanation on the grounding relation requires some kind of
hyperintensional account of tracking, and so the friend of the tracking conception
must come up with some alternative to grounding which meets this condition. In the
absence of such an account, we have reason to prefer a view whereby grounding is
an explanatory relation, and so there is no mystery as to how an understanding of
explanation might help elucidate ground.
A second reason for thinking that grounding is a relation of explanation concerns
the epistemology of ground. An underexplored question in the grounding literature
is that of how we come to know what grounds what. It is generally assumed that
we can know about grounding by recourse to our explanatory intuitions (see
e.g. Fine, 2001; 2012). Here’s one reason for thinking that explanatory intuitions
clue us in to grounding relations: grounding is hyperintensional, and knowledge
of hyperintensional notions requires a way of knowing which is sensitive to
hyperintensional distinctions. Explanatory knowledge is a paradigm example. We
can explain to someone why it costs a fortune to see Bob Dylan in concert in terms
naomi thompson 
of Dylan’s fame, but not (unless they know that Dylan and Zimmerman are the same
person) why it costs a fortune to see Robert Zimmerman in concert in terms of
Dylan’s fame. Similarly, understanding that Socrates grounds {Socrates}(but not that
{Socrates} grounds Socrates) requires an epistemology that is sensitive to fine-grained
distinctions between necessary co-existents.
If grounding is an explanatory relation, we have a neat explanation both of the
hyperintensionality of grounding, and of how we could come to know about the
grounding relations (explanations are precisely the kinds of things we can come to
know). If, on the other hand, explanations merely track grounding relations, it is an
open question how we can come to know what grounds what. Adopting a weaker
account of the connection between explanation and ground generates a requirement
for an epistemology of ground. Friends of grounding might be able to provide such
an epistemology, but until they do it seems safer to assume that if we can know the
grounding facts, it’s because grounding just is an explanatory relation.
Neither of the above arguments provides conclusive proof that we ought to adopt
the view that grounding just is an explanatory relation, but they provide reason
to think that the alternative account leaves some important questions unanswered.
My aim here is merely to support the claim that explanatory considerations can
legitimately be brought to bear on our account of grounding. I suspect that an account
of the tracking view that answered the questions raised above would also legitimize
this sort of argument, but I think we have reason for now at least to set that view aside,
and to proceed as though ground just is an explanatory relation.
Before continuing, I wish to respond to one perceived concern. In the next sections,
I argue that considerations about the structure of explanation push us towards accept-
ing metaphysical interdependence over foundationalism. But friends of grounding
don’t just say that grounding is an explanatory relation. They say that grounding is
related to explanation of a distinctively metaphysical sort (see e.g. Fine, 2012: 37).
It might be that metaphysical explanation functions very differently from the more
familiar explanatory notions. Were that the case, discussion of explanation generally
might not be relevant here.
In response, note that if we are to be able to use the connection between grounding
and explanation as it is generally assumed that we can (to shed light on the features of
grounding generally, as well as to settle questions about what grounds what), what we
mean by explanation in this context must be at least fairly familiar; not too far removed
from our ordinary understanding of what an explanation amounts to. We couldn’t
use an unfamiliar notion of explanation to shed light on anything. In the absence
of a properly developed theory of metaphysical explanation independent from our
theory of grounding, assuming that metaphysical explanation is divorced from our
ordinary conception of explanation would only serve to undermine the project
of highlighting connections between the two notions (see Daly, 2012; Thompson,
forthcoming). I therefore assume that metaphysical explanation is not different in
 interdependence, coherentism, and holistic explanation
kind from explanation in general (and so, for example, will exhibit the same formal
features).
6 Explanatory Considerations
In this section I argue that a foundationalist conception of the structure of explanation
is inferior to the sort of holistic account of explanatory structure favoured by friends
of metaphysical interdependence. I discuss two problematic features of the founda-
tionalist account, and give two examples of structures of metaphysical explanation
best described in accordance with holism.
6.1 Foundationalism
The idea that explanatory chains terminate is a familiar one. Take, for example, the
idea that higher level psychological facts are grounded in and explained by biological
facts, which are themselves grounded in and explained by chemical facts, in turn
grounded in and explained by physical facts, themselves dependent on lower level
microphysical facts. Eventually, some argue, we reach a level of facts for which no
further explanation can be given (or at least we will reach this level when physics
is complete). Advocates of this sort of conception of the structure of explanation
think that explanations terminate; there is a point at which (relative to a given fact)
the explanatory project is complete. There are some facts that don’t require further
explanation, or that are ungrounded.
When evaluating this account of the structure of explanation we might first
consider theoretical virtue.8 In evaluating competing theories, facts that cannot be
explained are called brute facts. Brute facts carry a theoretical cost; they represent
something that a theory leaves unexplained. Other things being equal, a theory that
carries a commitment to fewer brute facts is to be considered superior because it has
more explanatory power—it leaves fewer things unexplained. On a simple assessment
then, metaphysical foundationalism does much worse than interdependence, because
it carries a commitment to as many brute facts as there are fundamental facts.
Metaphysical interdependence (at least in the strong form under consideration here)
denies that there are any brute facts at all; each fact is (partially) explained by every
other fact in the system.
Arguments from theoretical virtue are notoriously difficult to evaluate, and a friend
of a foundationalist way of thinking about explanation might even take a non-
foundationalist view according to which there are no brute facts to be a reductio of
the idea that the number of brute facts in a theory is to be minimized. But I think this
would be a mistake, for the reasons mentioned above; brute facts weaken a theory. If
8
There is some precedent for this—see e.g. Cameron (2008).
naomi thompson 
we can have a theory according to which there are no brute facts, so much the better!
There is something unsettling about the idea that there are some facts which simply
obtain, and that’s all there is to say about it.
I’ll mention two ways to make this point a little more concrete. First, note again that
brute facts are not explained by the theory of which they are a part. A foundationalist
structure does not allow for the possibility that there might be metaphysically explana-
tory connections at the fundamental level of explanation, because all grounding
chains terminate in brute facts. Applied to our above example, that means that the
facts of fundamental physics are explanatorily independent of one another (at least in
terms of the sort of explanation relevant to grounding). This restriction on grounding
between fundamental physical facts in fact goes against our best current physics,
which points to the presence of holism and/or non-separability at the quantum level
(see e.g. Healey, 2016). This is naturally cashed out in terms of grounding (see e.g.
Schaffer, 2010; Ismael and Schaffer, forthcoming). Metaphysical interdependence is
the only theory of grounding (perhaps other than monist foundationalism) suited to
characterize this sort of metaphysical dependence.
A second unsettling feature of brute facts is that they represent apparent coun-
terexamples to the principle of sufficient reason (PSR); the principle stating that every
fact has an explanation. Bliss (2013: 415) argues that in requiring that every fact be
ultimately explained by some unexplained explainer (the brute facts) metaphysical
foundationalists reveal an implicit commitment to the PSR. Bliss contends that this
invites difficult questions about the modal status of the fundamental facts and their
ability to act as the requisite sort of explainers. At a minimum, unexplained explainers
should presumably be non-contingent, but foundationalists do not usually claim that
all foundational facts are thereby necessary facts. Foundationalists seem both to be
committed to the PSR and routinely to violate it.
6.2 Interdependence
An alternative approach to explanation in general involves thinking of explanation
as a holistic phenomenon. Explaining some fact, belief, or event is a matter of
connecting it with other facts, beliefs, or events in such a way that there can be
a mutual reinforcement between an explanation and what it explains (see Quine
and Ullian, 1978: 120). Holism in diverse areas of thought is enjoying something
of a resurgence (e.g. sociology, politics, ecology, psychology, physics, and medicine).
Holistic approaches to explanation are more often encountered in daily interactions
then in philosophical writing,9 and so I’ll illustrate what I have in mind with a
couple of examples. These I take to be examples of holistic metaphysical explanations,
but it is not just grounding explanations that force us to think of explanations as
9 Exceptions include discussions of holism about meaning, of quantum mechanics, and of living
organisms.
 interdependence, coherentism, and holistic explanation
having a holistic structure.10 More generally, there are reasons to think that holistic
explanations can be good explanations (see e.g. the example in §4 above) and so there
is nothing ad hoc about taking explanation to have this structure in the grounding
case. This matters if, as I suggested above, we are to appeal to our understanding of
explanation in general to elucidate grounding.
A first example of the relevant sort of explanatory structure is present in struc-
turalist approaches to mathematics. According to mathematical structuralists, math-
ematical objects are ontologically dependent on one another and on the mathematical
structure of which they are a part (see e.g. Shapiro, 1997). The following provides a
good characterization of the view:
The number 2 is no more and no less than the second position in the natural number structure;
and 6 is the sixth position. Neither of them has any independence from the structure in which
they are positions, and as positions in this structure, neither number is independent of the
other. (Shapiro, 2000: 258)
According to the mathematical structuralist, mathematical objects depend, for

their existence and their identity, on the other objects that are part of the structure.
The identities of those objects are determined ‘by their relationships to other positions
in the structure to which they belong’ (Resnik, 1982: 95). An obvious candidate
for understanding the relevant sort of dependence is in terms of grounding; facts
about the identity and nature of any given mathematical object are explained by, and
grounded in, facts about the structure of which that object is a part, and in facts about
the structure as a whole.
If grounding is an explanatory relation, the nature of the number 2 is to be explained
in terms of the natural number structure, and so it both helps explain the nature
of other numbers in the structure, and is itself explained in terms of the nature
of those other numbers which make up the structure. That structure itself is to be
explained in terms of the numbers that are a part of it. Mathematical structuralism
can only be adequately described by appeal to holistic explanation and to metaphysical
interdependence.
Consider another example. Structural realism is a now-popular form of scientific
realism which holds that rather than asserting that the nature of things is correctly
described by our best scientific theories, the realist emphasis should be on the struc-
tural content of those scientific theories. The motivation for the position is that it is
structural content that is retained across theory change (Ladyman and Ross, 2007: 93).
Structural realists of a metaphysical variety take structure and relations not to
be merely derivative of the entities or structural nodes they relate, but to be more
ontologically fundamental than has traditionally been assumed by scientific realists
(Ladyman, 2014). Esfeld and Lam (2010) develop a version of metaphysical structural
10 I use the terms ‘metaphysical explanation’ and ‘grounding explanation’ synonymously throughout.
naomi thompson 
realism that they term ‘moderate structural realism’ and which holds that relations
require relata, but that it is not the case that the relata necessarily have intrinsic
properties over and above the relations they bear to one another (Esfeld and Lam,
2010: 13). In other words, there are objects, but those objects are wholly characterized
by the relations in which they stand.
It follows from this characterization of moderate structural realism that there is
a mutual dependence between relations and relata—the objects are characterized by
the relations that relate them, and the relations themselves are characterized by the
objects that stand in the relations. There is, therefore,
. . . a mutual ontological as well as conceptual dependence between objects and structure
(relations): objects can neither exist nor be conceived without relations in which they stand,
and relations can neither exist in the physical world nor be conceived as the structure of the
physical world without objects that stand in the relations. (Esfeld and Lam, 2010: 13–14)
Explaining facts about the nature of a given object is a matter of citing facts about
the relations that object bears to other objects, facts about which are themselves to be
explained in terms of the relations they bear to further objects (and to the object which
was our starting point). Since facts about the structure through which the objects
are related themselves depend on facts about the objects themselves, a moderate
structural realist picture has it that giving an explanation will be a matter of pointing
towards a mutual dependence between explanans and explanandum. The relevant sort
of explanation here is a grounding explanation; facts about objects are grounded in
facts about relations, and vice versa. Moderate structural realism thus gives us a case of
symmetric metaphysical explanation, and therefore of metaphysical interdependence.
These two examples strengthen the case for thinking that explanation, and in
particular metaphysical explanation, is at least sometimes a holistic affair. A defence
of the strong version of metaphysical explanation under discussion here requires the
stronger claim that all metaphysical explanations are holistic explanations. I make the
case for this in §7, after first dispelling an objection to the idea that any explanations
are holistic.
7 Holistic Explanation
There is an obvious and immediate objection to the idea that explanation could be
a holistic phenomenon; explanations are asymmetric, and holistic systems licence
symmetric explanation. But there are good reasons to think that not all explanations
are, in fact, asymmetric. The examples given above each provide some reason to
resist the assumption that explanation is asymmetric. Numerous other examples are
available. To mention just two: recent developments in quantum mechanics have led
many to believe that some quantum states can only be explained holistically (see e.g.
Teller, 1986; Healey, 2016); and Achinstein (1983) gives a number of examples of
identity explanations (i.e. explanations that identify two things, such as water and
 interdependence, coherentism, and holistic explanation
H2 O) which seem to require us to accept that explanation is non-symmetric. Note

too that it is a well-documented feature of Hempel’s influential (e.g. 1965) Deductive-
Nomological (DN) account of explanation (albeit usually invoked as a criticism) that
it licences instances of symmetric explanation. These are reasons to think that our
familiar understanding of explanation includes at least some instances of symmetric
explanation.
It is in fact a difficult task to identify a basis for the intuition that explanations must
be asymmetric. Attempts to locate that asymmetry in the asymmetry of causation have
been called into question (see e.g. Persson, 2001), and in any case there are many rivals
to causal theories of explanation. In particular, metaphysical explanation is generally
assumed to be non-causal.11
An alternative way to account for the intuition that explanation is asymmetric is
to claim that explanation is an epistemic phenomenon, and symmetric explanations
are unlikely to advance our understanding. On this epistemic basis, we generally
object to symmetric explanations. This indeed is the position of van Fraassen (e.g.
1980), recent champion of the pragmatic theory of explanation. Van Fraassen argues
that ‘asymmetries of explanation result from a contextually determined relation
of relevance . . . [we] account for specific asymmetries in terms of the interests of
questioner and audience that determine this relevance’ (1980: 130). According to
van Fraassen, explanations are only explanatory in a context, where that context is
determined by the interests and the background commitments of whoever is seeking
the explanation.
In most cases, context dictates that asymmetric explanations are satisfying, but
sometimes the context will be such that holistic explanations are required. We have
seen some compelling examples already, but another case might help bring out the
point. In order to argue for his claim that explanatoriness is dependent on context,
van Fraassen (1980, §3.2) amends an example used to criticize Hempel’s DN account
of explanation. The example involves deriving the length of the shadow s cast by a
flagpole from the height h of the flagpole, the angle θ of the sun above the horizon
and laws l about the rectilinear propagation of light. According to Hempel’s theory,
we can deduce, and thus explain s by appeal to h, θ, and l. However (so the criticism
goes), we can also deduce, and thus explain, h from s, θ, and l. But our explanatory
intuitions are such that the direction of explanation only runs one way—s should be
explainable in terms of h, but s shouldn’t come out as an explanation of h—lengths of
shadows don’t explain heights of flagpoles.
Van Fraassen adapts the example and constructs a story according to which the
context is such that we do take the length of the shadow to explain the height of the
11 Note that there is no tension here with the claim that grounding is an explanatory relation. Not
all explanations are grounding explanations, and so there may be some features of other varieties of
explanation not shared with grounding explanations (though it is important that grounding explanations
are familiar enough from our ordinary understanding of explanation to be illuminating).
naomi thompson 
flagpole. The flagpole was designed to have height h in order that the shadow s would
be cast on a particular spot at a particular time.12
Van Fraassen’s point is that as a response to a particular explanatory request in
a particular context, it is both appropriate and satisfying to give an explanation in
terms of the length of the shadow. But as a response to a slightly different explanatory
request, it would be equally appropriate and satisfying to give an explanation of
the length of the shadow in terms of the height of the flagpole. Local explanatory
asymmetries are the consequence of the context-sensitivity of explanation, but are
not built in to what it is to be an explanation.
We can now make two points about explanation and asymmetry. The first is that
if we understand the asymmetry of explanation in terms of its connection to under-
standing and its relevance to the explanation seeker, there is no principled restriction
on explanation such that explanations must always go in one direction. Holistic
systems of explanation are certainly not ruled out, and are only to be considered
unacceptable if they fail to advance the epistemic position of the explanation seeker;
if they fail to make something intelligible, or to advance an agent’s understanding. We
have seen a number of examples above which demonstrate that holistic explanations
at least sometimes advance an agent’s epistemic position.
Second, the context sensitivity of explanation allows us to recognize that though an
entire explanatory system might have a holistic structure, it is very rarely appropriate
to cite the whole system in response to a specific explanation request. Accepting
that explanations can be holistic doesn’t require us to give up on the idea that local
explanations very often only work in one direction. In this way we can in fact maintain
that all explanatory systems have a holistic structure, as is required by the strong
version of metaphysical interdependence. In some cases (such as in the examples
above) the context dictates that providing a satisfactory explanation requires citing
enough of the explanatory system that its holistic structure is evident. In others,
requests for explanation are limited enough that a satisfying explanation cites only
a seemingly asymmetric portion of the structure.
The claim that it is generally appropriate to cite only a portion of an explanatory
system is far from unique to holistic explanations. In providing an adequate causal
explanation of some event, we are only required to cite information about some portion
of the causal history of that event (see e.g. Salmon, 1984; Lewis, 1986). The causal
history of each event can (presumably) be traced all the way back to the Big Bang, but
in very few contexts is an explanation which mentions the Big Bang appropriate. The
cause that really explains depends on our interests (Lipton, 1990: 249).
So, even if giving a maximally complete explanation would involve describing a
holistic system, and so we ought to think of explanations as having a holistic structure,
local cases of providing an explanation are generally far more restricted. Consider
12
Actually van Fraassen’s example involves a tower, but that need not concern us here.
 interdependence, coherentism, and holistic explanation
again the case of coherentism in epistemology. When we ask a coherentist why she
holds a given belief, she will offer support for that belief in terms of other beliefs to
which the questioned belief is closely connected in her belief system. Though it is an
aspect of her view that justification for the disputed belief depends on the entire belief
system, she could hardly be expected to communicate the entire system in response
to a request for justification (any more than a foundationalist is to be expected to
follow his or her chain of beliefs right down to the foundation when giving reasons
for a particular questioned belief). The same is true when providing a grounding
explanation.
At this stage, one might object again that this is all very well for ordinary cases of
explanation, but our concern here is supposed to be with metaphysical explanation.
One might worry that there is no place for a discussion of context when our concern
is with this supposedly more objective notion of explanation, and argue as follows:
Metaphysical explanation is not supposed to be about getting somebody to understand
something—it’s about characterizing reality’s structure. Perhaps it’s acceptable to take
ordinary or scientific explanations to have a holistic structure and to explain the illusion
of asymmetry in terms of our explanatory interests, but metaphysical explanations
are different; they’re divorced from our explanatory interests, and so if they appear
asymmetric it must be that the relevant structure is asymmetric!
I have already discussed some cases where it seems clear that metaphysical explana-
tion is not asymmetric (e.g. in structuralist approaches to mathematics and science),
but I’ll make a couple more points in favour of holistic metaphysical explanation. First,
note again that in distancing metaphysical explanation from ordinary explanation, the
friend of grounding is playing a dangerous game. The connection between grounding
and explanation is key to elucidating the somewhat opaque notion of grounding,
as well as to evaluating particular grounding claims. In order for that connection
to play the required role, grounding must be related to a notion of explanation
familiar enough that our explanatory intuitions are deemed relevant, and that we can
understand grounding in terms of it. That familiar notion of explanation is connected
to understanding, and is constrained by context.
But for the friend of grounding who wishes to resist that cautionary note and insist
that metaphysical explanation is not connected to understanding and that context
has no role to play, we can remind them that it is hard to locate any preference for
asymmetric explanation outside of the epistemic constraints on explanation in gen-
eral. If metaphysical explanation really is non-epistemic, those reasons to worry about
symmetric explanation don’t apply. In order to resist the possibility that metaphysical
explanations might be holistic, one must both resist the above putative examples of
symmetric grounding and holistic metaphysical explanation, and come up with a non-
epistemic account of the alleged asymmetry of (metaphysical) explanation.
I have argued that the supposed asymmetry of explanation is tied to the epistemic
character of explanation, and that that epistemic character is consistent with a view
whereby explanation can have a holistic structure. Not only can holistic explanations
naomi thompson 
be informative (and so it is a mistake to think that only asymmetric explanations can

increase understanding), but it is consistent with holding that explanatory structures
are holistic that context dictates that many local requests for explanation will elicit a
response which cites only a fraction of the holistic structure, and which runs only in
one direction.
8 Concluding Remarks
Foundationalism has been the dominant view of metaphysical structure since the
inception of the contemporary debate, but there are good reasons to challenge the
foundationalist orthodoxy. I have described some of those reasons here. In particular,
that in accepting that holistic explanations are good explanations, we break down a
significant barrier to serious consideration of theories such as metaphysical interde-
pendence, which are alternatives to metaphysical foundationalism.13
References
Achinstein, P. (1983). The Nature of Explanation. Oxford: Oxford University Press.
Audi, P. (2012a). A Clarification and Defense of the Notion of Ground. In F. Correia &
B. Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality
(pp. 101–21). Cambridge: Cambridge University Press.
Audi, P. (2012b). Grounding: Toward a Theory of the In-Virtue-Of Relation. Journal of Philos-
ophy, 109, 685–711.
Barnes, E. (2013). Explanation and Fundamentality. In M. Hoeltje, B. Schnieder, & A. Stein-
berg, Varieties of Dependence (Basic Philosophical Concepts Series) (pp. 211–42). Munich:
Philosophia Verlag.
Bliss, R. (2013). Viciousness and the Structure of Reality. Philosophical Studies, 166, 399–418.
Cameron, R. (2008). Turtles All the Way Down: Regress, Priority and Fundamentality. The
Philosophical Quarterly, 58(230), 1–14.
Daly, C. (2012). Scepticism about Grounding. In F. Correia & B. Schnieder (eds), Metaphysical
Grounding: Understanding the Structure of Reality (pp. 81–100). Cambridge: Cambridge
University Press.
Dasgupta, S. (2014). On the Plurality of Grounds. Philosophers’ Imprint, 14, 1–28.
Esfeld, M. & Lam, V. (2010). Holism and Structural Realism. In R. Vanderbeeken &
B. D’Hooghe (eds), Worldviews, Science and Us: Studies of Analytical Metaphysics. A Selection
of Topics from a Methodological Perspective (pp. 10–31). Singapore: World Scientific.
Fine, K. (2001). The Question of Realism. Philosopher’s Imprint 1, 1–30.
Fine, K. (2012). A Guide to Ground. In F. Correia & B. Schnieder (eds), Metaphysical Grounding:
Understanding the Structure of Reality (pp. 37–80). Cambridge: Cambridge University Press.
13 Thanks to Darragh Byrne for helpful comments and discussion, and to two anonymous referees for
detailed and thought-provoking comments.

 interdependence, coherentism, and holistic explanation
Healey, R. (2016). Holism and Nonseparability in Physics (ed. E. Zalta). Retrieved March 2016
from The Stanford Encyclopedia of Philosophy: http://plato.stanford.edu/archives/spr2016/
entries/physics-holism.
Hempel, C. (1965). Aspects of Scientific Explanation and Other Essays in the Philosophy of
Science. New York: Free Press.
Ismael, J. & Schaffer, J. (forthcoming). Quantum Holism: Nonseparability as Common Ground.
Synthese.
Jenkins, C. (2011). Is Metaphysical Dependence Irreflexive? The Monist, 94, 267–76.
Kim, J. (1994). Explanatory Knowledge and Metaphysical Dependence. Philosophical Issues, 5,
51–69.
Ladyman, J. (2014). Structural Realism (ed. E. Zalta). Retrieved August 2015 from The Stanford
Encyclopedia of Philosophy: http://plato.stanford.edu/archives/spr2014/entries/structural-
realism.
Ladyman, J. & Ross, D. (2007). Every Thing Must Go: Metaphysics Naturalized. Oxford: Oxford
University Press.
Lewis, C. I. (1946). An Analysis of Knowledge and Valuation. LaSalle: Open Court.
Lewis, D. (1986). Causal Explanation. In Philosophical Papers, Volume II. Oxford: Oxford
University Press.
Lipton, P. (1990). Contrastive Explanation. Royal Institute of Philosophy Suppliment, 27, 247–66.
Litland, J. (2013). On Some Counterexamples to the Transitivity of Grounding. Essays in
Philosophy, 14(1), 19–32.
Morganti, M. (2009). Ontological Priority, Fundamentality and Monism. Dialectica, 63(3),
271–88.
Olsson, E. (2017). Coherentist Theories of Epistemic Justification (ed. E. Zalta). Retrieved
November 2017 from The Stanford Encyclopedia of Philosophy (spring 2017 edn): https://
plato.stanford.edu/archives/spr2017/entries/justep-coherence/.
Persson, J. (2001). Why is there Explanatory Asymmetry? Explanatory Connections (ed. M.
Kiikeri & P. Ylikoski). Retrieved from: http://www.helsinki.fi/tint/matti/.
Quine, W. & Ullian, J. (1978). The Web of Belief (2nd edn). New York: McGraw-Hill.
Raven, M. (2012). In Defence of Ground. Australasian Journal of Philosophy, 90, 687–701.
Raven, M. (2015). Ground. Philosophy Compass, 10(5), 322–33.
Resnik, M. (1982). Mathematics as a Science of Patterns: Epistemology. Noûs, 16, 95–105.
Rodriguez-Pereyra, G. (2015). Grounding is Not a Strict Order. Journal of the American
Philosophical Association, 1(3), 517–34.
Rosen, G. (2010). Metaphysical Dependence: Grounding and Reduction. In R. Hale and A.
Hoffman (eds), Modality: Metaphysics, Logic, and Epistemology (pp. 109–36). Oxford: Oxford
University Press.
Salmon, W. (1984). Scientific Explanation and the Causal Structure of the World. Princeton, NJ:
Princeton University Press.
Schaffer, J. (2010). Monism: The Priority of the Whole. Philosophical Review, 119, 31–76.
Schaffer, J. (2012). Grounding, Transitivity, and Contrastivity. In F. Correia & B. Schnieder
(eds), Metaphysical Grounding: Understanding the Structure of Reality (pp. 122–38).
Cambridge: Cambridge University Press.
Shapiro, S. (1997). Philosophy of Mathematics: Structure and Ontology. Oxford: Oxford Univer-
sity Press.
naomi thompson 
Shapiro, S. (2000). Thinking about Mathematics. Oxford: Oxford University Press.

Teller, P. (1986). Relational Holism and Quantum Mechanics. British Journal for the Philosophy
of Science, 37(1), 71–81.
Thompson, N. (2016). Metaphysical Interdependence. In M. Jago (ed.), Reality Making. Oxford:
Thompson, N. (forthcoming). Grounding, Metaphysical Explanation, and the Structure of
Reality. Proceedings of the Aristotelian Society.
Trogdon, K. (2013). An Introduction to Grounding. In M. Hoeltje, B. Schnieder, & A. Steinberg
(eds), Varieties of Dependence (pp. 97–122). Munich: Philosophia Verlag.
van Fraassen, B. (1980). The Scientific Image. Oxford: Oxford University Press.
Wilson, J. (2014). No Work for a Theory of Grounding. Inquiry, 57(5–6), 1–45.
Woodward, J. (2003). Making Things Happen: A Theory of Causal Explanation. Oxford: Oxford
University Press.
6
Buddhist Dependence
Graham Priest
1 Introduction: Orientation
Many issues in Western philosophy were discussed with great sophistication in the
Eastern philosophical traditions. A prime example of this is metaphysical depend-
ence.1 This is absolutely central to Buddhist metaphysics.2 Indeed, there is a wide
variety of views about, in particular, the structure of metaphysical dependence.
In this essay, I will explain some of these views, and some of their ramifications.
The aim is neither to give a scholarly account of any of these views, nor to argue
for or against any one of them. Rather, the point of the essay is to open the eyes of
philosophers who know little of the Eastern philosophical traditions to important
possibilities of which they are likely to be unaware.
In Section 3 of this essay, I will explain three Buddhist positions concerning
metaphysical dependence: those of Abhidharma, Madhyamaka, and Huayan.3 In
Section 4, I will turn to some ways in which these positions engage with some Western
debates. But first, for those readers whose knowledge of the history and development
of Buddhist philosophy may be incomplete, I will explain enough of this in Section 2 to
situate what is to follow.
1 In contemporary Western philosophy, the topic is discussed under a variety of names, such as
ontological dependence and grounding. Moreover, there seems to be little unanimity as to whether there
is just one relationship here, or, if not, how the different varieties of the species are related. For general
discussions, see Tahko and Lowe (2015), and Bliss and Trogdon (2014). I use the term metaphysical
dependence as a catch-all term, to cover any sort of relationship concerning how some things depend for
whatever form of being they have on other things. More fine-grained distinctions are unnecessary for our
purposes.
2 I certainly do not want to suggest that the topic is unimportant in other Asian traditions, such as the
Vedic and Daoist ones. However, it is better for people who know more about these traditions than I do to
write about these matters.
3 Again, I do not want to suggest that there are not relevant and interesting matters in other parts of the
tradition, such as Yogācāra and Tiantai; but one can do only so much in one essay.
graham priest 
2 A Little History and Geography

Buddhist thought started with the historical Buddha, Siddhārtha Gautama. His dates
are uncertain, but he flourished around 450 bce; and his ideas were developed in a
canonical way for the next 500 years or so. The philosophical part of this development
was called Abhidharma (higher teachings). There were many Abhidharma schools.
The only one to survive to this day is Theravāda (Way of the Elders).
Around the turn of the Common Era, novel ideas emerged, which were critical
of the older tradition. This generated a new kind of Buddhism: Mahāyāna. The
foundational philosopher of this kind of Buddhism was Nāgārjuna. Dates are, again,
uncertain; but he flourished around 200 ce. He founded the version of Mahāyāna
Buddhism called Madhyamaka (Middle Way).
Buddhist thought died out in India around the twelfth century, but by that time
it had spread to the rest of Asia; Theravada going South East, and Mahāyāna going
North West into central Asia, and thence, across the Silk Route, into East Asia.
It entered China around the turn of the Common Era, where it met the indigenous
philosophical traditions: Confucianism and Daoism. Daoism, in particular, exerted
a crucial influence on Chinese Buddhist thought.4 This resulted in the emergence of
distinctively Chinese forms of Mahāyāna Buddhism, around the sixth century.
Some of these, such as Chan (Jap: Zen) are still extant. But perhaps the most
philosophically sophisticated of these flourished in China for only a few hundred
years (though it still has a presence in Korea and Japan), many of its ideas being
incorporated into other forms of Buddhism (and indeed, into Neo-Confucianism).
This was Huayan (Skt: Avatam . saka; Kor: Hwaeom; Jap: Kegon; Eng: Flower
Garland) Buddhism, named after the sūtra it took to be most important. The most
influential philosopher in this tradition was Fazang, traditionally dated as 643–712.5
3 Metaphysical Dependence in the Buddhist Traditions

With this background, let us turn to our three views concerning metaphysical depen-
dence.
3.1 Well-founded Buddhism
It is common to all Buddhisms that the world of our common experience is a world of
dependent origination, pratı̄tyasamutpāda. Nothing is permanent: things come into
existence when causes and conditions are ripe, and go out of existence in the same way.
Now, how should one think of a person in this context?6
4 Buddhism (Mahāyāna) entered Tibet relatively late in the piece, in the eighth century. The indigenous
Tibetan views did not have an impact of such magnitude.

5 For good introductions to the history of Buddhist thought, see Mitchell (2002), Siderits (2007), and
Williams (2009).
6 For what follows, see Siderits (2007), chs 3 and 6.
 buddhist dependence
The understanding of a person that developed in the Abhidharma literature was as

follows. Consider a car. This comes into existence when its parts are put together. The
parts interact with each other and the environment; they wear out and are replaced;
and they finally fall apart entirely. Persons are just like that. True, their parts (skandas),
unlike the car’s, are both material (rūpa) and mental. But otherwise the story is the
same. Of course we can think of this dynamically evolving bunch of parts as a single
thing, a person; we can even give it a name, say ‘Bertrand Russell’; but this is just a
matter of convenience.
The Abhidharma philosophers could see nothing special about people in this way.
Anything with parts, like our friend the car, is exactly the same. Indeed, what anything
in our common world of experience is, depends on what its parts are and how we think
about them.
So take the car, again, as an example. This depends on its wheels, engine, chassis,
and so on. The engine depends on its combustion chambers, fuel-injection system,
and so on. If we keep decomposing in this way, do we come to things where no
further decomposition is possible? The Abhidharma philosophers thought that the
answer was obviously: yes. If something is a conceptual construction, there must be
something, dharmas, out of which it is constructed. You can’t make something out of
nothing. This would seem to be the point when Asaṅga (fl. 4th century ce), in a late
Abhidharma text, says:
Denying the mere thing with respect to dharmas such as rūpa and the like, neither reality nor
conceptual fiction is possible. For instance, where there are the skandhas of rūpa etc., there
is the conceptual fiction of the person. And where they are not, the conceptual fiction of the
person is unreal. Likewise if there is a mere thing with respect to dharmas like rūpa etc., then
the use of convenient designators concerning dharmas such as rūpa and the like is appropriate.
If not then the use of convenient designators is unreal.7
There was some dispute about the nature of the dharmas. (A common view was
that they are tropes of some kind.) But, as all agreed, they are just as impermanent
as anything else; what distinguishes them is the fact that they are what they are
independently of anything else (parts, concepts, each other). They have svabhāva
(self-being).
The Abhidharma philosophers described the picture as one of two realities.8 There
is the fundamental reality composed of dharmas—ultimate reality (paramārtha-
satya); then there is the conceptual reality constructed out of this—conventional
reality (sam. vr.ti-satya).
Clearly, the whole picture paints a story concerning ontological dependence. Where
does it lie in the taxonomy of the Introduction to this volume? It is obviously some
7 Bodhisattvabhumi, 30–2. Translation by Mark Siderits.

8 The Sanskirt word is satya. This can mean either truth or reality. It is standard to translate the word as
truth. Of course if there are two realities, there are also two (sets of) truths: one about each of the realities.
But in the present context, and others that we will come to soon, the best translation is ‘reality’.
graham priest 
kind of foundationalism, the foundational elements being the dharmas. Does it

endorse anti-reflexivity, anti-symmetry, and transitivity? There is, as far as I know, no
explicit discussion of these matters in the texts, but let us extrapolate. The Abhidharma
philosophers would probably have endorsed transitivity. If the car depends on it its
engine, and the engine depends on its fuel injector, the car depends on its fuel injector.
Moreover, a whole would appear to depend on its parts, in a way that the parts do not
depend on the whole.9 So the dependence relation would seem to be anti-symmetric.
Since anti-symmetry entails anti-reflexivity, we have that as well. So this puts us in
case 2 of the taxonomy laid out in Section 2 of Chapter 0 in this volume.
3.2 Non-well-founded Buddhism
We now turn to Madhyamaka. Madhyamaka entirely rejected the notion of the
dharmas. Nothing has svabhāva. Everything is what it is by relating to other things. The
Madhyama philosophers accepted the Abhidharma view that the relations in question
could be mereological and conceptual, but also added a third important dimension:
causal. (Thus, persons are what they are, for example, because of their relations to their
parents, their genetic structure, etc.) Everything depends on other things in some or
all of these ways. That is, all things are empty (śūnya) of self-being.10
In much of his enormously influential text, the Mūlamadhyakamakārikā (MMK,
Fundamental Verses of the Middle Way) Nāgārjuna mounts the case that nothing has
svabhāva.11 He does this by running through all the things one might suppose to have
it (causation, consciousness, space, and so on), and rejecting each one. Many of the
arguments are reductio arguments. We assume that something has svabhāva and show
that this cannot be.12 We will not consider the arguments in any detail here.
More to the point in this context, one might expect Nāgārjuna to have rejected the
distinction between the two realities. But he does not (MMK XXIV: 8–10):
The Buddha’s teaching of the Dharma
Is based on two truths:
A truth of worldly convention
And an ultimate truth.
Those who do not understand
The distinction between these two truths
Do not understand
The Buddha’s profound truth.13
9 By ‘part’ here, I mean proper part, i.e. a part distinct from the whole.
10 For a discussion of this and what follows, see Siderits (2007), ch. 9, and Williams (2009), ch. 3.
11 It must be said that this is a highly cryptic text, and there can be significant differences as to how to
understand its claims. I try not to go beyond a general consensus in what follows.
12 The arguments themselves are often by cases, though the cases are not the ones familiar to Western
philosophy—true and false—but the four delivered by the catus.kot.i (Eng: four corners)—true, false, both,
and neither.
13 Translations from the MMK are from Garfield (1995). In this context, ‘Dharma’ means correct doctrine.
 buddhist dependence
Conventional reality is the world of our familiar experience. But if there are no things
with svabhāva, what is ultimate reality?
Though hardly explicit in the MMK, the view that emerged in Madhyamaka was
that ultimate reality is what is left if one takes the things of conventional reality, and
strips off all conceptual overlays: emptiness (Skt: śūnyatā; Chin: kong ) itself. One
might well think that this ultimate reality provides some foundational bedrock.14 It
does not. According to Madhyamaka, everything is empty, including emptiness itself.
In perhaps the most famous verse of the MMK (XXIV:18), Nāgārjuna says:
Whatever is dependently co-arisen
That is explained to be emptiness.
That, being a dependent designation,
Is itself the middle way.
Emptiness, as the verse says, is a dependent designation. That is, emptiness depends
on something. Conventional reality clearly depends on ultimate reality. But what does
ultimate reality depend on? It is hard to extract a clear answer to this question from
the MMK; let us set it aside for the moment.
We are now in a position to see how the Madhyamaka view fits into the taxon-
omy described in Section 2 of Chapter 0 in this volume. In general it takes over
the Abhidharma view, but simply rejects its foundationalism. That is, it endorses
Exendability. We are therefore in case 1.
3.3 Buddhist coherentism
Let us now turn to Huayan.15 This, like all Chinese Buddhisms is Mahāyāna, and
so inherited Madhyamaka thought. But whilst Madhyamaka held that all things
depend on some other things, the Huayan universalized: all things depend on all other
things. How did they get there? Come back to the question of what ultimate reality
depends on.16
As we have noted, Chinese Buddhism was indebted to Daoism. According to a
standard interpretation of this, behind the flux of phenomenal events, there is a
fundamental ineffable principle, dao , which manifests itself in the flux. To Chinese
Buddhist eyes, it was all too natural to identify the flux with conventional reality, and
the dao with ultimate reality. That is exactly what happened. Moreover, just as one
cannot have manifestations without whatever it is of which they are a manifestation,
one cannot have something whose nature it is to manifest, without the manifestations.
So conventional reality depends on ultimate reality, and ultimate reality depends on
conventional reality: they are two sides of the same coin. In his Treatise on the Golden
14 In which case, we still are in case 2 of the taxonomy described in Section 2 of Chapter 0 in this volume,
but G is true. Ultimate reality is the unique foundation.

15 For the following, see Williams (2009), ch. 6.
16 It must be said that these thoughts were available, in principle, to Madhyamaka, but no one ever
articulated them.
graham priest 
Lion, Fazang explains the point in this way.17 Imagine a statue of a golden lion. The
gold is like ultimate reality. The shape is like conventional reality. One cannot have the
one without the other.
By this time in the development of Buddhist thought, the objects of phenomenal
reality are called shi , and ultimate reality is referred to as li , principle. Hence we
have the Huayan principle of the mutual dependence of li and shi: lishi wuai .
The matter is put this way by the Huayan thinker Dushun (557–640) as follows:
Shi, the matter that embraces, has boundaries and limitations, and li, the truth that is embraced
[by things], has no boundaries or limitations. Yet this limited shi is completely identical,
not partially identical, with li. Why? Because shi has no substance [GP: svabhāva]—it is
the selfsame li. Therefore, without causing the slightest damage to itself, an atom can embrace
the whole universe. If one atom is so, all other dharmas should also be so. Contemplate on
this.18
But if every shi depends on li, then by the transitivity of dependence, every shi
depends on every other shi. Hence we have the Huayan thesis of the dependence
(interpenetration) of every shi on every other shi: shishi wuai . Chengguan
(738–839?), another Huayan thinker, puts the matter thus:
Because they have no Selfhood [GP: svabhāva], the large and the small can mutually contain
each other . . . Since the very small is very large Mount Sumeru is contained in a mustard seed;
and since the very large is the very small, the ocean is included in a hair.19
We therefore arrive at this: all things, whether li or shi, depend on each other.
The situation is depicted in what is arguably the most famous image in Huayan: the
Net of Indra. A god has spread out a net through space. At each node of the net there is
a brightly polished jewel. Each jewel reflects every other jewel, reflecting every other
jewel, reflecting . . . to infinity. Fazang puts the metaphor thus:
It is like the net of Indra which is entirely made up of jewels. Due to their brightness and
transparency, they reflect each other. In each of the jewels, the images of all the other jewels are
[completely] reflected . . . Thus, the images multiply infinitely, and all these multiple infinite
images are bright and clear inside this single jewel.20
Each jewel represents an object. And it is the nature of each jewel to encode every
other jewel, including that jewel encoding every other jewel, and so on.
So where is the Huayan picture in the taxonomy described in Section 2 of Chapter 0
in this volume? Clearly, this is coherentism, C, and we are in category 13 (since there
is more than one object).
17 The Treatise is translated into English as pp. 409–14 of Chan (1969).

18 Quoted in Chang (1972), pp. 144–5. The character translated as ‘identical’ is better translated in this
context as ‘interpenetrating’. See Priest (2015).
19 20
Quoted in Chang (1972), p. 165. Quoted in Liu (1982), p. 65.
 buddhist dependence
4 Western Connections
So much for the three Buddhist positions. As is clear, they are significantly different.
This is a striking feature of the tradition compared with Western philosophy, which,
for all its variety views, has been almost entirely foundationalist.21 For all that,
there are many interesting connections between the Buddhist views, and debates and
problems to be found in Western philosophy. In this section of the essay, I want to turn
to some of these. There is certainly no attempt to be comprehensive here: I have just
chosen some of the most obvious connections. I will structure this section by three
subsections mirroring those in Section 3.
4.1 Mereology
So let us return to the Abhidharma picture. Clearly, this is some kind of mereological
atomism, with the atoms being the dharmas—whatever they are. Why should one be
an atomist? The Abhidharmikas, as I noted, produced no real arguments for this: they
just seemed to think it obvious. But it isn’t. Just consider the real line, and let its parts
be all the nonempty sub-intervals. One is a part of another if it is a proper subset. Then
any part has parts, since any interval can be divided into a left and a right part. The
picture is perfectly coherent. So how might one argue that reality is not like that?
One famous answer was given by Kant, in the Second Antinomy of his Critique of
Pure Reason, and goes like this:
Let us assume that composite substances are not made up of simple parts. If all composition
then be removed in thought, no composite part, and (since we admit no simple parts), also
no simple parts, that is to say, nothing at all will remain, and accordingly, no substance will
be given. Either, therefore, it is impossible to remove in thought all composition, or after its
removal there must remain something which exists without composition, that is, the simple.
In the former case the composite would not be made up of substances; composition, as applied
to substances, is only an accidental relation in independence of which they must still persist as
self-subsistent beings. Since this contradicts our supposition, there remains only the original
supposition, that a composite substance is made up of simple parts.22
Kant’s argument is both dark and tangled—and, it should be remembered, he is

going to argue that it does not work (since the simple is a noumenon, and so the
categories cannot be applied to it). However, in nuce, it would appear to be this.23
Given any substance, it is always possible to decompose any compound part, at least
in thought. This is because the fact that something is arranged (composed) in a certain
way is always a contingent one. Now, take any substance, and suppose that it is not
composed of simples. Decompose it through and through. Nothing will be left, which
is impossible since, in that case, the substance would have had no substance.
21 See Bliss and Priest (2017), from which much of the above material comes.
22 23
A434 = B462ff. Translation from Kemp Smith (1933). See Priest (2002), p. 90.
graham priest 
But the argument would not seem to work—even setting Kantian scruples about
the noumenal. Take a substance, say the table on which I write. It is composed of
cells of wood. These are composed of molecules, which are composed of atoms,
which are composed of protons and electrons, which are composed of quarks,
which . . . Whether this regress does eventually terminate, we may never, in fact, know.
But there is nothing logically absurd about supposing that physics will find indefinitely
smaller and smaller particles (or maybe better, more and more fundamental kinds of
thing). The table is a substantial entity for all that.
Of course, none of this shows that atomism is false: merely that, at least as far as
this argument goes, there is no particular reason to suppose it true.
But let us grant its truth, at least for the sake of argument. Another Abhidharma
claim is that it is only the atoms which are ultimately real. The table, for example, has
no being over and above its atoms. It is simply a bunch of atoms ‘arranged table-wise’.
Again, I know of no very focused Abhidharma arguments for this view,24 but it does
comport with a certain intuition. Suppose I have a hydrogen atom composed of a
proton and electron, how many objects do I have: two or three? Say ‘three’ if you like,
but the hydrogen atom itself would seem to be, in the words of David Armstrong, an
“ontological free lunch”: no addition to being.25
Still, the atom would seem to have some kind of reality, unlike ghosts and
phlogiston. How are we to understand this? It is here that the Abhidharma distinction
between the two satyas kicks in. The table has a conventional reality, but not an
ultimate one. But how are we to understand this?
Recall that for the Abhidharmikas, the conventionally real objects are conceptual
constructions. That is, we have a concept, table, which we use to organise our thinking
about the world. Recall, also, that in mereology there is a debate concerning the
question of compositionality: when does a bunch of parts ‘fuse’ to form another
object?26 There are two extreme answers. The first is never: mereological nihilism.
This seems too extreme. In some sense, I am a perfectly good object, partite though
I be. The other is always: unrestricted composition. Every bunch of objects fuses to
form an object. This seems equally counter-intuitive. What sort of object is one whose
parts are: the number π , the rings of Saturn, and the Buddha’s left earlobe?
The most natural answer is a middle way, sometimes: special composition. But
when? Abhidharma provides a simple and natural answer to this: when the objects
fall under some concept. So tables and people are in; the bizarre object of the last
paragraph is not.
Of course, it is always possible to gerrymander a concept (e.g. simply by listing a
bunch of objects). The concepts must be ones which we actually employ to find our
way around the world (which does not entail that they have to be ‘common sense’ ones;
24 The nearest I know is to be found in the Malindapañha dialogue, part of which is translated in
Radhakrishnan and Moore (1957), pp. 281–4.

25 26
Armstong (1997), p. 12. See Varzi (2015) for discussion and references.
 buddhist dependence
the concepts of science also satisfy this rubric). The objects of conventional reality
are, then, those non-atoms delivered by the mereological principle of conceptually
constrained special composition.
This picture—whether or not it is correct—at least provides, at once, a natural
answer to the question of what, exactly, a conceptual construction is, and an answer
to the question of compositionality.
4.2 Causation
Let us turn next to the Madhyamaka picture. Madhyamaka took over the Abhidharma
picture, with a couple of very significant changes. First, and perhaps most importantly,
it ditched the dharmas, the things with svabhāva. The picture is quite coherent. Come
back to our model of the subsets of the real line; but now restrict the intervals in
question to those which are definable, say in the first-order language of real number
theory. (This gets rid of most of them, since there is only a countable number of
first-order definitions.) It is still true that every interval has sub-intervals, but now
every interval is also linguistically/conceptually isolable. Reality for a Madhyamaka is
like that.27
Neither is this an argument for idealism.28 True, things get to be in the domain
in virtue of there being certain concepts. But there is more to idealism than this.
For idealism holds not only that objects are conceptually dependent, but also claims
an ontological priority for the conceptual. This most certainly is not the case in
Madhyamaka. For concepts are as empty of svabhāva as anything else. They are what
they are in virtue of other things. What other things? Whilst, again, one does not find
a clear answer in the MMK, it is easy enough to produce one with the help of the
contemporary philosophy of language.29 What makes the concept dog the concept
it is, rather than, say, the concept cat? The fact that it relates in a certain way to the
canine creatures wandering the world. (If it related to feline creatures in the same way,
it would mean cat instead.) What exactly that relationship is, we might argue about;
but all that matters here is that the concept depends for its identity on being related
to things in the world in this way.30
And so an argument for the emptiness of all things emerges. Things in the world
depend on language and vice versa. But the picture is more complicated than that.
As noted, Madhyamaka takes over mereological and conceptual dependence from
Abhidharma, but adds a third kind of dependence: causal. And this brings us to the
second Madhyamaka break with Abhidharma.
It is absolute Buddhist orthodoxy that everything in the world is in a state of
pratı̄tyasamutpāda, coming into existence when caused to do so, and going out of
existence when caused to do so. Just as much as Madhyamaka, then, Abhidharma
27 In particular, we have another argument for the emptiness of ultimate reality. For as Nāgārjuna says,
emptiness is a “dependent designation”—thing denoted.
28 Though there was another school of Indian Mahāyāna that was idealist: Yogācāra.
29 30
e.g. Putnam (1973): “meanings ain’t in the head”. See Priest (2013), and (2014), §13.5.
graham priest 
philosophers took it that the dharmas were caused to exist and to cease to exist.
They just did not take these causal relations to be (partly) constitutive of the nature of
a dharma. The Madhyamaka did. (The first chapter of the MMK is a long analysis of
causation.)
This might certainly look like a mistake. It is fairly standard to distinguish between
causal dependence and metaphysical dependence.31 Causation determines when
something comes into existence, but not what it is.
But not so fast. What makes me the very person I am? Answer (in part): the way my
parents treated me, the education I had, my professional experiences, and so on. These
are causal factors. It might be thought that people are special in this way. Not so. What
makes something an oak tree? The fact that it grows out of an acorn, delivers acorns,
and so on. If it grew out of an onion, and delivered, not acorns, but goldfish, it would
not be an oak tree. So maybe it’s just biological entities that are like this? Again, no.
Take an electron. This is the kind of thing which repels particles of the same kind,
which is annihilated by positrons, and so on. If it were attracted by other particles of
the same kind, and annihilated by neutrons, it would not be an electron. Causation, it
would seem, can determine the nature of things, quite generally. One might certainly
contest the above considerations, but they have a certain persuasiveness.
So let us turn to another matter: the vexed issue of what to make of the notion
of ultimate reality and its relationship to conventional reality, once the notion of
svabhāva has gone out of the window. A distinctive view of the matter was given
by Candrakı̄rti (fl. first half of the seventh century), one of the most authoritative
commentators on Nāgārjuna, as follows:
The Buddhas, who have an unmistakable knowledge of the nature of the two truths, proclaim
that all things, outer and inner, as they are perceived by two kinds of subject (deluded
consciousness on the one hand and perfectly pure wisdom on the other), possess a twin
identity . . . They say that the object perceived by authentic primordial wisdom is the ultimate
reality, whereas the object of a deluded perception is the relative truth.32
That is, there is only one reality, but it has a dual nature, a double aspect. When
perceived correctly, its ultimate aspect is seen. When perceived incorrectly—that is,
by ordinary benighted beings like you and me—only its conventional aspect is seen.
But if these are both objective aspects of the one reality, what makes the one
any better (more ultimate) than the other? Come back to Kant again. According to
his transcendental idealism, our perceptions (‘intuitions’)—say of a table—are the
product of two things: a raw sensory input and a mental imposition: the forms of
space and time, and the concepts of the understanding. The empirical object, then,
has these dual aspects, and one of these involves conceptual imposition.
Candrakı̄rti’s account of reality may be thought of in the same way. (Though one
should not push the analogy too far. There is no suggestion in Candrakı̄rti that
31
See e.g. the first few sentences of Tahko and Lowe (2015).
32 Padmakara Translation Group (2004), p. 192.
 buddhist dependence
the concepts are universal and a priori. So much the better for Candrakı̄rti.) The
difference between Candrakı̄rti and Kant concerns neither the dual nature, nor the
conceptual overlay, but in our access to the conceptually naked. Seeing such a thing
is an impossibility for Kant. Our perceptual apparatus just doesn’t work that way. But
it is what you would see if, per impossible, you were able to do this. For Candrakı̄rti
there is no such impossibility. Difficult it may be; impossible it is not. It is exactly what
training in certain meditative practices gives you. Here is a serious and important
difference between our two philosophers; I do no more than note it here. More
importantly, and to return to our question of the previous paragraph, the one aspect of
reality is more ultimate than the other, precisely because it dispenses with an extrinsic
conceptual overlay.
4.3 The missing link
So let us turn finally to the Huayan picture. This moves from the claim that all things
depend on some other things to the claim that all things depend on all other things.
The crucial move here is to find something on which all things depend, and which
depends on all things. Transitivity then does the rest. In the Huayan story, it is li (an
intellectual descendent of dao) that plays this role. Are there other things which might
plausibly be thought to do so, so that the move in the argument does not need to be
underpinned by specifically Buddhist ideas?
As a first approach, come back to causation. Everything, physics tells us, is causally
derived from the Big Bang. Everything depends for what it is, then, at least in part, on
this. That gives us half the story we need. What of the other half? Could the Big Bang
depend for its identity on the things it produces? That is not so obvious. And even if
it is, in fact, the case, we have to worry about not just things in space and time, but
abstract objects, such as numbers—assuming there to be such. For these, we do not
have even one-way dependence on the Big Bang.
As a more promising candidate for a link, take the object which is the mereological
sum of everything: the whole of what is, W. The existence of this would seem to be
delivered by our account of special composition. We certainly have a conception of
such a totality: it does not seem to be at all gerrymandered. Now, a natural view is
that a whole depends on its parts. (Maybe not necessarily the very parts it has now.
Arguably the parts of a car can change while it remains that very car. But you could
not have a car without parts.) So W depends on all its (proper) parts.
What about dependence in the other direction: do the parts depend on the whole?
One can certainly make a case for this in some cases. Thus, Aristotle argued that a
hand, for example, would not be a hand unless it were integrated into the functioning
whole of a body.33 Aristotle’s claim has been generalized by Schaffer, who argues
33
See Parts of Animals, esp. 640b 34–641a 10.
graham priest 
that any object depends on W.34 At root, one might think of the whole as a single
functioning entity: we just normally fail to appreciate the deep inter-dependences.
Schaffer, it is true, holds that such dependence is anti-symmetric. So he would not
endorse the claim that the dependence also goes the other way. However, he gives no
argument for this, but simply assumes anti-symmetry. The considerations suggesting
dependence in the other direction still obtain. Given these, W depends in the objects
that are its parts, and these depend on W.35
A completely different route to a missing link comes from another direction.36
Any object is an object. It could not be the very thing it is unless it were an object.
Hence its nature depends on a certain relationship with the property of being an
object, or objecthood, to make the reference to universals explicit.37 Now, if one is
a Platonist about universals, there is no hope of getting a dependence in the other
direction. Plato’s forms epitomize beings with svabhāva. But if one is an Aristotelian
about universals, the matter is different. For Aristotle, there can be no uninstantiated
universals. The universal of being human depends on the humanity of Socrates, and so
Socrates, the humanity of Plato, and so Plato, and so on. And the universal objecthood
depends on objects. We have, then, the symmetric dependence relations we need.
We may recast the whole matter in Heideggerian terms. The driving question
behind Heidegger’s thought is exactly ‘What is being?’, and to be for Heidegger is
exactly to be an object.38 And being is that in virtue of which beings are. As Heidegger
puts it in Sein und Zeit:
What is asked about in the question to be elaborated is being, that which determines beings
as beings, that in terms of which beings have always been understood no matter how they are
discussed.39
Beings, then, depend on being. But Heidegger is an Aristotelian about the matter.
Being is always the being of some object. Again as he puts it:
If we think of the matter just a bit more rigorously, if we take more heed of what is in contest in
the matter, we see that Being means always and everywhere: the Being of beings.40
So being depends on beings as well. We have our symmetrical dependence.
34 Schaffer (2010). Though Schaffer is careful to restrict his concern to just the physical. He also uses a
version of the causal argument from the Big Bang. This delivered an entangled quantum state, in which
every object is dependent on the whole.
35 I note that this is exactly what the Huayan accepted. For them, any whole depends on its parts, and
any part depends on the whole. See Jones (2009).

36 I note that there is also at least one more candidate for the required linking concept: nothingness.
I have explored that matter in Priest (2014), ch. 13. As we are about to see, the universal of oneness, that is,
of being an object, could also play this role, though this had not occurred to me when I wrote the book.
37 The relationship between a universal and its instantiations is not exactly the same as that between
a whole and its parts, but it is very similar. For a gentle introduction to the matter, see Garrett (2006),
pp. 37ff.
38 For a general discussion of the matter, see Priest (2014), ch. 4.
39 40
Stambaugh (1996), pp. 4f. Stambaugh (2002), p. 69.
 buddhist dependence
I note that, for Heidegger, one cannot say what being is. (For to do such would be
to treat it as a being, which it is not.) Being shows itself in the way that beings present
themselves—to those with the eyes to see it. I note that at this point we are not so far
away from the dao which cannot be described, but which manifests itself in “all under
heaven”.41
5 Conclusion
So ends our somewhat whistle-stop tour of some Buddhist views on ontological
dependence, and some of their Western connections. Most of this has been written
with Western philosophers who know little of Eastern traditions in mind; but I hope
that some of it will be of interest to those who know of Buddhist philosophy, but,
perhaps, less of Western philosophy. I have done nothing here to try to evaluate the
views we have met, or determine their truth. The exercise has been one of urban
geography. To use a metaphor I have used before:42 Philosophy is like a city. It has
relatively self-contained suburbs, such as metaphysics, ethics (each with their own
neighborhoods). But only relatively: the connections spread in a network over the
city, sometimes in the most surprising of ways. Nor is this a finished city: remarkable
new buildings are going up all the time. All I have done in this essay is to describe one
of the Eastern neighborhoods, and explore some of the connections which cross the
city’s Berlin Wall. Of course, a philosopher can live quite happily in just one half of
the city—or even just one of its suburbs—the whole of their thinking lives. But their
philosophy cannot but be richer and deeper, the more they know of the city. Such is
the spirit in which this piece is written.
References
Armstrong, D. M. (1997), A World of States of Affairs, Cambridge: Cambridge University Press.
Bliss, R. L. and Priest, G. (2017), ‘Metaphysical Grounding, East and West’, pp. 63–85 of
S. Emmanuel (ed.), Buddhist Philosophy: a Comparative Approach, Wiley-Blackwell, 2017.
Bliss, R. L. and Trogdon, K. (2014), ‘Metaphysical Grounding’, in E. Zalta (ed.), Stanford
Encyclopedia of Philosophy, http://plato.stanford.edu/entries/grounding/.
Chan, W. T. (ed.) (1969), A Sourcebook in Chinese Philosophy, Princeton, NJ: Princeton
University Press.
Chang, G. C. C. (1972), The Buddhist Teaching of Totality: the Philosophy of Hwa Yen Buddhism,
London: George Allen & Unwin Ltd.
Garfield, J. (1995), The Fundamental Wisdom of the Middle Way, New York, NY: Oxford
University Press.
Garrett, B. (2006), What is this Thing Called Metaphysics?, Abingdon: Routledge.
41 As the famous opening lines of the Daodejing put it, “The Dao that can be described in language is not
the constant Dao; the name that can be given it is not the constant name”: Lynn (1999), p. 51.
42 Priest (2011).
graham priest 
Jones, N. (2009), ‘Fazang’s Total Power Mereology’, Asian Philosophy 19: 199–211.
Kemp Smith, N. (1933), Immanuel Kant’s Critique of Pure Reason, 2nd edn, London: Macmillan
& Co.
Liu, M. W. (1982), ‘The Harmonious Universe of Fazang and Leibniz’, Philosophy East and West
32: 61–76.
Lynn, R. J. (1999), The Classic of the Way and Virtue: A New Translation of the Tao-te Ching of
Laozi as Interpreted by Wang Bi, New York, NY: Columbia University Press.
Mitchell, D. (2002), Buddhism: Introducing the Buddhist Experience, Oxford: Oxford University
Press.
Padmakara Translation Group (2004), Introduction to the Middle Way: Candrakı̄rti’s Madhya-
makāvatāra with a Commentary by Jamgön Mipham, Boston, MA: Shambala.
Priest, G. (2002), Beyond the Limits of Thought, 2nd edn, Oxford: Oxford University Press.
Priest, G. (2011), ‘Why Asian Philosophy?’, pp. 211–21 of G. Oppy and N. Trakakis (eds),
The Antipodean Philosopher. Volume 1: Public Lectures on Philosophy in Australia and
New Zealand, Lanham, MD: Lexington Books.
Priest, G. (2013), ‘Between the Horns of Idealism and Realism: The Middle Way of Madhya-
maka’, ch. 13 of S. M. Emmanuel (ed.), A Companion to Buddhist Philosophy, Chichester:
Wiley-Blackwell.
Priest, G. (2014), One, Oxford: Oxford University Press.
Priest, G. (2015), ‘The Net of Indra’, pp. 113–27 of K. Tanaka, Y. Deguchi, J. Garfield, and
G. Priest (eds), The Moon Points Back, Oxford: Oxford University Press.
Putnam, H. (1973), ‘Meaning and Reference’, Journal of Philosophy 70: 699–711.
Radhakrishnan, S. and Moore, C. (eds) (1957), A Sourcebook in Indian Philosophy, Princeton,
NJ: Princeton University Press.
Schaffer, J. (2010), ‘Monism: The Priority of the Whole’, Philosophical Review 119: 31–76.
Siderits, M. (2007), Buddhism as Philosophy, Aldershot: Ashgate.
Stambaugh, J. (trans.) (1996), Being and Time, Albany, NY: State University of New York Press.
Stambaugh, J. (trans.) (2002), Identity and Difference, Chicago, IL: University of Chicago Press.
Tahko, T. E. and Lowe, E. J. (2015), ‘Ontological Dependence’, in E. Zalta (ed.), The Stanford
Encyclopedia of Philosophy, http://plato.stanford.edu/entries/dependence-ontological/.
Varzi. A. (2015), ‘Mereology’, in E. Zalta (ed.), Stanford Encyclopedia of Philosophy, http://plato.
stanford.edu/entries/mereology/.
Williams, P. (2009), Māhāyana Buddhism: The Doctrinal Foundations, 2nd edn, London:
Routledge.
7
Bicollective Ground
Towards a (Hyper)Graphic Account
Jon Erling Litland
1 Introduction
Most authors on grounding hold that grounding is left-collective: there are some truths
γ0 , γ1 , . . . such that, without any one of the γi grounding φ on its own, taken together
the truths γ0 , γ1 , . . . nevertheless ground φ. (A standard example is the grounding of
a conjunctive truth in its conjuncts.) Could grounding also be right-collective? In the
simplest case: could a truth φ ground some truths γ0 , γ1 , . . . taken together without
the truth φ grounding any of the truths γi on its own? More generally, let us say that
grounding is bicollective if it is both left- and right-collective.1
If bicollective ground is intelligible an interesting kind of metaphysical coherentism
becomes a live option. Just like an epistemological coherentist may say that it is
a mistake to ask, about a particular belief, what makes it justified, a metaphysical
coherentist may say that it is a mistake to ask, of a particular truth, what grounds it.
In the epistemological case we should rather ask of some beliefs, taken together,
what makes them justified; in the metaphysical case, we should rather ask, of some
truths, what grounds them. A considerable advantage of this version of metaphysical
coherentism is that one does not have to countenance circles of ground in order to be
a metaphysical coherentist.2
The intelligibility of bicollective ground was first argued for by Dasgupta (2014b) in
the course of defending various structuralist theses. Recently, Litland (2016) showed
how some sense can be made of the notion by developing a logic of bicollective
ground using Fine’s truthmaker semantics. While the truthmaker semantics is by far
1 A note about terminology. Dasgupta (2014b) speaks of grounding being “irreducibly plural” and
Litland (2016) speaks of grounding being “many–many”. I believe the present terminology is better; the
issue is not whether there are grounding operators that take many (or a plurality) of arguments on the
right; the issue is whether grounding is collective (non-distributive) on the right. The point is purely
terminological—nobody has been confused on this point.
2 Compare the discussion of “reciprocal” essence in (Fine 1994, pp. 65–6).
jon erling litland 
the most developed approach to the logic of ground,3 it is nevertheless problematic as

a semantics for ground.4 These problems, as we will see, are particularly pronounced
in the case of bicollective ground.
A different—in many ways more natural—approach to the logic of ground is
(hyper)graph-theoretic.5 The main contribution of this paper lies in showing how to
extend the (hyper)graph-theoretic account of ground to the bicollective case. (Along
the way we correct some minor infelicities in previous presentations of the graph-
theoretical account of ground.) We also—on the philosophical side—sketch how
bicollective ground is naturally applied to mathematical structuralism.
1.1 Overview
We begin in §2 by introducing the central notion of immediate strict full ground.
In §3 we develop some ways of making sense of the characteristic non-distributivity
of bicollective ground and argue that mathematical structuralists should avail them-
selves of bicollective ground. In §4 we rehearse the truthmaker semantics for bicol-
lective ground and point out some problems that arise in the bicollective case.
In §5 we recall the graph-theoretic account for the left-collective case and argue
against Fine’s principle of Amalgamation. The main contribution of the paper comes
in §6 where we develop the graph-theoretic account of bicollective ground. We discuss
how to define acyclic graphs, mediate ground, the notions of partial ground, and what
it is for two collections of truths to be ground-theoretically equivalent. We conclude
with some questions for future research (§7).
2 Notions of Ground
The central notions of ground for the graph-theoretic approach are strict full mediate
(<) and immediate ground ().6 For notational convenience, we here treat ground as
a binary relation between multisets of truths. Claims of ground are then generated as
follows: whenever , are multisets of truths < and are claims of ground.7
3 Apart from its applications to ground (Fine 2012b,c Litland 2016) the truthmaker theory has an
impressive range of further applications: to counterfactuals (Fine 2012a), intuitionistic logic (Fine 2014),
and partial content (Fine 2016).
4 For the case of left-collective ground this has been forcefully argued by deRosset (2013, 2015).
5 This is the approach taken (for the case of partial ground) by Schaffer (2009) (a more general approach
is sketched in (Schaffer 2016), by deRosset (2015), and by Litland (2015, 2017). The talk of “trees” in (Correia
2014) and (Rosen 2010) is, as we shall see, closely related to the graph-theoretical approach.
6 We introduce further notions of ground later.
7 Three notes on grammar. First, note that and are multisets, not sets. Unlike a set a multiset is
sensitive to repetion: while the set {a, a} is identical to the set {a}, the multiset {a, a} differs from the set
{a} in that it has two occurrences of the member a. It turns out that it is important to work with multisets.
(See §5.1 below.) Second, we allow both and to be empty. As in (Fine 2012b, 47–8) we distinguish
sharply between a truth’s being ungrounded and a truth’s being zero-grounded, that is, grounded, but by the
empty collection of truths. (While zero-grounding will play no positive role in this paper it is important to
develop the theory leaving room for zero-grounding.) Third, grounding, of course, is not a relation between
 bicollective ground: towards a (hyper)graphic account
These notions of ground are to be understood as follows. If < then the truths
provide a full explanation of the truths —nothing needs be added to the truths in
order fully to explain why the truths are the case. The explanation is strict in the
sense that cannot in turn be part of an explanation of . It is strict full ground in
its left-collective form that has been the focus of most of the recent work on ground.
is an immediate strict full ground for if ’s grounding “does not have to
be seen to be mediated”. The guarded phrase is required since there are cases where a
truth φ is both a mediate and an immediate ground for a truth ψ. A standard example
is φ and the disjunction φ ∨ (φ ∨ φ). The ease with which we can treat immediate
ground is one of the main advantages of the graph-theoretic approach.
3 Distribution Failure
That some sense can be made of grounding many truths is not in question.8 What
might be problematic about bicollective ground is that it is non-distributive: how can
some truths ground some truths δ0 , δ1 , . . . even though for each i and every ⊆
, the truths do not ground δi ? But there are pictures of grounding that make it
comprehensible how distributivity could fail.
3.1 The wall
One way of thinking about grounding is that the grounded “rests on” the grounds;
the grounds “support” the grounded. It is widely accepted that the grounds have to
be “relevant” to the grounded. (Consider, e.g. the consensus that mere necessitation is
not sufficient for grounding.) If the grounds “support” the grounded the support has
to be “relevant”.
Seen in light of this metaphor non-distributive grounding arises quite naturally.
Figure 7.1 depicts a wall made of bricks, the upper row being supported by the
lower row. There is sense in which no one brick in the upper row rests on any
collection of bricks from the lower row. Consider, for example, the bricks a, b,
and c. b and c provide support for a; but since b, c take up more space than is needed
stably to support a they do not relevantly support a. Taken as a whole, however, the
upper row is relevantly supported by the lower row (taken as a whole).
(multi)sets; it is either a multigrade relation between truths or (better) it should be treated as a variable arity
sentential operator. We can accommodate this by reading the notation as follows. “If and are multisets
of sentences then the result of writing the sentences in (in any order) followed by a grounding operator
followed by the sentences in (in any order) is a sentence.” (Similarly, if one prefers to treat ground as a
multigrade relation between truths.)
8 The notions of simultaneous and distributive many–many ground are both definable in terms of
left-collective ground (Fine 2012b, p. 54). simultaneously grounds ψ0 , ψ1 , . . . iff grounds each ψi .
distributively grounds if and have decompositions into 0 , 1 , . . . and δ0 , δ1 , . . . respectively such
that i grounds δi for each i.
jon erling litland 
b c
Figure 7.1. Grounds as support
This is obviously quite metaphorical, but the metaphor is useful for forming the
right intuitions about bicollective ground. As we will see, it provides a useful way of
thinking about the hypergraphs.
3.2 Ground as (holistic) explanation

Many philosophers have tied grounding closely to explanation.9 It will be useful to
think of the explanations associated with ground as constituted by a special class
of explanatory arguments, where the explanatory arguments in turn are understood
as composed of explanatory inferences. (The explanatory inferences correspond to
immediate ground, while the explanatory arguments correspond to mediate ground.)
In the bicollective setting we have to generalize our conception of argument and
inference: we have to allow inferences with many conclusions:
• q0 , q1 , . . . . Therefore. p0 , p1 , . . .
Multiple conclusion inference is, of course, not new; what is distintive about the use
to which we will put it here is that the conclusions are read conjunctively.
If there is “holistic explanation” then this connection between grounding and
explanation should alert us to the possibility of bicollective grounding. By “holistic
explanation” I do not mean circular explanations; what I have in mind is rather the
following. We have some truths φ0 , φ1 , . . . (that describe some complex system S, say)
and it is impossible to explain these truths one by one; rather, what we have to do is
to explain the truths φ0 , φ1 , . . . together, and all at once.
A natural place to look for the right sort of holistic explanation is in various
structuralist metaphysical views. Indeed, when Dasgupta introduced bicollective
ground it was in order to make sense of structuralist views. His chosen examples
were qualitativism about individuals and relationalism about quantities; to my mind,
however, it is mathematical structuralism that provides the cleanest motivation for
bicollective ground.
9 Some—Dasgupta (2014a,c), Litland (2015, 2017)—hold that to say that grounds φ simply is to say
that explains φ in a distinctive way; others—Schaffer (2012, 2016), Audi (2012)—hold that grounding is a
relation “underlying” or “underwriting” a distinctive type of explanation. While the difference is important
nothing said here will turn on this.
 bicollective ground: towards a (hyper)graphic account
3.3 Mathematical structuralism and bicollective grounding

There are many positions that can be described as structuralist, but a natural thesis
for a grounding theorist is that all mathematical truths are grounded in wholly
structural truths. To state this more precisely we require some terminology. Consider a
mathematical system S and a truth φ(a0 , a1 , . . . , an ) that concerns some objects in S.10
We need a notion of two truths being identical.11 Say that a truth φ is wholly structural
if for all automorphisms π of S the truth (expressed by) φ(π(a0 ), . . . , π(an )) is
identical to the truth (expressed by) φ(a0 , a1 , . . . , an ).
The precise thesis is then:
(Structuralism) For every system S and every truth φ about some objects in S,
either φ is wholly structural or φ is grounded in wholly structural truths.
A famous problem for structuralist views is presented by the complex field (Burgess
1999; Keränen 2001). The two square roots of −1, that is, i and −i, are indiscriminable
in the sense that there are automorphisms interchanging them. This makes it difficult
to see how the truth that i exists can be grounded in wholly structural truths: there is
no wholly structural truth about the complex field that bears on the existence of i (as
opposed to the existence of −i). For the same reason, it is very plausible that there is
no wholly structural truth about the complex field that grounds the existence of −i.
The structuralist has two ways out. The first way out invokes bicollective ground:
while neither the truth that i exists nor the truth that −i exists are individually
grounded in wholly structural truths, the truths that i exists, −i exists, taken together,
are grounded in purely structural truths. There is no problem about how wholly
structural truths may bear on the pair of truths i exists, −i exists: since i and −i are
the unique square roots of −1, the pair of objects (i, −i) is uniquely characterized in
structural terms. The second way out holds that the truths that i exists and −i exists
have exactly the same wholly structural grounds: they are, we may say, commonly
grounded in the wholly structural.12
There are, I believe, strong reasons to opt for the first view and accept bicollective
ground; I will indicate some of those reasons below. A full defense of this claim,
however, lies beyond the scope of this paper; for present purposes it suffices that the
first view has some plausibility and that its formulation requires bicollective ground.13
10 We make may take a system to be a tuple S = D, c , c , . . . , f , f , . . . , R , R , . . .. Here D is the

0 1 0 1 0 1
domain of objects, c0 , c1 , . . . some designated objects, f0 , f1 , . . . some functions, and R0 , R1 , . . ., some
relations. For more detail about this way of setting things up and some related discussion of the problems
raised by non-trivial automorphisms, see (Linnebo and Pettigrew 2014).
11 We do not strictly speaking need to reify truths: instead of a relation of equality between truths we
could use a sentential operator like “what it is for . . . to be the case just is for . . . to be the case”.
12 We should also add that they do not have any non-structural grounds that are not themselves
grounded in the wholly structural.

13 One might think that there is a third possibility. One might think that there is a single truth: (i, −i)
exists; and that it is this truth that is grounded in the wholly structural. (One may understand this single
truth either in terms of a variably polyadic existence predicate or as ascribing existence to the plurality
jon erling litland 
4 The Truthmaker Account of Bicollective Ground

4.1 The theory presented
The truthmaker semantics is based on the idea that truths can obtain in different
(determinate) ways: the ways in which a truth can obtain are its metaphysical verifiers.
The
metaphysical verifiers are mereologically
structured. For every set of verifiers
f0 , f1 , . . . there is a verifier ( f0 , f1 , . . . ) that is the fusion of f0 , f1 , . . .. We can
then say that two sets of verifiers f0 , f1 , . . . , and g0 , g1 , . . . are equivalent if the
fusion of f0 , f1 , . . . is identical to the fusion of g0 , g1 , . . ..
This allows us to characterize a notion of bicollective (weak) ground. We say that
φ0 , φ1 , . . . ground ψ0 , ψ1 , . . . iff for any ways f0 , f1 , . . . for φ0 , φ1 , . . . to be the case
there
are ways g ,
0 1 g , . . . for ψ0 , ψ1 , . . . to be the case such that f ,
0 1f , . . . is equivalent
to g0 , g1 , . . . .
More precisely, a state space is a pair F, . Here F is a set of objects and
: P (F) → F is a total function from the powerset of F to F. The members of F
represent metaphysical verifiers and represents the fusion operator. The order and
the number of times one fuses some verifiers makes no difference to the resulting
fusion; more formally, is associative in the sense that when each Xi is a collection
of verifiers and each Xj is a collection of verifiers, then

( Xj ∪ { (Xi ) : i ∈ I}) = ( Xk )
j∈J k∈I∪J
If X = {a0 , a1 , . . .} we write (X) as a0 · a1 · . . . .

Let V ⊆ F. We say that V is closed if for all non-empty V0 ⊆ V, (V0 ) ∈ V.
A proposition over F, is simply a closed subset of F. Let {Pi }i∈I be a collection
of propositions (that is, closed subsets of F). We define (Pi )—the fusion of the
propositions Pi —as the set of pointwise fusions:
i∈I (Pi ) = { ({a0 , a1 , . . .}) : ai ∈ Pi }
We may write (P0 , P1 , . . .) for i∈I (Pi ).
We can now define the notions of weak full (≤), strict full (<), weak partial (),
and strict partial (≺) ground familiar from Fine (2012c). (In the following definition
we assume that P0 , P1 , . . . and Q0 , Q1 , . . . are propositions over F.)
Definition 4.1.
(i) P0 , P1 , . . . ≤ Q0 , Q1 , . . . iff (P0 , P1 , . . .) ⊆ (Q0 , Q1 , . . .).
(i, −i).) While I have no objection to there being a single truth that (i, −i) exists, I do not think this gives
us a third alternative. For the question arises: what is the relationship between the one truth that (i, −i) exist
and the two truths that i exists and that −i exists? Again we either have to hold that the truths that i exists,
that −i exists are commonly grounded in the wholly structural or we have to hold that they are grounded
in the wholly structural, even though they are not individually grounded in the wholly structural.
 bicollective ground: towards a (hyper)graphic account
(ii) P0 , P1 , . . . Q0 , Q1 , . . . iff there are R0 , R1 , . . . such that

P0 , P1 , . . . , R0 , R1 , . . . ≤ Q0 , Q1 , . . .
(iii) P0 , P1 , . . . < Q1 , Q2 , . . . iff P0 , P1 , . . . ≤ Q0 , Q1 , . . . and it is not the case that
Q0 , Q1 , . . . P0 , P1 , . . ..
(iv) P0 , P1 , . . . ≺ Q0 , Q1 , . . . iff P0 , P1 , P2 , . . . Q0 , Q1 , . . . and it is not the case that
Q0 , Q1 , . . . P0 , P1 , . . . .
Just as in the left-collective case the basic notion of ground for the truthmaker
semantics is weak full ground. Strict full ground is understood in terms of weak full
ground—as “irreversible” weak full ground.14 strictly fully grounds iff weakly
fully grounds and does not in turn contribute to grounding ; that is, does
not weakly partially ground .
4.2 Structuralism and the truthmaker semantics
The truthmaker semantics identifies a truth φ with the set of states verifying φ. It
is worth mentioning that if one accepts the truthmaker semantics for bicollective
ground one is forced to adopt the bicollective development of mathematical struc-
turalism. For suppose one held that the truths that i exists, −i exists were commonly
grounded. (That is, that the grounds for the truth that i exists are exactly the same
as the grounds for the truth that −i exists.) One would then be forced to accept the
absurd conclusion that the truth that i exists is identical to the truth that −i exists.
By going bicollective one avoids this problem: on their own the truths that i exist
and that −i exist are not grounded in the wholly structural: it is only together that the
two truths are grounded in the wholly structural.
While this is a pleasing result the problematic features of the truthmaker semantics
mean that one should not put too much weight on it. Let us now turn to discussing
some of these problematic features.
4.3 Problems with the truthmaker semantics
The most important problem with the truthmaker semantics is its incapability of
dealing with immediate ground. We can see this as follows (working in the left-
collective case). For let R be a truth having as its verifiers p, q, and p·q. Let P be verified
by p. Is R immediately or only mediately grounded in P? There is no way of telling. If
R is the proposition P ∨ (P ∨ Q) then R is immediately and mediately grounded in P;
if R is the proposition (P ∨ P) ∨ Q, then R is only mediately grounded in P. As we will
see, the graph-theoretical approach deals with cases like this with ease.
In case one is dubious about the distinction between immediate and mediate
ground, however, it is worth noting that the truthmaker semantics also is problematic
as an account of mediate ground.
14
For more details about the truthmaker semantics see (Litland 2016).
jon erling litland 
An initial worry arises from the extensional character of the semantics. The content
of a collection of propositions is the set of fusions of the verifiers of the γ ∈ .
By inspection of the clauses for ≤ we see that ≤ is true if the content of is
included in the content of . But from the fact that the content of is included in the
content of why think that the obtaining of in any way explains the obtaining of
? In particular, why think that one can get from to by means of a sequence of
explanatory inferences?
A decent response to this worry is to hold that while the truthmaker semantics may
not give us the intended semantics of the logic of ground it still provides a serviceable
model theory for the (pure) logic of ground. This response might work in the left-
collective case.15 In the bicollective case, however, the truthmaker semantics validates
some problematic principles.
Let me begin with the principle of Self-Ground. As one can easily check, this
principle is valid:
<
Self-Ground
< ,
In other words, a collection of truths can strictly ground collections of truths that
contain .
That some notion of strict ground allows for instances of Self-Ground is not absurd.
After all, there is no conflict with the definition of asymmetry. To say that strict full
ground is asymmetric is to say that if < , then does not ground . It follows
that for no do we have that strictly fully grounds . But it does not follow from
this that does not strictly fully ground some
, where ⊂
.
What is problematic is that in the truthmaker semantics one cannot hope to define
a notion of strict ground that does not validate Self-Ground.16 If one ties grounding to
explanatory arguments one would not want Self-Ground to be validated: if there is an
explanatory argument from to why think that there is an explanatory argument
from to , ?
Much the same problem arises with the principle of Squeezing:17
< < ,
,
Squeezing
< ,
In words: if a collection of truths ,

is “squeezed” between two collections and
,
, that both are strictly fully grounded in , then ,
is grounded in .
15 Though, as we will see, there are worries about Amalgamation.

16 Strengthening the requirement of asymmetry to require that if strictly fully grounds then there
are no 0 , 1 such that , 0 weakly fully grounds , 1 will not work. Since we may always set 0 = 1
to be the totality of all propositions, this strengthening would ensure that there are no cases of strict full
ground.
17 For a proof of Squeezing, see (Litland 2016).
 bicollective ground: towards a (hyper)graphic account
But this is problematic. This is, again, perhaps most easily seen if we tie grounding
to explanation. Why think that there is an explanatory argument from to ,
when
we have explanatory arguments from to and also to ,
, ? What guarantees
that there is such an argument?
All these problems can be overcome by developing a hypergraph-theoretic account
of bicollective ground.
5 The (Hyper)Graphic Approach: The

Left-Collective Case
Since the bicollective case leads to complications we first rehearse the theory for
the left-collective case. In passing we correct some minor infelicities in previous
presentations of the graph-theoretic approach for left-collective ground (deRosset
2015; Litland 2015). If V is a set let P (V) be the set of multisets of V.18
Definition 5.1. A left-collective directed hypergraph is a tuple G = VG , AG , tG , hG . Here VG
is a collection of vertices. AG is a collection of hyperarcs. tG , hG are functions AG → P(VG ).
If A ∈ AG , tG (A) is the tail of A and hG (A) is the head of A. We demand that the cardinality of
hG (A) is 1 for each A ∈ AG .
We allow t(A) to be of any cardinality; in particular, we allow t(A) = ∅, in this way

we make room for Fine’s notion of zero-grounding. When no confusion is likely to
arise we freely drop the subscripts from VG , AG , hG , tG .
Intuitively, the vertices represent truths and an arc A represents that the truths t(A)
immediately ground the truth h(A).19 We occasionally refer to heads and tails as limbs.
In having h(A) and t(A) be multisets the present account differs from the previous
ones. Since we take V to consist of truths we will often use φ, ψ, . . . to range over the
elements of V and use , ,
, . . . to range over multisets of elements of V.
The following graphical depiction of a hyperarc with tail φ0 , φ1 , . . . and head ψ
might be useful:
ϕ0
ψ
ϕ1
…
Since we are interested in capturing notions of strict ground we impose an acyclicity

condition.
18 For cardinality reasons we impose a limit on the amount of repetition that is allowed. (What limit
we impose does not matter as long as we allow arbitrary finite repetition of elements.) For definiteness,
letting λ be the least strongly inaccessible cardinal greater than the cardinality of V, we allow an element
in a multiset to have κ-many occurrences for each κ < λ.
19 Strictly speaking h(A) is a multiset containing a single truth, but for ease of expression we will refer
to h(A) as a truth. No confusion should ensue.

jon erling litland 
Definition 5.2. Let G = V, A, t, h be a graph. A path in G is either a sequence φ where φ ∈ V

or a sequence φ0 , A0 , φ1 , A1 , . . . , An−1 , φn such that for each i, φi ∈ t(Ai ) and φi+1 =
h(Ai ). The length of a path φ0 , A0 , φ1 , A1 , . . . , An−1 , φn is n.
We say that φ lies on a path from ψ if there is path φ0 , A0 , φ1 , A1 , . . . , An−1 , φn where
n ≥ 1 and φ0 = ψ and φn = φ.
Definition 5.3. Let G = V, A, h, t be a graph. We say that G is acyclic if for all φ ∈ V there is
no path from φ to φ.
We extend this to an account of mediate ground. Rather than working directly

with the grounding graphs we extract certain labeled trees and read ground off
of those trees. This is convenient for two reasons. First, the graphs themselves are
quite messy: many truths have several (immediate) grounds, and there will be a
multitude of ways for a given truth to be mediately grounded. Each tree, on the
other hand, isolates a unique way for a given truth to be grounded. Second, the
trees may be viewed as representing explanatory arguments from the grounds to the
grounded.
For our purposes a tree is a certain type of hypergraph.
Definition 5.4. A hypergraph T = T, A, h, t, is a tree if T satisfies the following conditions:
(i) t(A) is a set (not a multiset) for each A ∈ A.
(ii) For every u ∈ T there is at most one A ∈ A such that u ∈ t(A).
(iii) For every u ∈ T there is at most one A ∈ A such that u ∈ h(A).
(iv) There is a unique c ∈ T such that each u ∈ T lies on a path ending in c.
(v) There is P ⊆ T such each p ∈ P is not in the head of any A ∈ A and such that T is the
closure of P under A.
We refer to P as the premiss nodes of the tree and c as its conclusion node. For short
we will say that P are the premisses of T and c its conclusion. To have an explicit
notation we use T, P, c, A, h, t to indicate that T, A, t, h is a tree with premisses
P and conclusion c.
Suppose we label the nodes of a tree T, P, c, A, h, t with propositions. We may then
take the label on c to represent a grounded truth; the labels on P represent the truths
that ground (the truth that labels) c. The tree as a whole, then, depicts an “explanatory
derivation” of the grounded truth c from its grounds P.
Of particular interest are the labeled trees that arise from grounding graphs.
Definition 5.5. Let G = V, A, h, t be a grounding graph. A labeled (directed) tree over G is a
tuple T = T, P, c, AT , L, tT , hT . Here T, P, c, AT , tT , hT is a tree as above; L : T → V is a
function assigning labels from V to the nodes of T. The sets of arcs A and AT are related as
follows.
• For all A ∈ AT thereis A ∈ A such that

for all φ ∈ V,
– the cardinality of ψ ∈ t(A ) : ψ = φ is the cardinality of {u ∈ tT (A) : L(u) = φ}.
– If u ∈ hT (A) then L(u) ∈ h(A ).
(Remember that while tT (A) is a set of nodes; t(A ) is multiset.)

 bicollective ground: towards a (hyper)graphic account
We refer to the labeled trees over G as G -trees. We now define the notion of strict
full mediate ground.
Definition 5.6. Let G = V, A, h, t be a grounding-graph. Let ∪ {φ} ⊆ V. We say that < φ
if there is G-tree T = T, P, c, AT , L such that L(P) = and L(c) = φ.
Since we think of trees as arguments from the grounds to the grounded it is

convenient to adopt some standard proof-theoretic notion. We write

E
φ
to indicate that E is a tree with premisses exactly and conclusion φ. (“E ” for
“explanatory argument”.)
To understand the definitions it is helpful to consider Figure 7.2. The trees show
perspicuously how both (φ, θ) and (ψ, θ) and (φ, ψ, θ) mediately ground (φ ∨ψ)∧θ .
(We here assume that if both φ, ψ is true, then φ, ψ together ground φ ∨ ψ.)
5.1 Amalgamation failure

The principle of (Strict) Amalgamation says that if i <φ for each i ∈ I then i ∈ I i <
φ. This principle is invalid—according to Definition 5.6. In fact, we can establish
something stronger.20
Say that a grounding graph G = V, A, t, h is amalgamating if whenever we have
Ai ∈ A such that h(Ai ) = h(Aj ) for all i, j ∈ I, then there is A ∈ A such that

h(A) = h(Ai ) for each i ∈ I and such that t(A) = i∈I t(Ai ). If a grounding graph
G is amalgamating then the principle of Amalgamation holds for immediate ground.
However, Amalgamation for mediate ground does not follow from Amalgamation for
immediate ground.
(ϕ∨ψ)∧θ
ϕ∨ψ θ
ϕ ψ
ϕ ψ ϕ ψ
ϕ∨ψ θ ϕ∨ψ θ ϕ∨ψ θ
(ϕ∨ψ)∧θ (ϕ∨ψ)∧θ (ϕ∨ψ)∧θ
Figure 7.2. A G -graph and (some of) its G -trees

20
Correia (2014, n. 17) gives a similar counterexample.
jon erling litland 
To see this let φ be a truth immediately grounding a truth ψ. Contrast the following
two cases. First, the truth ψ immediately grounds the truth ψ ∧ ψ; second, the truth
ψ grounds the truth ψ ∨ θ, where θ is some unrelated truth.
In the first case φ is a mediate ground for ψ ∧ ψ; moreover, φ, ψ taken together
also form a mediate ground for ψ ∧ ψ. In the second case, φ is a mediate ground for
ψ ∨ θ, but φ, ψ taken together do not constitute a mediate ground for ψ ∨ θ.
The reason is that in grounding ψ ∧ ψ, ψ is used twice; we can elect to use φ to
ground only one of the occurrences of ψ, leaving the other occurrence of ψ available
to make a contribution to grounding ψ ∧ ψ. In the grounding of ψ ∨ θ, on the other
hand, ψ is used only once. If we use φ to ground ψ there is no occurrence of ψ “left
over” to make a separate contribution to grounding ψ ∨ θ . The difference between
the two situations is depicted in Figure 7.3.
Amalgamation for immediate ground expresses a choice about how to deal with
overdetermination. In cases of overdetermination not only are the individual can-
didate immediate grounds in fact grounds, they are also immediate grounds taken
together. But if two grounds are “on the same path” to the grounded we should
not count them as together forming a ground, which is what Amalgamation in full
generality forces us to do.21
It is not just Amalgamation that fails for this reason. The following “structural”
principles of ground also fail:
, φ < ψ , φ, φ < ψ
Mingle Contraction
, φ, φ < ψ , φ < ψ
But all of Amalgamation, Contraction, and Mingle are validated by the truthmaker
semantics for the left-collective pure logic of ground. The truthmaker semantics
cannot even serve the instrumental role of characterizing the validities of the pure
logic of ground.
ϕ ϕ
ψ ψ ψ
ψ∧ψ ψ∨θ
Figure 7.3. Amalgamation failures
21 Consider the causal parallel. If Billy and Suzy each throw a rock at the window we might count the
event of them both throwing as a cause of the window’s shattering. In contrast, consider just the rock Billy
threw. If the rock’s momentum at t0 is a cause of its momentum at t1 and its momentum at t1 is a cause of
the window’s shattering we might reasonably count both the rock’s momentum at t0 and its momentum at
t1 as causes for the window’s shattering. But it would be wrong to count the momentum of the rock as t0
together with its momentum at t1 as a joint cause of the window’s shattering.
 bicollective ground: towards a (hyper)graphic account
ϕ0 ψ0
ψ1
ϕ1 ψ2
…
Figure 7.4. A hyperarc
6 The Hypergraphic Account: The Bicollective Case

To study bicollective ground proper we consider grounding graphs V, A, h, t where
both t(A) and h(A), for A ∈ A, may have any cardinality. A graphical representation of
a hyperarc A with t(A) = {φ0 , φ1 , . . .} and h(A) = {ψ0 , ψ1 , . . .} is given in Figure 7.4.
The definition of immediate ground is the obvious one: holds in a graph
G = V, A, h, t if there is A ∈ A such that t(A) = and h(A) = .
Since we allow both h(A) and t(A) to be empty we have to ensure that the notion
of the empty ground behaves properly. Whatever sense can be made of the empty
ground one thing is at least clear: the empty ground is minimal. We therefore impose
the following condition:
(∅-Minimality) If G = V, A, h, t is a graph then for all A, if h(A) = ∅, there is
B such that t(B) = ∅ and h(B) = t(A).22
6.1 Acyclicity
Since we are interested in strict immediate ground we impose an acyclicity condition
on the graphs. Unlike in the left-collective case there is some choice about what the
right notion of acyclicity is.
First, we must define the notion of a path.
Definition 6.1. Let G = V, A, t, h be a graph. A path in G is either a sequence V0 , where
V0 ⊆ V or an alternating sequence of (sets of) vertices and arcs V0 , A0 , V1 , A1 . . . , An−1 , Vn
such that t(A0 ) = V0 and h(A0 ) = V1 , t(A1 )∩V1 = ∅ and h(A1 ) = V2 . . ., t(An−1 )∩Vn−1 = ∅
and h(An−1 ) = Vn . The length of a path V0 , A0 , V1 , A1 . . . , An−1 , Vn is n. If there is a path
V0 , A0 , V1 , A1 . . . , An−1 , Vn of length n ≥ 1 we say that there is a path from V0 to Vn .
To see what a path looks like, it helps to return to (a variation on) the picture of the
wall (in Figure 7.1):
Figure 7.5.
22
This says that if t(A) grounds the null fact then t(A) is immediately zero-grounded.
jon erling litland 
We may consider Figure 7.5 as a graph by letting each small box be a vertex. We
let there be an arc the tail of which is the entirety of the bottom row; and the head of
which is the entirety of the middle row. For each box in the middle row, we let there
be an arc from that box to the box directly above it on the upper row. For instance,
there is an arc with tail {b} and head {a}. In this case there is a path from the box
labeled b to the box labeled d. d does rest on the lower row as a whole, and so in
part on b.
Definition 6.2. Let G = V, A, t, h be a graph.
(i) G is weakly acyclic if for all V1 ⊆ V0 ⊆ V there is no path from V0 to V1 ;
(ii) G is strongly acyclic if G is weakly acyclic and for all V0 and V1 such that V0 ∩ V1 = ∅
there is no path from V0 to V1 .
In (weakly or strongly) acyclic graphs nothing grounds the empty ground.

Observation 6.3. Let G = V, A, h, t be a (weakly, strongly) acyclic graph. Then for no A ∈ A
do we have h(A) = ∅.
Proof. Suppose h(A) = ∅. Then by (∅-Minimality) there is B such that h(B) = t(A) and t(B) = ∅.
We then have a path from ∅ to ∅, contradicting weak (strong) acyclicity.
Weak acyclicity is too weak to capture what we want in a strict notion of ground.
To see this consider the graph H depicted in Figure 7.6. H = V, A, h, t where
V = {a, b, c, d} and A = {A, B, C} where t(A) = a and h(A) = c; t(B) = b and h(B) = d
and t(C) = {c, d} and h(C) = {a, b}. H is weakly acyclic according to Definition 6.2,
but intuitively H is cyclic: it is natural to think that H represents that a, b (mediately)
grounds a, b.
I am inclined to think that strong acyclicity is the notion we are after, but some
might think that it demands too much. For consider the plurality of all truths: if we
require strong acyclicity we rule out that the totality of all truths is grounded. (If
the totality of all truths is grounded it has to be grounded in a subplurality, but this
would contradict strong acyclicity.) This might be an unwelcome result for someone
who wants to defend a version of the principle of sufficient reason for bicollective
ground.
Once we have defined mediate ground we can define a notion of acyclicity inter-
mediate between strong and weak. Before I go on to characterize mediate ground, let
me return to structuralism—and i and −i.
a b
c d
a b
Figure 7.6. The insufficiency of weak acyclicity

 bicollective ground: towards a (hyper)graphic account
6.2 Structuralism again

In §4.2 above we noted that the truthmaker semantics rules out that the truths that
i exists and −i exists are commonly grounded. (Their being commonly grounded
would make the truth that i exists identical to the truth that −i exists.) This type of
argument fails in the graph-theoretical framework. For it is possible for two truths
to have exactly the same immediate grounds, but be grounded in them in different
ways. Formally, this comes down to there being distinct co-tailed arcs. (As a plausible
example of this one might want to distinguish the truths φ ∨ φ and φ ∧ φ.23 ) One
might think that one way to be a structuralist is to hold that while the truths that
i exists and −i exists are commonly grounded, they are nevertheless grounded in
different ways.24
However, one might argue that this is impossible. For i and −i are absolutely
indiscriminable, there is no feature that tells i apart from −i. Suppose now that the
truths that i exists and −i exists are commonly grounded, but that the truths are
grounded in different ways. Then it would be possible to distinguish i and −i after
all; for i is characterized by the fact that the truth that i exists is grounded in way w0
(rather than way w1 ); whereas −i is characterized by the fact that the truth that −i
exists is grounded in way w1 (rather than way w2 ). But then we could discriminate
between i, −i after all. Since they are not discriminable, we conclude that the truths
that i exists, −i exists are not grounded in different ways. If this argument succeeds
the structuralist has no choice but to accept bicollective ground.25
6.3 Mediate ground
We begin by generalizing the notion of a tree.
Definition 6.4. An edifice is a hypergraph E = E, A, h, t such that
(i)t(A) and h(A) are sets (not multisets) for every A ∈ A.
(ii)For every u ∈ E there is at most one A ∈ A such that u ∈ t(A).
(iii)For every u ∈ E there is at most one A ∈ A such that u ∈ h(A).
(iv) There is a set C such that if c ∈ C then c is not in the tail of any A ∈ A and such that
each path through E terminates in some C0 ⊆ C.
(v) There is P ⊆ E such each p ∈ P is not in the head of any A ∈ A and such that E is the
closure of P under A.
Note that any graph with the empty collection of arcs counts as an edifice.
We may think of an edifice as an argument from the premisses P to the conclu-
sions C. Whereas a tree represents an argument with a single conclusion an edifice
23 Admittedly, the distinction here can be drawn in terms of the multiplicity of grounds; but there are
more complicated (infinitary) examples where this move will not work.
24 We here invoke the notion of a way of grounding. Putting this notion on a rigorous footing is, I believe,
one of the most pressing issues in the theory of ground and one to which I hope to return elsewhere.
25 Much more needs to be said, of course, to make this rigorous, but the above should be sufficient for
the present, largely motivational, purposes.

jon erling litland 
represents an argument with several conclusions. (Recall that the conclusions of an

argument are to be read conjunctively.) We write E = E, P, C, A, h, t to make it
explicit that E has premisses P and conclusions C. A labeled edifice is an edifice where
every node has been assigned a label. Just as in the left-collective case we are mainly
interested in the labeled edifices that are generated by grounding graphs.
Definition 6.5. Let G = V, A, h, t be a graph. An edifice over G is a tuple E =
E, P, C, AE , L, hE , tE such that E, P, C, AE , hE , tE is an edifice and L : E → V is a labeling
function such that
• For all A ∈ AE thereis A ∈ A such that for all φ ∈ V,
– the cardinality of ψ ∈ t(A ) : ψ = φ is the cardinality of {u ∈ tE (A) : L(u) = φ}.

– the cardinality of ψ ∈ h(A ) : ψ = φ is the cardinality of {u ∈ hE (A) : L(u) = φ}.
If E is an edifice over G we also say that E is a G -edifice.

It might help to think of the relationship between a graph G and the G -edifices in
the following way. Think of the vertices of a graph G = V, A, h, t as a collection of
various specialized bricks. Think of the arcs in A as specifying how the bricks may be
combined to produce more complicated structures (e.g. walls). One can think of the
edifices as representing the particular ways in which the blocks have been put together
in accordance with the rules.
How can we use edifices to define mediate ground? The natural definition is to
say that < holds in G iff there is a G -edifice E = E, P, C, AE , L, h, t such that
L(P) = and L(C) = . However, even in strongly acyclic graphs this does not
define a notion of strict ground.
To see this consider the grounding graph Z = V, A, h, t where V = {φi : i ∈ Z},
and where for each i ∈ Z there is Ai with t(Ai ) = {φi } and h(Ai ) = {φi+1 }. Z has the
following infinitely ascending and descending chain of immediate ground:
. . . φ−2 φ−1 φ0 φ1 φ2 . . .
Consider the two sets:
even = {φ−4 , φ−2 , φ0 , φ2 , . . .}
and
odd = {φ−3 , φ−1 , φ1 , φ3 , . . .}
These sets give us infinitely ascending and descending chains of mediate ground in
the obvious way.
It is easy to see that there is an edifice with premisses even and conclusions
odd and also one with premisses odd and conclusions even . The following edifice
witnesses the first case.
. . . φ−2 φ0 φ2 . . .
. . . φ−1 φ1 φ3 . . .
 bicollective ground: towards a (hyper)graphic account
But then we get both even < odd and odd < even , contradicting the asymmetry
of strict ground.
To define mediate ground we define the classes of weak, strict, and immediately
strict edifices. The idea is that the strict edifices are generated from the immediately
strict edifices by composing them with weak edifices in a constrained way. This
allows us to capture the idea that strict ground is the closure of immediate strict
ground.
To state the next definitions perspicuously we need some notation. Let us use
D, E , . . . to range over edifices. Consider the edifice {t(A)} ∪ {h(A)} , {A} , h, t.
We may depict this edifice both proof-theoretically and graph-theoretically. (Typo-
graphic considerations dicate which depiction we choose.)
t(A)
t(A)
A
h(A)
h(A)
More generally if D is an edifice with premisses {γ0 , γ1 , . . .} and conclusions
{δ0 , δ1 , . . .} we may write this in either of the following ways.
γ0 γ1 γ2
...
γ0 , γ1 , . . .
D D
δ0 , δ1 , . . .
δ0 δ1 δ2 . . .
Having both notations allows us to conveniently express how to compose edifices.
Suppose

D
0 , 1 , . . . , ϒ
is an edifice with conclusions 0 , 1 , . . . , ϒ. Suppose further that for each i,
i , i
Ei

i
is an edifice with premisses i , i and conclusions
i . These edifices may be com-
posed to yield an edifice with premisses , 0 , 1 , . . . and conclusions
0 ,
1 , . . . , ϒ.
We will depict such an edifice as follows.
jon erling litland 
Θ
D
Δ0 Γ0 Δ1 Γ1 Δ2 Γ2 Υ
...
ε0 ε1 ε2
Σ0 Σ1 Σ2 ...
Definition 6.6. Let G = VG , AG , tG , hG be a grounding graph. The immediately strict, strict,
and weak edifices over G are the least classes of G-edifices satisfying the conditions in Figure 7.7.
To see the idea behind Definition 6.6 look first at the immediately strict edifices. In
an immediately strict edifice the conclusions are all in the head of a single arc A. We
move from the many premisses t(A) to the many conclusions h(A) by a single step of
ground. (Many things are grounded, but in a single step of ground.)
Suppose that E (s) is a strict edifice with conclusions C. Suppose that C = C0 ∪ C1
and we have t(A0 ) = C0 and t(A1 ) = C1 . Then the edifice
C0 C1
A0 A1
h(A0 ) h(A1 )
is not strict. However, the edifice
E (s)
C0 C1
A0 A1
h(A0 ) h(A1 )
is strict.
The following proposition pinpoints an important difference between the strict and
the merely weak edifices.
Proposition 6.7. Let E = E, P, C, A, L be a strict edifice. Then for every p ∈ P and c ∈ C there
is some P ⊆ P and some C ⊆ C such that there is a path from P to C and such that p ∈ P
and c ∈ C .
Proof. The proof proceeds by induction on the complexity of the strict edifice E =
E, P, C, A, L. If E is immediately strict the result is immediate. Suppose that E is the result of
an application of (Strict Right Composition). Then E is of the form:
 bicollective ground: towards a (hyper)graphic account
(Weakness) Every edifice over G is a weak edifice

(Inclusion) Every immediately strict edifice is a strict edifice
t(A)
(Immediacy) A is an immediately strict edifice for each A ∈ A.
h(A)
Γ
ε
(Strict Left Composition) If is a weak edifice and D is a strict edifice then any
Γ
Σ
ε
Γ
edifice of the form is a strict edifice.
D
Σ
(Strict Right Composition) If
Θ
D
Δ0, Δ1, . . . , Υ
is a strict edifice and
Δi, Γi
εi
Σi
is a strict edifice for each i then any edifice of the form
Θ
D
Δ0 Γ0 Δ1 Γ1 Δ2 Γ2 Υ
...
ε0 ε1 ε2
Σ0 Σ1 Σ2 ... is a strict edifice

(Weak Composition) If
Θ
D
Δ0, Δ1,...
is a strict edifice and
Δi
εi
Σi
is a weak edifice for each i then any edifice of the form
Θ
D
Δ 0 Δ1 . . .
ε0 ε 1 . . .
Σ0 Σ1 . . .
is a strict edifice
Figure 7.7. Strict and weak edifices

jon erling litland 
Θ
D
Δ0 Γ0 Δ1 Γ1 Δ2 Γ2 Υ
...
ε0 ε1 ε2
Σ0 Σ1 Σ2 ...
So let p ∈ P and c ∈ C be given. If p ∈ i for some i, the result follows by the induction
hypothesis applied to Ei . If p ∈ we reason as follows. Suppose first that c ∈ ϒ. The edifice D,
that is,
Θ
D
Δ0 Δ1 Δ2 Υ
...
is strict. It then follows by the induction hypothesis that there is some P ⊆ P and C such that
there is a path from P to C and such that p ∈ P and c ∈ C .
Suppose then that c ∈
i , for some i. Let d ∈ i be given. Since Ei is a strict edifice there
is D ⊆ i ∪ i and C ⊆
i such that there is path from D to C where d ∈ D and c ∈ C .
Since D is strict there is also D ⊆ 0 ∪ 1 ∪ · · · ∪ · · · such that there is a path from P to D
where p ∈ P ⊆ P and d ∈ D . But since D ∩ D = ∅ this shows that there is a path from P
to C which is what we have to show.
To prove the cases of (Weak Composition) and (Strict Left Composition) we observe that if
D = D, P, C, AD is a weak edifice then (i) for all c ∈ C, there is some p ∈ P ⊆ P such that for
some C ⊆ C there is a path from P to C ; and (ii) for all p ∈ P, there is some c ∈ C such that
for some P ⊆ P there is a path from P to C . Having made this observation the proof proceeds
similarly to the case of (Strict Right Composition).
We can finally define an intermediate notion of acyclicity.

Definition 6.8. A graph G = V, A, h, t is acyclic if for all 0 ⊆ V there is no strict G-edifice
E = E, P, C, AE , L, h, t such that L(P) = 0 and L(C) = 1 ⊆ 0 .
Proposition 6.9. If G is strongly acyclic, G is acyclic.

Proof. Suppose E = E, P, C, L, A, h, t is a strict G-edifice such that L(P) = 0 and L(C) = 1 ,
where 1 ⊆ 0 . Let γ ∈ 0 . Let p ∈ P and c ∈ C be such that L(p) = L(c) = γ . By Proposition 6.7
there is P ⊆ P and C ⊆ such that there is a path from P to C . This path in E induces,
in G, a path from L(P ) to L(C ). Since γ ∈ L(P ) ∩ L(C ) this contradicts the strong acyclicity
of G.
 bicollective ground: towards a (hyper)graphic account
There are, however, acyclic graphs that are not strongly acyclic. Consider for
instance the graph G defined as follows. V = {a, b}, A = {A}, where t(A) = {a} and
h(A) = {a, b}. G is not strongly acyclic but it is acyclic. It would be of some interest to
determine under what conditions acyclic graphs are strongly acyclic.
We can finally define the various notions of full ground.
Definition 6.10. Let G = V, A, h, t be a graph.

(i) in G if there is A ∈ A such that t(A) = and h(A) = .
(ii) < holds in G if there is a strict G -edifice with premisses and conclusions .
(iii) ≤ holds in G if there is a G -edifice with premisses and conclusions .
We note the following principles about the interaction of weak and strict full
ground.
Proposition 6.11.
(i) ≤ for all . (Identity)
(ii) (a) If then < . (Subsumption (/<))
(b) If < then ≤ . (Subsumption(</≤))
(iii) If < 0 , 1 , . . . and i ≤
i , for each i, then <
0 ,
1 , . . . (Weak Right Cut)
(iv) If < 0 , 1 , . . . and i , i <
i , for each i, then , 0 , 1 , . . . <
0 ,
1 , . . . (Strict
Right Cut)
(v) If
0 ,
1 , . . . < and i ≤
i for each i, then 0 , 1 , . . . < . (Left-Cut)
(vi) For no and do we have , < 0 where 0 ⊆ . (Irreflexivity)
(vii) If G is strongly acyclic: for no , do we have < , γ , where γ ∈ .
(viii) if < holds in G , then for no
do we have ,
≤0 , where 0 ⊆ . (Irreversibility)
Proof. We prove some of the cases. The proofs are by and large just unpacking the definitions.
Identity is immediate. Consider the edifice with vertices (labeled by) and the empty set
of arcs.
To prove Weak Right Cut suppose that < 0 , 1 , . . .. There is then a strict edifice E with
premisses and conclusions 0 , 1 , . . .. For each i = 0, 1, 2, . . . let Di be a weak edifice with
premisses i and conclusions
i . Then by (Weak Composition) the following is a strict edifice
with premisses and conclusions
0 ,
1 , . . ..

D
0 1 . . .
E0 E1 . . .

0
1 . . .
Strict Right Cut is immediate from (Strict Right Composition).

Left Cut follows immediately from (Strict Left Composition)
To prove (Irreversibility) we reason as follows. Suppose, < . Let E be a strict edifice
witnessing this. Suppose, for contradiction, that there is a weak edifice F with premisses ,
and conclusions 0 , where 0 ⊆ . Let 1 = \ 0 . Then the following is a strict edifice with
premisses ,
, 1 and conclusions , contradicting the acyclicity of G.
jon erling litland 
F
0 1
E

6.4 Partial ground

We can also define various notions of partial ground.
Definition 6.12. Let G be a graph.
(i) holds in G if there is a G -edifice with premisses ,
and conclusions , for
some
.
(ii) ≺∗ holds in G if there is a strict G -edifice with premisses ,
and conclusions ,
for some
.
(iii) ≺ holds in if holds in G and does not hold in G .
≺∗ is the notion of partial strict ground. is a partial strict ground for when
is part of a strict full ground for . ≺ is the notion of strict partial ground. is a
partial strict ground for when weakly partially grounds but does not weakly
partially ground . (Strict partial ground is irreversible weak ground.)
In the truthmaker semantics strict partial and partial strict ground notoriously
come apart. Interestingly, they come apart in the graphical approach as well. To
see this, consider again the graph Z from above. Recall that we have the following
ascending and descending sequence of immediate ground:
. . . φi−2 φi−1 φi φi+1 . . .
Let = {φi : i ∈ Z}. Then we have φi ≺ , for each i, but we do not have φi ≺∗ , for
any i.
Z also provides a counterexample to the principle Reverse Subsumption. This
principle says that if γ0 , γ1 , . . . ≤ and γi ≺ , for each i, then γ0 , γ1 , . . . < .
The following principles about partial ground are easily established.
Proposition 6.13.
• If , ≤ then (Subsumption(≤/))
• If , < then ≺ . (Subsumption(</≺))
• If and then . (Transitivity(/))
• If ≺ and then ≺ (Transitivity(≺/))
• If and ≺ then ≺ (Transitivity(/≺))
6.5 Equivalent collections

It is tempting to use the notion of weak full ground to express identity in the object
language. This, however, will not work in the bicollective case. Let us use ≈ to
mean that ≤ and ≤ . There are collections , such that ≈ but and
differ in what they immediately ground.
 bicollective ground: towards a (hyper)graphic account
Consider, for example, the collection even and odd from above. In §6.3 above we

noted, in effect, that we have Even ≈ Odd . Consider the conjunctions Even and

Odd . On the plausible assumption that the immediate grounds for a conjunction

are all and only the conjuncts, Even and Odd differ in what they immediately
ground. Since any notion of identity of collections of propositions has to satisfy (the
analogue of) Leibniz’s law this shows that mutual weak full ground is not the right
notion of identity between collections of propositions.
Fortunately, the correct notion of identity between collections of propositions is
not hard to come by. Say that ≈I if for all
, , if ,
then
,
(and vice versa). Equivalent collections of propositions agree on what they (help to)
immediately ground.
7 Conclusion
In this paper I have shown how we can develop a graph-theoretic account of bicol-
lective ground and indicated how it might be useful in formulating mathematical
structuralism. As should be apparent, bicollective ground is much more complicated
than left-collective ground. The main contribution of this paper has been finding
the right definitions and establishing the (fairly rudimentary) results showing that
the definitions work. While lots of work remains to be done, both on the technical
and philosophical side, I hope to have done enough to convince the reader both
that bicollective ground might be useful and also that it can be developed rigor-
ously.
In closing, let me briefly mention two outstanding technical issues. First, can one
devise a calculus for a pure logic of bicollective ground and establish soundness
and completeness with respect to the hypergraph models constructed here? Sec-
ond, assuming one has done this, what is the relationship between the graphical
semantics for bicollective ground and the truthmaker semantics for bicollective
ground?26
References
Audi, Paul (2012). “Grounding: Towards a Theory of the in Virtue of Relation”. In: Journal of
Philosophy 109.12, pp. 685–711.
Burgess, John P. (1999). “Book Review: Stewart Shapiro. Philosophy of Mathematics: Structure
and Ontology”. In: Notre Dame Journal of Formal Logic 40.2.
Correia, Fabrice (2014). “Logical Grounds”. In: Review of Symbolic Logic 7.1, pp. 31–59.
Correia, Fabrice and Benjamin Schnieder, eds. (2012). Metaphysical Grounding. Cambridge
26 Relatedly, what is the relationship between the graphical semantics and the truthmaker semantics for
left-collective ground?.
jon erling litland 
Dasgupta, Shamik (2014a). “Metaphysical Rationalism”. In: Noûs 50.2, pp. 379–418.
Dasgupta, Shamik (2014b). “On the Plurality of Grounds”. In: Philosophers’ Imprint 14.20,
pp. 1–28.
Dasgupta, Shamik (2014c). “The Possibility of Physicalism”. In: Journal of Philosophy 111.9/10,
pp. 557–92.
deRosset, Louis (2013). “What is Weak Ground?” In: Essays in Philosophy 14.1. Ed. by Paul
Hovda and Troy Cross, pp. 7–18.
deRosset, Louis (2015). “Better Semantics for the Pure Logic of Ground”. In: Analytic Philosophy
56.3, pp. 229–52.
Fine, Kit (1994). “Senses of Essence”. In: Modality, Morality , and Belief: Essays in Honor of
Ruth Barcan Marcus. Ed. by N. Ascher, D. Raffman, and W. Sinnott-Armstrong. Chicago:
University of Chicago Press, pp. 53–73.
Fine, Kit (2012a). “Counterfactuals without Possible Worlds”. In: The Journal of Philosophy
109.3, pp. 221–46.
Fine, Kit (2012b). “Guide to Ground”. In: Metaphysical Grounding. Ed. by Fabrice Correia and
Benjamin Schnieder. Cambridge University Press. Ch. 1, pp. 37–80.
Fine, Kit (2012c). “The Pure Logic of Ground”. In: The Review of Symbolic Logic 5.1,
pp. 1–25.
Fine, Kit (2014). “Truth-maker Semantics for Intuitionistic Logic” In: Journal of Philosophical
Logic 43.2–3, pp. 549–77.
Fine, Kit (2017). “A Theory of Truthmaker Content I: Conjunction, Disjunction and Negation”.
In: Journal of Philosophical Logic 46.6, pp. 625–74.
Keränen, Jukka (2001). “The Identity Problem for Realist Structuralism”. In: Philosophia Math-
ematica 9, pp. 308–30.
Linnebo, Øystein and Richard Pettigrew (2014). “Two Types of Abstraction for Structuralism”.
In: Philosophical Quarterly 64.255, pp. 267–83.
Litland, Jon Erling (2015). “Grounding, Explanation, and the Limit of Internality”. In: Philo-
sophical Review 124.4, pp. 481–532.
Litland, Jon Erling (2016). “Pure Logic of Many–Many Ground”. In: Journal of Philosophical
Logic 45.5, pp. 531–77.
Litland, Jon Erling (2017). “Grounding Ground”. In: Oxford Studies in Metaphysics 10,
pp. 279–316.
Rosen, G. (2010). “Metaphysical Dependence: Grounding and Reduction”. In: Modality: Meta-
physics, Logic, and Epistemology. Oxford University Press, pp. 109–35.
Schaffer, Jonathan (2009). “On What Grounds What”. In: Metametaphysics: New Essays on the
Foundations of Ontology. Ed. by D. J. Chalmers, D. Manley, and R. Wasserman. Oxford
Schaffer, Jonathan (2012). “Grounding, Transitivity, and Contrastivity”. In: Metaphysical
Grounding. Ed. by Fabrice Correia and Benjamin Schnieder. Cambridge University Press.
Ch. 4, pp. 122–38.
Schaffer, Jonathan (2016). “Grounding in the Image of Causation”. In: Philosophical Studies
173.1, pp. 49–100.
PA R T II
The Fundamentality Thesis
8
Indefinitely Descending Ground
Einar Duenger Bohn
We often say that some facts obtain in virtue of others, for example that semantic facts
obtain in virtue of facts about language-use, or that normative facts obtain in virtue of
descriptive facts, or that mental facts obtain in virtue of physical facts. The question
I’m interested in is: must such in-virtue-of chains eventually end in some facts that
don’t obtain in virtue of any other facts? Or can they go on indefinitely without end?1
In other words (to be clarified below), must the in-virtue-of relation be well-founded?
In what follows, I argue that it must not, and point to some reasons for it even
actually not being so. More specifically, in Section 1, I introduce what is perhaps the
closest we get to a standard notion of the in-virtue-of relation, namely a relation of
grounding; in Section 2, I argue that there is no good reason to think that this relation
of grounding must be well-founded; and in Section 3, I argue more directly that it’s
not necessarily well-founded, and further that there are reasons to think it’s actually
non-well-founded.
1 The Standard Notion of Grounding

Everything about in-virtue-of talk can and has been questioned, but the following is
perhaps the closest we get to a standard underlying notion these days (Rosen, 2010;
Bliss & Trogdon, 2014; Raven, 2015). Saying that a fact obtains in virtue of some others
is to say that the fact is grounded in those other facts. Such grounding is taken to be a
one–many relation between one fact and a plurality of (one or more) facts, imposing
what we might (with some slack) call a strict partial order on its domain:2 no fact even
1 Note that there can be infinite chains that are limited, but I wish to talk about infinite chains that are
unlimited; I here and throughout use the term ‘indefinite’ for that purpose. This should not be confused with
the way ‘indefinite’ is sometimes used in the philosophy of mathematics, where there is a constructional or
potential aspect to it, nor should it be confused with the way ‘indefinite’ is sometimes used in debates over
vagueness, where there is an aspect of, well, vagueness to it.
2 I thus adopt the so-called predicate approach, not the operator approach. I also assume grounding to
be factive.
 indefinitely descending ground
partially grounds itself (irreflexivity); if a fact p is grounded in some facts qq, then no
one of qq is even partially grounded in p (asymmetric); and if a fact p is grounded
in some facts q,rr, and q is grounded in some facts ss, then p is grounded in rr,ss
(transitivity). Also, if a fact p is grounded in some facts qq, then, necessarily, if qq
obtain, then p obtain (necessitation); if a fact p is grounded in some facts qq, then it is
not the case that for any r, p is grounded in qq,r (non-monotonicity); and if a fact p is
grounded in some facts qq, then qq metaphysically explain p (explanatoriness).3 The
latter kind of metaphysical explanation amounts to constitutively explaining what a
fact consists in. The underlying notion of grounding is a notion of full as opposed to
partial grounding, where a fact p is partially grounded in some facts qq iff there are
some facts rr such that p is fully grounded in rr and qq are among rr; p being fully
grounded in rr being our official primitive, but intuitively characterized as providing
a complete metaphysical explanation of p.
I henceforth call this the standard notion of grounding (SNG). I will not further
discuss or defend SNG or any of its abovementioned features.4 Instead, I will here
simply assume SNG in order to argue against it being necessarily well-founded, in
favor of some reasons for it actually being non-well-founded.
What is the notion of well-foundedness in play? There has been some ambiguity in
the literature with respect to what it is more exactly, but my arguments below will be
directed at the following notion identified in Schaffer (2009; 2010), and made more
precise in Dixon (2016) and Rabin & Rabern (2016):5 every non-fundamental fact p
is fully grounded in some fundamental facts qq, where a fact q is fundamental iff there
are no facts rr such that q is (partially or fully) grounded in rr.6 I henceforth call this
notion of well-founded grounding WF, regimented as follows:
(WF) : ∀p(∼ Fp → ∃qq(Fqq ∧ qqGp))
where ‘qq’ range over pluralities of (one or more) facts, ‘p’ over single facts, ‘F’
expresses being fundamental, and ‘G’ expresses grounding.7 The question then is
simply this: must SNG obey WF?
3 One might here distinguish between, on the one hand, grounding being explanation, and on the other
hand, grounding backing explanation. I assume the former, unless noted otherwise.
4 For references to further discussions on each of these features, see Bliss & Trogdon (2014) and Raven
(2015). For a general criticism of an overall notion of grounding, see Wilson (2014) and Koslicki (2015).
See also Dasgupta (2014a), where grounding is argued to be many–many, rather than one–many.
5 Note that the following notion of well-foundedness is not the mathematical (set-theoretical) notion of
well-foundedness. The above well-foundedness of grounding is a distinctively metaphysical notion. Note

also that this is just one among several notions of metaphysical well-foundedness. Though I think much of
what I go on to say does not hinge on this particular notion, my conclusions should be hedged accordingly.
6 See Dasgupta (2014b) for some problems with the above definition of fundamentality. Fortunately,
those problems will not affect my overall argument in this paper.

7 I take ‘F’ to be distributive, but see fn.13, Section 2 below.
einar duenger bohn 
2 Why Believe Grounding Must Be Well-Founded?

What are the arguments for WF? There aren’t many. It seems to be more of an assumed
metaphysical axiom (or metaphysical law) supported by intuition. The underlying
intuition is perhaps most forcefully identified and endorsed in Schaffer (2009:376;
2010:37), where he claims that ‘[t]here must be a ground of being. If one thing exists
only in virtue of another, then there must be something from which the reality of the
derivative entities ultimately derives.’ Schaffer (2010:62) further claims that if there
is no ultimate ground, then ‘[b]eing would be infinitely deferred, never achieved.’
Finally, Schaffer (2016: section 4.5) claims that ‘[g]rounding must be well-founded
because a grounded entity inherits its reality from its ground, and where there is
inheritance there must be a source.’ And (2016: section 4.5): ‘the grounded exists in
virtue of its ground. This is why a source of reality is needed, in order for there to
be anything to transfer.’8 Cameron (2008:6) too finds Schaffer’s intuition appealing
(though he ultimately rejects that it must hold in favor of it only most likely actually
holding for methodological reasons): ‘if we never reach a bottom level, then it is hard
to see why there are any complex objects at all.’9
The intuition thus seems to be something like this: reality, or being is transferred
from the ground to the grounded, so all facts gain their being from their ground, so if
there is no bottom ground, there is nowhere from which the transfer of being initially
comes, nowhere from which to gain being to begin with, so SNG must obey WF on
pain of there not being any (being to) facts at all.10
Here is one way to make the intuition into an argument for WF: by definition,
something is a fundamental fact if and only if it has no ground; hence, if there are
no fundamental facts, all facts have ground; and if all facts have ground, there are no
facts; but obviously there are some facts; hence there are some fundamental facts. Let’s
call this the argument for WF.
The argument is valid; the main question is why we should believe the next to last
premise: if all facts have ground, there are no facts? I for one simply don’t feel the
intuitive pull here.11 An intuition pump might come by the more or less dynamical
metaphors often used to explain the grounding relation: the ground transmits its being
to the grounded, the grounded gains, achieves, or derives its being from the ground.
8 Thomas Aquinas’ (1266–8/1993:200–2), in his five ways towards the necessary existence of God,
shares Schaffer’s intuition concerning the need for a source of being. In general, there are great similar-
ities between Aquinas’ foundationalism in his five ways and Schaffer’s foundationalism with respect to
grounding.
9 Note: Schaffer doesn’t and Cameron might not endorse the full package of SNG, but officially I’m here
only interested in whether SNG in particular must be well-founded (though I do believe my arguments
generalize to other notions of grounding as well).
10 For more on this picture of transference, inheritance, or source of being, see Trogdon (ms). I treat the
notions of reality and being as interchangeable, but try to keep the notion of existence separate; though in
fact nothing hinges on this for present purposes.
11 And, in any case, why believe my intuitions match deep metaphysical truths?
 indefinitely descending ground
So, if there is no bottom ground, there is nothing that can thus transmit being to
the grounded, nothing from which the grounded can thus gain, achieve, or derive
its being. The dynamical talk makes this sound somewhat plausible. It’s as if you are
to fill up a swimming pool by the use of a hose: if nothing comes out of that hose,
then of course the swimming pool will not be filled up. Or, at least in my experience,
if I don’t at some definite point start writing my paper, it just won’t be written. But
the problem is that grounding is nothing like such dynamical processes. There is
no such definite dynamical starting point of the grounding chain. Grounding is like
a synchronic, static mathematical relation (like in arithmetic), not like a diachronic,
dynamic physical relation (like in thermodynamics, or action theory). Grounding is
an explanation of what the obtaining of a fact consists in, atemporally; grounding is
not an explanation of the causal history of that fact (cf. Fine, 2001; 2012; Rosen, 2010).
So, any intuitive pull we might feel from the dynamical metaphors is of no help in a
defense of SNG necessarily obeying WF.
Now, before I look at other ways to try to defend the above premise in the argument
for WF, consider the argument that simply drops it: by definition, something is a
fundamental fact iff it has no ground; hence, if there are no fundamental facts, all
facts have ground. Now that is a solid conditional argument for what I henceforth call
indefinitely descending ground (IDG):
(IDG) :∀p∃qq(qqGp)
Given our assumption that G is a strict partial order as per SNG, IDG gives us infinite,
non-ending chains of grounding.12 Given some very minor, plausible assumptions,
IDG is incompatible with WF. Proof : assume WF and ∼Fp. Then, ∃qq(Fqq∧qqGp).
Let q1 be one of those qq and assume the distributivity of F. It then follows that
Fq1 ; but by definition of fundamentality, ∀p(∼Fp ↔ ∃ss(ssGp)); hence ∼ ∃ss(ssGq1 );
but by IDG, ∃qq(qqGq1 ); hence, contradiction. Now, the minor assumptions that
fundamentality is a distributive property and the definition of fundamentality is here
taken for granted (though neither one is beyond dispute13 ), so WF and IDG are
incompatible.
Now, one way to indirectly defend the controversial premise in the above argument
for WF (i.e. the premise that if all facts have ground, there are no facts) is to argue
that IDG amounts to an appropriately vicious infinite regress. I take it a benign infinite
regress won’t do (e.g. it’s not a problem that we have: ‘s’ is true; “s’ is true’ is true; ‘ “s’ is
true’ is true’ is true; and so on ad infinitum). I also take it a mere infinite regress won’t
do (e.g. it’s not a problem that zero has infinitely many successors). Rather, IDG must
12 Cf. Bliss (2013); Tahko (2014); Morganti (forthcoming). With G being transitive but failing irreflex-
ivity, we could have loops of ground; see Bliss (2014). I’m not in principle opposed to such loops, but SNG
is for present purposes assumed to be irreflexive, so, by assumption, no loops.
13 If there are fundamental facts, but fundamentality is a collective property, things will look slightly
different. Though that is an interesting idea, I will ignore it in what follows since I believe that there most
likely are no fundamental facts. See Dasgupta (2014b) for criticism of the notion of fundamentality.
einar duenger bohn 
amount to an infinite vicious regress, and it must be of the appropriate kind to support
the argument for WF. But what could such an appropriately infinite vicious regress be
in this case?
I can think of at least two candidates (adapted from Nolan, 2001; Bliss, 2013). The
first candidate deals with reductive explanations. Now, ‘reduction’ is said in many ways,
but assume SNG is reductive in the sense that if qqGp, then qq explain away p as being
in some sense non-real. Given IDG, one would then explain away facts, but never
explain them away in terms of something that is not thus further explained away.
One might then get the feeling that every fact p is somehow explained away, but into
nothing, so to speak; that every fact is ‘infinitely deferred, never achieved’ (Schaffer,
2010:6314 ).
But, first, why think grounding is thus explanatorily reductive? It is not a common
view of grounding (Fine, 2001; 2012; Schaffer, 2009; 2010; Rosen, 2010). In fact, much
initial motivation for appealing to grounding is to be able to truly say that some things
obtain without thereby admitting that they are as real as all other things obtaining
(Schaffer, 2009). For example, we want to say—truly—that there are tables and chairs,
but not thereby admit that tables and chairs are as fundamental as particles. So, as
far as I can tell, a proponent of IDG could and should simply reject that grounding is
explanatorily reductive. Grounding is explanatory, but non-reductive.15 Grounding
metaphysically explains a given fact in terms of other facts, but grounding does not
thus explain away that fact as somehow non-real.16
Second, assume grounding is thus explanatorily reductive. Then the above argu-
ment, corresponding intuition, and supposed vicious regress get things backwards.
By SNG, the grounded has its being (or nature, or existence) in virtue of the ground.
But if so, the ground surely cannot have less being than the grounded. Either the
ground has the same degree of being as the grounded, or the ground has more being
than the grounded. But then, as we approach infinity towards ground, we either
stay with the same degree of being, or we approach infinite being! In neither case
is being infinitely deferred, never achieved, as per Schaffer’s intuition. If anything,
being is always deferred, but infinitely achieved! So, if grounding is reductive, we’re
not explaining facts away only to approach nothingness, but rather we’re explaining
facts away in terms of other facts that have equal or more being.17
14 Though note that to the extent one is concerned with grounding objects rather than facts, the notion of
grounding seems to fail the assumption of explanatoriness, in which case one loses this particular defense
of his intuition. One must then, as e.g. Schaffer (2016) does, appeal to a notion of grounding that backs
explanation, without grounding itself being explanation.
15 One might of course still accept Rosen’s (2010:123) principle that if p reduces to qq, then qqGp,
without accepting that grounding is reductive in the above sense.

16 Note that this latter notion of being real is not necessarily Fine’s (2001) notion of being real. Note
also that to the extent that Fine’s notion of being real is taken as a guide (though no guarantee) to being
fundamental, Fine’s (2001:26) definition of reduction is incompatible with IDG.
17 See also Fine (2001:27 and fn.38). Morganti (forthcoming) argues that rather than disappearing as per
the above intuition, being emerges from infinite grounding chains. Cameron (2008:10) too raises a similar
 indefinitely descending ground
Now, one might object that infinite being makes no sense, so at best we have
an equal degree of being along the grounding chain. But then, as we go down the
grounding chain, being is infinitely deferred, never achieved because without WF that
equal degree of being is zero! But, of course, that would be begging the question.
It is very important not to confuse various perspectives of the supposed explana-
toriness involved in SNG (cf. Bliss, 2013; Morganti, forthcoming). Assume IDG, its
corresponding chains of grounding, and the corresponding totality of all such chains
of grounding, which I’ll call the chains of being.18 We then have three questions, what
we might call a local, a regional, and a global question. Local Question (LQ): for any
fact p, what grounds p? Regional Question (RQ): what grounds the chain of grounding
of p?19 Global Question (GQ): what grounds the chains of being? A proponent of IDG
should commit to answering LQ, and might or might not commit to answering RQ (cf.
Dasgupta, 2014b). But should she commit to answering GQ? No! Given SNG+IDG,
GQ is incoherent. First of all, the chains of being is presumably a plurality of facts, not
a single fact, which is what it needs to be to be grounded as per SNG. Second, assume
the chains of being amount to an all-encompassing global fact; call it g. Then, by IDG,
there are some facts that ground g; but, by SNG being a strict partial order, those facts
must be distinct from g; so g is not an all-encompassing global fact, contradicting the
initial assumption. So, we might put it like this: each fact individually has a ground,
but all facts taken together (speaking unrestrictedly!) cannot have a ground.
One might object that by invoking many–many grounding (as per Dasgupta,
2014a), one has the logical resources to answer GQ. But note first that one is then
changing the assumptions of the argument, by switching from SNG to another slightly
different notion of grounding, call it SNG*. Second, that might be the right thing to
do, perhaps for independent reasons too, but, given IDG, it does not help answering
GQ. Assume the chains of being amount to an all-encompassing plural global fact; call
it gg. Then, by IDG, there are some facts that ground gg; but, by SNG* being a strict
partial order, those facts must be distinct from gg; so gg is not an all-encompassing
plural global fact, contradicting the initial assumption. So, this time, we might put it
like this: each fact individually has a ground, but all facts taken collectively (speaking
unrestrictedly!) cannot have a ground.
Neither does it help to invoke Fine’s (2012) distinction between being ungrounded
and being zero-grounded. We might put the distinction like this: something is
point: ‘Why could not everything get a bit more real as we progress down the chain, without anything being
wholly real?’ But his intuition ‘rules this out,’ though, as he himself points out, ‘this just is the intuition
that there must be a fundamental level.’ Note that an appeal to an analogy with causation doesn’t seem to
help here: there seems to be no more to an effect than what’s in its cause(s). In fact, since WF postulates
unexplained facts at the bottom, it seems to me it’s ultimately WF that fails to explain the being of our facts,
not IDG. See Section 3 below. See also Tahko (2014).
18 Recall, grounding is supposed to impose a partial order, not a total order, so the chains of being need
not be a single connected grounding chain, but can be many disconnected ones.
19 Note that RQ is not the question of what grounds a grounding fact. The question of what grounds a
grounding fact falls under LQ.

einar duenger bohn 
zero-grounded iff it is grounded, but in the empty collection of facts, and something
is ungrounded iff it is not grounded at all. Now, SNG is assumed to be explanatory,
so while being ungrounded amounts to having no explanation, being zero-grounded
amounts to having an explanation, but in terms of the empty collection of facts. But,
I say, a metaphysical explanation of a fact in terms of the empty collection of facts
is no metaphysical explanation of that fact, at least not as per SNG. Saying that the
obtaining of a fact p consists in the obtaining of the empty collection of facts, is not
metaphysically explaining p unless p is the empty fact. So, by trying to answer GQ
by invoking zero-ground, one either fails due to not being appropriately explanatory,
or one must be invoking a different notion of grounding from SNG.20 So, assuming
SNG, invoking Fine’s (2012) distinction between being ungrounded and being zero-
grounded just does not help, at least not on pain of changing the assumptions. Of
course, maybe that is the right thing to do in the end, but not for present purposes.21
Admittedly, the above argument is quick, but, whether or not it’s too quick, there
is another more decisive argument for not invoking zero-grounding to assist us with
respect to GQ: by our present assumptions, it entails necessitism, the view that all
actual facts are necessary.22 Assume the chains of being are zero-grounded, that is,
grounded in the empty collection of facts. Then, since the empty collection of facts
necessarily exists, by the assumption of necessitation, all actual facts exist necessarily;
hence necessitism. But, I claim, necessitism is false, so the chains of being are not
zero-grounded.
So, I conclude, whether grounding is reductive or not, by IDG, each fact has a
ground, and hence a metaphysical explanation in terms of some other facts, and
no fact is thereby lost into nothingness; that is, no fact is ‘infinitely deferred, never
achieved.’ Given IDG (and SNG, or SNG*), there is no sensible question about what
the ground of being as such is; the global question (GQ) of what grounds the chains
of being themselves is incoherent.
The point that LQ (and maybe RQ) is what we want answered, with GQ being
incoherent, resembles a point made by Hume (1779:IX) in his objection to cosmo-
logical arguments for the existence of God. According to Hume, if one has explained
each step in a perhaps infinite causal chain, there is nothing more to be explained,
and in particular no need to postulate a first cause to explain it all. Now, Pruss
(2012:81–2) objects to Hume by the following example. Consider the flight of a
cannonball between 12:00 and 12:01, and let pt be the state of the cannonball at time t.
20 Presumably, one could do better with respect to zero-grounding by switching to an operator approach
towards G, rather than our predicate approach. But then, again, one is moving away from SNG, which is
our present concern.
21 See Litland (2017) for some further work on zero-grounding. Unfortunately, Litland’s assumptions are
different from ours (e.g. G is a non-factive operator), so his account does not obviously help in answering
GQ. Despite his insistence to the contrary, I also fail to see how Litland’s account of zero-grounding is
metaphysically explanatory, but let that be as it may.
22 I owe this argument to Jon Litland, though of course he is not to blame. Note that the necessitism in
play is not to be confused with our earlier assumption of necessitation.

 indefinitely descending ground
Let p be the conjunction of all pt such that 12:00<t<12:01. Then each pt might explain
its successor pt+ , such that each pt is explained, but still, Pruss claims, the flight of the
cannonball itself is not explained, so Hume is wrong.
Is there a similar objection in the vicinity with respect to IDG? No. The flight case is
misleading because there is an external perspective on the flight of the cannonball, but
there isn’t any such external perspective on all the chains of being (witnessed by GQ
being incoherent). So, adapted to our global case of grounding, I think Hume is right.
The second candidate for a vicious regress in indirect support of the controversial
premise in the argument for WF is similar to the prior one in terms of reductive
explanations, but claims that there is another kind of failure. Whether grounding is
reductive or not, the supposed failure can be seen by analogy with the toy example of
the homunculus theory of perception (Cf. Nolan, 2001; Bliss, 2013). Suppose someone
attempts to explain the perception of the fact that p as follows. There are outside signals
coming into the eye, received by a homunculus sitting on the inside of the eye, who
interprets the signals as being of the fact that p before sending them off to the brain.
How does the homunculus so interpret the signals? Well, there is another homunculus
sitting inside the first homunculus, who interprets the signals as being of the fact that
p before sending them off to the first homunculus. How does the second homunculus
so interpret the signals? Well, there is a third homunculus sitting inside the second
homunculus . . . and so on ad infinitum. Obviously, here the supposed explanation of
the perception that p fails; we might plausibly blame the explanation for creating a
vicious regress.
Now, the worry this time is not so much that we explain away facts into nothingness,
but that the explanation fails in the sense of having to re-ask the same question with
respect to the same kind of object over and over again ad infinitum. The explanation
thus gets us nowhere, so to speak. But there is no such problem in our case of
grounding. IDG does not create any such vicious regress. If qqGp, then, by SNG, qq
provide a full metaphysical explanation of p, so there is no further question about
what grounds p. So, we do get somewhere. We do of course have the further question
of what grounds the facts among qq (not to mention what grounds the fact that
qqGp), and so on ad infinitum, but that is not analogous with the homunculus theory
of perception. In the homunculus-case, one has the same content over and over
again, but not so in the IDG-case of grounding. To make the homunculus-case more
analogous to the IDG-case, we would have to explain the perception that p in terms
of the first homunculus, but then explain this homunculus in terms of something
non-homunculus-like, or at least something different-homunculus-like and so on ad
infinitum. But then the reason for the initial failure in explanation has gone away. The
analogy simply doesn’t hold up to scrutiny.
In general, we might say that an infinite regress is vicious if something we want to
explain cannot be explained because of the regress (Nolan, 2001; Bliss, 2013), so, in
the particular case of grounding, IDG’s infinite regress is vicious if something we want
to ground cannot be grounded because of it. But, as we have just seen, given IDG, it
einar duenger bohn 
makes no sense to ask for the ground of the chains of being themselves, that is, there
is no ‘global’ fact to be grounded, and by IDG each fact is grounded, so it is not the
case that some fact cannot be grounded because of IDG’s regress.23
I conclude that IDG does not create a vicious regress of the appropriate kind, and
hence a defense of the controversial premise (that if all facts have ground, there are
no facts) based on IDG creating a vicious regress of the appropriate kind fails.
3 Against Grounding Being Well-Founded

Here is what I take to be the best argument for why SNG need not obey WF, that is,
why it is possible that SNG is true, but WF is false.
First, assume proper parthood is a strict partial order, and that the concrete world
U is gunky: every entity in U has a proper part.24 Then U is indefinitely divisible into
proper parts, that is, unlimitedly so (so U contains neither extended nor unextended
mereological atoms). Assume further that for any x, if x is part of (or in) U, then the
fact that x exists is (at least partly) grounded in the fact that its proper parts exist.
Then we have in effect a needed case of IDG, contradicting WF (cf. Rosen, 2010:116).
So, if SNG must obey WF, then either proper parthood is not a strict partial order or
such a gunky scenario is metaphysically impossible; but proper parthood is a strict
partial order and such a gunky scenario is metaphysically possible, so, SNG need
not obey WF.
Second, assume proper parthood is a strict partial order, and that the concrete
world U is junky: every entity in U is a proper part.25 Then U is indefinitely extendable
along the proper parthood chains, that is, unlimitedly so (so U contains no maximal
fusion that is not a proper part). Assume further that for any x and y in U, if x is a
proper part of y, then the fact that x exists is (at least partly) grounded in the fact
that y exists.26 Then we have in effect another case of IDG, contradicting WF. So, if
SNG must obey WF, then either proper parthood is not a strict partial order or such
a junky scenario is metaphysically impossible; but proper parthood is a strict partial
order and such a junky scenario is metaphysically possible (Bohn, 2009b), so SNG
need not obey WF.
Finally, assume proper parthood is a strict partial order, and that the concrete world
U is hunky: every entity in U both is and has a proper part.27 Then U is indefi-
nitely extendable and indefinitely divisible along the proper parthood chains (so U
contains neither a minimal nor a maximal entity). Then, whichever mereological
23 See also Schaffer (2016: section 4.5). My arguments above amount to a rejection of Schaffer’s ‘transfer
model’ in the sense that I argue there is no need for a source of being.
24 On gunk, see e.g. Sider (1993), Schaffer (2003), and Arntzenius (2008).
25 On junk, see Schaffer (2010) and Bohn (2009a; 2009b; 2012; ms).
26 See Schaffer (2010), where this direction of grounding is defended. If junk is possible, Schaffer’s
version of WF in terms of priority monism fails to be necessarily true (Bohn, 2012).

27 On hunk, see Bohn (2009b; 2012).
 indefinitely descending ground
direction a chain of grounding goes (though of course it cannot go in both directions

at once!), we have in effect another case of IDG, contradicting WF. So, if SNG must
obey WF, then either proper parthood is not a strict partial order or such a hunky
scenario is metaphysically impossible; but proper parthood is a strict partial order
and such a hunky scenario is metaphysically possible (Bohn, 2009b; 2012), so SNG
need not obey WF.
These three arguments seem very good to me; in fact, they seem as good as
arguments get in philosophy. Rhetorically: why exactly is a gunky, a junky, or a
hunky scenario metaphysically impossible? I know of no convincing reason to believe
they are.28, 29
Assuming SNG appropriately tracks the mereological hierarchy as above (which is
no minor assumption, but we only need its mere possibility here), I conclude that the
claim that SNG must obey WF is simply false.
In fact, there are even some (highly defeasible!) reasons to believe that the concrete
world is actually hunky. We are faced with a general cosmic pattern that so far has no
clear end points in sight. Starting high up, the universe is partly composed of clusters
of galaxies; the clusters of galaxies are partly composed of galaxies; the galaxies are
partly composed of solar systems; the solar systems are partly composed of planets
and stars; the planets and stars are partly composed of various chemicals; the various
chemicals are partly composed of molecules; the various molecules are partly com-
posed of atoms; the atoms are partly composed of electrons, protons, and neutrons;
the protons and neutrons are partly composed of various quarks; the various quarks
are partly composed of . . . to be continued? Starting low down, the various quarks
partly compose the protons and neutrons; the electrons, protons, and neutrons partly
compose atoms; the atoms partly compose molecules; the molecules partly compose
chemicals; the chemicals partly compose planets and stars; the planets and stars partly
compose solar systems; the solar systems partly compose galaxies; the galaxies partly
compose clusters of galaxies; the clusters of galaxies partly compose super-clusters of
galaxies; the super-clusters of galaxies partly compose . . . to be continued? Or perhaps
they partly compose the universe, which in turn partly composes a multiverse;30
which in turn partly composes . . . to be continued?
As science has progressed, we have again and again discovered that U is both
bigger (cf. the development of cosmology) and smaller (cf. the development of particle
physics) than we thought before. Considering that overall cosmic pattern, we are faced
with some inductive/abductive reasons to think there is no end in either direction;
28 Arguably, Sider’s (2011; 2013) reasons rest on a too deflationary notion of modality.
29 In Bohn (2009a; 2009b; 2012) I argued that if composition is not identity and gunk is possible, then
junk is possible. I now believe (i) that composition is identity and entails unrestricted composition (Bohn,
2009c; 2011; 2014); (ii) that hunk is possible, likely actual, if not necessary; (iii) that plural comprehension
is false; and (iv) that because of (iii), (i) and (ii) are compatible (Bohn, ms).
30 Cf. Carr (2009).
einar duenger bohn 
dismissing these reasons out of hand, and especially on a priori grounds, seems
scientifically and theoretically irresponsible.31
By extrapolation on the above cosmic pattern, we are simply forced to question
whether U is in fact open-ended in both directions, and, arguably, we have some
inductive/abductive reason to think it is, no convincing reason to think it’s not. First,
arguably, there is neither a convincing a posteriori reason nor a convincing a priori
reason for believing that the cosmic pattern does eventually have a stopping point; at
least none that cannot be equally well explained as being our (hopeful) theoretical
idealization rather than a real existent.32 Second, concerning the junk-direction in
particular, there is, again arguably, neither a convincing a posteriori reason nor a
convincing a priori reason to believe there is an all-encompassing infinite whole rather
than infinitely many bigger and bigger wholes.33 Third, concerning the gunk direction
in particular, there is, again arguably, neither a convincing a posteriori reason nor a
convincing a priori reason to believe there are mereological atoms.34 Rhetorically, why
think an end point is anything but our (hopeful) theoretical idealization in any case?
So, there are some reasons to think the concrete world is actually hunky, not just
merely possibly so. If so, again assuming SNG appropriately tracks the mereological
hierarchy, SNG even actually fails to obey WF (no matter which direction a grounding
chain goes).
Of course, all this leaves open whether there might be fundamentality somewhere
else in the mereological hierarchy than at the top or at the bottom, or whether there
might be fundamentality in a way that fails to appropriately track the mereological
structures at all. Raven (2018) defends one such alternative, where there can be
fundamentality without a foundation à la WF. The core idea is that fundamentality
amounts to ineliminability from a grounding chain, rather than being a fact at an
end-point of it. As such, there can, at least in one sense, be fundamentality even
if the world is hunky. Raven’s account is subtle, and I cannot for reasons of space
adequately discuss it here, but merely note two things: first, the fundamental is then
not necessarily facts, so we are beyond SNG, and second, it seems even such a
watered-down notion of fundamentality can be indefinitely descending, leaving us
with nothing truly fundamental at all.35
Life is short, with no time to take deep breaths, so let’s move on to some other argu-
ments. As we saw earlier, WF and IDG are mutually incompatible, so the proponent of
WF must deny IDG (and proponents of IDG must deny WF). Given that grounding
31 See also Schaffer (2003), which provides a similar inductive argument for gunk in particular. Note
that, contra Sider (2011; 2013), the above inductive/abductive reasons are not just a handful of cases of
unpacking particles, but rather they are based on a much more general cosmic pattern. Note also that
Sider’s (2011; 2013) other criticisms of the actuality of gunk rest on the assumptions that fundamentality
is an all-or-nothing matter and that fundamentality is well-founded; both assumptions are rejected by our
picture of SNG+IDG.
32 Of course, appealing to WF at this point is a non-starter.
33 34
See Bohn (2009b) and Carr (2009). See e.g. Arntzenius (2008).
35 Raven (2018) does not deny this second point.
 indefinitely descending ground
is explanatory (as per SNG), IDG is equivalent to a statement of what we might call
the metaphysical principle of sufficient reason (MPSR): every fact p has a metaphysical
explanation (see Guigon, 2015; cf. Della Rocca, 2010). So, the proponent of WF must
deny MPSR too. That should not be too surprising: WF postulates that there are some
facts that have no ground, so by SNG neither do they have a metaphysical explanation.
But consider what we might call the well-founded grounding riddle: assume WF, that
is, that all grounding chains end in some fundamental, ungrounded facts. Consider
these ungrounded facts. Either they have a metaphysical explanation (as per SNG) or
they don’t. If they do, they are of course not ungrounded, in which case grounding
is not well-founded after all. If they don’t, then they have no ground. But then the
obvious question arises: whence these fundamental facts?36
Not being able to answer this question fails to provide a natural resting point for
thought.37 There seems to be at least three kinds of answers. First, Brutalism: the
ungrounded facts just don’t have an explanation, and there is no explanation for it
beyond that. Second, Indefinitism: there are no ungrounded facts because every fact
has an indefinitely descending ground in all directions, so the riddle never arises.
Third, Loopism: there are no ungrounded facts because if you go far enough down
any grounding chain you’ll end up where you started.38
But just like not answering the above question is no natural resting point for
thought, so Brutalism too is no natural resting point for thought. This is so because
there is no non-ad hoc way to draw the line between facts that do and do not
have a metaphysical explanation, and drawing an ad hoc line is surely no natural
resting point for thought (cf. Della Rocca, 2010).39 So, all else being equal, unless it
contradicts our evidence, we should prefer Indefinitism or Loopism over Brutalism.
Assuming the transitivity of grounding, Loopism contradicts the irreflexivity of
grounding, so assuming SNG, according to which grounding is both transitive and
irreflexive, Loopism is out.40 So, unless it contradicts our evidence, we should prefer
Indefinitism over Brutalism. As argued throughout the present paper, all else is equal
and Indefinitism doesn’t contradict any evidence, so we should prefer Indefinitism
over Brutalism. That is, on our picture, WF is Brutalism and IDG is Indefinitism, so we
36 Note that this latter question is what I earlier called a local question (LQ), not a regional (RQ) or a
global question (GQ). It resembles the most common objection to Aquinas’ foundationalism: if God is the
cause of all things, what caused God?
37 Thanks to Ralph Henk Vaags.
38 For a defense of Brutalism, see e.g. Aquinas (1266–8/1993); Schaffer (2009; 2010; 2016); for a defense
of Loopism, see e.g. Bliss (2014); and for a defense of Indefinitism, see e.g. Bliss (2013); Tahko, (2014);
Morganti (forthcoming); not to mention the paper you are currently reading. Note that where I use
‘indefinitism,’ others use, misleadingly in my mind, ‘infinitism.’
39 On this point, among others, it is worth comparing Brutalism to Epistemic Foundationalism, but,
unfortunately, I have no space to do so here. Thanks to Jonathan Schaffer and an anonymous referee for
this volume for raising this question.
40 I don’t want to rule out Loopism as such, but only for present purposes, where we are assuming SNG.
Maybe the mental is grounded in the physical, but the physical in turn is grounded in the mental. I find
such a picture of Loopism well worth exploring.
einar duenger bohn 
should prefer IDG over WF. We have assumed SNG, so we should prefer SNG+IDG
over SNG+WF.
One might object that there are various alternative ways one might try to explain
the ungrounded facts, making Brutalism less brute, and thus a more natural resting
point for thought.41 For example, one might think it somehow lies in the essences
of ungrounded facts that they are ungrounded (Rosen, 2010:128–33; Dasgupta,
forthcoming), which thus might somehow explain them, but in a different sense from
grounding. But then again, I say, it might just not thus lie in their essences; we simply
don’t know what the ungrounded facts are, so we simply don’t know what does and
does not have an explanation in terms of their essences. As far as I can tell, we simply
have no convincing reason to believe there actually is such an alternative explanation
of ungrounded facts in terms of their essences.
Dasgupta (forthcoming) is perhaps the best attempt at this sort of account, accord-
ing to which both MPSR and WF are true, but the fundamental facts are not apt for
metaphysical explanation. But the fact that all and only the fundamental truths just
happen to necessarily not be apt for metaphysical explanation is incredible.42 The
idea that ungrounded facts are by their essence ungrounded is an interesting idea,
but I just see no convincing reason to believe it is actually true, not to mention to
rest the necessity of SNG+WF on it. The same goes, I say, in one admittedly too big
sweep, for other ways of trying to alternatively explain the ungrounded (e.g. causally,
or teleologically). So, it still seems to me we should prefer SNG+IDG over SNG+WF,
if not only for theoretical purposes.
Note that MPSR is often thought to entail necessitism, the view according to
which all actual facts are metaphysically necessary (see Della Rocca, 2010; Dasgupta,
forthcoming). But, as shown in Guigon (2015), neccessitism is in fact no implication
of MPSR as such, so necessitism is neither here nor there for SNG+IDG as such.
Here is a related methodological argument for SNG+IDG over SNG+WF: all else
being equal, a theory that respects MPSR is better than a theory that violates MPSR;
there is no particular reason to think that WF is true; there is no particular reason to
think that IDG is false; and IDG respects MPSR, but WF violates MPSR; so the theory
of SNG+IDG is an overall better theory than the theory of SNG+WF. Therefore, to
the extent we should prefer the better of two theories, we should prefer SNG+IDG
over SNG+WF.43
The first premise is not claiming that MPSR is true, just that it is better to not
violate it unless one has a good reason to. In short, neutrality and non-ad hocness
are theoretical virtues. The second and third premises are basically what the entire
Section 2 of the paper has been defending. The fourth premise is as we have seen
41 Thanks to Jon Litland, Alex Skiles, and Kelly Trogdon for pushing me on this.
42 Thanks to Mike Raven.
43 See also Della Rocca (2010). Note, importantly, that we’re talking about metaphysical theories here.
 indefinitely descending ground
provable: IDG is logically equivalent with MPSR, and WF is logically incompatible with
IDG, so WF is also incompatible with MPSR.44
Most generally, the proponent of WF faces the problem of having to explain a
non-ad hoc restriction on MPSR, which is no easy task (cf. Della Rocca, 2010). The
proponent of IDG faces no such problem; and, at least for theoretical purposes, to
the extent SNG doesn’t need to restrict MPSR, it shouldn’t; it doesn’t need to, so it
shouldn’t. The latter is just good old scientific and theoretical practice.
Now, contra my methodological arguments above, Cameron (2008) argues that
SNG+WF is theoretically better than SNG+IDG because SNG+WF provides a more
unified theory than SNG+IDG: by WF there are some fundamental facts in terms of
which all others are metaphysically explained, but by IEG there is not; and a more
unified theory is better than a less unified theory. But, why believe that there is more
theoretical unity with fundamental facts than without? First, by going far enough
down the grounding chains, there could be as much unity without fundamental facts
as with. Second, the fundamental facts might come in separate pluralities having
little or nothing in common, in which case there could be as much disunity with
fundamental facts as without. That is, there is a big difference between saying that
some fundamental facts ground all else and saying that all non-fundamental facts are
grounded in some fundamental facts. The former might bring about more unity, but
the latter need not. I have construed things in terms of the latter, not the former. Why
think the former is the case? As far as I can tell, SNG+WF thus need not provide any
more theoretical unity than SNG+IDG.45
References
Aquinas, T. (1266–8/1993). Selected Philosophical Writings. Oxford University Press.
Arntzenius, F. (2008). Gunk, Topology, and Measure. In D. Zimmerman (ed.), Oxford Studies
in Metaphysics, Vol. 4. Oxford University Press.
Bliss, R. (2013). Viciousness and the Structure of Reality. Philosophical Studies, 166,
pp. 399–418.
Bliss, R. (2014). Viciousness and Circles of Ground. Metaphilosophy, 45(2), pp. 245–56.
Bliss, R. & Trogdon, K. (2014). Metaphysical Grounding. Stanford Encyclopedia of Philosophy,
Nov. 25, 2014.
Bohn, E.D. (2009a). An Argument Against the Necessity of Unrestricted Composition, Analysis,
69(1), pp. 25–9.
Bohn, E.D. (2009b). Must There Be a Top Level? Philosophical Quarterly, 59(235), pp. 193–201.
Bohn, E.D. (2009c). Composition as Identity. PhD Dissertation, University of Massachusetts
Amherst.
Bohn, E.D. (2011). Commentary on Parts of Classes. Humana.Mente, 19, pp. 151–8.
44 Note, importantly, that I have not said that IDG is one and the same claim as MPSR.
45 Thanks to Jon Litland, Matteo Morganti, Mike Raven, Jonathan Schaffer, Alex Skiles, Kelly Trogdon,
and two anonymous referees for this volume.
einar duenger bohn 
Bohn, E.D. (2012). Monism, Emergence, and Plural Logic. Erkenntnis, 76(2), pp. 211–23.
Bohn, E.D. (2014). Unrestricted Composition as Identity. In D. Baxter & A. Cotnoir (eds.),
Composition as Identity. Oxford University Press.
Bohn, E.D. (ms). Plural Comprehension, Composition, Sets, and Junk.
Cameron, R. (2008). Turtles All the Way Down. Philosophical Quarterly, 58(230), pp. 1–14.
Carr, B. (ed.) (2009). Universe or Multiverse? Cambridge University Press.
Dasgupta, S. (2014a). On the Plurality of Grounds. Philosophers’ Imprint, 14(20), pp. 1–28.
Dasgupta, S. (2014b). The Possibility of Physicalism. Journal of Philosophy, 111(9), pp. 557–92.
Dasgupta, S. (forthcoming). Metaphysical Rationalism. Noûs.
Della Rocca, M. (2010). PSR. Philosophers’ Imprint, 10(7), pp. 1–13.
Dixon, T.S. (2016). What is the Well-Foundedness of Grounding? Mind.
Fine, K. (2001). The Question of Realism. Philosophers’ Imprint, 1(2), pp. 1–30.
Fine, K. (2012). A Guide to Ground. In F. Correia & B. Schnieder (eds), Metaphysical Grounding:
Understanding the Structure of Reality. Cambridge University Press.
Guigon, G. (2015). A Universe of Explanations. In K. Bennett & D. W. Zimmerman (eds),
Oxford Studies in Metaphysics, Vol. 9. Oxford University Press.
Hume, D. (1779/1993). Dialogues and Natural History of Religion. Oxford University Press.
Koslicki, K. (2015). The Coarse-Grainedness of Grounding. In K. Bennett & D. W. Zimmerman
(eds), Oxford Studies in Metaphysics, Vol. 9. Oxford University Press.
Litland, J. (2017). Grounding Ground. Oxford Studies in Metaphysics. Oxford University Press.
Morganti, M. (forthcoming). Dependence, Justification and Explanation: Must Reality be Well-
Founded? Erkenntnis.
Nolan, D. (2001). What’s Wrong with Infinite Regresses? Metaphilosophy, 32(5), pp. 523–38.
Pruss, A.R. (2012). The Leibnizian Cosmological Argument. In W.L. Craig & J.P. Moreland
(eds), Natural Theology. Wiley-Blackwell.
Rabin, G.O. & Rabern, B. (2016). Well Founding Grounding Grounding. Journal of Philosoph-
ical Logic, 45(4), pp. 349–79.
Raven, M. (2015). Ground. Philosophy Compass, 10(15), pp. 322–33.
Raven, M. (2018). Fundamentality without Foundations. Philosophy and Phenomenological
Research.
Rosen, G. (2010). Metaphysical Dependence: Grounding and Reduction. In B. Hale &
A. Hoffmann (eds), Modality. Oxford University Press.
Schaffer, J. (2003). Is There a Fundamental Level? Noûs, 37(3), pp. 498–517.
Schaffer, J. (2009). On What Grounds What. In D. Chalmers, D. Manley, & R. Wasserman (eds),
Metametaphysics. Oxford University Press.
Schaffer, J. (2010). Monism: The Priority of the Whole. Philosophical Review, 119(1), pp. 31–76.
Schaffer, J. (2016). Grounding in the Image of Causation, Philosophical Studies.
Sider, T. (1993). Van Inwagen and the Possibility of Gunk. Analysis, 53, pp. 285–9.
Sider, T. (2011). Writing the Book of the World. Oxford University Press.
Sider, T. (2013). Against Parthood. In K. Bennett & D. W. Zimmerman (eds), Oxford Studies in
Metaphysics, Vol. 8. Oxford University Press.
Tahko, T.E. (2014). Boring Infinite Descent. Metaphilosophy, 45(2), pp. 257–69.
Trogdon, K. (ms). Inheritance Arguments for Fundamentality. This volume.
Wilson, J. (2014). No Work for a Theory of Grounding. Inquiry, 57(5–6), pp. 535–79.
9
Inheritance Arguments for
Fundamentality
Kelly Trogdon
1 Introduction
I’m a proponent of grounding—I think that the notion of grounding is coherent,
figures in ordinary thinking, and deserves a place in our philosophical toolkit due
to its theoretical utility.1 For the purposes of this chapter I assume that grounding
is a relation between entities of various ontological categories and their categories
needn’t match. Where capital English letters take singular values and capital Greek
letters plural values, grounding statements have the following form: ground A.2
It’s a familiar idea that there must be a foundation for being. How can we sharpen
up this idea? As is customary, let’s say that an entity is fundamental just in case it isn’t
partially grounded. With recourse to fundamentality so understood we can put the
idea as follows:
Metaphysical foundationalism: necessarily, any non-fundamental entity is fully

grounded by fundamental entities.
Proponents of grounding tend to accept metaphysical foundationalism or similar the-
ses. Here are three examples. First, Cameron (2008) argues that general considerations
involving theoretical virtues suggest that metaphysical foundationalism without its
modal preface is true. He writes, “If we seek to explain some phenomena, then, other
things being equal, it is better to give the same explanation of each phenomenon than
to give separate explanations of each phenomenon” and we can do that only if “every
1 For introductions to and overviews of recent literature on grounding, see Bliss and Trogdon (2014),
Clark and Liggins (2012), Correia and Schnieder (2012), Raven (2015), and Trogdon (2013).
2 See Schaffer (2009) for more on this view of the logical form of grounding statements. Audi (2012)
and Rosen (2010) claim that the relata of the grounding relation instead are restricted to facts. I’m actually
sympathetic with this view, but as Schaffer works with the category-neutral view and I’m going to engage
with his argument for metaphysical foundationalism (a thesis I define shortly), I too will work with the
category-neutral view.
kelly trogdon 
dependent entity has its ultimate ontological basis in some collection of independent
entities” (12).
Second, Fine (2010) claims that metaphysical foundationalism is supported by the
plausible idea that all non-fundamental entities have complete or satisfactory explan-
ations. He writes, “But there is still a plausible demand on ground or explanation
that we are unable to evade. For given a truth that stands in need of explanation,
one naturally supposes that it should have a ‘completely satisfactory’ explanation, one
that does not involve cycles and terminates in truths that do not stand in need of
explanation” (105).
Third, Schaffer (2016), taking a page from Leibniz, argues that considerations
involving the notion of reality inheritance support metaphysical foundationalism.
Schaffer’s argument in outline is this: since non-fundamental entities inherit their
reality from what grounds them (equivalently: entities transfer reality to what they
ground), it must be that non-fundamental entities are fully grounded by fundamental
entities; otherwise, there would be no source of their reality, and the idea that there is
reality inheritance in the absence of a source for reality is incoherent.
In this chapter I take up the question of how we might appeal to the notion
of inheritance in arguing for metaphysical foundationalism. I first clarify Schaf-
fer’s inheritance argument sketched above, suggesting that it relies on a heavy-duty
metaphysical principle I call the inheritance principle. I show that Schaffer’s argu-
ment is unsuccessful even granting the principle. Then I explore what consequences
the inheritance principle might have for metaphysical foundationalism granting for
the sake of argument that the principle is true. I show how in this case we can
deploy the notion of causal capacity inheritance in arguing for two special cases
of metaphysical foundationalism, what I call causal foundationalism and concrete
foundationalism. I conclude that if considerations involving inheritance are to provide
a route to metaphysical foundationalism, the route will be indirect—in this case
we would need to argue for the thesis in a piecemeal fashion (say, by arguing for
concrete foundationalism plus a corresponding thesis about abstract entities). There
may indeed be plausible inheritance arguments for fundamentality—it’s just that
we need to be careful to focus on the appropriate theses as well as the appropriate
inherited properties.
2 Reality Inheritance
An object is gunky just in case each of its proper parts has proper parts. Leibniz can be
read as rejecting the possibility of gunk, as he claims that the reality of an aggregate
is “borrowed” from its proper parts, and “. . . where there is no reality that is not
borrowed, there will never be any reality, since it must belong ultimately to some
subject” (Adams 1994, 335). In other words, every being of aggregation “. . . has its
reality only from that of its components, so that it will have none at all if each being
 inheritance arguments for fundamentality
of which it is composed is again a being by aggregation . . .” (336). This line of thought

plays a crucial part in Leibniz’s argument for monads.
In a Leibnizian spirit, Schaffer articulates what I call the reality inheritance argument
as follows:
. . . a grounded entity inherits its reality from its grounds, and where there is an inheritance
there must be a source. One cannot be rich merely by having a limitless supply of debtors,
each borrowing from the one before. There must actually be a source of money somewhere.
Likewise something cannot be real merely by having a limitless sequence of ancestors, each
claiming reality from its parents. There must actually be a source of reality somewhere. Just as
wealth endlessly borrowed is never achieved, so reality endlessly dependent is never realized.
(2016, 95)
Schaffer sees the primary conclusion of the argument as the claim that grounding is
well-founded in the sense that downwardly non-terminating chains of partial grounds
are impossible. He claims that, were there such a grounding chain, the being of
any object in that chain would be “infinitely deferred, never achieved” (2010, 62).
Returning to Leibniz, he claims that in the event that every object is a being of
aggregation “one never arrives at any real being” (Adams 1994, 333).3
If grounding is well-founded in Schaffer’s sense it follows that metaphysical foun-
dationalism as I’ve characterized it is true. But metaphysical foundationalism is
compatible with downwardly non-terminating chains of partial grounds, so long as
each link in such a chain is fully grounded by fundamental entities. Here’s a potential
example of such a grounding structure. Let S1 be a spherical region of space, S2 a
proper sub-region of S1 , S3 a proper sub-region of S2 , and so on without end. You
might claim that S1 is partially grounded by S2 , S2 is partially grounded by S3 , and
so on without end. But suppose that there are spatial points. You might claim in this
case that spatial points are fundamental, and each of the Ss is fully grounded by spatial
points.4
Pace Schaffer, I think that the reality inheritance argument is better interpreted as
an argument for metaphysical foundationalism as I have formulated the thesis than as
an argument for the claim that downwardly non-terminating chains of partial grounds
are impossible. This is for the simple reason that the argument seems to proceed upon
the idea that there is something wrong about inheritance in the absence of a source,
not that there is something wrong about endless inheritance per se. Returning to the
region/point case, this grounding structure isn’t objectionable so far as the reality
3 While she doesn’t discuss the reality inheritance argument, Bennett (2011) also claims that down-
wardly non-terminating chains of partial grounds are impossible.

4 This case is adapted from cases discussed by Bliss (2013) and Rabin and Rabern (2016). See Dixon
(2016) for potential examples of what he calls fully pedestalled chains, grounding structures that contain
downwardly non-terminating chains of partial grounds each of whose links is fully grounded by the same
fundamental entities. In the example above, each region in the downwardly non-terminating chain is
grounded by different (but overlapping) collections of points. Rabin and Rabern’s (2016) version of the
region/point example is intended to be an example of a fully pedestalled chain.
kelly trogdon 
inheritance argument is concerned. It’s true that the spherical region is real by way of
endless reality inheritance, but there is nonetheless a source of its reality—the points
that it contains. While their focus isn’t the reality inheritance argument, Rabin and
Rabern make essentially the same point: “Only when there is no origin for being, no
fundamental substratum to the hierarchy of ground, is . . . being ‘infinitely deferred
and never achieved’. In [cases like the region/point case] being is infinitely deferred
but nevertheless achieved” (2016, 363).
Provided that the conclusion of the reality inheritance argument is that metaphys-
ical foundationalism is true, I see the argument as consisting of the following three
theses:
The reality inheritance premise: necessarily, if A is non-fundamental then A inherits

its reality from whatever fully grounds it.
The source of reality premise: necessarily, if A inherits its reality then there are that
are a source of A’s reality (i.e. A inherits its reality from , and no entity among
inherits its reality).
The reality/fundamentality premise: necessarily, if are a source of A’s reality then
the entities among are fundamental and fully ground A.
Two points of clarification are as follows. First, I assume that talk of reality and talk of
existence in this context come to the same thing—for something to be real is just for it
to exist. Second, a background assumption of the argument is that reality (existence)
is a property—the book on the table, for example, has the property being real. The
reality inheritance argument appeals to the idea that being real is transferable—when
fully grounds A the latter inherits this property from the former.5
Metaphysical foundationalism follows from the premises of the reality inheritance
argument. What can be said on behalf of each premise? I begin with the reality
inheritance premise: necessarily, if A is non-fundamental then A inherits its reality
from whatever fully grounds it. I take it that for Schaffer it’s a stipulative matter that
fully grounds A just in case the latter inherits its reality from the former. Viewed in
this way, the reality inheritance premise isn’t something that really stands in need of
argument, as it’s a consequence of the particular conception of grounding that Schaffer
has in mind.
Now I turn to the source of reality premise: necessarily, if A inherits its reality
then there are that are a source of A’s reality (i.e. A inherits its reality from
, and no entity among inherits its reality). I take it that the rationale for this
premise doesn’t proceed upon considerations about grounding, explanation, or reality
inheritance per se. Instead, it appeals to general considerations about inheritance.
5 Here we can follow Salmon (1987) and claim that the property expressed by ‘exists’ and ‘is real’ is the
property being identical to something or other, and an entity has this property just in case it’s not the case
that everything is distinct from it. For a different conception of reality—one whereby existence and reality
can come apart—see Fine (2001).
 inheritance arguments for fundamentality
In particular, I see the rationale for the source of reality premise as appealing to the
following principle, where lower case Greek letters range of properties:
The inheritance principle: necessarily, if A inherits ϕ then there are that are a
source of A’s ϕ-ness (i.e. A inherits ϕ from and no entity among inherits ϕ).
To motivate the inheritance principle you might begin by claiming that there are
ordinary properties that seem to conform to it, where ϕ conforms to the inheritance
principle just in case ϕ is inheritable and it’s necessary that something inherits ϕ only
if there is a source of that entity’s ϕ-ness. Schaffer’s example is wealth. For another
example, consider the property being royal. This property is inheritable, and it seems
necessary that people are royal only if there is a source for their royalty. As Brzozowski
(2008, 200) puts the idea, in the absence of a source for royalty no one is royal, for in
this case “. . . there is nothing in the world that makes it the case that someone is royal
in the first place. . . ”
You might also claim that regresses that violate the inheritance principle seem
vicious. Consider the following version of Bradley’s regress. Suppose that a has F.
So the fact that a is F obtains. A fact obtains only if its constituents are unified or
connected. Let U be the property of having unified constituents, and I the instantiation
relation. So the fact that a is F has U. The fact that a is F inherits its U-ness from
a further fact—the fact that a stands in I to F. As you must have U to transfer U-
ness, the latter fact also has U. The fact that a stands in I to F inherits its U-ness from
another fact—the fact that <a, F> stands in I to I. As you must have U to transfer
U-ness, the latter fact also has U. And the fact that <a, F> stands in I to I inherits
its U-ness from another fact—the fact that <<a, F>, I> stands in I to I. This goes on
without end. This regress violates the inheritance principle—each fact in the series
inherits its U-ness, and for any fact in the series there is no source for its U-ness (i.e.
anything from which it inherits its U-ness has its own U-ness by way of inheritance).
And many have claimed that the regress interpreted along these lines is vicious.6
Finally, I turn to the reality/fundamentality premise: necessarily, if are a source of
A’s reality then the entities among are fundamental and fully ground A. Suppose
that are a source of A’s reality. Hence, (i) A inherits its reality from , and (ii) no
entity among is real by way of reality inheritance. Given (i) and the stipulation
that some entities fully ground another just in case the latter inherits its reality from
6 There are two ways for a regress not to violate the inheritance principle. First, it might not involve
transference. Consider the benign regress of natural numbers. Why do we think that there is such a regress?
Well, we don’t think so because we think that the property having a successor is an inheritable feature,
something transferred from one natural number to another. Indeed, this regress doesn’t seem to involve
the transference of any particular property. Instead, we think that there is an infinite regress of natural
numbers because we accept Peano’s axioms of arithmetic, and it follows from them that there is such a
regress. Second, the regress might involve the transference of ϕ where, for any inheritor of ϕ in that regress,
there is a source of that entity’s ϕ-ness. Consider, for example, Tarski’s benign truth regress: supposing that
sentence S is true, ‘S is true’ is true; “S is true’ is true’ is true; and so on without end. While ‘S is true’ inherits
its truth from S, the latter is a source of the former’s truth. The same applies to the iterated sentences.
kelly trogdon 
the former, A is fully grounded by . Given (ii) and the reality inheritance premise
(which itself follows from the stipulation about grounding), the entities among are
fundamental.7
3 A Problem for the Reality Inheritance Argument

Should we accept metaphysical foundationalism on the basis of the reality inheritance
argument? Morganti (2015) is unconvinced—he questions in particular the reality
inheritance premise. He suggests that, rather than entities transferring reality to what
they ground, it may be that reality emerges from grounding chains instead. What
Morganti says here is modeled after what Klein (2007) says in defense of epistemic
infinitism. Many claim that epistemic properties are inheritable. It’s a familiar idea that
there are conditions under which knowledge is transferred via testimony from one
person to another and justification is transferred via inferential connections from one
of your beliefs to another. Klein claims, however, that justification (and presumably
knowledge) isn’t a transferable property. Instead, we should think of justification “as
emerging whenever there is an endless, non-repeating set of propositions available as
reasons” (2007, 16).
Suppose that the relevant notion of reality emergence is consistent, as Morganti
argues. Remember, however, that the reality inheritance premise is plausibly viewed
as a consequence of a stipulation about grounding. So the proponent of the reality
inheritance argument might respond that, while the notion of grounding*—a notion
characterized partly in terms of the notion of reality emergence rather than that of
reality inheritance—is consistent and perhaps worth exploring further, it simply isn’t
the relevant concept here. Metaphysical foundationalism and the reality inheritance
argument appeal to the notion of grounding rather than grounding*, so to take up the
latter in this context is just to change the subject.8
The source of reality premise is the most substantive premise of the reality inher-
itance argument. So the most obvious strategy to pursue in responding to the
argument is this: grant that the inheritance principle, if true, secures the source of
reality premise, but reject this heavy-duty metaphysical principle. Bliss (2013), for
7 While Schaffer’s reality inheritance argument is Leibnizian in character, you might think that its
historical roots trace further back, at least as far as Aquinas’s Five Ways. I read Aquinas, however, as rejecting
the inheritance principle. Consider, for example, a chain of royalty inheritance in the absence of a source
for royalty, what Aquinas would call a causal series ordered per accidens without a first cause. Apparently
Aquinas thinks that such a causal series is possible but argues that there is another type of causal series—
a causal series ordered per se—that must have first causes (see Davies 2014, Ch. 3). If this is right then
the standard responses to Aquinas’s arguments concerning first causes don’t transfer in a straightforward
manner to the reality inheritance argument. Thanks to Einar Bohn for helpful discussion here.
8 To be fair, however, Morganti’s (2015) ultimate goal is to develop an alternative to metaphysical
foundationalism—what he calls metaphysical infinitism—rather than simply object to the reality inheri-
tance argument. And I take it that Morganti would reject the idea that he and Schaffer are simply talking past
each other and claim instead that Schaffer is just wrong to claim that grounding involves reality inheritance.
 inheritance arguments for fundamentality
example, considers a principle similar in spirit to the inheritance principle—roughly,

it’s necessary that if A has ϕ conditionally then there are that have ϕ categorically—
and argues that the principle is unmotivated.
Another critical approach to the inheritance principle is this: grant that the prin-
ciple seems plausible on first inspection but claim that this appearance of plausibility
can be explained away. Here’s one proposal along these lines. As a matter of fact,
whenever A has ϕ by way of diachronic inheritance—inheritance of ϕ that unfolds
over time—the inheritor is such that there is a source of its ϕ-ness. But this is simply
because the universe as a matter of fact is temporally finite. As the universe has a
beginning, the chain of diachronic inheritance by which A has ϕ also has a beginning.
There are possible worlds, however, in which the universe is temporally infinite.
In some of these worlds Bill’s wealth, for example, was inherited from Bob, Bob’s
wealth was inherited from Bert, and so on, without anyone having amassed wealth
by honest work in the first place. What has happened here is that we’ve mistaken
a contingent fact—the fact that if A has ϕ by way of diachronic inheritance then
there is a source of A’s ϕ-ness—for a necessary fact, one flowing from the nature of
the inheritance relation, of which diachronic and synchronic inheritance (the type
of inheritance at issue with grounding) are species. But in reality neither diachronic
inheritance nor synchronic inheritance is such that where there is inheritance of some
feature there must be a source of that feature. This explains why we find the inheritance
principle plausible when in fact it’s false.9
Rather than argue that the inheritance principle is false, I’m going to approach
the reality inheritance argument differently—I’ll show that the source of reality
premise is either unmotivated or false even granting the inheritance principle. The
problem can be stated in terms of a dilemma. Either the relevant notion of reality
inheritance is primitive or it’s molecular in nature, consisting of the notion of reality
and that of inheritance. If the former holds then the source of reality premise is
unmotivated, and if the latter holds the premise (as well as the reality inheritance
and reality/fundamentality premises) is false. And this holds despite the fact that the
inheritance principle (we will assume) is true.
Taking the first horn, suppose that the notion of reality inheritance at issue in the
reality inheritance argument is primitive in nature. As Schaffer’s comparison between
reality and wealth above indicates, the rationale for the source of reality premise is
supposed to proceed upon general considerations about inheritance. In particular,
it seems that the rationale appeals to the inheritance principle, as I noted above.
But if the notion of reality inheritance is primitive, it’s unclear why considerations
regarding the general notion of inheritance should be relevant here—despite the
common terminology, the notion of reality inheritance and that of inheritance per se
might have nothing to do with each other! Conclusion: in this case the source of reality
premise is unmotivated.
9
Thanks to Ben Jantzen for helpful discussion here.
kelly trogdon 
The point here isn’t that there can’t be interesting connections between the (presum-
ably) primitive notion of grounding and other conceptual primitives, as there can be
such connections. The point instead is that if the relevant notion of reality inheritance
is primitive then it’s unclear just what the concept comes to. So in this case it’s unclear
just what theoretical payoff there would be to claiming that grounding involves reality
inheritance.
Interestingly there are other places in the literature on grounding in which similar
reasoning applies. Dasgupta (2014), Fine (2012), and Litland (2015) identify ground-
ing with what they call metaphysical explanation. And they claim that by reflecting
on the nature of explanation per se we can draw conclusions about grounding. For
example, Fine (2010) suggests that since explanation is asymmetric, grounding is
asymmetric as well. And Dasgupta (2014) argues that, as any part of an explanation
is relevant to what it explains, any part of a ground is relevant to what it grounds.
But note that if the notion of metaphysical explanation is primitive, it’s unclear why
considerations regarding explanation should be relevant here—despite the common
terminology, the notion of metaphysical explanation and that of explanation per se
might have nothing to do with each other!10
I now turn to the second horn, which is a bit more complicated. To begin, as
Bennett (2011) and others note, grounding is generative. As we will understand
the notion here, a relation is generative just in case its instantiation brings things
into existence.11 Grounding is generative given that grounded entities exist due to
grounding. Non-Humeans about causation think that the same is true of causation—
caused events exist or occur due to causation. The relation of being-taller-than, by
contrast, is non-generative—supposing that Sally is taller than Barbara, neither Sally
nor Barbara exists due to being-taller-than. It seems that the relation of inheritance is
also non-generative—it’s never the case that an inheritor exists due to inheritance. It’s
an inheritor due to inheritance, but it exists for some other reason, if for any reason at
all. In other words, it’s never the case that a transferee—something to which a property
is transferred—exists due to transference. Again, it’s a transferee due to transference,
but it exists for some other reason, if any reason at all.12
10 Thanks to an anonymous referee for suggesting that a comparison to discussions about the con-
nection between grounding and explanation might be useful. I think that there is actually a problem
for Dasgupta et al. here. They claim that grounding is a primitive notion. But if grounding is primitive
then when they claim that grounding just is metaphysical explanation they must be taking the notion of
metaphysical explanation as primitive as well. But if the relevant notion of metaphysical explanation is
primitive it’s unclear that it’s appropriate to draw lessons about the nature of grounding from considerations
about explanation per se.
11 See Bennett (2017), Ch. 3 for further discussion of the relevant notion of generation.
12 I should also note that the relevant notion of inheritance here is quasi-technical—while there are
similarities between its conditions of application and those of the ordinary notion(s) of inheritance, there
may be differences as well. Example: while the quasi-technical notion is such that synchronic inheritance
is possible, perhaps the ordinary conception is such that all inheritance is diachronic. So what I’m claiming
above in effect is this: one crucial similarity between ordinary notion(s) and the relevant quasi-technical
notion of inheritance is that inheritance is non-generative in the sense that inheritors don’t exist due to
inheritance.
 inheritance arguments for fundamentality
Generative relations, as I understand them, are ‘full-time’ generators—their

instances always generate. But might there also be relations that, while not full-time
generators, are ‘part-time’ generators, relations such that, while some of their instances
don’t involve generation, others do? You might think that transference in particular is
such a relation. It’s true that transference isn’t a full-time generator—when, say, wealth
is transferred the existence of the transferee isn’t due to this instance of transference.
But its existence nevertheless might be due to another instance of transference. The
idea is that when being real instead is transferred, the existence of the transferee is
indeed due to this instance of transference.
I have no problem with the distinction between full- and part-time generators—
the notion of a part-time generator seems perfectly coherent. Instead, the ultimate
problem as I see it is this: the idea that the property being real in particular is
transferrable just doesn’t make sense. If something exists by way of the transference
of being real then the entity’s status as a feature recipient is explanatorily relevant to its
existence. But it seems that the explanation goes in the other direction: the fact that
the entity exists instead is explanatorily relevant to its status as a feature recipient.13
Returning to the second horn, now suppose that the relevant notion of reality
inheritance in the reality inheritance argument is molecular, consisting of the notion
of reality and that of inheritance per se. The notion of reality inheritance in this case
is inconsistent, as the rules governing its use themselves are inconsistent. On the one
hand, if we judge that an inheritor exists due to inheritance then we’re using the notion
of inheritance incorrectly, as I pointed out above. And, as this concept is a component
of the notion of reality inheritance, the same applies to the molecular concept—if
we judge that an inheritor exists due to reality inheritance (or any other type of
inheritance) then we’re likewise using the notion of reality inheritance incorrectly.
(Compare: supposing that if we judge that cows are numbers we’re using the notion of
cows incorrectly, it follows that if we judge that brown cows are numbers we’re using
the notion of brown cows incorrectly as well, on the assumption that the latter is a
molecular concept consisting of the notion of the color brown and that of cows.) On
the other hand, the application conditions for the molecular concept are such that it
applies when an inheritor exists due to reality inheritance. Hence, the notion of reality
inheritance is inconsistent. Conclusion: in this case the source of reality premise (as
well as the reality inheritance and the reality/fundamentality premises) is false as it’s
semantically defective.
4 Causal Capacity Inheritance

Above I argued that the reality inheritance argument fails even granting the inher-
itance principle. In this section I argue that this principle nevertheless may have
13
Thanks to Cruz Davis and two anonymous referees for helpful discussion here.
kelly trogdon 
important consequences for metaphysical foundationalism. I show in particular that

if we grant the inheritance principle and focus on the appropriate inheritable property,
it may be that we have the resources to establish special cases of metaphysical
foundationalism. This is good news for those both sympathetic with the inheritance
principle and the idea that there are fundamental entities. And, even if you’re skep-
tical of the inheritance principle, it’s still worth knowing what its consequences for
fundamentality might be, were the principle true.
If we’re going to rely on the inheritance principle then the property being real
isn’t the property that we should focus on, given what I argued above. In this case
what property should we target? I propose having the capacity for causal activity
(causal capacity for short). An entity has this property just in case it has causal
powers, dispositions to enter into particular sorts of causal transactions. To specify
the causal powers of an entity is to specify just which sorts of effects that entity is apt
to causally contribute to. What sorts of entities can have causal powers? I assume that
property instances are the primary bearers of causal powers. Property instances are
like ordinary objects in that they’re concrete, but they’re like properties in that they ‘go
with’ entities. Following Lowe (2006, Ch. 2), a property goes with an entity when the
latter instantiates the former and a property instance does so when it characterizes that
entity. As Pereboom puts the idea, property instances are “concrete ways particular
things are” (2011, 141). Consider, for example, the property being magnetic. Instances
of this property both characterize certain objects (magnets) and are apt to enter into
particular sorts of causal transactions (e.g. attracting nearby pins).
Suppose that an instance of ϕ inherits its causal capacity from an instance of ψ.
There are different ways of thinking about just how the causal powers of these property
instances are related in this case. Following Shoemaker (2007, Ch. 1), you might claim
that, for any instance of a causal power that characterizes the ϕ-instance, there is some
instance of a causal power that characterizes the ψ-instance such that the former is
identical to the latter. Or, following Pereboom (2011, Ch. 7) you might instead claim
that any instance of a causal power that characterizes the ϕ-instance is constituted by
instances of causal powers that characterize the ψ-instance. I won’t take a stand on
the matter here.14
With the notion of causal capacity inheritance on the table, consider the following
special case of metaphysical foundationalism:
Causal foundationalism: necessarily, any non-fundamental entity with causal capac-

ity is fully grounded by fundamental entities.
14 Whatever we end up saying here, presumably it should be compatible with the idea that some property
instances have type-distinct causal powers from the causal powers of the property instances from which
they inherit their causal capacity. The Pereboom-inspired proposal apparently has this feature. It may be,
however, that the Shoemaker-inspired proposal doesn’t have this feature, given reasonable assumptions
about property individuation.
 inheritance arguments for fundamentality
If metaphysical foundationalism is true then so too is causal foundationalism. But

the truth of the latter is compatible with the falsity of the former—while it may be
necessary that any non-fundamental entity with causal capacity is fully grounded
by fundamental entities, perhaps it’s not necessary that any non-fundamental entity
without causal capacity is so grounded. Putting aside metaphysical foundationalism
for the moment, what reason, if any, do we have to think that causal foundationalism
is true? Here is an argument for the thesis that appeals to causal capacity inheritance:
The causal inheritance premise: necessarily, if A is non-fundamental and has causal

capacity then A inherits its causal capacity from whatever fully grounds it.
The source of causal capacity premise: necessarily, if A inherits its causal capacity
then there are that are a source of A’s causal capacity (i.e. A inherits its causal
capacity from , and no entity among inherits its causal capacity).
The causality/fundamentality premise: necessarily, if are a source of A’s causal
capacity then the entities among are fundamental and fully ground A.
Causal foundationalism follows from these premises. What can be said on behalf
of each premise of the argument? I begin with the causal inheritance premise. The
motivation for this premise as I see it is this: it plays an essential role in a plausible
response to a generalized version of Kim’s (2007, Ch. 1) causal exclusion argument,
an argument that challenges the very idea that non-fundamental property instances
can enter into causal relations.15
The argument I have in mind appeals to two substantive principles, the first of
which is Kim’s causal exclusion principle. There are various formulations of this
principle, but for our purposes the following formulation will do: no property instance
has simultaneous full causes.16 The second principle is a grounding analogue to the
familiar principle of physical causal closure, where the latter says that if a property
instance has a full cause then it has a simultaneous full physical cause. The grounding
analogue—the causal closure of grounding—is (roughly) this: if a property instance
has a full non-fundamental cause then whatever fully grounds that cause is also a full
cause of the property instance. Turning to the argument, suppose, for reductio, that an
instance of ϕ is both non-fundamental and fully causes an instance of property ψ. By
the causal closure of grounding, there is something—a property instance—that fully
grounds the ϕ-instance and fully causes the ψ-instance. As grounding is irreflexive,
this something and the ϕ-instance are distinct. Given that grounding is synchronic,
the ψ-instance has two simultaneous full causes. By the causal exclusion principle, no
15 Kim speaks of inheritance in his work on realization, but he focuses on the inheritance of par-
ticular causal powers rather than causal capacity in general. He claims, for example, that “. . . higher
states . . . inherit their causal powers from the underlying states that realize them” (1993, 335).
16 The argument can be recast with a weaker formulation of the causal exclusion principle—one
that merely rules out systematic causal overdetermination—but I won’t take up that version of the
argument here.
kelly trogdon 
event has two simultaneous full causes. By reductio, it’s not the case that there is an
instance of ϕ that is both non-fundamental and fully causes an instance of ψ.
A plausible response to this argument—versions of which have been proposed
as responses to Kim’s original version of argument—is that there are exceptions to
the causal exclusion principle when causal capacity inheritance is involved. The idea
in essence is that property instances can have simultaneous full causes so long as
one of the causes inherits its causal capacity from the other. The intuition behind
this proposal is that where there is causal capacity inheritance there isn’t causal
competition and the exclusion principle applies only when there is causal competition.
Provided that the causal inheritance premise—necessarily, if A is non-fundamental
and has causal capacity then A inherits its causal capacity from whatever fully grounds
it—is true, we can avail ourselves of this response.
Now, the issues here are obviously complex and controversial. Just what we should
ultimately think about all of this isn’t something that we need to decide here. For my
purposes, the following modest claim is enough: issues concerning causal exclusion
suggest that the causal inheritance premise is worth taking seriously.
Next, the source of causal capacity premise: necessarily, if A inherits its causal
capacity then there are that are a source of A’s causal capacity (i.e. A inherits its
causal capacity from , and no entity among inherits its causal capacity). This
premise is motivated by the inheritance principle, which we’re assuming for the sake
of argument is true. A crucial difference between the source of causal capacity premise
and the source of reality premise, then, is this: while the inheritance principle doesn’t
support the latter, it does support the former.
Finally, the causality/fundamentality premise: necessarily, if are a source of A’s
causal capacity then the entities among are fundamental and fully ground A.
Suppose that are a source of A’s causal capacity. Hence, (i) A inherits its causal
capacity from , and (ii) none of the entities among have their causal capacity by
way of inheritance. Given (ii) and the causal inheritance premise, the entities among
are fundamental. Why think, however, that A is fully grounded by in this case?
Given (i) we get this result if the following thesis is true: it’s necessary that if A inherits
its causal capacity from then the latter fully ground the former. What reason, if
any, do we have for thinking that this additional thesis is true? Well, let’s start with a
mundane observation: there are different ways of traveling from one town to another
that require distinctive sorts of links between the towns. Travel by car, for example,
requires that roads connect the towns, while travel by foot requires that footpaths,
sidewalks, and so on, connect them. Similarly, there are different inheritable prop-
erties whose transference between entities requires distinctive sorts of links between
them. Returning to two examples mentioned above, the transference of knowledge
from one individual to another requires that testimony link those individuals (it must
be that one testifies on some matter to the other), and the transference of justification
from one of your beliefs to another requires there to be an inferential link between
your beliefs. Returning to the additional thesis about the connection between causal
 inheritance arguments for fundamentality
capacity inheritance and grounding set out above, the idea is that grounding is the
distinctive sort of link between property instances that the transference of causal
capacity requires.
What else can be said on behalf of the additional thesis? Well, paradigmatic cases of
causal capacity inheritance are such that inheritors are fully grounded by the entities
from which they inherit, and this gives us reason to take seriously the idea that it’s
necessary that if A inherits its causal capacity from then the latter fully ground the
former. I’ll consider three sorts of cases. In the first case some of the relevant property
instances characterize the same entity, in the second some characterize distinct but
materially coincident entities, and in the third some characterize objects at different
levels of mereological aggregation.
First, suppose that various DNA molecules are embedded in a chemical system
in such a way that they play the causal role characteristic of genes. In this case, an
instance of the property of being a gene inherits its capacity for causal activity from
instances of various properties concerning these molecules and the system in which
they’re embedded. And the instance of being a gene is fully grounded by these property
instances. Second, suppose that people are related to a lump of clay on a pedestal
in such a way that there is a statue coincident with it. In this case, an instance of
the property being a statue inherits its capacity for causal activity from instances of
various properties concerning the clay and the people who are relevantly related to the
clay. And the instance of being a statue is fully grounded by these property instances.
Third, suppose that there are carbon molecules arranged in such a way that there is a
diamond. In this case, an instance of the property being a diamond inherits its causal
capacity from instances of various properties concerning the molecules that compose
the diamond and their arrangement. And the instance of being a diamond is fully
grounded by these property instances.
I’ve shown how we can appeal to the notion of causal capacity inheritance in
arguing for causal foundationalism. The notion can also be deployed in defending
another special case of metaphysical foundationalism:
Concrete foundationalism: necessarily, any non-fundamental concrete entity is fully

grounded by fundamental concrete entities.17
Consider the following familiar thesis:
The concrete principle: necessarily, an entity is concrete just in case it has causal
capacity.
Note that the considerations I outlined above that support causal foundationalism
also support the following thesis: necessarily, any entity with causal capacity is fully
17 In defending priority monism, Schaffer’s (2010) discussion of metaphysical foundationalism is
restricted to concrete entities—the operative principle in his discussion is this: any concrete entity that
is grounded by concreta is ultimately grounded by concreta that aren’t themselves grounded by concreta.
See Trogdon (2017) for related discussion.
kelly trogdon 
grounded by fundamental entities with causal capacity. If this thesis and the concrete
principle are true, concrete foundationalism follows.
Three additional points worth noting that go beyond causal and concrete foun-
dationalism are these. First, note that from the causal inheritance premise it follows
that no non-concrete entity can ground a concrete entity, provided that the concrete
principle is true. Since non-concrete entities lack causal capacity, they can’t transfer
this capacity to entities they ground. So in this case, if ‘concrete’ and ‘abstract’ are
contraries (so any abstract entity is non-concrete), it follows that abstract entities can’t
ground concrete entities.18
The second point concerns the neo-Aristotelian view according to which proper-
ties, while abstract, are fully grounded by concrete entities. Consider the following
thesis: necessarily, if an entity is fully grounded by entities with causal capacity
then the former inherits this capacity from the latter. This thesis is related to the
causal inheritance premise—while neither thesis entails the other, they form a natural
package. If this thesis is true then the neo-Aristotelian view is false, provided that the
concrete premise is true and any abstract entity is non-concrete.
The third point concerns the connection between causal capacity and reality.
Armstrong (1978, Chs 12, 16) and others endorse eleaticism, which I understand to
be the view that the property of existing and the property of having causal capacity
are the same property. Now, the discussion above indicates that the property of having
causal capacity is an inheritable feature. Hence, eleaticism has the consequence that
reality is an inheritable feature, as the property of existing and the property of being
real are the same property. But I’ve argued that reality isn’t an inheritable feature. If
I’m right about that, it follows that eleaticism is false.
5 Conclusion
I’ve argued that it’s more fruitful to focus on causal capacity inheritance than reality
inheritance in arguing for foundationalist theses in metaphysics. First I argued that
Schaffer’s reality inheritance argument for metaphysical foundationalism is unsuc-
cessful even granting the inheritance principle. Then I showed how to appeal to
this principle and the notion of causal capacity inheritance in arguing for special
cases of metaphysical foundationalism. What, however, about the general thesis of
metaphysical foundationalism?
Consider the following thesis:
18 For an alternative view, see Carmichael (2016) for an argument that facts about concrete objects are
grounded by facts about abstract objects.

 inheritance arguments for fundamentality
Abstract foundationalism: necessarily, any non-fundamental abstract entity is fully

grounded by fundamental abstract entities.19
Supposing, pace Williamson (2013, Ch. 1) and others that ‘concrete’ and ‘abstract’
are contradictories (so it’s necessary that any entity be concrete or abstract), if
both concrete and abstract foundationalism are true it follows that metaphysical
foundationalism is also true. I’ve argued that concrete foundationalism is worth
taking seriously provided that the inheritance principle is true. What about abstract
foundationalism?
I’m not going to provide an argument for the view here. Instead, I’ll conclude with
a recipe for constructing such an argument. Recall the concrete principle discussed
above. It identifies a positive theoretical role for the notion of being concrete to
play—concrete entities are those that have the capacity for causal activity. And as I
argued above, the property at issue here—having the capacity for causal activity—is
inheritable. So one way to argue for abstract foundationalism is to identify a positive
theoretical role for the notion of being abstract to play, extract an inheritable property
from this role, and then provide an inheritance-style argument for abstract founda-
tionalism in terms of this property. Such an argument would be structurally similar
to the one I provided for causal foundationalism. Whether a plausible argument for
abstract foundationalism along these lines can be given is something worth further
thought.20
So is metaphysical foundationalism plausible? The verdict, so far as inheritance
arguments are concerned, is still out. It seems fairly clear, however, that if considera-
tions involving inheritance are to provide a route to metaphysical foundationalism,
the route will be indirect—we would need to argue for the thesis in a piecemeal
fashion.21
References
Adams, R.M. 1994. Leibniz: Determinist, Theist, Idealist. Oxford University Press.
Armstrong, D.M. 1978. Universals and Scientific Realism, Vols 1 and 2. Cambridge University
Press.
Audi, P. 2012. “Grounding: Toward a Theory of the In-Virtue-Of Relation,” Journal of Philosophy
109: 685–711.
19 See Trogdon and Cowling (manuscript) for further discussion of metaphysical foundationalism
restricted to abstract entities—the operative principle in their discussion is this: any abstract entity that
is grounded by abstracta is ultimately grounded by abstracta that aren’t themselves grounded by abstracta.
20 One complication: Williamson (2013, Ch. 1) argues that there being a positive theoretical role for the
notion of being abstract to play counts against the idea that ‘concrete’ and ‘abstract’ are contradictories.
21 Thanks to my audience at the Canadian Philosophical Association conference at the University of
Ottawa (6/1/15) for their comments and suggestions, particularly Sam Cowling, Michaela McSweeney,
and Joshua Spencer. Thanks also to Einar Bohn, Cruz Davis, Ben Jantzen, Ted Parent, and two anonymous
referees for helpful comments.
kelly trogdon 
Bennett, K. 2011. “By Our Bootstraps,” Philosophical Perspectives 25: 27–41.

Bennett, K. 2017. Making Things Up. Oxford University Press.
Bliss, R. 2013. “Viciousness and the Structure of Reality,” Philosophical Studies 166: 399–418.
Bliss, R. and Trogdon, K. 2014. “Metaphysical Grounding.” In E. Zalta (ed.), The Stanford
Encyclopedia of Philosophy: https://plato.stanford.edu/archives/win2016/entries/grounding/.
Brzozowski, J. 2008. “On Locating Composite Objects.” In D. Zimmerman (ed.), Oxford Studies
in Metaphysics, Vol. 4. Oxford University Press.
Cameron, R. 2008. “Turtles All the Way Down: Regress, Priority and Fundamentality,” Philo-
sophical Quarterly 58: 1–14.
Carmichael, C. 2016. “Deep Platonism,” Philosophy and Phenomenological Research 92: 307–28.
Clark, M. and Liggins, D. 2012. “Recent Work on Grounding,” Analysis 72: 812–23.
Correia, F., and Schnieder, B. 2012. “Grounding: An Opinioned Introduction.” In F. Correia
and B. Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality.
Dasgupta, S. 2014. “On the Plurality of Grounds,” Philosophers’ Imprint 14: 1–28.
Davies, B. 2014. Thomas Aquinas’s Summa Theologiae: A Guide and Commentary. Oxford
University Press.
Dixon, S. 2016. “What Is the Well-Foundedness of Grounding?” Mind 125: 439–68.
Fine, K. 2001. “The Question of Realism,” Philosophers’ Imprint 1: 1–30.
Fine, K. 2010. “Some Puzzles of Ground,” Notre Dame Journal of Formal Logic 51: 97–118.
Fine, K. 2012. “A Guide to Ground.” In F. Correia and B. Schnieder (eds), Metaphysical
Grounding: Understanding the Structure of Reality. Cambridge University Press.
Kim, J. 1993. “The Nonreductivist’s Troubles with Mental Causation.” Reprinted in Kim’s
Supervenience and Mind (1993). Cambridge University Press.
Kim, J. 2007. Physicalism or Something Near Enough. Princeton University Press.
Klein, P. 2007. “Human Knowledge and the Infinite Progress of Reasoning,” Philosophical
Studies 134: 1–17.
Litland, J. 2015. “Grounding, Explanation, and the Limit of Internality,” Philosophical Review
124: 481–532.
Lowe, E. J. 2006. The Four-Category Ontology: A Metaphysical Foundation for Natural Science.
Morganti, M. 2015. “Dependence, Justification and Explanation: Must Reality Be Well-
Founded?” Erkenntnis 80: 555–72.
Pereboom, D. 2011. Consciousness and the Prospects for Physicalism. Oxford University Press.
Rabin, G. O. and Rabern, B. 2016. “Well Founding Grounding Grounding,” Journal of Philo-
sophical Logic 45: 349–79.
Raven, M. 2015. “Ground,” Philosophical Compass 10: 322–33.
Rosen, G. 2010. “Metaphysical Dependence: Grounding and Reduction.” In R. Hale and
A. Hoffman (eds), Modality: Metaphysics, Logic, and Epistemology. Oxford University Press.
Salmon, N. 1987. “Existence.” Reprinted in Metaphysics, Mathematics, and Meaning. Oxford
University Press, 2005.
Schaffer, J. 2009. “On What Grounds What.” In D. Chalmers, D. Manley, and R. Wasserman
(eds), Metametaphysics. Oxford University Press.
Schaffer, J. 2010. “Monism: The Priority of the Whole,” Philosophical Review 119: 31–76.
Schaffer, J. 2016. “Grounding on the Image of Causation,” Philosophical Studies 173: 49–100.
 inheritance arguments for fundamentality
Shoemaker, S. 2007. Physical Realization. Oxford University Press.

Trogdon, K. 2013. “An Introduction to Grounding.” In M. Hoeltje, B. Schnieder, and
A. Steinberg (eds), Varieties of Dependence. Philosophia Verlag.
Trogdon, K. 2017. “Priority Monism,” Philosophy Compass 12: 1–10.
Trogdon, K. and Cowling, S. Manuscript. “Prioritizing Platonism.”
Williamson, T. 2013. Modal Logic as Metaphysics. Oxford University Press.
10
From Nature to Grounding
Mark Jago
1 Problems with Grounding

Metaphysics aims to articulate the structure of reality: to discover the dependencies
between things, their identities, and the possibilities they afford. In asking how things
metaphysically depend on other things, for their identities and for the ways they are,
we are asking about how things are grounded: ‘investigation into ground is part of the
investigation into nature’ (Fine, 2012, 76). An understanding of grounding has broad
applications throughout metaphysics: to the nature of truth and truthmaking (Liggins,
2008; Schaffer, 2010a), to physicalism about the mind (Fine, 2012), to metaphysical
fundamentality (Schaffer, 2010b), and to what intrinsic properties are (Barker and
Jago, 2015; Rosen, 2010). So it is no surprise that the last few years have seen an
explosion of philosophical literature on grounding. (Bliss and Trogdon 2014; Clark
and Liggins 2012; Trogdon 2013 survey the literature.)
All of which is mere waffle, unless we understand clearly what we mean by
grounding. And, quite predictably, there are those who think we don’t have any clear
grasp on the notion. Daly (2012) argues that the notion is incoherent. Hofweber
(2009); Koslicki (2015); Sider (2011); Wilson (2014) are all sceptics of some kind about
grounding.
The dilemma for the grounding fans is familiar. It won’t be enough to claim some
‘intuitive’ grasp on the notion of grounding, for objectors will simply claim not to have
those intuitions. (And even gung-ho grounding fans will have to concede that those
intuitions are vague and limited in scope.) Nor does it seem easy to define a suitable
notion of grounding from accepted concepts. If we could (say, by cobbling together
some modal notions), then we could hardly claim that grounding is at the heart of
the metaphysical enterprise. (And indeed, grounding is supposed to help us in areas
where pure modal notions, such as supervenience, have trouble: truthmaking and the
nature of mental states come to mind.)
The underlying complaint is that there’s no way to understand the general fea-
tures of grounding. The clearest examples are the simple logical cases: conjunctions
are grounded by their conjuncts (taken together); disjunctions by their disjuncts
 from nature to grounding
(individually or together, accordingly); and existentials by their instances. (This

applies equally whether we’re discussing logically complex propositions or logically
complex facts.) But, the worry runs, there’s just no way we can scale that understand-
ing up to the general case. The real world is messy and complicated, and quite unlike
the neat cases of the logical connectives.
I’m going to suggest an approach to grounding on which these logical cases take
centre stage. We can develop both a general theory of grounding and a theory of how
particular things are grounded. The key link between the simple logical cases and the
difficult ones—involving material objects, mental states, truth, and so on—concerns
the natures of those entities. I’ll argue for a certain view of what makes those entities
what they are, and then show how this provides us with information on how they are
(or could be) grounded. If we can get a grasp on the natures of things (in the sense to
be articulated below), then the simple logical cases give us what we need to understand
the grounding conditions for those entities.
2 Nature and Grounding

In many cases, there is a very appealing connection between a thing’s nature and the
ways in which it is grounded. (Some authors talk of essence in place of nature. I’ll use
the two terms interchangeably.) The most straightforward of these are the ‘easy’ logical
cases mentioned above. It is of the nature of being a conjunctive entity (in general)
that a conjunctive entity obtains in virtue of both its conjuncts obtaining. So it is the
nature of that entity which grounds its grounding condition: that conjunctive entities
are grounded by their conjuncts.
Much the same goes for disjunctive entities. It is of their very nature that they obtain
in virtue of either disjunct obtaining (or both). Their grounding condition is grounded
in the nature of what it is to be a disjunctive entity. And again for existential entities. It
is of the very nature of an entity that there is something such that A that it be grounded
by any of its instances: any entity that Ac. This grounding condition is itself grounded
in the nature of what it is to be an existential entity.
On the theory I’m proposing here, we treat these cases as instances of a general
connection between a thing’s nature and its grounding conditions. (Clearly, the logical
cases don’t provide a sound inductive base for the general claim. They provide the
clearest instances of what I take to be a general feature. But to evaluate the general
claim, you’ll have to consider the theory on its overall merits.) On this theory, the
thing’s nature specifies its grounding conditions. Given the natures of some class
of entities, we can then define the grounding relations which hold between them.
(Compare Rosen’s (2015, 198) Grounding-Definition Link principle, which operates
in the opposite direction: from grounding to nature/real definition.) If every non-
fundamental entity has a nature of this kind, completely specifying grounding con-
ditions for those entities, then we may hope to have achieved an understanding of
grounding in terms of nature. It is equally true on this account that, given all the
mark jago 
grounding facts for a particular entity, we can thereby say what its nature is. Our
understanding of nature and grounding is intertwined, on this approach. (So scep-
ticism about the coherence of talking about natures would similarly infect grounding
talk. I hope to allay some worries along these lines in §5, by taking natures to be
constructions, of a certain kind, from common-or-garden properties, relations, and
states of affairs.)
Fine (2012) cautions against any such reductive temptation:
[I]t seems to me that there is a similar error – but writ large over the whole metaphysical
landscape – in attempting to assimilate or unify the concepts of essence [i.e. nature] and ground.
The two concepts work together in holding up the edifice of metaphysics; and it is only by
keeping them separate that we can properly appreciate what each is on its own and what they
are capable of doing together. (Fine, 2012, 80)
There is what a thing is—its nature—and then there are the ways in which it is
grounded. These notions are distinct, each irreducible to the other. (Fine (2015b)
revises this opinion, and provides a framework in which ‘the two notions thereby
complement one another’ (296). But this approach is rather different than the one
I will develop here.)
But suppose, with Fine (2012), that nature and grounding are independent of one
another. Then we can make sense of holding fixed some x’s nature whilst altering
whether some possible y is a possible ground for x. But I don’t think that makes sense.
It clearly doesn’t make sense in the simple logical cases: given what disjunction is, we
can’t avoid a disjunction being grounded by its disjuncts (individually or collectively).
I’ll argue below (§5) that, in general, natures are logically complex constructions.
Given that understanding of what natures are, the point automatically generalizes to
cover natures in general.
Can’t priority monists (Schaffer, 2009, 2010b) and their pluralist opponents agree
on what some whole is, but disagree on whether it grounds or is grounded in its
parts? No: they must disagree on the nature of being a whole. As I see things, we
define mereological summation as a kind of worldly conjunction of entities: a b
is made up of a and b (and all their parts, but no more), taken together. (I discuss
the issue further in §3.) If that’s right, then the issue of part/whole priority goes with
conjunct/conjunction priority. And, as indicated above, I see no sense in claiming that
conjunctions are not grounded by their conjuncts.
Here’s a further argument in favour of my proposed nature-grounding link. Every-
thing is either grounded or else is fundamental. (Grounding sceptics should agree.
‘Fundamental’ means ‘not grounded’; so, if there’s no grounding, then everything is
fundamental.) That includes facts about grounding. But no grounding fact is funda-
mental, since each involves some grounded, hence non-fundamental, entity. So all
grounding facts must themselves be grounded, ultimately by something fundamental
(Bennett, 2011). I suggest that the only entities that could play this role are the
fundamental natures.
 from nature to grounding
Clearly, some natures are fundamental. There’s no further explanation of why

conjunctions are grounded in their conjuncts other than: that’s what it is to be a
conjunction. The nature of conjunction is fundamental. Such fundamental natures
are best placed to serve as the ultimate grounds of grounding facts. But if grounding
is (or is correlated with) a kind of metaphysical explanation, as is often supposed, then
there is a kind of metaphysical explanation of grounding in terms of nature.
The strategy I have in mind for linking grounding to nature goes like this. First, we
set out how conjunctive, disjunctive, and existential entities are grounded. (I’ll discuss
the plausibility of logically complex entities in §4.) Next, we associate each non-
fundamental entity x with a logically complex entity, x∗ . This is a recursive process,
bottoming out in logically complex entities whose non-logical constituents are all
fundamental. We already have an account of grounding for all such x∗ s. So, if we can
assign an x∗ to each non-fundamental x, we need only add this: x is grounded in just
the way x∗ is.
My strategy is to treat x∗ as x’s nature. (Fine (1995, 55) calls this an ‘algebraic’
approach to essence/nature.) I’ll claim that there is a way to understand the
nature of an entity which provides us with a suitable construction x∗ for each
non-fundamental x. So that is the plan: first, to treat an entity’s nature as a structured,
logically complex construction, and then to understand grounding for that entity in
terms of its associated construction.
3 Real Definition as Formal Constitution

Why would I want to understand the nature of an entity in terms of some logically
complex construction? I’m going to draw on the analogy between nature and defin-
ition, brought to prominence in contemporary philosophy by Fine (1994) and Lowe
(2012). On Fine’s view,
the activities of specifying the meaning of a word and of stating what an object is are essentially
the same; and hence each of them has an equal right to be regarded as a form of definition.
(Fine, 1994, 14)
What goes for linguistic definition goes for real definition: the definiens gives the
nature of the defined entity. A real definition tells us what the object is (or what it
would be): it gives us the object’s nature. In giving that nature, we give ‘a proposition
which is true in virtue of the identity of the object’ (Fine, 1994, 13). In a similar vein,
E. J. Lowe says,
A real definition of an entity, E, is to be understood as a proposition which tells us, in the most
perspicuous fashion, what E is–or, more broadly, since we do not want to restrict ourselves
solely to the essences of actually existing things, what E is or would be. (Lowe, 2012, 104–5)
My proposal is to use this notion of nature-as-real-definition to provide grounding

conditions for an entity, by taking the definiendum to be grounded just in the ways
mark jago 
the definiens is grounded. We assign the conjunctive condition being unmarried and
being a man to being a bachelor, the former being the real definition of the latter. We’ve
already said that the conjunction being unmarried and being a man is grounded by its
conjuncts, working together. So the grounds for something’s being a bachelor are any
ground for its being a man, together with any ground for its being unmarried. This is
not a complete analysis of those grounds, in terms of fundamental reality. That would
require analysis of the nature of being unmarried and being a man. The claim is that
each such non-fundamental entity has a nature, understood as a logically complex
structure on the model of real definition.
The claim here is not that we can always give a definition of an entity. As I see things,
a real definition is not merely a proposition describing the essential features of the
entity in question. Rather, a real definition associates an entity with a construction
from more basic entities. (Rosen (2015) offers a similar concept of real definition,
as ‘a structured complex, built from worldly items in roughly the sense in which a
sentence is built from words’ (198).) Given the basic, fundamental constituents and
the logical modes of construction (which I’ll discuss in §4), the constructions thereby
exist, independently of us. Our task, on this way of looking at things, is to articulate
how the logical modes of construction work, and how the resulting constructions
relate to the entities with which they are associated by real definition.
In general, the entity in question is not numerically identical to the construction
associated with it by real definition. (I’ll discuss whether there is numerical identity
in special cases in §5.) The construction set out in a real definition constitutes the
nature of the defined entity, which makes it the very thing it is. This is what Fine
calls a constitutive, in contrast to a consequential, notion of nature. The former, but
not the latter, ‘is directly definitive of the object’ (Fine, 1995, 57). Now, this kind of
constitution is not the kind we mean when we say a house is made of bricks. In that
case, we are talking about the house’s material constitution. We can contrast this with
a thing’s formal constitution. A thing’s material constitution is usually inessential to
that thing’s identity. The house can have its bricks replaced bit-by-bit over time, or
a wall knocked down, but it remains the same house. (Moreover, many entities have
no material constitution: properties, numbers, pure sets, and perhaps mental states.
They nevertheless have natures.) By talking about a thing’s formal constitution, we
point to the way other entities fix that thing’s nature. A thing’s formal constitution is
its nature: a construction from fundamental entities via logical constructors, as set out
in the entity’s real definition. So, on this picture, an entity has both a material and a
formal constitution. The former is the entity’s matter, the latter its nature.
One may object at this point as follows. If entities have both a formal and a material
constitution, then they have both formal and material parts (or at least constituents of
some kind). So we require multiple notions of part. But there is just one notion of
part, the one given by the theory of mereology. A part which is not a mereological
part ‘is a contradiction in terms’ (Lewis, 1992, 213). I don’t think this objection can
be sustained. Multiple notions of parthood are unavoidable. (See also Fine 2010.)
 from nature to grounding
For what is a mereological whole, say, a b? It is an entity whose nature is given

by its parts: it is the entity which overlaps whatever overlaps either a or b (but no
more). That’s the explicit definition of ‘sum’ in axiomatic mereology. (If the system
is extensional, sums are unique if they exist; if there is a universal element, they are
guaranteed to exist.) An axiomatic definition of a term counts as a real definition if
any does. So overlapping just whatever overlaps a or b is of the nature of a b. But this
implies that a b essentially has a and b as parts.
No ordinary object is essentially the sum of its parts, in this sense of ‘part’. This
thin shard of perspex was a part of the coffee table, until I broke it off. Thankfully,
the table survived. Since the shard was, but no longer is, a part of the table, we cannot
understand the table as the mereological sum of its parts. The shard never was a part of
the table in this mereological sense of ‘part’. Ordinary parthood is not the same concept
as mereological parthood. But it’s a bit rum to think that mereological parthood has
nothing to say about ordinary parthood. If that were so, we’d have very little use for
the notion of mereological parthood. Moreover, it would be a remarkable coincidence
that ordinary and mereological parthood shares their key logical features: both are
partial orders, both have pairwise least upper bounds (and probably have generalized
least upper bounds too).
So here’s a suggestion. To be a material part of something is to be a (mereological)
part of its matter. The clay is the statue’s matter. Parts of the clay are material parts of
the statue. Statue and clay agree on all their material parts, and hence on their physical
microstructure, their mass, and so on. (That’s why both together don’t weight twice
as much as taken individually!) But statue and clay differ in their formal constitution:
only the former has being a statue as a constituent of its nature. By contrast, each little
portion of clay is a mereological part of, hence essential to, the lump of clay as a whole.
But none of them, taken individually, are essential to the statue, since none of them are
mereological parts of its nature. They are not written into it by definition. (The statue
must have material parts, of course; but it needn’t have any particular material part.)
Together, those parts materially constitute but do not formally constitute the statue.
It’s these differences that underlie the modal differences between (and in particular,
the different persistence conditions of) the statue and the clay.
This notion of material part doesn’t yet capture what I above called ordinary part.
My hands are parts of me (in the ordinary sense of ‘part’), but they’re neither formal
nor material parts of me. They’re not parts of my nature, but neither are they parts
of my matter, for they too do not have their material parts essentially. (Shedding skin
cells doesn’t change your hand’s identity.) So let’s say that one ordinary object is an
ordinary part (that is, part-in-the-ordinary-sense) of another ordinary object when
the former’s matter is a (mereological) part of the latter’s. Then my hand is an ordinary
part of me. Just what counts as an ordinary object is left open. Perhaps the notion is
conventional, response-dependent, or otherwise a secondary quality. The proposal
tells us what ordinary parts are, whatever we might count as an ordinary object. If
there are none, or if every material object is one, then no problem.
mark jago 
On this suggestion, we understand ‘material part’ and ‘formal part’ (or ‘part of —’s
nature’) in terms of the underlying mereological notion of parthood, coupled with an
independent characterization of a thing’s matter and its nature. This provides a strict
notion of being part of a thing’s nature. But a nature is a construction—a sentence of
our worldly language—which has constituents other than its (mereological) parts. The
constituents of an entity’s nature stand to its (mereological) parts much as the states of
affairs a proposition is about stand to its truthmakers. In the case of propositions, we
might adopt the following principles. A negated proposition ¬A is about whatever
A is about; and both A ∧ B and A ∨ B are about whatever A is about plus
whatever B is about (Fine, 2015a; Yablo, 2014). The same goes for the constituents of
worldly sentences, and hence for essences: that A and that ¬A share their constituents,
as do that A ∧ B and that A ∨ B. (This brings out one way in which the analogy with
subsentences breaks down: ¬¬A and A differ in their subsentences, whereas that ¬¬A
and that A do not differ in what they are about.)
4 The Language of Reality

In §3, I set out two metaphysical tasks needed by the theory of nature as formal
constitution. The first is to articulate how the logical modes of construction work.
The second is to understand how the resulting constructions relate to the entities with
which they are associated by real definition. I’ll deal with the first in this section and
the second in §5. Once we have done that, we will have a clearer picture of how all this
helps us understand grounding.
A key ingredient of my approach is that an entity’s nature is a construction from
more basic entities. Above, I spoke of natures as logical constructions. So let’s clear
up one misunderstanding right now: these are not linguistic, conceptual, or mental
entities. On my view, there is logical complexity in the world, not only in our thoughts
and language. The constructions in question include logically complex states of affairs
and logically complex properties. I’ll explain these constructions in the remainder of
this section, and then put them to work in §5.
Are logically complex states of affairs at all plausible? I think they are. I’ve argued
elsewhere (Barker and Jago, 2012; Jago, 2011) that there are good metaphysical
reasons to believe in negative states of affairs. They are causally efficacious: the
fact that the US and UK governments did not properly regulate their banks is a
partial cause of the 2008 global financial crisis. They are perceptible: Kevin Peterson
played the switch hit against Muttiah Muralitharan because he saw that there was
no fielder at cover. And they enter into the constitution of material objects: a ring-
doughnut isn’t a ring-doughnut unless it features some doughnut-dough surrounding
an absence of doughnut-dough. And if one has a prior reason for accepting truth-
maker maximalism—the view that every truth has a truthmaker—then negative states
of affairs become a highly welcome addition to one’s ontology.
 from nature to grounding
So what are negative states of affairs like, metaphysically speaking? Suppose you
believe in positive states of affairs, and take them to have constituents: particulars,
properties, and relations. Somehow, those constituents get together to form a state of
affairs. We can’t account for this merely by throwing a further relation of instantiation
in. That just gives us more potential constituents, whereas what we want to know is
how they bind together to form a state of affairs. Armstrong’s (1997) response is that
there must be some form of composition other than the usual mereological, part–
whole one. This is non-mereological composition.
Now, there’s pretty much nothing we can say about the metaphysics of non-
mereological composition, other than this: the non-mereological composition of a
property and a particular, F and a, gives us the state of affairs that a is F. Non-
mereological composition has to be taken as a metaphysically primitive notion. We have
to take it on its own terms, and if we don’t like doing that, we have to do without it
altogether. Suppose we adopt the notion, even though we can’t explain it further. If it’s
ok to accept a primitive notion of non-mereological composition which takes F and a
and gives us the state of affairs that a is F, then it can’t be intrinsically objectionable to
accept a further primitive notion of non-mereological composition which takes F and
a and gives us the state of affairs that a isn’t F. If the first accounts for instantiation,
then the latter accounts for anti-instantiation. Introducing two primitive notions is
more costly than accepting one, of course. But that’s a relative cost, to be evaluated
against the good work that negative states of affairs may do for us in our theory. That’s
the line we take in Barker and Jago (2012): if we can accept Armstrong-style positive
states of affairs via non-mereological composition, then there’s no absolute objection
to also accepting negative states of affairs.
What of other logically complex states of affairs? In Barker and Jago (2012), we
take conjunctive states of affairs to be mereological sums of their conjuncts. So given
prior acceptance of atomic states of affairs, acceptance of conjunctive states of affairs
is unexceptionable. But if we also have the kinds of negative states of affairs I propose
in Jago (2011), then we also have disjunctive states of affairs. They are identified with
negative conjunctions via the usual De Morgan equivalence: the disjunctive state that
A ∨ B is identified with the negative conjunctive state that ¬(¬A ∧ ¬B).
When it comes to existential and universal states, we have some options. Turner
(2016) takes an existential state to be a possibly infinite disjunction: that ∃xAx is the
disjunction of all states that Ac, for all c such that Ac. Similarly, universal states are
possibly infinite conjunctions. For this to be at all plausible, we’d need to consider
disjunctions and conjuncts which include merely possible states. After all, it could be
that something is F, even if none of the actual Fs are Fs. In that situation, the existential
state that ∃xFx exists, but the disjunctive state that Fa1 ∨ Fa2 ∨ · · · , where a1 , a2 , . . .
are all the actual Fs, would not. Similarly (for some choice of F and G), it could be that
all the actual Fs are Gs without all Fs being Gs, for there could have been more Fs than
there actually are, which needn’t be Gs. We avoid these worries if we treat existential
and universal states as infinite disjunctions and conjunctions which involve merely
mark jago 
possible as well as actual particulars. But that is quite a commitment. It would be

preferable if logically complex states didn’t require us to accept that merely possible
entities exist.
The solution in Barker and Jago (2012) is to invoke a higher-order property of being
instantiated. F is instantiated just in case something is F. So the positive state that F
is instantiated can play the role of the existential state that something is F, and the
negative state that F isn’t instantiated can play the role of the negative existential state
that nothing is F. We then capture the universal state that everything is F in terms of
the negative existential state that nothing is non-F.
This approach gives us a rich ontology of logically complex states of affairs. We
can say much the same about complex properties and relations. In general, we can
conceptualize a complex property or relation as a complex state of affairs with gaps
in place of particulars. (See Jago (2011) for the technical details.) Complex properties
and relations stand to complex states of affairs much as open sentences stand to closed
sentences.
We can consider this rich ontology as constituting a worldly language. The natures
of particulars, properties, and relations act as names and predicates of the language.
Each is interpreted as referring to the entity of which it is the nature: my nature
is the name for me, being human’s nature is the predicate for being human, and so
on. Positive non-mereological composition is interpreted as concatenation of terms
into sentences; negative non-mereological composition as concatenation of terms
preceded by a single negation; and conjunction, disjunction, and existential quan-
tification are interpreted as above. (This is something like Lewis’s (1986) Lagadonian
worldmaking language, stripped of its ersatz connotations.)
The nature of reality is written in this worldly language. Worldly vocabulary is
combined into complex names, predicates, and sentences, which are the natures of
derivative particulars, properties, and states of affairs. This is the sense in which the
nature of reality is a logical construction from fundamental entities.
One issue raised in this discussion is: just what are the primitive non-logical
constituents, the non-logical ‘vocabulary’, from which this language is constructed?
These are the terms which admit of no further definition. As they have no logical
constitution, our theory accords them no grounding condition. It treats them as
ungrounded, and hence as fundamental, entities. Strictly speaking, our theory says
only that if there are such primitive pieces of non-logical vocabulary, then they are
the fundamental entities. I don’t see that our general notion of a logical construction
requires there to be primitive non-logical vocabulary. (I come back to this issue in
§6.) All well and good, I say: theories of nature or grounding should leave open what
fundamental reality is like, or even whether there are any fundamental entities.
Nevertheless, it often helps to have some idea of what fundamental reality might be
like, on the theory proposed. Fundamental reality might be a distribution of properties
over space–time. This distribution may be atomist in nature—a scattering of distinct
qualitative entities—or it may be a holistic affair, in which a single distributional
 from nature to grounding
property gets together with space–time as a whole to form a single fundamental

state of affairs. If the former, then atomistic fundamental properties will serve as
primitive predicates. If the latter, then the primitive predicates will be aspects of the
big fundamental distributional property. Perhaps there are objects at this fundamental
level of reality; perhaps not. (I suspect not: see Barker and Jago (2015); Jago (2016).)
Perhaps it’s not physical space–time that we find at this level, but some other kind of
space: phase space or configuration space.
(Here’s a general principle I find attractive: we should adopt as the fundamental
space of our metaphysical theory whatever kind of theoretical space best suits funda-
mental physical theory. If you object that configuration space is purely a mathematical
construct designed for an elegant formulation of current physics, reflect on how we got
used to talking about space–time, initially on the basis of some mathematical–physical
theory.)
5 Constructing Natures
That’s the big picture of how reality is constructed. Each derivative entity has a real
definition, written in the worldly language. This is the defined entity’s nature: it is its
formal constitution, contrasted with its material constitution.
As mentioned above, many entities—including properties, numbers, and pure
sets—have no material constitution at all. We might think that such entities are
nothing beyond their formal constitution. That is to say, they are numerically identical
to their nature, and their parts are their essential properties. On this view, non-
fundamental properties are constructions from fundamental properties. They are
logically complex predicates, or open sentences, of the worldly language (§4).
A non-material entity is identical to its nature qua formal constitution. Not so
for material entities. This raises the question: do we find material entities in the real
definitions of other material entities? Or are real definitions in all cases constructed
purely in terms of non-material entities? The thing with material entities is that
they are contingent beings. Any construction involving a material constituent will
likewise be a contingent entity. But real definitions should not be like this. The job
of a real definition is to tells us what an entity is or would be (Lowe, 2012, 104–5).
The definition should exist, even if the defined entity does not. So we must take
every real definition to be a construction from non-material entities: properties and
mathematical entities, say.
Lowe (2006, 2008) objects to this ontic notion of nature and essence, arguing that ‘it
is simply incoherent to suppose that essences are entities’ (Lowe, 2006, 3). I think that’s
wrong. In the linguistic case, we can define a new term from old and, in so doing, we
make it mean what it does. The new term has its meaning in virtue of its definition. The
definiens, considered as a structured entity built up from more basic meanings, is the
meaning of the definiendum. The meaning of bachelor is built up from the meanings of
unmarried and man. Only on that picture can we say that the meanings unmarried and
mark jago 
man are each parts of the meaning of bachelor. So we should take linguistic meanings
to be entities. Now, if we follow Fine in holding that ‘the activities of specifying the
meaning of a word and of stating what an object is are essentially the same’ (Fine,
1994, 13–14), then we should also view the natures of worldly entities as entities in
their own right.
Lowe’s specific worry with taking essences to be entities is that
if the essence of an entity were just some further entity, then it in turn would have to have
an essence of its own and we would be faced with an infinite regress that, at worst, would be
vicious and, at best, would appear to make all knowledge of essence impossible for finite minds
like ours. (Lowe, 2006, 8–9)
We can deal with this worry. Taking natures to be entities in their own right doesn’t
commit us to a vicious regress, for we can identify the nature of a nature with itself.
The nature of the nature of any x is just the nature of x. The nature of being a bachelor
is given by the conjunctive entity, whose conjuncts are being unmarried and being a
man. That’s just the mereological sum, being unmarried being a man. The nature of
that entity is to be the sum with parts being unmarried and being a man (and all their
parts), but no more. But that entity is being unmarried being a man itself. The nature
of the nature is just the nature itself. This gives us a neat and non-ad hoc response to
Lowe’s regress worry.
It may be thought that this proposal quickly runs into contradiction, as follows.
Being a bachelor has being a material entity as part of its nature. Yet the nature of
being a bachelor is a property, and properties are essentially non-material entities. So
that nature cannot have being material as part of its nature. This implies that being a
bachelor’s nature’s nature is distinct from being a bachelor’s nature.
To avoid the worry, we have to pay careful attention to how we ascribe essential
properties to a property F, based on F’s nature. We must carefully distinguish between
what’s essential to possessing F, on the one hand, and what’s essential to F’s identity, on
the other. Suppose F is analysed as a conjunctive entity, G H. Then possessing G and
possessing H are both essential to possessing F: they are individually necessary (and
jointly sufficient) conditions for F-possession. What’s essential to F itself—that is, to
F’s being the very entity it is, rather than some other—is that it has just G and H (and
all their parts) as parts. It is of F’s nature to have G as a part. That doesn’t imply that
F possesses G. In just the same way, it’s essential to (the identity of) being a bachelor
that it has being a man, and hence being a material being, as a part. That implies that
being a material being is essential to bachelorhood-possession. But that doesn’t imply
that the property being a bachelor is itself a material being, essentially or otherwise.
Given our notion of nature as formal constitution as real definition, the story about
grounding is then relatively straightforward. How an entity is grounded can simply be
read off its real definition, given that we understand grounding for the logical cases:
conjunction, disjunction, quantification, and so on. (Rosen (2015, 198) goes in the
opposite direction, defining real definition in terms of ground.) Note that this view,
 from nature to grounding
combined with our identification of a nature’s nature with that very nature, does not
imply that grounding is reflexive. A nature specifies its own grounding conditions, but
that does not imply that it grounds itself. Indeed, as I’ll show in §6, the theory implies
that no entity grounds itself.
Let’s see how the idea might work in practice. Take a non-fundamental property,
floccinaucinihilipilification. (Say it out loud!) This, according to my dictionary, is ‘the
act or habit of describing or regarding something as unimportant’. That definition
involves conjunction, disjunction, and existential quantification. The claim is that
the property floccinaucinihilipilification is a logical construction—via conjunction,
disjunction, and existential quantification—from the more fundamental properties
in the definiens (being an act, being a habit, and so on). They in turn are identified
with the worldly constructions given by their real definitions, right down to the
fundamental level, if there is one. (Even if there is a fundamental level, the definition
nevertheless has the non-fundamental properties being an act, being a habit, and so
on as constituents, just as ‘p ∧ q’ is a constituent, but not a primitive constituent,
of ‘(p ∧ q) ∨ r’.)
We can read the grounding story for floccinaucinihilipilification-possession from its
definition. The definition is disjunctive, and so each disjunct gives us a (full) ground
for something’s possessing floccinaucinihilipilification. The first disjunct, for example,
tells us that x’s being an act of regarding something else as being unimportant is a (full)
ground for x’s possessing floccinaucinihilipilification. And since the real definition
doesn’t ‘bottom out’ until the fundamental level of reality (if there is one), it also tells us
that whatever grounds x’s being an act of regarding something else as being unimportant
thereby also grounds x’s possessing floccinaucinihilipilification. (More on this in §6.)
Now let’s consider an example for which philosophers genuinely want an informa-
tive grounding story: pain. Suppose the role-functionalists are right: pain is a property
identified (or otherwise closely associated) with a certain causal role, typically taking
damage in certain organisms to certain types of behaviour: call that role R. What is
a causal role? We know how to describe one in terms of typical causal inputs and
outputs; but what kind of entity is it? It must be an entity of some kind, else we can’t
identify pain with it. It must be a complex entity, somehow built from the typical input
and output properties: being physically damaged, displaying pain-behaviour, and so
on. And there must be some further constituents of the role, signifying the direction
of typical causation: a causal role in which pain-behaviour typically causes physical
damage would be something else altogether.
It isn’t clear what kind of entity those ‘typical causation’ constituents might be.
Here’s one suggestion. According to some theorists of dispositions, reality contains
a primitive form of modality, somewhere between metaphysical necessity and con-
tingency (Mumford and Anjum, 2011a,b). This modality is a tendency, and it links
various states (or perhaps, events) when the one tends to the other. A dropped-vase
state tends to a smashed-vase state; a solid-salt-in-water state tends to a dissolved-
salt state. Perhaps these state-linking tendencies obtain in virtue of the essences of
mark jago 
the properties (fragility, solubility) in question; perhaps not. Central to this kind of
theory is the idea that at least some of these dispositional tendencies are fundamental.
Perhaps fragility and solubility can be explained in terms of categorical molecular
structure but others—such as having unit negative charge—are fundamental and
irreducible to categorical properties. So the theory goes.
Making best sense of this kind of theory requires realism about the dispositional
tendency itself: the bit of ontology that links states when the one tends to the other.
If a state α has a tendency to state β, and a further state γ has a tendency to state
δ, then there’s something both those dispositional states α and γ have in common:
having a tendency (to some state or other). That’s what makes them dispositional states
(and that’s what makes the properties involved dispositional properties.) Realists
about properties in general buy this line of argument for ontological commitment:
they take ‘there’s something x and y have in common’ as an indicator of ontological
commitment. So, on the dispositional account, there is some part of fundamental
reality which links states when one tends to the other. Since it is fundamental, it serves
as a piece of primitive vocabulary in our worldly language, which we can use to express
causal tendencies. That is one way in which our worldly language might be expressive
enough to define-up causal roles in which we find causal tendencies, rather than strict
causal relationships.
If the causal role R can be expressed in this way in our worldly language, the
resulting construction is a metaphysical definition of R, specifying its constituents and
how they are put together. The resulting entity is a property (a worldly sentence with
a free variable): the property of occupying causal role R. Let’s suppose it is a property
of states of affairs. (Take it to be a property of events, or of properties, if you prefer to
think of causal relata in those terms.) We can then take the property being in pain to
be the existential property λx (that some state of x’s occupies causal role R). What, then,
grounds my being in pain? We’ve identified being in pain with an existential property.
And in general, a ground for an existential state that something is F is some particular
thing’s being F: the state that a is F, for example. So what grounds my being in pain
involves whatever state of mine occupies causal role R. If this is the state of my C-fibres
firing, then the ground is that my C-fibre-firings occupy causal role R.
In general, the construction of a thing’s nature in our worldly language may be a
very complex affair. But given that nature, written in our worldly language, then the
grounding-story for that entity will be relatively straightforward. Since that nature
is a logical construction, the entity’s grounds will be given by the logical cases of
grounding set out above.
These clauses give us the full ground: each disjunct (on its own) fully grounds a
disjunctive state, for example. We can also give a neat definition of partial ground.
Partial grounds are related to the constituents of a real definition (§3): each constituent
of a defined property plays a role in grounding instances of that property. But there is
a crucial difference in the notions. Being married is a constituent of being a bachelor,
but it is not being married which partially grounds being a bachelor. A constituent
 from nature to grounding
may appear in a definition positively or negatively: being a man appears positively,

but being married negatively, in being a bachelor.
So consider the negation normal form of a real definition. (This is the worldly
sentence obtained from the definition by ‘pushing’ all negations down the syntax
tree via De Morgan and quantifier duality rules, cancelling double negations as we
go, until all remaining negations apply immediately to atomic sentences.) Negative
occurrences of a state are those that appear negated in the definition; otherwise, an
occurrence is positive. Then, x’s being F partially grounds y when the state that x is F
has a positive occurrence in y’s real definition; and x’s not being F partially grounds y
when the state that x is F has a negative occurrence in y’s real definition.
6 The Shape of Grounding

We now have a story about how grounding works. What does it tell us about the shape
of the grounding relation? In particular, which formal properties does grounding pos-
sess? The standard account has it that grounding is irreflexive, asymmetric, transitive,
and well-founded. On this picture, each grounding chain is a strict order with a least
element. (Here I use ‘well-founded’ to mean that each grounding chain eventually ter-
minates; or equivalently, that every set of entities has a grounding-minimal element.
But see Dixon (2016) for a comparison of different characterizations of grounding
well-foundedness.) By contrast, Thompson (2016) defends a view she calls metaphys-
ical interdependence, on which grounding is non-symmetric. On that view, entities can
mutually ground one another. Priest (2014) defends a stronger view: everything plays
a role in grounding everything else. Grounding is symmetric, on this view. Schaffer
(2012) argues that grounding is not transitive, and Rodriguez-Pereyra (2015) argues
that it is not irreflexive, not asymmetrical, and not transitive. Morganti (2009) argues
that, whilst grounding is asymmetric, it is not well-founded: every entity is grounded
by some other(s), and so there is no fundamental level of ungrounded entities.
The story about grounding I’ve given here supports irreflexivity, asymmetry, and
transitivity; but it does not imply well-foundedness. Let’s consider each feature in turn.
There are two arguments we can give for the irreflexivity of grounding. The first turns
on real definition. A good definition cannot include its definiendum in its definiens.
No definition, taken on its own, should be circular. Nor can we offer a package of
definitions which, taken together, are circular. Avoiding circularity is part of the
success conditions on definitions. Now suppose some x is a partial ground of itself.
Then, on our picture, x appears (as a positive constituent) in its own real definition
(§5). This makes x’s real definition circular. But there can be no such definition, and so
there can be no such x. (Note that the conclusion is not that x lacks a real definition. If it
did then, on our picture, it would not have grounds: it would be a fundamental entity.)
The response to this is likely to be that the ban on circular definitions is merely
an epistemic requirement, from which a purely metaphysical conclusion cannot
be drawn. (Barnes (2017) and Thompson (2016) raise a similar objection to those
mark jago 
arguing from properties of explanation to properties of grounding.) But the objection

is incorrect: a circular definition is not merely deficient on epistemic grounds. In
the linguistic case, a circular definition fails to define a meaning, and so is not a
definition at all. It is the nature of such definitions to give the meaning of some
term. Something that fails to provide a meaning is just not a definition at all. It is to
genuine definitions as rubber ducks are to real ducks. The same goes for real definition.
Something that fails to specify what an entity is, is not a real definition at all. This is
a purely metaphysical point. Purely on metaphysical grounds, real definitions cannot
be circular. So the conclusion stands: nothing (partially or wholly) grounds itself.
A similar argument can be run to establish the asymmetry of grounding. Suppose
x partially grounds y and y partially grounds x. Then, by the reasoning above, each
appears in the other’s real definition. Taken together, we have a circular definition
package. But there are no such definitions, and hence there are no such x and y.
Grounding is asymmetrical. (Irreflexivity and transitivity entail asymmetry. This
direct argument shows that grounding is asymmetric, independently of whether it
is transitive.)
The second argument for the irreflexivity (and asymmetry) of grounding turns
on the fact that natures are constructions from fundamental entities. If x partially
grounded itself, then it would be a constituent of its own nature, and hence a
proper (mereological) part of its formal constitution. But, as we saw above (§5), only
necessary, non-material entities are constituents of natures. Hence x must be a non-
material entity and hence numerically identical to its own formal constitution. This
implies x is a proper (mereological) part of itself, which is impossible. So again, no
entity partially grounds itself. Again, a similar argument can be used for asymmetry.
Suppose x partially grounds y and y partially grounds x. Then, reasoning as above,
x is a proper (mereological) part of y and y of x, which is impossible. So again, partial
grounding is asymmetrical.
Let’s move on to transitivity. This is easily established. Suppose x partially grounds
y and y partially grounds z. Then by construction, x’s formal constitution is a positive
constituent of y’s, and y’s is a positive constituent of z’s (§5). But by definition, each
positive constituent of x’s formal constitution is also a positive constituent of z’s,
and hence any partial ground of y is a partial ground of z. It follows that x partially
grounds z. So partial grounding is transitive.
The final component of the standard picture is well-foundedness, which (given
asymmetry and transitivity) rules out infinitely descending grounding chains. As far
as I can see, the account developed here does not support well-foundedness. (Neither
does it support Dixon’s (2016, 8) alternative formulation of well-foundedness, which
requires each non-fundamental entity to be grounded by some fundamental entity.)
Suppose that there are possible entities, every (mereological) part of which has
proper (mereological) parts. Such entities are atomless; they have no smallest parts.
Lewis (1991) called such entities gunk. Suppose that the nature of a mereological
whole is to be the entity with precisely those mereological (proper) parts (§3). (That’s
 from nature to grounding
plausible, but by no means mandatory.) We might consider the real definition of a

particular mereological sum on the model of the formulas we use to specify fusions:
a = b c. (If a is atomless, no such finite (or even countably infinite) formula gets to
the whole truth about what a is. But there is no requirement on worldly real definitions
to be countable.) Or we might instead consider a real definition of mereological sums
in general: they are the entities defined by their (proper) parts. Either way, our theory
will then imply that mereological sums are grounded by their (proper) parts (and
nothing else). So if gunky objects are possible, then infinitely descending grounding
chains are possible too. Object-gunk leads to non-well-founded grounding structures.
(And if nothing grounds a whole but its parts, the gunk example would also be a
counterexample to well-foundedness in Dixon’s (2016) sense.)
Now, I don’t claim that the nature-grounding link forces this non-well-founded
picture on us. Gunk may be impossible. The natures of mereological sums might not
be as I’ve suggested. So, for all I’ve said, it may be that grounding is well-founded after
all. Either way is consistent with, but not implied by, the nature-grounding link I’ve
proposed.
7 Conclusion
Material objects have both a formal and a material constitution: the former is their
nature, the latter their matter. Non-material entities (including properties, relations,
numbers, and pure sets) have only a formal constitution: they are identical to their
natures. In both cases, natures are constructed from more basic constituents via logical
constructors. We can conceptualize those constructions as being part of a worldly
language. If there is a fundamental level to reality, then it provides the primitive non-
logical vocabulary of this language. If there are no fundamental ungrounded entities,
then we have a language without primitive non-logical vocabulary.
The association of some entity with its nature takes the form of a real definition. We
then read a real definition as setting out grounding conditions. The possible grounds
for an entity are specified by its nature, together with the recursive clauses setting out
the grounds for conjunctive, disjunctive, and existential entities. We thus have a tight
relationship between nature and constitution, on the one hand, and an entity’s logical
constituents and its grounds, on the other. On this approach, grounding is irreflexive,
asymmetrical, and transitive; but it need not be well-founded.
This approach seems to me to provide a powerful framework for understanding
nature, grounding, and the relationship between them. As such, it is worthy of more
detailed investigation.
References
Armstrong, D. (1997). A World of States of Affairs, Cambridge University Press, Cambridge.
Barker, S. and Jago, M. (2012). Being positive about negative facts, Philosophy and Phenomeno-
logical Research 85(1): 117–38.
mark jago 
Barker, S. and Jago, M. (2015). Essential bundle theory. Unpublished manuscript.

Barnes, E. (2017). Symmetric dependence. This volume.
Bennett, K. (2011). By our bootstraps, Philosophical Perspectives 25(1): 27–41.
Bliss, R. and Trogdon, K. (2014). The Stanford Encyclopedia of Philosophy (Winter 2016 Edition),
ed. Edward N. Zalta, https://plato.stanford.edu/archives/win2016/entries/grounding.
Clark, M. J. and Liggins, D. (2012). Recent work on grounding, Analysis 72(4): 812–23.
Daly, C. (2012). Skepticism about grounding, in F. Correia and B. Schnieder (eds), Metaphysical
Grounding: Understanding the Structure of Reality, Cambridge University Press, Cambridge,
pp. 81–100.
Dixon, S. (2016). What is the well-foundedness of grounding?, Mind doi: 10.1093/mind/fzv112.
Fine, K. (1994). Essence and modality: The second philosophical perspectives lecture, Philo-
sophical Perspectives 8: 1–16.
Fine, K. (1995). Senses of essence, in W. Sinnott-Armstrong, D. Raffman, and N. Asher (eds),
Modality, Morality, and Belief: Essays in Honor of Ruth Barcan Marcus, Cambridge University
Press, Cambridge, pp. 53–73.
Fine, K. (2010). Towards a theory of part, Journal of Philosophy 107(11): 559–89.
Fine, K. (2012). Guide to ground, in F. Correia and B. Schnieder (eds), Metaphysical Grounding:
Understanding the Structure of Reality, Cambridge University Press, Cambridge, pp. 37–80.
Fine, K. (2015a). Subject matter. Talk given at Yablo on Aboutness workshop, Hamburg, 3rd
August.
Fine, K. (2015b). Unified foundations for essence and ground, Journal of the American Philo-
sophical Association 1(2): 296–311.
Hofweber, T. (2009). Ambitious, yet modest, metaphysics, in D. Chalmers, D. Manley, and
R. Wasserman (eds), Metametaphysics, Oxford University Press, Oxford, pp. 260–89.
Jago, M. (2011). Setting the facts straight, Journal of Philosophical Logic 40: 33–54.
Jago, M. (2016). Essence and the grounding problem, in M. Jago (ed.), Reality Making, Oxford
University Press, Oxford, pp. 99–120 .
Koslicki, K. (2015). The coarse-grainedness of grounding, Oxford Studies in Metaphysics
9: 306–44.
Lewis, D. (1986). On the Plurality of Worlds, Blackwell, Oxford.
Lewis, D. (1991). Parts of Classes, Blackwell, Oxford.
Lewis, D. (1992). Critical notice of D.M. Armstrong’s A Combinatorial Theory of Possibility,
Australasian Journal of Philosophy 70(2): 211–24.
Liggins, D. (2008). Truthmakers and the groundedness of truth, Proceedings of the Aristotelian
Society 108: 177–96.
Lowe, E. J. (2006). Metaphysics as the science of essence. Keynote address, Metaphysics of
E. J. Lowe conference, Buffalo, NY, 8 April 2006. Available at: http://ontology.buffalo.edu/06/
Lowe/Lowe.pdf.
Lowe, E. J. (2008). Two notions of being: Entity and essence, Royal Institute of Philosophy
Supplement 62: 23–48.
Lowe, E. J. (2012). Essence and ontology, in L. Novák, D. D. Novotný, P. Sousedík, and
D. Svoboda (eds), Metaphysics: Aristotelian, Scholastic, Analytic, Walter de Gruyter, Berlin,
pp. 93–112.
Morganti, M. (2009). Ontological priority, fundamentality and monism, Dialectica
63(3): 271–88.
Mumford, S. and Anjum, R. L. (2011a). Dispositional modality, in C. F. Gethmann (ed.),
Lebenswelt und Wissenschaft, Deutsches Jahrbuch Philosophie 2, Meiner Verlag, Hamburg.
 from nature to grounding
Mumford, S. and Anjum, R. L. (2011b). Getting Causes from Powers, Oxford University Press,
Oxford.
Priest, G. (2014). One, Oxford University Press, Oxford.
Rodriguez-Pereyra, G. (2015). Grounding is not a strict order, Journal of the American Philo-
sophical Association 1(3): 517–34.
Rosen, G. (2010). Metaphysical dependence: Grounding and reduction, in R. Hale and A. Hoff-
man (eds), Modality: Metaphysics, Logic, and Epistemology, Oxford University Press, Oxford,
pp. 109–36.
Rosen, G. (2015). Real definition, Analytic Philosophy 56(3): 189–209.
Schaffer, J. (2009). On what grounds what, in D. Chalmers, D. Manley, and R. Wasserman (eds),
Metametaphysics, Oxford University Press, Oxford, pp. 347–83.
Schaffer, J. (2010a). The least discerning and most promiscuous truthmaker, The Philosophical
Quarterly 60(239): 307–24.
Schaffer, J. (2010b). Monism: The priority of the whole, The Philosophical Review 119(1): 31–76.
Schaffer, J. (2012). Grounding, transitivity, and contrastivity, in F. Correia and B. Schnieder
(eds), Metaphysical Grounding: Understanding the Structure of Reality, Cambridge University
Press, Cambridge, pp. 122–38.
Sider, T. (2011). Writing the Book of the World, Oxford University Press, New York.
Thompson, N. (2016). Metaphysical interdependence, in M. Jago (ed.), Reality Making, Oxford
University Press, Oxford, pp. 38–56.
Trogdon, K. (2013). An introduction to grounding, in M. Hoeltje, B. Schnieder, and A.
Steinberg (eds), Varieties of Dependence: Ontological Dependence, Grounding, Superven-
ience, Response-Dependence (Basic Philosophical Concepts), Philosophia Verlag, Munich,
pp. 97–122.
Turner, J. (2016). The Facts in Logical Space: A Tractarian Ontology, Oxford University Press,
Oxford.
Wilson, J. M. (2014). No work for a theory of grounding, Inquiry 57(5–6): 535–79.
Yablo, S. (2014). Aboutness, Princeton University Press, Princeton.
11
Grounding in Mathematical
Structuralism
John Wigglesworth
The relation of ground has been used to describe the metaphysical structure of the
world. This relation is thought to have certain structural properties: irreflexivity,
asymmetry, transitivity, and well-foundedness. This paper examines a putative case of
grounding that serves as a counterexample to almost all of these structural properties.
The case in question concerns grounding claims made by mathematical structuralists.
We focus on two of these claims: (i) that mathematical objects are grounded in the
structure that they belong to, and (ii) that mathematical objects are grounded in the
other objects in that structure. These cases of ground are particularly interesting for
the claim that the grounding relation is well-founded. If they are taken as genuine
cases of ground, as we argue they should be, then they provide cases that involve
infinitely descending chains of ground. These chains, however, are bounded from
below. So they are non-well-founded in one sense, but well-founded in another.
The paper is structured as follows. In Section 1 we describe the relation of ground,
discussing the nature of the relata, the structural properties of the relation, and some
varieties of the relation. Section 2 focuses on specific cases of ground that appear
in mathematical structuralism. These cases involve the identities of mathematical
objects, a notion that we articulate in Section 3. Section 4 makes the grounding claims
in mathematical structuralism more precise and argues that they are true by giving an
account of mathematical structures in terms of unlabelled graphs. Section 5 concludes
with a discussion of the structural properties of the relations of ground that occur in
mathematical structuralism, in particular their well-foundedness.
1 The Relation of Ground

Proponents of the notion of ground argue that the world has a hierarchical structure.
This view suggests a picture according to which the world is divided into different
levels. The grounding relation is what gives the world this structure, as it represents
the basic metaphysical relation that holds between the different levels. It does this
 grounding in mathematical structuralism
by tracking a notion of metaphysical explanation: lower levels in the hierarchy

metaphysically explain higher levels. Accordingly, the lower levels ground the higher
levels.
Given that the notion of ground should be understood as a relation, as we assume
it should be, there are debates over what its relata are.1 The position that we take in
this paper is that ground is a relation that holds between facts.2 For those who are
metaphysically averse to facts, alternative accounts of ground take the relata to be
propositions or sentences. One may also think that the relata can be taken from any
of these categories and mixed accordingly, as, for example, Ross Cameron (2008) and
Jonathan Schaffer (2009) do.
Taking ground to be a relation between facts, there are several recognized varieties
of ground, which map onto certain distinctions. One particular distinction is of
relevance to the present discussion, that between full ground and partial ground.3
The distinction is easiest to grasp by way of example. A conjunctive fact A1 ∧ A2 is
grounded in its conjuncts, fact A1 and fact A2 . The conjunction is partially grounded
in each conjunct taken on its own. But when the two conjuncts are taken together
as a pair, the conjunction is fully grounded in this pair. This is not to say that the
conjunctive fact is fully grounded in the set of facts {A1 , A2 }. There is no need to
invoke set theory here (though you may if you are comfortable doing so). The two
conjuncts can be taken together as a plurality of facts.
A full ground for a fact B is a sufficient ground for that fact, that is, a ground for
B such that nothing else is required to provide a ground for B. Accordingly, we take
full ground to be a many–one relation, as a single fact may not always be sufficient
to provide a full ground. A partial ground captures the idea of one fact contributing
to the ground of another. The fact B1 partially grounds the fact B2 when B1 , possibly
together with some other facts, provides a full ground for B2 . Partial ground is then
a binary relation, and on this understanding a (single) full ground is also a partial
ground.
The grounding relation, in both its full and partial varieties, is taken to have
certain structural properties, including irreflexivity, asymmetry, transitivity, and well-
foundedness. Some have questioned whether the grounding relation should have all of
these properties. Carrie Jenkins (2011) argues that there are cases of reflexive ground,
and Jonathan Schaffer (2012) argues that there are counterexamples to the transitivity
of ground. We will not be primarily concerned with irreflexivity and transitivity,
though the cases we consider are examples that involve a reflexive grounding relation.
We are primarily interested in the assumption that the grounding relation is well-
founded. Some have challenged the well-foundedness of ground, and plausible cases
1 For those who reject the idea that ground should be understood as a relation, one can give an account
of ground as a sentential connective. See e.g. Fine (2012) or Schnieder (2011).

2 Others who take this position include Audi (2012) and Rosen (2010).
3 Other distinctions include those between weak and strict ground and between mediate and immediate
ground. These will not be relevant in the present discussion (though see fn. 6).
john wigglesworth 
of non-well-founded ground have been investigated.4 In what follows, we discuss

a natural case of non-well-founded ground that has been sparsely discussed in the
literature. This particular case involves grounding claims made by mathematical
structuralists. Interestingly, this case also provides a natural counterexample to the
asymmetry of ground.
The difference between the cases of ground in mathematical structuralism and
other cases of non-well-founded ground in the literature is that many of these latter
cases are developed through thought experiments specifically designed to generate
counterexamples to the well-foundedness of ground. In the case of mathematical
structuralism, however, the grounding claims involved are made independently of
any formal study of the relation of ground. For this reason, the cases involving
mathematical structuralism might be seen as more natural examples of non-well-
founded ground. These cases are also interesting because they occur in the abstract
world, as opposed to the concrete world where many studies of ground are focused.
Before looking at these cases in detail, we note that whether a relation is well-
founded depends on the particular notion of well-foundedness that one appeals to.
There are several different acceptable ways in which a relation may be considered
well-founded, three of which will be relevant in the following discussion: (i) being
finitely grounded, (ii) being bounded from below, and (iii) having a foundation.5
To understand the differences between these notions of well-foundedness it is
helpful to consider chains of ground. For simplicity, we focus on chains of partial
ground, though the same notions of well-foundedness can be applied to full ground.
We symbolize the claim that fact A1 partially grounds fact A2 as A1 ≺ A2 . A chain of
partial ground is then a collection of facts such that for any two facts, A1 and A2 , in
the collection, either A1 ≺ A2 , A2 ≺ A1 , or A1 = A2 . A chain of partial ground then
has the form:
. . . A−2 ≺ A−1 ≺ A0 ≺ A1 ≺ A2 . . .
In this case, An−1 partially grounds An , which partially grounds An+1 , and so on.
One version of well-foundedness says that every chain of partial ground is finitely
grounded, where a chain is finitely grounded when it does not contain an infinitely
descending chain of partial ground. The progression to the left of the above chain
must terminate with an ungrounded fact after finitely many steps for the chain to
be finitely grounded. A second version of well-foundedness says that every chain of
partial ground must be bounded from below. In this case, for every chain of partial
ground, there is some fact, F, such that each fact in the chain is either partially
grounded by F or identical to F. Note that the grounding fact need not be part of
the chain. A third version of well-foundedness does not necessarily invoke the notion
of a chain of ground, but simply requires that there be a set of facts S that are not
4 See Bliss (2013, 2014), Dixon (2016), Rabin and Rabern (2016), Trogdon (2013).
5 See Dixon (2016) and Rabin and Rabern (2016). In what follows, we make use of the notation and
terminology from Rabin and Rabern (2016).
 grounding in mathematical structuralism
grounded by any fact in the totality of facts, and such that each fact in this totality is
either identical to some fact in S or is partially grounded by some fact in S.
Given a collection of facts that is closed under the relation of ground, one can show
that if every chain in the collection is finitely grounded, then every chain is bounded
from below, and if every chain is bounded from below, then the collection of facts
has a foundation. However, the converse implications do not hold. In the case of
mathematical structuralism, we show that there are structures that contain infinitely
descending chains of ground. However, these structures are bounded from below. It
follows that they have a foundation.
2 Grounding and Dependence in Mathematical

Structuralism
Given this understanding of the grounding relation, we turn now to an examination
of some particular claims made by mathematical structuralists that we argue are
claims involving the notion of ground. The claims that we examine concern grounding
relations between facts that involve both mathematical objects (e.g. individual natural
numbers) and mathematical structures (e.g. the natural number structure). In what
follows, we reserve the term “mathematical entity” to refer to both mathematical
objects and mathematical structures.
Mathematical structuralism is roughly the view that mathematics is the study of
structure. There are different ways to make this view more precise, ways that can be
divided into eliminative and non-eliminative approaches. Eliminativists appeal to a
background ontology that does not include abstract structures. One might take the
background ontology to be sets, and see structures as realized in the universe of sets
(Benacerraf 1965). Alternatively, one can take structures to be models of concrete
objects, or possible models of concrete objects (Hellman 1989). Non-eliminativists,
on the other hand, accept a background ontology of abstract mathematical structures.
The grounding claims that we investigate are those of the non-eliminativist, and
in what follows, we use the term ‘structuralism’ and its cognates to mean non-
eliminativist mathematical structuralism.
Structuralists appeal to several kinds of claims that sound a lot like grounding
claims. Stewart Shapiro, for example, makes these kinds of claims when describing
the version of structuralism that he endorses:
The number 2 is no more and no less than the second position in the natural number structure;
and 6 is the sixth position. Neither of them has any independence from the structure in which
they are positions, and as positions in this structure, neither number is independent of the
other. (2000, p. 258)
This passage appeals to a notion of dependence, a notion that is often associated with
the notion of ground. Indeed, dependence and ground are sometimes, though not
always, used as synonyms for each other. But it’s not obvious that this is Shapiro’s
john wigglesworth 
intention, or the intention of other structuralists who make similar claims (e.g.
Resnik 1982).
The question then is whether these dependence claims entail grounding claims
that the structuralist would agree to. Shapiro claims that certain natural numbers are
dependent on other natural numbers, and that natural numbers are dependent on
the natural number structure. These claims express relations of dependence between
mathematical entities. Two varieties of dependence between entities are relevant here:
existence dependence and identity dependence. The most straightforward version of
existence dependence appeals to a notion of relative existence across (metaphysically)
possible worlds: the existence of x depends on the existence of y iff there is no
metaphysically possible world where x exists but y doesn’t. Shapiro (2008) rejects this
version of existence dependence as unsuitable for the characterization of dependence
relations in mathematical structuralism.
[T]o say that the A’s and the B’s depend on each other is to say that there could not be A’s without
B’s, and vice versa. Despite a proposal I broached tentatively in Shapiro (2006), it seems that this
will not do here, since mathematical objects exist of necessity if they exist at all (or so we shall
assume). There is no sense of the natural-number structure existing without its places, nor vice
versa for that matter. Nor is there any sense of the number four existing without the number
six. Or so say our intuitions, or at least my intuitions. (p. 302)
Shapiro’s alternative is to propose that the dependence relation between a mathemat-

ical object and the structure it belongs to is that of non-mereological constitution
(2008, p. 303). But the details of this kind of constitution are not provided, nor is it
clear whether the account of the dependence of one mathematical object on the others
in the same structure is also one of non-mereological constitution. So there is more
work to be done on this proposal.
The standard alternative to existence dependence is identity dependence, where
one entity depends on another for its identity. For the structuralist, this amounts
to the view that the identity of each natural number (for example) depends on the
identity of every other natural number in the same structure, and on the identity of
the natural number structure itself. This account of dependence in structuralism is
most prominently defended by Øystein Linnebo (2008). It is also the view we will
develop in Section 3, though in a way that differs from Linnebo in order to address
certain issues discussed below.
Linnebo argues that mathematical objects depend on other objects in the same
structure, and on the structure itself, by appealing to what he calls a weak notion
of dependence. According to this notion of dependence, x weakly depends on y iff
any individuation of x must make use of entities which also suffice to individuate
y.6 An individuation for Linnebo is an explanation of the identity of an entity. The
dependence involved thus qualifies as a form of identity dependence.
6 Linnebo’s notion of weak dependence is not the same as the notion of weak ground that is often seen
in the literature, according to which a fact may be a weak ground for itself. See e.g. Fine (2012).
 grounding in mathematical structuralism
Furthermore, if the explanation involved is a metaphysical explanation, then this

form of dependence also qualifies as a form of grounding. We argue that the struc-
turalist case involves the notion of metaphysical explanation that is relevant to the
grounding relation. Consider Kit Fine (2001) on the connection between grounding
and explanation. “We take ground to be an explanatory relation: if the truth that P is
grounded in other truths, then they account for its truth; P’s being the case holds in
virtue of the other truths’ being the case” (p. 15). Similarly, Jon Litland (2013) argues
that grounding should be understood as explanation how. “[G]rounding corresponds
to (metaphysical) ‘explanation how’ in the following sense: when φ is grounded in
ψ then ψ is a way for it to be the case that φ . . . . What’s in question is constitutive
explanation: if ψ grounds φ then its being the case that φ consists in its being the case
that ψ” (pp. 19–20). Litland argues that this form of (metaphysical) explanation has
a “reasonable claim” to be an appropriate articulation of the notion of ground.
On this understanding of ground as an explanatory relation, we argue that the
structuralist cases qualify as legitimate cases of ground. A mathematical structure
having the particular identity that it has (we will explore what this really amounts to in
Sections 3 and 4) provides a way, perhaps the only way, for an object in that structure
to have the identity it has. Or, in Fine’s terms, the object has its identity in virtue of
the structure having its identity. Similarly, a mathematical object has its identity in
virtue of every other mathematical object in the same structure having the identities
they have.
In fact, the connections between objects and structures that these grounding
claims make give reasonable interpretations of structuralism’s distinguishing theses
about the identity of mathematical entities. Independent of Linnebo’s account of
dependence/grounding in structuralism, arguably many structuralists would agree
that the identity of a mathematical object derives from the mathematical structure
that it belongs to. For structuralists, the structure is what provides the metaphysical
explanation of the objects in the structure. And as every other number in the structure
belongs to the structure as well, the identity of each object is explained, at least in part
by these other objects and how they are all related. At the very least, there is nothing
outside of the structure that explains the identity of the objects inside the structure.
So if there is any metaphysical explanation of their identities, it must come from the
structure itself. Of course the full nature of ground is still being explored, and there are
as yet no adequate necessary and sufficient conditions to appeal to. But the connection
between grounding and metaphysical explanation tells us that these are cases that
involve grounding.7
To show that mathematical objects in a structure weakly depend on the other
objects in that structure, and on the structure itself, Linnebo starts with a system
that comprises a domain D of (not necessarily mathematical) objects, and relations
7
In the remainder of this section, however, we follow Linnebo and use the term “dependence”.
john wigglesworth 
R1 , . . . , Rn on D. Take R to be the product relation R1 ×. . .×Rn . The product relation R

is then a complete representation of the original system. Linnebo denotes the abstract
structure of R as R̄, and invokes what he calls a Dedekind abstraction principle to
provide identity conditions for abstract structures: R̄ = R̄ iff R ∼ = R (that is, iff R is
8
isomorphic to R ). If x is an object in the domain of the system represented by R, then
the corresponding position in the abstract structure R̄ is denoted by τ (x, R). Identity
conditions for these positions, which according to the structuralist are mathematical
objects, are given by another abstraction principle: τ (x, R) = τ (x , R ) iff ∃f [f : R ∼=
R ∧ f (x) = x ].9
Identity conditions are important for Linnebo’s account of weak dependence. The
account relies crucially on the notion of individuation, where an individuation is a
metaphysical explanation of the identity of an entity. The respective identity condi-
tions are supposed to supply individuations of mathematical structures and objects
in those structures. With this setup established, Linnebo argues that mathematical
objects weakly depend on other mathematical objects in the same structure.
[A]n abstract office [i.e., a position in an abstract structure] can be individuated via an ordered
pair x, R, where R is some particular system that realizes the abstract structure S, and where x
is an object in the field of R. As part of the system R we are also given occupants of all the other
R-offices. We thus have available the entities needed to individuate any of the other abstract
offices. This means that each office weakly depends on all the others. (2008, p. 79)
Linnebo’s notion of weak dependence is interesting, and one that, as Linnebo points
out, has received little attention in the literature. However, this notion of dependence
is problematic for some non-eliminativist structuralists. Recall the definition of weak
dependence: x weakly depends on y iff any individuation of x must make use of
entities which also suffice to individuate y, where an individuation of an entity is
an explanation of its identity. According to this definition, the individuation of a
mathematical object in an abstract structure must make use of a particular system
R that exhibits that structure, as well as an object x in the domain of R. How is
this “must” to be understood? Either it means that this particular system R and this
particular object x are crucial, in which case the condition fails, because any other
system that exhibits this structure would suffice. Or the claim is weaker, requiring
only that some system or other that exhibits the appropriate structure is required. This
weaker position may in fact be the one that Linnebo endorses. But if it is, the position
commits him to a version of non-eliminativist structuralism called in re structuralism.
In re structuralism is often contrasted with ante rem structuralism. There are
different interpretations of this distinction in the literature. On a coarse-grained
8 The abstraction principle must be restricted to avoid paradox. For Linnebo, and for us, it is enough to
let relation variables range only over sets and have their isomorphism types be non-sets. See Hazen (1985).
9 Linnebo (2008), pp. 75–6. The function f : R ∼ R is an isomorphism. Note that Linnebo is only
=
considering rigid structures, where a structure is rigid iff the only automorphism on the structure is the
identity mapping.
 grounding in mathematical structuralism
interpretation, this distinction is simply the same as the eliminativist (in re) and non-
eliminativist (ante rem) distinction. But on a more fine-grained interpretation, in re
and ante rem structuralists are both non-eliminativist structuralists that appeal to a
background ontology of structures. The difference between them lies in the relation-
ship between a structure and the systems or realizations that exemplify that structure.
According to in re structuralism, the existence of an abstract structure depends on
the existence of some system or other that exhibits that structure. Without some
realization of the structure, the structure would fail to exist. Ante rem structuralists
deny this dependence, such that the existence of a structure is independent of the
existence of any system or realization that exemplifies the structure. On this view, a
structure can exist without any exemplifications.
Linnebo’s commitment to in re structuralism, on the fine-grained interpretation, is
consistent with his account of dependence. In fact the core thesis of in re structuralism
is a dependence claim: for all abstract structures S, the existence of S depends on
the existence of some system exemplifying the structure S.10 This dependence thesis
differentiates in re structuralism from ante rem structuralism, according to which
no such dependence between structures and systems holds. Ante rem structuralism
argues that an abstract structure exists independently of any system that exemplifies
the structure, and could exist without any such system existing.
Linnebo’s account of dependence appears to commit him to a version of in re
structuralism, which can be seen as follows. In addition to the weak dependence
relation between mathematical objects in a structure, Linnebo argues for a weak
dependence relation between mathematical objects and the structure they belong to.
It can also be shown that every office of an abstract structure weakly depends on the structure
itself. For in order to individuate such an office we need a realization of the structure. But this is
also all we need to individuate the relevant abstract structure itself. (For completeness, I remark
that the converse also holds. For the realization needed to individuate the abstract structure
contains all the entities needed to individuate its abstract offices as well.) (2008, p. 79)
Accordingly, mathematical objects weakly depend on the structure they belong to.
In addition to his weak notion of dependence, Linnebo describes a strong version
of dependence, such that x strongly depends on y iff any individuation of x must
proceed via y. Returning to Linnebo’s argument above, to individuate a mathematical
object, we need a system or realization. But the realization is all that’s needed to
individuate the structure. In other words, any individuation of the structure must
proceed via a realization. According to the condition for strong dependence, it follows
that a mathematical structure strongly depends on the existence of some system
10 We mention in passing that this is another interesting claim made by some mathematical structuralists
that sounds a lot like a claim involving the notion of ground. We leave exploration of this connection
between grounding and structuralism for another occasion.
john wigglesworth 
that exemplifies the structure, which is precisely the dependence thesis that in re
structuralism endorses and ante rem structuralism rejects.
In itself the commitment to in re structuralism is not problematic. What is problem-
atic is that the account of identity dependence, and so the account of ground related
to identity, that Linnebo provides is off limits to ante rem structuralists (e.g. Stewart
Shapiro) who endorse the dependence of an object on the structure it belongs to and
on the other objects in that structure. In Sections 3 and 4, we provide an alternative
account that is available for both ante rem and in re structuralists to use in order to
evaluate grounding claims involving mathematical entities.
3 Identity in Mathematical Structuralism

On our view, grounding is a relation that holds between facts. We argue that struc-
turalists are committed to grounding relations that hold between facts involving the
identities of mathematical entities. In terms of facts, the claim that the identity of one
entity grounds the identity of another amounts to something like: the fact that one
entity has the identity it has grounds the fact that another entity has the identity it
has. The crucial notion then is that of having a particular identity.
In a structuralist context, mathematical objects in a structure are not supposed to
have any internal nature. What matters are the relations that an object instantiates in
that structure. The identity of an object is then given by the relations that it has to
objects in the same structure.
The essence of a natural number is its relations to other natural numbers. . . . The number 2,
for example, is no more and no less than the second position in the natural number structure;
6 is the sixth position. . . . There is no more to the individual numbers “in themselves” than the
relations they bear to each other. (Shapiro 1997, pp. 72–3, emphasis in original)
This passage can reasonably be interpreted as endorsing the view that there is nothing
more to the identity of an object in a structure than its structural relations, the
relations it bears to other objects in the same structure. This account of the identity
of objects in a structure has generated much discussion and Shapiro (2008) has
since backed away from it.11 The main point of contention is that it seems to have
counterexamples. These include any structures that have any non-trivial automor-
phisms, that is, isomorphisms from the structure to itself that are not the identity
mapping. For example, the field of complex numbers has a non-trivial automorphism
that maps each number to its conjugate. This automorphism maps, for example, i
to −i, implying that i and −i instantiate the same structural relations. If an object’s
11 See Burgess (1999), Hellman (2001), Keränen (2001, 2006), MacBride (2006), Shapiro (1997, 2006,
2008). One thing to come out of the debate over structural relations is the recognition that a precise
explication of the notion of a structural relation is needed. No satisfactory account has been given. But
the discussion in what follows does not rely on any specific account of structural relations.
 grounding in mathematical structuralism
identity is given solely by these structural relations, then i and −i would have the
same identity, that is, would be identical, on this account. But of course they are not,
even for the structuralist. On the structuralist view, they are two distinct positions in
the complex number structure.
This objection has been resisted by those who argue that the relations of identity
and non-identity are structural relations.12 On this view, the identity of, or difference
between, places in a structure is given by the structure itself. The argument for this
claim is made by endorsing a connection between structuralism and graph theory.
Graph theory, in particular the theory of unlabelled graphs, is of particular interest to
structuralism. Indeed, a structure can be seen as an unlabelled graph. Labelled graphs
are graphs that include labels for each node, thus making all nodes distinguishable
from one another (by their labels), and effectively giving each node an identity
independent of its relations to any other nodes in the graph. Unlabelled nodes drop
the label, and so if we ignore any relations between the nodes, all of the nodes
are indistinguishable from one another. This is very much how structuralists view
mathematical objects. Mathematical objects have their identity strictly in virtue of
the relations they bear to the objects in the same structure. But to be consistent with
mathematical practice, these must include the relations of identity and non-identity.
Otherwise, perfectly legitimate mathematical structures would be ruled out. Consider,
for example, what Shapiro (1997) calls the cardinal 2 pattern, consisting of two objects
with no relations between them. This structure can be represented by the graph with
two nodes and no edges (Figure 11.1).
Though there are no relations between the two different nodes, there are neverthe-
less two different nodes in this graph. This is just part of what the graph is, a graph
with two distinct nodes. They may be interchangeable; you can permute them and
are left with the same graph. But you cannot collapse them into one node. Doing
so would result in a different graph. The two node edgeless graph is a perfectly
legitimate mathematical entity. And if we associate mathematical structures with
graphs, then the structure that this graph represents, the structure with two objects
and no relations, is a perfectly legitimate structure. The fact that the two objects are
distinct is given by the structure itself.
The upshot is that we can give an account of the identity of an object in a
mathematical structure by appealing only to the structural relations of that object.
And we can do this without the risk that two distinct objects will be given the same
Figure 11.1. The two node, edgeless graph (cardinal 2 pattern)
12
See Leitgeb and Ladyman (2008).
john wigglesworth 
identity. The easiest way to give such an account is to associate with each object the
set of structural relations that it instantiates. This set will be the identity of the object.
At the beginning of this section, we formulated the kind of grounding claim we are
interested in as: the fact that one entity has the identity it has grounds the fact that
another entity has the identity it has. In the context of mathematical structuralism,
we can now formulate the claim that mathematical objects are grounded in the other
objects in the same structure, where the grounding involves facts concerning the
identities of the mathematical objects. Given two mathematical objects n1 and n2 ,
we denote the collection of structural relations that each object instantiates as En1
and En2 , and take these collections to be the identities of the respective mathematical
objects. The grounding claim is then: the fact that the identity of n1 is En1 grounds the
fact that the identity of n2 is En2 . We argue in Section 4 that this grounding claim is
true, when the notion of ground involved is partial ground.
Recall that the structuralist makes two grounding claims, one involving mathemat-
ical objects, as above, and one involving mathematical objects and structures. The
second grounding claim will then involve the identities of mathematical structures,
and so we will need some account of these identities. This will also be made more
precise in Section 4.
4 Mathematical Structuralism, Graph Theory, and

Grounding
In this section we develop the view that mathematical structures can be understood as
unlabelled graphs, and then use this theory to specify and argue for grounding claims
involving mathematical entities.13 A graph is usually represented set-theoretically
as an ordered pair G = N, E where N is a non-empty collection of nodes, and E
is a collection of edges that connect elements of N. If we take graphs to represent
mathematical structures, then the nodes represent positions in the structure, that is
mathematical objects, and the edges represent relations between mathematical objects
in the structure. The edges of a graph can be represented set-theoretically as well.
If the relations in the structure are symmetric, then E is a collection of unordered
pairs. But as we usually do not require relations to be symmetric, we can take E to
be a collection of ordered pairs. However, all of the set-theoretic representation we
have used to describe graphs is merely an explanatory convenience. The theory of
unlabelled graphs can be developed by taking the predicates ‘Graph(G)’, ‘Node(n, G)’,
and ‘Connected(n1 , n2 , G)’ to be primitive.
An axiomatic theory of graphs based on these primitive predicates can be developed
in a second-order language. The full details of this theory are not completely relevant
13 The theory of structures as unlabelled graphs, which can be generalized to multigraphs and hyper-
graphs, has been articulated by Hannes Leitgeb (2014). The account of grounding developed within this
theory here, with whatever faults it may contain, is due to the author.
 grounding in mathematical structuralism
for the present purposes, but here are some of the highlights. In a structuralist context
we have the following natural identity conditions for first-order objects of the theory.
Two first-order objects, x and y, are identical iff they instantiate exactly the same
properties and relations: x = y ↔ ∀X[X(x) ↔ X(y)]. Importantly, in this theory
of unlabelled graphs, first-order objects include both graphs and nodes in graphs.
Second-order objects include properties and relations of these first-order objects.
These identity conditions are obviously controversial, especially given the dis-
cussion in Section 3 and in the work referenced there. All we really need for our
purposes is the less controversial left-to-right direction of the biconditional: x = y →
∀X[X(x) ↔ X(y)]. If two first-order objects differ with respect to the properties or
relations they instantiate, then they are not identical. For nodes in a graph, that
is, mathematical objects, this will manifest itself in a difference with respect to the
objects’ identities as defined above. The condition ensures that if you change the
identity of an object, as we will do when arguing for the relevant grounding claims,
then what results is a different mathematical object.
We also have additional identity conditions for graphs, that is, mathematical
structures. Two graphs, G and G , are identical iff they are isomorphic: G = G ↔
G ∼ = G (the relation ∼= can be defined in the theory of graphs with the primitive
predicates mentioned above). This is a standard definition of identity for structures in
mathematical structuralism.
[W]e stipulate that two structures are identical if they are isomorphic. There is little need to
keep multiple isomorphic copies of the same structure in our structure ontology, even if we
have lots of systems that exemplify each one. (Shapiro 1997, p. 93)
This definition of identity differentiates the current axiomatic theory of graphs from
the usual set-theoretic interpretation. Two ordered pairs N, E and N , E might be
set-theoretically distinct, for example, because they may have different sets of nodes,
yet still be isomorphic.
Identifying structures that are isomorphic suggests an account of the identity of
a structure. We equate the identity of a structure with its isomorphism class. Given
the account of identity between graphs, this will amount to the singleton of the graph
itself. This will do for our purposes below, where we use this account of the identity
of a structure to discuss grounding relations between mathematical objects and the
structures they belong to.
There are standard operations that one can perform on a graph to generate a
different graph. The operations we need are those of adding and removing an edge
from a graph. Removing an edge e from a graph G results in the largest subgraph G−e
that contains all of the nodes of G and all of the edges of G except e. Adding an edge, on
the other hand, between nodes n1 and n2 in graph G results in the smallest supergraph
of G containing an edge between these nodes (Harary 1969, pp. 11–12). For each node
n, we have the set En ⊆ E of edges that involve n: En = {n1 , n2 ∈ E | n = n1 ∨n = n2 }.
Take En to be the identity of n. If graphs are understood as mathematical structures,
john wigglesworth 
removing or adding edges equates to an object changing the relations it bears to other
objects in the same structure, that is, changing the particular identity it has.
With these tools, we argue for two grounding claims in the context of mathematical
structuralism that involve mathematical entities. The first claims that, given two
mathematical objects in a structure, the identity of one grounds the identity of the
other. As grounding is a relation between facts, we have:
ODO: For any two mathematical objects, n1 and n2 , in the structure G, the fact
that the identity of n1 is En1 partially grounds the fact that the identity of n2 is En2 .
ODO is a claim of partial ground.14 The fact that the identity of n1 is En1 , together with
other facts , fully ground the fact that the identity of n2 is En2 . The set contains
facts about the identities of all of the other objects in G, as the identity of n2 is partially
grounded in these identities too (that is, the choice of n1 as a partial ground of n2 in
this example is an arbitrary choice).
To evaluate the full grounding claim, we observe that the relation of full ground is
usually taken to entail a corresponding necessity claim: if the facts in fully ground
the fact that F, then it is necessary that if all of the facts in obtain, then F obtains.15
Applied to ODO we have: it is necessary that if the identity of n1 is En1 , and all the
facts in obtain, then the identity of n2 is En2 . Contrapositively, it is necessary that if
the identity of n2 is not En2 , then either the identity of n1 is not En1 or some fact in
fails to obtain. This is what we show.
The notion of necessity here is crucial. Necessity is usually defined as truth in
every situation, and the situations involved are usually metaphysical possibilities. This
interpretation gives us the notion of metaphysical necessity. But this notion will not
do when applied to grounding and mathematical structuralism. In this context, we
have to be able to consider what happens when mathematical objects change their
identities, something that could not happen in any metaphysical possibility. So the
situations involved must be something else. We are asking what happens when we
change a particular structure G so that a node n2 in that structure changes with respect
to its structural relations En2 . The situations that are relevant are those structures that
are just like G, but which modify En2 by adding or removing edges between n2 and
some other node. To evaluate the necessity claim, we look at graphs that differ from
G only in what is required to make it the case that the identity of n2 is different from
En2 , and we see what follows from this change.
14 As formulated, ODO is a claim involving distributive ground, such that each object depends on each
of the other objects in a structure. There is, however, an alternative formulation that takes ground to be
irreducibly plural. According to this view, there are pluralities of facts grounded in further facts, though no
single fact in the plurality is grounded in the further facts (see Dasgupta 2014). Applying irreducibly plural
grounds to the structuralist case, it may be that the identities of all of the mathematical objects are grounded
together as a plurality of facts. Many thanks to an anonymous referee for suggesting this alternative, a full
treatment of which we leave for another occasion.
15 The connection between grounding and necessity has been disputed. See e.g. Leuenberger (2014) and
Skiles (2015).
 grounding in mathematical structuralism
Consider some arbitrary graph G just like G, but which differs from G only in what
is required to make it that the set of edges that involve n2 is not En2 . If, at G either
the identity of n1 is not En1 , or some fact in fails to hold, then the relevant necessity
claim is true, as the choice of G was arbitrary. Let En 2 be the set of edges at G that
involve n2 , which is the result of adding or removing at least one edge, e, to/from En2 .
This edge must be between n2 and some other node. Suppose it’s between n2 and n1 .
This would entail that the identity of n1 is no longer En1 , but is rather the set of edges
that results from adding or removing e to/from En1 . So the identity of n1 at G is no
longer En1 , and we are done. On the other hand, suppose e is between n2 and some
other node, say n3 . The fact that the identity of n3 at G is En3 is a fact contained in .
But this fact would no longer hold at G . The identity of n3 at G is no longer En3
because an edge has been added to or removed from the identity of n3 . So some fact
in fails to hold in the relevant situation. We have then, that in any relevant situation,
if the identity of n2 is not En2 , then either the identity of n1 is not En1 or some fact in
fails to obtain, because the identity of some other node has changed. As n1 , n2 , and
G were arbitrary, this holds for any two objects in a structure.
We have argued for the truth of a necessity claim. Does this necessity claim
entail a corresponding grounding claim? In broader discussions of ground this is
not generally the case. For example, it is (metaphysically) necessary that if Socrates
exists, then the number 2 exists. But neither of these facts grounds the other. The
strict conditional holds because the consequent holds necessarily. However, we are in
a context where we allow metaphysically necessary truths to fail. Whatever structural
relations mathematical objects like n1 or n2 instantiate (i.e. whatever identities these
objects have), presumably they instantiate these relations necessarily. But we are
considering situations where we suppose these objects fail to instantiate some of
these relations. We could even consider a situation where such objects fail to exist
(in the structure), another metaphysically impossible situation. This can be done by
performing the graph-theoretic operation of removing a node, a perfectly legitimate
mathematical operation.
We are therefore in a context where we allow situations that might be considered
metaphysically impossible. However, these situations, given by the space of possi-
ble graphs, are governed by the axioms of the relevant graph theory. And so we
can make precise claims about what happens when we consider these situations.
Furthermore, by evaluating necessity claims in the space of possible graphs, we are
no longer threatened by counterexamples such as those involving the existence of
Socrates and the number 2. In the relevant situations, the number 2 does not exist
in every mathematical structure. So its existence will not trivially be grounded by
the existence of any object whatsoever. Socrates, on the other hand, doesn’t exist
in any mathematical structure. It may seem like the threat of triviality resurfaces
here. As Socrates doesn’t exist in any situation, his non-existence is grounded in
any fact whatsoever. But really this just shows that it is inappropriate to evaluate
grounding claims involving the existence of Socrates by appealing to what’s true in
john wigglesworth 
the space of possible graphs, something most people would agree with. The account
of ground given here is proposed only as an account of grounding in mathematical
structuralism, and so it is not threatened by these kinds of counterexamples. And as
this threat has been removed, there is then reason to think that the necessity claim
we have argued for entails, or at the very least gives evidence for, the corresponding
grounding claim ODO.
We note that the argument for ODO shows that the identities of any two objects
in a structure are grounded in one another. This follows from the fact that contains
facts about the identities of all objects in the relevant structure. This version of ODO is
consistent with other formulations of ODO in the literature (e.g. Linnebo 2008, p. 67).
A weaker version of ODO might claim that only the identities of connected objects
are grounded in one another, in which case a modified version of the argument is
required. However, if one accepts that the non-identity relation is a structural relation,
as discussed in Section 3, then all distinct objects in a structure are connected by
this relation.
Having argued that the identities of mathematical objects are grounded in the
identities of other objects in the same structure, we consider the second grounding
claim relevant to mathematical structuralism: that the identity of a mathematical
object is grounded in the identity of the mathematical structure it belongs to. To
evaluate this grounding claim, we need an account of the identity of a structure, just
as we needed an account of the identity of a mathematical object. This account will be
given by the identity conditions for graphs.
Recall that identity conditions for graphs are given in terms of isomorphisms. Based
on these identity conditions, a natural understanding of the identity of a graph is in
terms of its isomorphism class, the collection of graphs that are isomorphic to it. Let
G be the isomorphism class of the graph G. We then say that G has the identity G iff
G ∈ G. That is, we equate the identity of G with its isomorphism class G. As we are
taking graphs to be structures, this gives us an account of the identity of a structure.
And so we can formulate the relevant grounding claim.
ODS: For any mathematical object, n, in the structure G, the fact that G ∈ G fully
grounds the fact that the identity of n is En .
As before, this grounding claim is usually taken to entail a corresponding necessity
claim, where necessity ranges over the space of relevant possible graphs. In the
contrapositive form, this comes to: for all graphs G that are like G except where
the identity of n is not En , it follows that G ∈ / G. Here G is the isomorphism class
of the graph G that we started with (in effect, the “actual” graph).
Consider some arbitrary graph G just like G but which differs from G only in what
is required to make it that the set of edges that involve n is not En . If it follows that G ∈
/
G, then the relevant necessity claim is true, as the choice of G was arbitrary. Recall that
G is the isomorphism class of the graph G. And isomorphic graphs are identical. So the
isomorphism class of G is simply {G}. It follows that all we have to show is that G = G.
 grounding in mathematical structuralism
Let En be the set of edges in G that involve n, which is the result of adding or
removing at least one edge, e, to/from En . This edge must be between n and some
other node m. There are two cases. Either there is an edge between n and m that is
removed, so that n and m were connected in G but not connected in G . Or an edge is
added between n and m that was not there before, so that n and m were not connected
in G but are connected in G .
Recall the identity conditions for first-order objects, which hold for graphs as
well: x = y ↔ ∀X[X(x) ↔ X(y)]. We only need the left-to-right direction, or more
precisely, the contrapositive of the left-to-right direction. In either of the above
cases, there is a property that holds of G but fails to hold of G . That is, either
Connected(n, m, G) and ¬Connected(n, m, G ), or vice versa. So these graphs are not
identical. The necessity claim under consideration is therefore true. As before, this
gives us reason to think that the corresponding grounding claim holds as well.
5 Conclusion
We have argued for two grounding claims involving mathematical entities that are
relevant to the mathematical structuralist: that the identity of a mathematical object
is grounded in the identity of the structure it belongs to, and in the identities of other
mathematical objects in that structure. This argument has proceeded by describing
mathematical structures in terms of unlabelled graphs. With this account of structure
to hand, we present standard identity conditions for objects in a structure and for
structures themselves, which allow us to articulate the notion of the identity of a
mathematical entity in the context of structuralism. We then interpret grounding
claims involving these entities as claims about what happens in the space of possible
mathematical structures. This is an interpretation which makes no reference to any
particular systems or realizations that exemplify the structures in question. And so,
unlike Linnebo’s account, it is an account of grounding that is available to both the
ante rem and in re non-eliminativist structuralists. On this interpretation, we argue
that the grounding claims are true. Their truth follows from, or is at least evidenced
by, the truth of the relevant corresponding necessity claims, claims ranging over the
space of possible mathematical structures.
The notion of ground we appeal to is similar to what has been called the modal
account of ground, according to which the fact that A fully grounds the fact that
B iff necessarily, if A obtains then B obtains. This account of ground, as a general
analysis, is widely agreed to fail when the necessity involved is metaphysical necessity.
The failure results from counterexamples to the right-to-left direction, as discussed
in Section 4. But in certain contexts, this account can be successful, given that the
necessity operator is adequately interpreted for the context at hand. We argue that
mathematical structuralism is an appropriate context for understanding the relevant
grounding claims in terms of necessity claims that range over the space of possible
john wigglesworth 
graphs, because the kind of counterexamples that arise in the general case do not
arise here.
Given this understanding of the grounding relation in the context of mathematical
structuralism, we can evaluate the structural properties of this relation. It is transitive,
at least in the context of mathematical structuralism, as we have argued that each
object in a structure is grounded in every other object. But more interestingly, it
also allows for non-well-founded chains of partial ground. That is, there are chains
of partial ground that are not finitely grounded.
Consider the natural number structure. The arguments from Section 4 show that
the identity of each number is partially grounded by the identity of its predecessor
(for simplicity we denote the identity of a number by the standard numeral for that
number):
1 ≺ 2 ≺ 3 ≺ ...
This chain of partial ground tracks the usual ordering on the natural numbers and is
well-founded. But equally, the identity of each number is partially grounded in the
identity of its successor, because the identity of each number is partially grounded in
the identities of every other number in the structure. So we have:
... ≺ 3 ≺ 2 ≺ 1
This chain is clearly not finitely grounded, as it progresses infinitely to the left. These
examples also give us symmetric cases of ground, as we have that, for example, 1 ≺ 2
and 2 ≺ 1. These cases in themselves give rise to chains of partial ground that are not
finitely grounded:
... ≺ 1 ≺ 2 ≺ 1 ≺ 2...
The relation is also reflexive, generating further chains of partial ground that are not
finitely grounded:
... ≺ 1 ≺ 1 ≺ 1 ≺ 1...
But what is also interesting about these chains is that they are all bounded from
below. Recall, a chain is bounded from below when there is some fact F (not
necessarily in the chain) such that each fact in the chain is either partially grounded
by F or identical to F.
The chains in question are actually bounded from below in two ways. As the identity
of every number is partially grounded in the identity of every other number, each
number serves as a lower bound of partial ground for the chain.16 But there is also a
lower bound of full ground for the chain. This lower bound isn’t the identity of any
particular number in the structure, as the identity of each number only serves as a
partial ground for the others. But the identity of the structure itself is a full ground
for each member of the chain, and so it is a lower bound of full ground.
16 This result suggests that the accepted definition of having a lower bound, as given here, is inadequate,
or that an alternative definition is worth exploring in the examination of the well-foundedness of ground.
 grounding in mathematical structuralism
This latter kind of well-foundedness, being bounded from below with a lower
bound of full ground, will arguably hold for any structure, as the arguments in
Section 4 are widely generalizable and not restricted to any particular structure. And
so we have that the identity of each object in a structure is fully grounded in the
identity of the structure. Even if we have infinitely descending chains of ground
involving the identities of the objects within a structure, the identity of the structure
itself will serve as a lower bound for each chain. As each chain in the structure is
bounded from below, each chain has a foundation, a set of facts, S, such that each
fact in the chain is grounded by some fact in S. In the cases relevant to mathematical
structuralism, S will simply contain the single relevant fact concerning the identity of
the structure. So while mathematical structuralism gives rise to a relation of partial
ground that is non-well-founded in that it is not finitely grounded, the relation does
appear to be well-founded in two other important senses, that of being bounded from
below and having a foundation.17
References
Audi, P. (2012). Grounding: Toward a theory of the in-virtue-of relation. Journal of Philosophy,
109: 685–711.
Benacerraf, P. (1965). What numbers could not be. Philosophical Review, 74: 47–73.
Bliss, R. L. (2013). Viciousness and the structure of reality. Philosophical Studies, 166:
399–418.
Bliss, R. L. (2014). Viciousness and circles of ground. Metaphilosophy, 45: 245–56.
Burgess, J. (1999). Review of S. Shapiro (1997), Philosophy of Mathematics: Structure and
Ontology, Oxford University Press, Oxford. Notre Dame Journal of Formal Logic, 40:
283–91.
Cameron, R. (2008). Turtles all the way down: Regress, priority and fundamentality. The
Philosophical Quarterly, 58: 1–14.
Chalmers, D. J., Manley, D., and Wasserman, R., editors (2009). Metametaphysics: New Essays
on the Foundations of Ontology. Oxford University Press, Oxford.
Correia, F. and Schnieder, B., editors (2012). Metaphysical Grounding: Understanding the
Structure of Reality. Cambridge University Press, Cambridge.
Dasgupta, S. (2014). On the plurality of grounds. Philosophers’ Imprint, 14(20): 1–28.
Dixon, T. S. (2016). What is the well-foundedness of grounding? Mind, 125: 439–68.
Fine, K. (2001). The question of realism. Philosophers’ Imprint, 1: 1–30.
Fine, K. (2012). Guide to ground. In F. Correia and B. Schnieder, editors (2012), Metaphysical
Grounding: Understanding the Structure of Reality. Cambridge University Press, Cambridge,
pages 37–80.
Hale, R. and Hoffman, A., editors (2010). Modality: Metaphysics, Logic, and Epistemology.
Oxford University Press, Oxford.
17 The research leading to this paper was carried out within the research project “Mathematics:
Objectivity by representation”, which is funded by L’Agence Nationale de la Recherche (ANR) and the
Deutsche Forschungsgemeinschaft (DFG). I wish to thank Hannes Leitgeb and Graham Priest for helpful
discussions of this material and two anonymous referees for comments that helped to improve this paper.
john wigglesworth 
Harary, F. (1969). Graph Theory. Addison-Wesley, Reading, MA.

Hazen, A. P. (1985). Review of C. Wright (1983), Frege’s Conception of Numbers as Objects,
Aberdeen University Press, Aberdeen. Australasian Journal of Philosophy, 63: 250–4.
Hellman, G. (1989). Mathematics Without Numbers. Oxford: Clarendon Press.
Hellman, G. (2001). Three varieties of mathematical structuralism. Philosophia Mathematica,
9: 184–211.
Hoelte, M., Schnieder, B., and Steinberg, A., editors (2013). Varieties of Dependence: Ontological
Dependence, Grounding, Supervenience, Response-Dependence. Philosophia Verlag, Munich.
Jenkins, C. (2011). Is metaphysical dependence irreflexive? The Monist, 94: 267–76.
Keränen, J. (2001). The identity problem for realist structuralism. Philosophia Mathematica,
9: 308–30.
Keränen, J. (2006). The identity problem for realist structuralism II: A reply to Shapiro.
In F. MacBride, editor (2006), Identity and Modality. Clarendon Press, Oxford, pages
146–63.
Leitgeb, H. (2014). On mathematical structuralism. Paper presented at First Symposium on the
Foundations of Mathematics, Vienna.
Leitgeb, H. and Ladyman, J. (2008). Criteria of identity and structuralist ontology. Philosophia
Mathematica, 16: 388–96.
Leuenberger, S. (2014). Grounding and necessity. Inquiry, 57: 151–74.
Linnebo, Ø. (2008). Structuralism and the notion of dependence. Philosophical Quarterly, 58:
59–79.
Litland, J. E. (2013). On some counterexamples to the transitivity of grounding. Essays in
Philosophy, 14: 19–32.
MacBride, F. (2006). What constitutes the numerical diversity of mathematical objects? Ana-
lysis, 66: 63–9.
Rabin, G. O. and Rabern, B. (2016). Well founding grounding grounding. Journal of Philosoph-
ical Logic, 45(4): 349–79.
Resnik, M. D. (1982). Mathematics as a science of patterns: Epistemology. Noûs, 16: 95–105.
Rosen, G. (2010). Metaphysical dependence: Grounding and reduction. In R. Hale and A.
Hoffman, editors (2010), Modality: Metaphysics, Logic, and Epistemology. Oxford University
Press, Oxford, pages 109–36.
Schaffer, J. (2009). On what grounds what. In Chalmers et al., editors (2009), Metameta-
physics: New Essays on the Foundations of Ontology. Oxford University Press, Oxford, pages
347–83.
Schaffer, J. (2012). Grounding, transitivty, and contrastivity. In F. Correia and B. Schnieder,
editors (2012), Metaphysical Grounding: Understanding the Structure of Reality. Cambridge
University Press, Cambridge, pages 122–38.
Schnieder, B. (2011). A logic for “because”. Review of Symbolic Logic, 4: 445–65.
Shapiro, S. (1997). Philosophy of Mathematics: Structure and Ontology. Oxford University Press,
Oxford.
Shapiro, S. (2000). Thinking About Mathematics: The Philosophy of Mathematics. Oxford: Oxford
University Press.
Shapiro, S. (2006). Structure and identity. In F. MacBride, editor (2006), Identity and Modality.
Clarendon Press, Oxford, pages 109–45.
Shapiro, S. (2008). Identity, indiscernibility, and ante rem structuralism: The tale of i and −i.
Philosophia Mathematica, 16: 285–309.
 grounding in mathematical structuralism
Skiles, A. (2015). Against grounding necessitarianism. Erkenntnis, 80: 717–51.

Trogdon, K. (2013). An introduction to grounding. In M. Hoelte et al., editors (2013), Varieties
of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence.
Philosophia Verlag, Munich, pages 97–122.
Wright, C. (1983). Frege’s Conception of Numbers as Objects. Aberdeen University Press,
Aberdeen.
12
Fundamentality and Ontological
Minimality
Tuomas E. Tahko
1 Introduction
This chapter deals with the idea that reality comes with a hierarchical structure of
‘levels’.1 The idea has a long history and it remains popular, despite some recent
challenges. Our everyday experiences as well as scientific practice seem to, prima
facie, strongly support such a view, since the sciences are naturally interpreted as
operating at different levels, different scales.2 Usually, the reference to scale becomes
apparent when talking about parts and wholes, which are studied in mereology: we
talk about subatomic particles constituting atoms, atoms constituting molecules, and
molecules constituting everything we see around us. A typical view is that composite
objects depend for their existence on their parts—at least in the sense of having
some parts. Fundamentality comes in when we ask whether there is an end or a
beginning to this hierarchical structure, the relevant chain of dependence. That is,
is there a fundamental level or does the hierarchical structure of reality continue ad
infinitum?
The received view has long been that there indeed is a fundamental level. The
fundamental level is usually thought to be at the smaller end of the spectrum:
mereological atomism suggests that certain indivisible ‘atoms’, such as subatomic
particles, are fundamental. But this does not mean that the fundamental level must
necessarily be at the bottom of the scale—the fundamental end could also be at the
top, that is, the universe as a whole could be considered fundamental. Moreover, this
conception of fundamentality associates it very closely with some sort of ontological
1 Some of the material presented here is based on the discussion in Tahko 2015a, Ch. 6.
2 I do not wish to put much weight on the notion of a ‘level’, the idea is that the ‘levels’ reflect an order
of priority.
 fundamentality and ontological minimality
independence—the mereological atoms are fundamental in the sense that everything

else depends on them, is grounded in them, or even reducible to them. But there
is another traditional way of understanding fundamentality, or the intuition that
drives it: it’s the world’s intrinsic structure that is fundamental (see Fine 2001, 26).
Now, it may not be immediately clear what this is supposed to amount to, but here’s
another way of putting the idea: whatever God had to bring about to make the
world is what counts as fundamental; the fundamental entities are the basic building
blocks of reality, like axioms in a theory (see Wilson 2014, 560 and Wilson 2016).
It is this conception of fundamentality—attributable at least to Kit Fine and Jessica
Wilson—which I find worth exploring in more detail, and indeed to hold much more
promise when attempting to develop a general theory of fundamentality.3 On this
view, fundamentality itself is primitive and not something that is fixed by the relevant
chain of dependence. That is, we must first fix the fundamental level and only then do
we know what the correct direction of priority is. But before we examine this view of
fundamentality in more detail, we will first have to wade through the complexities of
the more mereologically inclined approach, which is perhaps more familiar.
Recent work by scientifically minded philosophers, taking heed of contemporary
physics, has gone some way towards refuting the idea of levels altogether (e.g. Lady-
man and Ross 2007, 4, 53–7, 178–80). Without the hierarchical view introduced by
the levels metaphor, talk about a fundamental level does not seem possible. So, before
we can get started, we must clarify several issues, such as the idea of a hierarchical
structure itself. Further complications are introduced when we try to make sense of
the possibility that there could be a hierarchy and no fundamental level. An important
question here is whether it is possible that the levels not only go on ad infinitum, but
are also infinitely complex.
Consider Figure 12.1 (which demonstrates a mereological hierarchy). The cone
represents the hierarchical structure of reality, with the smallest thing(s) at the bottom
and the universe as a whole at the top. Why a cone? Because it reflects the idea that
there is a scale to the structure of reality. Another reason: it seems natural to think
that there are fewer different kinds of things at the smaller end of the spectrum. The
Standard Model of particle physics postulates that there are sixty-one elementary or
fundamental particles (if we count particles and their corresponding antiparticles as
well as the various colour states of quarks and gluons)—everything else is composed
of these fundamental things. In contrast, the larger end of the spectrum may be
considered to encompass all the different kinds of things there are in reality—perhaps
even infinitely many different kinds. It should be noted though that putting this in
3 For further discussion, see Bennett 2017, Ch. 5, where different senses of absolute fundamentality are
discussed. I got my hands on the book too late to discuss it here in detail, but the view that I support takes
fundamentality as a primitive. Bennett as well associates this view with Fine and Wilson.
tuomas e. tahko 
JUNK?
GUNK?
Figure 12.1.
terms of kinds is entirely optional and the resulting picture will of course depend
on one’s views regarding kinds; one option might even be to consider the cosmos
as a whole as just one kind, in which case the previous description of fewer kinds
at the smaller end would not be accurate. Leaving that aside, the lines at each end
of the cone represent the hypothetical bottom and top levels. The dotted sections
represent the possibility of infinite descent (gunk) and infinite ascent (junk), beyond
the hypothetical bottom and top levels.
Four options immediately arise regarding the hierarchical structure of reality:
(1) Closed (i.e. the chain of dependence terminates) at both ends of the cone.
(2) Open at the top, but not at the bottom.
(3) Open at the bottom, but not at the top.
(4) Open at both ends.
If both ends of the cone are closed, then we might consider either one of them as
fundamental, fixing the direction of the relevant dependence relation accordingly. If
only one end is closed, then it seems, prima facie, more plausible that the closed end
is fundamental, whether it is the top or the bottom. The first three options enable
different varieties of metaphysical foundationalism, which, in its broadest sense, states
simply that there is a fundamental level. The fourth option, however, cannot support
metaphysical foundationalism, as there is neither a top nor a bottom level. In this case,
only some sort of metaphysical infinitism would seem to be available, regardless of the
direction of the relevant dependence relation. To give names to the different types of
metaphysical infinitism at play, let’s start with (3) (assuming that the order of priority
is from top to bottom), which is best known: an ontology embracing (3) is typically
 fundamentality and ontological minimality
called ‘gunky’, where ‘gunk’ is a term familiar from mereology, referring to the idea
that all wholes have further proper parts, ad infinitum (Lewis 1991, 20). Similarly, an
ontology embracing (2) is ‘junky’, where ‘junk’ is the converse of gunk, i.e., everything
is a proper part of something (Schaffer 2010, 64). Note though that, strictly speaking,
gunk does not rule out junk, nor does junk rule out gunk, contrary to (2) and (3). This
brings us to (4), which indeed combines gunk and junk, resulting in what has been
labelled a ‘hunky’ ontology (Bohn 2009, 193).
These are the basic options, but this description is somewhat simplified, as we
have not yet said anything about what ‘fundamentality’ amounts to. In particular,
we should clarify the role of mereology in discussions of fundamentality, which we’ll
do in Section 2. In Section 3, a common objection to the possibility of metaphysical
infinitism is outlined, with some critical remarks. In Section 4, a more general sense of
fundamentality will be explicated with the help of the idea of ontological minimality—
this more general approach to fundamentality considers it as an ontological minimality
thesis. Section 5 examines the tension between the mereological, object-oriented
ontology and structuralism. It is suggested that fundamentality understood as an
ontological minimality thesis can accommodate both. Finally, in Section 6, we will
briefly consider whether fundamentality understood as an ontological minimality
thesis rules out metaphysical infinitism.
2 Mereological Fundamentality
The options outlined above were presented in mereological terms: the cone in Figure
12.1 is taken to represent a mereological hierarchy, where mereological complexes
are at the top and (supposed) mereological atoms at the bottom. The ‘levels’ of reality
according to this picture are thus mereological levels. This brings us to Mereolog-
ical Fundamentality, which may be described as a combination of the following
two theses:
Mereological Hierarchy (M): The world is organized into mereological levels,

resulting in a hierarchical structure governed by mereological dependence
relations.4
Fundamentality Thesis (F): There is a fundamental ‘level’ at one end of the hierar-
chical structure of reality, that is, the relevant chain of dependence is closed at least
at one end. Together, (M) and (F) entail (simplified) Mereological Fundamentality.
Mereological Fundamentality (MF): The world is organized into mereological
levels and there is a fundamental, mereologically independent level which is at one
end of the mereological scale.
4 I take the notion of ‘mereological dependence’ from Kim, which he prefers to the term ‘mereological
supervenience’ and contrasts with causal dependence: ‘the properties of a whole, or the fact that a whole
instantiates a certain property, may depend on the properties and relations had by its parts’ (2010, 183).
tuomas e. tahko 
The idea, according to (MF), is that fundamentality is a thesis about mereological

dependence, or to put it another way: (absolute) fundamentality is mereological
independence. On this view, the type of dependence that we track with the notion
of (relative) fundamentality is mereological dependence. On the face of it, (MF) is
compatible with three of the options listed in the previous section: (1), (2), and (3).
So, the only option ruled out is (4), where both ends of the hierarchical structure are
open. (MF) comes in two primary forms, depending on which end of the mereological
scale is considered fundamental. An additional commitment that those who support
(MF) may have is that the entities at the fundamental level have the highest degree of
‘reality’, that is, they are the most fundamental. But we need not dwell on the idea
of degrees of reality or fundamentality here.5 Metaphors abound, but to put it in
terms of grounding, x is fundamental or ontologically independent in the relevant
sense if and only if nothing grounds x (Schaffer 2009, 373). Note that this is of course
a much more general thesis than (MF), which ties fundamentality specifically with
mereological dependence; grounding may involve other types of dependence or it can
be considered primitive.6 But again, the view I favour is that fundamentality itself is
primitive and should not be defined in terms of grounding or any other dependence
relation. In what follows I will avoid the grounding terminology and will attempt to
find a somewhat more neutral way to explicate (but not define!) fundamentality.
Given (M), (MF) is naturally combined with the idea that the mereological depen-
dence relation is asymmetric. The direction of the asymmetric dependence is fixed
by which end of the hierarchy is considered fundamental. But this is a question
that must be settled separately. What determines the direction of the dependence
relation is the distinction between pluralism and monism.7 The pluralists hold that the
direction of the dependence is from the larger to the smaller, resulting in mereological
atomism—this is, perhaps, the standard view.8 The monists hold that the parts are
dependent on the whole and hence that there is only one fundamental entity, namely,
the universe. However, pluralism and monism are independent of (MF), which does
not by itself fix the direction of priority. So, different combinations of these views may
be possible. One proponent of the monistic view, Schaffer, considers only substances
to be fundamental, and further, that there is exactly one substance, namely the
cosmos. On this understanding, a substance is a basic or fundamental, ontologically
independent entity. This is roughly how Aristotle might characterize the notion of
‘substance’. So, for Schaffer, it is the cosmos, conceived as a substance, which is prior
to its (arbitrary) parts. Schaffer’s Spinoza-inspired priority monism, however, does not
5 The idea can be understood in many ways. For discussion, see e.g. McDaniel 2013.
6 For an introduction to grounding, see, for instance, Bliss and Trogdon 2014. Moreover, another
possible view here is that fundamentality tracks (and can be defined in terms of) what Bennett (2017)
calls ‘building’, which also encompasses several dependence relations, including grounding.
7 For further discussion on the monism/pluralism issue as well as an alternative account of fundamen-
tality, see Trogdon 2009.

8 Cotnoir 2013 labels this as the ‘default’ view in a recent note.
 fundamentality and ontological minimality
assume mereological atomism, even though he operates in mereological terms. In fact,

one of Schaffer’s arguments in favour of priority monism draws on its compatibility
with the possibility of gunk (2010, 61ff.).9 The question of gunk is of some importance
here, as those committed to (MF) and the idea that the dependence relation goes
from the larger to the smaller (pluralism) will struggle to accommodate the infinite
divisibility of matter entailed by a gunky ontology—we will return to this shortly.
(MF) combined with mereological atomism may once have been the default view
when it comes to fundamentality, but its popularity has dwindled. For instance,
Ladyman and Ross (2007) strongly oppose the view; so does Markosian (2005). Most
opponents of (MF) are worried about the (additional but common) commitment to
mereological atomism. The worry that Ladyman and Ross have regarding this picture
is that none of the various atomistic conceptions can stand the test of contemporary
physics—but note that this is strictly speaking a case against (M) rather than (MF).10
The core of the critique by Ladyman and Ross is that there is nothing in fundamental
physics that corresponds with the atomistic idea of ‘simples’. Note, however, that even
if their critique is correct, it only applies to (MF) combined with pluralism, which
together entail a commitment to mereological atomism. Even though it won’t get
around the problems associated with the idea of mereological hierarchy, one could
be tempted to defend (MF) but adopt (priority) monism when faced with the line
of criticism from Ladyman and Ross, hence taking the fundamental level to be at
the top end of the cone in Figure 12.1. This way one could endorse the possibility
of a gunky ontology without losing (MF). In any case, pluralism still enjoys wider
support, at least partly because it seems to be a promising way to articulate (MF). Yet,
there is a wide agreement about the incompatibility of (MF) + pluralism + gunk. This
has, naturally enough, resulted in some hostility towards gunk, given the popularity
of both (MF) and pluralism. But if we are happy to leave the door open to some
form of metaphysical infinitism (and hence gunk) then we could instead drop (MF)
while keeping pluralism. The question that remains is: how strong is the case against
gunk/metaphysical infinitism and in favour of (MF)?
3 The Problem with Metaphysical Infinitism

There is a well-known paper by Cameron (2008), which summarizes the basic
argument in favour of (MF)—or against metaphysical infinitism.11 Cameron resists
the idea that there could be infinite chains of ontological dependence because of the
9 This might be a good place to note that Schaffer may have changed his views after the 2003 paper,
where fundamentality is discussed in mereological terms—he may now prefer to define fundamentality in
terms of grounding.
10 The phenomenon they appeal to is famous: ‘quantum entanglement’. However, it is not at all clear that
the quantum entanglement objection is successful. For discussion of this issue, see Dorato and Morganti
2013, and O’Conaill 2014.
11 See also Paseau 2010.
tuomas e. tahko 
unintuitive implications that such chains would entail. Applying the idea to the case
of composition, Cameron notes that in a gunky world, composition would never
‘get off the ground’. In other words, if complex objects are ontologically dependent
on their mereological parts, then composition never bottoms out in gunky worlds.
The intuitive result, according to Cameron, is that complex objects are not possible
in gunky worlds—and this is a strong result! It rules out the possibility that we live
in a gunky world, at least if we accept the plausible idea that complex objects are
ontologically dependent on their parts. But how strong is the intuitive appeal that
this argument rests on? This question is pressing, given that Cameron and many
others writing about fundamentality and the idea that composition couldn’t ‘get off
the ground’ without a fundamental level, typically acknowledge that this appeal to
intuition hardly constitutes a proper argument in favour of fundamentality.
One complication concerns the way in which we think about the structure of space–
time. If space–time itself is made up of zero-dimensional space–time points and hence
has no internal structure, no smaller parts, then this ‘pointy’ space–time would seem
to rule out a certain type of infinite regress by its very nature: surely those space–
time points themselves do not depend for their existence on anything and hence are,
in one sense, fundamental. It’s standard to think that any given collection of space–
time points composes a space–time region. These regions are identified in terms of
their composite space–time points. Yet, there are two different ways to conceive of
the relationship between the points and the regions: we can either think of points as
derivative and facts about extended objects as more fundamental or we can hold that
the space–time points and facts about their arrangement are more fundamental than
space–time regions (see Hawthorne 2008, 264). The latter is perhaps the typical way of
looking at things nowadays, but this question regarding the relative fundamentality
of the regions and the points would seem to precede Cameron’s case. The relevant
intuition that Cameron appeals to would thus seem to presuppose the view that points
are more fundamental.
If I may, I’d also like to make an anecdotal point. Having presented versions of this
paper and the related arguments to various audiences around the world, the most
common reaction seems to be that people do not find themselves drawn towards
the foundationalist intuition that Cameron and various others have attempted to
defend. Now, this may be because I have not done justice to the idea or presented the
argument clearly enough. But the challenge here is that no systematic argument has
been put forward in the first place—I take it that Cameron is mainly just reporting the
fundamentality intuition that he thinks many metaphysicians share. My impression is,
however, that the intuition is currently not that widely shared. There are good reasons
for this. One of the most important of them is the worry pushed by Ladyman and Ross,
namely the strong association with mereological atomism, which does not fare so well
in the light of contemporary physics. Yet, personally I thought that I do have at least
something very much like the fundamentality intuition. But I am now less certain that
it’s the same fundamentality intuition that Cameron has in mind. It may be closer to
 fundamentality and ontological minimality
what Wilson (2014, 560) has in mind when she compares entities in the fundamental
base to the axioms of a theory. Importantly, the understanding of fundamentality
that I find appealing is not accurately captured by (MF), because I wish to drop the
mereological connotations. It is, however, captured by the Fundamentality Thesis
(F). It’s crucial to recognize that (F) does not specify which type of dependence is
at issue when it comes to the hierarchical structure. All it requires is that there is a
hierarchical structure governed by (presumably) an asymmetric dependence relation
that terminates at least at one end. This is still a thesis that would require an argument
in its support, but it is not immediately subject to the same counterarguments that
(MF) is. However, on its own, (F) is really not fine-grained enough to constitute
an interesting metaphysical thesis at all. The question that we now face is whether
anything more can be said about (F) without a commitment to (M), hence leading
back to (MF).
4 Fundamentality and Ontological Minimality

I think we should resist the idea that the appeal of metaphysical foundationalism rests
solely on (MF), since the problems associated with mereological atomism have been
known for some time. Moreover, even though I have not discussed the grounding-
based approach to fundamentality here, I don’t think that it does much better (see
Wilson 2014 for criticism). However, it’s clear that (F) on its own is not going to
do the trick either.12 Let’s see if something more can be said about it. Rather than
attempting to list all the possible ways to understand levels and the relevant chains
of dependence that structure them, I will here try to explicate (but not define) a
more general sense of fundamentality, which could capture (MF) but also many, if
not all, other conceptions of fundamentality. To do so, recall that we distinguished
the Fundamentality Thesis (F) from the idea of Mereological Hierarchy (M). The first
is neutral about the sense of ontological dependence that does the structuring. But if
(M) is already too restrictive about the relevant sense of dependence, then we better
adopt the broadest possible sense of ontological dependence at the outset.13 Even this
is not enough though: it must be recognized that what is fundamental and hence what
fixes the direction of priority is independent of (M) and (F). If one accepts (M) +
pluralism, the fundamental level consists of mereological atoms. But if we drop (M),
the nature of the fundamentalia—the elements that are fundamental—is unrestricted.
This is where the notion of ontological minimality finally comes in. With the help
12 Some recent notable discussions which introduce further complications (being favourable to meta-
physical infinitism) are Bliss 2013, 2014 and Morganti 2014, 2015. The alternative formulation that I am
sketching has its origins at least in Fine 2001 and Wilson 2014—although I don’t mean to suggest that
there are no differences between these conceptions. New work on metaphysical foundationalism is also
constantly in the pipeline; see, for instance, Bennett 2017, Dixon 2016, and Raven 2016.
13 For an introduction to the various senses of ontological dependence, see Tahko and Lowe 2015.
tuomas e. tahko 
of ontological minimality, we can explicate a general conception of fundamentality

which is neutral regarding the nature of the fundamentalia:
Generic Ontological Fundamentality (GOF): The world is organized into ‘levels’ of

ontological elements and the fundamental ‘level’ consists of ontologically minimal
elements.
This understanding of fundamentality is only provisional and it is obviously in need
of further explication, as nothing has yet been said about ‘ontological elements’ in
general and ‘ontologically minimal elements’ in particular. Moreover, I don’t wish to
put too much weight on the notion of ‘level’ used here—it could be regarded merely
as a placeholder for whatever notion one wishes to use to parse the hierarchical
structure of reality. In fact, even the idea of a hierarchy could be dropped here, as
there could be just one ‘level’, a flat world. But I will continue to speak as if there are
several levels. Since (GOF) is supposed to be able to express (MF) as well, we can
already identify at least one thing that could take the place of ‘ontological elements’
in (GOF), namely, mereological elements. In that case we would get the familiar
hierarchy of mereological levels and the ontologically minimal elements would be
mereological atoms—if (MF) is combined with pluralism. But what else could the
ontological elements be? That would seem to depend on three things. Firstly, what
kind of hierarchy are we interested in? Secondly, what kind of dependence structures
that hierarchy? Thirdly, how is the hierarchy manifested in the world, that is, what is
the relevant science? These are difficult questions, especially the last one. If the critical
remarks by Ladyman and Ross that we discussed above are correct, then much of the
relevant work in determining the correct answers to these three questions will come
from science. It may also be worth noting that (GOF) is neutral between pluralism
and monism. It can certainly be combined with pluralism, for there could be multiple
ontologically minimal elements (e.g. mereological atomism). But it could feasibly also
be combined with monism, for it could turn out that there is only one ontologically
minimal element (e.g. the universe as a whole, like in Schaffer’s priority monism).
However, even if we will eventually need input from science, this does not mean
that we can’t say anything more at the moment. Further specification of ontologically
minimal elements—the theoretical background of ‘ontological minimality’—is possi-
ble.14 To get the idea started, ontologically minimal elements could be compared to
‘minimal truthmakers’, familiar from Armstrong (see also O’Conaill and Tahko 2016):
If T is a minimal truthmaker for p, then you cannot subtract anything from T and the remainder
still be a truthmaker for p. (Armstrong 2004, 19–20)
14 Compare the idea also with one of Schaffer’s (2004) criteria for sparsity, namely ‘minimality’, according
to which sparse properties serve as the minimal ontological base, i.e. the fundamental properties that
macro-properties supervene on (assuming that they do supervene!) form the minimal ontological base.
Schaffer himself argues against this criterion with appeal to the possibility of gunk, but as we will see, gunk
may not be a similar problem for the notion of ontological minimality introduced here.
 fundamentality and ontological minimality
A minimal truthmaker for a given proposition can be understood as the smallest or

least encompassing portion of reality that fully grounds the truth of that proposition.
This analogy between minimal truthmakers and ontologically minimal elements
suggests that the fundamental level consists of the least encompassing portion of
reality. In usual accounts of minimal truthmakers, the least encompassing portion
of reality is understood in terms of parthood, which would naturally take us back
towards (MF) rather than (GOF). However, it may be possible to avoid the mereolog-
ical connotations by invoking the more general notions of composition and of being
integral as intended by Kit Fine:
When one object is a part of another, there is a sense in which it is in the other—not in the sense
of being enclosed by the other, as when a marble is in an urn, but more in the sense of being
integral to the other. When parts are in question, it is also appropriate to talk of a given object
being composed of or built up from the objects that it contains. (Fine 2010, 560)
This more general notion of ‘part’ may allow one to avoid the unintuitive conse-
quences of infinite chaining of parts more narrowly understood, that is, construed
according to classical mereology. On this view, all manner of things, from sentences
to symphonies to sets, can be composed of other things. It seems plausible that the
components of a composite object one level down are typically integral to the object
in the sense suggested by Fine. But it is less clear that this has to be the case when
we proceed further down. Fine suggests that a person would somehow change if her
kidney was replaced with a new one, and perhaps it’s plausible that a kidney changes
when some of its cells die and are replaced by new ones—it certainly does if some
of these cells become cancerous. But given that our cells die and are replaced by new
ones all the time, it doesn’t seem that any individual cell is really integral to us. I think
that this is the case regardless of the fact that the cells may be integral to whatever
organ they are a part of or whatever biological function they participate in. But if it is
possible that the components several levels down are not integral to the object at the
higher level, then the infinite regress causing the typical unintuitive consequences of
infinite chaining (transitivity) wouldn’t follow.
The more general line of thought here is that the liberal sense of parthood is not
obviously transitive—this may have further consequences regarding what counts as a
minimal truthmaker on one hand and an ontologically minimal element on the other.
So there could, perhaps, be an understanding of ontological minimality whereby
we should only be concerned with the parts that are integral to an object, even if
these parts are composed of further parts. This would be in clear violation of (MF),
of course, for the fundamental level might not be the mereologically fundamental
level. But note that arguments from gunk (or junk) would no longer apply, since
infinite mereological descent would not necessarily violate ontological minimality.
Moreover, on the face of it this would seem to imply that not everything is grounded
in the fundamental, because there could be cases where we do not need to follow the
grounding chain beyond the immediate ground, to use Fine’s (2012, 50) terminology.
tuomas e. tahko 
The discussion surrounding the transitivity of grounding more generally is beyond

the scope of this paper though.15 The same goes for an analysis of the possible
compatibility between gunk and ontological minimality (although we will return to
this in Section 6). But let’s consider Fine’s discussion of transitivity briefly.
Consider sets and their members. On a traditional conception of parthood, sug-
gesting that the members of sets are parts of sets is problematic: since the parthood
relation is traditionally conceived as transitive, but the relation of member to set is not
transitive, the members of a set cannot be its parts. But according to Fine, this confuses
two claims: that the relation a member of a set bears to the set is the parthood relation,
and that the members of a set are its components. We can consistently deny the first
of these and accept the second. This suggests that the transitivity of parthood, when
parthood is understood in the more liberal sense that Fine suggests, is a question that
must be settled separately, that is, it is not built in to the notion itself:
[I]f part is understood as component, it will be a substantive question whether or not the
relation is transitive. In the case of sums, for example, it will be transitive, since components
of components are components, while in the case of sets, it will fail to be transitive, since com-
ponents of components (that is, members of members) will sometimes fail to be components
(members). (Fine 2010, 569)
Of course, even if Fine is correct about the case of sets and their members, this
does not yet show that the relevant dependence relation at work in the hierarchical
structure of reality is not transitive. It merely shows that there is at least one conception
of parthood which may not be transitive. In any event, this opens the possibility
of interpreting ontologically minimal elements quite liberally indeed: the smallest,
minimal ‘parts’ of reality do not need to be mereological elements at all, they can be
anything that count as components, such as structures, relations, objects, or whatever.
This may enable us to overcome at least some of the complications regarding the
tension between a classic, object-oriented metaphysics such as mereological atomism
and the type of view defended for instance by Ladyman and Ross, according to which
reality is fundamentally relational or structural. We turn to this topic in Section 5.
5 Fundamentality, Structuralism, and Ontological

Minimality
Perhaps the primary reason to search for a general notion of fundamentality such
as (GOF) is exactly because ontic structural realism (OSR) familiar from the work
of Ladyman and Ross (see also Esfeld and Lam 2010, French 2014, and others)
consider (MF) to be based on a completely skewed sense of priority: it’s not objects
that are fundamental, but structure or relations. The question is, could these two camps
nevertheless agree on the relevant notion of fundamentality? Kerry McKenzie (2014a)
15
For critical discussion regarding grounding and transitivity, see Schaffer 2012 and Tahko 2013.
 fundamentality and ontological minimality
has done some important work in order to clarify the issue; she distinguishes the
extreme version of (OSR) defended, for example, by Ladyman and French, and a more
moderate version defended, for example, by Esfeld and Lam. The first takes structure
as prior to objects, whereas the latter eschews talk about priority between structure
and objects altogether (because they are on a par, ontologically speaking) (McKenzie
2014a, 355–6). McKenzie spends some time arguing that instead of supervenience, we
should focus on ontological dependence in articulating the relevant sense of priority.
I do not consider this particularly helpful, as supervenience as well is very naturally
considered as a species of ontological dependence. That said, McKenzie is right about
the importance of finding the relevant sense of dependence and it’s clear that, given
that we’re looking for a conception of fundamentality compatible with (OSR), this
can’t be mereological dependence, which is naturally associated with Mereological
Fundamentality (MF). According to McKenzie’s analysis, we would do well to adopt
Fine’s (1995) essentialist notion of dependence and clarify (OSR)’s understanding
of priority in these terms. McKenzie focuses on what we may call generic essential
existential dependence (GEED), defined as follows:
(GEED) x depends essentially for its existence upon Fs =df . It is part of the
essence of x that x exists only if something is an F.16
Applying (GEED) to (OSR), McKenzie observes that the extreme version of (OSR)
as developed by Ladyman and French turns out to be in trouble because although the
dependence of objects on structure can be derived, so can the dependence of structure
on objects. Hence, McKenzie takes the analysis to favour the moderate version of
(OSR) as developed by Esfeld and Lam instead. But it should be noted that all this rides
on the relevant understanding of what is considered essential. The relevant essentialist
content on McKenzie’s analysis concerns the nature of relations—whether we conceive
of them extensionally or intensionally, and it is on the extensional conception that
the dependence between objects and structure emerges as symmetric, whereas the
intensional conception is considered inferior for this purpose. The further details of
McKenzie’s account are not crucial for our purposes, but what is important is that
(GEED) is indeed able to produce the relevant dependence structure. Note that on this
view what is fundamental is not strictly speaking primitive, but it is rather determined
by what is considered essential—(GEED) does not presuppose an object-oriented
metaphysics nor the extreme or the moderate version of (OSR). So, here we have a
sense of dependence that is fairly neutral, albeit there are no doubt those who would
take issue with the fact that (GEED) relies on the notion of essence.17
16 Compare with Tahko and Lowe 2015, section 4.3. McKenzie uses a formal definition due to Fine and
Correia rather than (GEED), but the content is the same.

17 However, this may not be a huge problem, since (GEED) entails the modal-existential notion of
generic existential dependence, which could perhaps be adopted as well: (EDG) x depends generically for
its existence upon Fs =df . Necessarily, x exists only if some F exists (Tahko and Lowe 2015, section 2).
In fact, (EDG) is more straightforwardly compatible with symmetric dependence anyway, since there is
tuomas e. tahko 
We clearly need our notion of ontological minimality to be quite flexible indeed,

as evidenced by the need to apply it to (OSR) as well. It might be useful to introduce
another example, one without the complications of contemporary physics. One reason
for this is that I’d like to better demonstrate the idea that there can be several ontologic-
ally minimal descriptions and it may not always be quite straightforward to compare
them—the choice between mereological atomism and (OSR) is plausibly decided
based on other criteria than considerations having to do with ontological minimality.
But here’s a case where the choice is less clear: consider two different ways to describe
the colour space: RGB vs YUV.18
The RGB (red, green, blue) system is perhaps the more familiar system, but YUV
is useful in some cases, as it takes into account human perception of colour, which
depends on lighting conditions. Hence, the Y stands for luminance (brightness)
and U and V are the chrominance (colour) components; there are also other lumi-
nance/chrominance systems. Although there are differences between the systems,
both RGB and YUV can give a complete description of a given colour value and it’s
also possible to convert an RGB value to a YUV value. But what’s the connection
to ontological minimality? Well, since both systems can completely describe a given
colour C and may be considered to consist of three minimal elements (at least
if we simplify a little), then each description is a minimal description of C. This
should make clear that, just as with minimal truthmakers, there could be several
mutually consistent ontologically minimal descriptions of reality. Of course, colour is
unlikely to feature in the ontologically minimal description of reality, so we should
not be misled by this example. But there is a clear need to illustrate the idea of
ontological minimality with examples such as this, because it seems quite likely that
any description of reality we might give based on current science will be incomplete
and hence not genuinely ontologically minimal. The point is that we must resort to
other criteria to assess two competing physical theories if they both give a complete,
ontologically minimal description of reality. However, unless the theories differ with
regard to some further theoretical virtues that we consider important, the choice
appears to be arbitrary. There may, of course, be practical or contextual reasons to
prefer one system over the other, quite like in the case of RGB vs YUV. This is just
to say that there could be several ontologically minimal descriptions of reality, based
on different sets of ontologically minimal elements. But if one description needs to
no requirement of building the existential dependence into the essences of the involved entities. All we
need here is a way to model the symmetric dependence between objects and structure. Note also that if
one wishes to hold fundamentality as primitive rather than fixed by what is essential, an even more neutral
sense of dependence may be needed. I will leave this issue aside here, since the potential connection between
essentiality and fundamentality is a topic that deserves a paper of its own.
18 I owe this example to Tim Button and Hugh Mellor. Note that I do not mean to suggest that colour is
a fundamental feature of reality, or indeed to put much weight on the example—the example is primarily
heuristic, to get a better grasp on the abstract idea of ontological minimality.
 fundamentality and ontological minimality
postulate more ontologically minimal elements than the other one to reach a complete
description, then the first is clearly not an ontologically minimal description at all.19
Finally, recall that on one interpretation the cone in Figure 12.1, in the beginning of
this paper, could be considered to represent different kinds of things, where ‘kinds’ are
considered as natural kinds. If we apply the idea of ontological minimality here, what
we get is something like the following: an ontologically minimal description identifies
all and only the most fundamental natural kinds. These natural kinds could be kinds
of fundamental particle, like those listed in the Standard Model, or they could be the
structures identified by (OSR). Moreover, they could perhaps even be symmetries,
as the idea that symmetries are fundamental is now emerging as a candidate view,
supported by physics (cf. McKenzie 2014b). The association with fundamental natural
kinds is in fact quite an interesting understanding of the ontological minimality thesis,
but since specifying this option would require much more detail about the nature of
natural kinds, I will not be relying on this reading here (but see Tahko 2015b).
Summarizing, it appears that much remains to be done in order to specify the
range of options with regard to fundamentality. The generic notion (GOF) sketched
here is designed to be as liberal as possible, for the various possible conceptions
of fundamentality to be captured with one general notion. For this purpose, one
possible sense of dependence that could do the trick has been identified as (GEED),
which McKenzie has already applied to (OSR). The upshot is that understanding
fundamentality as an ontological minimality thesis does seem to hold at least some
promise.
6 Ontological Minimality and Metaphysical Infinitism

We will end with a more radical note, for I propose that even some versions of
metaphysical infinitism may satisfy the ontological minimality thesis and hence
Generic Ontological Fundamentality (GOF).20 I have in mind what Schaffer (2003,
510) calls ‘infinitely boring descent’, whereby the same structure (pattern, description)
repeats infinitely. To distinguish this from metaphysical infinitism proper, consider
a scenario where reality is structured mereologically and each time we split a sup-
posedly fundamental particle, a set of different kinds of particle emerge. This would
constitute infinite complexity, which would not appear to be reconcilable with (GOF).
In contrast, a ‘boring’ or repetitive descent may allow for an ontologically minimal
description in the sense that a description of the repetitive part only needs to be
supplemented with an instruction to continue as before; for example, ‘the world stands
on four elephants, the four elephants stand on a turtle, the turtle stands on two camels,
19 Note that other considerations could even trump the minimality constraint, which is a purely
qualitative notion. Balancing between different theoretical virtues is notoriously difficult; see e.g. Nolan
1997 for discussion about qualitative vs quantitative parsimony.
20 I have previously discussed this possibility in Tahko 2014.
tuomas e. tahko 
the camels stand on four elephants, the four elephants stand on a turtle . . . and repeat
ad infinitum’. No other terms than these four elephants, a turtle, and two camels can
be introduced that would describe reality more minimally—they carve perfectly at
the joints and hence constitute the fundamental level in the sense of (GOF).21
In this toy example, the elephants, turtles, and camels may be considered as three
different fundamental natural kinds, for instance. But a repetitive structure can be
pictured otherwise too. Perhaps more in line with the structuralist conception, we
could picture the world as a fractal and zoom into that fractal to find further structure.
We can keep doing this indefinitely and always find further structure, more and
more elaborate, stunning patterns. But all these patterns, all the structure, is already
contained in the initial fractal. The function that produced the initial fractal contains
(recall the idea of fundamental entities as axioms, as suggested by Wilson), as it were,
all the further iterations that emerge when we zoom in (or indeed if we zoom out!).
To use Fine’s notion of component introduced in Section 4, we could perhaps conceive
of the iterations of the function to be components of the fractal, and in this sense the
fractal picture would have at least the appearance of minimality.
Admittedly, such abstract toy examples will do little to convince anyone of the
possibility of actual boring infinite descent, even though the idea itself shouldn’t be
all that alien, given the now familiar case of structuralism (think of relations all the
way down). Much of the real work would have to come from science, of course, and
one example that has received some attention is Hans Dehmelt’s (1989) Nobel lecture,
where he speculates about the possibility of a quark/lepton substructure based on
the model of the triton—the nucleus of hydrogen’s radioactive isotope tritium. This
tripartite substructure could, in Dehmelt’s view, be infinitely repeated, and since it is
the same structure that repeats, Dehmelt’s model is a prima facie candidate for boring
infinite descent—a gunky ontology. However, as far as I’m aware, the idea has not
been picked up by physicists and one might think that quantum field theory would
be a more promising line of research in this regard.22 In any case, here I only wish to
indicate the possibility of ‘boring infinite descent’ being compatible with ontological
minimality. The upshot is that the idea of ontological minimality, rather than (MF),
21 Cf. also Raven’s 2016 recent paper, where he distinguishes between the fundamental, the foundational,
and the eliminable. Eliminability implies a lower bound to the grounding chain—this is a version of
the idea typically expressed in terms of well-foundedness (e.g. Bennett 2011, 30; see also Dixon 2016)—
but Raven proposes that it’s possible to reject foundationalism, the claim that necessarily, something is
fundamental if and only if it is foundational. On Raven’s terms, infinitely boring descent would seem to
be a case of unbounded ineliminables, which ‘persist’, i.e. infinitely repeat. On my reading, I take it that
boring infinite descent would still be a case of foundationalism, while on Raven’s proposal it would qualify
as fundamentality without foundations, although more work needs to be done in order to properly compare
our accounts. Finally, see Wilson 2016 and especially the idea that there could be a ‘convergence’ to a
fundamental level at a limit. This idea could perhaps be understood as another type of ‘boring’ infinite
descent.
22 But for a philosophical study of Dehmelt’s model, see Tahko 2014.
 fundamentality and ontological minimality
may be a promising way to capture what I take to be the core of the fundamentality
intuition.23
References
Armstrong, D.M. (2004). Truth and Truthmakers. Cambridge University Press.
Bennett, K. (2011). By Our Bootstraps. Philosophical Perspectives, 25(1), 27–41.
Bennett, K. (2017). Making Things Up. Oxford University Press.
Bliss, R.L. (2013). Viciousness and the Structure of Reality. Philosophical Studies, 166(2):
399–418.
Bliss, R.L. (2014). Viciousness and Circles of Ground. Metaphilosophy, 45(2), 245–56.
Bliss, R.L. and Trogdon, K. (2014). Metaphysical Grounding. In E.N. Zalta (ed.), The
Stanford Encyclopedia of Philosophy (Winter 2014 Edition). [Online] Available from:
http://plato.stanford.edu/archives/win2014/entries/grounding/.
Bohn, E.D. (2009). Must There Be a Top Level? Philosophical Quarterly, 59, 193–201.
Cameron, R.P. (2008). Turtles All the Way Down: Regress, Priority and Fundamentality.
Philosophical Quarterly, 58, 1–14.
Cotnoir, A.J. (2013). Beyond Atomism. Thought, 2, 67–72.
Dehmelt, H. (1989). Triton, . . . Electron, . . . Cosmon, . . .: An Infinite Regression? Proceedings
of the National Academy of Sciences, 86, 8618–9.
Dixon, S. (2016). What Is the Well-Foundedness of Grounding? Mind, 125(498), 439–68.
Dorato, M. and Morganti, M. (2013). Grades of Individuality: A Pluralistic View of Identity in
Quantum Mechanics and in the Sciences. Philosophical Studies, 163, 591–610.
Esfeld, M. and Lam, V. (2010). Ontic Structural Realism as a Metaphysics of Objects. In
A. Bokulich and P. Bokulich (eds), Scientific Structuralism (pp. 143–59). Dordrecht: Springer.
Fine, K. (1995). Ontological Dependence. Proceedings of the Aristotelian Society, 95, 269–90.
Fine, K. (2001). The Question of Realism. Philosophers’ Imprint, 1(1), 1–30.
Fine, K. (2010). Towards a Theory of Part. Journal of Philosophy, 107(11), 559–89.
Fine, K. (2012). ‘A Guide to Ground’. In F. Correia and B. Schnieder (eds), Metaphysical
Grounding: Understanding the Structure of Reality. Cambridge University Press.
French, S. (2014). The Structure of the World. Oxford University Press.
Hawthorne, J. (2008). Three-Dimensionalism vs. Four-Dimensionalism. In T. Sider, J.
Hawthorne, and D.W. Zimmerman (eds), Contemporary Debates in Metaphysics. Blackwell.
Kim, J. (2010). Essays in the Metaphysics of Mind. Oxford University Press.
Ladyman, J. and Ross, D. (2007). Every Thing Must Go: Metaphysics Naturalized. Oxford
University Press.
Lewis, D. (1991). Parts of Classes. Oxford: Blackwell.
McDaniel K. (2013). Degrees of Being Philosophers’ Imprint 13(19), 1–18.
23 I would like to thank audiences at Barcelona, Cambridge, Helsinki, Hong Kong, London, and
Singapore, for helpful discussion on the material of the paper. Thanks also to Ken Aizawa, Jason Bowers,
Carl Gillett, Matteo Morganti, Donnchadh O’Conaill, and Jessica Wilson for comments on some of the
material, and to two anonymous referees for this volume. The research for this paper was made possible by
two grants from the Academy of Finland, decisions #266256 and #274715.
tuomas e. tahko 
McKenzie, K. (2014a). Priority and Particle Physics: Ontic Structural Realism as a Fundamen-
tality Thesis. British Journal for the Philosophy of Science, 65(2), 353–80.
McKenzie, K. (2014b). On the Fundamentality of Symmetries. Philosophy of Science, 81(5),
1090–102.
Markosian, N. (2005). Against Ontological Fundamentalism. Facta Philosophica 7, 69–84.
Morganti, M. (2014). Metaphysical Infinitism and the Regress of Being. Metaphilosophy 45(2),
232–44.
Morganti, M. (2015). Dependence, Justification and Explanation: Must Reality Be Well-
Founded? Erkenntnis 80(3), 555–72.
Nolan, D. (1997). Quantitative Parsimony. British Journal for the Philosophy of Science 48(3),
329–43.
O’Conaill, D. (2014). Ontic Structural Realism and Concrete Objects. Philosophical Quarterly,
64(255), 284–300.
O’Conaill, D. and Tahko, T.E. (2016). Minimal Truthmakers. Pacific Philosophical Quarterly,
7(2), 228–44.
Paseau, A. (2010). Defining Ultimate Ontological Basis and the Fundamental Layer. Philosoph-
ical Quarterly 60, 169–75.
Raven, M. (2016). Fundamentality without Foundations. Philosophy and Phenomenological
Research 93(3), 607–26.
Schaffer, J. (2003). Is There a Fundamental Level? Noûs 37, 498–517.
Schaffer, J. (2004). Two Conceptions of Sparse Properties. Pacific Philosophical Quarterly, 85,
92–102.
Schaffer J. (2009). On What Grounds What. In D. Chalmers, D. Manley, and R. Wasserman
(eds) Metametaphysics (pp. 347–83). Oxford University Press.
Schaffer, J. (2010). Monism: The Priority of the Whole. Philosophical Review, 119(1), 31–76.
Schaffer, J. (2012). Grounding, Transitivity, and Contrastivity. In F. Correia and B. Schnieder
(eds), Metaphysical Grounding (pp. 122–38). Cambridge University Press.
Tahko, T.E. (2013). Truth-Grounding and Transitivity. Thought: A Journal of Philosophy, 2(4),
332–40.
Tahko, T.E. (2014). Boring Infinite Descent. Metaphilosophy, 45(2), 257–69.
Tahko, T.E. (2015a). An Introduction to Metametaphysics. Cambridge University Press.
Tahko, T.E. (2015b). Natural Kind Essentialism Revisited. Mind, 124(495), 795–822.
Tahko, T.E. and Lowe, E.J. (2015). Ontological Dependence. In E.N. Zalta (ed.), The
Stanford Encyclopedia of Philosophy (Spring 2015 Edition). [Online] Available from:
http://plato.stanford.edu/archives/spr2015/entries/dependence-ontological/.
Trogdon, K. (2009). Monism and Intrinsicality. Australasian Journal of Philosophy, 87(1),
127–48.
Wilson, J.M. (2014). No Work for a Theory of Grounding. Inquiry, 57(5–6), 535–79.
Wilson, J.M. (2016). The Unity and Priority Arguments for Grounding. In K. Aizawa and
C. Gillett (eds), Scientific Composition and Metaphysical Ground (pp. 171–204). Palgrave
Macmillan.
13
The Structure of Physical Reality
Beyond Foundationalism
Matteo Morganti
1 Introduction
The interest in the notion of grounding has grown exponentially in recent years.1 In
a rough summary, grounding is understood as a sui generis ‘in virtue of ’ relation—
in particular, one that is arguably different from causation, necessitation, and super-
venience. Grounding may not be a unique, monolithic relation, yet it subsumes a
number of more specific relations—such as truth-making, mereological part–whole
relations and others—which share the peculiar feature that the relata on one end non-
causally determine the relata on the other. Grounding is an objective relation, that
is, one that exists in the world independently of human practices,2 and is intimately
connected, if not identified, with ‘metaphysical explanation’.3 Grounding is (generally!
more on this later) understood as a relation that gives rise to strict partial orders, that
is, hierarchies in which no element grounds itself (irreflexivity); if a grounds b then b
does not ground a (asymmetry); and if a grounds b and b grounds c then a grounds
1 For an overview, see Bliss and Trogdon (2014).

2 Note, however, that the ‘sentential’ or ‘operational’ view of grounding, according to which grounding
is an operator connecting sentential expressions, allows one to remain neutral on ontological questions.
This is, in particular, the view advocated by Kit Fine.
3 This is a sui generis sort of explanation, though: in particular, it is non-epistemic in the sense that
it need not be the case that the grounding is what allows us to gain knowledge of the grounded; and that
grounding explanations need not obey the constraints that seem to apply in other cases, e.g. to explanations
provided in the domain of practical reasoning. Also, the notion of metaphysical explanation brings with
itself the idea that grounding is non-monotonic, i.e. it could be the case that a and b together ground c, but
a,b, and d together fail to ground c.
matteo morganti 
c (transitivity);4 and in most cases well-foundedness is also assumed, meaning that

grounding chains have an ultimate, ungrounded basis.5
There is intense discussion, however, on each one of the above elements. In partic-
ular, some authors have questioned the transitivity of grounding (Schaffer (2012)); its
being asymmetric and well-founded (Thompson (2016)); its being irreflexive (Jenkins
(2011) and Correia (2014)); and, recently, all of these in one take (Rodriguez-Pereyra
(2015)). On the other hand, some of the attempts just mentioned have been rebutted
(for instance, replies to Schaffer (2012) have been provided by Litland (2013)). Others
crucially rely on the assumption that grounding is a unitary concept, so that cases of
more specific relations that appear not to have the features of a strict-partial order
are used to question the canonical view of grounding in general (Griffith (2014),
for instance, suggests that truth-making is at best a peculiar species of grounding
involving a unique form of dependence, hence it is wrong to draw general conclusions
about grounding from putative counterexamples involving truth-making). Reasons
for doubt linger in other cases as well (Jenkins’ (2011) arguments against irreflexivity,
for example, seem to trade on an ambiguity between facts and linguistic items
referring to them).
The present paper also discusses the specific features of grounding relations. The
goal, however, is not to make general claims about grounding per se but, rather, to
expand on the idea that grounding may give rise to structures that differ significantly
from those normally associated to the notion. I will look at the characteristics that
non-standard grounding structures may possess; and at possible interesting uses that
they could be put to.6 Specific examples will be considered, intended to lend support
to the view that there may be good reasons for conceiving of specific aspects of reality
in, non-foundationalist, terms.
An important thing to be made explicit at the outset is that, in the sections to
come, I will presuppose that the relevant grounding relations hold between worldly
entities,7 and determine in virtue of what (some of) the things that make up material
4 To be precise, explicitly assuming a many–one conception of ground, transitivity must be replaced by
a ‘cut’ rule: if a grounds b, c grounds d, . . . , and some together with b and d ground e, then a, b, . . . , and
together ground e; also, irreflexivity and transitivity make sense if one explicitly introduces partial grounds
(where something partially grounds something else iff it is part of the latter’s grounds) and appropriate
‘subsumption rules’ for moving from full to partial ground.
5 The ground-theoretic notion of well-foundedness is not easily defined despite (or maybe because of)
the similarities it bears with mathematical and set-theoretical well-foundedness. For good discussions, see
Scott Dixon (2016) and Rabin and Rabern (2016).
6 In so doing, I will by and large gloss over several important aspects of the debate which are, however,
irrelevant for present purposes: among these, the questions whether grounding facts are themselves
grounded and, if they are, whether a notion of ‘grounding ground’ can be coherently developed; what the
relata of the grounding relation are/could be; whether grounds necessitate the grounded; and whether the
metaphysical structure of the actual world is necessary or contingent.
7 These will more or less interchangeably be taken to be objects, facts, or, more vaguely, ‘levels’, under
the assumption that object-language and fact-language are easily translated into one another, and that the
levels metaphor that is often used (at least in the context of the traditional, hierarchical view of reality)
simply refers to groups of objects/facts that are in some sense equally derivative/close to the fundamental.
 structure of physical reality: beyond foundationalism
reality exist as the very things that they are. This is tantamount to saying that I
will assume that ‘a grounds b’ is more or less a synonym of ‘b depends on a for its
existence/essence/identity’. More precisely, I will assume that, at least in the domains
that are relevant for the present discussion, grounding claims coincide with claims of
(possibly generic) ontological dependence between physical entities.8
This might be taken to deprive the present inquiry of interest, on the basis that
ontological dependence is usually regarded as a broader notion than grounding
and, consequently, equating the two might be a cheap way of arguing for non-
standard grounding structures. However, first, considerations concerning (non-)well-
foundedness are untouched by this objection, as both grounding and ontological
dependence are usually regarded as giving rise to well-founded structures. Yet, it
is mostly the well-foundedness assumption that will be questioned in what follows.
Second, it is true that counterexamples have already been provided to the claim that
ontological dependence relations give rise to strict partial orders, based on putative
cases of symmetric dependence. However, such counterexamples are irrelevant here:
for instance, it may be the case, say, that Dante ontologically depends on his life but
also the converse, yet this says nothing in favour of the symmetry of the dependence
between Dante and his material constituents. But it is facts of the latter type, that is,
facts about the structure of material reality, that are of interest for us here. Third, I take
the widespread claims to the effect that, obviously, ontological dependence is reflexive
while grounding is not to be definitely unconvincing: for, either one is making a trivial,
uninformative claim, in which case there is no difference between ‘x is P in virtue
of x’s being P’ and ‘x has existence in virtue of x’s being an existing thing’; or, one is
attempting to formulate an informative statement, and then again, say, ‘x is P in virtue
of x being Q’ and ‘x depends for its identity on x’s existence’ are on a par.
There is, of course, a lot more to say about these issues, not lastly because of ter-
minological ambiguity in the extant literature.9 At any rate, in the rest of the paper I
will assume that, given the above restrictions and caveats, it is indeed an interesting
task to inquire into the possibility and significance of non-standard metaphysical
8 Unlike its specific counterpart that holds between particular entities, generic dependence holds
between entity-types. The specification concerning generic dependence takes care of counterexamples
such as the following: a true disjunct grounds a disjunction, but the disjunction is not dependent on
that particular disjunct for its truth (the other disjunct might be/have been true); or, a material object
is grounded in its constituent parts, but it is not existentially dependent on them, as it can survive the loss
or replacement of one or more of its specific constituents. It remains true that disjunctions depend on there
being some true disjuncts, and that a material object depends on there being some material constituents.
9 Casual mention of the fact that dependence and grounding are ‘very close’, ‘akin’, ‘cognate’ . . . notions
is often made. At the same time, there is no agreement on which properties the two relations need to
have, hence what exactly makes them similar/differentiates them. A common reason for separating the two
notions is that grounding appears to be more closely related to explanation than ontological dependence.
But, apart from the fact that grounding is more or less defined in terms of metaphysical explanation,
ontological dependence too appears intimately connected to explanation: isn’t the claim that a ontologically
depends on b ultimately an explanatory claim of the ‘in virtue of ’ type? For a good discussion of ontological
dependence, including some thoughts on its connection with grounding, see Tahko and Lowe (2015).
matteo morganti 
models of the structure of physical reality framed in terms of ontological priority/

dependence relations between concrete entities.10
The plan of the paper is as follows. Section 2 briefly discusses the traditional view,
based on grounding relations as determining strict partial orders and well-founded
structures—so-called ‘metaphysical foundationalism’. The discussion then focuses on
the prospects of non-standard models of the metaphysical structure of (parts of)
physical reality. Section 3 looks at ‘infinitist’ models, where the well-foundedness
assumption is dropped. Section 4 discusses ‘coherentist’ models, in which grounding
relations fail to be irreflexive and symmetric and grounding structures give rise to
‘loops’ and/or ‘webs’. Section 5 concludes the paper by considering the plausibility of
what one may call ‘hybrid’ models and, more generally, of pluralism with respect to
the metaphysical structure of reality.
2 Foundationalism
By far the most widespread view of the metaphysical structure of reality is foundation-
alism: the view that the world has a ‘vertical’, hierarchical structure and a fundamental
level. The latter could be identified with the entire cosmos as a Parmenidean One
(monistic foundationalism) or be somehow situated at an intermediate level (an as
yet virtually unexplored, but admissible alternative). However, the most popular
form of metaphysical foundationalism is no doubt pluralistic foundationalism. In
analogy with foundationalism in epistemology—where a basis of non-further justified
reasons is invoked to justify our beliefs—this view has it that there is a multitude of
fundamental constituents of reality which are not themselves grounded. In particular,
pluralistic foundationalism is normally understood in terms of a finite set of types of
entities, tokens of which may then be finite or infinite in number, making up the whole
of physical reality. A paradigmatic example is the atomistic picture we are all familiar
with, introduced by Leucippus and Democritus (as well as by some thinkers in ancient
India, such as those belonging to the Nyaya–Vaisesika school, which flourished
between the 6th and 1st centuries bc) and revived by Renaissance philosophers.11
Importantly, atomism identifies the relations that make up the structure of reality
with mereological relations.12 This picture is seemingly vindicated nowadays by the
10 Clearly, the above entails that I am assuming that grounding need not be a monolithic notion, to which
all ‘in virtue of ’ metaphysical explanations have to be reducible. Borrowing Wilson’s (2014) expressions
while disagreeing with her conclusions, I believe that there may be many distinct ‘small-g’ grounding
relations without the study of ‘capital-G’ Grounding necessarily being devoid of interest—the former
relations just need to be sufficiently analogous to each other.
11 It is interesting to notice that thinkers such as Gassendi, Hobbes, Boyle, and even Newton rather
subscribed (at least at some point in their life) to corpuscularianism, a view according to which the
supposedly basic particles could in principle be divided.
12 While it seems obvious that part–whole relations give rise to strict partial orders, the existence of
a fundamental level of entities that do not have proper parts is not an axiom of mereology. Also, that
there is a fundamental, all-encompassing whole is only an axiom in classical mereology, where unrestricted
composition holds. This is crucial for foundational monists.
 structure of physical reality: beyond foundationalism
Entities at level 1
Entities at level 2
Entities at level 3
…
…
F
Figure 13.1.
Standard Model of elementary particles, according to which there are seventeen kinds
of basic entities (plus the as of now still undetected gravitons) and everything is
derivative on these, in the sense of being composed of them.
According to foundationalism, then, reality looks something like Figure 13.1, F
being the fundamental level occupied by the atoms (or, alternatively, by the cosmos),
acting as the foundation for every grounding chain.13
Foundationalism is certainly close to common sense. In particular, the search for
a ground seems to intuitively require a terminating point, hence a fundamental,
ungrounded level. However, there is obviously no reason for concluding that reality
has a foundation just because we cannot think otherwise without strong-arming our
intuitions; nor for ruling out the possibility that an explanation that explicitly rules
out the existence of a fundamental level may turn out to be preferable, all things
considered, in the particular case of metaphysical structure. Indeed, arguments have
been put forward to the effect that no compelling reasons for endorsing founda-
tionalism can be found in the history of either science or philosophy (see Schaffer
(2003) and Bohn (2009)), and thus the prevalence of foundationalism may well be
due exclusively to the intrinsic limitations of our cognitive faculties. On the other
hand, it must be noted that, in fact, the case against foundationalism rests mostly on
reasoning of a meta-inductive kind: that is, from the fact that (a) in the past we have
always abandoned hypotheses that were formulated concerning the fundamental, and
accepted the groundedness of what we used to regard as ungrounded, the conclusion
is inferred that (b) probably, there is no fundamental level. This sort of inference is
notoriously problematic. In light of the available historical data, one may conclude
equally well that we probably haven’t got to the fundamental level yet. In addition to
this, it must be taken into account that a foundationalist presupposition appears to
be generally at work in contemporary physics, considerations of a purely historical
nature consequently appearing insufficient for mounting a compelling case against
foundationalism.14 This is certainly a complex set of issues, calling for a careful
13 Of course there may be more than one set of fundamental entities, each one constituting the basis for
a proper subset of all grounding chains.

14 For authoritative examples, see, for instance, Feynman (1994, ch. 1, pp. 2, 20) and Rovelli (2016, ch. 1).
It goes without saying that, even in the case of the most prominent thinkers, such a presupposition might
just be based on common-sense.
matteo morganti 
Entities at level …
Entities at level 1
Entities at level 2
…
…
…
Figure 13.2.
analysis of a number of different elements, and for some key decisions at the level
of general methodology.15
In what follows, I will take for granted that foundationalism has an intuitive pull,
and that the case against it cannot rest simply on the historical fact of theoretical revi-
sion concerning the (allegedly) physically fundamental. Also, I will assume that overly
general claims as to truth or falsity of foundationalism (or any other theory of the
metaphysical structure of physical reality) per se are likely to be highly contentious—
why, after all, should every corner of reality have the same structure as all the others?
Against this background, I will explore some alternatives to foundationalism, looking
in particular at the surplus explanatory resources that they may provide compared
to foundationalism, especially when applied ‘locally’, that is, to specific domains of
inquiry.
3 Infinitism
Suppose that one drops the well-foundedness assumption, and concedes that ground-
ing chains may fail to terminate in fundamentals. If one does this and preserves the
idea of a strict partial order, one obtains the ‘infinitist’ alternative to foundationalism:
one whereby reality has a hierarchical structure (possibly, but not necessarily, based on
mereological relations) which is open-ended either in just the direction towards which
the asymmetric relations of dependence go, or in both directions (Figure 13.2).16
This is something akin to the view of Anaxagoras, who posited fundamental ‘seeds’
of reality but also explicitly stated that ‘There is no smallest among the small and no
largest among the large, but always something still smaller and something still larger’
(fragment 15).
15 McKenzie (2011), for instance, argues that philosophical naturalists—that is, those who require
philosophy to be responsive, in some sense, to science—had better ground their case for or against
foundationalism exclusively on a careful study of specific theories rather than on a general consideration
of the history of science.
16 In the case of mereological relations, one obtains so-called ‘gunky’ worlds in case everything has a
proper part, and ‘junky’ worlds in case everything is a proper part of something. ‘Hunky’ worlds contain
both gunk and junk.
 structure of physical reality: beyond foundationalism
The immediate objection is obvious: reality cannot have an infinitist structure

because, if it did, everything would simply fail to ‘obtain’ either its existence and/or
the distinctive features that make it the existent that it is. However, it has already been
argued in the literature that the infinite regress that arises in infinitist frameworks
need not be vicious (see e.g. Bliss (2013)). In Morganti (2014), an analogy with
epistemological infinitism was suggested and an emergentist model of being was put
forward, whereby the grounding relations do not ‘transmit’ existence, identity, or
whatever other property, from one level to the other; rather, they allow such features
to emerge from the relevant chains of grounding as a whole (be such chains finite or
infinite). Without getting into the details, let us assume that something along these
lines works, infinitism consequently constituting a relevant philosophical option,17
and consider next two specific examples of infinitist structure—one rather briefly and
another in a bit more detail.18
First, let us look at S-matrix theory. Dating back to Wheeler’s and Heisenberg’s work
in the late 1930, the S-matrix theory avoids the explicit assumption of space and time,
and makes do with the abstract mathematical features of a particular mathematical
entity (the S-matrix, of course) which connects the infinite past to the infinite future in
one step, without any possible decomposition into intermediate steps corresponding
to time-slices. What is important for present purposes is that, as shown by McKenzie
(2011), the theory does not need to define a priori the properties of fundamental
elementary particles. More strongly, it turns out that it cannot provide a complete
and coherent description and understanding of processes involving certain types of
particles, if not by making reference to all other types of particles; and that all particles
described by the theory are necessarily composed of other particles.19 This sort of
endless compositeness, clearly, only makes sense in a metaphysical infinitist context.
Therefore, if there’s anything correct in the fundamental structure of S-matrix theory
(hence, in string theory, which is the modern-day development of S-matrix theory as
applied to the strong interaction), metaphysical infinitism should be taken seriously
in the light of contemporary physics.
Let us consider next so-called ‘hierarchical’ cosmological models of the universe.
The basic idea is to conceive of the universe as a continuously repeating structure,
17 One may ask whether infinite grounding chains should be ‘boring’—the same structure repeating
itself ad infinitum. This would provide a foundation at least at the level of the types of entities that are
chained, and/or of the way in which the chaining takes place, so assuaging at least some of the conceptual
worries that arise from infinite regresses. For a discussion of this issue, and a defence of ‘boring infinite
descent’, see Tahko (2014).
18 For another example, Dehmelt (1989) proposes an (allegedly superior) alternative to the Standard
Model of elementary particles, moving beyond the level of quarks and in fact postulating a possibly infinite
series of levels of subatomic structure. Arguably, though, Dehmelt’s model is infinitely layered yet well-
founded.
19 In particular, all hadrons, i.e. the particles normally understood as composites of quarks held together
by the strong force, must be described as composed by other hadrons, so that quarks (or any other kind of
particles, for that matter) cannot be regarded as fundamental.
matteo morganti 
whereby, for instance, stars group together to form galaxies, galaxies group together to
form clusters, clusters group together to form superclusters and so on without an upper
limit. In some versions of this view, one speaks of ‘fractal cosmologies’, postulating that
the universe assumes the features of a fractal across infinitely many scales.20 A uni-
verse of this sort was first envisaged by the German mathematician Johann Lambert
and by Immanuel Kant, based on some observations made by the Swedish astronomer
Emanuel Swedenborg. Lambert and, later, astronomers such as John Herschel and
Richard Proctor and then Fournier d’Albe and Carl Charlier used this model to solve
what is known as ‘Olbers’ paradox’ besetting stationary cosmological models.21, 22
In this case too, the interest of the model is not merely historical, nor is it limited
to the analysis of a specific problem concerning the brightness of the night sky. To
the contrary, the discovery of evidence of clustering at several levels led to further
development of the hierarchical model in recent years, reinvigorating stationary
cosmologies as an alternative to the Big Bang, free from the issue of explaining
the possibility and status of ‘boundaries’, and independent of the very idea of a
temporal beginning.23 Perhaps more importantly, a novel project in theoretical
physics, which has gained a lot of attention and caused much discussion in the
last few years, rests on very similar ideas: it is the conjecture that our universe
originated from a black hole in a larger universe and, more generally, black holes
may be the source of new universes at many (possibly, infinitely many) levels. This
proposal has a strong, well-defined motivation: according to Big Bang cosmology,
the universe originated from a point of infinite density (singularity) with no cause,
and this fact is strictly speaking unphysical. Moreover, a mechanism of cosmic
inflation must be introduced in order to account for the density, homogeneity,
and isotropy of the actual universe; and the presence of antimatter and of a
definite direction of time also have to be simply assumed. Alternative models of
the hierarchical type promise to offer a more informative and unitary answer to all
these interrogatives. Poplawski (see e.g. Poplawski (2010)), for instance, recently
got back to the so-called Einstein–Cartan (or Einstein–Cartan–Sciama–Kibble)
20 A fractal being a mathematical structure that exhibits a repeating pattern at every scale, ad infinitum.
By extension, fractals are also said to be found in nature, e.g. in the case of snowflakes or other crystals,
or even some vegetables. Not surprisingly, in this case the repetition of the pattern is (usually) taken to be
finite.
21 That is, the problem of explaining why the sky is (largely) dark at night in spite of the number of
stars—under the assumption that the universe is infinite, static, and eternal—supposedly being infinite.
Despite its name, Olbers’ paradox really is a non-paradoxical riddle that sets certain constraints on workable
cosmological models. (Incidentally, the name is not appropriate also in another sense, as the issue was first
discussed not by the late 18th- and early 19th-century German astronomer Heinrich Wilhelm Olbers, but
rather by Thomas Digges in 1576.)
22 The fractal solution is based on the fact that, in the appropriate fractal structure, as distances increase
the amount of incoming light becomes smaller, and in the limit light fails to reach the observer. Thus, only
a finite amount of light reaches the observer in spite of the universe being infinite in space and time and
the number of stars being also infinite. For a more detailed discussion, see Harrison (1987; ch. 11).
23 See Pietronero (1987), Nottale (1993), and Guth (2007).
 structure of physical reality: beyond foundationalism
theory of gravity. This theory extends general relativity, allowing the geometrical
part of space–time called torsion to be different from zero (like curvature). In
an appropriately updated form, Poplawski argues, the Einstein–Cartan theory
affords models that correctly account for the quantum-mechanical, intrinsic angular
momentum (spin) of elementary particles called fermions, and in which, crucially,
torsion can be shown to be responsible for all the puzzling facts mentioned above in
relation to Big Bang cosmology. In particular, Poplawski suggests, it is due to torsion
that inside black holes one doesn’t get singularities but rather new ‘bounces’ and the
creation of new universes.24
Clearly, the sort of hierarchical cosmological models that have just been discussed
describe junky universes, in which the part–whole relation gets reiterated in the
‘upward’ direction ad infinitum. Noticing that at least some approaches to quantum
gravity—the much sought unification of quantum mechanics and general relativity—
introduce a fractal nature also at the other end,25 the fractal hypothesis can be easily
seen to fit infinitism also in the sense of reality being gunky as well as junky and,
consequently, hunky (see footnote 16 above).
Now, there is of course a lot more to say about fractal cosmologies, infinitist models
of subatomic structure, and other possible forms of infinitism in contemporary
physics. Here, however, it is sufficient to point out that (i) such models in fact exist,
and are elaborated upon and discussed by physicists; and (ii) they appear, or at least
promise, to have a non-negligible amount of explanatory power, which is likely, at least
in some cases, to be essentially due to the role infinite series of dependence relations
play in them. At the very least, then, the acceptance of infinitism provides a wider
range of options in terms of explanatory hypotheses for particular domains of inquiry,
and this alternative to foundationalism consequently deserves serious attention on the
part of philosophers.
4 Coherentism
Arguably, the most underdeveloped conception of the metaphysical structure of
reality is metaphysical coherentism. As with coherentist solutions to the problem
of justification in epistemology, metaphysical coherentism abandons the idea of a
pyramidal structure of directed, asymmetric relations with an ultimate foundation.
As in Quinean webs of beliefs, the entities that make up reality are instead deemed
24 It is worth noticing that essentially the same idea underlies Smolin’s recent account of the universe
whereby physical laws evolve in time through some sort of natural selection mechanism for universes
(see e.g. Smolin (2013)). Thus hierarchical models of the universe may also provide answers to problems
surrounding the nature of physical laws and the specific values of the cosmological constants. See also Smith
(1990).
25 This seems to happen, for instance, in the case of Connes’ approach based on noncommutative
geometry. See e.g. Connes (1996).

matteo morganti 
Physical
Object 1
Physical
…
Object 2
Physical Physical
Object 3 Object 5
Physical
Object 4
Figure 13.3.
to be mutually related in structures that are at least partly composed by ‘cycles’ or

‘loops’.26 Figure 13.3 represents schematically a possible coherentist structure.
But why should one take coherentist models seriously at all? What features
can/should these have, exactly? Very little has been said, if anything, on this so far.
To begin with, coherentism appears to violate irreflexivity: since cycles of priority
and dependence relations ultimately lead back to the starting point, it looks as though
they necessarily make something self-grounded. Of course, it could be responded that
everything trivially depends on itself, and therefore this is not a problem. The point,
however, is that what is at stake here are non-trivial relations underpinning genuinely
informative metaphysical explanations. And indeed, such explanations are normally
expected to be based on irreflexive relations. Not surprisingly, grounding is also
normally taken to be an irreflexive relation. However, first, the idea that grounding
must be irreflexive can be questioned (Jenkins (2011) and Correia (2014) have been
mentioned earlier). Secondly, and perhaps more importantly, it is in fact not true that
cycles entail the reflexivity of the relevant relations. As a matter of fact, it can be shown
that irreflexivity may hold in a cycle, provided that one has either antitransitivity (the
relevant relation is never transitive) or intransitivity proper (the relevant relation is not
26 There is in fact a relevant difference in epistemology between scenarios in which justification goes
in circles and more sophisticated ones whereby structures with the appropriate features—in particular,
internal coherence—are justified as a whole. This latter, holistic form of coherentism is probably more
credible as a model of metaphysical structure; the former, however, is also relevant (consider, for instance,
the circular dependence that is established between substrata and property instances on at least some
accounts of the ontology of material objects; or the mutual dependence between triplets of quarks inside
protons or neutrons).
 structure of physical reality: beyond foundationalism
always transitive).27 While the former option seems too high a price to pay, the latter
appears worth exploring further, as it allows for cycles in which grounding relations
are generally transitive, and yet nothing grounds itself.
In connection to this, think about quasi-transitivity as understood in social choice
theory and microeconomics. The notion was introduced by Sen (1969) in order to
account for the consequences of Arrow’s theorem—according to which, when subjects
have three or more distinct alternatives to choose from, no system can convert the
ranked preferences of individuals into a shared, generally valid, ranking while meeting
certain sensible criteria. Roughly, quasi-transitivity allows one to consistently have
that (for a given subject) a is not better than b, and b is not better than c, and yet
c is better than a; or, applied for instance to the sorites paradox, that x is not more
a heap than y, y is not more a heap than z, and yet z is more a heap than x. In the
grounding case, mutatis mutandis, one may analogously have that a grounds b, b
grounds c, but c does not ground a. In the case in which a is identical to c, one obtains
the desired result, that is, a structure or substructure in which everything grounds
(or may ground) everything else except itself.28 It must be noted that, in fact, quasi-
transitive relations are acyclic, that is, ground as described above, like preference in
scenarios such as those studied by Sen, does not admit cycles. Yet, it could be that a
more general, quasi-transitive relation ‘coexists’ with another one which is transitive
and reflexive, hence possibly cyclic. For instance, in the present case one may postulate
that full ground relations only connect entities/levels that are next to each other in the
cycle, while only partial ground relations can connect entities/level that are further
away from each other in the cycle. So, for instance, it could be that a fully grounds b,
b fully grounds c, but a only partially grounds c (and does so exactly in virtue of the
fact that it fully grounds c’s full ground and does not ground c directly). If c then fully
grounds a, one obtains that a only partially grounds itself.
A ‘third way’ of sorts may also be introduced when it comes to reflexivity versus
irreflexivity. In particular, one can have recourse to the notion of quasi-reflexivity:
something is related to (in our case, partially grounds) itself if and only if it is related
to something other than itself.
27 Transitivity is the claim that ∀xyz((Rxy∧Ryz)⇒Rxz), and one could either negate the universal
quantifier or the conditional that is quantified over. It is worth emphasizing that something analogous
holds for all formal features of relations, including symmetry, asymmetry, anti-symmetry, etc., from which
a greater variety of options than one may think immediately follows.
28 Of course, our intuition goes in a different direction, and tells us that if something grounds its own
grounds, then it grounds itself. But intuition is not, at least not obviously, a particularly helpful guide in phi-
losophy. In the present case, our prejudices might be at least partly rooted in the fact that we tend to identify
grounding relations with mereological relations. However, non-classical mereologies can be developed, and
even the transitivity of parthood is not entirely indisputable. Incidentally, results in this sense may be of
relevance in physics, where the opinion is becoming widespread that composition/decomposition relations
are not amenable to an interpretation in terms of classical mereology (see e.g. Healey (2013) and Caulton
(2015)). More in general, there is no reason for identifying grounding relations with mereological relations,
and indeed the coherentist case might be best made on the basis of completely different assumptions (more
on this in a moment).
matteo morganti 
It goes without saying that a precise formal account of the above notions should be
given. Here, however, suffice it to say that (i) quasi-transitivity reduces to asymmetry
(without anti-symmetry) plus transitivity, and leaves irreflexivity untouched; (ii) the
abandonment of irreflexivity is already accepted, as mentioned above, in trivial cases
of ontological self-dependence; and (iii) there is only one non-negligible change
required in the set of the allegedly fundamental principles governing ground once
one drops (or weakens) reflexivity—one has to give up the strong asymmetry principle
according to which it is never the case that A together with some grounds A -, but
this is arguably compensated in terms of explanatory capacity of the resulting model.
With respect to this last, crucial point, an interesting case study for metaphysical
coherentism is constituted, I take it, by so-called ontic structural realism. It is by
now a traditional dispute in the philosophy of science that between scientific realists,
who consider the (approximate) truth of scientific theories the best explanation
of their empirical success; and anti-realists, who, emphasize instead—among other
things—the discontinuity across theory-change in the history of science. What is
known as structural realism attempts to re-establish the connection between success
and truth by pointing to the structural continuity that exists between (some parts
of some) subsequent theories across theory change. ‘Epistemic structural realists’
endorse the epistemological view that we can be realist about whatever is described
by the (preserved) structure of our theories. More or less recently, many authors
have subscribed instead to ‘ontic structural realism’ (OSR), the view according to
which structure is not only all we can be realist about but also all there is. Setting
aside the realist component of the theory, and the issues whether (a) this move from
epistemology to ontology is justified in the first place, and (b) OSR can do all the work
it is expected to do as a metaphysical view,29 an interesting question concerns what
structuralism exactly amounts to as a metaphysical claim.
In the extant literature, OSR has been understood as an eliminative position,
whereby objects are reduced to the relational structures that constitute the real-
world counterpart of the relevant theoretical apparatus that defines and describes
those objects. On weaker readings, either objects and relations are ontologically on
a par (‘moderate’ OSR)—so sidestepping the infamous ‘no relations without relata’
problem—or what is relational is really only the identity of objects which otherwise
preserve their intrinsic qualitative features (‘contextual’ OSR). Now, it has been argued
(Wolff (2012), McKenzie (2014)), that eliminative OSR is hard to turn into a complete
and consistent metaphysical picture. For, either structures are defined extensionally,
but then objects are not eliminated; or they are defined intensionally, but then we seem
to lack a well-defined way of determining their identity conditions and, consequently,
their grounding role with respect to objects (as objects and their properties appear in
turn necessary to define the relevant relations). Additionally, and more importantly,
29 On the first problem, see Morganti (2011). On the second, French (2014) is a useful general discussion.
 structure of physical reality: beyond foundationalism
a complete structural reduction of all intrinsic properties of particles has not been
carried out yet. The remaining, weaker options also appear problematic. Moderate
OSR vindicates a form of scientific structuralism at the price of introducing objects
as mere placeholders, deprived of any intrinsic characterization, and consequently
akin to something like Lockean bare particulars. And even if this were not regarded
as unacceptable, the view also establishes a necessary relation of mutual ontological
dependence between objects and relations that many regard as in principle bad. Of
course, this need not be so, and surely is something coherentists would disagree with.
Notice, however, the specific worry that arises in this particular case: why introduce
a sui generis, additional ontological category at all if it only comprises entities that
are necessarily dependent on tokens of another category, and are essentially deprived
of properties and even identity? Lastly, contextual identity structuralism also has
shortcomings—at least three. First, those subscribing to it have rested their case exclu-
sively on the permutation-invariance of certain fundamental theories of contemporary
physics: very roughly, merely exchanging two exactly similar quantum particles or two
generally relativistic space–time points with each other is physically irrelevant, that is,
it doesn’t give rise to a new state of affairs, and this is taken to point to the necessary
extrinsicness of identity. However, this seeming violation of haecceitism simply does
not have to be explained by dropping intrinsic identities: other explanations have been
shown to be available.30 Secondly, the cost of this form of generalized contextualism
is that identity must be regarded as extrinsic even in cases in which it does not
correspond to any qualitative physical feature of the relevant structure. And this opens
the questions whether the ensuing form of structuralism is still genuinely physics-
based and, if it is, whether the attribution of non-structural, primitive intrinsic
identities is not an acceptable option too. Lastly, the limitation of structuralism to
identity facts hardly vindicates the initial structural intuition, which concerned the
description of reality provided by physical theory at a much more general level.
Overall, then, it looks as though the claim of priority of physical structures over
physical objects distinctive of OSR is quite difficult to flesh out—whatever exact
formulation of the view one chooses.
However, consider now a different understanding of the situation: one whereby the
key claim is not that physical relations are prior to physical objects—or, at least, as
fundamental as the latter; but, rather, that objects can be regarded as fundamental,
provided that an essential part of their being what they are31 is not taken to derive
from ‘lower’ or ‘upper’ more fundamental levels, and is instead conjectured to stem,
so to put it, from ‘horizontal’, that is, same-level, structures of mutual dependence
relations intended in the coherentist sense. This, I submit, is a good move for the
defender of OSR. For, it makes sense of the typical structuralist claim that science
30 Haecceitism being exactly the view that two distinct worlds may differ de re without differing
qualitatively, that is, they may differ merely with respect to which entity is which in each of them.
31 In particular, their identities and/or (allegedly) intrinsic, state-independent properties.
matteo morganti 
urges us to drop (at least partially) the hierarchy metaphor in favour of some form of
interconnectedness; and that, in particular, the role played by symmetries, mathemat-
ical groups, permutation invariance, and the like at the theoretical level truly has onto-
logical import. At the same time, the proposed view does not require the structuralist
to substantiate the more contentious claim that physical relations are ontologically
fundamental. In particular, the coherentist point of view is arguably superior to elim-
inative OSR, in that it does not eliminate objects in the problematic attempt to make
physical relations self-standing and fundamental; it is preferable to moderate OSR,
because it establishes a form of mutual dependence between objects, and not between
relations and ‘pseudo-objects’ actually deprived of any intrinsic characterization; and
it is better than contextual identity OSR because it does not put forward a claim
exclusively about identity, and instead allows for more nuanced assertions.32
Let us now look, albeit briefly, at how this translates in practice in physical terms.
The key fact ontic structuralists refer to is that fundamental physical theories tend
to be ‘gauge theories’, that is, theories containing more variables than the actual
number of degrees of freedom of the relevant physical systems, from which the
physically meaningful degrees of freedom are ‘selected’ as invariants under certain
transformations. The relevant transformations are, in particular, those characterizing
mathematical entities known as ‘groups’. Based on this, building up on results dating
back to the work of Wigner, objects are reconceptualized by ontic structuralists “as
representations of symmetry groups, where the symmetries reflect spatiotemporal,
i.e. external, and internal degrees of freedom as well as permutation invariance” (Lyre
2004; 662). The irreducible representations of the relevant mathematical groups, and
ultimately the groups and their internal symmetries themselves, consequently become
ontologically prior to objects. The unitary group SU(2) × U(1) × SU(3), in particular,
is regarded as a fundamental ontological basis due to the fact that it underpins
the Standard Model of elementary particles. There, the ontic structuralist says, the
SU(2) × U(1) group gives rise to those bosons that mediate electromagnetic and
weak interactions (photons and W and Z bosons), and SU(3) determines the bosons
that transmit the strong interactions described by quantum chromodynamics (i.e.
the eight kinds of gluons). This has been extended to fermions and to all allegedly
intrinsic properties: Muller (2009), for instance, points out that the fundamental
symmetry group of quantum mechanics, the Galilei group, determines the admissible
displacements, rotations, and phase transformations; and that, associated to this there
is an algebra of a specific type (a Lie algebra) which has certain invariants (the Casimir
invariants) that immediately correspond to allegedly intrinsic, essential properties of
32 Of course, one may object that this is not OSR at all, as it posits objects, not relations, as the
fundamental ontological items. If this is granted (and I, for one, think that it should be), then metaphysical
coherentism should be regarded as an alternative to OSR that (allegedly) preserves the theoretical advan-
tages of structuralism, rather than as an interpretation of it. In any case, this is not crucial for the present
discussion, as we are not interested in structuralism per se, but rather in coherentism as an alternative to
foundationalism.
 structure of physical reality: beyond foundationalism
particles, such as mass and spin magnitude.33 Mathematical groups being essentially
analysable in terms of relations, OSRists conclude, the foregoing grounds the claim
that relations are ontologically fundamental.
Granting the above at least for the sake of the present discussion, the idea here
is simple. All the physical facts that are pointed at by structuralists, such as those
briefly illustrated a moment ago, do not require an ontological interpretation in terms
of the priority of physical relations over physical objects (even less do they require
the postulation of symmetries, etc. as Platonic entities that are prior to concrete
object-types and tokens). That particle-types and, consequently, the properties of
every particular particle are constrained by the symmetries of the relevant groups
can instead be interpreted in terms of holistic grounding structures of the coherentist
type exclusively involving physical objects. That is, in terms of specific sets of objects
being interrelated and mutually ontologically dependent in such a way that (i) their
joint existence determines the existence of specific structures (of grounding, not of
physical relations!) and (ii) the resulting structures constrain the features of each
particular element of the set. This kind of mutual dependence can be regarded as
constitutive of both qualitative differentiation between different families of particles
and numerical distinction between different tokens, without at any point postulating
anything beyond objects and non-hierarchical dependence relations between them.
This, notice, also allows one to account for state-independent properties that, as men-
tioned earlier, appear not to be amenable to structuralist reduction: for the coherentist
simply doesn’t have to assume that all properties of the mutually dependent objects
derive on such mutual dependence.
Based on our earlier remarks, further specifications may then be added concerning
the relevant grounding structures. Personally, I favour the view that the relations
emphasized by OSRists are transitive, symmetric, and reflexive relations of merely
partial ground. For, the very idea of structure has it that each object is the very object
that it is thanks to the fact that it belongs to the structure it in fact belongs to. Hence,
objects cannot fully ground themselves (if not in a trivial sense, more on this in a
moment). But it is also the case that the dependence also goes in the other direction,
as the structure in question would be different if it did not connect the very entities (or
structure-places) it in fact connects. And this entails that each object in the structure
is part of the ground for what grounds it, in other words, is among its own partial
grounds.34
One-place structures represent an interesting limiting case, in which grounding
relations, if they hold at all, hold trivially exactly in the same sense that it is trivial that
a = a for any a. It seems to me that this ‘collapse’ of ground onto identity suggests
33 We need to limit ourselves to this brief outline here. For more details, see Kantorovich (2003, 2009),
Lyre (2004), Muller (2009), and Roberts (2011).
34 It should be clear that the partial self-grounding in question is far from being a trivial fact that only
corresponds to uninformative claims that fail to truly explain. Exactly the opposite seems to be the case.
matteo morganti 
that no real explanation is forthcoming in those cases. For this reason, I would opt
for quasi-reflexivity (see earlier in this section) here, and claim that exactly because
the structure is a one-object structure, the object constituting it is not structurally
grounded, but instead possesses the relevant features primitively. Indeed, it seems
to me that a lone object cannot be informatively said, for example, to possess the
identity that it has because it exists (and no other reason), and it is more appropriate
to postulate primitive intrinsic identities and qualities in those cases. OSRists are of
course free, and likely, to disagree with this. Notice, however that a non-negligible
consequence of such a disagreement would be that they should accept the reflexivity
of full ground.35
Be this as it may, the foregoing case study shows that coherentism might be
preferable in metaphysical scenarios where a significant form of non-hierarchical,
most probably non-mereological, holism is in play, so that everything is intimately
related to everything else with respect to identity and/or essential properties.
5 ‘Hybrid’ Views
Summing up, infinitism and coherentism both abandon the idea that chains of ground
must terminate in a foundation, but do so in rather different ways: infinitism sticks
to the idea of grounding chains as strict partial orders, while coherentism questions
the very idea of a hierarchy of levels of reality. However, there are also elements
of similarity—interestingly mirrored, as the reader can see, by analogies that exist
between the S-matrix case and the group-theoretic structuralist view of physical real-
ity.36 For instance, especially against the background of what we called the emergence
model of being—whereby the grounded gains its existence/identity from the whole
series of grounding entities and at no point is anything ‘transmitted’ in its entirety
from the grounding to the grounded—both infinitism and coherentism can be said
to oppose foundationalism by suggesting a holistic view according to which it is the
relevant structure taken in its entirety that determines the existence/essence/identity
of what is grounded. Also, coherentism is compatible with the idea of a hierarchy of
levels each one of which is characterized by specific mutual dependence relations (for
an application of this idea to group-theoretic ontic structural realism, see Roberts
(2011)). And, of course, such a hierarchy need not be well-founded. This is no
doubt interesting and rather unexplored terrain, and invites one to undertake further
work on the relationship between infinitism and coherentism.37 More generally, the
investigation of more articulated views on metaphysical structure, whereby one model
35 Of course there is room for discussion here, but it is not essential for us to reach a final verdict on this.
36 In both cases, particles are symmetrically dependent on other particles both at the token- and the
type-level.
37 Here, the analogy with the debate about justification proves useful again. For an interesting discussion
of the complex relationship between foundationalism, infinitism, and coherentism in epistemology, and in
particular of the similarities between infinitism and coherentism, see Herzberg (2014).
 structure of physical reality: beyond foundationalism
does not necessarily rule out the other, promises to be of interest. Why not think that
some aspects of reality have a structure of one type, and others a different one?
The variety of case studies examined in this paper constitutes just one hint going
in this direction: perhaps the universe has no ultimate level and possesses instead an
infinitist mereological structure both in the direction of the small and of the large; and
perhaps, at the same time, it exhibits an essential non-mereological holistic structure
when it comes to the nomological constraints determining the possible typologies
of (some of) the objects that inhabit it. In connection to this, it is perhaps useful
to point out that the sort of pluralism just envisaged is by no means harmful for
the idea that grounding is a useful philosophical concept. For, exactly in the same
way in which, as mentioned earlier, a non-monolithic notion of grounding is ok,
provided that one is truly dealing with a non-causal relation with sufficiently well-
defined general features and clear explanatory power, one should have no problem
with a multifaceted account of the structure of reality, provided that, by endorsing it,
one obtains good explanations for the domains of things one aims to account for. After
all, to repeat, why should reality possess a uniform, all-encompassing metaphysical
structure just because (maybe) we would like it to?
Be this as it may, as stated in the introduction the present paper was intended
to offer an illustration of potentially fruitful interactions between science and meta-
physics rather than arguments as to the true structure of reality. Hopefully, more work
dealing with the uncharted domain of non-foundationalist metaphysical structures
will be done in the near future.38
References
Bliss, R.L. (2013). Viciousness and the Structure of Reality, Philosophical Studies, 166, 399–418.
Bliss, R.L. and Trogdon, K. (2014). Metaphysical Grounding, in E.N. Zalta (ed.), The Stan-
ford Encyclopedia of Philosophy (Winter 2014 Edition), http://plato.stanford.edu/archives/
win2014/entries/grounding/.
Bohn, E.D. (2009). Must There Be a Top Level?, Philosophical Quarterly, 59, 193–201.
Caulton, A. (2015). Is Mereology Empirical? Composition for Fermions, in C. Wüthrich and
T. Bigaj (eds), Metaphysics in Contemporary Physics, Poznan Studies in the Philosophy of the
Sciences and the Humanities, Amsterdam, Rodopi, 293–321.
Connes, A. (1996). Gravity Coupled with Matter and Foundation of Noncommutative Geom-
etry, Communications in Mathematical Physics, 182, 155–76.
Correia, F. (2014). Logical Grounds, The Review of Symbolic Logic, 7, 31–59.
Dehmelt, H. (1989). Triton, . . . Electron, . . . Cosmon, . . .: An Infinite Regression?, Proceedings
of the National Academy of Sciences, 86, 8618–9.
38 I thank audiences in Groningen, Helsinki, Neuchatel, Padua, Paris, Rome, and Turin for their feedback
on material related to that presented in this paper. I am also grateful to Ricki Bliss and Graham Priest for
their invitation to contribute to this volume.
matteo morganti 
Feynman, R.P. (1994). Six Easy Pieces: Essentials of Physics Explained by its Most Brilliant
Teacher, Reading, MA, Perseus Books.
French, S. (2014). The Structure of the World, Oxford, Oxford University Press.
Griffith, A.M. (2014). Truthmaking and Grounding, Inquiry, 57, 196–215.
Guth, A. (2007). Eternal Inflation and its Implications, Journal of Physics A: Mathematical and
Theoretical, 40, 6811–26.
Harrison, E. (1987). Darkness at Night: A Riddle of the Universe, Cambridge, MA, Harvard
University Press.
Healey, R. (2013). Physical Composition, Studies in History and Philosophy of Modern Physics,
44, 48–62.
Herzberg, F. (2014). The Dialectics of Infinitism and Coherentism: Inferential Justification
versus Holism and Coherence, Synthese, 191, 701–23.
Jenkins, C. (2011). Is Metaphysical Dependence Irreflexive? The Monist, 94, 267–76.
Kantorovich, A. (2003). The Priority of Internal Symmetries in Particle Physics, Studies in
History and Philosophy of Modern Physics, 34, 651–75.
Kantorovich, A., (2009). Ontic Structuralism and the Symmetries of Particle Physics, Journal
for General Philosophy of Science, 40, 73–84.
Litland, J.E. (2013). On Some Counterexamples to the Transitivity of Grounding, Essays in
Philosophy, 14, 19–32.
Lyre, H. (2004). Holism and Structuralism in U(1) Gauge Theory, Studies in History and
Philosophy of Modern Physics, 35, 643–70.
McKenzie, K. (2011). Arguing Against Fundamentality, Studies in History and Philosophy of
Modern Physics, 42, 244–55.
McKenzie, K. (2014). Priority and Particle Physics: Ontic Structural Realism as a Fundamen-
tality Thesis, British Journal for the Philosophy of Science, 65, 353–80.
Morganti, M. (2011). Is There a Compelling Argument for Ontic Structural Realism?, Philoso-
phy of Science, 78(5), 1165–76.
Morganti, M. (2014). Metaphysical Infinitism and the Regress of Being, Metaphilosophy, 45,
232–44.
Muller, F. (2009). Withering Away, Weakly, Synthese, 180, 223–33.
Nottale, L. (1993). Fractal Space-time and Microphysics, Singapore, World Scientific Press.
Pietronero, L. (1987). The Fractal Structure of the Universe: Correlations of Galaxies and
Clusters, Physica A: Statistical Mechanics and its Applications, 144, 257–84.
Poplawski, N.J. (2010). Cosmology with Torsion: An Alternative to Cosmic Inflation, Physics
Letters B, 694(3), 181–5.
Rabin, G.O. and Rabern, B. (2016). Well Founding Grounding Grounding, Journal of Philo-
sophical Logic, 45(4), 349–79.
Roberts, B.W. (2011). Group Structural Realism, British Journal for the Philosophy of Science,
62, 47–69.
Rodriguez-Pereyra, G. (2015). Grounding is Not a Strict Order, Journal of the American
Philosophical Association, 1(3), 517–34.
Rovelli, C. (2016). Reality Is Not What It Seems: The Journey to Quantum Gravity, London,
Penguin Books.
Schaffer, J. (2003). Is There a Fundamental Level?, Noûs, 37, 498–517.
Schaffer, J. (2012). Grounding, Transitivity, and Contrastivity. In F. Correia and B. Schnieder
(eds), Metaphysical Grounding Cambridge, Cambridge University Press, pp. 122–38.
 structure of physical reality: beyond foundationalism
Scott Dixon, T. (2016). What is the Well-Foundedness of Grounding?, Mind, 125, 439–68.
Sen, A. (1969). Quasi-Transitivity, Rational Choice and Collective Decisions, Review of Eco-
nomic Studies, 36, 381–93.
Smith, Q. (1990). A Natural Explanation of the Existence and Laws of Our Universe, Aus-
tralasian Journal of Philosophy, 68, 22–43.
Smolin, L. (2013). Time Reborn: From the Crisis in Physics to the Future of the Universe, Toronto,
Alfred A. Knopf.
Tahko, T.E. (2014). Boring Infinite Descent, Metaphilosophy, 45, 257–69.
Tahko, T.E. and Lowe, E.J. (2015). Ontological Dependence, in E.N. Zalta (ed.), The Stan-
ford Encyclopedia of Philosophy (Spring 2015 Edition), http://plato.stanford.edu/archives/
spr2015/entries/dependence-ontological/.
Thompson, N. (2016). Metaphysical Interdependence, in M. Jago (ed.), Reality Making, Oxford:
Oxford University Press, 38–56.
Wilson, J. (2014). No Work for a Theory of Grounding, Inquiry, 57, 535–79.
Wolff, J. (2012). Do Objects Depend on Structures?, British Journal for the Philosophy of Science,
63, 607–25.
PA R T III
The Contingency
and Consistency Theses
14
On Shaky Ground? Exploring the
Contingent Fundamentality Thesis
Nathan Wildman
The past decade and a half has seen an absolute explosion of literature discussing the
structure of reality. One particular focus here has been on the fundamental. However,
while there has been extensive discussion, numerous fundamental questions about
fundamentality have not been touched upon. In this paper, I focus on one such lacuna
that emerges when we consider the interaction between fundamentality and modality,
about the modal strength of fundamentality. More specifically, I am interested in
exploring the idea that the fundamentalia are only contingently fundamental—or,
put in property-terms, that the property of being fundamental is not a (weakly)
necessary property (where a property is weakly necessary iff the things that have the
property do so in every world in which they exist).1 Call this claim the contingent
fundamentality thesis. While I think this thesis is plausible—indeed, as I show later, it
lurks in the unexamined shadows/assumptions of some fairly prominent positions—
as far as I can tell, nothing has been said about it. Here, I hope to fix this by giving the
thesis a proper airing. In this way, this paper represents a first pass at exploring not
only the modal status of fundamentality, but also offers a starting point for examining
broader issues about the relationship between fundamentality and modality.
In particular, after fixing some preliminaries in Section 1, I’ll discuss in Section 2
three reasons for taking the contingent fundamentality thesis seriously. I then evaluate
some objections in Section 3 intended to show that taking fundamentality to be
contingent is wrong-headed; I argue that these objections can be dealt with, leaving
the contingent fundamentality thesis at least prima facie plausible. In Section 4, I then
look at how the thesis relates to views about the possibility of contingently existing
fundamentalia, pulling some of the various packages apart, and making the case for
adopting what I call the Shifty Shaky view. I then conclude in Section 5 by indicating
further areas about the thesis ripe for fruitful future exploration.
1
For more on weak necessity, see Kripke (1971), Davies (1981), and Wildman (ms).
 exploring the contingent fundamentality thesis
To be clear: the main aim of this paper isn’t to necessarily convince readers of the
truth of the contingent fundamentality thesis. Rather, I want to use this opportunity
to explore the thesis—look at some reasons for thinking it true, some objections
to it, and evaluate how it relates to other debates about the nature of fundamentality.
In so doing, I hope to prompt further discussion—not only about the contingent fun-
damentality thesis, but about the relationship between modality and fundamentality
more generally.
1 Preliminaries and Clarifications

A good place to start is by clarifying what exactly the contingent fundamentality
thesis is. As I understand it, the thesis isn’t meant to be the claim that every fundamen-
tal entity is contingently fundamental. This is too strong, as it seems there are some
entities that are necessarily fundamental if fundamental at all—for example, assuming
that the nul set is fundamental, then it looks necessarily so. Similarly, if God exists,
then she will likely be necessarily fundamental.
Relatedly, the thesis shouldn’t be necessitated, as in:
N-CFT Necessarily, for some x, x is fundamental and possibly, x exists and is not
fundamental.2
This also seems obviously false, since it’s extremely plausible that everything that’s fun-
damental in an abstracta only world is necessarily fundamental; similarly for lonely
God worlds—everything that’s fundamental there looks necessarily fundamental.
For these reasons, the proper formulation of the contingent fundamentality thesis
is a possibility claim, along the lines of:
CFT Possibly, for some x, x is fundamental and possibly, x exists and is not
fundamental.
This gets at the main point—that is, that being fundamental isn’t a weakly nec-
essary property. Further, it also allows the contingentist a bit of wiggle room, in
that, even if it turns out that all the actual fundamentalia are (weakly) necessar-
ily fundamental, something like the contingent fundamentality thesis can still be
true, provided there’s something out there in modal space which is contingently
fundamental.
A second point that must be addressed concerns clarifying the metaphysical notion
of fundamentality. As it happens, this is a fairly difficult task, for there are about as
many different conceptions of fundamentality as there are authors writing about it—
and there are a lot of authors.
That said, the standard notion—or at least the one that most of the literature
works with—roughly defines fundamentality in terms of metaphysical grounding.
2 If you think that fundamentality only applies to facts, then replace ‘exists’ with ‘obtains’, and if you
think properties can be fundamental, then add ‘is instantiated’.

nathan wildman 
According to this standard conception, x is fundamental iff x exists/obtains and

there is no y such that y grounds x. This account of fundamentality is discussed
by, for example, Audi (2012), Correia and Schnieder (2012), deRossett (2010), and
Rosen (2010).3
A close, albeit distinct characterization comes from Schaffer (2009, 2010a, 2010b,
2010c, 2013), who also defines fundamentality in terms of ‘grounding’, but might be
better described as defining it in terms of ontological dependence. For Schaffer, x is
fundamental iff x exists and there is no y such that x depends on y.4
I take it that something along one of these two lines is what most have in mind
when they talk of fundamentality. Yet these two certainly don’t exhaust the options! In
fact, fundamentality has also been characterized in terms of: the ‘familiar theological
metaphor: the fundamental entities are all and only those entities which God needs
to create in order to make the world how it is’ (Barnes 2012: 484); truthmakers
(Cameron 2008, Heil 2003); structure (Sider 2011); what holds ‘in reality’ (Fine 2001);
metaphysical explanation (Jenkins 2013); and the ‘ineliminable’ (Raven 2015). It’s also
been taken as a primitive, hyperintentional notion (Wilson 2014, 2016).5
Given this plethora of options, settling upon the ‘right’ conception of fundamental-
ity looks like a daunting task, and one that, to be frank, falls well beyond the scope of
this paper. Thankfully, little of what I go on to say hinges upon adopting any particular
conception—from what I can tell, the following discussion remains much the same
regardless of how you understand fundamentality. With that in mind, I’ll make do
with a rough-and-ready characterization, which underlies both the standard and the
Schafferian conceptions, that x is fundamental when it sits at the bottom of a hierarchy
of ground/dependence, ‘grounding’ derivative things but being itself ungrounded.
This is the notion of fundamentality that I’ll be working with. And this isn’t because
I think this is the only viable conception or account of fundamentality—far from
it! Rather, I hope that in formulating my characterization of fundamentality in such
generic terms, I leave sufficient space for those with more specific notions to engage
with the following discussion, modifying and adapting the details as required.
A third matter that must be addressed concerns what sorts of entities are (eligible
to be) fundamental. And, much like with our previous question, lots of answers
have been given; specifically, metaphysicians have suggested that ‘is fundamental’
can apply to truths (Sider 2011, Williams 2010), facts (Audi 2012, deRossett 2010,
Fine 2012, Raven 2015, Rosen 2010), states of affairs (Armstrong 1997), concrete
3 This account has also been criticized for being too broad, since there might be some entities that are
ungrounded but not properly fundamental; see e.g. Dasgupta (2014a, 2014b).
4 See e.g. Steinberg (2015) for a discussion of how Schaffer’s idiosyncratic notion of ‘ground’ relates to
both standard notions of ground and ontological dependence.

5 In fact, Wilson invokes several of these conceptions when she spells out her primitive notion: ‘The
fundamental is, well, fundamental: entities in a fundamental base play a role analogous to axioms in a
theory—they are basic, they are ‘all God had to do, or create’. As such—again, like axioms in a theory—
the fundamental should not be metaphysically defined in any other terms, whether these be positive or
negative’ (2014: 560).
 exploring the contingent fundamentality thesis
and abstract objects (Cameron 2008, Schaffer 2009, 2010c), properties and relations
(Armstrong 1997, Lewis 1986), and logical operators/quantifiers (Sider 2011). In fact,
one of the major differences between the standard and the Schafferian conceptions
of fundamentality concerns what entities can be fundamental. Schaffer allows his
notion of ‘grounding’ to apply to entities of any category—for example, material
objects, facts, properties, and events—that are at the bottom of his partially ordered
hierarchy of being. Meanwhile, those who favour the standard conception tend to split
Schaffer’s hierarchy into one characterized by ontological dependence, which relates
entities of any kind but doesn’t necessarily result in a partial ordering, and a hierarchy
characterized by metaphysical grounding, which mono-categorically applies to facts,
and generates a partial ordering thereof.6 On this conception, ‘is fundamental’ pri-
marily applies to facts—specifically, those facts that sit at the bottom of the grounding
hierarchy, serving as grounds for other facts, but which are not grounded in anything.
Entities from other categories are then said to be only derivatively fundamental, in
that they feature in fundamental facts (alternatively, if they feature in fundamental
facts of the form [x exists]).
While settling what the exact range of possible satisfiers for the ‘is fundamental’
predicate are is important, for present concerns I would prefer to remain as ecu-
menical as possible; in particular, I’d like to avoid making any clearly controver-
sial assumptions about which sort of fundamentality statements are (or are not!)
acceptable. Thankfully, we can skirt around the issue (and avoid stepping on anyone’s
toes) by employing some terminological place-holders. Thus, when talking about
the fundamentalia—be they facts, things, properties, or what have you—I’ll employ
category neutral terms like ‘entity’. This allows those of us who are undecided about
what kinds are eligible for being fundamental to continue discussing fundamentality’s
modal strength, while also leaving ample room for those who have a horse in the
race to slot in their preferred fundamental kinds where applicable. Further, when it
matters, I’ll try to make it clear how to extend/apply the relevant point to whatever
category is desired.7
One final point that must be discussed concerns whether being fundamental comes
in degrees. Some have argued that entities can be more or less fundamental (and
similarly more or less derivative). In contrast, others favour a ‘fundamentalist’ view,
according to which fundamentality is instead an all-or-nothing thing.8 Which con-
ception one favours—either fundamentality as quantitative or as non-quantitative—
will entail reading the central question in a slightly different manner. First, if we
understand fundamentality as purely binary (thus we adopt the latter, ‘fundamentalist’
view), then we can read the following discussion as concerning whether entities that
6 See e.g. Correia (2005), Fine (1995), Lowe (2010), and Schnieder (2006) on ontological dependence,
and Audi (2012), Fine (2012), Raven (2015), and Rosen (2010) on grounding.
7 See fn. 2.
8 See Barnes (2012: 875–9) for a useful comparison of the two views, and McDaniel (2013) for an
excellent discussion of degrees of being.

nathan wildman 
are fundamental are necessarily so. But if we allow for degrees of fundamentality,
then we’ve options: one can understand the matter as concerning all entities and
their respective degrees of fundamentality (e.g. supposing that x is fundamental to
degree n, could x be more or less fundamental?). Alternatively, one can focus just
on the absolutely fundamental entities—those entities that possess fundamentality to
the highest possible degree—and read the following as concerning whether these elite
entities are necessarily elite (e.g. suppose that x is absolutely fundamental, could it be
non-absolutely fundamental?). More generally, regardless of what one thinks about
degrees of fundamentality, there’s a sensible version of the contingent fundamentality
thesis in the vicinity.
So, we’ve said something about how best to formulate the contingent fundamen-
tality thesis, what fundamentality is, what kinds of things are eligible for being
fundamental, and what fundamentality is like (in particular, whether it comes in
degrees). And, as I hope I’ve made clear, I am here trying to stay as inclusive as
possible. This is partially because I want to avoid any unnecessary disagreements. But,
more importantly, this is also because I think that, regardless of one’s position on these
issues, the question of fundamentality’s modal strength is one that’s worth exploring.
This is a (radically underexplored) point that everyone should say something about;
that alone justifies adopting as neutral a position as possible.
2 For Contingent Fundamentality?

For all that, we might wonder why we should care about the contingent fundamen-
tality thesis. What reasons are there for taking this particular view about the modal
strength of fundamentality seriously?
The first reason for taking contingent fundamentality seriously is that structurally
similar claims have been offered in other debates. For example, priority presentism is
the view according to which (i) only present entities fundamentally exist, and (ii) past
and future entities exist, but are grounded in present entities (Baron 2015, Lopez de
Sa ms). So, on this view, Caesar might have been fundamental at the moment when he
crossed the Rubicon, but he is derivative now.9 This temporally variable fundamental-
ity thesis closely parallels the contingent fundamentality thesis.10 Similarly, several—
in particular, Cameron (2007), Miller (2009), and Nolan (2005)—have argued for
the contingency of composition, a view that nicely complements the contingent
fundamentality thesis.
9 Or, in terms of facts, [Caesar crossed the Rubicon] was, but is no longer, fundamental. Admittedly,
while a rather important man, it’s not likely that Caesar is (or was) a fundamental entity. But this doesn’t
take away from the point that priority presentism is committed to there being a host of entities which are
fundamental at one time and derivative at another.
10 Further, if we assume that individual moments correspond to possible worlds, priority presentism
entails the contingent fundamentality thesis.

 exploring the contingent fundamentality thesis
A second reason is that some individuals have apparently committed themselves

to it. Specifically, priority monism is the view according to which the cosmos, the
whole all concrete objects are part of, is the one and only concrete object that is not
grounded in (or depends upon) any other object. This isn’t to say that the smaller
parts don’t exist—they do! However, they do so derivatively, being grounded in the
fundamental cosmos. Thus, in a quip, priority monism says that priority goes from
(biggest) whole to part, not the other way around.11
In a recent discussion, Schaffer has suggested that the priority monist can allow
for big-to-bigger embedding: take the cosmos c, which is actually fundamental,
and embed it in a larger whole, c’. In the embedded worlds, c is grounded in c’.
Consequently, things can cross the ‘categorical divide’ between ‘fundamental and
dependent’ (2013: 81)—that is, being fundamental is, at least for cosmoi, a merely
contingent property. Interestingly, shrinking cases also deliver the same result: take r,
the sub-region of the cosmos that corresponds to my dog Ohle. In the actual world,
r is derivative. However, there are (quite small!) worlds where r exhausts the whole
of the cosmos. And, in these worlds, since there is nothing larger than r—that is,
nothing that r is a proper part of—r is itself fundamental. Thus we’ve another case
where fundamentality is a merely contingent property.
And there’s no problem squaring Schaffer’s apparent commitment to contingent
fundamentality with his explicit endorsement of the necessity of monism (see e.g.
Schaffer 2010c: 56), provided we read the necessity of monism as being phrased in de
dicto, not de re, terms—that is, every possible world w is such that, at w, the cosmos at
w is fundamental, where ‘the cosmos’ is here used as a non-rigid definite description,
picking out whatever thing happens to be the cosmos at w. This is perfectly compatible
with some particular object a being the cosmos at w (and hence being fundamental
at w ) but being a mere sub-region of the cosmos in (and hence derivative at) w .
Of course, priority monism does not entail contingent fundamentality—a priority
monist could readily deny contingent fundamentality, perhaps because they endorse
a kind of mereological essentialism, such that objects essentially stand in the mereo-
logical relations that they in fact do. This would block embedding cases—c essentially
isn’t a proper part of anything, so whatever exists in the embedded world isn’t c—as
well as shrinking cases—r essentially is a proper part of the cosmos, so the entity that
is the mini-cosmos in the shrunk worlds isn’t r. But that doesn’t take away from the
fact that the priority monist—Schaffer—seems to endorse contingent fundamentality,
which is enough to indicate that it’s worthy of consideration.
Interestingly, the contingency of fundamentality is also compatible with priority
pluralism, according to which there are at least two fundamental entities, neither of
which are the cosmos. In fact, it can help the pluralist handle de re modal objections
11 Schaffer (2010c: 44) also claims priority monism is compatible with priority relations going from the
cosmos to mereological atoms, then upwards from there to all the non-cosmos wholes; Steinberg (2015),
however, raises some potent objections to this.
nathan wildman 
to their position. Suppose that Quarky is one of the actual fundamental entities.
Assume that gunky worlds seem possible—that is, it’s possible that the material
world is such that every object has proper parts. In the gunky worlds, Quarky looks
derivative, grounded in its (infinite series of) proper parts. So, goes the objection,
because Quarky fails to be fundamental in the gunky worlds, Quarky isn’t actually
fundamental after all. However, if fundamentality is contingent, this argument doesn’t
go through: that Quarky isn’t fundamental in gunky worlds says nothing about its
fundamental status in the actual world.12
This highlights a third reason for taking contingent fundamentality seriously: the
thesis might prove useful, in the sense that it can be helpful when it comes to deflating
or avoiding objections to various positions, as in the previous paragraph.
Taken together, this trio motivate at least giving the contingent fundamentality
thesis the time of day. But that’s not to say that we’ve said enough to secure its truth!
After all, there might be good reasons for dismissing it as implausible. With that
in mind, Section 3 looks at several possible objections one might raise against the
contingent fundamentality thesis.
3 Fundamentally Mistaken about Contingent

Fundamentality?
The first objection to contingent fundamentality is that priority claims are metaphys-
ical theses, and metaphysical theses are, if true at all (metaphysically) necessarily
true.13 Obviously, this would render the contingency of fundamentality a non-starter.
However, there’s no reason to think that all metaphysical claims have such modal
force. Perhaps the most famous example of a metaphysical thesis that is only intended
to be contingently true is Lewis’s doctrine of Humean Supervenience. In this sense,
the would-be contingent fundamentalist is in good company.14
A more potent objection comes from the tight link between the structuring rela-
tions that characterize fundamentality and essence. For example, one might think that
Socrates’s existence grounds the existence of {Socrates}, and that, if so, this explanatory
link must somehow emerge from {Socrates}’s essence; it should, as Fine puts it, ‘be part
of the nature of singleton Socrates that its existence is determined in this way from
the existence of Socrates’ (2015: 279). More generally, it’s very plausible to say that it
12 Cameron (2007: 13) makes a similar point, but in a different context. Further, Schaffer seems to suggest
that pluralists should take fundamentality to be contingent, when he says that ‘the pluralist who treats, say, a
given electron as [fundamental] can grant that it may be divisible into small constituents, and then it would
no longer (by her lights) be [fundamental]’ (2013: 81). One interesting area I hope to explore in future work
is how contingent fundamentality relates to Schaffer’s (2010c) modal objection to priority presentism.
13 See e.g. Schaffer (2010c: 56; 2013: 84) and van Inwagen (2002: 28).
14 A related version of the objection says that the fundamental claims are necessary because they’re a
priori. However, I take it that not all fundamentality claims are a priori—empirical investigation certainly
plays a role in determining what is or is not fundamental. Further, as Kripke (1980) and Evans (1985) have
shown, that something is a priori doesn’t mean it’s necessarily true.
 exploring the contingent fundamentality thesis
is part of the essence of derivative entities that they depend upon whatever it is they
in fact do derive from.15
But this seems to conflict with contingent fundamentality. For suppose that Tommy
the Atom is in fact fundamental, but only contingently so—possibly, Tommy’s exist-
ence is grounded in the existence of Quarky. If we grant the ground–essence con-
nection, then Tommy essentially is grounded in Quarky. And, given that essentiality
entails necessity, it follows that Tommy is necessarily so grounded. Yet, if Tommy is
necessarily grounded in Quarky, he can’t be fundamental since being non-derivative
is, at minimum, a necessary condition for being so. Consequently, assuming the
ground–essence connection, it looks like we must reject contingent fundamentality.
An immediate contingentist response is to simply reject the ground–essence con-
nection; however, such connections look extremely plausible, and it would be better
to avoid such drastic measures if at all possible.16 With that in mind, a more measured
response is preferable. Is there one?
What’s problematic for the contingentist is if the relational property being derivative
from Quarky is essential to Tommy; for this would entail that, in every world where
Tommy exists, he’s derivative (and hence not fundamental). But what wouldn’t be a
problem is if Tommy was essentially such that, if Quarky exists, Tommy is derivative
from it. Tommy’s essentially—and hence necessarily—possessing this conditional
property doesn’t prevent him from being fundamental in worlds where Quarky isn’t
around.17 It does preclude Tommy being fundamental in worlds where Quarky also
exists, but that’s fine—all we need to protect contingent fundamentality is a world
where Tommy exists and is fundamental!
More generally, the contingentist can accommodate the ground–essence link by
conditionalizing the relevant properties, such that x essentially is such that if y exists,
then x depends upon it. Possessing these conditional essential properties preserves the
ground–essence link (it’s still part of Tommy’s essence that he’s derivative from Quarky
whenever it’s around), but is also perfectly compatible with x being fundamental in
worlds where y doesn’t exist.18 The upshot is that the essence–ground connection
alone doesn’t tell against contingent fundamentality.19
Admittedly, this conditionalizing move won’t be available to those who think that
(i) the fundamental entities must be ontologically independent, and (ii) that an entity
is ontologically dependent on whatever appears in facts about its essence. However,
15 Trogdon (2013) also argues for the grounding–essence link.

16 Though cf. Fine (2012), Rosen (2010), and Skiles (2015) for some doubts about the essence–ground
link.
17 Note that the use of conditional relational properties is just for convenience; the same point can be
made using only relations. This ensures that sparse modalists about essence can make use of this manoeuvre
too. For more on sparse modalism about essence, see Wildman (2013, 2016).
18 This existential conditionalization move isn’t available for someone who doesn’t allow for contingently
existing fundamentalia, like either of the two ‘Fixed’ views discussed in Section 4.
19 A similar move blocks an objection from the internality of dependence/grounding: if we condition-
alize the relevant dependence/grounding claims to those worlds where both the grounds and the ground
exist, then we can preserve internality while allowing for contingent fundamentality.
nathan wildman 
I see little reason to accept the latter—instead, I’d suggest that an entity ontologically
depends on those things that both appear in facts about the entity’s essence and exist.
This leaves room for contingent dependence (and hence contingent fundamentality),
and is, I think, a close, contingentist friendly, surrogate for (ii).
There is one potential drawback to this conditionalizing move: it might turn out
that it’s part of Tommy’s essence that, if the world was populated with some weird
ectoplasmic goo, then Tommy is derivative from it.20 But, one might sensibly object,
this simply isn’t part of Tommy’s essence—Tommy is in no way related to this
weird goo! However, note that we only said this might be part of Tommy’s essence.
Nothing said so far commits the conditionalizer to gooey-Tommy worlds, and hence
to including such a property in Tommy’s essence (indeed, we might take our vitriolic
response as a good reason for thinking there are no such worlds). In other words,
not all conditional properties will make it into Tommy’s essence. Of course, we’ll
need some reason for including/excluding the relevant ones, but that’s a task at least
partially to be determined by our epistemology of essence. And, until we can show that
no conditional properties make it in, there’s space to maintain the conditionalizing
reply. So it seems we can preserve the ground–essence link and still be contingentists
about fundamentality.
A third objection to contingent fundamentality concerns negative explanations.21
Grant, for the sake of argument, that tables are derivative, but possibly fundamental.
Now, go to any world where tables don’t exist. What explains the fact that there are
no tables? The standard answer is something like the fact that there are no table-wise
arrangements of simples. But, given that tables are possibly fundamental, this isn’t a
complete explanation—after all, it might be that, even though there are no table-wise-
arranged simples, some fundamental tables exist. Consequently, given contingent
fundamentality, our explanations for the non-existence of entities are compromised.
One line of reply is to say that it’s part of what it is to be a table that it be composed of
tablewise-arranged simples—so, every table-world is also a table-arranged-simples-
worlds—but still maintain that tables are possibly fundamental—that is, there are
some worlds where tables exist and are fundamental, along with some worlds where
tables are derivative. This would block the above objection, but at the cost of denying
the idea that mereological structure mirrors priority structure, since it allows for
mereologically complex objects to be fundamental.22
An alternative response is to say that this is a specific instance of the more general
problem of explaining negative facts, which is a complicated matter for anyone,
contingentist or not. This looks enough to take the wind out of the objection’s sails,
leaving it a problem—we need to say something to explain why there are no tables—
but not a problem particular to contingent fundamentality.
20
Thanks to Alex Skiles for useful discussion on this and the following point.
21 Thanks to Tobias Wilsch for helpful discussion regarding this objection.
22 Such a move would likely appeal to priority monists.
 exploring the contingent fundamentality thesis
We can make the final objection clearest via the theological metaphor. When
she makes the world, what happens is that God sits down and then brings the
fundamentalia into existence—everything then emerges from this fundamental basis.
And, included within this derivative explosion are the modal facts. But, if the modal
facts emerge from the fundamentalia, then it’s nonsense to ask about the modal status
of the fundamenta—the fundamental comes before modal matters even get going.
On this conception, the fundamental are like Dasgupta (2014b)’s autonomous facts:
they are the scaffolding that supports everything else. Or, to draw another metaphor,
the fundamental are like axioms in a theory; asking about their modal status within
said theory is to fail to understand the role they play in fixing what is or is not true
therein. Thus the objection is that the notion of contingent fundamentality rests upon
a complete misunderstanding of how fundamentality works.
I must confess that I have a hard time wrapping my head around this objection.
For one, those who are happy to advance it must commit themselves to modal gaps—
that is, there are certain facts/propositions/things which are entirely amodal. That is,
for lack of a better term, really weird. Further, it seems like we can make sense of
contingent fundamentality: I can ask, of some object that is actually fundamental,
‘could this thing have been derivative?’ Nothing seems to have misfired there. For
these reasons, I’m happy to set this objection aside, and progress onwards with
exploring contingent fundamentality. To steal a line from Luther Ingram, if doing so
is wrong, I don’t want to be right.
The general upshot of this section is that the notion of contingent fundamentality
doesn’t look entirely off base. Admittedly, the above represent just a small sampling of
the various objections one might offer against the idea. But their dismissal, combined
with the points made in Section 2, give us a prima facie motivation for adopting
the position.
With that in mind, Section 4 examines how contingent fundamentality interacts
with another point of intersection between modality and fundamentality: namely, the
matter of whether contingently existing (or obtaining) entities can be fundamental.
4 Contingent Fundamentality and Contingent

Fundamentalia: Shifting and Shaking
The necessitarian thesis says that the things that are fundamental necessarily exist—
that is, for all x, if x is fundamental, then x necessarily exists. In contrast, the
contingentist thesis says that at least some of the fundamentalia contingently exist—
that is, for some x, x is fundamental and possibly, x does not exist.23 So the necessi-
tarian thesis makes necessary existence a necessary, but not sufficient, condition for
23 These theses are phrased in terms of existence, but they can be readily altered to accommodate those
who, like Williamson (2013), think that all objects necessarily exist by e.g. changing ‘exists’ to ‘is concrete’.
nathan wildman 
being fundamental, while the contingentist thesis allows for some necessarily existing
fundamentalia; it just denies that necessary existence is a necessary condition for
being fundamental.
So, how does commitment to either the necessitarian or the contingentist thesis
interact with the contingent fundamentality thesis? There are four possible combina-
tions. The first holds that the fundamentalia consist entirely of necessary existents,
and that the fundamental necessarily existing entities are necessarily fundamental.
In other words, this position accepts the necessitarian thesis, but denies contingent
fundamentality. The second package also accepts that only necessary existents can be
fundamental, but denies that being fundamental is a necessary property. Thus this
view accepts the necessitarian thesis, but also accepts the contingent fundamentality
thesis. Meanwhile, the final pair of packages both allow for contingently existing
fundamentalia—that is, they accept the contingentist thesis—though they differ with
regards to contingent fundamentality. The first denies contingent fundamentality,
and holds that while the fundamentalita (can) include contingent existents, what-
ever is fundamental is so in every world in which it exists. In contrast, our final
package agrees that some fundamentalia contingently exist, but also claims that the
grounding structure of the world is similarly fundamental—that is, this position
endorses the contingent fundamentality thesis, and takes being fundamental to be a
non-necessary property.
We can make the various positions a bit clearer by thinking about them as four
options emerging from the two distinctions. To apply some labels: the first distinction
is between fixed vs shifty foundations—the fixer says that there aren’t any contingent
fundamentalia, the shifter says there are. Meanwhile, the second is between firm
vs shaky foundations—the firmer says fundmentality is weakly necessary, while the
shaker says it isn’t. We can represent the various combinations with the following table.
Fixed Firm Fixed Shaky

Necessary fundamentalia that are nec- Necessary fundamentalia that are con-
essarily fundamental tingently fundamental
Shifty Firm Shifty Shaky
Contingent fundamentalia that are Contingent fundamentalia that are
(weakly) necessarily fundamental only contingently fundamental
Are there any reasons to prefer any one of these package views over the others? Let’s
start with Fixed Firm, which adopts the necessitarian thesis but denies the contingent
fundamentality thesis. As far as I can tell, there’s nothing logically inconsistent
about this view. However, that’s damning with faint praise, for the view faces some
major problems. In particular, it is hard pressed to account for modal variation: if
the fundamental entities are the same in every world—which is what this view is
 exploring the contingent fundamentality thesis
committed to—how can the various worlds differ concerning what exists/what these
things are like?24 Suppose that A and B are the fundamental entities. On this view,
they both necessarily exist and are necessarily fundamental. So, they exist and serve
as the metaphysical foundation for every world. But let C be some merely contingent
existent. C must be derivative from some combination of A and B. So, in worlds
where C exists, it is grounded in (e.g.) A. But given that A is necessary, every world
is an A-world. So why isn’t every world also a C-world? We might try to explain the
A-but-not-C-worlds by appeal to the existence of some blocker D in those worlds,
but that won’t help because D must also be derivative from some combination of
A and B, and unless we want to say that D is necessary (which, if it’s a blocker
for C, entails C can’t possibly exist), we’re stuck with the same problem as regards
the existence of D. Without a satisfactory story to tell here, this package looks like a
non-starter.25
Similar points apply to Fixed Shaky that accepts the necessitarian thesis along with
the contingent fundamentality thesis. Like its predecessor, this view too faces some
significant difficulties. For one, it is unclear how it is any better placed with regards
to resolving the modal variation explanatory challenge: while this view allows that
the priority ordering of the worlds vary, this looks irrelevant when it comes to (e.g.)
explaining the variable existence/qualitative nature of contingent entities like C. But
there’s another problem specific to this view: the things that are eligible for being
fundamental all necessarily exist, so all of them are available at all the worlds. However,
the view also requires modal variation regarding their priority ordering—for example,
it says that x is fundamental in w, but derivative in w’. Yet what explains this? What
explains the different possible priority structures the worlds have?
We can make this problem more acute by thinking about Shifty Shaky, which
allows for contingent fundamentalia that are only contingently fundamental. This
position can explain variation in priority structure in terms of variation in existence:
x is derivative in w’ because y exists in w’ and, in that world, x depends upon/is
grounded in y, while x is fundamental in w because y—the thing that x would depend
upon—doesn’t exist in w. So Quarky the quark is fundamental in the actual world
because there’s no gunk for Quarky to depend on, but in gunky worlds, Quarky
depends upon—and hence is derivative from—a certain glob of gunk. Packages that
are committed to the necessitarian thesis have no recourse to this kind of explanation.
For them, everything that is possibly fundamental exists in every world, so we can’t
explain priority variation in terms of existential variation.
24 I suspect that worries along these lines are what motivate many to reject the necessitarian thesis in
the first place.

25 One could salvage the position by saying there’s only one possible world; this avoids the problem at
the cost of entailing that there is no modal variation, which is a clear case of throwing the baby out with
the bathwater.
nathan wildman 
While these problems aren’t insurmountable, they suffice to push us away from the
two necessitarian views and into the waiting arms of the latter contingentist pair. But
what might settle the difference between these two?
It seems to me that the key difference between these two packages will be disagree-
ments over certain cases. For example, suppose that some object o is fundamental
in world w, in part because, in w, o is mereologically simple—that is, nothing is a
proper part of o in w. Further, it’s natural to read this as an existential claim; that is,
as something like, it is not the case that there exist some xxs such that they are proper
parts of o in w. But nothing said so far precludes o’s existing in another world w’,
wherein the xxs exist and serve as proper parts of o. Then, assuming that mereological
structure and priority ordering are correlated, we get the result that o is merely
contingently fundamental, since it is fundamental in w but non-fundamental in w’.
Shifty Firm denies that such cases are possible. The closest one can come would be to
have a qualitatively similar but distinct object o’, which is derivative in w’; however,
you’d never get o itself, since it is necessarily fundamental. Meanwhile, as we saw
above, Shifty Shaky embraces such cases.
So, should we allow for such cases? As far as I can see, the best case for rejecting
them stems from certain essentialist assumptions—specifically, we think it essential
to o that it be mereologically simple, and it is this which blocks the possibility
of o-as-dependent-worlds (or at least o-as-mereologically-dependent-worlds). How-
ever, the mereological essentialist assumption is questionable. Further, as we saw
earlier (Section 3), the shifty shaker can agree to something like this essentialist
assumption in the form of o’s essentially having a relevant conditional property like
being mereologically simple provided the xxs don’t exist.26 Consequently, we’ve prima
facie reasons for being shift shakers over shifty firmers. Of course, there is more to be
said here, but this suffices for a first pass.
5 Conclusions
As stated at the outset, the aim here was to prompt further discussion of the so-far-
neglected question of fundamentality’s modal status, which I have tried to initiate
by exploring the contingent fundamentality thesis. In particular, I’ve focused on
clarifying the essential groundwork and specified my interpretation of the core thesis,
offered a few points in favour of it, replied to some initial objections, and made a
first pass at seeing how it relates to another point in the intersection of modality and
fundamentality.
But this is just a rough, exploratory survey of a small part of the terrain—there are
many further points to evaluate with regards to contingent fundamentality. We might,
26 It might be necessary to complicate the property, adding conditions beyond the mere existence of
the xxs to include e.g. that they are arranged in such and such a manner. Such complications can be easily
added without undercutting the main point.
 exploring the contingent fundamentality thesis
for example, wonder how it relates to debates about the necessitation of grounding
relations.27 Additionally, assessing what the contingentist can say about the modal
freedom of fundamentalia would be interesting.28 The standard assumption seems
to be that we’re free to combine the various fundamental entities in whatever way
we like—that is, they are modally independent—but exactly how to square this with
contingent fundamentality looks potentially problematic. Finally, it would be nice to
have a proper argument for adopting the contingent fundamentality thesis, something
that I’ve not given here (alternatively, formulating more objections/arguments against
the thesis looks like a potentially fruitful area too).
However, these are all projects for another day. For now, I hope that I’ve convinced
some of my readers that the contingent fundamentality thesis—and, more generally,
the matter of fundamentality’s modal strength—are points worth thinking about.29
References
Armstrong, D. (1997). A World of States of Affairs. Cambridge: Cambridge University Press.
Audi, P. (2012). ‘A Clarification and Defense of the Notion of Grounding’. In F. Correia
and B. Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality.
Cambridge: Cambridge University Press, pp. 101–21.
Barnes, E. (2012). ‘Emergence and Fundamentality’. Mind 121(484): 873–901.
Baron, S. (2015). ‘The Priority of the Now’. Pacific Philosophical Quarterly 96: 325–48.
Cameron, R. (2007). ‘The Contingency of Composition’. Philosophical Studies 136(1): 99–121.
Cameron, R. (2008). ‘Truthmakers and Ontological Commitment’. Philosophical Studies 140:
1–18.
Correia, F. (2005). Existential Dependence and Cognate Notions. Munich: Philosophia Verlag.
Correia, F. and Schnieder, B. (2012). ‘Introduction’. In F. Correia and B. Schnieder (eds), Meta-
physical Grounding: Understanding the Structure of Reality. Cambridge: Cambridge Univer-
sity Press, pp. 1–36.
Dasgupta, S. (2014a). ‘On the Plurality of Grounds’. Philosophers’ Imprint 14(20): 1–28.
Dasgupta, S. (2014b). ‘The Possibility of Physicalism’. Journal of Philosophy 111(9–10): 557–92.
Davies, M. (1981). Meaning, Quantification, Necessity: Themes in Philosophical Logic. London:
Routledge.
deRossett, L. (2010). ‘Getting Priority Straight’ Philosophical Studies 149: 73–97.
Evans, G. (1985). ‘Reference and Contingency’. In G. Evans, Collected Papers Oxford: Clarendon
Press, pp. 178–213.
Fine, K. (1995). ‘Ontological Dependence’. Proceedings of the Aristotelian Society 95: 269–90.
27 See e.g. Leuenberger (2014), Skiles (2015), and Trogdon (2013).

28 See e.g. Wang (2016).
29 I’d like to thank the members of the phlox research group—in particular Yannic Kappes and Jonas
Werner—along with Alex Skiles, Jennifer Wang, and Tobias Wilsch for fruitful discussion concerning this
paper. Thanks also to Amanda Cawston and Ohle for their generous help and support. The research for
this paper was partially funded by the Swiss National Science Foundation Sinergia project, ‘Grounding:
Metaphysics, Science, and Logic’ (Project 147685), and the DFG Emmy Noether Research group, ‘Ontology
after Quine: Fictionalism and Fundamentality’.
nathan wildman 
Fine, K. (2001). ‘The Question of Realism’. Philosophers’ Imprint 1: 1–30.

Fine, K. (2012). ‘A Guide to Ground’. In F. Correia and B. Schnieder (eds), Metaphysical
Grounding: Understanding the Structure of Reality. Cambridge: Cambridge University Press,
pp. 37–80.
Fine, K. (2015). ‘Unified Foundations for Essence and Ground’. Journal of the American
Philosophical Association 1(2): 296–311.
Heil, J. (2003). From an Ontological Point of View. Oxford: Oxford University Press.
Jenkins, C.S.I. (2013). ‘Explanation and Fundamentality’. In M. Hoeltje, B. Schnieder, and
A. Steinberg (eds), Varieties of Dependence. Munich: Philosophia Verlag, pp. 211–42.
Kripke, S. (1971). ‘Identity and Necessity’. In M.K. Munitz (ed.), Identity and Individuation
New York: New York University Press, pp. 135–64.
Kripke, S. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press.
Leuenberger, S. (2014). ‘Grounding and Necessity’. Inquiry 57(2): 151–74.
Lewis, D. (1986). On the Plurality of Worlds. Oxford: Blackwell.
Lopez de Sa, D. (ms). ‘Priority Presentism’. [Unpublished manuscript.]
Lowe, E.J. (2010). ‘Ontological Dependence’. In E.N. Zalta (ed.), The Stanford Encyclope-
dia of Philosophy (Spring 2010 Edition), http://plato.stanford.edu/archives/spr2015/entries/
dependence-ontological/.
McDaniel, K. (2013). ‘Degrees of Being’. Philosophers’ Imprint 13(19): 1–19.
Miller, K. (2009). ‘Defending Contingentism in Metaphysics’. Dialectica 63(1): 23–49.
Nolan, D. (2005). David Lewis. Chesham: Acumen Publishing.
Raven, M. (2015). ‘Fundamentality without Foundations’. Philosophy and Phenomenological
Research 90(3): 607–26.
Rosen, G. (2010). ‘Metaphysical Dependence: Grounding and Reduction’. In R. Hale and
A. Hoffman (eds), Modality: Metaphysics, Logic, and Epistemology. Oxford: Oxford Univer-
sity Press, pp. 109–35.
Schaffer, J. (2009). ‘On What Grounds What’. In D. Chalmers, D. Manley, and R. Wasserman
(eds), Metametaphysics. Oxford: Oxford University Press, pp. 347–83.
Schaffer, J. (2010a). ‘The Internal Relatedness of All Things’. Mind 119(474): 341–76.
Schaffer, J. (2010b). ‘The Least Discerning and Most Promiscuous Truthmaker’. Philosophical
Quarterly 60: 307–24.
Schaffer, J. (2010c). ‘Monism: The Priority of the Whole’. Philosophical Review 119(1): 31–76.
Schaffer, J. (2013). ‘The Action of the Whole’. Proceedings of the Aristotelian Society Supplemen-
tary Volume 87(1): 67–87.
Schnieder, B. (2006). ‘A Certain Kind of Trinity: Dependence, Substance, Explanation’. Philo-
sophical Studies 129: 393–419.
Sider, T. (2011). Writing the Book of the World. Oxford: Oxford University Press.
Skiles, A. (2015). ‘Against Grounding Necessitarianism’. Erkenntnis 80(4): 717–51.
Steinberg, A. (2015). ‘Priority Monism and Part/Whole Dependence’. Philosophical Studies
172(8): 2025–31.
Trogdon, K. (2013). ‘Grounding: Necessary or Contingent?’ Pacific Philosophical Quarterly
94(4): 465–85.
van Inwagen, P. (2002). Metaphysics. Boulder, CO: Westview Press.
Wang, J. (2016). ‘Fundamentality and Modal Freedom’. Philosophical Perspectives 30(1):
397–418.
 exploring the contingent fundamentality thesis
Wildman, N. (2013). ‘Modality, Sparsity, and Essence’. Philosophical Quarterly 63(253): 760–82.
Wildman, N. (2016). ‘How (Not) To Be a Modalist about Essence’. In M. Jago (ed.), Reality
Making. Oxford: Oxford University Press, pp. 177–96.
Wildman, N. (ms). ‘What’s Wrong with Weak Necessity?’ [Unpublished manuscript.]
Williams, J.R.G. (2010). ‘Fundamental and Derivative Truths’. Mind 119: 103–41.
Williamson, T. (2013). Modal Logic as Metaphysics. Oxford: Oxford University Press.
Wilson, J. (2014). ‘No Work for a Theory of Grounding’. Inquiry 57(5–6): 535–79.
Wilson, J. (2016). ‘The Unity and Priority Arguments for Grounding’. In K. Aizawa
and C. Gillett (eds), Scientific Composition and Metaphysical Ground. London: Palgrave
Macmillan, pp. 171–204.
15
Heidegger’s Grund
(Para-)Foundationalism
Filippo Casati
1 Introduction
There is a world. There are chairs, stars, dreams, and many other things. That’s a fact.
Another fact is that many of these things depend on something else. For instance,
chairs depend on their parts, stars depend on some specific chemical reactions, and
dreams depend on dreamers. Many things have whatever form of being or existence
they have because they depend on other things. Interpreting this dependence relation
as a grounding relation, we can also say that many things have whatever form of being
or existence they have because they are grounded in other things.1 In the contempo-
rary literature, such a grounding relation has been both understood and spelled out in
many different ways. Nevertheless, as it is shown in the Introduction of this volume by
Bliss and Priest, if we abstract from the very many different metaphysical dependence
relationships, it is still possible to produce a taxonomy which covers all the possible
understandings of grounding. Such a taxonomy makes use of four structural proper-
ties: (i) Anti-Reflexivity [AR]: nothing depends on itself; (ii) Anti-Symmetry [AS]: no
things depend on each other; (iii) Transitivity [T]: everything depends on whatever
a dependent depends on; (iv) Extendability [E]: everything depends on something
else. These four structural properties, and their relative negations, give rise to ten
different possible grounding theories, which may be gathered together in three major
clusters: Foundationalism (F), Infinitism (I), and Coherentism (C). It is interesting to
remark that, even though all these theories have different features, because they make
use of different combinations of structural properties, they also share one important
characteristic: they are all consistent.
1 We are aware that the contemporary literature on grounding is vast and that many philosophers draw
finer distinctions within this notion. Nevertheless, the broad and intuitive understanding presented here
will be sufficient for our purposes. For a general overview of the area, see Bliss and Priest 2017, Bliss and
Trogdon 2014, and Takho and Lowe 2015.
 heidegger’s grund: (para-)foundationalism
This paper presents two new grounding theories (called para-foundationalism 1.0
and para-foundationalism 2.0) that, in virtue of their being inconsistent (but not
trivial) theories, do not fit in the taxonomy presented by Bliss and Priest. In order
to do so, we will develop some metaphysical ideas proposed by Martin Heidegger.
Consistently with a vast part of the current literature, he thought that all things have
whatever form of being they have because they depend on other things. In particular,
he believed that every thing is because every thing depends on being.2 Heidegger’s
being is the ground [Grund] of literally everything because being is what makes any
entity an entity. Chairs, stars, dreams, and the world are in virtue of being.3
In Section 2, we introduce Heidegger’s concept of ground by distinguishing
between an ontic ground and an ontological ground. In Section 3, we focus our
attention on the ontological ground. We present Heidegger’s idea according to
which being is the ground of every entity and being is itself ungrounded. We also
discuss its relation with the Principle of Sufficient Reasons (PSR), and we describe its
structural properties. Finally, we show that these structural properties are the same
ones that characterize a particularly strong form of foundationalism. In Section 4 and
Section 5, we show how Heidegger’s characterization of being leads to a contradiction,
according to which being both is and is not an entity. After that, assuming that such a
contradiction is a dialetheia (namely a true contradiction), we show how Heidegger’s
foundationalism should be revised in order to do justice to the antinomic nature
of being itself. Thus, we introduce two forms of para-foundationalism, which is an
inconsistent version of foundationalism. In Section 6, using para-foundationalism,
we try to give an interpretation of one of the most obscure concepts of the so-called
late Heidegger, namely the last God. Finally, in the Appendix, we propose two formal
models that show how, working in a paraconsistent setting, para-foundationalism
does not lead to logical triviality.
2 Sein as the Ontological Ground

During the last months of 1928, while Heidegger was writing his well-known lecture
entitled What is metaphysics?, another manuscript was sitting on his desk. Its title
was On the Essence of Ground. Heidegger was aware that, even though the notion
2 Before developing the main arguments of the paper, let’s make two terminological clarifications. [1]
Heidegger uses different terms to talk about entities (for instance, thing [Ding] and object [Objectum]).
All these terms have different phenomenological meanings. Nevertheless, for simplicity, the present paper
uses all these terms as synonyms. [2] The majority of the translators have decided to follow the convention
according to which capital ‘B’ is used for being [Sein] and a lower case ‘b’ for beings or entities. However,
this convention seems to be an arbitrary artifact because, in German, both words, being nouns, begin with
capitals [Sein, Seindes]. For this reason, in this paper, we do not follow this practice.
3 We are aware that, following some interpreters, Heidegger endorses a specific form of ontological
pluralism, according to which everything exists but some things exist in different ways than others (see
McDaniel 2009, 2017). Since I believe that this position is contentious, I do not want to endorse it here. For
this reason, I prefer to use to be rather than to exist.
filippo casati 
of ground “is, in general, bound up with the central questions of metaphysics”

(Heidegger 1998, p. 99), its meaning has never been properly understood. So, this
second essay was meant to answer the following question: what does ground [Grund]
mean?
Heidegger proposes two complementary characterizations of the notion of ground.
On the one hand, appealing to Aristotle, he takes it to be the first thing or the beginning
on which what is grounded depends: these expressions seem to suggest that there is
a hierarchical structure of elements in which the primary ones behave as basis for
the secondary ones. In this sense, the ground is the cause which metaphysically and
logically determines what is grounded in it (cf. Heidegger 1998, p. 98).4 On the other
hand, appealing to Leibniz, Heidegger takes ground to be the reason for something to
be how and what it is: something obtains because of its ground or something holds in
virtue of its ground (cf. Heidegger 1998, p. 99).5
According to Heidegger, grounding is not unitary; there is no single dependence
relation in play. Indeed, he discusses two kinds of grounding: the first one is concerned
with the reason why some entities are the entities that they are (for instance, why
some entities are hammers or chairs), while the second one is concerned with the
reason why all entities simply are. The first kind is an ontic ground (a ground which
deals exclusively with specific kinds of entities), while the second one is an ontological
ground (a ground which deals only with entities as entities). Since, in what follows,
we will focus on the latter kind of ground, let’s describe it in more detail.
The ontological ground, namely what Heidegger labels as the ground of something,
is “the ultimate and primary” one (Heidegger 1998, p. 130) because it is not concerned
with one specific kind of entity but with entities in general. According to Heidegger,
everything we can refer to with an intentional activity is an entity: everything we
can think about, speak about, or reason about is an object.6 “Such grounding of
things [namely the ontological ground] lies ‘at the ground’ of all comportment toward
beings, and in such a way that (. . .) beings become manifest in themselves (i.e. as the
beings they are and in the way they are)” (Heidegger 1998, p. 130). The ontological
ground makes “possible the manifestation of beings in themselves” (Heidegger 1998,
p. 129). Exactly the entities manifested in themselves are grounded in the ontological
ground—namely entities simply as entities, as things that are. For instance, since we
can think about a hammer, a number, the redness of the rose, the idea of the infinite,
4 Concerning this first characterization of the notion of ground, see Bliss and Priest 2017 and Corkum
2013.
5 Concerning this second characterization of the notion of ground, again see Bliss and Priest 2017.
6 This definition of ‘entity’ relies on a strong account of intentionality: there is always something (namely
an object, an entity, or a thing) towards which an intentional activity is directed. We are aware that not all
the interpreters of Heidegger would agree on this matter; however, concerning this point, we follow Moore
in believing that, according to Heidegger, “for any perception, there is an object of perception; for any flash
of understanding, an object of understanding (. . .)” (Moore 2012, p. 439). Finally, this interpretation of
Heidegger has important exegetical virtues because it can explain why, for Heidegger, being constitutes a
problem. See Section 4 of the present paper and Casati and Fujikawa 2015.
 heidegger’s grund: (para-)foundationalism
and even God, they are all entities and, as such, they are grounded in the ontological
ground, which makes them be.
Since the ontological ground is the ground in virtue of which entities are entities
and since, according to Heidegger, what makes entities entities is being only, then the
ground in virtue of which entities are entities is being itself. Entities are because of
being. Being grounds entities. The ground, understood as the ontological ground, is
being itself. This is also the reason why “grounding something means making possible
the why-question in general. (. . .) Why is this in this way and not otherwise? Why this
and not that? Why something at all and not nothing?” (Heidegger 1998, pp. 129–30).
We can ask why a number (which is an entity) has this property or that property,
only because there is a number (an entity) in the first place, and there is a number
(an entity) in the first place if and only if the ontological ground, namely being,
makes that number an entity. We can worry why there is something (the hammer, the
redness of the rose, the idea of the infinite) and not nothing, exactly because there is
something and not nothing at all. However, there is something if and only if something
is grounded in an ontological ground which makes that something something. Such
ontological ground is being. In order to ask something about entities, entities are
needed. Therefore, being is needed as well because being ontologically grounds all
entities—being makes all entities entities.
As Heidegger points out, the ontological ground is metaphysically more funda-
mental than the ontic ground. Indeed, the latter is the reason why a specific kind
of entity is that specific kind of entity, while the former is the reason for any entity
to simply be an entity. It is the reason why everything is. As an example, consider
a hammer. The fact that a hammer is a specific kind of entity (namely, according to
Heidegger’s phenomenology, a piece of equipment) is primarily grounded in the fact
that a hammer is an entity. Without being ontologically grounded in being, a hammer,
as any other equipment, cannot be that specific entity that a piece of equipment is—
without being ontologically grounded in being, a hammer would not be an entity in
the first place. The ontological ground is always more radical than the ontic ground:
it is metaphysically prior because it is the conditio sine qua non for anything to be.
Heidegger: “Yet because such grounding of something [namely the ontological ground]
prevails from the outset throughout all becoming-manifest of beings (ontic truth),
all ontic discovery and disclosing must in its way be a grounding of something”
(Heidegger 1998, p. 130).
In the secondary literature, Heidegger’s discussion about being and the ontological
ground is often interpreted as the mystical element of his philosophy. The ontological
ground, namely being itself, is not something we can logically think about or rationally
discuss—it has to be perceived, experienced without turning it into the subject matter
of philosophy. As Caputo claims, “Heidegger’s grounding is without a why; it is
the renunciation of concepts and representations, of propositions and ratiocinations
about Being; it lets Being be Being” (Caputo 1986, p. 191). In Section 3, we give an
filippo casati 
alternative interpretation revealing the philosophical argument that can rationally do

justice to Heidegger’s understudying of the ontological ground.
3 Sein as Abgrund: Foundationalism

As we have already seen, the ontological ground is being. Since being makes entities
entities, being ontologically grounds all entities. If something is not grounded in
being, it is not an entity. More explicitly, this metaphysical dependence relation
between being and all entities can be spelled out in the following way:
(Gr): x (ontologically) grounds y if and only if x makes y an entity.
However, since being is what makes all entities entities and since nothing else makes
all entities entities, being is the only ground for all entities as such. The foundational
element that makes all entities entities (namely being itself) is unique. Therefore, (Gr)
can be reformulated as:
(Gr ): b (ontologically) grounds y if and only if b makes y an entity.
where b is being and y can be any entity. Nevertheless, Heidegger not only holds the
belief that any entity is an entity in virtue of being, he also thinks that, according to
the so-called ontological difference [ontologische Differenz], being itself is not an entity.
“The being of beings ‘is’ not itself a being. The first philosophical step in understanding
the problem of being consists in avoiding telling the mython tina diegeisthai, in not
‘telling a story’, that is, not determining beings as beings by tracing them back in their
origins to another being – as if being had the character of a possible being” (Heidegger
1996, p. 5). Being is not an entity and, exactly because of this, Heidegger claims that the
ground, understood as the ontological ground, is transcendental. Since being grounds
all entities, being is the ground on which all entities are grounded. Because of the
ontological difference, being is transcendental in the sense that it is beyond entities.
So, if being is the ground of all entities and being is transcendental, the ground of all
entities is transcendental. Heidegger: “Ground belongs to the essence of being and
being (not beings) is given only in transcendence” (Heidegger 1998, p. 132). Thus,
“such grounding prevails transcendentally from the outset throughout all becoming-
manifest of beings” (Heidegger 1998, p. 130).
From these observations, two important consequences follow. First of all, given
(Gr ) and the ontological difference, being is ungrounded. If all entities (and only
entities!) need to be grounded in being and being is not an entity, then being is neither
grounded in something other than being itself nor is it self-grounded; otherwise it
would be an entity. Therefore, being is simply ungrounded.7 In Heidegger’s jargon,
7 Following McDaniel (2015) and Priest (2015), someone may object that, in Heidegger, it is not the
case that being is ungrounded because being always depends on entities. As Heidegger himself writes: “if
we think of the matter just a bit more rigorously, (. . .) we see that being means always and everywhere:
 heidegger’s grund: (para-)foundationalism
being is the abyssal ground. Being is the “abyss of the ground” (Heidegger 1998, p. 134)
because, since being grounded means being an entity, and since being is not an entity,
being is ungrounded. Nevertheless, being grounds everything else. In a footnote of
On the Essence of Ground, this idea is clearly stated: “Where does the necessity lie for
grounding? In the abyss of, in the non-ground” (Heidegger 1998, p. 100). Being lies on
the abyss because it is the non-ground: it does not have a foundation. Metaphorically,
there is no thing (no ground) that supports (grounds) being.
The second consequence is that the so-called Principle of Sufficient Reason (PSR)
does not unrestrictedly hold. Heidegger takes (PSR) to be that “fundamental principle
[Grundsatz]” according to which “nihil est sine ratione, nothing is without reason”
(Heidegger 1998, p. 100) or, transcribing it positively, “omne ens habit rationem, every
being has a reason” (Heidegger 1998, p. 100). Moreover, Heidegger thinks that having
a reason means being in virtue of something, being because of something or, more
generally, being grounded in something. From this point of view, the fundamental
principle [Grundsatz] is a principle about what is fundamental [Der Satz von Grund].
“The fundamental principle is a principle about the ground [Der Satz von Grund
ist ein Grundsatz]” (Heidegger 1998, p. 100). If this is the case, (PSR) can be also
read as ‘nothing is without a ground’ or ‘everything is with a ground’. Now, having
said so, it is easy to see why Heidegger’s metaphysics leads to a constrained version
of (PSR). Indeed, since every entity is grounded in being, every entity has a reason
to be something (and not nothing). However, since being is not an entity, being is
ungrounded. This means that it is not true that everything has a reason (understood
as ground) and, consequently, it is not the case that (PSR) holds unrestrictedly.
Indeed, according to Heidegger himself, every entity has a reason but not being. Since
being is ungrounded, “the principle of [sufficient] reason (or ground) is valid for
beings [entities]” (Heidegger 1998, p. 132). (PSR) holds for entities but not for being.
Everything is grounded in being (which is why every entity has a reason to be) but
being remains the ‘groundless ground’ of everything that is (cf. Braver 2012).
Even though, given the contemporary analytic debate, the position described
until now may seem weird and obscure, it could be easily categorized as a kind
of foundationalism. Following Bliss and Priest (2017), we take foundationalism to
be the view according to which everything grounds out in foundational elements.
An element is a foundational one (let’s call it FEx) if there is no y on which x
depends, other than perhaps itself. Such an element can be formally described in
the following way: ∀y(x → y ⊃ x = y). We read x → y as x depends on y.
Since being behaves as a foundational element because there is no element on which
the being of beings” (Heidegger 2002, p. 61). However, the kind of dependence or grounding relation in
place here cannot be the dependence or grounding relation discussed in this paper. Indeed, (Gr) is only
that kind of (ontological) dependence or grounding relation that makes all entities entities. Since, from the
beginning to the end of his philosophical trajectory, Heidegger has always endorsed the position according
to which being is not an object, being cannot be grounded (in the sense of (Gr)) in entities. If it is the case
that being depends on entities, such dependence relation is not a (Gr) one.
filippo casati 
being depends, Heidegger’s metaphysics is a foundationalist one. However, Heidegger

also endorses a particularly strong form of foundationalism in which there is only
one foundational element, which is being, and everything depends upon this unique
fundamentalium. Taking b as being, we can formally express this thought in the
following way: ∃b(FEb ∧ ∀y(y = b ⊃ y → b)). On the one hand, being is unique
because there is only one being which grounds all entities. On the other hand, being
is not grounded in anything else and is not self-grounded either; otherwise it would
be an entity.
It is also possible to precisely describe Heidegger’s foundationalism appealing to
its structural properties. For simplicity, let’s consider the graph in Figure 15.1, which
describes the grounding relation between being (represented by the node labeled b)
and one entity only (represented by the node labeled e). The solid arrows indicate what
depends on what, while the dashed arrows indicate where there is no dependence
relation.
As we can see from the graph, entity e depends on being but it does not depend
on itself. Moreover, being does not depend on entity e and it does not depend on
itself either. Being does not depend on anything at all. From this simplified picture,
the structural properties of Heidegger’s foundationalism should appear clear. First
of all, Heidegger’s foundationalism endorses anti-reflexivity or [AR], according to
which nothing depends on itself. Formally, this structural property is expressed in
the following way:
[AR]: ∀x¬(x → x).
Since neither entity e nor being depend on themselves, nothing depends on itself.
More generally, since every entity depends on being (and only on being!), and being
does not depend on anything (not even on itself), nothing depends on itself. Secondly,
it endorses anti-symmetry or [AS], according to which no things depend on each
other. Formally, it is expressed as:
[AS]: ∀x∀y(x → y ⊃ ¬y → x).
Once again, consider the graph. Entity e depends on being but being does not depend
on entity e. Since there are only two elements and since neither of them depend on
each other, nothing depends on each other. Of course, [AS] does not rule out the
Figure 15.1.
 heidegger’s grund: (para-)foundationalism
possibility that something depends on itself because, as the formula shows, x could
be y. Nevertheless, according to the graph presented above, this is not the case. Since
[AR] holds, no things depend on each other, not even on themselves. More generally,
let’s recall that, in Heidegger’s metaphysics, we deal with entities and being only.
Entities depend on being in order to be entities, but being does not depend on entities.
Moreover, since being does not depend on anything, it does not depend on itself either.
Thus, no things depend on each other. Finally, Heidegger’s foundationalism endorses
the negation of extendability or [¬E], according to which something does not depend
on anything else. Formally,
[¬ E]: ∃x∀y(x → y ⊃ x = y).
As we can easily see looking at the graph, being does not depend on entity e and it does
not depend on itself either. Thus, being does not depend on anything. According to
Heidegger, there is an element that does not depend on anything; this element is being.
It does not depend on anything because, since everything that depends on being is an
entity and being is not an entity, being does not even depend on itself. Once again,
being is the ungrounded ground or the groundless ground. Using a poetic expression
borrowed from Angelus Silesius’ Cherubinic Wanderer, Heidegger claims that, as any
other entity, the rose is without a wherefor [warum]—it blooms because it blooms
(see Caputo 1986; Vannini 2007). Asking for the ultimate reason of the rose is useless
because there is no ultimate reason for its blooming. Of course, the rose has some
reasons to be what the rose actually is; however, the rose, as all the other entities, is
without an ultimate reason. Its reason, namely being, does not have any reason. Since
all entities are in virtue of being and since being is not in virtue of anything, all entities
are not in virtue of anything either. This does not mean that entities do not have any
ground at all. It simply means that all entities rest on a ground which is ungrounded.
The ground on which every entity relies is, indeed, an abyss. The rose, as with all the
rest, is ultimately groundless.
4 Sein as Seyn: Para-Foundationalism 1.0

At this point, a problem should appear evident. The issue in question does not directly
concern the foundationalist thesis according to which the rose, with all the other
entities, is grounded in an ungrounded fundamentalium; on the contrary, it is about
the fundamentalium itself. The problem concerns being and it is easy to see why.
As we have seen, Heidegger’s metaphysics works under two main assumptions. First
of all, being is not an entity: “The being of beings ‘is’ not itself a being [an entity]”
(Heidegger 1996, p. 5). Secondly, an entity is whatever we can refer to. It follows
that it should be impossible to refer to being; otherwise being would be an entity.
Nevertheless, we did refer to being describing it as such and such. For instance, we
have argued in favor of the idea that being is the ungrounded ground of every entity.
However, since being is not an entity at all, being cannot be the ungrounded ground
filippo casati 
either. Being can be neither thought nor spoken. The price paid in doing so is that,
referring to it, being itself would be turned into exactly what being is not, namely
an entity. Therefore, being is ineffable: whatever is the grounding structure which
has being as a foundational element, nothing can be said about it. Being, the ground,
cannot be discussed, argued for or discovered. “The essence of ground cannot even
be sought [or] let alone found” (Heidegger 1998, p. 131).
The situation is even more complicated than this, though. Indeed, not only do the
two assumptions endorsed by Heidegger lead to the ineffability of the ground, but they
also lead to an aporia. This is actually the case because, exactly in saying that being is
ineffable, we say something about it. In other terms, exactly in saying that being is not
an entity, being is an entity because we refer to it. Heidegger’s assumptions imply that
being is an entity and not an entity, namely a contradiction.8 Now, facing this problem,
two questions seem inevitable: what did Heidegger do? And, more interestingly, what
is to be done?
Heidegger has tried to solve the problem in two different ways. [1] In the first place,
realizing the inevitable inconsistency implied by his own metaphysics, he simply gives
up the problem: he admits that, about being, namely about the ground of everything,
we need to be silent. Since the contradiction of being is due to the fact that, even
though being is ineffable, being can be said, let’s not speak about it anymore. About
being, we should simply be silent. This first attempt is not very convincing, though.
More than a solution, this idea looks like a strategy to avoid the problem. Even if
we grant that the silence about being allows us to avoid facing the inconsistency of
being, that inconsistency is still there because we have already spoken about being.
In one way or another, being has been spoken of. Moreover, Heidegger himself,
in his Contributions to Philosophy (2012a), suggests that the silence about being is
not completely silent: it communicates the impossibility of saying something about
being itself. If this is the case, the silence refers to being (and its impossibility of
being spoken of). If the silence refers to being, being is an entity as well. Therefore,
being remains contradictory. The foundational element is still aporetic. Thus, Heideg-
ger’s first solution fails. [2] In the second place, Heidegger suggests the idea that, even
though being cannot be spoken of, it can be shown by art. “The art work opens up
in its own way the being of beings. (. . .) In the art work, the truth of beings has set
itself to work” (Heidegger 1977, p. 166). However, even assuming that a work of art
actually behaves in this way, this solution does not work either because the problem of
the inconsistency of being remains. It is not the case that, because someone can grasp
being through art or poetic language, then being is not contradictory anymore. Once
again, the foundational element is still inconsistent.
Given that Heidegger’s attempts do not seem successful, is there any solution left
on the table? One possibility is to simply accept the brutal fact that, given Heidegger’s
8 This paradoxical feature of Heidegger’s metaphysics is extensively discussed in Section 2 of Casati and
Wheeler 2016.
 heidegger’s grund: (para-)foundationalism
assumptions, being is both an entity and not an entity. As Priest has already suggested
(2002; 2015), Heidegger should have endorsed a dialetheic solution according to
which being is a dialetheia—a true contradiction. In other places, we have argued
that this is the position that Heidegger actually endorses in his late period.9 More
specifically, Heidegger’s Ereignis (the Event of being) is exactly the realization that
“‘being’ does not (. . .) mean objective presence in itself, and non-being does not (. . .)
mean complete disappearance. Instead, non-being [is] a mode of being: it is and
yet is not. And likewise being, which is permeated with the ‘not[-being]’ and yet it
is” (Heidegger 2012a, p. 80). Therefore, being is (an entity) and is not (an entity).
This is also the reason why the late Heidegger claims that “the contradiction is
essentially a fundamental proposition about being and its truth” (Heidegger 2012b,
p. 3—my translation). We believe that the difference between being [Sein], used by
the early Heidegger, and beyng [Seyn], used by the late Heidegger, is that the former
is consistent (because it is simply not an entity) and the latter is inconsistent (because
it is both an entity and not). However, what Heidegger has really claimed is not a
relevant issue here. What is important is the following question: regardless of whether
the dialetheic solution to the problem of being was or was not endorsed by Heidegger
himself, how does such a metaphysical position affect his account of grounding?
It is natural to think that the switch from a consistent account of the foundational
element (call it being [Sein]) to an inconsistent account of the foundational element
(call it beyng [Seyn]) would imply a switch from a consistent account of grounding
to an inconsistent account of grounding. And this is what happens. Let’s recall that
beyng [Seyn] is an entity and not an entity. As we have already seen in Section 4, from
the fact that beyng is not an entity, it follows that beyng grounds all entities but beyng
itself is ungrounded. From the fact that beyng is an entity, it follows that, in virtue of
its being an entity, beyng needs to be grounded. Since beyng is the ontological ground
that makes every entity an entity, then beyng grounds itself. Thus, since beyng is both
an entity and not, beyng depends on itself and, at the same time, it does not depend
on anything. Beyng is grounded and ungrounded at the same time. This is shown in
the graph in Figure 15.2:
Figure 15.2.
9 For an interpretation according to which Heidegger himself endorses dialetheism, accepting as true
the contradiction of being, see Casati 2016.

filippo casati 
Once again, beyng is represented by the node labeled b and the entity grounded in
beyng is represented by the node labeled e. The solid arrows indicate what depends on
what, while the dashed arrows indicate where there is no dependence relation. As we
can see, entity e depends on beyng and it does not depend on itself. Moreover, beyng
does not depend on entity e but it both depends and not on itself.
Now, let’s have a look at the structural properties that this new inconsistent ground-
ing structure has. From the fact that beyng is not an entity, everything previously
stated still holds. Since all entities depend on beyng and beyng does not depend on
anything (not even on itself), nothing depends on itself. Therefore, [AR] holds. Since
all entities depend on beyng (and only on beyng) but beyng neither depends on all
entities nor depends on itself, no things depend mutually on each other. Therefore,
[AS] holds. Finally, from the fact that all entities depend on beyng but beyng does not
depend on anything else, it follows that something does not depend on anything else.
Therefore, [¬E] holds. However, in contrast with the dependence relation grounded
in the consistent foundational element (being), this new approach has more structural
properties than the ones described until now. Indeed, given the fact that beyng is
an entity as well, two other structural properties hold. First of all, since beyng is
(also) an entity and since beyng is the only foundational element that makes all
entities, something depends on itself, namely beyng. Beyng makes itself an entity. The
structural property according to which something depends on itself is the negation of
the anti-reflexivity property ([¬AR]) and it is formalized in the following way:
[¬AR] : ∃x(x → x)
Secondly, since beyng is (also) an entity and beyng is the only foundational element
which makes all entities entities, the negation of [AS] holds as well. We formally
describe it as
[¬AS] : ∃x∃y(x → y ∧ y → x).
Strictly speaking, [¬AS] says that some things depend on each other without ruling
out the possibility that some things depend on themselves. Now, [¬AS] would be
incompatible with the fact that there are things depending of themselves if and only
if [¬AR] holds. However, this is not the case with Heidegger because, as we have
seen, according to him, both [AR] and [¬ AR] hold. Therefore, [¬AS] holds because
beyng does depend on itself: it is self-grounded. At this point, it is easy to see why
the inconsistency of the foundational element beying spreads to the structure of the
grounding dependence. Since beyng is an entity and not an entity, the grounding
dependence relation has an inconsistent characterization as well: both [AR] and
[¬AR], and [AS] and [¬AS] hold.
On the one hand, beyng’s theory of grounding remains a foundationalist one, as in
the case of being. Since we follow Bliss and Priest (2017) in defining foundationalism
as the theory which includes foundational elements (FEx) and since beyng behaves as
a foundational element, this new account of grounding can be seen as a particularly
 heidegger’s grund: (para-)foundationalism
extreme version of foundationalism. On the other hand, such a foundationalist

approach is inconsistent because it has inconsistent structural properties: it endorses
[AR] and its negation [¬AR], and [AS] and its negation [¬AS]. We call this incon-
sistent form of foundationalism, para-foundationalism.
5 Seyn as Being Self-Identical: Para-Foundationalism 2.0

Heidegger’s foundationalist approach is very different than the para-foundationalist
one: as we have seen, the former is a consistent theory and it uses a consistent
foundational element (being) while the latter is an inconsistent theory and it uses an
inconsistent foundational element (beyng). However, these two grounding theories
have something in common. Both of them take the foundational element (being or
beyng) to be what determines entities as entities. Nevertheless, in the very late period
of his career, Heidegger starts to characterize the foundational element in a slightly
different way. Appealing to the principle of identity, he equates beyng with being self-
identical. In his Identity and difference (1969), Heidegger claims that all entities are
in virtue of the fact that each entity is itself (cf. Heidegger 1969, p. 28). This idea is
expressed by the formula A = A, which should be read as “every A is itself ” or as “A is
A” (Heidegger 1969, p. 25). Heidegger also specifies that “with this ‘is’ [namely the ‘is’
contained in ‘A is A or ‘every A is itself ’], the principle [of identity] tells us how every
being is, namely: it itself is the same with itself. The principle of identity speaks of
the being of beings” (Heidegger 1969, p. 26). Even more explicitly, Heidegger thinks
that: “As a law of thought, the principle [of identity] is valid only insofar as it is a
principle of Being that reads: to every being as such there belongs identity, the unity
with itself ” (Heidegger 1969, p. 26). Now, beyng (interpreted as being self-identical)
still determines entities as entities because all entities are entities in virtue of the fact
that they are exactly what they are. Any entity is that specific and unique entity that
is: the reason for an entity to be an entity is its being identical to itself. However, if we
work with this new understanding of beyng, we have a new (and more extreme) form
of para-foundationalism. Let’s see why.
First of all, if beyng determines all entities as entities, and if beyng is being self-
identical, since all entities are entities, all entities are self-identical. Secondly, according
to the inconsistent account of the foundational element, since beyng is an entity
and not an entity, beyng is self-identical and not. Moreover, since, in Heidegger’s
metaphysics, everything is an entity only, with the exception of beyng (which is an
entity and not), everything is self-identical only with the exception of beyng (which
is self-identical and not). Now, given this framework, all the structural properties
previously attributed to para-foundationalism hold. Beying grounds all entities and,
because beyng itself is not an entity, it is ungrounded. From here, [AR] holds. [AS]
holds as well because, if all entities depend on beyng and beyng does not depend
on any entity (not even on itself), no elements depend on each other. However, since
beyng is also an entity, beyng depends on itself and, thus, both [¬AR] and [¬AS] hold
filippo casati 
too. Finally, [¬E] holds because, since beyng is both ungrounded and self-grounded,
something (namely beyng) does not depend on anything else. In other words, the
negation of Extendability holds because, since beyng is ungrounded, it is not the case
that everything depends on anything else. Indeed, beyng itself does not.
Until here, everything is exactly the same as in the para-foundationalist case
previously described in Section 4. However, if we carefully look at the definition of [E],
we will see that, when beyng is understood as a synonym of being self-identical, this
structural property holds as well. So, consider [E], which is defined in the following
way:
[E] : ∀x∃y(y = x ∧ x → y).
Now, this structural property says that everything depends on something else. It
is important to underline that, according to [E], all entities depend on something
which is not themselves (and this is guaranteed by the first conjunct of the for-
mula, that is y = x). Exactly for this reason, in both foundationalism and para-
foundationalism 1.0, the negation of [E] holds. According to foundationalism, being
is simply ungrounded and this means that it is not the case that everything depends
on something else. Indeed, being does not. According to para-foundationalism 1.0,
the negation of [E] holds because beyng is both ungrounded and self-grounded. This
means that it is still not the case that everything depends on something else. In other
words, beyng does not depend on anything else (in virtue of its being ungrounded) and
it depends on itself (in virtue of its being self-grounded). In both cases, the negation
of [E] holds because there is something that does not depend on something else,
namely beyng. However, the situation changes if beyng (or what determines entities as
entities) is interpreted as being self-identical. In this case, beyng itself is self-identical
(because it is an entity) and not-self-identical (because it is not an entity). Now, on the
one hand, since beyng is self-identical, the negation of [E] holds; even though beyng
depends on itself, it is not true that everything depends on something else. On the
other hand, since beyng is not self-identical as well, [E] holds. From the fact beyng is
not self-identical, it follows that beyng grounds something other than itself. In other
words, if beyng is not self-identical, extendability holds because the first conjunct of
the formula for [E] (namely y = x) is satisfied. Therefore, since beyng is both self-
identical and not self-identical (as discussed in the present Section), both [E] and
[¬E] hold.
Let’s sum up. On the one hand, when the foundational element (beyng) is just
characterized as what determines entities as entities, we have a form of para-
foundationalism in which [AR] and its negation, and [AS] and its negation, hold.
Moreover, [¬E] holds. On the other hand, when the foundational element (beyng)
is understood as being self-identical, we have a particularly extreme form of para-
foundationalism. Such a new form is stronger than the previous one for two main
reasons. First of all, it has more inconsistent structural features: beside the features of
the previous version of para-foundationalism, it has also the structural property [E].
 heidegger’s grund: (para-)foundationalism
Secondly, it is more radical than the previous form of para-foundationalism because

it is inconsistent on that specific structural property, namely [E], the negation of
which was meant to characterize all forms of foundationalism.10 This second form of
para-foundationalism shows that all forms of foundationalism do not have [E], even
though, in some extreme forms, both [E] and its negation hold. Finally, from both
forms of para-foundationalism also follow a new consideration about Heidegger’s
account of (PSR). In the case of being, Heidegger has correctly claimed that (PSR)
cannot unrestrictedly hold. As we have seen, if we interpret the principle ‘nothing is
without a reason’ as ‘nothing is without a ground’, there is, at least an element, namely
being, for which (PSR) does not hold: indeed, being is ungrounded. Being is without
any reason. However, according to both forms of para-foundationalism, this is not
the case anymore. As we have seen, since beyng is inconsistent, it is both ungrounded
and self-grounded. On the one hand, because beyng is ungrounded, it is without any
reason. Therefore, (PSR) fails. On the other hand, because beyng is grounded (or,
more precisely, self-grounded), it has a reason (namely itself). Since all entities are
grounded in beyng and beyng is both ungrounded and self-grounded, all entities
have a reason in beyng, and beyng has its reason in itself. Even though something
(namely beyng) both has a reason and not, it is still true that everything has a reason.
Therefore, (PSR) holds.
6 Seyn as the Last God

Before concluding, it may be interesting to point out that the present discussion
about being (with its entailed foundationalism) and beyng (with its entailed para-
foundationalism) can help us to have some new insights into one of the most
enigmatic figures of Heidegger’s philosophy, that is, the last God.
Very often, at least in the Western philosophical tradition, the ground of all entities
(or the reason why human beings, trees, sunsets, stars, planets, and everything else
are) goes under the name of God. For instance, according to Aquinas, God wills
the existence of all things, while, according to Leibniz, God is the Perfection Prior,
which generates the world as a necessary logical consequence of his essential nature
(cf. Lovejoy 1936, p. 319). Schelling believes that God is that eternal realization
[Realwerden] or Genesis of the world (cf. Lovejoy 1936, p. 320), while Meister Eckhart
(who has highly influenced Heidegger) thinks that God is the ground [Grund] of
everything (cf. Vannini 2007). Even though, in all these cases, God is thought of
as the universal ground, Heidegger has never borrowed any term from theology.
His conception of the universal ground always goes under the name of ‘being’ or
10 This is true because if we have the structural property [E], it follows that it is impossible to have
foundational elements. Since foundationalism is characterized as the view according to which there are
foundational elements, [E] is incompatible with foundationalism.
filippo casati 
‘beyng’. The reason for his divorce with the theological tradition is clearly explained in
Phenomenology and theology (Heidegger 1998, pp. 39–63). Since Heidegger holds the
idea that the ground of every entity is not an entity itself, he thinks that God cannot
be such a universal ground because, in the theological framework, God is treated as
an entity, namely as something that is. God is what wills, generates, or realizes every
thing; God is the ground of all entities. Of course, God is not a normal entity among
entities. For instance, God is eternal and normal entities are not. Even more simply,
God is the ground of all entities while a normal entity (as a table) is not. So, even
though God is often conceptualized as a super-ens, God is still thought of as an ens.
Since we refer to God, God is an entity and this is why we can pray to Yahweh, we
can move a war in the name of Christ, and we can distinguish what is wrong from
what is right following the commandments of Allah. We refer to God because, in
the first place, God is treated as an entity. Such an account of God is incompatible
with the Heideggerian assumption that the ground of everything is not an entity and,
from this, also follows the necessity of separating God (the super-ens) from being (or
beyng), which is not an ens. Therefore, being (or beyng) is not God. In Heidegger’s
jargon, we could say that traditional theology is ultimately an onto-theology, namely
the study of the divine element as an entity. As such, theology is just another science.
Heidegger writes: “We call the sciences of beings as given – of a positum – positive
science. (. . .)” (Heidegger 1998, pp. 41–2). Since theology is about a positum, namely
God, “which is essentially disclosed in faith” (Heidegger 1998, p. 49), then “theology
is science” (Heidegger 1998, p. 45). From this point of view, theology deals with God
as an ‘object of study’ while the ground of all entities is not an object at all. Therefore,
God is inappropriate as a characterization of being (or beyng).
Having said that, the approach proposed by Heidegger in Phenomenology and
theology (1998) does not seem to hold for the whole trajectory of his philosophy.
Indeed, in Contributions to Philosophy (2012a), Heidegger himself returns to theology
accepting what he calls the last God. Unfortunately, he does not give any clear account
of what the last God really is. On the one hand, we know that it is not simply another
God because it is “wholly other than past [Gods] especially other than the Christian
one” (Heidegger 2012a, p. 319). The last God is characterized, paraphrasing Hölderlin,
as the closest but, at the same time, the most remote one. “The extreme remoteness
of the last God is a peculiar nearness” (Heidegger 2012a, p. 326). On the other hand,
we also know that the last God is tightly connected with the idea of beyng [Seyn]
because “the last God is the extreme venture of the truth of beyng” (Heidegger
2012a, p. 326). Since Heidegger never rejected his condemnation of theology, at
this point, some questions look inevitable: how can Heidegger hold the position
that ‘God’ cannot be another name to refer to beyng, believing, at the same time,
that the last God is actually a synonym of beyng? And, if so, what is the difference
between the traditional account of God and the last God? These questions remain
unanswered.
 heidegger’s grund: (para-)foundationalism
One possible way of making sense of the thoughts expressed in Contributions

to Philosophy (2012a) is to appeal to para-foundationalism. The last God may be
interpreted as the dialetheic foundational element that we have called beyng. If this
is the case, since beyng (or the last God) is not an entity, Heidegger can still hold his
previous critique against theology. However, since beyng is also an entity, it is actually
possible to refer to it as God because, consistently with the theological tradition, God
is that entity in virtue of which everything is. From this point of view, the last God is
the closest one because it is an entity and the world of human beings is constituted
by entities. Entities are the things we are most familiar with. Nonetheless, the last
God is also the most remote one because it is not an entity as well and, as we have
seen at the beginning of Section 4, what is not an entity leads to an aporia, which
is accepted as true in the para-foundationalist view. In other words, the last God is
phenomenologically present to the human beings and, at the same time, it is not. On
the one hand, it is phenomenologically present as all the other entities are: the last
God is present to the mind of the people thinking or speculating about it and, thus,
referring to it. On the other hand, the last God is not present as all the other entities are:
since it is not an entity and everything that is phenomenologically present to human
beings (or everything human beings can refer to) is an entity, the last God cannot be
phenomenologically present. Paraphrasing Heidegger, if it is true that only a God can
save us, this God must be the last one—it has to be the God that grounds everything
including itself without being grounded at all.
7 Conclusion
To conclude, let’s summarize what we have done in this paper. As we have claimed
in the Introduction, even though the contemporary debate has given rise to a lot
of different grounding theories, all these theories have a common feature: they
are consistent. Based on a development of Heidegger’s metaphysics, this paper has
suggested two new accounts of the grounding relation, which are inconsistent. First
of all, we have summarized Heidegger’s account of ontological ground. Secondly,
we have focused our attention on Heidegger’s foundationalism, according to which
every thing is grounded in being and being is ungrounded. After that, we have also
shown that Heidegger’s foundationalism faces a problem because, according to his
own metaphysical premises, being is inconsistent (being is an entity and not). Thirdly,
working under the assumption that it is true that beyng is an object and not an
object, we have revised Heidegger’s foundationalism in order to have two ground-
ing theories, which can accommodate the fact that what grounds every thing (in
Heidegger’s terms, beyng) is a dialetheia. We have labeled these theories ‘para-
foundationalism’. Finally, we have used ‘para-foundationalism’ to interpret one of the
most obscure parts of Heidegger’s philosophy, namely his understanding of the so-
called last God.
filippo casati 
8 Technical Appendix
Someone could worry that the two inconsistent theories of grounding presented in
Section 4 and Section 5, exactly in virtue of the fact that they are contradictory,
fall into logical triviality. The following two models (one for each kind of para-
foundationalism) show that this is not the case. These models are set up using a first-
order interpretation for the paraconsistent logic LP (Priest 1979). Now, in order to
show that para-foundationalism 1.0 and para-foundationalism 2.0 are not trivial, we
adopt the following strategy. First of all, we confirm that all the structural properties of
para-foundationalism 1.0 and para-foundationalism 2.0 hold in the relative models.
This means that these structural properties take either value t (true) or value b (both
true and false). Secondly, we show that there is at least one sentence which takes value
f (false).
Let’s start by discussing the model of para-foundationalism 1.0. To define such a
model MPF1.0 = D, V
, let the domain of interpretation D be {e, b}. As in the graphs
presented above, e represents an entity and b represents beyng. V(→) and V(=) are
as described in table (b) and table (c). As we can see in Figure 15.3, comparing the
graph (a), which represents para-foundationalism 1.0, and table (b), entity e does not
depend on itself (e → e is false) and beyng b does not depend on entity e either
(b → e is false). However, entity e depends on beyng (e → b is true) and beyng both
depends and does not depend on itself (b → b is both true and false). Moreover, as
is clear comparing graph (a) and table (b), identity behaves in a consistent way. Thus,
both entity e and beyng are self-identical (e = e and b = b are true) and beyng is not
identical to entity e (b = e is false).
At this point, it is easy to show that, in our model, the structural properties of para-
foundationalism 1.0 (namely, [AR], [¬AR], [AS], [¬AS], [¬E]) hold, taking either
→ e b
e f t
e b f b
(b)
b = e b
e t f
(a) b f t
(c)
Figure 15.3.
 heidegger’s grund: (para-)foundationalism
value t or value b. Let’s see. For Anti-Reflexivity ([AR]: ∀x¬(x → x)), we consider all
the possible values of x:
• V¬(e → e) = t
• V¬(b → b) = b
As it is clearly understandable looking at graph (a), it is true that entity e does not
depend on itself and it is both true and false that beyng depends on itself. In all cases,
[AR] holds. For the negation of Anti-Reflexivity ([¬ AR]: ∃x(x → x)), it is enough
to consider one value of x for which [¬ AR] holds. As we have already noticed from
graph (a), beyng depends on itself and it does not depend on itself. Thus, V(b → b)
takes value b. For Anti-Symmetry ([AS]: ∀x∀y(x → y ⊃ ¬y → x)), we consider all
the possible values of x and y:
• V(e → b ⊃ ¬b → e) = t
• V(b → e ⊃ ¬e → b) = t
• V(e → e ⊃ ¬e → e) = t
• V(b → b ⊃ ¬b → b) = b
The fact that Anti-Symmetry holds is intuitively understandable from graph (a) as
well. First of all, entity e depends on beyng but beyng does not depend on entity e.
Secondly, even considering the case where x is y, entity e does not depend on itself
and beyng does and does not depend on itself. Thus, [AS] holds. For the negation of
Anti-Symmetry ([¬ AS]: ∃x∃y(x → y ∧ y → x)), it is enough to show that there is at
least a value of x and a value of y for which [¬ AS] takes value t or b. Now, as is shown
by graph (a), beyng depends on itself and does not. Thus, V(b → b ∧ b → b) takes
value b. [¬ AS] holds as well. Finally, let’s consider the last structural property of para-
foundationalism 1.0. For the negation of Extendability ([¬ E]: ∃x∀y(x → y ⊃ x = y)),
we consider all the values of y and at least one value of x for which [¬ E] takes either
value t or value b:
• V(b → b ⊃ b = b) = t
• V(b → e ⊃ b = e) = t
As we can see, [¬ E] holds. Indeed, since the negation of Extendability claims that
something does not depend on anything else, from graph (a) it is clear that, on the
one hand, beyng depends and does not depend on itself; on the other hand, beyng
does not depend on anything else other than itself.
At this point we know that all the structural properties of para-foundationalism
1.0 hold. In order to show that the model presented here is not trivial, we need
to show that there is a sentence which is false only. This is not difficult because,
as we have already argued in Section 4, we know that, in this first kind of para-
foundationalism, Extendability [E] does not hold. Indeed, as graph (a) shows, it is
false that everything depends on something else: for instance, beyng does and does
filippo casati 
not depend on itself. In both cases, it does not depend on something else. Thus, in
MPF1.0 , ∀x∃y(y = x ∧ x → y) is false because V(¬b = b ∧ b → b) takes value f.
Now, what about a model of para-foundationalism 2.0? This second model,
MPF2.0 , is not very different than the one presented above and it is defined as D, V
.
As in the previous case, the domain of interpretation D is {e, b} where e represents
an entity and b represents beyng while V(→) is as described in table (b). This is
the case because para-foundationalism 2.0 has exactly the same features of para-
foundationalism 1.0: in both cases, entity e does not depend on itself (e → e is false)
and beyng b does not depend on entity e either (b → e is false). Moreover, entity e
depends on beyng (e → b is true) and beyng both depends and does not depend on
itself (b → b is both true and false). What is different is that, in para-foundationalism
1.0, both entity e and beyng are self-identical (as table (c) shows, e = e and b = b
are true) while, in para-foundationalism 2.0, even though entity e is still self-identical
(e = e still takes value t), beyng is both self-identical and not (b = b takes value b).
Thus, for para-foundationalism 2.0, V(=) is described by the following table:
= e b
e t f
b f b
At this point, as we have done before, we start checking if all the structural prop-
erties of para-foundationalism 2.0 ([AR], [¬AR], [AS], [¬AS], [E], and [¬E]) hold,
taking either value t or value b. Now, since the only thing that changed from para-
foundationalism 1.0 to para-foundationalism 2.0 is the behavior of identity, we need
to check only Extendability and its negation. Besides Extendability and its negation,
all the other structural properties (already verified in para-foundationalism 1.0) do
not make use of identity; therefore, they are verified in para-foundationalism 2.0
too. Let’s begin considering Extendability. So, according to Extendability, everything
depends on something else. On the one hand, entity e depends on beyng. On the other
hand, beyng both depends and does not depend on itself. In this second case, it may
seem that Extendability does not hold: indeed, in this model, either beyng does not
depend on anything or beyng depends on itself. In both cases, beyng does not depend
on anything else. However, we should not forget that, in para-foundationalism 2.0,
beyng is not self-identical. This means that, when beyng depends on itself, beyng
depends on something other than itself because beyng is also not itself—beyng is
also not self-identical. This is the reason why Extendability holds. Formally, for [E]:
∀x∃y(y = x ∧ x → y), we consider all the value of x and at least one value of y for
which [E] takes either value t or value b:
• V(¬b = b ∧ b → b) = b
• V(¬b = e ∧ e → b) = t
 heidegger’s grund: (para-)foundationalism
For the negation of Extendability, namely [¬ E]: ∃x∀y(x → y ⊃ x = y), we consider

all the values of y and at least one value of x for which [¬ E] takes either value t or
value b:
• V(b → b ⊃ b = b) = b
• V(b → e ⊃ b = e) = t
As we can see, [¬ E] holds. To conclude, we show that, also in para-foundationalism
2.0, there is a sentence which is simply false. This is easy. In MPF2.0 , consider the
sentence ∀x(x → x). This sentence is simply false because V(e → e) takes value f.
References
Bliss, R. and Priest, G. (2017), ‘Metaphysical Dependence, East and West’, in S. Emmanuel (ed.),
Buddhist Philosophy: a Comparative Approach, Basil Blackwell, pp. 63-87.
Bliss, R. and Trogdon, K. (2014), ‘Metaphysical Grounding’, in E. Zalta (ed.), Stanford Encyclo-
pedia of Philosophy, http://plato.stanford.edu/entries/grounding/.
Braver, L. (2012), Groundless Grounds, A Study of Wittgenstein and Heidgger, Cambridge, MA:
MIT Press.
Caputo, J.D. (1986), The Mystical Element in Heidegger’s Thought, New York: Fordham
University Press.
Casati, F. (2016), ‘Being. A dialetheic interpretation of the late Heidegger’, Doctoral Thesis,
University of St Andrews.
Casati, F. and Fujikawa, N. (2015), ‘Better than Zilch?’, Logic and Logical Philosophy 24: 255–64.
Casati, F. and Wheeler, M. (2016), ‘The Recent Engagement between Analytic Philosophy and
Heideggerian Thought: Metaphysics and Mind’, Philosophy Compass, 11: 486–98.
Corkum, P. (2013), ‘Substance and Independence in Aristotle’, in B. Schnieder, A. Steinberg,
and M. Hoeltje (eds), Varieties of Dependence: Ontological Dependence, Supervenience, and
Response-Dependence, Basic Philosophical Concepts Series, Munich: Philosophia Verlag,
pp. 65–96.
Heidegger, M. (1969), Identity and Difference, New York, NY: Harper and Row.
Heidegger, M. (1977), Martin Heidegger: Basic Writings, New York, NY: Harper and Row.
Heidegger, M. (1996), Being and Time, Albany, NY: State University of New York Press.
Heidegger, M. (1998), Pathmarks, Cambridge: Cambridge University Press.
Heidegger, M. (2012a), Contributions to Philosophy (of the Event), Bloomington, IN: Indiana
University Press.
Heidegger, M. (2012b), Storia dell’essere, Milan: Marinotti.
Lovejoy, A.O. (1936), The Great Chain of Being: A Study of the History of an Idea, Cambridge,
MA: Harvard University Press.
McDaniel, K. (2009), ‘Ways of Being’, in D.J. Chalmers, D. Manley, and R. Wasserman (eds),
Metametaphysics, Oxford: Oxford University Press, pp. 290–319.
McDaniel, K. (2015), ‘Heidegger and the “there is” of Being’, Philosophy and Phenomenological
Research, 92(2): 1–15.
McDaniel, K. (2017), The Fragmentation of Being, Oxford: Oxford University Press.
filippo casati 
Moore, A.W. (2012), The Evolution of Modern Metaphysics: Making Sense of Things, Cambridge:
Priest, G. (1979), ‘The Logic of Paradox’, Journal of Philosophical Logic, 8: 219–41.
Priest, G. (2002), Beyond the Limits of Thought, Cambridge: Cambridge University Press.
Priest, G. (2015), ‘The Answer to the Question of Being’, in J.A. Bell, A. Cutrofello, and P.M.
Livingston (eds), Beyond the Analytic-Continental Divide: Pluralist Philosophy in the Twenty-
First Century, New York, NY: Routledge, pp. 249–58.
Takho, T. and Lowe, E.J. (2015), ‘Ontological Dependence’, in E. Zalta (ed.), Stanford Encyclo-
pedia of Philosophy, http://plato.stanford.edu/entries/dependence-ontological/.
Vannini, M. (2007), Mistica e Filosofia, Florence: Le Lettere.
Index of Names
Achinstein, P. 119 Charlier, C. 261

Adams, R. M. 183–4 Chengguan 131
Aikin, S. E. 86n.47 Clark, M. J. 182n.1, 199
Anaxagoras 259 Cobreros, P. 102
Aquinas, T. 2, 79–81, 169n.8, 178nn.36, 38, Connes, A. 262n.25
187n.7, 304 Corkum, P. 293n.4
Aristotle 2, 62, 136–7, 241, 293 Correia, F. 72, 141n.5, 150n.20, 182n.1,
Armstrong, D. 14, 56n.16, 57–8, 67, 133n.25, 248n.16, 255, 263, 277, 278n.6
195, 206, 245, 277–8 Cotnoir, A. J. 241n.8, 264n.28
Asaṅga 128 Cowling, S. 196n.19
Audi, P. 108n.2, 113n.7, 114, 143n.9, 182n.2,
218n.2, 277, 278n.6 d’Albe, F. 261
Daly, C. 108n.2, 115, 199
Bacon, A. 264n.28 Dasgupta, S. 12n.21, 18, 28, 83nn.39–40,
Barker, S. 199, 205–8 85n.42, 86n.48, 87n.49, 113n.7, 140, 143,
Barnes, E. 5n.9, 6, 13–15, 38, 43, 48–9, 55n.12, 168n.4, 168n.6, 170n.13, 172, 179, 189,
103, 108n.2, 212, 277, 278n.8 229n.14, 277n.3, 284
Benacerraf, P. 220 Davies, B. 187n.7
Bennett, K. 3n.3, 52, 53nn.4, 7, 54, 55n.14, 62, Davies, M. 275n.1
81n.34, 184n.3, 189, 201, 238n.3, 241n.6, Davis, C. 190n.13, 196n.21
244n.12, 251n.21 Dehmelt, H. 251, 260n.18
Bliss, R. I. 2n.2, 4n.6, 7n.13, 12, 17nn.32–3, Della Rocca, M. 178–80
18n.38, 24n.41, 38, 49, 53n.6, 72n.5, 74n.12, Democritus 257
77n.23, 85nn.43–5, 86n.47, 103, 117, 126n.1, Denkel, A. 58n.21
132n.21, 167, 168n.4, 170n.12, 171–2, 174, deRosset, L. 38, 141nn.4–5, 148, 277
178n.38, 182n.1, 184n.4, 187, 199, 219n.4, Descartes, R. 2
241n.6, 244n.12, 254n.1, 260, 291–2, Digges, T. 261n.21
293nn.4–5, 296, 301 Dixon, T. S. 6n.12, 7n.14, 9n.15, 74n.11, 168,
Bohn, E. D. 26n.46, 27, 175–6, 177n.33, 187n.7, 184n.4, 212–14, 219nn.4–5, 244n.12,
240, 258 251n.21, 255n.5
Borges, J. L. 94, 104 Dorato, M. 242n.10
Boyle, R. 257n.11 Dreier, J. 108
Brzozowski, J. 186 Dushun 131
Buddha (Siddhārtha Gautama) 127, 129,
133 Einstein, A. 91
Burgess, J. P. 144, 225n.11 Escher, M. C. 98
Button, T. 249n.18 Esfeld, M. 118–19, 247–8
Evans, G. 281n.14
Cameron, R. 18, 25, 53, 54nn.8–9, 75n.19, 76,
80n.31, 81n.36, 110, 111n.5, 116n.8, 169, Fazang 127, 131
171n.17, 180, 182, 218, 242–3, 277–9, Fine, K. 6, 16n.30, 18, 20, 45, 51–2, 54,
281n.12 56n.15, 62, 64, 72, 74–5, 85n.43, 92n.1, 96,
Candrakı̄rti 135–6 108, 113n.7, 114–15, 140–1, 142n.8, 145, 148,
Caputo, J. D. 294, 298 170–3, 183, 185n.5, 189, 199, 201–3, 205,
Carr, B. 176n.30, 177n.33 209, 218n.1, 221n.6, 222, 238, 244n.12,
Casati, F. 19, 29, 31, 293n.6, 299n.8, 246–8, 251, 254n.2, 277, 278n.6, 281,
300n.9 282n.16
Caulton, A. 264n.28 Frege, G. 84
Chan, W. T. 131n.17 French, S. 59n.24, 247–8, 265n.29
Chang, G. C. C. 131nn.18–19 Fujikawa, N. 293n.6
 index of names
Garrett, B. 137n.37 Litland, J. E. 16, 39, 113n.7, 140, 141n.3,

Gassendi, P. 257n.11 141n.5, 143n.9, 146n.14, 147n.17, 148,
Gödel, K. 91 173nn.21–2, 179n.41, 189, 222, 255
Gratton, C. 85n.45 Lombard, L. 60n.26
Griffith, A. M. 255 Lopez de Sa, D. 279
Guigon, G. 178–9 Lovejoy, A. O. 13n.23, 304
Lowe, E. J. 5, 25n.44, 53nn.5–6, 63, 65, 66n.30,
Haraway, D. 67n.31 126n.1, 135n.31, 191, 202, 208–9, 244n.13,
Harding, S. 67n.31 248n.16, 256n.9, 278n.6, 291n.1
Haslanger, S. 67n.31 Lynn, R. J. 138n.41
Hawthorne, J. 243 Lyre, H. 267, 268n.33
Healey, R. 117, 119, 264n.28
Heidegger, M. 2, 19, 29, 31, 137–8, 291–302, MacBride, F. 66–7, 225n.11
304–6 Maitzen, S. 20n.40, 24nn.42–3
Heil, J. 277 Maraldo, J. C. 30n.51
Heisenberg, W. 260 Markosian, N. 242
Hellman, G. 220, 225n.11 Maurin, A.-S. 58n.18
Hempel, C. 120 McDaniel, K. 54, 241n.5, 278n.8, 292n.3,
Herschel, J. 261 295n.7
Herzberg, F. 269n.37 McKenzie, K. 247–8, 250, 259n.15, 260, 265
Hippolytus of Rome 42n.3 Mellor, H. 249n.18
Hobbes, T. 257n.11 Mitchell, D. 127n.5
Hornsby, J. 60n.25 Moore, C. 133n.24, 293n.6
Hume, D. 20, 173–4 Morganti, M. 3, 28, 72n.5, 74n.12, 77n.23, 110,
170n.12, 171n.17, 172, 178n.38, 187, 212,
Ismael, J. 117 242n.10, 244n.12, 260, 265n.29
Mortensen, C. 98n.4
Jago, M. 27, 199, 205–8 Muller, F. 267, 268n.33
Jantzen, B. 188n.9
Jenkins, C. 11n.19, 38, 44, 48–9, 75, 77, Nāgārjuna 127, 129–30, 134n.27, 135
88n.50, 103, 109n.3, 218, 255, 263, 277 Newton, I. 257n.11
Nishida, K. 2, 30
Kant, I. 84, 132, 135–6, 261 Nolan, D. 15–16, 97n.2, 171, 174, 250n.19, 279
Keefe, R. 12n.22, 24n.43, 71n.1
Kemp Smith, N. 132n.22 O’Conaill, D. 242n.10, 245
Keränen, J. 144, 225n.11 Olbers, H. W. 261n.21
Kim, J. 72n.6, 113–14, 192–3, 240n.4 Olsson, E. 109
Klein, P. 187
Koslicki, K. 51, 54, 62, 168n.4, 199 Parsons, J. 94
Kripke, S. 275n.1, 281n.14 Paseau, A. 75, 80n.31, 242n.11
Pasnau, R. 64n.29
Ladyman, J. 118, 226n.12, 238, 242–3, 245, Passmore, J. 24n.41, 86n.47
247–8 Patzia, M. 42
Lam, V. 118–19, 247–8 Paul, L. A. 58n.19
Lambert, J. H. 261 Pereboom, D. 191
Leibniz, G. 2, 12, 82–3, 162, 183–4, Pettigrew, R. 144n.10
293, 304 Plato 2, 28, 96, 137
Leitgeb, H. 226n.12, 227n.13 Poplawski, N. J. 261–2
Lesher, J. H. 42 Priest, G. 2n.2, 7n.13, 14, 16, 18n.38,
Leucippus 257 29nn.49–50, 30n.52, 73n.8, 76, 131n.18,
Leuenberger, S. 81n.33, 229n.15, 288n.27 132nn.21, 23, 134n.30, 137nn.36–7, 138n.42,
Lewis, D. 84, 109, 113, 121, 203, 207, 213, 240, 212, 291–2, 293nn.4–5, 295n.7, 296,
278, 281, 283 300–1, 307
Liggins, D. 182n.1, 199 Proctor, R. 261
Linnebo, Ø. 59, 144n.10, 221–5, 231–2 Pruss, A. R. 173–4
Lipton, P. 121 Putnam, H. 134n.29
index of names 
Quine, W. V. O. 117 Smolin, L. 262n.24

Spinoza, B. 2, 241
Rabern, B. 6n.12, 38n.2, 74n.11, 168, 184n.4, Stambaugh, J. 137nn.39–40
185, 219nn.4–5, 255n.5 Steinberg, A. 277n.4, 280n.11
Rabin, G. O. 6n.12, 16, 38n.2, 74n.11, 168, Suarez, F. 64n.29
184n.4, 219nn.4–5, 255n.5 Swedenborg, E. 261
Radhakrishnan, S. 133n.24
Raven, M. 6n.11, 11n.18, 15n.27, 75n.15, 108, Tahko, T. E. 5–6, 27–8, 74n.12, 126n.1,
113n.7, 114, 167, 168n.4, 177, 179n.42, 135n.31, 170n.12, 172n.17, 178n.38, 237n.1,
180n.45, 182n.1, 244n.12, 251n.21, 277, 244n.13, 245, 247n.15, 248nn.16–17, 250,
278n.6 251n.22, 256n.9, 260n.17, 291n.1
Rayo, A. 55 Teller, P. 119
Resnik, M. 118, 221 Thompson, N. 3, 5, 14–15, 73n.7, 110–11, 115,
Restall, G. 102 212, 255
Roberts, B. W. 268n.33, 269 Trogdon, K. 5, 27, 28n.47, 72, 79n.29, 108,
Rodriguez-Pereyra, G. 110, 212, 255 113n.7, 114, 126n.1, 167, 168n.4, 169n.10,
Rosen, G. 52–3, 54n.8, 62, 113n.7, 141n.5, 167, 179n.41, 182n.1, 194n.17, 196n.19, 199,
170–1, 175, 179, 182n.2, 199–200, 203, 209, 219n.4, 241n.6, 254n.1, 282n.15, 288n.27,
218n.2, 277, 278n.6, 282n.16 291n.1
Ross, D. 118, 238, 242–3, 245, 247
Rucker, R. 93–4, 98 Ullian, J. 117
Salmon, W. 113, 121, 185n.5 Vaags, R. H. 178n.37

Sanford, D. 94, 97n.3 van Fraassen, B. 120
Schaffer, J. 4nn.5, 7, 15, 18, 23, 26n.46, 27, 38, van Inwagen, P. 97n.3, 281n.13
46, 49, 50n.1, 52–4, 58n.20, 62, 66n.30, 72, Vannini, M. 298, 304
77nn.24, 27, 81n.36, 85n.43, 103, 111, 113n.7, Varzi, A. 133n.26
117, 136–7, 141n.5, 143n.9, 168–9, 171, von Neumann, J. 100
175nn.23–6, 177n.31, 178nn.38–9, 182n.2,
183–6, 187nn.7–8, 188, 194n.17, 195, 199, Wheeler, M. 260, 299n.8
201, 212, 218, 240–2, 245, 247n.15, 250, 255, Wigglesworth, J. 28
258, 277–8, 280, 281nn.12–13 Wildman, N. 28–9, 275n.1, 282n.17
Schnieder, B. 52–3, 56, 66n.30, 182n.1, 218n.1, Williams, D. C. 58n.18
277 Williams, J. R. G. 277
Schopenhauer, A. 83n.38 Williams, P. 127n.5
Sen, A. 264 Williamson, T. 196, 284n.23
Shapiro, S. 59n.22, 118, 220–1, 225–6, 228 Wilsch, T. 283n.21
Shoemaker, S. 191 Wilson, J. M. 16n.29, 54n.8, 108n.2, 168n.4,
Sider, T. 6, 175n.24, 176n.28, 177n,31, 199, 199, 238, 244, 251, 257n.10, 277
277–8 Witt, C. 67n.31
Siderits, M. 127n.5–6, 128n.7, 129n.10 Wolff, J. 265
Silesius, A. 298 Woodward, J. 113
Simons, P. 58n.21
Skiles, A. 28n.47, 179n.41, 229n.15, 282n.16, Xenophanes 42
283n.20, 288n.27
Slaney, J. K. 102 Yablo, S. 98, 205
General Index
Abgrund 295 beings 3n.3, 12, 19n.39, 23, 28–9, 83, 132,
abstract 137–8, 208, 292n.2, 293–6, 298–9, 302,
entities 183, 195–6 304–5, 308
foundationalism 196 beliefs 3, 12, 77, 97, 109–10, 112–13, 117, 122,
idea of ontological minimality 249n.18 140, 187, 193, 257, 262, 295
mathematical features 260 beyng 300–9
mathematical structures 220 bicollective ground 16, 140–8, 152–4, 161–2
objects 95, 136, 195n.18, 278 Big Bang 91–2, 99, 104, 121, 136, 137n.34,
offices 223–4 261–2
structures 220, 223–4 biological 47
subject matter 10 camouflage 41
world 219 entities 135
abstracta 196n.19, 276 facts 116
abstraction 14, 223 functions 146
action theory 170 phenomena 40, 46
acyclicity 141, 148–9, 152–3, 155, 159–60, 274 processes 40, 48
anti-foundationalism 3, 27, 42n.3 remains 48
anti-instantiation 206 states 37
antimatter 261 systems 1
anti-reflexivity (AR) 2–3, 7–12, 16, 22, 29, biology 37, 39–41, 43–4, 48, 91
129, 207, 209, 291, 297–8, 301–3, black holes 261–2
307–8 bootstrapping 11, 62, 71, 81–2
anti-symmetry (AS) 2–3, 7–16, 29, 38–9, boundaries 131, 261
42–3, 45, 48, 91, 94, 97, 109n.3, 129, Bradley’s regress 186
137, 264n.27, 265, 291, 297–8, 301–3, brain 46, 48, 75, 100, 174
307–9 dorsal stream 44
Aristotelian universals 62, 64–5, 127 identity theories 88
Arrow’s theorem 264 phenomena 44
asymmetry 13, 39, 51, 54, 57, 63, 67, 70, 74–5, states 38, 75, 87–8
80, 96, 97n.3, 99, 101–5, 108–12, 119, Brutalism 178–9
121–2, 147, 156, 168, 189, 212–14, 217–19, brute facts 82, 116–17
254–5, 262, 264n.27, 265 Buddhism
of causation 120 Abhidharma 126–30, 132–4
of dependence 50, 53–5, 61–2, 64, 81, 241, Chan (Zen) 127
244, 259 Huayan 14, 126–7, 130–1, 136, 137n.35
of explanation 66n.30, 121–3 Madhyamaka 126–7, 129–30, 134–5
atomism 8, 10, 28, 132–3, 237, 241–5, 247, Mahāyāna 127, 130, 134n.28
249, 257 non-well-founded 129
atoms 92, 111, 131–4, 175–7, 237–8, 240, Theravada 127
244–5, 258, 280n.11, 282 well-founded 127
automorphism 144, 223n.9, 225 Buddhist
autonomous facts 87n.49, 284 coherentism 130
dependence 126–7
bare mass 58–9 ideas 136
being 2–3, 11–12, 18–19, 21–5, 27, 30, 54, li 131, 136
56–7, 72, 75, 81–2, 93, 95, 97, 108, 126n.1, metaphysics 126
128, 133, 137–8, 169–75, 182–5, 190–1, orthodoxy 134
194–5, 201, 203, 207, 209–12, 260, 269, philosophy 126, 138
278, 291–2, 293n.6, 294–306 positions 132
general index 
shi 131 chemical

thought 127, 131 facts 40, 116
traditions 14, 16, 127 phenomena 46
views 132, 138 processes 40
building relation 52 properties 40–1
bundle theory 58–9, 62 reactions 291
states 37
categories 4–5, 8, 26, 72n.3, 107n.1, 132, 182, systems 1, 194
218, 278 chemicals 176
category-neutral view 182n.2 chemistry 37, 39–41, 43, 91
causal Chinese
activity 191, 194, 196 Buddhism 127, 130
capacity 183, 190–5 traditions 2, 14
chain 163 circles of ground 70–1, 73–5, 77, 79, 84, 87, 96,
closure 192 99, 140
competition 193 circular
dependence 17, 38, 42n.3, 50, 134–5, accounts 71n.1
240n.4 arguments 53n.6, 66, 70
exclusion principle 192–3 definitions 212–13
explanation 20–2, 108, 113, 120–1 dependence 54, 263n.26
factors 135 explanation 13, 65, 88, 143
foundationalism 27, 183, 191–2, 194–6 reasoning 70
generalizations 104 circularity 65–6, 70–2, 88, 212
history 121, 170 clusters 54, 261
inheritance premise 192–3, 195 of dependence 58
inputs 210 of explanatory worries 84, 87
law 83n.38 of galaxies 176, 260–1
loops 91 of interdependence 59
mechanism 108 of theories 291
overdetermination 192n.15 of views 2
parallel 151n.21 coherence 105, 109, 201, 263n.26
powers 191, 192n.15 of models of grounding 91
relations 113, 129, 135, 192, 211 theory of truth 70
roles 210–11 coherentism 3, 8–10, 14, 25, 31, 65, 77, 107,
series 187n.7 110, 112, 122, 130–1, 140, 262–3, 265,
structure 4 267n.32, 268–9, 291
tendencies 211 coherentist 3, 112, 122, 140, 264n.28, 266, 268
transactions 191 approach 109–10
causality/fundamentality premise 192–3 models 257, 263
causation 4n.5, 11–12, 38n.1, 79, 104, picture 14n.25
113, 120, 129, 134–6, 172n.17, 189, solutions 262
210, 254 structures 16, 28, 263
C-fibres 24, 211 colour 190, 249
chains 15, 21–3, 26, 65, 73, 91–2, 102–3, 110, composition 132–4, 136, 157–60, 176n.29,
116, 155, 167, 188, 214, 220, 233, 243, 246 206–7, 243, 246, 257n.12, 264n.28, 279
of beliefs 3, 109, 122 computation 41–2, 44
of dependence 1, 17–18, 21, 31, 237–40, conception see layered conception
242, 244 conceptual
of entities 1, 3, 6 connections 4
of ground/grounding 28, 31, 76, 84–5, 95, constructions 128, 133–4
100, 109, 111–12, 117, 170, 171n.17, 172, dependence 50, 119, 134
176–8, 180, 184, 187, 212–13, 217, 219, entities 205
233–4, 251n.21, 255, 258–60, 269 falsehood 97n.3
chains of being 172–5 fiction 128
global question (GQ) 172–4, 178n.36 imposition 135
local question (LQ) 172–3, 178n.36 landscapes 93
regional question (RQ) 172–3, 178n.36 overlays 130, 136
 general index
conceptual (cont.) Dedekind abstraction principle 223

possibility 75 Deductive-Nomological account of
primitives 189 explanation 120
reality 128 dependence 13, 37–8, 52, 54n.8, 68, 77, 199,
stipulation 97 220, 248, 250, 256n.8
tools 97 chains 1, 17–18, 21, 31, 237–40, 242, 244
truths 97 clusters 58
concreta 194n.17 relations 2, 4, 6–7, 10–12, 14–16, 18, 30–1,
concrete 40, 50n.1, 51, 66–7, 70, 79, 113–14, 119,
biological happenings 40 137, 221, 224, 239–42, 244, 247, 256–7,
chemical happenings 40 262–3, 266, 268–9, 291, 293, 295, 296n.7,
entities 194–6, 257 297, 301
foundationalism 27, 183, 194–6 structure 2, 6, 29, 39, 245, 248
objects 95, 191, 195n.18, 220, 268, see also Buddhist dependence, causal
277–8, 280 dependence, circular dependence,
premise 195 conceptual dependence, identity
principle 194–6 dependence, mereological dependence,
world 175–7, 219 metaphysical dependence, ontological
Confucianism 127 dependence, symmetric dependence
conjunctions 22, 38, 140, 162, 174, 199–203, dependence/fundamentality
206–7, 209–10, 214, 218 debates 10
conscious experience 45 literature 17n.31
consciousness 2, 30, 44, 129, 135 dependent 7, 10, 13, 18, 52, 53n.4, 55, 57,
consistency 2, 28–30 59–60, 64, 67, 73, 116, 118, 120, 134, 184,
constitution 96, 101, 203, 205, 214 204, 221, 241, 243, 256n.8, 266, 269n.36,
formal 202–5, 208–9, 213–14 280, 282, 287, 291
logical 207 designation 130
material 203, 208, 214 entities 6, 19, 21–5, 55, 86, 183
non-mereological 221 facts 85–6
of the One 96 objects 268
contingency 2, 28–9, 52, 83, 210, 279–81 origination 127
contingent fundamentality thesis 275–6, determinate–determinable relation 92
279–88 determination 45, 108, 151, 192n.16
contingentist 276, 283, 287–8 dharmas 2, 128–9, 131–2, 134–5
response 282 diachronic
thesis 284–5 dynamic physical relation 170
contradiction 8, 29–30, 132, 153, 156, 159–60, inheritance 188, 189n.12
170, 172, 175–6, 178, 196, 207, 209, 292, dialetheia 30, 104, 292, 300, 306
299–300 Ding 292n.2
corpuscularianism 257n.11 disjunction 142, 199, 201, 206–7, 209–10,
cosmic 256n.8
inflation 261 entities 200, 202, 214
loops 16, 91–9, 101–5 states 206, 211
pattern 176–7 dispositions 191, 210–11
cosmological distribution 99
arguments 19–20, 22, 173 failure 142
models 260–2 of properties 207–8
questions 19–20
cosmology 176, 261–2 ecology 117
cosmos 94–5, 99, 101, 239, 241, 257–8, 280 economics 1, 37, 39, 264
cycles 18, 153, 183, 263–4 Einstein–Cartan theory of gravity 261–2
eleaticism 195
Daoism 126n.2, 127, 130 eliminability 177, 251n.21
De Morgan eliminative
equivalence 206 approach 220, 265
rules 212 distinction 224
decomposition 128, 142n.8, 260, 264n.28 ontic structural realism 265, 267
general index 
emanation 101 knowledge 114

emergentist model of being 260 models 14, 67
epistemic notions 115
coherentism 3, 77, 107, 112, 140 power 27, 116, 262, 270
dimensions 78 praxis 15, 79
explanation 113, 122 principle 62
foundationalism 3, 77 progress 87
grounds 213 projects 19, 116
infinitism 3, 77, 187, 260 reasons 12, 71, 78, 84, 89
innocence 79 regress 85–6
justification 109, 112 relations 5, 13, 70, 107–8, 113–15, 118, 222
phenomenon 120 resources 259
position 121 structure 64–5, 107, 116, 118, 123
problems 13 systems 121
properties 187 targets 19, 22, 86
reasons 66, 78–9 worries 84, 87
requirement 212 extendability 6–10, 291, 298, 303, 308–10
structural realists 265
view 265 Finean account of dependence 55–6, 59
epistemology 3, 65, 67n.31, 77, 107, 113–15, Forms 2, 95
122, 257, 262, 263n.26, 265, 269n.37, 283 foundation 38n.2, 42, 122, 130n.14, 177,
Ereignis 300 182, 219–20, 234, 258, 260n.17, 262, 269,
essence 12, 27–8, 52, 55, 82, 84, 87, 140n.2, 286, 296
179, 193, 200–2, 205, 208–10, 225, 248, foundational 130, 251n.21
249n.17, 256, 269, 281–3, 295, 299 elements 8–9, 129, 295–7, 299–303, 304n.10,
essentialism 52, 56, 60n.26, 62, 64, 248, 306
280, 287 epistemology 3, 77
ethics 17n.31, 138 facts 107, 109–10, 112, 117
explanation 4–5, 11–15, 17–25, 27, 37, 55, metaphysics 3, 77
59n.24, 65, 70–3, 76, 78–9, 81–3, 85–9, monists 257n.12
108, 112–23, 142–3, 148, 168n.3, 170–1, objects 9
173–4, 178–9, 182–3, 189–90, 202, 213, foundationalism 2–4, 6–10, 13, 19–21, 24–7,
221–3, 254n.3, 256n.9, 258, 265–6, 269–70, 29, 31, 38–9, 42, 65, 76–7, 108–9, 115–17,
283, 286 see also causal explanation, 123, 129–30, 169n.8, 178n.36, 251n.21,
circular explanation, grounding 254, 257–8, 259, 262, 267n.32, 269, 291–2,
explanation, holistic explanation, 295–8, 301–4, 306 see also abstract
metaphysical explanation, reductive foundationalism, causal foundationalism,
explanation, symmetric explanation concrete foundationalism, epistemic
explanatoriness 120, 168, 171n.14, 172 foundationalism, metaphysical
explanatory foundationalism, para-foundationalism,
arguments 143, 147–50 pluralist foundationalism
breakdown 82 Fregean view of structured propositions 88
chains 21, 23, 116 fundamentalia 3n.3, 6, 12, 21, 25, 28–9, 31, 76,
challenge 286 81–3, 244–5, 275–6, 278, 282n.18, 284–6,
circularity 88 288, 297–8
claim 56 fundamentality 1–2, 5–6, 10, 13–14, 18, 25,
concerns 17, 71 47–8, 53–5, 58, 64, 68, 75n.19, 80–1,
connections 117 86n.48, 168n.6, 170, 177, 182, 185–6, 188,
considerations 13, 20–1, 115–16 191–2, 199, 237–8, 240–5, 247–8, 249n.17,
convenience 227 250, 275–81, 284
demands 89 absolute 238n.3, 241
dimension 110 arguments in defence 19–21, 25, 85
failure 71, 84, 86–7 as a primitive 238n.3
holism 50, 58, 65, 67 as lower boundedness 6
inferences 85, 143, 147 as non-quantitative 278
interests 122 as quantitative 278
intuitions 114, 120, 122 as well-foundedness 6
 general index
fundamentality (cont.) graphs 40–1, 49, 149–50, 152–3, 155, 157

contingent 275–6, 279–88 left-collective 140–2, 146, 148, 151–2,
general theory 238 155, 162
inheritance arguments 182–3 loops 73, 78, 89, 91–2, 103–4, 178
intersection with modality 29 operators 140n.1, 142n.7
intuition 243, 252 orthodoxy 37, 44, 48, 108
layering 37, 47 relations 5, 16n.29, 40–1, 43, 46–8, 67,
mereological 240, 248 72, 74, 82, 85, 91, 99, 102–3, 105,
modal status 275, 287 107–15, 169, 182n.2, 200, 212, 217–18,
modal strength 278, 288 220, 222, 225, 228, 233, 255, 257,
paradigmatic accounts 28 259n.16, 260, 264, 268, 288, 291, 296n.7,
relative 47, 54, 80–1, 241, 243 297, 306
standard accounts 28 right-collective 140
statements 278 standard account 212
thesis 2, 7, 17–19, 27, 31, 240, 244, 275–6, standard notion (SNG) 167, 169–80, 268,
279, 287 276, 277n.4
without foundations 251n.21 structures 6, 29, 48, 85, 91, 110, 184, 214,
255–7, 268, 285, 299, 301
gauge theories 267 theories 30–1, 93, 144, 291–2, 302, 306
general relativity 91, 98, 261–2 Grounding-Definition Link principle 200
generic essential existential dependence ground-theoretic notion of
(GEED) 248, 250 well-foundedness 255n.5
generic ontological fundamentality Grund 291–3, 304
(GOF) 245–7, 250–1 Grundsatz 296
geological gunk 27, 112, 176–7, 183, 213, 239–40, 242,
phenomena 42 245n.14, 246–7, 259n.16, 262, 286 see also
properties 48 hunk, junk
geology 41–2, 44, 48 gunky
God 2, 6, 9, 12–13, 22, 28, 51, 62, 82–4, 87, objects 183, 214
94, 169n.8, 173, 178n.36, 238, 276–7, 284, ontology 242, 251
294, 307 region 95
last God 292, 304–6 scenario 175
graphical worlds 111–12, 243, 259n.16, 281, 286
approach 161
depiction 148 haecceitism 266
representation 46, 152 hierarchical
semantics 162 conception of the world 37
graphs 40–1, 46, 49, 141, 149, 152–3, 155–7, cosmological models 260–2
159–62, 217, 226–33, 297–8, 300, 307–8 structure 37, 46–7, 81, 217, 241, 244, 257
graph-theoretic structure of elements 293
account 162 structure of levels 237–8, 240
approach 141–2, 146, 148, 156 structure of reality 1–2, 11, 14–16, 27, 80,
framework 154 237–40, 245, 247
operation 230 view of reality 255n.7, 259
graph theory 226, 230 hierarchy 14, 46–7, 80, 92, 109, 176–7, 185,
Great Chain of Being 1, 13n.23, 31 218, 238, 240–2, 245, 254, 267
Greeks 13, 42 of being 278
grounding of dependence 38
as a unitary concept 255 of levels 37, 269, 277
as generative 189 thesis 2, 7, 11, 16, 28
as not unitary 293 holism 50, 58, 65, 67, 113, 116–17, 207,
chains 31, 85, 111–12, 170, 172, 177–8, 184, 260, 269
246, 251n.21, 258 holistic
claim 227, 229–31 explanation 50, 53n.6, 58, 59n.24, 65, 67,
conditions 27, 200, 202, 207, 210, 214 107, 112–13, 116, 118–23, 143
explanation 118n.10, 113, 117, 118n.10, 119, form of coherentism 263n.26, 268
120n.11, 122, 168n.3, 185 metaphysical explanation 117, 119, 122
general index 
structure 116, 118, 121–3, 268–9, 270 irreflexivity 11, 38–9, 44–5, 48, 50, 70–1, 74–5,
system 113, 119, 121 77–81, 89, 91, 96, 99, 101, 103–5, 108, 160,
homunculus theory of perception 174 168, 170n.12, 178, 192, 212–14, 217–18,
Humean supervenience 281 254–5, 257, 263–5
hunk 27, 175n.27, 176 see also gunk, junk isomorphism 223, 225, 228, 231
hunky 262
ontology 240 Japanese traditions 2
scenario 176 junk 27, 112, 175n.25, 176n.29, 177, 239–40,
worlds 175, 177, 259n.16 246, 259n.16 see also gunk, hunk
hypergraphic account junky 111, 240
bicollective case 140, 152 scenario 175–6
left-collective case 148 universes 262
hyperintensionality 52–3, 88n.50, 108 worlds 112, 175, 259n.16
account of dependence 51
account of tracking 114 Kantian scruples 133
of grounding 115 Kyoto School 2
idealism 134–5 Lagadonian worldmaking language 207

identity 16, 45, 81, 88, 113, 118–19, 134–6, layered conception 16, 37–49
160–1, 176n.29, 203–4, 209, 222, 226–31, layering 37–8, 41, 43–9
233, 256, 260, 266–9, 302, 307, 309 Leibniz’s law 162
conditions 223, 228, 231–2, 265 levels 26–7, 37, 80, 92, 94, 194, 217–18,
dependence 53, 63, 221, 225 237–40, 244–6, 255n.7, 260n.18, 261, 264,
mapping 223n.9, 225 266, 269
of entities 221–3, 225, 232 Lockean bare particulars 266
of numbers 233 Loopism 178
of objects 202, 222, 225–8, 231–2, 234, 265 loops 7–9, 16, 31, 38, 48, 73, 78, 84–6, 89,
of propositions 162 91–105, 170n.12, 178, 257, 263, 264n.28
of structures 228, 231–2, 234
relations 71 mathematical
structuralism 25, 44–5, 75, 88, 266 entities 208, 220–2, 225–7, 229, 232,
in virtue of 21, 53–4, 59, 67–8, 72, 75, 108, 111, 260, 267
134, 137, 167, 169, 171, 200, 202, 208, 210, objects 118, 217, 220–32
222, 226, 254–6, 257n.10, 264, 292–6, 298, ontology 43, 55, 59
300, 302–3, 306–7 structuralism 28, 66, 118, 141, 143–4, 146,
indefinitely descending ground (IDG) 167, 162, 217, 219–21, 224n.10, 225, 227–9,
170–80 231–4
Indefinitism 178 structures 59, 63, 118, 217, 220, 222–4,
independence 6, 53n.4, 54, 64, 78, 118, 132, 226–8, 230–2, 261n.20
220, 238, 241 mental 37, 50–1, 128, 178n.40
Indian traditions 2 entities 205
induction 157, 159, 177n.31, 200, 258 facts 72, 167
inductive/abductive reasons 176–7 imposition 135
ineliminability 177, 251n.21, 277 phenomena 44
infinitism 3, 9–10, 25–8, 31, 77, 110, 112, states 75, 199–200, 203
178n.38, 187, 239–40, 242, 244n.12, 250, mereological
257, 259–60, 262, 269–70, 291 atomism 132, 237, 241–5, 247, 249
inheritance 27, 169, 188–95 atoms 175, 177, 238, 240, 244–5, 280n.11
arguments 182–5, 187–8, 190, 195 dependence 134, 240–1, 248
premise 101, 183, 185, 186–8, 191–3, elements 245, 247
195–6 essentialism 52, 280, 287
Intelligences 95 fundamentality (MF) 240–2, 244–8, 251
intransitivity 16, 46, 99, 102, 263 hierarchy 176–7, 238, 240–2, 244
intuition 10–11, 13, 17, 104, 111, 114, 120, independence 241
122, 133, 135, 141n.3, 143, 169, 171, levels 94, 240, 245–6
172n.17, 193, 199, 221, 238, 243, 252, nihilism 133
258, 264n.28, 266 parthood 204–5
 general index
mereological (cont.) freedom 288

parts 203–6, 213, 243, 254 notions 52, 114, 199, 248n.17
relations 129, 257, 259, 264n.28, 280 objections 280, 281n.12
scale 240–1 properties 72
structures 145, 177, 250, 270, 283, 287 space 276
sums 136, 201, 204, 206, 209, 214 status 28, 72, 117, 284, 287
whole 204, 213 strength 275, 278–9, 288
mereology 95, 104, 132–4, 194, 203–4, 237, variation 285–6
240, 246, 257n.12, 264n.28 modality 28–9, 52, 64, 81, 176n.28, 210, 275–6,
meta-inductive reasoning 258 284, 287
metaphysical monism 62, 76, 112, 117, 175n.26, 194n.17,
coherentism 3, 14, 77, 140, 262, 265, 267n.32 201, 241–2, 245, 257n.12, 280, 283n.22
dependence 2, 4–6, 11–12, 14–17, 28, 30, Mūlamadhyakamakārikā (MMK) 129–30,
37–8, 42n.3, 61, 70, 107, 117, 126–7, 135, 134–5
291, 295 mutual dependence relations 266, 269
explanation 5–6, 18–21, 25, 45, 50, 52,
65–7, 70, 72–3, 78–9, 81, 86–7, 88n.50, natures 51n.3, 73n.9, 82, 95, 200–3, 205, 207–9,
89, 107–8, 113, 115–17, 118n.10, 119–20, 213–14
122, 168, 171, 173–4, 178–80, 189, 202, necessitarian
212, 218, 222–3, 254, 256n.9, 257n.10, thesis 284–6
263, 277 views 287
foundationalism 2–3, 10, 20, 22, 25, 27, necessitarianism 83
76–7, 86, 108–10, 112, 116–17, 123, 182–5, necessitation 28n.47, 64n.28, 142, 168, 173,
187, 191–2, 194–6, 239, 244, 257 254, 288
fundamentality 199 necessitism 173, 179
impossibilities 97, 112, 175–6, 230 neo-Aristotelian 56, 195
infinitism 3, 25–6, 76–7, 110, 187n.8, neo-Confucianism 127
239–40, 242, 244n.12, 250, 260 neo-Platonist 95–6
interdependence 107–8, 110–12, 115–19, never-ending questions objection 22–3
121, 123, 212 non-eliminativist structuralism 59, 223–4, 232
necessity 96, 109, 112, 179, 210, 229–30, non-foundationalist
232, 281 metaphysical structures 270
possibilities 4, 10, 91, 97n.2, 175, 221, 229 terms 255
principle of sufficient reason (MPSR) 27, view 116
178–80 non-fundamental 185, 202, 287
relations 75, 97, 104, 217 entities 182–3, 191–2, 194, 196, 201,
structural realism 118–19 203, 213
structure 4, 28, 31, 107–8, 123, 217, 255n.6, facts 168, 180
257–9, 262, 263n.26, 269–70 properties 192–3, 210
theories 14, 25, 55, 61, 179n.43, 208 non-hierarchical
verifiers 145 dependence relations 268
metaphysicians 13n.24, 50–1, 61, 76, 81, holism 269
243, 277 non-Humeans 189
metaphysics 1, 13, 44, 53, 56–8, 60, 62, 65, non-material entities 208–9, 213–14
70–1, 76, 82, 84, 105, 138, 195, 201, 206, non-mereological
247–8, 270, 288n.29, 292–4, 296–9, composition 206–7
302, 306 constitution 221
mind/brain identity theories 75, 88 holism 269–70
mind-independence 78–9 non-monotonicity 108, 168, 254n.3
modal non-orthodox conception of ground 48–9
accounts 64, 232 non-symmetry 14–15, 54n.11, 55, 61–2, 64–5,
analyses 51–2 67, 105, 120, 212
concepts 51 non-well-foundedness 28, 71, 75, 78,
connections 56n.15 84, 110, 112, 129, 167–8, 214, 217,
definitions 52 219, 233–4
differences 204 nothingness 2, 30, 137n.36, 171, 173–4
force 281 Nyaya–Vaisesika school 257
general index 
Objectum 292n.2 principle of

Olbers’ paradox 261 Amalgamation 141, 150
ontic physical causal closure 192
ground 292–4 Self-Ground 147
structural realism (OSR) 247–50, 265–9 Squeezing 147
ontological (Strict) Amalgamation 150
categories 72n.3, 107n.1, 182 sufficient reason (PSR) 20, 27, 83, 117, 153,
dependence 1, 4–6, 16n.30, 18, 50–1, 53–5, 178, 292, 296, 304
63n.27, 75n.19, 80n.31, 108, 126n.1, 128, proper–parthood relation 102
138, 242, 244, 248, 256, 266, 277–8, 296n.7 psychological
difference 295 facts 116
elements 245 phenomena 41–2, 46
emergence 55 states 37
ground 19, 76, 292–5, 300, 306 psychology 37, 39, 42, 51, 91, 117
independence 237–8
minimality 28, 237, 240, 244–7, 249–51 qualitativism 143
priority 71, 80n.31, 134, 257, 268 quantum
chromodynamics 267
para-foundationalism 29, 291–2, 298, 302–4, entanglement 242n.10
306–10 field theory 251
parthood 111, 175–6, 203–5, 246–7 gravity 262
particles 37, 93, 96, 133, 135, 171, 176, 177n.31, mechanics 104, 117n.9, 119, 262, 267
237–8, 250, 257n.11, 258, 260, 262, 266–8, particles 266
269n.36 states 119, 137n.34
particulars 40, 43, 48, 57–8, 63–4, 88n.52, Quinean webs of beliefs 262
206–7, 266
part–whole 92–4, 96, 97n.3, 102–5, 206, 254, realism 59, 63, 118–19, 211, 247, 265, 269
257n.12, 262 reality/fundamentality premise 185–6,
Peano’s axioms of arithmetic 186n.6 188, 190
perception 135, 174, 249, 293n.6 reality inheritance 27, 183–90, 195
permutation invariance 266–7 realization 192n.15, 224, 232, 300, 304
physicalism 24, 37–8, 199 reductionism 88
physics 28, 37, 39, 59n.24, 91, 104, 116–17, reductive
136, 176, 208, 238, 242–3, 249–50, 258, explanation 88, 171, 174
260–2, 264n.28, 266 grounding 171, 173
Platonic reflexivity 8, 11–13, 16, 22, 29, 44–6, 48, 70–1,
entities 268 73–5, 77–82, 85, 87–9, 103, 109n.3, 210,
heaven 56 218, 233, 256, 263–4, 268–9
pluralism 28, 76, 112, 241–2, 244–5, 257, 270, regress 9, 17–18, 71, 78, 84–6, 133, 170–1,
280, 292n.3 174–5, 186, 209, 243, 246, 260
pluralist 241, 280, 281n.12 relational 247
foundationalism 257 formulation 108
opponents 201 predicate 93
possible worlds 52, 56n.15, 65, 81, 92, 188, 221, properties 282
279n.10, 280, 286n.25 structures 276
pratı̄tyasamutpāda 127, 134 view 72, 74
predicate satisfaction principle 22, 24 relationalism 143
primitive 50, 52, 65, 168, 188–9, 227, 241, Rucker
249n.17 loop 93, 96, 101–4
constituents 207, 210 world 100, 103
fundamentality 238, 248
identities 266, 269 same kind objection 22, 24–5
metaphysical grounding 114 same questions objection 22, 23–4
notions 45, 189, 206, 277 Sein 292, 295, 298, 300
predicates 208, 227–8 self-being 128–9
relation 108 self-dependence 11–13, 70–1, 75–9, 82, 84,
vocabulary 211, 214 87, 265
 general index
sentential explanation 119–20, 122

connective 4, 72–3, 79, 87, 93, 96, 218n.1 ground 43–6, 48, 73, 122, 233
operator 108, 142n.7, 144n.11, 254n.2 symmetry 8, 11, 59n.23, 63, 66, 99, 103, 120,
Seyn 298, 300, 302, 304–5 250, 256, 264n.27, 267–8
Shifty Shaky 275, 285–7
simple principle 42–3, 45, 47–8 Tarski’s benign truth regress 186n.6
S-matrix 260, 269 taxonomy 7, 9–10, 29, 31, 128–31, 291–2
sorites paradox 101–2, 264 the last shall be first (TLSBF) 94–6
space–time 94–6, 98, 207–8, 243, 262, 266 transitivity 2–3, 7–11, 15–16, 29, 38–9, 44,
Standard Model of particle physics 238, 250, 46–8, 50, 70, 73–4, 80, 91, 94–7, 99, 101–5,
260n.18, 267 108, 110–11, 129, 131, 136, 161, 168,
standard notion of grounding (SNG) 168–80 170n.12, 178, 212–14, 217–18, 233, 246–7,
standard view 3, 10–11, 13, 28, 86, 241 255, 263–5, 268, 291
structural tropes 58–9, 62–5, 128
features 28, 303 truthmaker semantics 140–1, 145–7, 151, 154,
grounds 144, 151 161–2
properties 2, 7, 14, 217–18, 233, 291–2, 297, truthmakers 205, 245–6, 249, 277
301–2, 303, 304, 307–9
realism 59nn.23–4, 118–19, 247, 265, 269
universals 43, 56–7, 62–5, 137
relations 225–7, 229–31
universe 21, 91–6, 99, 131, 176, 188, 220,
similarities 13
237–8, 241, 245, 260–1, 262n.24, 270
truths 144
universes 93, 100, 261–2
structuralism 28, 59, 66, 118, 141, 143–4, 146,
153–4, 162, 217, 219–29, 231–4, 240, 247,
251, 265–6, 267n.32 vicious infinite regress 17–18, 71, 78, 85–6, 170
supervaluationist approaches 102 viciousness 71
supervenience 11, 16n.29, 81, 114, 199, 240n.4,
248, 254, 281 well-founded grounding (WF) 168–72,
svabhāva 128–31, 134–5, 137 174–80, 184, 214, 219
symmetric well-foundedness 6, 38n.2, 39, 71, 74, 109–10,
dependence 2, 14, 43, 48, 50, 59–60, 62–7, 112, 168–9, 212–14, 217–19, 233n.16, 234,
137, 248, 249n.17, 256 251n.21, 255–7, 259

Untitled

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Untitled

Uploaded by

Copyright:

Available Formats

OUP CORRECTED PROOF – FINAL, 6/4/2018, SPi

Reality and its Structure

Reality and its

List of Contributors vii

0. The Geography of Fundamentality: An Overview 1

Part I. The Hierarchy Thesis

Part II. The Fundamentality Thesis

13. The Structure of Physical Reality: Beyond Foundationalism 254

Part III. The Contingency and Consistency Theses

Index of Names 313

Elizabeth Barnes University of Virginia

Ricki Bliss and Graham Priest

Reality is a rather large place. It contains protons, flamingos, economies, headaches,

 the geography of fundamentality: an overview

ricki bliss and graham priest 

sufficient for a view to count as a kind of foundationalism. For proponents of what we

 the geography of fundamentality: an overview

In what remains of this introduction, we take up the mantle of introducing and

1 The Lie of the Land

according to which grounding is a kind of causing.

of any and all categories and cross-categorically.

ricki bliss and graham priest 

 the geography of fundamentality: an overview

possible ways of understanding fundamentality.

ricki bliss and graham priest 

AR AS T E Comments Special Cases

 the geography of fundamentality: an overview

Consider, next, the Comments column. Here’s what it means.

ricki bliss and graham priest 

follows.15 Let X0 = {x : FEx}, and for any natural number, n ∈ ω: x ∈ Xn+1

the obvious way.

 the geography of fundamentality: an overview

• A final special case. Let x y iff x → y ∨ y → x. Then x and y are connected

3 On the Metaphysical Possibility of the Alternatives

ricki bliss and graham priest 

In what follows, although appeal to intuition is often made in defence of one

18 See Raven 2013 for a well-articulated defence of the hierarchy thesis.

 the geography of fundamentality: an overview

for causation here does not (necessarily) go for metaphysical dependence.20 As

ricki bliss and graham priest 

Being and its centrality to the development of Western metaphysics.

 the geography of fundamentality: an overview

that reality is hierarchically structured that the structuring relation is anti-symmetric.

inspired, in part, by Huayan.

ricki bliss and graham priest 

necessarily anti-symmetric. Indeed, according to Barnes and Thompson, explanation

26 Schaffer 2012, p. 127.

 the geography of fundamentality: an overview

ontological dependence and grounding.

ricki bliss and graham priest 

3.2 The Fundamentality Thesis

.. regress arguments

a more sustained elaboration of these thoughts.

 the geography of fundamentality: an overview

unsatisfactory exactly when we have failed to explain everything that we need an

ricki bliss and graham priest 

 the geography of fundamentality: an overview

ricki bliss and graham priest 

 the geography of fundamentality: an overview

ricki bliss and graham priest 

 the geography of fundamentality: an overview

41 See Bliss 2013 and Passmore 1970.

ricki bliss and graham priest 

.. arguments from theoretical virtue