Fundamental Causation Physics, Metaphysics, and The Deep Structure of The World

Fundamental Causation
Fundamental Causation addresses issues in the metaphysics of deterministic

singular causation, the metaphysics of events, property instances, facts,
preventions, and omissions, as well as the debate between causal reductionists
and causal anti-reductionists. The book also pays special attention to
causation and causal structure in physics. Weaver argues that causation is a
multigrade obtaining relation that is transitive, irreflexive, and asymmetric.
When causation is singular, deterministic, and such that it relates purely
contingent events, the relation is also universal, intrinsic, and well-founded.
Weaver shows that proper causal relata are events understood as states of
substances at ontological indices. He then proves that causation cannot be
reduced to some non-causal base and that the best account of that relation
should be unashamedly primitivist about the dependence relation that
underwrites its very nature. The book demonstrates a distinctive realist and
anti-reductionist account of causation by detailing precisely how the account
outperforms reductionist and competing anti-reductionist accounts in that it
handles all of the difficult cases while overcoming all of the general objections
to anti-reductionism upon which other anti-reductionist accounts falter. This
book offers an original and interesting view of causation and will appeal to
scholars and advanced students in the areas of metaphysics, philosophy of
science, and philosophy of physics.
Christopher Gregory Weaver received his PhD in philosophy from Rutgers

University (2015), where he completed his dissertation Essays on Causation,
Explanation, and the Past Hypothesis under Barry Loewer (chair), David
Albert, Tom Banks (physicist), and Jonathan Schaffer. He has published (or
has forthcoming) two book chapters (one with Joshua Rasmussen) and many
peer-reviewed articles in such venues as Erkenntnis, the Journal for General
Philosophy of Science, Metaphysica, Synthese, and the International Journal
for the Philosophy of Religion. Weaver is currently Assistant Professor of
Philosophy at the University of Illinois at Urbana-Champaign.
Routledge Studies in Metaphysics
The Future of the Philosophy of Time

Edited by Adrian Bardon
Properties, Powers and Structures

Issues in the Metaphysics of Realism
Edited by Alexander Bird, Brian Ellis, and Howard Sankey
The Puzzle of Existence

Why Is There Something Rather Than Nothing?
Edited by Tyron Goldschmidt
Neo-Davidsonian Metaphysics
From the True to the Good
Samuel C. Wheeler III
Neo-Aristotelian Perspectives in Metaphysics

Edited by Daniel D. Novotný and Lukáš Novák
Nominalism about Properties

New Essays
Edited by Ghislain Guigon and Gonzalo Rodriguez-Pereyra
Substance and the Fundamentality of the Familiar

A Neo-Aristotelian Mereology
Ross D. Inman
Philosophy of Time and Perceptual Experience

Sean Enda Power
Physics, Metaphysics, and the Deep Structure of the World
Christopher Gregory Weaver
For a full list of titles in this series, please visit www.routledge.com

Physics, Metaphysics, and the Deep
Structure of the World
Christopher Gregory Weaver

First published 2019
by Routledge
711 Third Avenue, New York 10017
and by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an informa
business
© 2019 Taylor & Francis
The right of Christopher Gregory Weaver to be identified as author of
this work has been asserted by him in accordance with sections 77 and
78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced
or utilised in any form or by any electronic, mechanical, or other
means, now known or hereafter invented, including photocopying and
recording, or in any information storage or retrieval system, without
permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks
or registered trademarks, and are used only for identification and
explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
A catalog record for this book has been requested
ISBN: 978-1-138-21313-5 (hbk)
ISBN: 978-1-315-44908-1 (ebk)
Typeset in Sabon
by Apex CoVantage, LLC
To my mother, Jackie Weaver
“And we know that for those who love God all things work
together for good, for those who are called according to his
purpose.”
Romans 8:28 ESV
Contents
List of Figures viii

Preface ix
Acknowledgments xi
1 A Metaphysical Prolegomena for the Theory of

Fundamental Causation 1
2 In Defense of the Causal Relation 89
3 The Brute Asymmetry of Causation 110
4 On the Epistemological Isolation Objection to Causal

Hyperrealism152
5 Universal Causal Determination 182
6 On the Irreflexivity, Transitivity, and Well-Foundedness

of Causation 198
7 Causal Relata215
8 On the Argument From Physics and General Relativity 252
9 Fundamental Causation 287
Bibliography 309
Index 359
Figures
3.1 Newtonian Gravitational Force #1 119

3.2 Newtonian Gravitational Force #2 119
3.3 Illustration of David Lewis’s Similarity Metric Applied 140
3.4 Illustration of Circularity 142
6.1 Causal Loop 206
6.2 Complexity of Causal Structure 207
7.1 Quark Interaction 235
9.1 Late Preemption 298
9.2 Early Preemption 299
9.3 Symmetric Overdetermination 300
9.4 Trumping Preemption 300
Preface
Singular or token causation occurs when event tokens are related by causa-
tion. The proposition <Brandon’s fall caused Brandon’s wrist fracture.> is a
singular causal fact, if true. General causation occurs when event types are
related by causation. The proposition <Smoking causes lung cancer.> is a
general causal fact. Deterministic singular causation occurs when the laws
backing causation (between event tokens) are strictly deterministic and/or
the probability that the effect occurs, given that the cause occurs (perhaps
together with the truth of the relevant laws), is one. Indeterministic singu-
lar causation occurs when the backing laws are indeterministic. This book
will be limited in scope. It provides a philosophical analysis (as defined in
chapter 1: sect. 4.4.1) of deterministic singular (or token) (full, not partial)
causation solely.
Chapter 1 provides a lengthy metaphysical prolegomenon for my study.
It presents numerous methodological principles essential to the adopted
metametaphysic. It motivates numerous philosophical doctrines needed to
help bolster subsequent argumentation in later chapters of the work. It dis-
tinguishes my approach to the study of causation from other methods in the
causation literature.
In chapter 2, I argue that (a) there are instances of obtaining causal rela-
tions. I then defend (a) and the thesis that (b) necessarily all instances of
causation are instances of an obtaining causal relation. Against (a) stands
causal eliminativism. I show that arguments for causal eliminativism from
physical considerations incur too high a cost and that E. J. Lowe’s attempt to
make true causal facts with liabilities, directed causal powers, and manifes-
tation partners alone is problematic. I then critically evaluate David Lewis’s
use of the void to show that (b) fails.
I argue in chapter 3, that the causal relation is formally asymmetric. I then
make use of the theories of gravitation in Newtonian mechanics and general
relativity, plus fundamental interactions between gluons in quantum chro-
modynamics, to show that the directedness of causation has no explanation
from a non-causally interpreted physics.
In chapter 4, I argue, pace Huw Price and Brad Weslake, that if causal
direction is completely detached from a non-causally interpreted physics, we
can nonetheless have knowledge of causal structure.
x Preface
I provide two new arguments from weak causal principles that (if those
principles are true) show that the singular (or token) causal relation is univer-
sal with respect to causal relata I call purely contingent events in chapter 5.
In chapter 6, I proffer a theory of explanation. I subsequently use that
theory to show that causation is more than likely irreflexive. I then present
a second argument for irreflexivity called the master argument. I address
potential objections to irreflexivity from time travel and causal loops just
before articulating a positive case for the transitivity of causation. I then
turn to a defense of the doctrine of transitivity by providing responses to the
supposed counter-examples to transitivity. I conclude the chapter with an
argument for the well-foundedness of causation that draws upon some of
my previous work (Weaver 2017, 107–109).
I articulate my theory of causal relata after criticizing a multitude of
analyses in the literature in chapter 7. I include there a defense of the thesis
that Feynman diagrams in quantum field theory are causal models and that
fundamental forces/interactions in quantum field theory are therefore best
understood as causally efficacious processes.
In chapter 8, I argue that the best interpretation of general relativity has
need of a causal entity (i.e., the gravitational field) and causal structure that
is not reducible to light cone structure. I suggest that this causal interpreta-
tion of general relativity helps defeat a key premise in one of the most popu-
lar arguments for causal reductionism, viz., the argument from physics.
In chapter 9, I criticize both the interventionist manipulability account of
causation in the work of James Woodward and Stuart Glennan’s mechanis-
tic theory. I then present a new anti-reductive account of deterministic sin-
gular causation, subsequently showing how that theory relates to Jonathan
Schaffer’s theory of grounding.
Acknowledgments
I’d like to thank David Black, Eddy Keming Chen, Shamik Dasgupta, Ned
Hall, Daniel Korman, Barry Loewer, Laurie Paul, Joshua Rasmussen, Daniel
Rubio, Jonathan Schaffer, Peter van Elswyk, and Dean Zimmerman for their
comments on earlier drafts of portions of the current project. I’d also like
to thank George Ellis, John D. Norton, Don N. Page, Craig Roberts, Carlo
Rovelli, Robert Wald, Aron C. Wall, and Michael Weissman for valuable
correspondence on issues related to quantum field theory, and/or cosmology,
and/or general relativity. Special thanks to Tom Banks for answering my
many questions about various topics in physics. If any details about physical
matters in the current work are incorrect, please blame only me, the author.
The current project gleans much from work I did as a PhD candidate
at Rutgers University. I am therefore grateful to the members of my dis-
sertation committee, David Albert, Tom Banks, Barry Loewer (chair), and
Jonathan Schaffer.
I presented chapter 2 at the University of Mississippi. I thank the philoso-
phy department at Ole Miss for their hospitality and for their great ques-
tions and objections. Chapter 3 was presented at the 6th Midwest Annual
Workshop in Metaphysics (MAWM) at the University of Nebraska–Lincoln.
I thank my audience there for great feedback. An earlier version of c hapter 4
was presented at Baylor University. I thank the wonderful department there
for their hospitality and feedback. I thank the Rutgers University meta-
physics reading group for their comments on an earlier draft of chapter 5.
Various versions of chapter 8 were presented at numerous places over the
course of several years, including the 2014 Eastern American Philosophical
Association conference, Yale University, the University of Illinois at Urbana-
Champaign, and Mississippi State University. I thank the audiences at each
of these events and institutions for their wonderful feedback. A version of
chapter 8 was accepted for publication by Erkenntnis. I thank Springer
Nature for allowing me to republish much of the material in that chapter.
An early version of chapter 9 was presented at the 2015 Central American
Philosophical Association. I thank my audience there as well.
I would like to thank my loving wife, Christina, and my children, Celeste,
Grant, and Esmée, for their support and encouragement during the comple-
tion of the project.
1 A Metaphysical Prolegomena for the
Theory of Fundamental Causation1
To get the metaphysics of causation right, one will need to draw upon a
metaphysical worldview. A metaphysical worldview includes a collection of
purportedly true statements (perhaps supplemented with some directives or
principles) that accurately represent a truth-aimed metaphysical methodol-
ogy, that methodology’s norms, and its commitments. It likewise incorpo-
rates a set of theses that purportedly accurately describe and explain what
exists, the nature of reality, and the hierarchy of being. I will understand
worldview building and metaphysical inquiry in general as knowledge-seeking
inquiry. Metaphysical inquiry should be aimed at delivering to cognizers
warranted true beliefs about the contents of a metaphysical worldview. Fol-
lowing many others in the epistemology literature, I will assume that war-
rant is that which (when it is of sufficient degree) distinguishes beliefs that
are merely true from knowledge.2 My metaphysical inquiry will therefore
adopt a knowledge norm of assertion,
(K-A): One ought to (or “one must”)3 assert p, only if, one knows that p.4
I will impose this norm on others to such a degree that I will interpret their
non-elliptical, unqualified, sincere, and purportedly factual statements in
such a way that if one cannot actually be warranted with respect to one’s
belief in them, or if one cannot actually have knowledge of them, then that
is a strike against them or against the theory of which they are an indispens-
able part.5
To ensure coherence and the mitigation of inaccuracy, building a meta-
physical worldview ought to be methodical even if that building is restricted
in such a way that it solely serves the purposes of theorizing rightly about
causation. I will call a metaphysical worldview, or at least those parts of a
metaphysical worldview that are built to facilitate proper theorizing about
causation, a metaphysicalC system. The presuppositions of metaphysicalC
system building should be clear, as should any principles that facilitate that
building. Those presuppositions and principles (sometimes supplemented
with obligatory directives) will be understood as parts of the metaphysicalC
2 A Metaphysical Prolegomena
system itself. Those parts represent a metametaphysic for the study of cau-
sation (i.e., they constitute what I will call a metametaphysicC). In this
chapter, I articulate and motivate a metaphysicalC system that includes
facets of the metametaphysicC needed to construct the theory of singular
deterministic causation espoused in this book. Because one could write
a substantial amount on each tenet of any metaphysicalC system, a full-
fledged explication and defense of every doctrine of that system is not my
aim. I intend only to explicate and begin to motivate and defend the most
important doctrines, citing and referring to more full-length defenses along
the way.
Section 1: Avoiding Self-Stultification in Metaphysical

Methodology: Truth and Meaning Part I
The natures of meaning and truth become relevant to the study of causa-
tion at numerous junctures. If one is a verificationist and maintains that
a necessary condition for the meaningfulness of informative declarative
sentences is that they be empirically and scientifically verifiable, then one
will have a reason to dismiss some metaphysical theories of the causal
relation that include posits that are beyond the reach of science (e.g., that
a non-corporeal mind can cause a belief to form). Likewise, if one were an
advocate of an epistemic theory of truth, which says that necessarily what
is true depends upon the cognitive activity of human persons, then one
would be forced to forsake the existence of objective, mind-independent
causal factsP.6 And a fortiori, there could be no objectively true theory or
analysis of the causal relation.
My first metaphysically significant assumption will be the relatively
uncontroversial thesis (in contemporary analytic metaphysics, at least)
that truth or true things depend on reality or being (TDB).7 We should
accept every instance of the schema: Necessarily, for any true proposi-
tion <q>, <q> holds because q. For example, <Water is H2O.> because
water is H2O. And <Electrons have charge.> because electrons have charge
(and here I follow Jonathan Schaffer’s lead in (Schaffer, Truth 2008, 305);
Jeffrey C. King (Criticisms 2014, 146; emphasis in the original) said
“[s]urely, we want to say . . . P is true at w because w is a certain way . . .
this seems like a truism”). By affirming TDB, I do not necessarily intend
to endorse a truthmaker theory (TMT) (e.g., that states of affairs, or the
stuff of the world or indeed, the world itself de re necessitate the truth of
truth-bearers, such as propositions, and thereby metaphysically explain
truths (see e.g., Armstrong 2004, 5–6, and q.v., n. 199)). I am affirming
something considerably weaker than TMT, although truthmaker theory
certainly entails TDB.
That TDB is weaker than TMT is evidenced by two facts. First, TMT, but
not TDB, requires a commitment to the existence of truthmakers (Schaffer,
Truth 2008, 305). Second, leading theoreticians like Trenton Merricks reject
TMT, although they embrace TDB as obvious and doubted by no one,
A Metaphysical Prolegomena 3
That Fido is brown is true because Fido is brown. That the Trojans were
conquered is true because the Trojans were conquered. That hobbits do
not exist is true because hobbits do not exist. . . . And so we might say
that truth ‘depends on the world’. . . . No one would deny it.8
Given TDB, the importance of truth and meaning for metaphysicalC sys-
tem building is as follows. One’s metaphysicalC system ought to admit, on
pain of epistemic irrationality, that there exist truths, and therefore mean-
ingful things, or meaning in general.9 A necessary condition for the truth of
any sentence or statement p is that p be meaningful. If one’s metaphysicalC
system S were something that we knew, then it would be something that
is true, and therefore also something that is meaningful. But the proposi-
tion <S is true.> is true because S is true. Likewise, the proposition <S is
meaningful.> is true because S is meaningful. The proposition <S is true.>
depends for its truth on a specific way reality is, the way involving S’s being
true. The proposition <S is meaningful.> depends for its truth on a specific
way reality is, the way involving S’s being meaningful. The intuition (and
see sect. 4.5.2 on the role of intuitions in my argumentation) driving these
ideas can be made manifest by considering the following questions. How
could there be any truths about metaphysicalC systems without truths and
meaningful things? How could there be an accurate and meaningful meta-
physicalC system at all without things that are meaningful or true? More
explicitly (square brackets below function like parentheses),
(Methodological Principle #1 (MP1)): Necessarily, for any cognizer C,

and for any theory T, if [((i) C believes that T), and (ii) either [(T asserts
or entails that (-M): <T is without meaning.> and C believes that T
asserts or entails (-M)) or (T asserts or entails that (-Tr): <T cannot be
true.> or (-Tr*): <There are no truths.> and C believes that T asserts
or entails (-Tr) or (-Tr*)), or (T asserts or entails that (-A): <C should
refrain from believing that T is meaningful or true.> and C believes that
T asserts or entails (-A))]], then C has an actual mental state defeater for
their belief that T holds.
An actual mental state defeater for a cognizer’s belief B is a believed reason

in favor of ~B that robs C’s belief B of warrant, thereby rendering C epis-
temically irrational with respect to their forming or retaining B.10
MP1, and the questions that motivated it, suggest the following caution-
ary directives.
(Directive #1 (D1)): One’s metaphysicalC system ought not preclude or

be agnostic about the existence of meaningful things.
Likewise,

be agnostic about the existence of truths.
If one asserts one’s metaphysicalC system in a way that abides by K-A,
and if it is possible for one to actually know one’s metaphysicalC system,
then one ought to, on pain of epistemic irrationality, allow into one’s meta-
physicalC system mental states, mental events, and instances of warranted
belief.11 This is because knowledge requires instances of warranted, true
belief. Belief is a mental state. Thus,

be agnostic about the existence of mental states (specifically beliefs),
mental events, and instances of warranted belief.12
If the guiding principle behind D3 is correct, then one’s metaphysicalC sys-

tem ought not preclude or be agnostic about the existence of mental causa-
tion either. This is because mental events like belief formation—whether
voluntary or involuntary—are acts or behaviors that causally result in and
can result from reasoning to, thinking about, forming, or sustaining a dis-
tinct belief (q.v., chapter 2: sect. 2).13 Again, ‘formation’ is a causal locution.
And when we seek to build and employ a metaphysicalC system, or when we
form a belief about a metaphysicalC system, we are doing something insofar
as we are performing acts of mental causation such as belief formation.
Stephen Yablo put the more general point I’m after this way, “[d]eny men-
tal causation and you are denying that anyone ever does anything” (Yablo
1997, 251; emphasis in the original). I therefore recommend,

be agnostic about the existence of mental causation or instances of the
knowledge of “otherly”-mental causation.14
One might resist this last directive. Donald Davidson (Mental Events
2001) argued for anomalous monism (AM), the hypothesis that (a) there
“can be” no “strict laws connecting the mental and the physical,”15 (b) every
event token is identical to a physical event token, and (c) mental concepts
receive no conceptual reduction “to physical concepts.”16 One might think
that AM is a threat to mental causation, for one common objection to it is
that it has the unhappy consequence of rendering the mental epiphenom-
enal.17 That is, if the mental is epiphenomenal, then it cannot cause any-
thing. So, AM is a threat to mental causation. However, the realist about
mental causation need not fear AM. Indeed, Davidson’s argument for AM
included a premise that asserts that there are instances of mental causation
(Davidson 1993, 3). AM entails that mental event tokens are identical to
physical event tokens. Davidson’s idea was to avoid the problem that the
absence of strict psychophysical laws poses for mental causation by forcing
the mental to earn its causal place via the physical with which it is identical
(see Maslen, Horgan, and Daly 2009, 525). Worries about the mental being
epiphenomenal only creep in once one has abandoned the thesis that the
mental is something distinct from the physical.
Neither should the proponent of D4 and its underlying methodological
principle (i.e., MP2) be afraid of the subtraction arguments of Horgan (1987,
502) and Seager (1991) (cf. the discussion in Chalmers 1996, 150–171).
They attempted to move from the possibility of a qualia-absent world w*
that is a physical duplicate of ours (call it @) to the epiphenomenal character
of qualia. Given that nothing causally reaches into @ or w* from some tran-
scendent ontological index (i.e., there are no gods, or supernatural causal
interventions of any kind), then if one can causally explain everything at
w*, qualia become causally irrelevant at w*. That irrelevance should reap-
pear at @ because w* is a physical duplicate of @, and everything is causally
explained at w* by physical entities or things. So, qualia are actually epiphe-
nomenal. The price of saving one’s theory of qualia from epiphenomenalism
is the hypothesis that qualia supervene upon the physical in such a way that
all worlds that are consistent with the laws at @ and that are physical dupli-
cates of @ must be duplicates in all mental respects.
Realists about mental causation should not fear subtraction arguments.
If one identifies the mental with the physical, or if one maintains that
the mental logically supervenes upon the physical such that there are no
qualia-absent worlds that are physical duplicates of the actual world, then
the argument can be escaped.18 I believe that arguments like these (that are
sometimes understood as threats to mental causation) are really threats to
particular theories of the mental entities involved in mental causation.19
There are plenty of maneuvers for realists about mental causation to make
in light of such threats.
Consider now,
(Methodological Principle #2 (MP2)): Necessarily, for any cognizer C

and for any theory T, if [(i) T is meaningful, logically coherent, and
logically consistent, (ii) C believes that T, and (iii) either [(T asserts
or entails that <(-MS): There are no mental states, or beliefs, or men-
tal events, or instances of mental causation, or there are no instances
of knowledge of “otherly”-mental causation.> and C believes that T
asserts or entails (-MS)), or (T asserts or entails that <(-A-MS): C cannot
know whether there are mental states, or beliefs, or mental events, or
instances of mental causation, or instances of knowledge of “otherly”-
mental causation.> and C believes that T asserts or entails (-A-MS)),
or (T asserts or entails that one should refrain from believing any of
the following theses (a) there are mental states, (b) there are beliefs, (c)
there are mental events, (d) there are instances of mental causation, (e)
there are instances of knowledge of “otherly”-mental causation and C
believes that T asserts or entails that one should refrain from believing
(a) [or any of the other claims that are in the interval (b) through (e)])]],
then C has an actual mental state defeater for their belief that T holds.
We should also be sure that our metaphysical worldview does not lead us
to abandon the existence of the theory we are sincerely advancing (assuming
the theory is meaningful, coherent, and internally consistent), nor should it
lead one to agnosticism about the existence of the theory one is asserting in
accordance with K-A. Thus,

be agnostic about the existence of that very metaphysicalC system.
And the methodological principle underneath D5 is,
(Methodological Principle #3 (MP3)): Necessarily, for any cognizer C

and for any theory T, if [(i) T is meaningful, coherent, and internally
consistent, (ii) C believes that T, and (iii) either [(T asserts or entails that
(-ET): <T does not exist.> and C believes that T asserts or entails (-ET))
or (T asserts or entails that (A-ET): <C should refrain from believing T
or C cannot know T.> and C believes that T asserts or entails (A-ET))]],
then C has an actual mental state defeater for their belief that T holds.
Our directives and principles have real teeth. They should lead one to
a metaphysical worldview that will transparently recommend jettisoning
several substantive philosophical theses. For example, the principles adum-
brated thus far lead one away from any worldview or metaphysical theory
that delivers to one the truth of eliminative materialism, the hypothesis that
there are no mental states or mental events (defended in Churchland, Mate-
rialism 1995; Postscript 1995; 1998; 2007). Likewise, the above discussion
suggests that one should not appropriate a worldview or theory that deliv-
ers to one metaphysical or ontological nihilism, the hypothesis that nothing
exists. But there is more fruit.
Section 1.1: Sider’s New-Fangled Humeanism

My metaphysicalC system includes principles and directives that partially
constitute a metaphysical methodology that is at odds with other proposals,
including Theodore Sider’s new-fangled Humeanism (Sider 2011) (call it
SHM).20 For Sider, there is a privileged set of notions that are structural in that
they carve reality at the joints. These notions constitute the ideology of the
correct book or metaphysical theory of the world (TM).21 One arrives at the
relevant batch of notions by looking to the ideologies of our best and most
empirically successful theories such as mathematics, logic, and physics.22
This is because “the conceptual decisions of successful theories correspond
to something real: reality’s structure.”23 The ontology of TM should include
those entities one’s best theories require. However, there also exists some-
thing deeply primitive in reality that is not itself directly quantified over
by those best theories, viz., structure (that to which joint-carving notions
correspond). Structure is non-linguistic and non-conceptual (ibid., 5, n. 5),
and is both objective and worldly (ibid., 188).24 Again, the ideology of TM
will correspond to pieces of the world’s structure.25 And because structure is
fundamentally real and notions in the ideology of TM are representative of
objective structure, suggesting more (complicated) ideology affects in some
way TM’s ontology. Hence, “[a] theory with a more complex ideology posits
a fuller, more complex, world, a world with more structure. Thus[,] ideo-
logical posits are no free lunch” (ibid., 14). And “[a] theory’s ideology is
as much a part of its worldly content as its ontology” (ibid., 13). And still
further,
[T]he world according to an ideologically bloated theory has a vastly

more complex structure than the world according to an ideologically
leaner theory; such complexity is not to be posited lightly.
(ibid., viii)26
My reading is at odds with other ways Sider characterizes structure. For

example, in Sider (2011, 94), he remarks, “[j]udgments about structure con-
cern ideology, not ontology”. But there is clearly some connection between
the two (i.e., ideology and ontology), for how else could Sider claim to rid us
of fundamental causation, laws of nature, and modality because “there are
no causal, nomic, or modal notions in . . . [TM’s] ideology” (ibid., 293,
his account is Humean for reasons having to do with ideological choice)?
I conjecture that the relationship is this,
[Interpretive Conjecture]: Substantive notions of the ideology of the

book of the world correspond to the structure of the world. Structure
is a real fundamental existent (stuff) in that it is part of TM’s ontol-
ogy, and its features are that to which primitive ideological notions
correspond. Some entities can enter the ontology of TM by simply
being quantified over by the structural (fundamental) existential quan-
tifier (see Sider 2011, 188, 202–203). Updating Quine, Sider can say,
“[t]o be is, purely and simply, to be the value of a variable” per the
regimented (fundamental/structural) language of our best theories.27
What about structure? Sider believes that adding structure to the ide-
ology of TM adds to the explanatory power of TM (where TM is now
some unified best theory of fundamental inquiry (e.g., physics)). The
notion of structure is structural. So, it earns its keep as a member of
TM’s ontology.
Sider’s new-fangled Humeanism includes a recommendation for achiev-

ing Humean reductions. These reductions are acquired through the pro-
vision of a metaphysical semantics for derivative truths. A metaphysical
semantics is supposed “to show how what we say fits into fundamental real-
ity.”28 Because of a commitment to the doctrine of completeness (i.e., that
“[e]very sentence that contains expressions that do not carve at the joints
has a metaphysical semantics” (Sider 2011, 116)), Sider’s project will seek
a metaphysical semantics for derivative truths (truths that aren’t in purely joint-
carving terms/notions). A metaphysical semantics provides those fundamen-
tal truths or fundamental facts “in perfectly fundamental terms” by virtue of
which derivative truths hold (ibid., 118). The effort is essentially explanatory
(“[t]hey [metaphysical semantics] . . . must be explanatory . . .” ibid., 118).
A proper statement of TM should include only fundamental truths that
incorporate “only fundamental terms” (ibid., 116) dressed in the structural
garb of set-membership, predicates of fundamental physics, classical first-
order quantification, and structure itself (ibid., 292). This is the doctrine
of purity applied. Add to it the further claim that non-fundamental/deriva-
tive truths containing “expressions that do not carve at the joints” have
“a metaphysical semantics” (i.e., add the doctrine of completeness), and
one can clearly see how Sider is led to the view that one should attempt
to achieve the aforementioned reductions by providing truth-conditions, or
proof-conditions, or some other type of metaphysical semantics for non-
fundamental truths (ibid., 116; Sider uses truth-conditions in ibid., 123
inter alia loca).29 The ontology of TM is given by its axioms or laws. These
laws commit Sider to sets and space-time points as the lone fundamentally
existing entities. Talk of anything else, whether adulterated with sets and
space-time points, or completely impure, holds by virtue of the unadulter-
ated fundamental truths about sets and space-time points (I’m aware of the
fact that Sider (ibid., 292) seems to want to get away with not invoking ‘set’
in the ideology of TM. However, I’m unsure of whether it is friendly to
include that as part of his overall framework because while ‘set’ is not in the
relevant ideology, ‘set-membership’ or ‘∈’ is.).
SHM is problematic for at least two reasons. First, suppose there could
be a conjunction of all derivative truths without logical redundancies, (d1,
d2, d3, d4, . . . dn) = dC. By completeness and purity, dC has a metaphysical
semantics and so a metaphysical explanation (call it E). Note also that a
conjunct of dC will be the conjunction of some derivative truth d545 and E
itself. Because explanation is conjunction distributive, the explanation of dC
that is E will be an explanation of (d545 & E). But that E explains (d545 &
E) and that explanation is conjunction distributive entails that E explains E
reductio ad absurdum.
Consider now the second problem. It follows from SHM that,
(1) It is not the case that there are(fundamentally) truths.30
There are no fundamental axioms of Sider’s book of the world that report
on the existence of truths. This is because Sider’s metaphysical theory of the
world, and so also SHM, “includes no linguistic ideology, no notions of predi-
cate, sentence, conjunct, satisfies, true, means . . .” (ibid., 295). Sider believes
that these notions are not structural because they are not indispensable to the
ideologies of our best theories. Are they required by the ontologies of those
theories? Sider does not believe so, because for him, one can get by with
merely sets and space-time points. But again, a metaphysical semantics is sup-
posed “to show how what we say fits into fundamental reality” (ibid., 112;
emphasis mine). For example, on the truth-conditional way of implementing
a metaphysical semantics, one explains (1) by appeal to some fundamental
truth (call the metaphysical truth-condition specification in view (FT)) that
gives the truth-conditions for it in perfectly joint-carving terms. However, FT
does not exist(fundamentally). That is a consequence of (1). How then is it that we
are connecting, (1) itself, by reduction31 to fundamental reality? Answer: FT
gives the metaphysical truth-conditions for it in purely joint-carving terms.
But the bijunct (call it φ) in FT, that is in purely joint-carving terms, does not
enjoyfundamentally existence. The proponent of SHM might say that φ is about and
corresponds to the fundamental, and that is how we achieve the connection
between “what we say” and “fundamental reality.” However, on the SHM
the correspondence/aboutness relation does not fundamentally exist. No cor-
respondence or aboutness relation answers to indispensable notions of our
best theories (i.e., logic, mathematics, and physics), nor is such a relation part
of the ontologies of logic, mathematics, or physics (all, at least, by Sider’s
lights). Perhaps the way out is to maintain that structure is structural, and so
a notion’s being structural just amounts to its carving at the joints. It thereby
corresponds to the fundamental. But Sider’s conception of structural notions
does not fit well with this response. Structure attaches itself to notions like
the existential quantifier of first-order logic, or notions peculiar to our best
theories. It is not the correspondence/aboutness relation commonly employed
by proponents of the correspondence theory of truth, or, in the case of
aboutness, truthmaker theorists. The only sense in which we can say that
φ hooks up with the fundamental is by correctly asserting that its notions
are structural (metaphysical semantics is compositional (ibid., 118)). But
that tells us nothing about the relationship between the derivative entity that
bears(derivatively) truth and fundamental reality. Truth-bearers are not the mereo-
logical sums of their notions (and Sider is a mereological nihilist).
Consider now the following arbitrarily chosen derivative truth and its
metaphysical truth-condition specification,
(2) Statement (1) holds iff φ. [I’m assuming that φ contains only struc-
tural notions, and that the biconditionals involved are/or are at least
entailed by truth-condition specifications of the kind that are given by a
metaphysical semantics.]
is true only if,
(3) (φ ↘ FR) [where ‘↘’ is the corresponds to relation and where ‘FR’ stands
for structure]
That is to say, a (non-trivial) necessary condition for the truth of (2) is that
φ corresponds to fundamental structure. However, (3) is not composed of
purely structural notions since it includes the notion of correspondence or
‘↘’. Thus, (3) will require a metaphysical semantics like the following,
(4) (φ  FR ) iff γ .
But a non-trivial necessary condition for the truth of (4) is,
(5) ( γ  FR )
That is to say, γ corresponds to fundamental reality or structure. It follows

from the above that (5) is a non-trivial necessary condition for (2).32 But (5)
will need a metaphysical semantics as well. We therefore have an infinite
regress on our hands. That regress will suggest that there are infinitely many
substantive necessary conditions for the truth of any one metaphysical truth-
condition specification (q.v., n. 33). Not just that, but given our assumption
that metaphysical truth-condition specifications at least afford bicondition-
als between derivative truths and fundamental truths (i.e., truths with only
structural notions composing them), a necessary condition for the truth of φ
is any and every necessary condition for the truth of (1). However, given that
(2) is a true metaphysical truth-condition specification (i.e., at least the rel-
evant biconditional holds), it follows that (2) itself is a substantial necessary
condition for (1).33 But again, (2) has infinitely many necessary conditions,
and therefore so does (1). By the same reasoning, φ will inherit infinitely
many necessary conditions given that (2) really does hold. These necessary
conditions will include infinitely many metaphysical truth-condition specifi-
cations, or infinitely many reductions, or infinitely many explanations.
A consequence of Sider’s metametaphysic is not just that there is a
metaphysical self-explainer, but also that any and every truth couched in
non-purely joint-carving terms has infinitely many non-trivial necessary
conditions involving infinitely many metaphysical truth-condition specifi-
cations. The success of any one metaphysical explanation via metaphysi-
cal semantics specification rests upon infinitely many other metaphysical
explanations. That result seems quite contrary to the spirit of the reduction-
ist’s explanatory program. Given such consequences, we should ask if Sider
is ever really connecting the derivative with the fundamental. Is he really
explaining the derivative in terms of the fundamental?
My argumentation so far should lead one to believe that the project of
giving a metaphysical semantics for truths like (1) is in trouble. On SHM, if
there is no way to lower a drawbridge from the fundamental to (1) with-
out incurring the cost of infinite regresses, or infinitely many substantive
necessary conditions, then we should abandon identifying them as deriva-
tive. To ensure that they do in fact describe the way the world fundamen-
tally is (without the aforementioned costs), and to ensure that certain of
our descriptions of the world’s fundamental structure actually attach to the
fundamental furniture of the world (without the aforementioned costs), we
should regard them (truths like (1)) as fundamental* or derivative* truths
(see n. 34 for what’s meant here),34 or just truths tout court. However, on
this modified-SHM, statement (1) (without the Siderian qualification) will
ensure a violation of MP1. That is to say, if Sider is forced into modified-
SHM, he will have actual mental state defeaters for his belief in modified-SHM.
It cannot be epistemically rational for one to believe a metaphysicalC system
S that entails that there are no truths (given that one also believes S asserts
or entails there are no truths and that one believes S).
Section 1.2: Epistemic Structural Realism and Metaphysics of Science

Scientific structural realisms of various kinds have been the subject of much
discussion in the contemporary philosophy of physics and metaphysics of
science. Any one of them can be viewed as a position that is part of one’s
metaphysical methodology, metaphysical worldview, and metaphysicalC
system insofar as the adopted realism suggests ways for them to be informed
by science and vice versa. The version of structural realism I focus on here
is called epistemic structural realism (ESR). It says that the strongest sense
in which one can be rational with respect to one’s belief that some empiri-
cally successful scientific theory Ts is verisimilitudinous must be qualified in
such a way that the object of rational belief is strictly that Ts is structurally
approximately true.35 Proponents of ESR have in mind science in general
(see Worrall’s summary statement in (Worrall 1989, 99) and his use of the
phrase “and elsewhere”; cf. the reading of Zahar in (Ladyman 1998, 410),
plus the remarks in (Stein 1989, 57) and (Ladyman 2016, sect. 3)). ESR adds
that there are indeed successful scientific theories like quantum mechanics
for which it is rational to believe that they are structurally approximately
true. If Ts is empirically successful, and it accurately describes or expresses
via its underlying formalism structural relations that hold between objects,
then justification is provided for belief in those relations.36 For example,
by John Worrall’s lights, the structure of the world as described by classi-
cal electrodynamics includes dependence relations that hold between, e.g.,
“optical effects” and “that which oscillates at right angles to the direction
of [the] transmission of . . . light. . . .”37 It is not rational to believe in the
existence of electromagnetic radiation as posited by James Clerk Maxwell’s
(1831–1879) (fully developed) classical theory of that radiation, nor is it
reasonable to believe in the existence of some luminiferous aether behind
the effect (as in Augustin-Jean Fresnel’s (1788–1827) theory). What it is rea-
sonable to believe in is the existence of structure, and the observables (i.e.,
observable entities) posited by Ts.
What is structure according to Worrall’s ESR? Beyond actually existing
dependence relations, Worrall’s characterization of structure is fragmentary
at best.38 From what I can gather from his corpus, structure is likened to
form,39 it is (supposedly) preserved or at least approximated amidst theory
change (e.g., in the shift from Fresnel’s theory of light to Maxwell’s), (again)
it is described as a type of dependence40 that is tracked by a mathemati-
cal core, a set of equations that are themselves retained precisely when the
structure/dependence it tracks is approximated amidst preservation through
theory change.
Worrall considers it a consequence of ESR that “a theory’s full cogni-
tive content is captured by its Ramsey sentence,”41 whereas one acquires
Ts’s Ramsey sentence by performing existential generalization with second-
order variables on every theoretical predicate in a statement of Ts. What
motivates such a view is the epistemological thesis that “we know about”
theoretical stuff “only by description—that is, via” that stuff’s “role in our
theories.”42 I will soon return to this point in my criticism.
Consider the following implication of ESR,
(1) If ESR is true, then (a): the only theoretical stuff we have empirical epis-
temic access to are dependence structures and relations, (b): the empirical
epistemic access we have to theoretical stuff is solely through scientific
inquiry, and (c): everything else we can know through empirically suc-
cessful scientific inquiry pertains to the observational. [Premise]
According to ESR (quoting Worrall), “it is a mistake to think that we can

ever ‘understand’ the nature of the basic furniture of the universe.”43 What
content about unobservables we can know through empirically successful
science must be delimited to (quoting Worrall) “relationships between phe-
nomena expressed in the mathematical equations . . . the theoretical terms
of which should be understood as genuine primitives.”44
(2) (d): Brandon’s decision to ask a question about Newton’s theory of

gravitation in physics class caused Brandon’s hand to raise, and (e): the
content of (2d) reports on an instance of mental causation (call it Me)
that is a theoretical relation. [Premise]
Premise (2d) describes an actual instance (we are supposing) of mental cau-
sation. Brandon decided (a mental event) to raise his hand and brought
about a physical event (mental to physical causation).
On the supposition that one is in the classroom with Brandon, one does
not observe Brandon’s decision, although one does observe the hand raising
event.45 We do not have privileged access to the mental goings-on of Bran-
don’s mind, and so we cannot perceptually behold his decision. Also, we are
in no position to become directly acquainted with it in any way. And even
though some do maintain that we can perceive obtaining causal relations,
these scholars do not argue that all instances of causation are open to our
perceptually beholding them. For example, David M. Armstrong argued
that we perceive causation in cases involving pressure and forces impressed
upon the body (Armstrong 1997, 214), or when we exert our own wills
(Armstrong 1988, 225). Nancy Cartwright contended only that causation is
observable in certain forms of local work, explicitly acknowledging cases of
causation that we cannot observe (Cartwright 1993, 426–427). Like Arm-
strong, Stephen Mumford and Rani Lill Anjum suggest that some instances
of causation involving inward bodily sensation can and are directly veridi-
cally perceived (Mumford and Anjum 2011, 201–209, although they seem
to not regard causation as a relation).46 None of the appropriate literature
maintains that we can and do perceive instances of causation such as Me.
(3) If (2), then Brandon’s decision is best categorized as theoretical stuff

(e.g., of psychology) serving as a relatum in an obtaining theoretical
dependence relation. [Premise]
(4) If (1a, 1b, 1c), and Brandon’s decision is best categorized as theoretical
stuff serving as a relatum in an obtaining theoretical dependence relation,
then we cannot acquire knowledge of Brandon’s decision. [Premise]
(5) Therefore, if ESR is true, then we cannot acquire knowledge of Bran-
don’s decision. [Conclusion]
ESR precludes knowledge of (at least) instances of mental causation when

those instances are the result of the mental activity of agents other than
one’s self (the case of Brandon’s decision was somewhat arbitrary). In other
words, ESR lands us in the problem of other minds.47 The deliverances of
psychology cannot include plausible abductive inferences to the existence
of the mental states of cognizers given ESR. If it is appropriate for the psy-
chologist to use the typical analogical and abductive arguments to put us in
epistemic touch with Brandon’s mental decision, then why is that style of
argumentation systematically unsuccessful in the context of scientific theo-
rizing more generally? The proponent of ESR therefore faces a dilemma.
Either (horn 1) ESR holds for all our best scientific theories or (horn 2)
ESR holds only for theories of physics. If the proponent of ESR chooses
(horn 1), they will violate D4, keep themselves from the knowledge of ESR
itself due to MP2, and in light of argument (1)–(5), cut out from under
themselves knowledge of “otherly”-mental causation. If one chooses (horn
2), then one faces the difficult problem of how to allow for knowledge-
conducive inferences to unobservables in the domain of the special sciences
like psychology, but not in the domain of the physical sciences. The prob-
lem is particularly tough because empirical success in experimentation and
the like is more abundant and at times embarrassingly strong in the case of
our best physical theories. Why then are our inferences to unobservables
more secure in the special sciences where our evidence is often thinner than
in the physical sciences? The dilemma is so serious that I believe it warrants
an abandonment of ESR, and our directive D4 and methodological prin-
ciple MP2 have played an indispensable role in underwriting that aban-
donment (they each sharpened horn 1). Proper metaphysical methodology
has led us to jettisoning ESR’s take on how science informs metaphysical
commitment.
Section 2: Propositions and Truth-Conditions: Truth
and Meaning Part II
Every proposed metaphysicalC system will be dressed in the garb of a lan-
guage. D1 and its underlying methodological principle suggest that every
metaphysicalC system will need to walk together with a theory of meaning
and truth. Scott Soames has argued persuasively that “propositions are needed
to state the goals of semantic and pragmatic theories of any language” and
that “natural languages like English” require them “as referents of that-
clauses, arguments of attitude verbs, referents of some names and uses of
indexicals, members of the domains of some quantifiers, and so on.”48 The
metaphysicalC system I am currently developing is being proposed with both
English and formal languages like mathematics and logic. Thus, notwith-
standing those who make no room for propositions in their theory of lan-
guage,49 I will follow Bernard Bolzano (particularly his conception of truths
and propositions in themselves)50, Gottlob Frege (particularly his notion of
a thought)51, Bertrand Russell (specifically his view of propositions),52 and
the many scholars writing in their wake by assuming that there are proposi-
tions and that they are the fundamental bearers of truth-values, those enti-
ties that stand in logical relations, and those things that are the contents of
beliefs, assertions, and denials. I will also assume that declarative sentences
of various languages (even the language of mathematics) express proposi-
tions and that propositions are about things and therefore have representa-
tional properties. These properties metaphysically explain why propositions
have the truth-conditions they do.
One acquires both a reason to believe in propositions and a reason
to restrict their natures to something wholly beyond that which is con-
tingent upon our cognitive activity by giving attention to meaningful
equations of mathematics. What are equations? They are entities that are
expressed, or intended to be expressed, by various representation tokens
of the language of mathematics (and other languages) at various times.
These tokens designate or attempt to designate entities representative of
purported equality relationships between the two members of the equa-
tion (e.g., those entities intended to be represented by the left and right
sides of the equality sign in the language of mathematics or its appropri-
ate substitute in a natural language). Equations should not be identified
with our concepts, linguistic expressions, or propositional attitudes. My
reasons for giving up on such identity theses lie in the nature of certain
mathematical equations themselves. For example, the study of algebra
will introduce one to polynomial equations, that is, equations with the
following form, anxn + an-1xn-1 + . . . + a2x2 + a1x + a0 = 0 (as in Gowers
2008, 49). Solutions to polynomials are themselves given by equations or
equality relationships between variables and mathematical values. The
simple polynomial,53
(Eq. 1): x 2 = 2
has solutions ordinarily expressed as
(Sol. 1&2) x = ± 2 (Let (Sol. 1) include the plus/positive value.)
or x equals plus or minus the square root of 2. In this case, a certain positive
number is the principal root (although such roots are generally just non-
negative) of 2 (call it σ), and I will limit my discussion to it (i.e., when one
sees 2 please understand me to be referring to the positive square root of
2 as is the ordinary practice in mathematics with respect to this particular
case (Clapham 1996, 263)).
The number or value that is σ serves as content of (Sol. 1). But can we be sure
there is such a value? Yes. The existence of σ is an implication of results related
to the intermediate value theorem of mathematics (indicating that the solution’s
value lies somewhere between the real rational numbers 1 and 2).54 That a
positive square root of 2 exists is also a consequence of the necessary truth that
there exists at least one positive square root of any positive real number (ibid.).
Although we can prove that a σ exists and that it has a value between 1
and 2, the principal square root of 2 is an irrational number. It cannot be
expressed as a ratio of two integers. Its decimal representation possesses
neither a terminating expansion nor a recurring or finite expansion (the
digits in the decimal representation do not build up a repeating pattern).
Like the decimal value that equals ¹⁄³, any intended approximate represen-
tation of σ with decimal notation can only be stated in truncated form.
Like . 33, the decimal values of the principal square root of 2 is such that
we can, in principle, extrapolate out from known mathematical facts what
value any one arbitrary decimal place of the intended representation of σ
in decimal notation will take. However, its precise value and decimal rep-
resentation are not known, nor do we appear to be able to form a definite
concept of painstaking extrapolations of that decimal representation that
extend beyond 10 million digits (Nemiroff and Bonnell 2017).
One might reply that irrational numbers are themselves precise math-
ematical values in that the principal square root of 2 that is σ just is 2, and
nothing more. The locution ‘ 2’ is not shorthand at all, but in fact individu-
ates a precise mathematical value. But this response seems wrong. 2 is
indeed shorthand, indicative, at least in part, of a mathematical operation.55
Again, it represents a positive number or value σ, which when you square
it, the value you retrieve is 2. To assert all over again that σ just is the posi-
tive 2 is wrong-headed at least because it is uninformative. What is more,
some irrational numbers like π are not expressible in terms of non-nested
radicals. And to assert that the value of π just is π is radically uninformative.
We seem to be unable to form any determinate concept of σ. However, there
is such a value, and that value is part of the content of propositions like Eq. 2.
Thus, Eq. 2 must be something that outstrips our concepts because it possesses
representational features that lie beyond what we have concepts of.
(Eq. 2): σ = 2 (where, again, I have in mind the principal root of 2)

The entity that is <Eq.2> (using brackets around Eq. 2 to pick out the prop-
osition that is the equation) must also lurk beyond distinctive strings of
graphemes or phonemes such as written or uttered sentences of the public
language of mathematics. Neither an utterance of Eq. 2 (the sentence, not
the proposition), nor the appropriate symbols on a page give us the value
that is related to the outcome of the (at least in part) operation represented
by 2.56
Here is my argument thus far,
(1) If (a): all propositions have truth-conditions (see sect. 4.1) and are rep-
resentational, (b): for any proposition p, p has the truth-conditions it
does in virtue of the representational properties it possesses or has or
exemplifies, and (c): representational property instances connecting the
representational properties of propositions with the appropriate/rel-
evant propositions themselves are grounded57 in our cognitive and/or
linguistic activity, then <Eq.2> has the truth-conditions it does by virtue
of our cognitive and/or linguistic activity. [Premise]
(2) (d): It is not the case that <Eq.2> has the truth-conditions it does by
virtue of our cognitive and/or linguistic activity, and (e): both (1a) and
(1b) hold. [Premise]
(3) Therefore, it is not the case that (1c). [Conclusion]
The “in virtue of” talk in the above argument is meant to capture the
idea of metaphysical explanation through grounding. As I will make clear
in sect. 4.5.3, my assumed theory of grounding posits that grounding is
an asymmetric, transitive dependence (for positive ontological status and
nature) relation, whose relata can be stuff belonging to any ontological cat-
egory whatsoever (as in Schaffer 2009). An instance of (1b) would be that
<Eq.2> is such that it has the truth-conditions it does because (metaphysical
explanation) of the representational properties it bears. Or, <Eq.2>’s having
the truth-conditions it does is grounded by <Eq.2>’s having/exemplifying its
representational properties.
Statement (1a), if true, ensures that propositions have representational prop-
erties and truth-conditions. (1b), if true, ensures that their having (instances
of exemplification are what I’m concerned with) the truth-conditions they do
depends (in the grounding sense) upon property exemplifications involving
propositions and appropriate representational properties to which they are
tied. Statement (1c) says that instances of exemplification connecting repre-
sentational properties with propositions depend (in the grounding sense) upon
our cognitive and/or linguistic activity. The involved grounding-dependence
relations are transitive, thus, given (1a) through (1c), propositions (in gen-
eral) being such that they have/possess/exemplify the truth-conditions they do
depend upon (are grounded in) our cognitive and/or linguistic activity. And
so, given (1a) through (1c), <Eq.2> (in particular) has the truth-conditions it
does by virtue of our cognitive and/or linguistic activity.
Conjunct (2e) follows from the assumptions of this subsection. To amass
support for (2d), we note first that σ lies beyond our knowledge and cogni-
tive grasp. We can form no definite concept of it and all its features. It seems
difficult then to maintain that the representational property exemplifications
or instances involving <Eq.2> (I shall speak now of just the representational
properties and truth-conditions and not the relevant instances for brevity)
that have directly to do with constituent σ are grounded in our cognitive/
linguistic activity. At best we could only have created mental or linguistic
representation tokens that are mental or linguistic singular or general terms
that refer to or partially describe it. Nothing we do contributes all the rep-
resentational properties of <Eq.2> that are associated with that value. But
if nothing we do grounds all of <Eq.2>’s representational properties, then
nothing we do indirectly grounds <Eq.2>’s truth-conditions because <Eq.2>’s
representational properties, in toto, ground <Eq.2>’s truth-conditions.
One might counter that the relevant representational properties of <Eq.2>
are grounded in our successful mental or linguistic acts of referring to σ.
We use mental or linguistic representation tokens like the term ‘σ’ to suc-
cessfully refer to the entity that is σ. Why can’t we ground the σ-related
representational features of <Eq.2> in the success of the relevant referential
acts? Insofar as <Eq.2> is really about σ, it will have fine-grained σ-related
representational properties that we are completely insensitive to when we
successfully refer to σ or when we think or utter Eq.2. This is because many
of the relevant σ-related representational properties are connected to the
complexities of σ that we are unable to grasp and that we have no determi-
nate concept of. How does mere success of reference afford a metaphysical
explanation of even those representational features of <Eq.2>?
One might adopt a modified neo-RussellianP position regarding the nature
of propositions by insisting that singular terms like ‘σ’ contribute the entity
that is σ to <Eq.2>. The name directly refers to that entity. Indeed, that name
has that entity as its semantic value.58 All of this is what makes <Eq.2> a
singular proposition59 and what helps ground the σ-related representational
properties of <Eq.2>. <Eq.2> comes to have σ-related representational prop-
erties by virtue of having σ as part of its structure. But notice that on this
view, nothing we do grounds the σ-related representational properties of
<Eq.2>. It’s the structure of the proposition itself that grounds those prop-
erties. But this particular neo-RussellianP might counter that metaphysical
explanation is transitive such that our activity of referring in our mental
or linguistic use of Eq. 2 grounds, by contribution, <Eq.2>’s structure and
thereby indirectly grounds <Eq.2>’s σ-related representational properties.
But how could what we do ground, even indirectly, the structure of Rus-
sellian propositions? On one view of their precise nature, such entities are
ordered n-tuples.60 These types of sets (or concatenations thereof) do not
acquire their order from anything we do; they just are collections of n-items
with that order or conjoined order. Our acts do connect mental and linguis-
tic entities to such collections, but that is because with such acts, we express
those collections. The collections are not built out of our activity. Our activ-
ity is merely related to them via proposition expression. In addition, sets
are a se and abstract. We cannot influence them. Thus, if this particular
neo-RussellianP position is correct, our referential activity does not provide
a proper ground for <Eq.2>’s representational features.
<Eq.2> should not be identified with an event-type whose existence
depends upon our cognitive activity either (contra the view of propositions
in Soames (Meaning 2010; Cognitive Propositions 2014)). According to
Soames, “the perceptual and cognitive activity of agents is the conceptual
basis of all representation and . . . propositions are representational in virtue
of the relations they bear to this representational activity.”61 The event-type
Et of an agent’s predicating (cognitively) to George’s car that it is white
grounds the representational features of the proposition that <George’s
car is white.>. Event-types like Et both depend (at least in part) for their
existence, and derive (at least in part) their representational properties,
from actual tokened instances of agents representing George’s car as white
through appropriate cognitive acts of predication (ibid., 97 and Soames,
Meaning 2010, 104).62 This purported explanation of <Eq.2>’s represen-
tational features cannot succeed because it rests upon the identification of
propositions with event-types, and that is independently suspect.
Soames dismisses the view that propositions are act-types. That they fail
to be propositions is well-supported by the further fact that “ascriptions to
propositions of what can be said of act types, as well as ascriptions to act
types of what can be said of propositions, strike us as bizarre, or incoher-
ent.”63 A similar argument will undermine the thesis that propositions are
event-types. If there are instances of general causation, then event-types
occur or happen. I will assume that there are instances of general causa-
tion. If Soames does not wish to challenge that assumption, the follow-
ing should be troublesome for his approach. It is perfectly coherent to
say that event-types like beheadings happen. However, the assertion that
“The proposition that two plus two equals four happened (or happens),”
“strikes us as bizarre” (borrowing Soames’s wording). I can truthfully say
about or ascribe to certain event-types that they cause (in the sense of
general causation) certain other event-types, but I cannot, without strange
looks, ascribe to a proposition that it caused/causes something in the gen-
eral (or singular/token) sense (see my discussion of the factsP view of causal
relata in chapter 7: n. 113). Some propositions are true. That ascription
is above reproach, though it strikes us as bizarre to say of the event-type
picked out by the term ‘beheadings’ that it is true. Propositions are not
event-types. But Soames’s explanation of the representational properties
of propositions requires the posit that propositions are event-types, so
Soames’s explanation fails.
A more exhaustive case for (1a) and (2d) would require a more com-
plete critical survey of various naturalistic theories of propositions such as
those in King (2007; Naturalized 2014) and Speaks (2014). Speaks denies
that propositions are representational (i.e., he would deny (1a)), and King
attempts to explain the representational properties of propositions by
appeal to our activity. I believe the case for (1a) is quite strong and theo-
ries like King’s will be unable to properly ground the σ-related representa-
tional properties of <Eq.2>, but I leave an explication of the criticisms for
another time.
My conclusion is that the representational properties of propositions
generally (i.e., propositions exemplifying the representational properties
they do (the property-instances)), but <Eq.2> specifically, are not grounded
in what we mentally or linguistically do. And because I am assuming that
propositions are fundamental truth-bearers, (1b) holds, and as I will suggest
later, every proposition, including <Eq.2>, is either true or false, in the spirit
of Gottlob Frege, we should therefore also maintain that propositions are
true or false independent of what we say, do, or think.64
We can now add one further directive to our list,
(Directive #6 (D6)): One’s metaphysical worldview (and one’s metaphysicalC

system) ought not preclude or be agnostic about the existence of mind-
independent truths or falsehoods that are true or false propositions.
Section 2.1: Against the New Verificationism

Don Ross, James Ladyman, and David Spurrett65 in Ross, Ladyman, and
Spurrett (Causation 2007, 258–297) (henceforth RLS) are led to a specific
view of causation on the basis of a distinctive philosophical methodology
that violates D6. More specifically, RLS are verificationists in the pragmatist
(not the positivist) tradition. They affirm that,
(Verificationist Principle (VP)): (ai) Theories in metaphysics should have

some “identifiable bearing on the relationship between at least two”
scientific hypotheses deemed confirmed by the scientific community, or
(aii) else they should be “motivated and in principle confirmable by
such science”, and that (b) any hypothesis that science deems as beyond
our ken “to investigate should [not] be taken seriously.”66
Beside this Verificationist Principle, RLS add the Principle of Naturalistic

Closure,
(Principle of Naturalistic Closure (PNC)): One should only take seri-

ously those metaphysical theses that are solely motivated by their
explanatory contribution to showing how a hypothesis belonging to a
fundamental physical theory and at least one other hypothesis belong-
ing to a distinct scientific theory explain more than “the sum of what
is explained by the two hypotheses taken separately” (and without the
relevant metaphysical thesis).67
All metaphysical hypotheses and arguments that fail to abide by PNC should
be dismissed as hopeless speculation.
Earlier, I adumbrated a metaphysical theory of propositions according
to which propositions are the fundamental bearers of truth-values, those
entities that stand in logical relations, those objects that are expressed by
declarative sentences, those things that are the contents of beliefs, asser-
tions, and denials, and those entities that are true or false independent of the
cognitive and linguistic activity of us cognizers. Call this theory T-P.
Does one abide by the PNC if one believes T-P? No, one does not.
A metaphysical theory of propositions like T-P is not solely motivated by
what explanatory contribution it provides to the endeavor of connecting
fundamental physical hypotheses to other scientific hypotheses. Proposi-
tion theorists are interested in accounting for linguistic data, and pragmatic
and semantic facts. They are interested in answering questions like, “By
virtue of what is it the case that two sentences belonging to two different
natural languages mean the same thing?” These theorists want to know
about and account for the nature of meaning, truth, falsity, representation,
denial, assertion, belief content, reference, valid arguments, and other sub-
jects studied in logic, linguistics, and the philosophy of language.68 They
also want to provide metaphysical theories of the nature of propositions
themselves (King 2007, 4–5).
By the lights of the metaphysical methodology of RLS, we should not take
seriously theories of propositions like T-P. If, however, we do not have T-P
or anything like it, then neither VP nor PNC can come out true (continuing
to assume that propositions are fundamental truth-bearers). This is because
propositions whose natures are specified by T-P or theories like it are needed
to serve as both those entities that the sentences communicating VP or PNC
express (possibly thereby also acquiring their meanings) and those entities
(when true) that help metaphysically explain why sentences communicating
VP and PNC are true. Proponents of the new verificationism seem trapped
in a self-undermining position that disallows from the content of any meta-
physical worldview built upon its principles, metaphysical theories of prop-
ositions like T-P. They thereby cut out from under themselves that which is
needed to secure the truth of VP and PNC.
RLS may counter that even if their approach to metaphysics depends
upon “metaphysical assumptions” like T-P, “the metaphysical assumptions
in question are vindicated by the success of science.”69 But the PNC does
not make room for metaphysical theories that are indirectly justified by the
success of science and are not solely motivated by their explanatory contri-
bution to showing how a fundamental physical hypothesis and at least one
other scientific hypothesis explain more than “the sum of what is explained
by the two hypotheses taken separately” (ibid., 37). It is the lack of that sole
motivation that serves as justification for dismissing T-P and theories like it.
That dismissal lands the new verificationist in trouble, even given indirect
support for T-P from empirical success.
Section 3: The Default Setting
In sect. 2, I asserted that an assumption of the present work on causation
will be that propositions are fundamental truth-bearers. I argued that the
sense in which such entities are true is (qualifiedly) independent of our men-
tal states or mental activity. This non-epistemic approach to truth fits com-
fortably with a view of the external world of objects, properties, relations,
processes, and structures that affirms that those objects, properties, rela-
tions, and the like enjoy mind-independent existence.70 These two theses
together constitute a picture of reality called Realism.71 That picture has an
illustrious philosophical history. As Michael Loux wrote,
Virtually every major thinker in the ancient, medieval, and early modern
periods endorsed the themes making up the picture. Indeed, the picture
provided something like a framework within which traditional philo-
sophical inquiry took place; and it was so much a part of the assumed
backdrop for doing philosophy that it did not occur to philosophers to
give the picture a name.
(Loux 2001, 449)
Realism is incompatible with Hilary Putnam’s doctrine of conceptual rela-

tivity. According to that doctrine, “truth genuinely depends on the behavior
of things distant from the speaker.”72 However, “the nature of the depen-
dence changes as the kinds of language games we invent change.”73 There is
no “single metaphysically privileged description” of the world.74 Thus, for
Putnam facts about “what exists may depend on which of various conven-
tions we adopt”.75 Of course, the involved linguistic conventions afford,
descriptions which are cognitively equivalent . . . in the sense that any

phenomenon whose explanation can be given in one of the optional
languages involved has a corresponding explanation in the other . . . but
which are incompatible if taken at face value (the descriptions cannot
be simply conjoined).
(Putnam 2004, 48)76
But explanation is factive such that for any proposition q and any other propo-
sition p, necessarily if q explains p, then p is true. However, on not a few theories
of explanation, the explanans must hold as well.77 If that is the case and expla-
nation is factive, then there can be no conceptual relativity because there can-
not be true yet incompatible explanations of any one individual phenomenon.
There are other anti-realisms or objections to realism to worry about (see
e.g., Dummett 1978, 145–165), but I will seek shelter from them behind
critiques/replies in van Inwagen (2001) and Maudlin (2015).78 And so the
current work on causation assumes Realism.
Is hiding behind replies to objections to Realism the best that one can do?
Are there any good arguments for Realism? None that I believe dissenters
of Realism will find convincing. However, it seems to me to be good enough
to proffer a defense that consists of plausible replies to anti-realist argumen-
tation (compare the approach in Alston 2001, 32). I therefore suggest the
following directive,
(Directive #7 (D7)): One’s metaphysical worldview (and one’s meta-

physicalC system) should endorse as expansive a Realism about the stuff
of the world as is rationally permissible.
Section 4: Metaphysical Theorizing
Section 4.1: Quantifiers and Logic Choice

There are meanings. There are truths. There are mental states and mental
events. True statements like these involve existential quantification.79 When
I am intending to sincerely assert metaphysical and other theses about the
world that use quantifiers, I will appropriate a domain conditions (as in
Hofweber’s “external reading” 2016, 60, cf. 81–84) and objectual inter-
pretation of quantifiers. Thus, the statement, ‘some entity is F,’ holds, just
in case, an entity in the universe of discourse is such that it has F.80 As a
consequence, I reject, or at least I will not use, the substitutional interpreta-
tion of quantifiers according to which (for the case of quantifiers in English)
“some entity is G” is true if, and only if, there’s some term t belonging to
the English language that is such that “Gt” holds with respect to it (ibid.).
I doubt that interpretation captures part of the meanings of most quantifiers
used in ordinary English anyway (following Kripke 1976, 380, inter alios).
In English, a single quantifier can be polysemous. Here’s what I mean.
Leaning some on Sider’s notion of candidate meanings, the diverse meanings
of but one quantifier Q in English are assigned meanings. Let one assigned
meaning be R1, and let the other assigned meaning be R2. R1 and R2 fix dif-
fering truth conditions for some statements in which Q is embedded. These
meanings do not yield ungrammatical readings of those statements, and
I will require that every assigned meaning be both materially and inferen-
tially adequate (following Sider 2009, 391–392).81 I will return to the poly-
semous nature of quantifiers shortly.
Despite the existence of varying assigned meanings of quantifiers within
one natural language such as English, I will work with and assume the assigned
meanings for quantifiers given by the domains condition and objectual
interpretation of them in classical first-order logic connecting that assigned
meaning to quantifiers in higher-order logic, plural first-order logic (PLO or
PFL), and S5 quantified modal logic (S5), all supplemented with Zermelo-
Fraenkel set theory that includes the axiom of choice (ZFC). By consequence,
the inferences I can and do make with statements containing quantifiers are
(or at least should be) governed by the axioms and theorems of classical
first-order logic (CFOL) (and when appropriate set theory, higher-order logic,
PLO, and S5, all associated with or built atop CFOL). Drawing from some of
my previous work (Weaver forthcoming), I note here that CFOL is a formal
first-order language constructed on the back of classical propositional logic
(CPL), set up with the classical quantifier rules of inference82 and a certain
formal semantics.83 In addition, CFOL obeys the following four principles:
(Principle #1): Every well-formed formula’s (wff’s) truth-value on some

interpretation ℐ (where interpretations include non-empty universes
of discourse) is completely fixed via the extension of the parts of that
wff under ℐ.
(Principle #2): There are only two truth-values, truth and falsehood.
(Principle #3): Every wff on any assumed interpretation ℐ has exactly
one truth-value.84
(Principle #4): Every wff on any one interpretation ℐ is either true or
false.
(Weaver forthcoming)
Continuing to follow (Weaver forthcoming), I note how classical logic has

very elegant formal features. The propositional version is Post-complete (Church
1996, 109–110). And for CFOL, the theorem of Herbrand (1971) applies. Thus,
if one were to suppose that “there’s a valid wff of CFOL p and a suitably cor-
relative valid formula q of CPL,” “Herbrand’s theorem entails that p’s validity is
nothing over and above q’s validity (i.e., in some qualified sense, CFOL reduces
to CPL)” (quoting myself in Weaver forthcoming). “There are no applicable
and general analogs of this theorem in non-classical logics such as intuitionism
(Rumfitt 2015, 14). Moreover, with regard to CPL, there are ways of justifying
the inference rules of that logic from even a non-classical metalanguage” (quot-
ing myself once more).85 Examples could be multiplied.
The elegant formal features of CFOL constitute one leg of support for
embracing it as the choice logic for inquiry into the nature of reality. Another
support comes from considerations having to do with CFOL’s choice notion
of logical consequence. According to that notion (call it CLC for classical
logical consequence), a deductive argument is valid, just in case, there is no
possible world in which the premises are true and the conclusion false.86
A strong case can be made for the thesis that CLC is and ought to be that
exclusive notion of logical consequence at work in the fundamental inquiry
conducted in mathematical physics, and physics. This is because no other
notion of logical consequence (not even the notion in intuitionist-based
constructive mathematics) can bear all of the relevant fruit (e.g., support
inferences needed to derive important theorems of general relativity and
quantum mechanics).87 If any discipline aims to uncover the nature of real-
ity, it is physics. And because physical theorizing has enjoyed much success
uncovering the nature of reality, metaphysicians also interested in the nature
of reality would do well to borrow CLC for their theorizing. We therefore
have the following directive,
(Directive #8 (D8)): One’s deductive reasoning in one’s metaphysical

inquiry aimed at metaphysical worldview and metaphysicalC system
construction should be classical deductive reasoning.
Section 4.2: Quantifier Variantism

I have admitted that there do exist a number of meaning assignments for
quantifiers in both a single natural language, such as English, and in alterna-
tive languages. However, that admission alone does not commit me to quan-
tifier variantism. Quantifier variantism says that “quantifier expressions”
can vary “their meaning from one language to another,”88 and that this
fact renders ontological disputes about the existence of perceptual objects
merely verbal and therefore non-substantive. Let me elaborate.
According to the leading proponent of quantifier variantism, Eli Hirsch,
when philosopher A appropriates assigned quantifier meaning R1 and says
that desks existR1, she is not actually disagreeing substantively with philoso-
pher B who appropriates a different assigned quantifier meaning R2 and
says that desks do not existR2. Philosopher A’s use of R1 suggests that she is
speaking a different language (perhaps a different form of English, call it L1)
than that of philosopher B (call that philosopher’s language L2). Thus, “all
that these philosophers can be disagreeing about is whether the language we
speak is in fact the first imagined language or the second.”89 And “[e]ach
side can charitably interpret each other side’s position in terms of a language
in which all the other side’s assertions come out true.”90
Resolving the question of whether quantifier variantism is true is impor-
tant for the study of causation. For while Hirsch applies that doctrine to
“issues about the existence and identity of perceivable objects,”91 I will
argue in chapter 4: sect. 3 that causation is an obtaining relation that is
sometimes perceived. Thus, the debate between causal realists (i.e, those
who affirm that there are obtaining causal relations in the mind-independent
world) and causal eliminativists (i.e., those who affirm that there are no
obtaining causal relations in the mind-independent world) concerning or
about instances of causation that can be and are perceived, can be justifiably
dismissed as a merely verbal dispute, if Hirsch is right. My approach seeks
to avoid that consequence by defending the idea that the debate between
causal realists and causal eliminativists (appropriately restricted) is substan-
tive through an argument for the falsity of quantifier variantism.
How should one avoid quantifier variantism? I will follow van Inwagen
(2009, 492–499) and Sider (2009, 397–402) by affirming that there exists
a privileged quantifier that is none other than the existential quantifier used
in CFOL.92 When that quantifier is appropriated by language users in true
statements, it helps ensure that those statements correspond to objective
structure. By my use of the term ‘structure,’ I do not mean to appropriate
Sider’s notion of structure because that notion is very unclear.93 For me,
structure and reality amount to the same thing.
Substantive ontological disputes arise (at least) amidst meaningful and
cooperative communication exchanges in philosophical, scientific, and other
academic contexts (the standards of precision are appropriately high), when
such exchanges are about reality, the involved interlocutors are speaking the
same privileged language, and so are appropriating the privileged quantifier
in that shared language.94 Attempts to resist the very possibility of satisfying
the relevant conditions are refuted, I believe, by the empirical evidence for
the thesis that many a philosopher and academic do in fact satisfy these con-
ditions at times and enter into substantial disputes over perceivable objects,
else there is significant semantic blindness afoot.95
Hirsch maintains that in certain ontological disputes about perceivable
objects, there is simply no reason for two interlocutors who are using differ-
ent assigned meanings for the quantifier (and therefore are speaking alterna-
tive languages) to switch their language choice.96 It is here that Hirsch departs
drastically from the deflationism of Rudolf Carnap (1891–1970). Carnap
maintained that there could exist good reasons for one’s linguistic framework
choice, and that these reasons could be based on theoretical knowledge and
the fruitfulness of the chosen framework (Carnap 1956, 205–221). I therefore
find an unlikely ally against Hirsch in Carnap. The choice assigned meaning
for the quantifier in a language used to conduct metaphysical inquiry gener-
ally (i.e., inquiry aimed at uncovering the truth about what exists, the nature
of reality, and the hierarchy of being) should be that same assigned mean-
ing for the quantifier in the language used to conduct physical inquiry (i.e.,
inquiry in fundamental physics).97 The specific justification for this language
choice with respect to the assigned meaning of the quantifier and the sphere
of inquiry about which Hirsch is a deflationist (i.e., the domain of perceivable
objects) is at least the embarrassing riches of success that type of inquiry has
enjoyed uncovering the natures of observable/perceivable objects and struc-
tures. Not even the scientific anti-realist would object to the thesis that physi-
cal inquiry sometimes yields knowledge of the truth about observable objects
and structures. We therefore have strong reasons with which to persuade
those interlocutors with whom we are engaged in metaphysical inquiry and
who appropriate a different assigned meaning for quantifier expressions than
that which is provided by our best empirical inquiry (i.e., physics).
(Directive #9 (D9)): One’s reasoning and one’s statements in metaphysi-

cal inquiry aimed at knowledge of the true metaphysical worldview
and metaphysicalC system should appropriate the existential quanti-
fier whose assigned meaning is that which is provided by CFOL, when
appropriate (q.v., n. 92 for why I use this ‘when appropriate’ qualifica-
tion) (where I have in mind the domains condition and objectual under-
standing of existential quantifier expressions).
Section 4.3: The Problem of Projection
One might maintain that there are equally good ways of describing real-
ity with the privileged quantifier (now ignoring, for brevity, the very few
instances in which PFL is required). One could use very different predicates
than those that are standardly used in academic contexts. For example, some
statements (with sentences) of scientific theories could appropriate unwieldy
predicates like ‘grue’ (from Goodman 1955, and Goodman 1972, 357–362).
Say that for any entity x, x is grue, if x is observed and verified to have a
primary green hue before the year 2050 or else x is not observed prior to
2050 and has a primary red hue98 (Goodman 1972, 359). Following the
ordinary setup of a now classic problem in the philosophy of science99 called
the new problem or riddle of induction, I note that there are large collections
of emeralds in the world that have been examined and all of them prior to
2050. These gemstones were all found to have a primary hue of green before
2050. What’s to keep one from inductively inferring that every emerald is
grue? Notice that if one licenses such an inference, the inferred proposition
will imply that every emerald that has not been observed or appropriately
verified to have a primary green hue prior to 2050 has a primary red hue.
That conclusion is in direct conflict with the thesis everyone wants to be
able to inductively infer on the basis of the same empirical data, viz., that all
emeralds have a primary green hue. How does one privilege that conclusion
and preclude the illicit inductive inference to grue emeralds?
The problem here revolves around the predicate ‘grue’, and the important
question for metaphysical methodology is, why can’t one describe the world
in terms of grueness with the privileged existential quantifier and thereby
accurately represent reality? Are we not entitled to describe the world that
way given that the involved inductive inference to grue emeralds is strong?
If we let a projection be an inference that moves from a premise of the form,
x/y entities of type T were perceived and judged/verified to be R, to the con-
clusion that x/y entities of type T are in fact R (as in Plantinga, Proper Func-
tion 1993, 129), the question before us is, which predicates are projectable?
Or, what ensures that a predicate is such that it can be projected?
Grue-like predicates are not projectable. First, the inference involved in
the attempt to project the predicate grue is not cogent such that it confers
warrant upon the belief that the relevant collection of gemstones is grue.
The reason why lies in the nature of warrantK, or that which (in sufficient
degree) separates knowledge from true belief,
(WarrantK): Necessarily, for any cognizer C, and for any of C’s beliefs b,
b is warrantedK, just in case, (a) C has no actual mental state defeaters for
b in C’s noetic structure, (b) C formed b via properly functioning cogni-
tive faculties F1–Fn, (c) C formed b in an environment e (or environment
very similar to e) designed for the utilization F1–Fn, and (d) “the mod-
ules of the design plan governing the production of [b] are (1) aimed
at truth, and (2) such that there is a high objective probability that a
belief formed in accordance with those modules (in that sort of cogni-
tive environment) is true.”100
Our beliefs are sometimes indirectly warranted. What this means is that some-
times, some of our beliefs receive warrantK on account of warrantK transfer
from a belief or set of beliefs that are directly (or indirectly) warrantedK. For
example, belief b1 might deductively entail belief b3. Both belief b1 and one’s
belief that b1 entails b3 (call that belief b2) might be directly warrantedK, and
warrantK (given that the inference is valid) transfers to b3 when one forms b3
on the basis of b1 and b2 in a way that satisfies conditions (a)–(d). Directly
warrantedK beliefs are often those that are foundational to our noetic struc-
tures. For example, cognizers often form beliefs on the basis of perceptual
experiences they have. I might be appeared to redly (following Chisholm’s way
of putting this without endorsing his adverbialism) by a traffic light. I form
the belief (b4) that I’m appeared to redly, and warrantK for that belief arises by
virtue of that belief’s being formed in the right way, that is, via properly func-
tioning cognitive faculties in the right kind of environment (and absent defeat-
ers).101 Beliefs like b4 are what Alvin Plantinga calls properly basic beliefs.
The account of warrantK above is Alvin Plantinga’s with an added no-
defeater condition that follows the work of Michael Bergmann.102 Both
Plantinga and Bergmann understood the proper function of cognitive facul-
ties in such a way that “their functioning results in the cognitively healthy
doxastic response to the circumstances in which they are operating (which
will include, rather prominently, the input to the subject’s belief-forming
systems).”103 The solution (which is also Plantinga’s) to the new riddle of
induction, and therefore to the problem of projection, is that “[p]roperly
functioning human beings don’t typically project such properties” like being
grue or predicates like grue (Plantinga, Proper Function 1993, 133–134).
Taking Plantinga’s cue (Proper Function 1993, 133–135), we can extend the
proper functionalist solution to the new riddle of induction in such a way that
it helps us resolve issues in metametaphysics or metaontology generally. What
reasons do we have for preferring the language with a privileged quantifier
and non-grue-like predicates? Why should we think that grue-like predicates
fail to correspond with reality? It is because properly functioning human per-
sons do not appropriate such predicates when they accurately describe real-
ity. For searching inquiry into metaphysical matters, human persons ought
to use the privileged quantifier in a language outfitted with choice privileged
predicates. The choice predicates are those that properly functioning human
persons appropriate and can accurately describe reality and/or project in a
warrant-generating way in that type of inquiry with appropriate standards of
precision. At least, a great many of the predicates of the sciences are non-grue-
like and appropriate for metaphysicalC system building and warrant-produc-
ing projection in metaphysical inquiry. Is demonstration that one is using the
choicest predicates in one’s projections required for warranted beliefs in the
conclusions of such projections? Does one have to show that the predicates
one uses in descriptions of reality are distinguished in order to be warranted
in believing those descriptions? No. Proper functionalism about warrant is
externalist. A demonstration of the kind that this question references is unnec-
essary. One need not know or have access to that which generates warrant.
What matters is whether the conditions of warrant are satisfied.
Inquiry and debate is not stymied by an appropriation of the proper func-
tionalist solution to the problem grue-like predicates pose for knowledge-
conducive metaphysical inquiry. Defeaters can be proffered for claiming
that this or that predicate is that which properly functioning human persons
appropriate. Defeaters of the relevant kind may appropriate simplicity con-
siderations. Responses to such defeaters (proffering defeater-defeaters) will
sophisticate matters.
One’s belief b that this or that predicate is non-grue-like to the degree that
it is eligible to play a role in metaphysical inquiry can be a properly basic
belief. However, while b can be properly basic for some cognizer C, it can
also be indirectly warrantedK for that self-same cognizer. In other words, b
can be warrantK-overdetermined104 because in addition to b’s being prop-
erly basic, evidence can be amassed for regarding this or that predicate as
that which one ought to project and/or appropriate in academic contexts
in which one’s inquiry is knowledge-seeking metaphysical inquiry. Some of
this evidence will be in the form of theoretical knowledge. Some of the evi-
dence might be pragmatic. For example, one might argue (much like I did
in sect. 4.2 on quantifiers) that the simpler non-disjunctive predicates are
those used in empirically successful inquiry in the sciences, and that this suc-
cess provides justification for appropriating the very same (non-disjunctive)
predicates in metaphysical inquiry.
Some will say proper functionalism smacks of theism. These theorists will
argue that it is not at all in the spirit of a serious mainstream solution to a
very important problem. Or, at least, it is not in the spirit of a naturalistic
metaphysical methodology. But I should note that the proper functionalist
response is very much like Sider’s ontological realist reply to a similar prob-
lem. Sider wrote, “[a]n ideal inquirer must think of the world in terms of its
distinguished structure; she must carve the world at its joints in her think-
ing and language. Employers of worse languages are worse inquirers.”105
Properly functioning human persons are or at least approach ideal inquirers.
The solution is in the spirit of a mainstream approach to a similar problem.
But what about the complaint that proper functionalism is theistic? I think
theism is not entailed by the truth of proper functionalism about warrantK.
The best way to see this is by tackling the Swampman scenario, first articu-
lated by Donald Davidson and subsequently applied (as an objection) to
Plantinga’s proper functionalist account in the work of Ernest Sosa.106
Suppose lightning strikes a dead tree in a swamp; I am standing nearby.

My body is reduced to its elements, while entirely by coincidence (and
out of different molecules) the tree is turned into my physical replica.
My replica, The Swampman, moves exactly as I did; according to its
nature it departs the swamp, encounters and seems to recognize my
friends, and appears to return their greetings in English. It moves into
my house and seems to write articles on radical interpretation. No one
can tell the difference. But there is a difference.
(Davidson 1987, 443–444; emphasis in the original)
Swampman has knowledge once he comes into being. However, his faculties
were not designed, and so cannot function in accordance with a design plan.
Thus, Swampman’s faculties cannot function properly in the technical sense
Plantinga and Bergmann have in mind. We therefore have a counter-example
to proper functionalism about knowledge since that position requires that
we understand proper function in accord with a design plan as necessary
for warrantK, and warrantK as necessary for knowledge. Swampman has
knowledge but does not function properly in the relevant way, and so does
not have warrantK.
Plantinga’s initial response to Sosa insisted that Swampman is the type of
entity that can malfunction. For instance, Swampman’s organs can suddenly
fail, or he can contract a deadly fever. But if that’s right, then Swampman’s
faculties can also function properly. However, proper function is a notion
correlative with the notion of a design plan. Thus, by Plantinga’s lights, it
appears that an entity functioning in accordance with a design plan can pop
into existence by chance. That the entity in question was designed by some
literal creator or designer is not a necessary condition for proper function of
that entity in accordance with a design plan.
Whatever one makes of Plantinga’s reply to the proposed counter-example,
this much seems clear. Plantinga does not (or at least did not at one time)
maintain that proper functionalism entails theism or designing activity by
a literal designer. I believe this is correct, and Swampman worries go away
once we recognize, as Bergmann (2006, 148–149) has correctly noted, that
because Swampman is a molecule-for-molecule replica of Davidson (includ-
ing even Davidson’s DNA), Swampman is a full member of the species and
natural kind Homo sapiens. As such, he has a design plan, that which is in
place for Homo sapiens. Of course, one might still ask the question, how can
the atheist countenance truthful talk of a Homo sapiens’s faculties function-
ing the way they are supposed to function? They can do that by insisting on
the truth of a robust realism about supposed to facts. It is consistent with
atheism and proper functionalism that there exists an abstract Platonic sche-
matic for the proper function of various human faculties. That schematic is
composed of a set of brute abstract normative truths about how human fac-
ulties ought to function in our cosmos. When the faculties of human beings
function in accordance with those true principles, they function properly in
accordance with a design plan. They function the way they are supposed to
(the principles are normative) function.
The reasoning of this section should lead one to adopt the following
directive,
(Directive #10 (D10)): One’s metaphysical inquiry aimed at true meta-

physical worldview and metaphysicalC system building should preclude
predicating with grue-like (or unnecessarily complex) predicates, and it
should not project predicates that properly functioning human persons
do not typically project.
Section 4.4: MetaphysicalC System Building and Analysis
Section 4.4.1: Philosophical Analysis vs. Conceptual Analysis

MetaphysicalC system and metaphysical worldview building must be done
with a choice logic, choice quantifiers, and choice predicates. It must be
executed in a manner that respects the existence of meaningful entities,
truths, mental states, mental events, instances of rational belief, warrant/
epistemic justification, knowledge, knowledge of otherly-mental causation,
and mind-independent truth-bearers that are propositions. Such worldview
and system construction must also respect as expansive of a metaphysical
Realism as is rationally permissible, where the sense of rationality in play
here is that which is connected to cognitive proper function and cognitive
malfunction. None of this touches on the role of philosophical analysis in
metaphysicalC system building, and metaphysical worldview building more
generally. To get clear on that issue, we must first say something about what
philosophical analysis is.
I will follow to some extent King (1998; 2007; 2009; 2016, 254–256),
and to a lesser extent Sosa (1983), by envisioning philosophical analyses of
the kind that are pertinent for metaphysical theorizing as necessarily true
(when true) universally quantified biconditionals whose left bijuncts con-
tain (at least) some representation token that picks out (or intends to pick
out) some complex relation or property, and whose right bijuncts reveal
(a) that which (at least purportedly) makes manifest those constituents that
one must report on to specify the metaphysical nature of the target relation
or property, and (b) the way those constituents hang together. Or as King
put it, “an analysis” says “what the component properties and relations
are that make up the complex property being analyzed and how they are
combined to form this complex property” (King 2016, 255). For example,
David Lewis at one time provided what can be understood as a philosophi-
cal analysis of deterministic causation,
(Counterfactual Analysis of Causation (CAC)): Necessarily, for any

event c and any event e, c is a deterministic cause of e, just in case, c
and e are non-identical, c and e are both mereologically and logically
distinct or unrelated, and e counterfactually depends on c, or else there
is step-wise counterfactual dependence between e and c.107
Say that e counterfactually depends on c, just in case, were c to fail to
occur, e would fail to occur. Call the relation that is deterministic causation
referenced by the left bijunct of CAC the analysandum, and call the right
bijunct the analysans. The relation that is deterministic causation is here
being analyzed in terms of more basic properties and relations, although
philosophical analyses need not be reductive in this way.
I am understanding CAC in such a way that it is not providing an analysis
of the concept of deterministic causation. Rather, it is analyzing the rela-
tion itself. Some would object to this by insisting that proper philosophical
analysis reveals something directly about our concepts (conceptual analy-
sis). However, conceptual analysis will not be possible given certain views
about the nature of concepts.
According to the reigning consensus in contemporary philosophy of mind
and cognitive science, concepts are mental representations that are expressed
by constituents of compositional thoughts in larger internal representational
systems.108 And although constituents of every thought express concepts,
some of our thoughts express propositions. I will therefore associate with
this consensus view the idea that concepts themselves are part of the repre-
sentational contents of propositional thoughts, and that they are robustly
associated with the contents (if not part of the contents themselves) of prop-
ositions (following Burge 2010, 540; 2009, on propositional representa-
tional contents; Fodor and Pylyshyn 2015, 8). The aforementioned internal
representation systems incorporate beliefs and other propositional and psy-
chological attitudes. Those attitudes have mental representation structure
(this is part of what makes them mental representations themselves), and
that structure is where some concepts reside. One can distinguish between
psychological attitudes by giving attention to the various causal roles those
attitudes play (a point made by many others already cited).109
Following Fodor (1998), I will maintain that lexical concepts do not
enjoy structure forming and reference fixing intentional constituents. The
necessary conditions for the possession of a single lexical concept are not
the necessary conditions for the possession of some distinct individual lexi-
cal concept (to closely paraphrase ibid., 14). That is to say, I will assume
that conceptual atomism is true.110 However, I will remain silent on the
question of whether concept atoms are symbols for a language of thought
(as in Fodor 1975; 2008). As has been noted by Beck (2013), conceptual
atomism makes sense of the well-known fact that our track record for defin-
ing concepts in terms of other concepts through the provision of necessarily
true biconditionals is spectacularly horrible.
The philosophical methodology that is conceptual analysis seems to
demand precisely what conceptual atomism precludes. Consider the follow-
ing explications of the method,
The situation is entirely different in the case of philosophical defini-

tions. In philosophy, the concept of a thing is always given, albeit con-
fusedly or in an insufficiently determinate fashion. The concept has to
be analysed; the characteristic marks which have been separated out
and the concept which has been given have to be compared with each
other in all kinds of contexts; and this abstract thought must be ren-
dered complete and determinate.111
Conceptual analysis is an important means to an end. The means is

clarificatory; the end is the pursuit of truth and understanding. Intro-
spection on our concepts enables us to clarify them, and enables us
thereby to reach a more perfect understanding of them: of how they
hang together, and of how they apply to things and events in the world.
The standard way of trying to clarify a given concept is to provide an
analysis for it: that is, a statement of the individually necessary and
jointly sufficient conditions for its application. The statement of such
conditions usually involves other concepts, taken to be better under-
stood or previously clarified.112
We construct a sense out of its constituents and introduce an entirely new

sign to express this sense. This may be called a ‘constructive definition’
[‘aufbauende Definition’], but we prefer to call it a ‘definition’ tout court.113
These sources envision associated descriptive senses114 in the analysans as

specifications of more primitive, simple, and/or basic concepts (these are the
“characteristic marks which have been separated out” in the quotation from
Kant, and these are the concepts that are “taken to be better understood”
in the quotation from Tennant).115 If concepts have no substructure and no
constituents out of which one can “construct” descriptive senses, then con-
ceptual analysis of the kind envisioned previously will not be possible. But
what exactly is a descriptive sense? It turns out that conceptual analysis is
a methodology at home within a descriptivist tradition in the philosophy of
language.116 That tradition affirms, inter alia, that nonlogical singular terms
(and even natural kind terms like ‘water’) have meanings given by descriptive
senses. A speaker’s association of a descriptive sense with a singular term fixes
the meaning of that term. Descriptive senses deliver individually necessary
and jointly sufficient conditions that determine the references of such terms.
These senses, therefore, provide the conditions of application of the concepts
standing behind singular terms. Analyses of the meanings of terms and asso-
ciated concepts consist of proffering specifications of associated descriptions
(the descriptive senses) (see the discussion, without endorsement, of these
ideas in Soames 2005, 7–13, upon which I lean for my exposition).
Conceptual atomism coheres nicely with the work of a significant num-
ber of philosophers who migrated away from descriptivist theories of sin-
gular and natural kind terms. That movement was led by David Kaplan,
Saul Kripke, John Perry, and Hilary Putnam, and has been defended more
recently by Scott Soames.117 According to this anti-descriptivist move-
ment, the meanings of expressions are not given by conceptual contents
“in the head” of employers of those expressions but are, at least in part,
fixed by external matters (semantic externalism). That which determines
the references of singular terms are not descriptive senses. Rather, that
some singular term like the proper name ‘Barack Obama’ successfully
refers to the individual that is the 44th President of the United States
is a matter of becoming a part of the right causal chain, the chain that
began with an initial baptism in which a speaker, acquainted with Barack
Obama in their perceptual experience, first associated Barack Obama
with ‘Barack Obama’ (this is the causal theory of reference).118 And most
importantly, for my purposes, anti-descriptivists do not maintain that
a central aim of philosophical inquiry is conceptual engineering and
clarification through conceptual analysis as previously described. In the
domain of metaphysical inquiry and metaphysical worldview building,
there are important metaphysically necessary a posteriori truths about
the natures of things. Empirical evidence (particularly from the sciences)
is required to acquire warrant for belief in such truths. I would like to
situate my metaphysicalC system within this anti-descriptivist tradition
(i.e., I affirm all of the commitments of this paragraph). We, therefore,
have the following principle,
(Directive #11 (D11)) With respect to metaphysicalC system build-

ing and metaphysical inquiry aimed at uncovering knowledge of the
nature of causation, one should, when appropriate, invoke philo-
sophical analyses and refrain from invoking conceptual analyses in
one’s inquiry.
When is it appropriate to use philosophical analyses? When one discerns

that in the search for and/or construction of a philosophical analysis of cau-
sation one can acquire knowledge-conducive warrant for one’s belief in a
philosophical analysis essential to that search and/or construction.
A goal of the current project will be to articulate and defend a philosophi-
cal analysis of the causal relation. I hope to be able to at least defend a nec-
essarily true, universally quantified biconditional of the form,
(Causation Schema): Necessarily, for any event c and any event e, c is

a deterministic cause of e, just in case, Φ. (See chapter 9: sect. 4 for my
analysis.)
But that is not all. My theory/analysis of causation will also involve an

ascription of important properties of the causal relation, properties that
I believe exhaust its metaphysical nature. I believe that we can justify those
ascriptions by looking to certain empirical evidences provided by physical
inquiry. But before discussing my view of the relationship between meta-
physics and physics, I’d like to distance myself from a methodology called
empirical analysis.
Section 4.4.2: Empirical Analysis
One of the most able defenders of the conserved quantity theory of causa-
tion is Phil Dowe (q.v., my discussion of the theory in chapter 3: sect. 4.3.1).
His theory is an empirical analysis of causation (Dowe 2000, 1–13). Empiri-
cal analyses are metaphysically contingent truths (when true), empirically
adequate theories of objective features of the world. Unlike philosophical
analyses, they are not necessarily true if true, and unlike many conceptual
analyses of causation, they are not attempts to provide a necessarily true
analysis of our folk causal concept. The approach is adopted by a significant
number of theorists working on causation.119 Dowe’s specific conception of
it is problematic. On the one hand, Dowe maintains that empirical analyses
are not conceptual analyses because they aim “to map the objective world,
not our concepts” (Dowe 2000, 3). On the other hand, he is “happy to
grant” that “any empirical analysis will still be a kind of conceptual analy-
sis, for example, of the concept [of causation] implicit in scientific theories”
(ibid., 11). Dowe is also “happy to think of the task of empirical analysis
as . . . conceptual analysis of a concept inherent in scientific theories.”120
But if empirical analyses are a brand of conceptual analysis of scientific
concepts, then Dowe’s methodology will be precluded from proper meta-
physicalC system building by D11, which is unqualified.
Douglas Kutach’s recent work (Kutach 2013) is an attempt to provide an
empirical analysis of causation in the tradition of Dowe’s project, although
it adds conditions for the success of such analyses. For Kutach, to discover
the true empirical analysis of causation, one starts by using platitudes under-
stood as deliverances of common sense (e.g., <The match strike caused the
fire spark.>) (see ibid., 10, on these points). One uses these platitudes to
acquire a clue about where to look for empirical happenings whose obtain-
ing explains why we have the concept of causation in the first place (Kutach
focuses on effective strategies or “the empirical phenomena behind” them
(ibid., 17)). Subsequent to finding the relevant empirical happenings, one
passes systematic investigation of those happenings off to the scientist.
Experimentation ensues, and the metaphysician looks to the best scientific
explanation of the experimental results or outcomes. The concepts used to
help facilitate that best explanation are those that track causation-like rela-
tions and are therefore concepts that build a proper empirical analysis of
causation. These concepts will be scientifically regimented or “scientifically
improved concepts of” causation.121 The resulting picture of the metaphysics
of causation need not (and according to Kutach’s account, does not) respect
(a) the deliverances of common sense, or (b) our intuitions about metaphysi-
cally possible cases. For example, Kutach believes his theory need not respect
the principle that events do not bring about themselves. Likewise, Kutach
insists that “we should care little about how our concept applies to highly
unrealistic possibilities and not at all about whether it applies to absolutely
every possibility.”122
Kutach’s methodology prohibits one from using causal scientific explana-
tions of the experimental results he’s interested in. If the explanations (call
them ES) of those results explicitly employ our full notion of causation, then
nothing in the ES will reveal to us informative non-causal details about the
deep nature of the causal relation. The project of using non-causal, albeit,
causation-like notions that are scientifically “improved,” will never get off
the ground. This explains why Kutach insists on replacing our full notion
of causation, the notion that would be at work in the ES, with what he
believes are more scientifically respectable notions: determination, fixing,
and terminance.123
The causation of the metaphysicians (what I am interested in in this work)
is what Kutach calls culpable causation.124 That relation “plays no essen-
tial role in the metaphysics of causation” (ibid., 266). The metaphysician’s
causation is but a helpful concept. We use it to facilitate understanding
and learning, and we use it to help us manipulate the world as described
by the special sciences. Does this mean that culpable causation does not
exist? Apparently not (see ibid., 267). Culpable causation plays a role in
causal explanations, and in heuristics for learning (ibid., 275–277) in the
higher-level sciences. However, learning and explaining are epistemic pur-
suits (Kutach favors an epistemic conception of explanation; ibid., 277).
Thus, the causation of the metaphysicians seems to be relegated to the status
of an epistemic tool that would fail to exist were there no causal cogniz-
ers seeking explanations and knowledge. In other words, Kutach seems to
be a causal eliminativist. That is to say, Kutach rejects the existence of a
mind-independent obtaining causal relation, for culpable causation (quot-
ing Kutach) “does not play any role in how nature evolves,”125,126 Causal
“culpability is metaphysically superfluous . . .”.127
There are three problems with Kutach’s approach to the study of cau-
sation. First, Kutach’s method subsumes Dowe’s method. For that reason
it will, like Dowe’s methodology, violate D11. Second, several chapters of
the current work provide independent lines of support for the claim that
(culpable) causation is indispensable to fundamental physics. If I’m right,
then our notion of causation will not be replaced by scientifically respect-
able concepts because the concept of causation is a scientifically respectable
concept serving as the referent of our notion of causation in the ideologies
of some of our best physical theories (see sect. 4.5.5.2 for my understanding
of an ideology).
Third, causal explanation looks factive. An implication of this is that
when one causally explains that p, by appeal to q, it will be true that p. And
as has already been pointed out, not a few theories of explanation, includ-
ing theories of causal explanation, entail that the explanans (in this case,
q) must hold as well (q.v., sect. 3). Indeed, successfully causally explain-
ing p, with knowledge of q, while fully grasping the explanation relation
between p and q, yields knowledge of p (and therein lies one reason for
regarding causal explanation as factive). Presumably, the type of causation
backing causal explanations in psychology is culpable causation. Psychol-
ogy is a higher-level special science. And psychologists typically do regard
their discipline as one that is in the business of causally explaining.128 But if
the relation backing causal explanation is culpable causation, the explanans
and explanandum of a causal explanation in psychology are both true, and
we have reported on in the explanans and explanandum, the causes and
the effects, respectively, then why can we not infer that there exists a true
culpable causal fact, namely that the events reported on in q cause the event
reported on in p literally interpreted? Moreover, how could there exist a
legitimate causal explanation of p in terms of q that was backed by a causal
relation that does not actually exist in the mind-independent world irrespec-
tive of whether that relation itself stands in a reduction relation to some
more fundamental goings-on?
I encourage the adoption of the following directive,
(Directive #12 (D12)): With respect to metaphysicalC system building

and metaphysical inquiry aimed at uncovering knowledge of the nature
of causation, one should refrain from invoking Dowe-like empirical
analyses, or Kutach-style empirical analyses.
Section 4.5: The Epistemology of Metaphysics

Recall that metaphysical inquiry is an attempt to build a metaphysical
worldview. A metaphysical worldview provides (purported) truths about
what exists, the nature of what exists, and the hierarchy of being. But how
does one come to know about what exists? How does one discern correctly
the natures of existing things? And how does one acquire knowledge of
hierarchical structures or relations holding between existing things? These
are difficult questions that I cannot possibly answer in a chapter, let alone a
sub-section. I hope it will appease the reader to provide a preliminary out-
line of a sketch of those sources of knowledge that I think are important to
any attempt at answering the preceding questions.
Section 4.5.1: The Basic Sources of Justification

Let a basic source of epistemic justification (for some belief b) be a source
that a cognizer uses for that justification without positively relying upon
the employment or use of any other source of justification.129 Perception is
a basic source of justification. It is that capacity whereby properly function-
ing human persons can and sometimes do perceive reality. There are various
ways we do this (here I follow standard discussions, including Plantinga,
Proper Function 1993, 89–101). One way is sensuous in that it involves var-
ious faculties of human persons to sense their surroundings when those sur-
roundings impress upon sense organs.130 If sense capacities and perceptual
apparatuses are functioning correctly, when the environment affects sense
organs in a way that causes an individual to perceive, sensory phenomenol-
ogy is produced. Sensory phenomenology is what it is like to be enjoying
or having a sensory experience.131 There are ways of perceiving that do not
involve sensory phenomenology. These other ways constitute nonsensuous
perceptual experiences (as in Plantinga, Proper Function 1993, 92). The
paradigm example of nonsensuous perception is blindsight. In cases of
blindsight, individuals are able to judge fairly reliably that, for example,
there is an object present at a certain location, or that an object is arranged a
certain way, or even that it points in a certain direction although that object
is not in that individual’s visual field (see Boyer et al. 2005; Kentridge et al.
1999; Lormand 2006, 340, writes “[a]rguably there are perceptual states
without experience, in subliminal perception, ‘blindsight,’ and ‘early’ states
in processing in the retina, lateral geniculate nucleus, and (perhaps) primary
visual cortex”; Weiskrantz 1980; 1986; 1988).
I will not commit to any one specific theory of perception. All that my
epistemology of metaphysics will require is that perceptual experiences of
the sensuous and nonsensuous varieties (when perceptual and cognitive fac-
ulties are functioning properly) be understood as that which can serve as a
reliable basic source of epistemic justification for perceptual beliefs. Meta-
physical worldview and metaphysicalC system building should be open to
being informed by this basic source. In addition, non-basic sources of jus-
tification like the sciences will rely upon the more basic source of epistemic
justification that is perception.
Properly functioning human persons introspect, thereby giving attention
to some of their mental states. The introspective process is temporally proxi-
mate, immediate, and first-personal. As with perception, I will not assume
any fine-grained theory of introspection. I require only the doctrine that
properly functioning human persons who introspect acquire epistemic justi-
fication for beliefs about some of their own presently (or close to presently)
existing mental states.132 Beliefs formed by introspecting are introspective
beliefs. Those introspective beliefs that enjoy epistemic justification can help
with metaphysical worldview and metaphysicalC system building.
Memory is that capacity whereby properly functioning human persons
can and sometimes do remember truths. As such, it can only be a basic
source of justification. It is not also a basic source of knowledge. This is
because a necessary condition for remembering that p is having already
acquired one’s knowledge of p by way of a distinct source of knowledge.133
One might have reasoned to p, or someone might have testified to you that
p, or you might have formed p on the basis of perceptual experience. The
precise means whereby memory gives cognizers justification is disputed, but
very few doubt that it can provide something like epistemic justification.
Concerning metaphysicalC system building, theoreticians can and do depend
upon the proper use of memory for justification in many different ways.
I should add that some philosophers and scientists maintain that the process
of remembering (involving the source that is memory) requires the existence
of certain kinds of causal chains, or entails the existence of certain kinds
of causal activity in the brain (e.g., Alkon 1989; Audi 2011, 64–65; Moser
1989, 119). I believe they are right. When a properly functioning human
cognizer leans upon memory as a source for knowledge or epistemic justifi-
cation by remembering that p, certain causal facts come out true.
Properly functioning cognizers often use a basic source of justification
that is their faculty to reason in order to put their understanding of concepts
and/or propositions to work so as to ascertain truths. Reason can provide
a priori justification in the form of reasons to believe truths. For example, a
cognizer C can come to understand the proposition <Bachelors are unmar-
ried males.>. C can, after, or at the same time as understanding that propo-
sition, form a belief (b) with that proposition as its content. C’s belief b
can enjoy epistemic justification on account of C’s grasping the conceptual
relations involved in the structure of b’s content. Truths that can be the
contents of beliefs that receive their justification in this way are self-evident
or incorrigible truths.134
Some justified beliefs of properly functioning cognizers are inferentially
justified on the basis of deductive reasoning. Some other beliefs are inferen-
tially justified on the basis of inductive reasoning. Only the skeptic would
deny such claims. I will take it for granted that any skeptical problems of
deduction or induction can be solved and that reason can serve as a source
of justification by facilitating reliable deductive and inductive inferences.
(Directive #13 (D13)): With respect to metaphysicalC system build-

ing and metaphysical inquiry aimed at uncovering knowledge of the
nature of causation, one should use the basic sources of justification
(i.e., perception, introspection, memory, and reason) for that inquiry
when appropriate.
Section 4.5.2: Intuition and the A Priori

There is an extension of the capacity to reason that is a priori intuition. I will
follow the contemporary rationalist tradition and affirm that intuiting that
p is a type of conscious episode wherein it seems to a cognizer C that p. That
is to say, I agree with George Bealer and Ernest Sosa.135 Intuitions are “intel-
lectual seemings” (Sosa 2007, 60) that can serve as a priori justification
when those seemings are rational. However, unlike Sosa, I do not character-
ize rational intuitions in terms of manifestations of epistemic competencies.
And unlike Bealer, I do not maintain that intuitions are rational or evidential
on account of a “modal tie between” the deliverances of a seeming “and the
truth” (Bealer 2002, 102). Rather, I maintain that an intuition is rational for
a cognizer C who intuits that p, when C’s intellectual allure to p is due to
the proper function (in the sense, in the way, and in the type of environment
suggested by the Plantinga/Bergmann theory of warrant discussed in sect.
4.3) of those faculties of C that are responsible for C’s discriminating among
various propositions that C “understands well enough, the true from the
false” in the domain of the conceptually, logically, or metaphysically pos-
sible, impossible, and/or necessary.136 Rational intuition is a source of epis-
temic justification. And when its deliverances are undefeated, it is a source
of warrant. How do we become allured to propositions when they seem to
us to be the case? Sosa answers,
When we are intuitively justified in believing p, we are attracted to so

believe through the mere grasp of the content that p, which we then
entertain. So, it is the mere entertaining of that very content that
prompts attraction and perhaps assent.137
Rationally intuited truths are not incorrigible truths. They are not self-
evident. One is justified in believing incorrigible truths by virtue of grasping
the relevant conceptual relations. With respect to truths one rationally intu-
its, the justification issues forth from understanding propositional content.
Grasping conceptual relations and understanding a proposition are two
distinct types of states of affairs.
Rational intuition plays an extremely important role in justifying meta-
physical theses. In fact, one way I will use rational intuition is in defense of
descriptions of how one can acquire justification for believing in various
metaphysical possibility claims. As we will see in chapters 2 and 5, impor-
tant conclusions for the study of causation can be reached from very weak
metaphysical possibility claims.
Many philosophers of science and metaphysicians of science will object
to my use of intuition for the study of causation. I cannot address all of
their worries, nor can I adequately defend the characterization of intuition
I have adumbrated in the present work in the space allotted. Fortunately,
Bealer and Sosa have answered most of the objections that can be lodged
against the above account (notwithstanding my appropriation of proper
function).138 Despite their persuasiveness, I do feel that I should assuage the
worries of those empiricist and scientifically minded philosophers who have
expressed disdain for the a priori because it does not appear to play a role in
the context of justification and/or discovery in scientific inquiry.
SECTION 4.5.2.1: THE A PRIORI IN SCIENCE
Several philosophers have argued persuasively that some scientific thought

experiments can provide a priori reasons for belief in scientific inquiry.139
A potential example of just such a thought experiment is provided in the
work of Galileo Galilei (1564–1642).140 And I should add that we need not
fixate on just this case involving Galileo, for his employment of a rationalist
scientific methodology is not at all a sui generis deposit in the history of sci-
ence. There are not a few instances of rationalist scientific methodology in
the history of mechanics.141 In fact, many important theorists affirmed the
central tenets of quantum chromodynamics (QCD) on the basis of a priori
considerations that were quite independent of experimental results. And
here I have in mind acceptance of chiral symmetry and the indispensability
of color, though I do not mean to suggest that these tenets did not receive
some empirical support. The idea is that important physicists justified their
incorporation into the theory on the basis of what they understood to be a
priori considerations (see Peter Galison’s remarks in Galison 1997, 643).
Recall that Aristotle (Complete Works vol. 1, 366) maintained that a heavy
body bH undergoes faster terrestrial free fall than a lighter body bL (all other
things equal). In response, Galileo reasoned to a refutation of Aristotle and a
new more accurate theory of free fall without executing an experiment that
provided additional empirical information. In other words, from the armchair
(as it were), Galileo was able to show that Aristotle’s theory of free fall had
to be false and that a newer theory had to be true. He reasoned as follows:
If Aristotle is correct, then if bH is conjoined with bL, the resulting complex
object (bH & bL) will fall faster than bH, since the former is heavier than the
latter. However, (bH & bL) has both a lighter proper part and a heavier one.
The lighter proper part that is bL will give rise to drag during the terrestrial
free fall of the complex (bH & bL), causing it to fall slower than bH, reductio
ad absurdum. To avoid the reductio, Galileo gave merely mental attention to
a hypothetical scenario in which the relevant happenings obtained. He subse-
quently determined in an a priori manner that the best and only theoretically
unifying account of the facts of terrestrial free fall was that every object falls
at the same speed. That he regarded deliverances like these as a priori seems
clear, for in one place he stated, “[w]ithout experiment, I am sure that the
effect will happen as I tell you because it must happen that way.”142
John D. Norton has challenged the rationalist reading of this case (Nor-
ton 2004). Norton argues that thought experiments that yield knowledge of
non-logical truths should be thought of as arguments, and Galileo’s thought
experiment can be characterized as an argument. Galileo’s reasoning under-
stood as pure thought, and not as an argument, cannot produce knowledge.
This is because (quoting Norton) “[a]ll pure thought can do is transform
what we already know.”143 Epistemic justification for belief in Galileo’s
conclusion(s) issues forth from belief in the premises of the argument to
which Galileo’s reasoning reduces. Furthermore (quoting Norton),
The actual conduct of a thought experiment consists of the execution

of an argument, although this may not be obvious since the argument
may appear only in abbreviated form and with suppressed premises.144
Something like Norton’s objection can be given the following formulation,
(1) Arbitrary scientifically respectable thought experiment T can be repre-

sented as an argument A for conclusion C.
Norton adds to this initial premise/principle the further claim that
“[b]elief in the outcome-conclusion of a thought experiment is justified
only insofar as the reconstructed argument can justify the conclusion”
(ibid., 50). But this seems false if our concerns are general. Our justifi-
cation for a belief in such an outcome-conclusion may be epistemically
overdetermined and so enjoy multiple independent sources of justifica-
tion. I, therefore, cut premise (1) short and ameliorate (in my opinion,
at least) Norton’s ideas via the content of the consequent of (3).
(2) Performing an arbitrary scientifically respectable thought experiment T
is nothing over and above “the execution of” A (ibid., 50).
(3) If (1) and (2), then belief in C on the basis of the premises of A is epis-
temically justified, only if, the supporting beliefs in the premises of A are
epistemically justified.
(4) If belief in C on the basis of the premises of A is epistemically justified,
only if, the supporting beliefs in the premises of A are epistemically
justified, then epistemic justification for belief in C is not transmitted
from beliefs about the contents of a thought experiment understood as
something other than an argument (e.g., pure thought).
(5) Therefore, epistemic justification for belief in C is not transmitted from
beliefs about the contents of a thought experiment understood as some-
thing other than an argument (e.g., pure thought).
If epistemic justification is necessary for knowledge, and belief b in a con-

clusion arrived at on the basis of a thought experiment does not transfer or
generate epistemic justification for b, then believing on that basis will not
generate knowledge (assuming there is no other source of epistemic justifica-
tion or warrant in play).
Why believe premise (2)? Here is Norton’s justification,
if we accept . . . [premise (1)], we believe that the reach of thought exper-

imenting coincides exactly with the reach of argumentation. If thought
experimentation opens up some other channel to knowledge, how curi-
ous that it should impersonate argumentation so perfectly! How are we
to explain this coincidence, if not by the simple assumption that thought
experimenting merely is disguised argumentation?145
Norton motivates (1) by way of the thesis that every thought experiment
can have its content represented by an argument (more on this below). Nor-
ton affirms that the truth of premise (2) best explains (1) and that you can-
not appropriate that explanation if you maintain that thought experiments
are something other than arguments.
I have two responses to (1)–(5). First, suppose that argument (1)–(5) is
sound. With respect to Galileo’s specific T, it would remain the case that the
corresponding argument that is A (i.e., the argument that represents Gali-
leo’s thought experiment) includes at least one premise that is not supported
by Galileo’s experience. That one premise will constitute an a priori reason
to accept Galileo’s desired outcome-conclusion (viz., that bodies or objects
fall at the same speed). To see this, consider something close to one of Tamar
Szabó Gendler’s characterizations of Galileo’s A,146
(6) “Natural speed is mediative.” [meaning, natural speed is a property

“such that if a body A has natural speed s1, and a body B has natural
speed s2, the natural speed of the combined body A-B will fall between
s1 and s2.” (Gendler 1998, 404)]
(7) “Weight is additive.”
(8) “Not all weights and natural speeds are either zero or infinite.”
(9) Therefore, “[n]atural speed is not directly proportional to weight.”
[from (6)–(8)]
(10) The best unifying explanation of facts (6)–(9) is that “all natural speeds
are the same.” [Premise]
(11) Therefore, “all natural speeds are the same.”147 [by a type of
abduction]
The inference from (6)–(8) to (9) is deductively valid. The further infer-
ence all the way to (11) is ampliative in that it is an inference to a best
theoretical unification of facts (6)–(9). Theoretical hypotheses that unify do
so by drawing together disparate theoretical and/or observational facts. We
have this indirect and non-basic source of knowledge (in part) because we
have varying basic and non-basic sources of epistemic access to facts about
the world. For example, I might (and assume these are facts), on the basis of
perception, see a building on fire. I might hear in the distance a loud siren,
I might have seen (perception) and at the present moment recall (memory)
that my friend dialed the emergency phone number. In that moment, I will
faultlessly infer that firefighters are on their way in their emergency vehicle
to the scene of the fire. This belief is, for me, epistemically justified on the
basis of the unifying role it plays. This is, in part, because what (quoting
Moser et al.) “separates a good hypothesis from a bad hypothesis is not
just its truth value, but also its ability to unify.”148 In this case (we are sup-
posing), Galileo’s judgment that (9) holds issues forth from facts (6)–(8)
and that (11) explains in the unifying way. But on what grounds does one
believe (10)? Gendler correctly judges that “the only way to maintain” (6)–
(9) “simultaneously is to assume that all natural speeds are the same.”149
The best unifying explanation is seen by Galileo (on the assumed recon-
struction of the case, at least) to be the only unifying explanation in that it
is the only theoretical hypothesis that renders facts (6)–(9) compatible (as
Gendler has noted). But seeing that truth is not an empirical matter. It is
true that (6)–(8) are empirical deliverances. However, seeing/intuiting that
some explanation of them renders their seeming inconsistency, consistent,
is an a priori matter. No amount of raw empirical data (I exclude things
like testimony for the purposes of deliberation) directly about the natural
world will reveal to one the relevant truth. Thus, even given Norton’s view
of thought experiments, such entities depend upon a priori justification, and
that a priori justification is essential to the acquisition of new knowledge
about the world in Galileo’s case. My understanding of the case in question
does suggest that at least sometimes abductive inferences of this kind are a
priori, but that viewpoint has been ably defended in more than one way by
several renowned philosophers.150
Second, recall that Norton seems to be committed to the thesis that the
truth of premise (2) explains the truth of (1), and that you cannot appropri-
ate that explanation if you maintain that thought experiments are not argu-
ments, although they nonetheless provide epistemic justification by being in
some sense pure thought. But what exactly does the explanandum (premise
(1)) in this case really mean? What is a formulation of a thought experi-
ment? Is the idea just that one can, in principle, formulate an argument
that includes premises that serve as descriptions of various features of the
content of T, and that such an argument moves from those descriptions or
expressions about that content to the conclusion reached on the basis of the
content of T? If any such formulation is even approximately psychologically
realistic, it will be (often enough) enthymematic. With this Norton seems to
agree.151 However, if a cognizer (call her Diana) thinks through T, to a cer-
tain statement or purported factP, and that process of thought is psychologi-
cally no more than “the execution of an” enthymematic deductive/inductive/
abductive “argument” A (as Norton maintains), then epistemic justification
for Diana’s belief that C will not be transmitted from Diana’s beliefs in the
premises to her belief that C. This is because A is (psychologically) enthyme-
matic. The conclusion does not follow from the psychologically real beliefs.
Nor will the conclusion be adequately inductively or abductively supported
by the relevant premises because, again, the argument is enthymematic. In
the purportedly deductive case, Diana’s noetic structure would lack virtue
were it to include a belief in C based on the premises of an invalid argument
(especially if, in the case of a purportedly deductive argument, the set of
beliefs in the base included the erroneous belief that the premises of A entail
C, as one might require due to matters about closure). Thus, if Norton is
right about the content of thought experiments reducing to the contents of
psychologically real (and for that reason often enthymematic) arguments,
then few thought experiments deliver to cognizers epistemic justification.
Thus, I believe I’ve shown that if (1) and (2) are true, and (2) explains (1),
then a great many thought experiments (those that are psychologically real
and enthymematic) fail to provide cognizers with epistemic justification. How-
ever, Norton’s project is not a skeptical one. It is revisionary. Norton seeks to
salvage the avenue to epistemic justification that travels by way of thought
experimentation. We have therefore made trouble for Norton’s general under-
standing of thought experimentation.
We should make room in our epistemology of metaphysics for the a
priori, especially if we should make room in our epistemology for sound
scientific method that includes rationalist thought experimentation of the
kind Galileo was engaged in. So we have,
(Directive #14 (D14)): With respect to metaphysicalC system building and

metaphysical inquiry aimed at uncovering knowledge of the nature of
causation, one should invoke intuition and a priori reasons more gener-
ally as sources of epistemic justification for that inquiry when appropriate.
Section 4.5.3: Probability and Abduction

The present work will assume that probabilistic reasoning and inference are
an indirect part of a reliable and sound epistemology of metaphysics. I will
presuppose that the correct interpretation of probability in the context of
metaphysical inquiry is a modified brand of subjective Bayesianism. Degrees
of belief, or credences (or conditional versions thereof), can be accurately
represented by probabilities or probability functions. These credences, in
some sense, must dance to the rules of the probability calculus and should
be updated through the process of classical conditionalization such that
your current credence Prold(H) should be understood in such a way that
when evidence E for H comes in, Prold(H/E) is identical to Prnew(H), where
‘Prnew(H)’ represents the new degree of belief you have after conditionaliz-
ing on the evidence you’ve received for H. My use of the terms ‘must’ and
‘should’ underwrites a further aspect of subjective Bayesianism, viz., that
the rules of Bayesian probability theory are normative, not just descrip-
tive (Zynda 2016, 360). What I mean by this is not that the axioms and
theorems of probability are normative. Rather, the rules for how credences
should be structured and related to one another are normative. These rules
are backed by the axioms and theorems of probability theory (Stefánsson
2017, 573–574).
Often, some of the normative elements of subjective Bayesian approaches
to probability are understood in terms of rationality constraints, where the
thought is that if one’s credences do not abide by the normative rules backed
by probability theory, then one will fail to be rational. Likewise, if one’s
credences are not updated in accordance with classical conditionalization,
then one will be irrational. In other words, there’s a condition on rationality
that is probabilistic coherence (Howson 2014, 131–132). As Paul Horwich
put the thought, “the Bayesian approach rests upon the fundamental prin-
ciple . . . [t]hat the degrees of belief of an ideally rational person conform to
the mathematical principles of probability theory.”152 This is probabilism or
what is sometimes just called Bayesian epistemology (see Jeffrey 1992; Joyce
1998 and 2004, where Joyce connects probabilistic consistency to accuracy
considerations).
Although my choice interpretation of probability in the context of meta-
physical inquiry is a subjective Bayesianism, I am not a probabilist. I am
agnostic about whether the normative rules of subjective Bayesianism
support rationality constraints in general. This is because, at present,
I believe no one knows what kind of rationality the demand for probabilistic
coherence incorporates. No one appears to be able to justifiably answer the
question, on pain of what kind of non-pragmatic irrationality should our
credences be probabilistically coherent? I believe no one is yet justified in
believing that the rationality in question has anything directly to do with
conditions for knowledge or warrant. Here’s why. No one knows what the
precise relationship is between full beliefs (where full belief is a necessary
condition for knowledge) and credences. In the literature, there appear to be
three major positions. The first view says that the metaphysics of doxastic
states like beliefs and credences suggests that the two types of entities are dis-
tinct types of psychological states (this is the view I favor and will assume).
No one has connected credences (understood in this autonomous fashion)
with truth, warrant, or full belief. However, most Bayesians do not maintain
that beliefs and credences are distinct types of psychological states.153 A few
Bayesians affirm that one can acquire knowledge with credences and without
full belief. However, there is even less scholarly commitment to that posi-
tion than that credences and beliefs are distinct psychological states,154 and
because I am assuming that full belief is a necessary condition for knowledge,
there is no route to knowledge without full belief. A popular third position is
called the threshold view.155 On that account, credences turn into full beliefs
once they are connected to a high enough value, a value above a certain
threshold between 0 and 1. The threshold may depend upon contextual mat-
ters, and so there may not exist one invariant absolute value or interval of
values that is the threshold for full belief in every context. However, it is
well-known that this view runs into arbitrariness worries.156 The objection
from the arbitrariness of any imposed threshold whether that threshold is
context dependent is so severe that one leading formal epistemologist, Brian
Weatherson, invokes a threshold view and a vagueness response to the prob-
lem of arbitrariness (i.e., the question of where the precise value is that turns
credences into full belief is vague), and yet remarks that as far as he can tell,
the vagueness response is “the only prima facie plausible response and” it is
at the same time “unsuccessful.”157
The prospects for building a bridge that extends from knowledge or war-
rant (as understood by traditional epistemology) to credences (as under-
stood in Bayesian formal epistemology) are worse than what the preceding
text suggests. There is currently no agreement on how the credences or
partial beliefs of agents relate to the credence functions that are the rep-
resentations of partial beliefs in the formal apparatus of Bayesianism. For
example, leading probabilist James M. Joyce has argued that an appropriate
collection of credence functions represents an individual degree of belief. No
single credence function can do the job (Joyce 2010). Such a view would
complicate any picture of how credences relate to full belief on Bayesianism.
Others have argued against the position on the grounds of belief inertia and
the like (Vallinder forthcoming).
There are yet more problems that unsettle strategies for connecting the
rationality sought in formal Bayesian epistemology with a notion of ratio-
nality important to knowledge or warrant in traditional epistemology.
Knowledge is psychologically real, as are the beliefs and the supports rela-
tions between beliefs in a cognizer’s noetic structure. If Bayesianism is to
provide an accurate model of those states building up those structures (i.e.,
beliefs), then its representations of them and their roles/relations should
not be too psychologically unrealistic. In addition, if there is a robust con-
nection between Bayesian rationality (i.e., the rationality enjoyed by the
cognizer who is probabilistically coherent) and knowledge, then presum-
ably normal human cognizers who acquire a great deal of knowledge to
navigate the world, will sometimes do so, at least in part, because they are
very good at a brand of probabilistic reasoning that is modeled well by
Bayesianism (i.e., updating and shifting around values of credences as the
axioms and theorems of the probability calculus would dictate). Yet, there is
an abundance of evidence from psychology to the contrary. Human persons
are ofttimes very bad at probabilistic reasoning.158 Under certain conditions
(i.e., given certain requests), the evidence suggests that we regularly confuse
the Pr(H/E) with the Pr(E/H) (see the discussion of this in Smithson 2016,
486–487). Our probability estimates about categorical and conditional
probabilities are again and again less than what Bayes theorem would rec-
ommend (Phillips and Edwards 1966). The study of Phillips and Edwards
recommends that such conservatism is “unaffected by prior probabilities,”
remaining “constant as the amount of data” increases (ibid., 346). That we
consistently violate the conjunction rule (i.e., that Pr(H & E) ≤ min(Pr(H),
Pr(E))) in our probabilistic reasoning is affirmed by a general consensus
among psychologists who study probabilistic reasoning (Smithson 2016,
485). And a fortiori (quoting Smithson), “[t]here exist cultures with no
identifiable notion of probability, and some people in Western cultures dis-
avow any connection between probability and mental aspects of uncertainty
such as degrees of belief.”159 These facts constitute a cumulative case for
the thesis that probabilistic reasoning may play no role in a great deal of
instances of knowledge about the world. Thus, even if a belief b reduces to
a credence or partial belief satisfying special conditions of a certain kind,
that b comports with the rules of probability, or that it is updated in the way
classical conditionalization suggests, is not at all necessary for knowledge
(compare Harman’s reasons from complexity in Harman 1986, 27: “. . . the
actual principles we follow do not seem to be of that sort [the kind recom-
mended by Bayesians]”).
Neither is probabilistic coherence sufficient for knowledge. A cognizer C
rests atop a beautiful mountain. A random burst of electromagnetic waves
(or a wave packet) hits C’s brain and causes them (almost magically) to
become probabilistically coherent with respect to all of their credences. In
an instant, all of C’s partial beliefs abide by the probability calculus, and
everything is updated appropriately. C forms a number of credences that
have the value .99 in a way that respects the probability calculus. Even if we
were to assume the threshold view, and even if we could demonstrate that
the threshold in this context is .9, it seems clear that C will fail to come to
know anything as a result of C’s suddenly and randomly becoming probabi-
listically coherent and having C’s credence in a proposition p moved to .99
in the appropriate way. C’s noetic structure is caused to exhibit the features
it does by happenstance in ways that are completely out of C’s control. C
does not know p.
What role is there then for probabilistic reasoning as understood by sub-
jective Bayesianism? I believe there is some, as of yet, undiscovered full
account of conditional epistemic probability that underwrites the epistemic
and theoretical value that is explanatory power in inferences to the best
explanation (IBE). I will try to articulate what I consider to be the best
theory of just how explanatory power relates to epistemic probability, but
because of worries like those I’ve just expressed, my account will ultimately
be incomplete. I do not believe this should force us to be agnostic on a great
many metaphysical matters because the partial account I develop may help
indicate what truly is the best explanation of some facts in metaphysical
inquiry, and not all metaphysical inquiry and not all metaphysicalC system
building transpires on the back of IBE reasoning.
An IBE is an inference-type that is not reducible to inductive or deductive
inference making, although it does constitute a unique class of ampliative
inference that is not fully captured by the deliverances of probability theory
or confirmation theory.160
Metaphysicians who are after a metaphysicalC system ought to employ
abductive inferences to the best explanation in their methodology. They
should do this because physicists have and continue to successfully employ
such inference-types in their systematic investigations of what exists and
the nature of physical reality. That is to say, a strong case can be made for
the thesis that many scientists in the history of science employed inference
to the best explanation, reasoning in such a way that it secured epistemic
justification for beliefs about what there is, and the nature of what there
is.161 I therefore agree with Gilbert Harman (1965), there actually are justi-
fication-providing inferences to the best explanation,162 and these inferences
afford epistemic justification transmission to beliefs in the conclusions of
sufficiently strong versions of such inferences from true premises.
IBEs, in the context of metaphysical investigations (and so I’m not inter-
ested in mathematical explanation), are explanatory arguments. They have
logical forms expressed schematically as,
(1) Fact 1, Fact 2, . . . , Fact N require an explanation (where these facts are
all true propositions).
(2) There are an appropriate class of potential explanations E1, . . . , En of
Fact 1, Fact 2, . . . , Fact N that either are identical to or embedded in
metaphysical theories/hypotheses.
(3) E2 is the best potential explanation of Fact 1, Fact 2, . . . , Fact N.
(4) Therefore, it is credible that E2 explains Fact 1, Fact 2, . . . , Fact N.
Fact 1, . . . , Fact N report on some features of reality (call these results). The
results are the explananda.
What work does the locution ‘it is credible that’ do in (4)? That opera-
tor simply highlights the fact that the inference is not deductive. Premises
(1)–(3) do not entail the truth of (4). Rather, the inference from (1)–(3) to (4)
is risky. Some, such as Lycan (2002, 413), prefer to use the qualifier “prob-
able” or “it is probable that” in the conclusion of abductive arguments,
but I do not want to suggest the idea that ampliative inferences of this kind
reduce completely to probabilistic inferences.
What about the clause ‘appropriate class of potential explanations’?
What is a potential explanation? And what would constitute an appropriate
class of such entities?
(Desiderata for a Potential Explanation): A potential explanation of

Fact N is a meaningful, logically coherent, sincere, purportedly factive,
purportedly conjunction distributive, purportedly understanding pro-
ducing answer (α) to the question, “Why is Fact N true?”, where (1)
α invokes only existing objects, properties, relations, structures, pro-
cesses, and/or laws; (2) α does not invoke Fact N; and (3) α does not
fundamentally depend for its truth on any set of objects, properties,
relations, structures, processes, or laws that include Fact N. In addition,
α is the type of answer to the preceding question for which there is some
evidence for the thesis that for a properly functioning cognizer C who
grasps Fact N and α, α attenuates, reduces, or removes C’s cognitive
discomfort (when it is present), puzzlement, and surprise concerning
Fact N’s truth.163 And necessarily, if the potential explanation of Fact N
that is α explains Fact N, then the resulting explanation relation will be
formally asymmetric.
Having an appropriate class of potential explanations is important. One

does not want to allow for easy abductive inference making in the sense that
one’s range of competing explanations is significantly poor and therefore
qualitatively worse than E2 from the start. The field of competition must be
real competition. What is more, E2 itself cannot be intrinsically implausible.
It must pass over a threshold.164 The threshold, I believe, is one intimately
related to knowledge-why (i.e., knowledge concerning why something has
transpired or why some fact holds). Cogent IBEs provide one with war-
rant for one’s belief that the best explanation successfully and legitimately
reduces cognitive discomfort (when it is present), attenuates surprise, and
removes puzzlement about facts or results.165
Members of the appropriate class of potential explanations satisfy the
desiderata for a potential explanation for at least one member or conjunct
of the conjunction of facts to be explained, in this case, Fact 1 . . . Fact N.
Call facts (Fact 1. . . Fact N), the evidence E. And let our best potential
explanation E2 be a metaphysical thesis or hypothesis H. As I alluded to
previously, what supports premise (3) is an appropriation of a set of epis-
temic values that privilege H over the other potential explanations in the
appropriate set. Here is a list of the values I believe constitute a substantial
heuristic for testing competing hypotheses or potential explanations,
(1V) Explanatory power: First, the best potential explanation H is

incrementally confirmed by E (i.e., Pr(H) < Pr(H/E)). Second, if H has
the most explanatory power out of a set of competing potential expla-
nations, then E favors or better supports H than any other compet-
ing potential explanation, such that the probability of E given H (i.e.,
the Pr(E/H)) is greater in value than the probability of E given any
other competing potential explanation in the appropriate set of poten-
tial explanations (i.e., for any competing potential explanation in the
appropriate set CE, Pr(E/H) > Pr(E/CE)).166
The fact that the best potential explanation (H) among the appropriate set
of competing potential explanations is the most explanatorily powerful
underwrites the claim that the best (potential) explanation reduces a prop-
erly functioning cognizer’s surprise concerning E’s truth.
(2V) Coherence: The best potential explanation in the appropriate set

of potential explanations is one that is consistent with the set of con-
ceptually necessary truths (e.g., <Bachelors are unmarried males.>), the
set of metaphysical laws (e.g., <Necessarily, numbers cannot be prime
ministers.>), and the set of physical laws (e.g., Einstein’s field equations
properly or correctly interpreted).
According to this value, it shall count against a potential explanation or

hypothesis HN that it conflicts with the laws of physics. Employment of this
value should not be interpreted as a commitment to the idea that all meta-
physical hypotheses should be solely about the domain of reality accurately
described by physics. The theistic and metaphysical doctrine that God cre-
ated the cosmos ex nihilo does not violate any laws of physics because the
laws of physics do not quantify over entities like God, or actions like c reation
ex nihilo. The systems the laws are restricted to explaining and describ-
ing do not include supernatural entity-types, or supernatural action-types.
Employment of this value does not delimit the domain of the metaphysi-
cal to the domain of the natural. This is the case for another independent
reason. I’m interested in providing values that help with inferring to the
best explanation. Not all metaphysical hypotheses are discoverable via such
inference-making.
(3V) Simplicity: An indication that H is the best potential explanation in

the appropriate set is that H both (a) invokes less (in terms of amount)
fundamental objects, properties, relations, structures, processes, and/or
laws, and (b) depends for its truth (in the sense employed by the TDB
thesis of sect. 1) on less (in terms of amount) fundamental objects, prop-
erties, relations, structures, processes, or laws than any of the compet-
ing potential explanations in the appropriate set.
Here, I help myself to Jonathan Schaffer’s theory of grounding (Schaffer

2009). A fundamental thing is one that is not grounded, where an entity x
grounds another entity y, just in case, y depends for its positive ontological
status (its existence) and nature on x (where the dependence is asymmet-
ric, transitive, and well-founded). A derivative thing will therefore be some
entity that is grounded.167 This choice value is my way of ensuring that
the best potential explanations are those that don’t aggrandize or generate
fundamental stuff beyond what’s required (leaning on an idea in Schaffer
2015, 644).
(4V) Testability: An indication that H is the best potential explanation

in the appropriate set is that H non-trivially implies the most substan-
tive truths that are testable by the community of practicing scientists.
If a metaphysical hypothesis H that is in the appropriate set of potential

explanations has implications that can be tested by scientists, that is a virtue
of H. If a competing member of the appropriate set CE cannot be so tested,
that is evidence that it is not the best.
(5V) Explanatory Plausibility: An indication that a hypothesis H is the

best potential explanation in the appropriate set is that either it is a
non-trivial material implication of more metaphysical theses that have
consensus status among the community of scholars that are metaphysi-
cians, or it is well-supported by more consensus views of this kind than
any other competing potential explanation.
A scholar who is a metaphysician will be a member of the community of

individuals who has published on a metaphysical topic in a venue (usually
a journal, PhD dissertation, or book publication) that is respected by the
wider community of professional philosophers.
(6V) Theoretical Unification: An indication that H is the best potential

explanation in the appropriate set is that H explains more non-trivial
facts about scientific or metaphysical matters that are not identical to
facts Fact 1 . . . Fact N than the competing potential explanations.
It is a definite plus that one’s potential explanation is an explanation (gener

ally) of a number of other facts that aren’t part of the explananda. One
shouldn’t broaden one’s explanatory repertoire if one can get away with it.
Moreover, one will be put in a better position to unify if one seeks to extend
an explanation as far as it can be plausibly extended.
(7V) Explanatory Scope: An indication of the best potential explana-

tion will, in addition, be that it satisfies the desiderata for a potential
explanation (previously discussed) for all of the explananda (i.e., for the
conjunction of the explananda).
The way we bring all of these values together is by demanding of the best
potential explanation that it satisfy a best balance condition,
(8V) Best Balance: An important indication that H is the best potential

explanation in the appropriate set of potential explanations is that H is
indicated as best by the best balance of indications of the best potential
explanation in the appropriate set given by values (1V)–(7V).
What is the best balance? That is a very tough question. My best conjecture
is that one should weigh the preceding values as follows (from most impor-
tant to least important),
(Weights): (2V) tied (for equal importance or weightiness with) (7V)

> (3V) > (6V) > (1V) > (4V) tied (for equal importance or weightiness
with) (5V).
Evidence for this conjecture can issue forth from the types of verdicts it
yields. My hope is that the present work on causation will utilize IBE in a
way that respects this ordering and that the conclusions reached by my use
of IBE in the relevant way show off some type of prima facie plausibility of
the inference-type so applied.168
Abductive inferences that respect the dictums of this section are cogent
abductive inferences.

of causation, one should use cogent inferences to the best explanation
when appropriate.
Section 4.5.4: Other Non-Basic Sources

There are non-basic sources of justification. These are sources that, to be
utilized appropriately, require positive dependence upon basic sources.
Metaphysicians receive a great deal of justification about what exists and
the nature of reality from the non-basic source of knowledge that is testi-
mony (see Coady 1992). Knowledge carriers that are reliable testifiers bear
witness to various metaphysical truths. One of the most important groups
of such testifiers is scientists. It is entirely appropriate for the metaphysi-
cian to justify metaphysical theses by supporting them with the content of
testimony from proper scientific authorities. If you want to learn about the
nature of stars, or galaxies, or space-time, you will do well to glean valuable
knowledge of such entities from physicists.
Some philosophers may charge me with licensing the informal fallacy that
is argumentum ad verecundiam. But I believe this fallacy is often misun-
derstood. Standard critical thinking texts present the fallacy as one that is
committed when one makes use of an inappropriate authority’s testimony as
support for a hypothesis. Expert testimony about an issue that falls within
the purview of the expert’s field of study can and often does constitute epis-
temic justification for belief in the content of the testimony or that about
which the testimony bears witness.
I will take no particular stand on the many views about how it is, precisely,
that testimony yields knowledge or epistemic justification.169 I encourage
only a commitment to the following general principle irrespective of how
testimony is more precisely cashed out,

of causation, one should appropriate testimony as a non-basic source
of justification.
Section 4.5.5: The Propositional View of Physical Theory Structure

One of the most important non-basic avenues to epistemic justification in
metaphysical inquiry and metaphysicalC system building is that path that is
provided by science. So I affirm D17,

of causation, one should appropriate the non-basic sources of justifica-
tion provided by the sciences (mainly, but not solely, properly interpreted
physics or the deliverances of our best physical theories both properly
interpreted and endowed with the type of structure I’m about to detail).
Most of the scientific equipment and building materials I will need for
my metaphysicalC system will come from physics and physical inquiry,
although I will make use of conclusions in higher-level science as well. But
because metaphysical and physical inquiry are intertwined in ways that are
very complicated, and because many contemporary philosophers of physics
believe that physics ought to play no role in metaphysicalC system building
(because causation is absent from physics), I will spend some time in this
subsection, detailing both (a) how physics and metaphysics interact and (b)
how the deliverances of physics can be used in one’s metaphysicalC system
by articulating a new (the propositional) view of physical theory structure.
The resulting type of theorizing that both properly informs and is properly
informed by physical inquiry, I call fundamental metaphysical theorizing.
The theory of causation defended in this work provides both a philosophical
analysis of the causal relation and a fundamental metaphysical theory of its
properties and place in the world.
SECTION 4.5.5.1: FORMALISM
Theories of physics have formulations. For example, the general theory

of relativity has ADM and tetrad formulations.170 Classical mechanics has
Newtonian, axiomatic, Lagrangian, Hamiltonian, and geometric formula-
tions.171 Copenhagen non-relativistic quantum mechanics has both wave (or
wave function, or Schrödinger) and matrix (or Heisenberg) formulations.172
These formulations include various equations with auxiliary principles
(e.g., the specification of the Hamiltonian in wave mechanics, or specifica-
tions of energy conditions in GTR), or added terms needed for perturbative
methods and their exact, or numerical (or approximate), solutions. Some
formulations include postulates that are not stated in the language of mathe-
matics. Appropriate conjunctions of physical postulates, auxiliary principles,
and equations together with their substantive and explanatory deductive con-
sequences constitute formulations or formalisms of physical theories. With
some other equipment that I will soon discuss, formulations/formalisms are
used by physicists to describe and explain the kinematics and dynamics of
actual, and sometimes merely possible, physical systems.
The logical positivists espoused the so-called “Received View” (also
called the syntactic view) of physical theory structure.173 On the Received
View, theories are “partially interpreted calculi” and are therefore linguistic
structures.174 These structures have three parts: theoretical postulates (the
axioms), abstract formalisms (the calculi), and correspondence rules.175 The
calculus or formalism of a theory T is constructed out of a set of transforma-
tion and formation rules of first-order logic, plus a signature. That formal-
ism has both logical and nonlogical terms as constituents. The logical terms
come from first-order calculus with (when necessary) some extra bits of
machinery.176 The nonlogical terms are separated into two distinct cate-
gories: (i) nonlogical observational terms, and (ii) nonlogical theoretical
terms. Both the correspondence rules and theoretical postulates consti-
tute the partial interpretation of the formalism of T (French, Structure
of Theories 2014). Correspondence rules connect (i) and (ii). The partial
interpretation of T allows (quoting French’s summary, without endorse-
ment of the view),
for the derivation of certain sentences containing observation terms

from certain sentences containing theoretical ones. The interpretation is
partial because the theoretical terms are not explicitly defined and there
is room, as it were, for the addition of further correspondence rules as
science advances.177
Some of the axioms of T are its basic laws. Those basic laws are universally
quantified conditionals that are metaphysically contingent truths although
nomologically necessary and exceptionless.
I will refrain from endorsing the Received View’s characterization of the
formalisms of physical theories. This is because extremely successful physical
theories, like the general theory of relativity, require more than first-order
logic for axiom-stating purposes. The topological structures needed in general
relativity stand in need of second-order quantification.178 Moreover, “there is
no set . . . of first-order axioms such that the models of” that set “are precisely
the Hilbert spaces” of quantum mechanics (Halvorson 2014, 594).
The semantic view of theory structure gets the formalisms of physical
theories wrong as well. For according to that view, physical theories have
two fundamental parts: theoretical definitions and theoretical hypotheses.
The former scientifically represent, whereas the latter indicate what is repre-
sented (ibid., 592). Formalisms are given by theoretical definitions that are
themselves understood in terms of abstract mathematical objects (models,
or state-space models) as explicated by model theory.179 These models are
provided by solutions to the equations of physical theories.180 Interpreta-
tions of formalisms consist of detailing the ways in which models fit or are
isomorphic to the world.181 And therein lies several problems. First, some
parts of the formalisms of physical theories are not distinctively mathemati-
cal such that they can be rightly called mathematical structures or parts
thereof. Everyone agrees that Newtonian mechanics, as explicated by Sir
Isaac Newton (1643–1727) in the Principia (Newton 1999), is a physical
theory. However, Newton’s law of inertia is not and cannot be stated with
any extant formal (and purely) mathematical language (amidst the New-
tonian formulation) and is therefore not properly recognizable as part of a
mathematical structure.182 That law states that, for any body b, if b is free,
then it will either continue (absolutely) moving in a uniform manner involv-
ing its uniform velocity in a straight line or else it will remain at absolute
rest. This law is not equivalent to an instance of the second law of motion in
the special case in which F = 0 (contra Ter Harr 1971, 2). One has need of
the law of inertia to help define the very entity relative to which the second
law of motion holds, viz., an inertial frame (agreeing, in part, with Brown
2005, 37, n. 9).
Like the law of inertia, some foundational principles in principle-based
physical theories (e.g., some formulations of thermodynamics and special
relativity) include in their formalisms underlying postulates or principles that
cannot be expressed in an extant (strictly) mathematical language either. For
example, Einstein’s statement of the restricted principle of relativity (a part of
the original promulgation of special relativity) in his “Zur Elektrodynamik
bewegter Körper” (Einstein, Electrodynamics 1905) was put this way,
“[f]or every reference system in which the laws of mechanics are valid, the
laws of electrodynamics and optics are also valid.”183 There is no purely
mathematical statement of this restricted principle of relativity. For a second
example, consider Einstein’s postulate of light. It says that (relative to some
beholder of the relevant light) light’s speed c is constant and finite in a vac-
uum irrespective of the speed of the source of that light.184 Like the restricted
principle of relativity, the postulate of light’s content is not exhausted by
any one mathematical expression. This explains why presentations of these
principles in standard physics texts on special relativity never dress them in
mathematical garb (see Helliwell 2010, 29; Jackson 1999, 517–518; T aylor,
Zafiratos, and Dubson 2015, 12–13, inter alios). No one would doubt that
the aforementioned principles are part of (or a popular formulation of) the
special theory of relativity.185 Einstein understood them that way, and it
is from these principles that one can derive the Lorentz transformations
that are central to STR (Einstein 1952, 111); cf. the discussion in Lange
(2017, 100ff). Formalisms of physical theories should not be restricted to
that which can be expressed in the language of mathematics.
Second, quantum chromodynamics (QCD) is our best physical theory of
the strong interaction. To work with manageable calculations for the pur-
poses of finding approximate descriptions of certain systems in the low
momentum (low energy and large coupling constant) regime of QCD, sci-
entists appropriate non-perturbative lattice gauge QCD (Banks 2008, 108;
Gupta 1998). According to that theory/method, space-time is discrete. How-
ever, perturbative QCD in the exceedingly high momentum (high energy and
small coupling constant) regime yields descriptions of systems evolving on a
continuous space-time (Sterman et al. 1995). Thus, some models falling out
of solutions attained in lattice gauge theoretic QCD will suggest representa-
tions of the world that are incompatible with some other models that fall out
of solutions attained in the context of a perturbative QCD that assumes a
continuous space-time.186 If, as it is commonly assumed in the semantic tradi-
tion (e.g., van Fraassen 1989, with Savage 1990, vii–viii, reading semantic
proponents similarly), physical theories are sets of their models, it will fol-
low that QCD is an inconsistent theory. Perhaps the way out is to say that
approaches to the strong interaction like lattice gauge theoretic QCD and
perturbative QCD are all different physical theories because they appear to
require inconsistent assumptions. However, if we are forced to make that dis-
tinction in the context of studying QCD, we will be forced to make it in the
context of studying a more general quantum field theory (QFT) as well (QCD
is, after all, a quantum field theory). But non-perturbative QFT, if considered
as a theory in isolation from perturbative QFT, is significantly explanatorily
impotent. Consider the following historical episode. In 1927, P.A.M. Dirac
(1902–1984) thought that his quantum field theory was incomplete because
[r]adiative processes of the more general type considered by Einstein

and Ehrenfest, in which more than one light-quantum take part simul-
taneously [Dirac is thinking of absorptions and emissions involving lots
of photons], are not allowed on the present [quantum field] theory.187
Dirac would soon change his mind. He would later judge in the same year that
his quantum field theory could provide an approximate description of such
complicated systems if he used second-order perturbation theory.188 Without
perturbative methods, the more complicated systems were not in reach.
Proponents of the semantic tradition could argue that models of non-
perturbative lattice QCD and models of perturbative QCD should be under-
stood as representations of distinct possible worlds, much like solutions to
Einstein’s field equations represent varying worlds with alternative space-
time structures. This response is problematic for those semantic theorists
who, like van Fraassen, maintain that the Hauptsatz of scientific representa-
tion is the thesis that (quoting van Fraassen) “[t]here is no representation
except in the sense that some things are used, made, or taken, to represent
some things as thus or so.”189 Scientists use lattice gauge theory and per-
turbative QCD to represent two regimes of one and the same world (i.e.,
the actual world). Those who use lattice formulations of QCD, do so to
approximate actual hadronic physics (Richards 2000). Likewise, perturba-
tive QCD is used to approximate some actual strongly interacting physical
systems. This suggests that the representations they use are not models of
differing worlds, but one and the same world. Even if we could get away
with connecting lattice gauge theoretic approaches to QCD to worlds that
differ from those one connects to perturbative QCD, the resulting theory
of actual strong interactions would be significantly impoverished by such
a maneuver. The low (or high) momentum regime would become overly
difficult if not impossible to approximate. It would be a plus if we could
avoid delimiting the explanatory scope of QCD in this way. My approach to
physical theory structure facilitates such avoidance. Here is how it does so.
Formulations of physical theories are collections of propositions express-
ible by a multitude of varying (at least partially interpreted) coherent and
meaning-bearing linguistic devices, such as declarative sentences and the like,
called designative formulations that can be used to perform communicative
acts. Designative formulations must (at least) be partially interpreted if they
are to express propositions. An expression with the following symbols F = ma
does not carry the substantive meaning ordinarily associated with it if there
is no representation relation between the symbols or representation tokens
like ‘F’ and (at least) the entity-type that is force.
We escape the preceding problem involving lattice QCD and perturbative
QCD by requiring that designative formulations of physical theories some-
times incorporate complex ‘would’-counterfactual conditional claims.190
For example, with respect to lattice gauge theoretic approaches to QCD
(and here I assume that the actual world is endowed with a space-time that
is discrete), a designative formulation of QCD might include a subjunctive
conditional of the form,
(1) If it were the case that (a) space-time were continuous, and p, and q,
and, . . . then it would be the case that (b) entities E1–En (which pertur-
bative QCD for the high momentum regime studies or the appropriate
limit) in the very high momentum regime R behave y-ly.
Support for (1) will come by way of empirical/experimental support for
perturbative QCD. We save the consistency of the designative formulation
(and so also the formulation; see below) of QCD by precluding from it
the conjunct that our space-time is continuous. Instead, we allow the (non-
schematic version, which I will call (1*)) subjunctive conditional into the
designative formulation and await empirical evidence for it. The predictive
power of QCD is salvaged because (1*) sets down a rule whereby one can
anticipate the behaviors of the objects of study. Look to the world and see
if it comes close to satisfying (a). When it does, we can hypothesize that the
relevant objects will behave y-ly. This is rule-governed prediction.191 We
can make such predictions without asserting the truth of (a), and so we can
have the full predictive power of QCD without maintaining that the theory
is inconsistent.
The strategy just adopted for finessing would-be inconsistencies will help
with the problem of idealizations in physical theories as well. Laws of phys-
ics are sometimes about object-types that are highly idealized (e.g., point
masses, center masses, and frictionless surfaces). These laws might be best
understood as counterfactual conditionals. At nearby worlds with similar
histories and the same or similar laws as ours, point masses and the like
dance as the laws dictate. The approximate truth of these laws is captured
by the sense in which ‘would’-counterfactual conditionals are true tout
court. We rescue predictive success, and therefore also some of the empirical
success of the theory, by noting how the theory provides rules for hypothesis
formulation, where the hypotheses in question are tested and found to be
approximately correct.
At least one condition for formulation expression by a designative formu-
lation is that the user of the designative formulation employ some under-
standing of referential and representation relations between representation
tokens that are members of the designative formulation and those objects,
properties, relations, processes, structures, etc. (including their types), that
the representation tokens are about. In other words, and to repeat, a desig-
native formulation does not express a formulation unless that designative
formulation is partially interpreted. The correct partial interpretation of a
designative formulation will be the correct partial interpretation of the for-
mulation it expresses. This is because the correct partial interpretation of
the designative formulation will connect not just structured elements of the
designative formulation to the world, but also those structured elements
belonging to the formulation (a distinctive set of propositions), which are
expressed by elements in the designative formulation.
I will say more about partial interpretations in sect. 4.5.5.2. For now,
I note that my account rescues the sense in which formulations of physical
theories can be understood as language-independent entities. This is because
the propositions those formulations are identical with are language-inde-
pendent entities. The sense in which formulations of physical theories are
mind-independently true is captured well by the sense in which propositions
are mind-independently true. The sense in which (partially interpreted)
formulations of physical theories can be believed is captured well by the
sense in which propositions can be believed (i.e., the sense in which propo-
sitions can serve as the contents of beliefs). The sense in which (partially
interpreted) formulations of physical theories represent is captured well by
the sense in which propositions represent. What makes representations sci-
entific or physical depends upon the types of entities, structures, processes,
etc. (including their types), that are represented by the relevant propositions.
If some collection of propositions were about the legislative powers of the
U.S. Supreme Court, those propositions would not scientifically represent
anything. If, however, they were (truthfully) about atoms, or molecules, or
gases (and/or their types), they would scientifically represent. Truth-evaluable
sentences belonging to various languages can scientifically represent, but
only in derivative fashion. They inherit their content and representational
qualities from the propositions they express.
The question of whether two designative formulations have the same for-
mulation can be determined by giving attention to the correct answer to
the question of whether those designative formulations express one and the
same formulation or formalism. If they do, then they are formally equivalent.
To help illustrate the distinctions I am introducing, consider a statement
of the designative formulation of Newtonian mechanics in English with
some expressions of that designative formulation belonging to the language
of mathematics (call this DF1). One might have a statement of the law of
inertia in English, and then a statement of Newton’s second law of motion
in terms of the mathematical-physical expression F = ma, with the adden-
dum (in English) that that expression holds relative to an inertial frame, and
so on. One might also have a designative formulation DF2 of Newtonian
mechanics in German with some aspects of that formulation belonging to
the language of mathematics. There might also exist a third formulation
DF3 that is expressed by a designative formulation dressed in the garb of
the formal object language that is first-order logic perhaps supplemented
by set theory and some bits of mathematics. If these designative formula-
tions express the same conjunction of propositions, then they are all for-
mally equivalent.192 I take it that it is a desideratum of proper theorizing
about propositions that propositions be the sorts of entities that can be
expressed by synonymous meaning bearers.193
SECTION 4.5.5.2: PARTIAL INTERPRETATION
Propositions have truth conditions. The formulation F of a physical theory

T will therefore also have truth-conditions. What truth-conditions F has
depends (at least in part) upon the representational properties of the propo-
sitions that constitute F. Those representational properties will be in some
sense connected to the ontology of the partial interpretation of F. The partial
interpretation of a true or approximately true F (i.e., a PIF) is given by a
specification of the indispensable types and tokens of objects, properties,
relations, processes, and structures that (i) F directly corresponds to and that
(ii) F (and its respective designative formulation) is about. With respect to
(i) and (ii), we may say that the relevant indispensable actual token objects,
properties, relations, processes, and structures that fall under the respective
object, property, relation, process, and structure types, plus the indispens-
able types themselves, constitute the ontology of the partial interpretation.
The actual token objects, etc., are indirectly represented by the representa-
tion tokens for the respective types. Thus, if one’s goal is to specify the ontol-
ogy of a PIF, it is often good enough to specify the types in that ontology.
The indispensable notions used in the partially interpreted designative for-
mulation that represent the various indispensable members of the ontology
of the partial interpretation of F compose one part of T’s ideology (contrast
this with the notion of a primitive ideology introduced in chapter 8: sect. 3).
The aboutness relation referenced in conjunct (ii) has a nature that is best
explicated by appeal to examples. We say that the statement, ‘N.T. Wright
is a New Testament scholar.’ is about N.T. Wright. The statement, ‘Bowl-
ing is a difficult sport.’, is about bowling, and that ‘Electrons have negative
charge.’, is about electrons. Unfortunately, I cannot move beyond ostensive
definitions of aboutness, but I believe the idea is clear enough.
The direct correspondence relation posited by option (i) can be under-
stood as none other than the correspondence relation required by the corre-
spondence theory of truth.194 Or that relation could be a congruence relation
between propositions and facets of reality such as factsΣ.195 Whatever its pre-
cise nature, I will assume that correspondence is minimally that relation that
obtains between propositions bearing the property being true and facets of
reality they are about, such that propositions bearing truth do so by virtue
of standing in that distinctive relation to reality.196
The partial interpretation of a false but metaphysically possible formu-
lation is given by specifications of the types and metaphysically possible
tokens of objects, properties, relations, processes, and structures (i) that F
would directly correspond to were F true, and (ii) that F is about.197 I will
call false but metaphysically possibly true formulations of physical theories
theoretical or conjectural formulations. Specifications of the kind that are
described at the start of this paragraph give that which constitutes partial
interpretations of conjectural or theoretical formulations (call these PICs).
Because of the counterfactual and/or modal nature of PICs, and because sets
of equations have varying solutions, the relevant galaxies of nearest pos-
sible worlds used to evaluate the relevant counterfactual conditionals will
often include varying collections of objects, properties, processes, etc., that
constitute the ontologies of the various PICs from nearby world to nearby
world. For example, the de Sitter or vacuum solution to Einstein’s field equa-
tions, plus those equations and all of the other consistent principles and
equations commonly said to be part of the standard formulation of GTR (see
Wald 1984), have a PIC whose ontology is significantly slimmer than the ontol-
ogy of GTR’s standard formulation with the Friedmann–Robertson–Walker
solution or simplification instead. The latter ontology (of the PIF) includes
an abundance of various kinds of matter, whereas the former ontology (of
the PIC) does not.
The formulation of every physical theory has at least a PIC. This entails
that there are no metaphysically impossible formulations because PICs are
false but metaphysically possible formulations. You have done something
unscientific if you have proffered a would-be “formulation” of a physical
theory that implies a contradiction, or metaphysical absurdity.
If two formulations have the same truth conditions (and this is a necessary
truth), then so will their respective designative formulations. However, just
because necessarily, two formulations have the same truth conditions does
not mean that they will have the same partial interpretation. The fact that
two formulations are truth-functionally equivalent is not enough to gener-
ate shared PIFs or PICs. This is because two conjunctions that are truth-
functionally equivalent may have differing metaphysical explanations or
dependency bases, and may therefore be about or correspond to two distinct
sets or collections of tokens and types of objects, processes, structures, etc.
PIFs will include partially interpreted laws of nature. On the proposi-
tional account of theory structure, laws of nature (in general) are,198
• Exceptionless
• Naturally necessary
• Truths (or approximate truths)
• Propositions
• Supportive of counterfactuals
• Essential ingredients in scientific explanations of metaphysically contin-
gent regularities
• (If derivative), substantive logical consequences of other laws given that
those consequences themselves satisfy the conditions above
I will commit to nothing more than the above picture of laws. Any further
metaphysical specification of their nature or function that is consistent with
the above is also consistent with the propositional view of theory structure
I am seeking to motivate in this chapter.
The partial interpretation of laws of nature will require a metaphysics of
physical quantities, magnitudes, scales, and dimensions. Metaphysics enters
physics here. In what precise way laws of nature scientifically represent is
determined by the precise metaphysics of those quantities featured in the laws.
For example, if the metaphysics of the gravitational field and its interactions
suggests that such an entity is a causal one, then laws like Einstein’s field equa-
tions will be best (at least) partially interpreted as causal laws that causally rep-
resent the actual world (q.v., chapter 3 and chapter 8 for more on this motif).
SECTION 4.5.5.3: FULL THEORY INTERPRETATION
Physical theories have full interpretations. A full interpretation of (genitive

of possession) a physical theory T is a part of T that is given by a specification
of (a) the partial interpretation of T’s formulation, (b) the ontology of T,
(c) a specification of the complete (in the appropriate limits, scales, and
domains of the physical world) kinematics and dynamics of T, together with
(d) the initial conditions and (e) the background conditions employed by the
contents of T’s kinematics and dynamics so as to ensure approximate accu-
racy (perhaps from the perspective of the target world considered as actual).
The ideology of T’s partial interpretation, along with the indispensable
notions used to represent the various members of the ontology of the full
interpretation of T (without redundancy) compose T’s complete ideology.
The ontology of a physical theory, or the ontology of the full interpreta-
tion of a physical theory, includes (b-i) those entities (including relations,
processes, structures, and objects) that are indispensable to the ontology of
its partial interpretation (this includes entity-types), (b-ii) those entities and
entity-types that are indispensable to (c)–(e), plus either (b-iii) those entities
(including entity-types) upon which the theory directly depends for its truth
or (b-iv) those entities (including entity-types) that are required to directly
make it true. Getting clear on the deep natures of the stuff in a physical
theory’s ontology involves metaphysical theorizing. Moreover, the correct
articulation of the ways in which theories depend for their truth on stuff in
reality will be arrived at on the tail end of metaphysical inquiry. Metaphys-
ics enters physics here.
The involved direct truthmaking in (b-iv) can be understood as the kind
of truthmaking at the heart of the work of various truthmaker theorists.199
For example, one might espouse the view that some collection of entities C
makes a collection of truth-bearers F true, just in case, C is that by virtue of
which F is true.200 Or as in Lowe (2006, 203), one might argue that C makes
F true, just in case, the essences (or parts thereof) of the propositions that are
constituents of F are such that, given the existence of C, F is true. One might
also maintain that the existence and intrinsic nature of some collection of
entities C serves as a proper truthmaker in the sense that for any conjunct
of F, p, the existence of some constituents of C and their corresponding
intrinsic natures de re necessitates the truth of p.201 I will also leave open the
possibility of defining truthmaking in terms of some other metaphysically
distinctive relation that backs metaphysical explanation. And here I have in
mind theories of grounding.202, 203
The dependence relation I have in mind with respect to option (b-iii) is
something milder than truthmaking or grounding. It is the same type of
dependence I appealed to in my statement of the TDB thesis in sect. 1.
SECTION 4.5.5.4: ACAUSAL REPRESENTATION
We can illustrate (d)–(e) well by discussing two ways in which physical theo-
ries scientifically represent. For example, there are acausal representational
properties of those parts of formulations that are about correlations and
functional dependencies. Robert Boyle’s (1627–1691) ideal gas law rep-
resents a relation of functional dependence and/or a correlation relation
between the values of properties of an ideal gas and the state of that gas.
More specifically, the standard formulation of Boyle’s law is given by the
mathematical expression,
(Boyle’s Ideal Gas Law (BL)): PV = nRunivT [where Runiv is the universal
gas constant, T is absolute temperature, n is the mole number, P is pres-
sure, and V is the volume of the ideal gas204]
To appreciate more precisely what BL represents, one should envision a

case in which an appropriate ideal gas with a constant (appropriate) tem-
perature, mole number, and low density is confined to a container with a
piston over its top. If the pressure of the gas is 1 kpa and the initial involved
volume is 4 m3, then if one increases the pressure to 1.33 kpa, then the
volume will decrease to 3 m3. Likewise, if one increases the volume back
to 4 m3, the pressure of the gas will return to 1 kpa. As the content of BL
suggests, the product of the gas’s volume and pressure will always equal
nRunivT. And so here we have a functional dependency relation between the
pressure and volume of an ideal gas. The equation appears to remain silent
about causes and effects (i.e., what generates the increase in pressure or vol-
ume). The representation therefore appears to be acausal.205
Notice that in order for BL to provide the non-causal scientific rep-
resentation just adumbrated, we must assume that the type of gas it is
about has a low density, and that it is confined in some manner. These
assumptions appear nowhere in the formalism and must be added as
interpretational postulates to the theory that includes BL and thereby
describes ideal gases.206 These postulates constitute specifications of back-
ground conditions. They are important facets of the full interpretation of
thermodynamics.
Physical theories have need of initial conditions. For example, a realistic
Boltzmannian statistical mechanics of the cosmos (understood as a statisti-
cal mechanical system) has need of the past hypothesis, that is, the thesis
that the universe began in a low entropy state (see the arguments for this in
Albert 2000). That initial condition says nothing about what causes what.
When theories represent the world or a physical system by imputing to its
history some initial condition, it represents, albeit, acausally.
SECTION 4.5.5.5: CAUSAL REPRESENTATION AND CAUSATION HUNTING

IN THE HISTORY OF PHYSICS
Most contemporary philosophers of physics believe that causation reduces

to law-governed physical history, where both the laws and physical history
are non-causal. This is causal reductionism. It is therefore not surprising
that most also maintain that (i) our best physical theories do not causally
represent and that (ii) physicists are not (and should not be) in the business
of hunting causes. As Jonathan Schaffer put the thought,
causation disappears from sophisticated physics. What one finds instead
are differential equations (mathematical formulae expressing laws of
temporal evolution). These equations make no mention of causation.
Of course, scientists may continue to speak in causal terms when popu-
larizing their results, but the results themselves—the serious business of
science—are generated independently.207
Michael Redhead noted,
physicists long ago gave up the notion of cause as being of any par-
ticular interest! In physics, the explanatory laws are laws of functional
dependence.208
Christopher Hitchcock, summarizing what he calls principle P*, remarked

similarly,
there are advanced stages in the study of certain phenomena when it

becomes appropriate to eliminate causal talk in favor of mathematical
relationships (or other more precise characterizations).209
Likewise, J.T. Ismael opines that the dynamical laws of physics give func-
tional dependency relations and do not contain within them the neces-
sary directionality indicative of causation (Ismael 2016, 113).210 I will call
this line of thought (i.e., that causation can be eliminated from physics)
neo-Russellianism.211
Why is neo-Russellianism so prevalent in philosophy of physics? I believe
the reason is connected to a tendency in that sub-discipline to associate
the substantial content of physical theories with the mathematical formal-
isms of those theories (or perhaps the associated models). Hence Schaffer’s
activity of collapsing “the results . . . the serious business of science” into
the “differential equations . . . mathematical formulae.” And hence Hitch-
cock’s demand that one abandon causal talk for “mathematical relation-
ships” when seeking true descriptions of physical phenomena. The thought,
very much in the spirit of (Russell 1912–1913), seems to be that because
formalisms do not contain any causal notions (or because the associated
models are structures that do not contain causal parts), and because those
formalisms (or models) constitute the substantial content of theories, physi-
cal theories should not be understood causally.
The neo-Russellian approach is flawed, given the robust scientific realism
defined below and the metaphysical Realism defined above. Consider the
following principles,
(1) In the context of truth-aimed physical inquiry and theorizing, mathemat-

ical relationships between physical quantities should be posited for the
purposes of accurately describing and/or explaining natural phenomena.
With (i), the realist neo-Russellian should agree.
(2) The actual objects, relations, and processes represented by mathemati-

cal representation tokens in designative formulations of physical theo-
ries have natures. They are textured.
Again, the neo-Russellian should not resist.
(3) The designative formulations (and formulations simpliciter) of physical

theories should be fully interpreted in a manner that is sensitive to both
the natures of the things the formulations are about and the way those
things are textured.
If formulations, whether designative or not, fail to exhibit the type of sen-

sitivity (3) is about, accuracy will be delimited and truth-aimed physical
inquiry stifled.
Let F be the designative formulation of an approximately true physical theory
arrived at on the basis of the type of truth-aimed physical inquiry (1) is about.
Let F exhibit the sensitivity (3) is about, and add that F fails to incorporate the
notion of causation. Given (1)–(3) above, it does not follow from the mere omis-
sion of the notion of causation in the mathematical parts of F that F ought not
be interpreted, whether partially or fully, in some causal manner. Consider what
Christopher Pincock describes as “an uncontroversial instance of a causal repre-
sentation” (Pincock 2012, 48). The instance in question features mathematical
equations used to represent a simple traffic system of vehicles (with no collisions)
that travel on a single-lane highway. The designative formalism provided consists
of three equations for the case in which i is greater than 1 but less than or equal
to the number of vehicles traveling on a single-lane highway (or N),
x i (t ) − x i − t (t )
i (t+τ ) = λ
(Eq. 3): x
xi (t ) − xi −1 (t )
x i (t ) − x i − t (t ) d
(Eq. 3.5): If xi (t ) < xi −1 (t ) , then = ln xi (t ) − xi −1 (t )
xi (t ) − xi −1 (t ) dt
(which can be false and yet Eq. 4 still causally represents accurately).
(Eq. 4): x i ( t + τ ) = λ ln xi ( t ) − xi −1 ( t ) + ai (all three equations are from
ibid., 48–49)
A
where xi gives the position of the ith vehicle, a is a constant, t is time, λ =
,
m
A is a constant, and m is vehicle mass. Pincock interprets τ in such a way
that it represents “intuitively the time between when the driver sees a change
on the road ahead and” when the driver “begins her accelerating/braking
maneuver” (ibid., 48).212 He notes that after detailing the values of τ, λ, and
ai, one can use Eq. 3 and Eq. 4 to represent a traffic system involving deter-
minate trajectories of vehicles (ibid., 49).
Notice that no piece of the involved formalism includes the notion of cau-
sation, and yet only the staunchest causal eliminativist would disagree with
Pincock’s judgment that we have here, particularly in Eq. 4, “a dynamic
representation” that includes “information about how a given system will
evolve over time as a result of causal interactions between . . . constituents”
(ibid., 49). The mathematical formalism should be regarded as formalism
that causally represents because the formalism’s partial or full interpretation
is causal. The partial or full interpretation is causal because the mathemati-
cal formalism is about, or corresponds to, or depends for its truth upon or is
made true by the evolution of traffic, a causal phenomenon involving causal
interactions. It is a sensitivity to the nature of traffic systems (which are obvi-
ously causal) that leads us to causally interpret equations like Eq. 4 in the
manner Pincock interprets them. The absence of the literal notion of causa-
tion in the mathematical portions of the designative formalism is irrelevant.
The above view of the relationship between metaphysics (more specifically
the natures of things) and physics (specifically the partial and full interpreta-
tions of physical theories) is called phenomenon-first natural philosophy.
Phenomenon-first natural philosophy assumes a robust scientific realism
according to which many of the indispensable unobservable properties,
relations, processes, entities, and structures (including respective types)
embedded in our most successful physical theories of certain appropriate
physical domains (e.g., general relativity for space-time structure) enjoy
mind- independent existence, and most of these same physical theories
accurately predicate to those properties, processes, entities, etc.213 I cannot
defend robust scientific realism here. I merely assume it.
Neither the idea that (i) physical theories are more than their mathemati-
cal formalisms, nor the idea that (ii) formulations of physical theories receive
their most informed interpretations by first giving attention to the nature of
the phenomena that is the object of representation in each case, are new.
Many of the best physicists adhered to something close to them. In support
of (ii), Henri Poincaré (1854–1912) stated,
It may be asked why, in physical sciences, generalization so readily takes

the mathematical form. The reason is now easy to see. It is not only
because we have numerical laws to express; it is because the observable
phenomenon is due to the superposition of a great number of elemen-
tary phenomena all alike. Thus quite naturally are introduced differential
equations.214
Here, Poincaré reveals that the way the world is explains why mathematical–
physical attempts to describe it take the forms they do. Consider now
esteemed historian Olivier Darrigol’s description of James Clerk Maxwell’s
approach, which very clearly supports both (i) and (ii),
the equations were always subordinated to the physical picture. He

[Maxwell] sought consistency, completeness, and simplicity in the pic-
ture, not necessarily in the equations. The latter were symbolic tran-
scriptions of partial aspects of the picture, and therefore could not be
safely used without keeping the underlying picture in mind.215
Of some relevance to (ii), I note the report of Werner Heisenberg (1901–

1976) on how Niels Bohr’s (1885–1962),
insight into the structure of the [quantum] theory was not a result of
mathematical analysis of the basic assumptions, but rather of intense
occupation with the actual phenomena, such that it was possible for
him to sense the relationship intuitively, rather than derive them for-
mally. . . . Bohr was primarily a philosopher, not a physicist . . . I noticed
that mathematical clarity had in itself no virtue for Bohr. He feared that
the formal mathematical structure would obscure the physical core of
the problem, and in any case he was convinced that a complete physical
explanation should absolutely precede the mathematical formulation.216
Noted historian of science Peter Galison suggests that Einstein was a sup-
porter of (ii). He remarked,
He aimed for a depth between phenomena and the theory that underlay
them. Like Poincaré, Einstein believed that laws must be simple, not for
our convenience but because (as Einstein put it) ‘nature is the realiza-
tion of the simplest conceivable mathematical ideas.’ The form of the
theory therefore had to exhibit in its detailed form the reality of the
phenomena: ‘In a certain sense,’ Einstein later insisted, ‘I hold it true
that pure thought can grasp reality, as the ancients dreamed.’ Einstein
believed that a proper theory would match the phenomena in auster-
ity . . . Einstein insisted over and over that, insofar as they could, scien-
tists fashioned theories that seized some bit of the underlying, simple,
and harmonious natural order. Since Einstein believed that the phenom-
ena did not distinguish true from apparent time, neither, he insisted,
should the theory.217
Indeed, the phenomenon-first approach may explain important empirical

evidence from psychology regarding how cognizers acquire knowledge of
the world. Here is psychologist Steven Sloman on the relevant data,
[M]any studies have shown that students use general knowledge about
the way the world works, sometimes causal knowledge, to solve word
problems in physics and mathematics. Instead of thinking about prob-
lems in a purely abstract way, manipulating symbols until arriving at
the correct answer, people solve problems guided by an understanding
of the situation. If the situation is causal, then we use a causal model.218
The current project will attempt to show that, (i) with respect to the ideol-
ogy of the best full interpretations of our most successful physical theories,
we should employ causal notions because (ii) the ontologies of those inter-
pretations incorporate causal relations, and (iii) claim (ii) should be believed
precisely because much of the phenomena being represented in physical
inquiry are causal phenomena.
There exists strong precedent in the history of physics for a causal
approach to physical inquiry. That is to say, a great many of our very best
physicists either explicitly interpreted physics causally or else hunted for
causal structure in the world. In addition, many of the same giants of phys-
ics sought causal explanations of physical phenomena.
Prior to Galileo, most of those invested in uncovering scientific explana-
tions and descriptions of the natural world sought causes of observed states
of affairs such as the motions of celestial and terrestrial bodies. Not a few
historians and philosophers of science attest to this. For example, Julian
Barbour remarked, “[b]efore Galileo, the theory of terrestrial motions had
been dominated by the concept of cause.”219 Mathias Frisch went further
when he asserted that “a conception of forces as causes of motion appears to
have been widely endorsed up until the middle of the nineteenth century.”220
I agree with Frisch. Causal approaches to physics and causal structure-
seeking in physical inquiry did not stop at Galileo. In fact, Galileo himself
argued that explanations of motions reside in their dynamical causes.221
With respect to some natural occurrences, Galileo’s scientific methodology
required that he search for the “true, intrinsic, and entire cause.”222 For
René Descartes (1596–1650), “all change” is reducible “to the local motion
of particles, and . . . change of motion” is “caused by localized impacts
between corpuscles.”223 Christiaan Huygens (1629–1695) sought an expla-
nation of gravitation in motion, hypothesizing that the latter is the cause of
the former. The relevant motions were due to “Cartesian vortices in a ple-
num of subtle matter.”224 Causal talk in Huygens is far from mere descrip-
tive gloss. The causal structure of nature was vital to acquiring a physical
understanding of the world. In his Treatise on Light, Huygens remarked,
This is assuredly the mark of motion, at least in the true philosophy,

in which one conceives the causes of all natural effects in terms of
mechanical motions. This, in my opinion, we must necessarily do, or
else renounce all hopes of ever comprehending anything in physics.225
There is an abundance of evidence that Robert Hooke (1635–1703) hunted

for the cause of gravitation. This is supported by the great presupposition of
his gravitational physics, viz., that gravitation has a cause. He also clearly
maintained that gravitation is the cause of motion.226 In the work of Newton,
forces, in general, were the causes of motion,227 and a good methodological
rule for seeking out sound scientific explanations is to refrain from positing
more causes than are required (Newton 1999, 794).228 Daniel Bernoulli’s
(1700-1782) rationalist tendencies included the adoption and application of
a principle of sufficient reason (PSR). In some cases of its application (quot-
ing Darrigol’s study), the PSR “boils down to considerations of symmetry:
the effect must have the symmetry of the cause, if no asymmetric circum-
stance . . . intervenes” (Darrigol 2014, 25). Leonhard Euler (1707–1783)
thought that the gravitational force was an effect of some more fundamental
cause and that motions were caused by forces.229 And one need only read just
a handful of Henry Cavendish’s (1731–1810) memoirs, interpretations, and
explanations of experimental results to see that he clearly understood him-
self to be uncovering causes of natural phenomena. To cite just one exam-
ple among many, Cavendish read an interesting paper on the experiments
of a certain Mr. Hutchins on May 1st of 1783. In that work, Cavendish
states that the goal of his inquiry is to explain Mr. Hutchins’s experimental
results by making manifest the underlying causes.230 Charles Augustin de
Coulomb (1736–1806) desired to know the deep causes of friction (Gill-
mor 1971, 132). Joseph-Louis Lagrange (1736–1813) defined dynamics
as “the science of accelerating or retarding forces and the diverse motions
which they produce” (Lagrange 1997, 169). Lagrange clearly believed that
forces caused motions and says as much throughout his Analytical Mechan-
ics.231 In Joseph Fourier’s (1768–1830) Analytical Theory of Heat, external
causes determine states of systems, and constant ratios between quantities
are maintained on the basis of persistently acting causes.232 Heat arises due
to constant causes.233 Causes render states of bodies susceptible to certain
mathematical descriptions. These causes act on masses.234 Fourier is very
clear about the fact that mathematical formalism should be interpreted in
such a way that it reveals causal relations. He spoke generally of this type
of interpretive endeavor, describing it as an “example of the relations which
exist between the abstract science of numbers and natural causes.”235 For
Fourier, analysis demonstrates the “true cause[s]” of the results of experi-
mentation (Fourier 2007, 24). André Marie Ampère (1775–1836) “did not
exclude the search for physical causes. He himself speculated on various
mechanisms for the production of electrodynamic forces.”236 Ampère seemed
to interpret his own mathematical results in causal terms, musing that,
The proofs on which I base [“my theory”] mostly result from the fact
that they reduce to a single principle three sorts of actions which all
phenomena prove to depend on a common cause, and which cannot be
reduced in a different manner.237
According to the theologian, scientist, and philosopher William Whewell

(1794–1866) (quoting Whewell), “the motion of bodies from rest is produced
by a cause which we call Force”,238 and “we should attempt to discover
both the laws of phenomena, and their causes.”239 Indeed, Whewell thought
that there were causal laws. He wrote, “the laws of phenomena, in many
cases, cannot be even expressed or understood without some hypothesis
respecting their mode of production” (Whewell vol. 2 1967, 103). The great
experimental physicist Michael Faraday (1791–1867) would sometimes
uncover substantiation for certain of his hypotheses; however, “he wanted
more”, wrote physicist Carl Trueblood Chase (1932, 77), having in mind
the case of electromagnetic induction. “He desired to find out exactly how
the effect was produced” (ibid., 77). The hunt for causation in the case of
electromagnetic induction led to Faraday’s discovery of the principle of the
dynamo (ibid., 78). Herman von Helmholtz (1821–1894) could not have
been clearer about causation in physics (I quote Frisch’s translation and
quotation of him),
Our demand to understand natural phenomena, that is, to discover

their laws, is a different way of expressing the demand that we are to
search for the forces that are the causes of the phenomena. The lawful-
ness of nature is conceived of as causal relationship, as soon as we rec-
ognize nature’s independence from our thought and from our will. Thus
when we ask about the progress of science as a whole, we will have to
judge it according to the extent in which the recognition and the knowl-
edge of causal connections, encompassing all natural phenomena, have
progressed.240
In James Clerk Maxwell’s (1831–1879) Treatise on Electricity and Magne-

tism and in other work, various equations and forces are interpreted caus-
ally. For example, gravity is said to be that which causally produces its
effects (e.g., it causes iron filings to fall). The electromotive force is clearly a
causal force in Maxwell’s eyes. One equation featuring it is said to commu-
nicate how that force “causes the current to flow through the conductor.”241
When informing his readers about how to “find the electromotive forces in
a moving body,” he remarked, “[t]he variation of the velocity of the vortices
in a moving element is due to two causes—the action of the electromotive
forces, and the change of form and position of the element.”242 The causal
talk is not mere gloss. We learn from Maxwell’s “Essay for the Apostles on
‘Analogies in Nature’” (1856) that, “[w]hen the objects are mechanical,
or are considered in a mechanical point of view, the causes are still more
strictly defined, and are called forces.”243
In the twentieth century, we find in the work of the discoverers of rel-
ativity and quantum mechanics similar causal views of elements of both
classical and quantum physics. Albert Einstein (1879–1955) said that
“[t]he scientist is possessed by the sense of universal causation.”244 Einstein
causally interpreted both STR and GTR, and even maintained that the
final grand unified theory would be causal ((see Pais 1982, 465) and q.v.,
chapter 3: sect. 2.2 for more support). Bohr reported that Dirac judged that
quantum theory “describes the state of the world at any instant by a wave
function ψ, which normally develops according to a causal law so that its
initial value determines its value at any later instant.”245 John Stewart Bell
(1928–1990) explicitly posited a non-local and instantaneous action-at-a-
distance or causal dynamical explanation of the EPR correlations, con-
cluding that “. . . there must be a mechanism whereby the setting of one
measuring device can influence the reading of another instrument, however
remote. Moreover, the signal involved must propagate instantaneously . . .”
(Bell 1964, 199). And elsewhere he states, “[s]o the quantum correlations
are locally inexplicable . . . we could admit the input at one end as a causal
influence at the other end” (Bell, 2004, 153; with Healey, 2017, 59, reading
Bell similarly). Bernard d’Espagnat (1921–2015) held a similar view about
the correlations, musing that “[w]henever a consistent correlation between
such events is said to be understood, or to have nothing mysterious about
it, the explanation offered always cites some link of causality” (d’Espagnat
1979, 160). David Bohm (1917–1992) had an incredibly rich causal view of
the laws of nature. He said,
When we study any particular set of processes within one of its rel-
atively autonomous contexts, we discover that certain relationships
remain constant under a wide range of changes of the detailed behavior
of the things that enter into this context. Such constancy is interpreted
not as coincidence, but rather as an objective necessity inherent in the
nature of the things we are studying. These necessary relationships
are then manifestations of the causal laws applying in the context in
question. . . . [A]ctual experience shows that the necessity of causal
relationships is always limited and conditioned by contingencies aris-
ing outside the context in which the laws in question operate . . . the
categories of necessary causal connection and chance contingencies
are seen to represent two sides of all processes.
(Bohm 1957, 29; emphasis in the original)
Bohm explicitly causally interpreted both Newton’s laws of motion and

Maxwell’s mature field theory of electromagnetism in causal terms. He wrote,
The combined laws (Newton’s equations for the bodies plus Maxwell’s
equations for the fields) then form a unified and extended set of basic
causal laws, generalizing the laws of Newton, which, as we recall, were
expressed solely in terms of the motions of the bodies. Thus, the com-
plete causal laws now include both bodies and fields.246
The deterministic laws of Newtonian mechanics were not the only causal
laws for Bohm. Indeterministic physics incorporates statistical laws that
Bohm described as “approximate causal laws.”247 Erwin Schrödinger
(1887–1961) desired to (quoting Moore’s summary) “champion the cause
of strictly causal physics.”248 And we could go on with further quotations,
citations, and examples. The point of this historical discussion is to empha-
size the fact that many physicists adopted causal approaches to physics and
conceived of their inquiry as a searching evaluation of the world that should
uncover causes. The above historical discussion lends some prima facie
plausibility to the prospect of engaging in fundamental phenomenon-first
metaphysical theorizing about causation. The current project will seek to
engage in such inquiry at numerous junctures, and various arguments will
be provided for causally interpreting physics in chapters 2, 3, 4, 7, and 8.
Section 5: Conclusion
Metaphysical theorizing about causation in the current work will be gov-
erned by the directives and methodological principles of this chapter. My
metaphysicalC system will therefore allow for the existence of truths, mean-
ingful entities, mental states, mental events, instances of justification/war-
rant, instances of knowledge, instances of mental causation, instances of
knowledge of “otherly”-mental causation (and therefore other minds), the
metaphysicalC system itself, and the very philosophical analysis and funda-
mental metaphysical theory of causation I arrive at. My theory will be crafted
upon the assumed falsehood of Sider’s new-fangled Humeanism; scientific
epistemic structural realism; Ross, Ladyman, and Spurrett’s new verifica-
tionism; and various anti-realist theses such as Putnam’s doctrine of concep-
tual relativity. My metaphysicalC system building will presuppose objectual
domain condition interpretations of the quantifiers, a classical logic, and a
proper functionalist resolution to the problem of projection, plus the proper
functionalist theory of warrant. The theory I arrive at will be a philosophical
analysis of causation. The theorizing I use will include fundamental meta-
physical theorizing, and the theory I construct will be a fundamental meta-
physical theory of causation because the theorizing I perform will be done
hand-in-hand with the deliverances of a metaphysically informed physical
inquiry (i.e., physical inquiry associated with the construction of physical
theories with the types of structure I’ve detailed in this chapter). Last, the
sources of evidence I appropriate for my theorizing will at least include
perception, memory, introspection, reason (including deductive, inductive,
and abductive inference making), and a priori intuition understood as basic
sources of epistemic justification. Moreover, I will use testimony, and vari-
ous sciences as non-basic sources of epistemic justification.
Notes
1. A point about notation at the outset: (except where context renders matters
clear) I note, first (unless otherwise noted), that I represent specific propositions
with ‘< >’ brackets, and specific facts or obtaining (worldly) concrete states of
affairs with ‘[ ]’, a different set of brackets. With respect to the latter mode of
expression, I’m using a convention that is close to one adopted in Rosen (2010)
(although he uses such brackets to pick out the kinds of things that may be the
same as Russellian (structured) true propositions (ibid., 114 and n. 3 there)).
With respect to the former mode of expression, I follow the discipline that is
philosophy of language generally. I will also represent properties by emphasis
through bold italics. Important technical terms that are introduced for the first
time will be emphasized through the convention of underlined emphasis. I will
indicate emphasis in ordinary language use by using italics. Sometimes I will
italicize to emphasize the main logical connective in a sentence or sentence form.
2. Plantinga (2010, 678). I provide my assumed theory of warrant in sect. 4.3.
Throughout this work, epistemic justification will be warrant minus the no-
defeaters condition in the theory of warrant communicated in sect. 4.3. I situate
my general epistemology in a tradition that is at odds with Timothy Williamson’s
knowledge-first approach, and Stephen Hetherington’s thesis that one can get
a plausible theory of knowledge with mere justified true belief, so long as the
involved justification includes an awareness condition incorporating the notion
of knowledge itself (see Williamson 2000 and Hetherington 2016, 219, respec-
tively). I will assume that full belief, truth, and warrant are necessary and suffi-
cient for knowledge. Warrant therefore incorporates all that is needed to appease
Gettier (1963).
Later, I will make use of principles that assume the above picture of knowledge.
I wield these principles against positions like eliminative materialism (the thesis
that there are no mental states). Proponents of eliminativism will charge me with
begging the question. My metaphysical system simply takes a theory of knowl-
edge at home in folk psychology for granted, and then incorporates principles that
both presuppose the legitimacy of folk psychology and preclude eliminative mate-
rialism. I insist that knowledge requires full belief, and as I will go on to suggest
in chapt. 2, eliminative materialist replacement theories of propositional attitudes
like belief and other bits of the ontology of folk psychology are at best speculative.
Eliminative materialists have no replacement theory of knowledge because they
have no real replacements for propositional attitudes. Why then exclude them
if they are needed in an account of knowledge? The retort will be “because of
empirical considerations”. But even if Churchland (2007) is right about how cer-
tain of the deliverances of neuroscience are not captured by folk psychology, that
would not show that folk psychological notions are not required to best explain
our behavior or to best characterize knowledge. One can readily agree with
Churchland that neither reductionism nor an identity theory best solves problems
about the mental and the physical, and yet embrace the view that the goings-on
described by neuroscience realize (or microbase-determine) the higher-level mental
entities or activity that is properly capturable by folk psychology (assuming, for
now, that realization is not a reduction relation). Does the resulting view imply
that the mental is epiphenomenal? No. For as Peacocke (1979, 119) maintained,
the realization relation is such that the following conditional comes out true,
[i]f John’s having P [a psychological state] caused it to be the case that John
raised his arm, then John’s being in N [a neurophysiological state] caused it
to be the case that John raised his arm.
(ibid.)
Who’s afraid of causal overdetermination? Not me. Or perhaps the involved
causes are joint and therefore partial. There are options.
Empirical evidence for the thesis that folk psychological equipment does not
reduce to pieces of the ontology of neuroscience, etc. does not constitute a con-
clusive empirical case against folk psychology. There are successes of folk psy-
chology to recover (a point admitted by Churchland Materialism 1995, 154).
Eliminative materialist interpreted neuroscience is not yet up to the task. A for-
tiori, we come up short even in cases involving the movements of our limbs
(as Schiffer said, “no one can give complete neurophysiological explanations
of bodily movements” (1987, 147)). In Churchland’s own sketch of an elimi-
nativist theory of cognition, he appropriates all manner of intentional relations
such as representation relations (2007, 174, 175). He speaks of “perspective-
neutral portrayals of the abstract categorial and causal structure of the world”
(ibid., 176; emphasis mine). He speaks of maps that embody or encode a cogniz-
er’s understanding of none other than universals (ibid., 175). And he speaks of
learning as something that transpires and causally produces synaptic connection
arrangements (ibid., 174), etc. We are not told what really goes on in the neuro-
physiology of it all. One wonders if certain folk psychological items are not here
being borrowed for the sketch. In any event, I believe we can save mental states
and events by following Peacocke’s lead (this is one way).
3. This gloss is from Williamson (2000, 243).
4. See the case for K-A in Williamson (2000, 11–12, 249–260) and Turri (2015).
5. Unless of course they affirm some type of skepticism. I should confess that
I know many theoreticians do not intend to advance their theories in accordance
with the K-A. However, I do not see why we should take those theories seriously
if belief in them cannot actually be warranted or if we cannot actually have
knowledge of them.
6. Some have used the term ‘facts’ to pick out true propositions. However, others have
used it to pick out objects standing in relations or exemplifying properties (states of
affairs). Except when context makes matters clear, I will use the locution ‘factP’ to
pick out the former notion, and the term ‘factΣ’ to pick out the latter notion.
7. Aristotle affirmed this idea in his Categories. See Aristotle (Categories 1984, 22).
Compare D.K. Lewis (1999, 206).
8. Merricks (2007, xiii).
9. My use of the term ‘thing’ is intended to pick out all that stuff belonging to
the broadest ontological category (e.g., particulars, universals, properties,
relations).
My use of existential quantifiers abides by the directives and principles for
quantifiers articulated in sect. 4.
The relevant sufficient condition for a cognizer’s being epistemically irrational
with respect to belief retention or belief formation is as follows: A cognizer C is
epistemically irrational with respect to her retention or formation of belief b, if
C has an actual mental state defeater for b in her noetic structure. See the main
text just after MP1 for a definition of actual mental state defeaters.
10. Cf. the discussion of mental state defeaters in (M. Bergmann (2006, 158–159),
although he appeals to justification, not warrant), on which I lean only slightly.
One could, as Bergmann does, explicitly add in to the definition of mental state
defeaters time indices, but I leave that type of sophistication out of my discussion
for simplicity.
11. A similar point has been made by Baker (1987) in the context of eliminative
materialism in the philosophy of mind. For more on this issue, q.v., chapter 2:
sect. 2.
12. For similar reasons, your method should refrain from precluding or encouraging
agnosticism about the existence of phenomenal consciousness, concepts, concep-
tual content, and processes of thought.
13. On the point that even some involuntary actions are mentally caused, see Maslen,
Horgan, and Daly (2009, 523, n. 1); S. Yablo (1997, 251).
14. An instance of the knowledge of “otherly”-mental causation involves one’s com-
ing to know that some other individual mentally caused some event (e.g., Bran-
don’s decision to ask a question caused Brandon’s hand to raise). Cutting off
instances of that kind of knowledge puts one in a terribly skeptical place.
15. Davidson (Mental Events 2001, 212). Strict laws are exceptionless generaliza-
tions that back counterfactuals and that are part of some broader complete (with
respect to some sphere of inquiry) physical theory, see B. P. McLaughlin (1989,
112, and the sources cited therein).
16. Point (c) is little emphasized, but see Davidson (1993, 3).
17. See the discussion of this in Kim (2007, 229), although the complaint there is
localized around “mental properties, or mental descriptions of” mental events
(ibid.). If, however, mental properties are identical to physical properties (i.e.,
type-identity theory holds), then they too earn their causal place in the world
through the physical properties with which they are identical. The position does
not eliminate the mental by identifying it with the physical. Identity is a sym-
metric relation, and relations require relata. The mental exists on both type and
token-identity theories.
When some scholars called Davidson’s view “epiphenomenal,” they had in
mind type-epiphenomenalism (see on this type of epiphenomenalism McLaugh-
lin (1989)). Type-epiphenomenalism does not imply that mental events do not
cause physical events.
18. Q.v., n. 2.
19. In addition, there are coherent stories to tell about mental causation in the face
of other worries that are dubbed threats to mental causation as well (e.g., the
problem of mental quausation, i.e., causation by the mental qua mental; Hor-
gan 1989) (although Horgan defends the thesis that there are instances of men-
tal quausation, his presentation of the problem is quite clear), and the causal
or explanatory exclusion argument defended in Malcom (1968, 51–53) (cf.
MacIntyre (1957)), and then later addressed by K. Bennett (2003); Goldman
(1970, 157–169); Kim (1993); Peacocke (1979, 119–124); Schiffer (1987, 146–
150), and many others.
20. I interpret Sider’s Humeanism as the best implementation of what might be a
much weaker skeletal version of a broadly Quinean and Humean metameta-
physic. One can see throughout Sider’s treatise on metaphysical methodology the
intent to strive toward rendering Humeanism metaphysically plausible. I have
no qualms with this way of favoring one’s worldview. For as Sider himself put
it, “metametaphysical critiques are just more metaphysics” (Sider 2011, 83).
Hence my association of one’s metaphysical methodology with one’s metaphysi-
cal worldview in the opening paragraph of this chapter.
21. Sider (2011, vii–viii, 13) claims to have inherited his notion of an ideology from
Quine (1951) and Quine (1980). However, Quine (e.g., 1951, 14) believed that
the ideology of a theory includes both its primitive and derivative notions. Sider
seems to limit ideology to primitive/structural or joint-carving notions solely.
22. As Sider wrote, “believe in the fundamental ideology that is indispensable in our
best theories” (Sider 2011, 188). And Sider does have in mind “our most funda-
mental inquiries” (ibid., 272) when he speaks of “our best theories.”
23. Ibid., 12; emphasis mine. Notice the language of correspondence.
24. Given the expository points of the main text, it is somewhat strange to find
Sider’s admission that “structure is not an entity or stuff” (ibid., 5, attempting to
explain some of the ways he speaks about structure). I cannot make this cohere
with the supposition that structure is worldly, objective, non-linguistic, and non-
conceptual. Nor does it fit with the idea that there can be more or less of it (ibid.,
14). Nor is it consistent with the claim that it is a certain way (objective and
worldly). Nor is it a happy partner with the view that it stands in a relation with
the ideology of TM because relations require relata, some stuff.
25. Sider remarked, “[a] fundamental theory’s ideology is as much a part of its repre-
sentational content as its ontology, for it represents the world as having structure
corresponding to its primitive expressions” (ibid., viii).
26. My reading makes sense of the secondary literature. Schaffer’s review of Sider
states that “[a] second change in view: Sider’s preferred mereology has shifted
from the universalism of Sider 2001 to nihilism, driven by the demand for
ideological parsimony, which invites the elimination of mereological terms if
possible” (Schaffer 2014, 128, n. 4). How would the elimination of mereological
terms in the ideology of a theory T suggest something about T’s ontology (mereo-
logical nihilism) if there were no substantive relationship between the two?
27. Quine (1948/1949, 32). Although, see my comments on structure in the main
text, for Sider wants to be able to also say that certain notions like the existen-
tial quantifier of classical first-order logic are themselves structural. They cor-
respond to structure.
28. Sider (2011, 112; emphasis mine).
29. Well, given that his reasons for rejecting grounding, truthmaking, and other
approaches hold water.
30. When I qualify terms or phrases like ‘exists’ or ‘there are’ with ‘(fundamentally),’
I am noting that I intend to use the first-order existential quantifier in the expres-
sions in which such phrases or terms appear. Sider believes that the first-order
existential quantifier is structural.
31. Sider considers himself to be providing some type of reduction (ibid., 118, n. 17).
32. Because the following inference is valid,
(1) ~q→~p
(2) q↔r
(3) ~ s → ~ (q ↔ r )
(4) ∴ (~ s → ~ p)
33. This is because the following is a theorem of classical propositional logic:
( p ↔ q) → ~ ( p ↔ q) → ~ q  .
34. I will assume, for now, that a derivative* fact/truth is one that holds by virtue of
some other fact/truth, and that a fundamental* fact does not hold by virtue of
some other fact (departing now from Sider’s usage of such terminology by aban-
doning the claim that the “by virtue of” idea is captured well by metaphysical
semantics provision, and by leaving the “by virtue of” notion unanalyzed and
primitive although it will respect the formal constraints on general (non-mathematical)
explanation).
35. Worrall (2007, 125). Worrall calls the view “[s]tructural [s]cientific [r]ealism”
(ibid.). I will focus on Worrall’s distinctive brand of ESR. It is motivated by an
odd couple, viz., (a) the success or (no) miracles argument for scientific realism
(discussed in Devitt 2005, 772–776; Putnam 1979, 73; Putnam 2010, 18–19,
inter alios), and (b) the pessimistic meta-induction argument (discussed in its
very early form by Poincaré 1905, 114–115); cf. Putnam (2010, 25, 36–37)) for
scientific anti-realism. Worrall is therefore forced to defend forms of both argu-
ments (a) and (b). He does this in Worrall (1989; 2007; 2012).
ESR is criticized by Cei (2005), Chakravartty (2007, 35–39), French (Struc-
ture 2014), and Psillos (1999, 146–161). It is thought to succumb to the famous
Newman argument of Newman (1928) due to its similarities with Russell’s
structuralism—according to which what we know about the mind-independent
world beyond that with which we are directly acquainted is only its mathematical/
formal features, its structure—Russell (1927, “. . . structure is what can be
expressed by mathematical logic, which includes mathematics.” 254, “The only
legitimate attitude about the physical world seems to be one of complete agnosti-
cism as regards all but its mathematical properties.” 270–271); (cf. Demopoulos
and Friedman 1985).
I do not know the first place in which the distinction between ESR and ontic
structural realism (OSR) appeared. The underlying ideas behind the distinction
show up alongside one another in Ladyman (1998), although the term ‘ontic’
never appears in that paper. Ladyman (ibid.) notes that the fundamental content
was already in earlier work. Many believe that there was a dearth of important
scientists and philosophers in the history of thought who were structural realists
of one kind or another (see Gower 2000).
36. Quoting Poincaré (1952, 160, 161), Worrall (2007, 132–135) argues that Poin-
caré was a proponent of ESR.
37. Worrall (2007, 134). Worrall is here describing Fresnel’s theory, but he notes
what the theories of Fresnel and Maxwell have in common when he asserts what
I have quoted.
38. Structure is also said to be dependence that holds between quantum states
and constituents or effects of microsystems in quantum mechanics (Worrall
1989, 122).
39. Worrall (1989, 117).
40. Worrall (2007, 134).
41. Ibid., 147.
42. Ibid., 148.
43. Worrall (1989, 122; emphasis mine).
44. Ibid., 123. In the immediate context, Worrall has in mind “what Newton discov-
ered” (ibid.), but his point generalizes to other cases.
45. As Plantinga put the point more generally, one “cannot perceive another’s men-
tal states” (Plantinga 1967, 188).
46. Susanna Siegel argues that obtaining causal relations are like holes, in that they
can be represented in our visual experiences, but I take it that her case was not
intended to cover all actual instances of obtaining causal relations (Siegel 2009,
538–539). Cf. Siegel (2010, 117–139)
47. See on this problem, Audi (2011, 340–342); Chihara and Fodor (1965); P. M.
Churchland (1979, 89–95; 2013, 111–118); J. A. Fodor (1987, x, who discusses,
very clearly, the idea of justifiably attributing beliefs and mental states on the
basis of behavior); Gomes (2015, appeals to testimony to solve the problem);
Melnyk (1994); Pargetter (1984); and Plantinga (1967, 187–244). By far, the
orthodox solution invokes inference to the best explanation, understood as a
type of scientific inference (Melnyk (1994, 482) calls the IBE approach ortho-
doxy, although he does not endorse it).
48. Soames (Cognitive Propositions 2014, 92–93)
49. For example, Davidson (Truth and Meaning 1967, 304, 307), where by ‘mean-
ings’ of things like declarative sentences, he has in mind stuff like propositions
(see the reading of him in Soames (2008)). In Davidson (1984, 55), “a theory
of truth” suffices to “give the meanings of all independently meaningful expres-
sions . . .” (ibid.). A standard argument for rejecting theories of meaning that
require referents of sentences such as propositions is called “The Slingshot” (cf.
the discussions and readings of Davidson in Lepore and Ludwig (2011, 256);
and Soames (Meaning 2010, 33–48, 54; Language 2010, 109); cf. Soames (Prop-
ositions 2014)). Alston (2000, 118) discusses reasons for bemoaning the use of
propositions in one’s theory of meaning. Alston was not himself a detractor of
propositions (see ibid., 117, 119).
50. See Bolzano (Theory of Science: Volume 1 2014, 81, 84–85, 88; Theory of
Science: Volume 2 2014, 1–175).
51. Frege (Brief Survey 1997, 300); see also Frege (Thought 1997, 336–337).
52. See Russell (1903, 42–52). Cf. the discussion of Russell’s view in King (2007,
19–24).
53. I will assume that those entities flanking markers like (Eq. 1) are the propositions
or equations themselves. One could put those entities in brackets ‘< >’ to achieve
the same effect.
54. Gowers (2008, 49).
55. Mathematicians commonly understand it that way. Although I disagree with
his precise way of putting things, Kline said that the very meaning of 2 is “a
number whose square” (operation) equals 2 (Kline 1985, 67).
56. Following the Fregean tradition found in Frege (The Foundations of Arithmetic
1997, 88; Logic in Mathematics 1997, 312; Thought 1997, 343)
57. For the theory of grounding I am assuming, see sect. 4.5.3.
58. See N. Salmon (1989). A proper name’s value base is the entity to which it refers
on the modified naïve theory (ibid., 346). The naïve theory posits that “the infor-
mation value of a singular term, as used in a possible context, is simply its refer-
ent” (ibid., 337). Salmon attributes something like the naïve view to Frege and
Russell (ibid., 338) and Soames (1989; Meaning 2010, 111). Forbes (1989, 136)
reports that Russell (at least at one time) held the view that proposition constitu-
ents are sense-data.
59. Adams (1981, 6–7), although I’m not sure if Adams goes in for a full-blown Rus-
sellian view of propositions. Cf. the comments in G. Fitch (1994, 181) and the
criticisms of structured Russellian views of propositions in Forbes (1989, 136).
60. Although propositions for Soames may also be concatenations of n-tuples, see
the discussion in Soames (1987, 69–81), with King (2017, sect. 3.1) reading
Soames similarly.
61. Soames (Cognitive Propositions 2014, 96).
62. Soames exegesis is difficult here. I include the “at least in part” qualifications
because Soames stated, “[s]ince the proposition that o is red is the event type
in which an agent predicates redness of o, it represents o as being red because
[explanation] all conceivable instances of it are events in which an agent does
so” (Soames, Cognitive Propositions 2014, 97; emphasis in the original). At least
some of the time (certainly in cases involving propositions like our example in the
main text), the actual world includes an actual state of affairs in which an agent
predicates in the relevant manner. That actual state of affairs would be among
the conceivable instances of the relevant event(s) (actual occurrences are certainly
conceivable). So, according to Soames, what actually happens can, and often does
(again, certainly in the case of the proposition that is discussed in the main text), at
least in part ground or metaphysically explain the intentionality of propositions.
Disclaimer: I do not know if Soames would approve of my insertion of a
grounding reading of the type of explanation involved here. However, if ground-
ing is the right way to understand metaphysical explanation (and there would be
separate arguments for that consideration), then outfitting Soames’s work with
that reading would be amicable.
That Soames maintains that what agents do metaphysically explains at least
in part the representational properties of propositions is clear. He wrote that a
proposition has the inherent representational properties it does “by virtue of
[metaphysical explanation] a relation it bears to agents” (Soames, Cognitive
Propositions 2014, 97). Soames does appear to allow one to understand event-
types as necessary beings (Soames Meaning 2010, 104, although matters are
unclear there). Does this undercut my grounding reading? No. What we do (at
least in part) grounds (in the Schafferian sense (again see sect. 4.5.3)) the relevant
actual property-instances involving propositions (like the one in the main text)
actually bearing their representational properties. I can cover propositions unlike
the one in the main text (i.e., propositions that have no actual instances) by
changing the consequent of premise (1) (and by making other consistent changes
in the argument to preserve validity) from “<Eq.2> has the truth-conditions it
does by virtue of our cognitive and/or linguistic activity” to “<Eq.2> (where
I’m now temporarily assuming that the proposition stated here has no actual
token instances) has the truth-conditions it does by virtue of our possible or
actual cognitive activity,” where the disjunction here is the inclusive disjunction
of propositional logic. I could then substantiate the adjusted part of premise (2)
(call it 2d*) by arguing that it is impossible for us to ever perform the necessary
predication activity that would yield all of <Eq.2>’s representational properties
because it is impossible for us finite cognizers to ever have a determinate concept
of σ, or at least I don’t see how we ever could have such a determinate concept.
63. Soames (Meaning 2010, 101; emphasis in the original).
64. There are, of course, qualifications to make. I may believe that p and thereby
render true the proposition that <Chris Weaver believes that p.> There is a sense
in which the truth of that proposition depends upon my cognitive activity. This
type of dependency is benign. What I’m after is the additional point that even
with respect to such a case, <Chris Weaver believes that p.> holds independent
of what you the reader thinks, or believes, or does. And moreover, even if, in a
moment of confusion or self-deception, I thought that I didn’t believe that p, if
I really believe that p, then <Chris Weaver believes that p.> will hold regardless.
65. RLS qualify their case as something put forward as a proposal. I will evaluate
their remarks as if they are truth-evaluable and intended as something represen-
tative of what someone actually believes. I will engage their view as if it were
advanced by someone seeking to abide by the preceding K-A.
Some of the following work I reference is in collaboration with John Collier as well.
66. Ross, Ladyman, and Spurrett (Scientism 2007, 29).
67. Ross, Ladyman, and Spurrett (Scientism 2007, 37). Stipulative definitions that
help further explicate PNC are given in ibid. (38).
68. See the discussions in W.G. Lycan (2000, 80–86); Merricks (2015, xiii); Soames
(Meaning 2010); Speaks (2015, 9–11). None of these sources cite motivations
having to do with physics, or how scientific theories or hypotheses relate to one
another. I take it that their motivations and pursuits are still worthy endeavors.
For example, the nature of meaning and the like is a very important issue in its
own right.
69. Ross, Ladyman, and Spurrett (Scientism 2007, 7). Borrowing wording from
their response to Lowe.
70. I could add that objective truths are true by virtue of their standing in a cor-
respondence/correlation relation with the mind-independent world of objects,
properties, etc. The correspondence theory of truth is ably and adroitly defended
by Rasmussen (Correspondence 2014). A close cousin of the correspondence
theory I lean toward is defended by Alston (1996).
71. Following Putnam’s characterization in Putnam (2010, 125).
72. Putnam (1994, 309).
73. Ibid.
74. Ibid.; emphasis in the original.
75. Putnam (2004, 39).
76. See also Putnam (1983, 205–228). The “at face value” clause does not rob my
criticism (to follow) of its bite. There is real incompatibility. Why else would
Putnam prohibit conjoining the two explanations?
77. Any view according to which explanations are deductively sound or inductively
cogent arguments (as in Braithwaite 1968, 321–322, according to which scientfic
causal explanations issue forth from deductive systems; Hempel and Oppenheim
1948, 137–138; Nagel 1961, 29–46; Popper 1961, 59; Strevens 2008, 77) will
require the truth of the explanantia and (in the deductively sound case) the truth
of the explananda. Multifarious theories of causal explanation suggest that such
explanations include in their explanantia true reports on antecedent conditions
or occurring events that are causes perhaps also with true causal laws (e.g., R.
Miller 1987, 60–105; W. Salmon 1977, discussing the third dogma of empiri-
cism, which is that scientific explanations have the form of arguments; see W.
Salmon 1984, 132, on the ontic conception of explanation; Sober 1984, 96).
Other accounts require true reports on causal history even if the causal informa-
tion provided only accurately reports on an event’s failing to have causes (see
Elster 1983, 34, 70–71; Jackson and Pettit 1992, 12–13; D.K. Lewis, Explana-
tion 1986, 217 (for more on this see chapter 5: sect. 2). Lange’s (2017, 5–12)
examples of distinctively mathematical explanations or explanations by con-
straint involve true explanantia.
78. I believe there are good responses to Putnam (1987; 1990; 1994).
79. For a good philosophically informed discussion of quantifiers in natural lan-
guage, see Hofweber (2016, 55–101) and the sources cited therein. I will assume
the generalized quantifier theoretic approach to quantifiers in natural languages
(see ibid., 96; Barwise and Cooper 1981; Gamut 1991).
80. Or, more technically, (∃x)(Gx) holds for a model M, just in case, there’s an entity
e in the universe of discourse of the model M, such that e has G or satisfies G, as
with or in the characterization in Hofweber (2016, 81).
81. We say that R1 and R2 are inferentially adequate if they license standard infer-
ences with quantifiers in classical logic. We say that R1 and R2 are materially
adequate if they preserve the truth of a set of obvious (non-philosophical) state-
ments like ‘There are grapes in Napa Valley.’ and ‘There are gluons.’ (see ibid.).
We could broaden our attention to other types of linguistic expressions besides
statements. But I will focus, for simplicity, on statements or declarative sentences.
82. For a defense of the view that a system of logic should be specified (at least in
part) by appeal to that system’s rules of inference, see Rumfitt (2015, 31–65).
83. See Hodges (2001) and Hughes and Cresswell (1996, 235–244), although they
call it the “lower predicate calculus.”
84. I have paraphrased the first three principles from Grandy (2002, 531).
85. Weaver (forthcoming). See Rumfitt (2015, 153–219), who delivers a full justifi-
cation of CPL, given an intuitionist metalogic.
86. Following Beall and Restall (2006). They also allow for a characterization of
classical logical consequence that utilizes Tarskian models. I have no preference
on the matter. Either will work.
87. I argue for the claim made in the main text in Weaver (forthcoming), but see also
Hellman (1993; 1998). B.P. McLaughlin (1997, 219) has said, “no one knows
how to do calculus without classical logic, and no one knows how to do physics
without calculus.”
88. Hirsch (Ontological Arguments 2008, 373). Also see Hirsch (2009).
89. Hirsch (Ontological Arguments 2008, 368).
90. Ibid., 376.
91. Ibid., 377. I am not currently engaging, what appears to me to be, a much more
radical quantifier variantism thesis in Hirsch (Structure 2008, 513), which is
applied to all physical objects, and which is also extended to cover, not only
quantifier statements about such objects, but also higher-level statements about
how certain existential claims about physical objects carve nature at the joints
(ibid., 521–522). Sider has responded in Sider (2014).
92. The quantifiers of PFL are also privileged in this way. I do not have space to sup-
port this claim, however. Plural quantifiers enter my study in chapter 9: sect. 4.
93. Hirsch expresses a similar complaint about Sider’s notion of structure (Hirsch,
Ontological Arguments 2008, 378).
94. I am here merely providing at least one sufficient condition for a substantive
ontological dispute.
Although my objections are reminiscent of Sider’s, I do not mean to adopt
Sider’s Lewisian reference magnetism objection to Hirsch in Sider (2009).
The preceding reply resembles more closely that found in Sider (2014), which
argues that metaphysicians can choose to speak a privileged language, one
with the structure corresponding quantifier, even if at the cost of giving up a
principle of charity. My substantiation of this maneuver that follows is, as far
as I know, novel.
95. In other words, philosophers and other academics employ the same privileged
language (i.e., English, or another natural language, outfitted with the privi-
leged quantifier, perhaps with some other tweaks) at times. Many of us have
witnessed it and subsequently watched substantial disagreement arise. This
constitutes my response to Hirsch (Structure 2008, 521–522).
96. When Sider argued for the view that there’s a privileged language that carves
reality at the joints (Ontologese), Hirsch replied that, “[i]t seems unclear, in
any case, why it is necessary to switch to a new language” (Hirsch, Ontological
Arguments 2008, 377). See also Hirsch (Structure 2008, 523).
97. The normativity in play here is captured well by the claim that properly func-
tioning metaphysical researchers employ the privileged language(s) to conduct
their inquiry. The assigned meaning for the existential quantifier in physics
is the same one given to the existential quantifier in CFOL. I have already
noted how other approaches (involving other formal languages like intuitionist
logic) to the mathematics needed to do physics (e.g., constructive mathematical
approaches) fall short.
I should add that I will assume (as in van Inwagen (2009) and Sider (2009))
that there is no “space” between being and non-being. I deny Meinongianism.
98. I have changed the ordinary setup some. Some emeralds are blue-green, bear-
ing green as their primary hue. I have therefore chosen red instead of Good-
man’s blue in my setup. Using the term ‘grue’ now seems odd, but it’s OK to be
odd sometimes.
99. See, for example, the setup in Plantinga (Proper Function 1993, 129) on which
I lean.
100. Plantinga (Proper Function 1993, 19). On an environment’s being aimed at
truth, see ibid., 38–40. On design plans, see ibid., 22–24.
101. See the remarks in Plantinga (Proper Function 1993, 96).
102. See Plantinga (Proper Function 1993, 19, 41, 46–47; Why We Need Proper
Function 1993; 2000, 156–161). Some believe that in order for the account
to overcome Gettier-like counter-examples, warrant must entail truth. For a
defense of that idea, see Merricks (1995). For a defense of the related no-defeaters
condition, see M. Bergmann (2006).
103. M. Bergmann (2006, 134). I should note here that Bergmann takes himself to
be providing an account of epistemic justification, not warrant, because it is
admitted to lack sufficiency “for warrant” (ibid., 6).
104. The idea here is like epistemic overdetermination (see Casullo 2003, 37).
105. Sider (2009, 401; emphasis mine).
106. Davidson (1987, 443–444); Sosa (1993, 53–54). A similar worry appears also
in J. Taylor (1991). There are responses in Plantinga (Why We Need Proper
Function 1993) and Boyce and Moon (2016).
107. D.K. Lewis (Postscripts 1986). Lewis actually understood this to be a concep-
tual analysis (see following text), not a philosophical analysis. He was also
giving an analysis of partial and not full causation.
108. Fodor (1975; 1987; 1998; 2003; 2008); G. Harman (1987, 57); Millikan
(2000, 196); and see also the study in Pinker (1995, 55–82). Margolis and
Laurence (2007, 563) state, “[t]he Psychological View [that “[c]oncepts are
mental representations”] is the default position in many areas of cognitive
science and enjoys a good deal of support in the philosophy of mind.”
109. For criticism of competing views of concepts, see Fodor and Pylyshyn (2015,
19–63). The mental representation theory of concepts often walks hand-
in-hand with the representational theory of mind (see Fodor 1975; 2007;
2008).
110. This is a particularly controversial assumption, but I believe it is well-justified.
For a defense of conceptual atomism for lexical concepts, see Kwong (2007).
For criticism, see Prinz (2002, 89, calling the view “[i]nformational atomism”).
111. Kant (1992, 248–249).
112. Tennant (2015, 125; only the first and last emphases are mine).
113. This quotation is from Gottlob Frege, as quoted by Beaney (2016, sect. 2). The
bracketed clause was inserted by Beaney.
114. Here, I borrow the phrase ‘descriptive sense’ from Soames (2005, 2).
115. There are other brands of conceptual analysis for which see Glock (2017). But
see also Chalmers (1996, 52–71); Chalmers and Jackson (2001); and Jackson
(1998). These scholars use two-dimensional semantics and argue that concep-
tual analyses are specifications of primary intensions from the armchair. For
example, the primary intension of the natural kind term ‘water’ is a function
that maps a collection of centered possible worlds to the stuff that is in lakes and
streams and bays at those worlds (were any of those worlds actual). Concerning
a single centered world w, the function or primary intension of ‘water’ generates
the referent (the stuff in the lakes and streams) of the term ‘water’ as tokened in
w, were w actual, and maps the term to its referent. Terms like ‘water’ also have
secondary intensions, and these too are functions. However, these take the term
‘water’ and map its use at any world to H2O (the stuff in our lakes and streams),
under the assumption that the world we inhabit is the actual world. See the
helpful discussion (without endorsement) of two-dimensionalism in King (2007,
214–217).
116. For which see Frege (1980); Russell (1905); and Searle (1958).
117. See D. Kaplan (1979); Kripke (1980); Perry (1979); Putnam (1975); and
Soames (2005). Soames objects to current attempts at using two-dimensional
semantics to defend descriptivism. It is not a coincidence that modern-day
defenders of two-dimensionalism (and so a brand of descriptivism, too) are
also defenders of a type of conceptual analysis as a sound philosophical
methodology.
118. Kripke (1980). An initial baptism can transpire on the back of a definite
description as well.
119. See Aronson (1982, 302) and Fair (1979, 232–233) for a sample of some
of the flavors of empirical analysis, although I should add that Fair thinks
such analyses are necessarily true insofar as they are a posteriori identities
(ibid., 231).
120. Dowe (2000, 11).
121. Kutach (2013, 3).
122. Ibid., 5.
123. For more on these scientifically respectable causation-like relations, see Kutach
(2013, 67–75). Kutach (ibid., 10–11, n. 2) explicitly notes how the empirical
analysis of a concept can leave that concept behind. There is a type of elimina-
tivism in Kutach’s work (despite the remarks in ibid., 267).
124. Kutach wrote, “[l]et us say that a culpable cause of some event e is an event
that counts as ‘one of the causes of e’ in the sense employed by metaphysicians
who study causation” (Kutach 2013, 46; emphasis in the original).
125. Kutach (2013, 272).
126. Ibid., 273.
127. Ibid., 282. But cf., ibid., 318.
128. See the discussion in Sperber (1995, xvi).
129. See on this category, Audi (2002, 72), whom I follow. As Audi notes, one’s
justification for b might become defeated on account of the appropriation of
some other source, but that just shows off negative dependence or reliance for
justification of b on some other source of justification.
130. Sensation therefore involves causation J. Prinz (2006, 451). Sense perception
does as well Burge (2010, 376–377).
131. Siegel (2006, 484).
132. For more on the nature of introspection, see Lormand (2006) and Schwitz
gebel (2011).
133. Audi (2002, 74–75).
134. I follow Audi (2011, 104–114). I believe that in these cases, the beliefs are also
warranted.
135. Bealer (2002) and Sosa (2007, 60–61).
136. The quotations are from Sosa (2007, 61). Ernest Sosa noted in correspondence
that an instance of some cognizer’s rationally intuiting that p, as I understand
it, can be viewed or understood as a special case of the manifestation of an
appropriate epistemic competence, a competence of the kind to which his
account of rational intuition appeals.
137. Sosa (2007, 62).
138. Besides the sources just cited, please also see Sosa (2015). Cf., the strong case
for the a priori in BonJour (2005), and the defense of understanding intuitions
as evidence in philosophy in Climenhaga (forthcoming), although he has in
mind something much weaker than (but certainly consistent with) Bealer/Sosa-
style intuitions.
139. See particularly J.R. Brown (1986; 1991; 2004), who argues that Galileo’s
thought experiment regarding falling bodies involved the use of platonic intu-
ition and the acquisition, thereby, of a priori justification. Gendler (1998,
419–420) suggests that Galileo’s thought experiment is “a reconfiguration of
internal conceptual space” that “brings us to new knowledge of the world . . .
by means of non-argumentative, non-platonic, guided contemplation of a par-
ticular scenario” (ibid.) (this seems to suggest the deliverance is a priori).
140. Galileo (1967).
141. See the brilliant discussion of the relevant history in Darrigol (2014, 1–46). He
regards Daniel Bernoulli (1700–1782), René Descartes (1596–1650), Leonhard
Euler (1707–1783), and Hermann Ludwig Ferdinand von Helmholtz (1821–
1894) as, in some sense at least, practitioners of a rationalistic science.
142. Galileo (1967, 145; emphasis mine) in the mouth of Salviati.
143. Norton (2004, 49).
144. Norton (2004, 50).
145. Norton (2004, 51).
146. Gendler (1998, 417–418) seems to think her characterization does not depart
very far from Norton’s own formulation. I should add that Gendler rejects
the hypothesis that argument (6)–(11) has the same “demonstrative force”
(ibid., 404) as Galileo’s specific T. I should also add that I do not mean to
communicate that (6)–(11) is Gendler’s exact characterization of Galileo’s
argument.
147. All quotations in this argument are from Gendler (1998, 404).
148. Moser, Mulder, and Trout (1998, 115). In the immediate context, the authors
are summarizing an assumption of Robert Boyle’s (1627–1691).
149. Gendler (1998, 404).
150. See Biggs and Wilson (2017); BonJour (1998); Peacocke (2004); and Swin-
burne (2001).
151. Norton (2004, 50).
152. Horwich (1982, 12). Or as Plantinga put it, “. . . Bayesians propose coherence
as a necessary condition of rational belief” (Plantinga, Warrant: The Current
Debate 1993, 121).
153. “There is no evidence to believe that the mind contains two representational
systems, one to represent things as being probable or improbable and the other
to represent things as being true or false” (Weatherson 2005, 420). See also
Kotzen (2016, 625–626).
154. One notable exception being Moss (2013), who describes the view defended
there as “apparently radical” (ibid., 1).
155. M. Kaplan (1996, 125).
156. See Weatherson (2005, 420). Cf. related worries in G. Harman (1986, 106),
and Stalnaker (1984, 91).
157. Weatherson (2005, 421). See the discussions in Ganson (2008); Ross and
Schroeder (2014).
158. See Kahneman and Tversky (1982, 66–68); Harman (1986); Lipton (2004,
108–117); Nisbett and Ross (1980); Tversky and Kahneman (1983) “[s]ystem-
atic violations of the conjunction rule are observed in judgments of lay people
and of experts in both between-subjects and within-subjects comparisons.”
(ibid., 293). See the discussion in Smithson (2016), who cites some of these
sources as well.
159. Smithson (2016, 477).
160. See G. Harman (1965). Contra (Climenhaga 2017) and many others who
argue that IBE should be reducible to some form of Bayesian inference.
161. See Boyd (1984, 65–75) “. . . it is by no means clear that students of the sci-
ences, whether philosophers or historians, would have any methodology left if
abduction were abandoned” (ibid, 67) and see the remarks made about scien-
tists at the top of (ibid., 68). Raftopoulos (2016).
162. See also Weintraub (2013), although I reject Weintraub’s reductive thesis that
inductive inferences are a type of IBE.
163. That explanation involves the reduction of surprise or cognitive discomfort
is similar to an idea in Glymour (2007, 133). That “in the typical situation”
explanations, at least in part, remove puzzlement was suggested by Michael
Sherwood (1969, 13) and see (ibid., 14).
164. W.G. Lycan (2002, 413).
165. I provide my full definition of a cogent IBE in the following text.
166. H is a metaphysical hypothesis. So, H is (metaphysically) necessarily true,
if true. One axiom of probability theory says that if a hypothesis B is true
at every possible world, then Pr(B) equals 1. How then can some evidence
incrementally confirm it? There’s no problem here. It is common to think
of the necessity involved in the relevant axiom as logical necessity (Ear-
man 1992, 36). On my view, logical necessity is stronger than metaphysical
necessity. So, the axiom does not require that one’s credence in H be 1 if H
is metaphysically necessarily true. The second condition appropriates what
Malcolm Forster (2006, 321) calls the Law of Likelihood (LL). That law is
defended in Forster and Sober (2004), although there it is called the likeli-
hood principle. Forster (2006, 321, n. 4) notes that Forster and Sober had
in mind the law of likelihood, not the likelihood principle. See also Sober
(2002, sect. 3), who defends LL from an objection and who calls LL likeli-
hoodism in note 5.
167. Henceforth, all uses of grounding terminology should be given the definitions
provided in Schaffer (2009). Schaffer rejects the transitivity of grounding in
(Schaffer Transitivity 2012). I believe this was a mistake.
168. For brevity, I will refrain from explicating and explicitly using the preceding
criteria in subsequent discussion. But when I state that this or that conclusion
is supported by abductive inference, I will mean that there is an abductive
inference available that can be fully stated in a way that respects the outline for
abductive inference-making in this section.
169 See the collection of essays in Lackey and Sosa (2006) for more on the various
views on testimony.
170. For the ADM formulation of GTR, see Arnowitt, Deser, and Misner (2008).
For the tetrad formulation, see Rovelli (2004, 33–48).
171. For the Newtonian formulation, see Shankar (2014, 1–193). For the axiom-
atic formulation, see McKinsey, Sugar, and Suppes (1953). For the Lagrangian
formulation, see J.R. Taylor (2005, 237–325). For the Hamiltonian formula-
tion, see Ter Harr (1971). And for the geometric formulation, see Thorne and
Blandford (2017, 5–36).
172. See the discussions in Mahan (2009, 9–11) and Styer et al. (2002, 288–290).
For the Dirac transformation formulation, see Weinberg (2013, 21, 52–96). Cf.
the historical and philosophical discussion in Cushing (1998, 282–289).
173. See Carnap (1939); Feigl (1970); and Nagel (1961). Cf. the evaluation and
historical discussion in Suppe (1974, 6–118).
174. I’m quoting Putnam’s famous summary and critique of the view in Putnam
(1962, 250).
175. See the discussion (without endorsement) of the view in S. French (Structure of
Theories 2014, 301–304).
176. Craver (2008, 56). Craver is not a proponent of the syntactical view, although
he presents it well.
177. S. French (Structure of Theories 2014, 302).
178. Halvorson (2014, 587), although he also points out that Carnap (a principal
proponent of the Received View) was okay with second-order quantification.
There is a first-order axiomatization of GTR that uses the theory of locales
or pointless topology (Reyes 2011). But this approach is put forward as an
intended replacement of the standard theory of point-laden topological spaces.
Moreover, Reyes only recovers the vacuum field equations.
179. Models are “always a mathematical structure” (van Fraassen 1970, 327).
180. See Giere (1988), and van Fraassen says, “[w]hen equations formulate a scien-
tific theory, their solutions are the models of that theory” (van Fraassen 2008,
310; emphasis mine).
181. In van Fraassen (2008, 168), the idea is that interpretations relate (through
isomorphism) the world to surface models (or perhaps the surface models are
part of the interpretations), which are then themselves related (through iso-
morphism) to theoretical models. Isomorphism is that relation that is seen as
that which “is important because it is also the exact relation a phenomenon
bears to some model or theory, if that theory is empirically adequate” (van
Fraassen 1989, 219–220).
182. In their very thorough study of the law of inertia, Earman and Friedman dis-
cuss standard three-dimensional Newtonian mechanics and settle on a version
of that law that is equivalent to the formulation in Newton’s Principia. That
version has no strict mathematical expression (Earman and Friedman 1973,
336–338).
183. I’m quoting the translation of Einstein (Zur Elektrodynamik bewegter Körper
1905) by Arthur Miller in the appendix to A.I. Miller (1981, 392). P.G. Berg-
mann (1961) calls it the restricted principle of relativity.
184. Mook and Vargish (1987, 70). The discussion in Mermin (1989, 9–17) includes
an equivalent statement that does not mention light (ibid., 16). Elsewhere,
Mermin put it this way: “the speed of light has the same value c with respect
to any inertial observer” (Mermin 1989, 11).
185. In fact, many would suggest that they are laws of nature. As Marc Lange
documents, “[m]any scientists besides Einstein have distinguished constructive
theories from theories of principle, placing the principle of relativity along-
side various conservation laws, the laws of thermodynamics, and other laws
that are often taken to be ‘constraints’.” (Lange 2017, 115), noting earlier
how “physicists commonly characterize the principle of relativity as ‘a sort
of super law,’ ” as Lévy-Leblond (1976, 271) calls it (following Wigner 1985,
700) (Lange 2017, 101).
186. It is thought that lattice gauge QCD recovers the results of perturbative QCD
for the high momentum regime via a limiting process. So far as I’m aware, lat-
tice gauge theoretic QCD can only do this to low order of perturbation theory.
Thanks to Tom Banks for important correspondence on some of these issues.
187. Dirac (Emission and Absorption 1927, 261).
188. Dirac (Dispersion 1927). And see the historical discussion of these matters in
Pais (1986, 337–338).
189. van Fraassen (2008, 23). With this Frisch (2014, 24–47) and others agree.
190. The present work on causation will assume the standard account of the seman-
tics of counterfactual conditionals in D.K. Lewis (Counterfactuals 1973).
191. It is “rule-governed” because the prediction is not a deductive consequence of
the theory, nor is it a probabilistic prediction. Rather, a rule (a counterfactual)
is adopted that facilitates hypothesis formulation (i.e., a prediction because
the hypothesis is about the future behavior of objects of study). The rule and
prediction are objective because the counterfactual conditional is objectively
true (I am supposing). I borrow the title ‘rule-governed prediction’ from M.
Forster (2014, 450), although my specific understanding of this type of predic-
tion departs somewhat from Forster’s conception of it.
192. This example can be easily set up by stating the theory in first-order logical
terms (perhaps with a bit of mathematical jargon) and then by providing Eng-
lish and German translations of the first-order logical statements.
193. As suggested by the one-over-many argument for the existence of propositions
(see the discussion of the argument in McGrath 2014). I am not here intending
to suggest an inter-translatability account of synonymity.
I’m assuming that the designative formulations have partial interpretations
that enable their formulation expression.
194. Again see Rasmussen (2014).
195. As in Russell (1959, 119–130), although truth-bearers related by congruence
to factsΣ in that work appear to be beliefs.
196. Leaning partially on Alston (1996).
197. Again, I’m assuming the account of counterfactuals in D.K. Lewis (Counter-
factuals 1973).
198. PICs will include propositions that are partially interpreted purported laws.
These purported laws are propositions that satisfy the conditions of lawhood
articulated in the main text at some target metaphysically possible world(s)
that is/are, in part, represented by the content of the PIC in question.
199. On the motif of truthmaker theory, see Armstrong (2004); J.F. Fox (1987);
Lowe (2006); Rodriguez-Pereyra (2002); Truthmakers (2006); and the collec-
tion of essays in Beebee and Dodd (2005, inter alia).
200. As in Rodriguez-Pereyra (2002), although he has in mind sentences, not
propositions.
201. This is reminiscent of the spectral truthmaker theory discussed in Koons and
Pickavance (2017).
202. See the discussions in Dasgupta (2014); Fine (2001); Rosen (2010); and Schaf-
fer (2009).
203. For an example of how this might be done, see the discussions in Koons and
Pickavance (2017, 56) and Rodriguez-Pereyra (Entailment 2006, 960). Contra
Heil (forthcoming).
Truthmaking might also be understood in terms of aboutness. Such a view
is held up as the best approach to truthmaking in Merricks (2007, 33–34),
although it is later criticized by him.
204. See the discussion in Klein and Nellis (2012, 69–72).
205. There may be other kinds of acausal or non-causal scientific-mathematical rep-
resentations. See the discussion in Pincock (2012, 51–65).
206. This point was made by Nagel (1961, 344–345). Cf. Ross, Ladyman, and
Spurrett (Scientism 2007, 48). A similar point is made by Cartwright (1999,
59–64) in the context of a discussion of Coulomb’s law. However, her point is
intended to motivate the use of nomological machines.
207. Schaffer (Causation and Laws 2008, 92).
208. M. Redhead (1990, 146).
209. Hitchcock (2007, 56). Hitchcock does appear to endorse P* because he states
that “[w]e can explain the truth of P* without appealing to the incoherence of
causal notions” (ibid.).
210. See also Field (2003), who says he agrees with Russell regarding the elimina-
tion of causation in physics (ibid., 435), and Sider (2011, 15–17), where the
thought is that adding causation to our best theories does not bolster their
explanatory power; van Fraassen (1989, 282) dispels causation from Newto-
nian mechanics; and Smith (2013) with respect to classical mechanics gener-
ally (inter alios).
211. I borrow the title from Blanchard (2015).
212. Pincock adds to the representation two background assumptions about the
nature of braking force.
213. This is close to what Devitt calls scientific fact-realism (Devitt 2014, 257). See
also the studies in Devitt (1991; 2005).
214. Poincaré (1905, 113; emphasis mine). Cf. the discussion in Darrigol (2014,
202).
215. Darrigol (2000, 170).
216. Heisenberg (1967, 95, 98), as quoted and cited in and by Camilleri (2009,
115–116; italics from Camilleri removed).
217. Galison (2003, 318; emphasis mine).
218. Sloman (2005, 70).
219. Barbour (2001, 357).
220. Frisch (2014, 2). Although I agree with Frisch’s historical point, his (2014)
project differs significantly from mine. He argues that causal reasoning is indis-
pensable to physical representation, and merely that. I argue, inter alia, that
causation enters the ontologies of many of our best physical theories. Frisch’s
position is consistent with instrumentalism about causal structure (as Frisch
2014, 11, 244 admits). My view is not.
221. Q.v., chapter 3: note 24 for appropriate references.
222. As quoted in Heilbron (2010, 222). In De Motu, Galileo hypothesized that
(quoting Swerdlow’s summary) “[t]he cause of natural motion is . . . heaviness
and lightness” (Swerdlow 2013, 27). Causes subject to mathematical descrip-
tion are not the mysterious causes that Galileo described as “not ‘within the
limits of nature’” (ibid., 25).
Consider David Wootton’s reading of Galileo (which I quote),
Galileo is not interested in describing these systems [celestial orbits and opti-
cal phenomena] as causal because the causal mechanism is too obvious to
be interesting. Much of Galileo’s science is thus causal—but the causes are
trivial and transparent. What matters is that he is able to save or solve phe-
nomena by reducing them to mathematical principles.
(Wootton 2011, 23)
I’m not sure what it would mean for causation to reduce to mathematical
principles. Causation is an obtaining relation in the world, whereas mathe-
matical principles are clearly not, nor do they look apt for serving as reduc-
tion bases for causal relations. I think a better reading is this, Galileo is a
phenomenon- first natural philosopher. He believes (indeed it may be obvious
in the sense suggested by Wootton’s interpretation) that the physical phenom-
ena he is describing and explaining are causal, but he doesn’t stop there. He
wants a mathematically rigorous way of describing the phenomena, although
the resulting mathematical machinery should be interpreted causally precisely
because it is intended to track obviously causal phenomena. Perhaps that is
what Wootton meant by the idea that causal phenomena reduce “to math-
ematical principles” (i.e., he meant that the math tracks causal structure in the
world for Galileo).
223. Guicciardini (2013, 228).
224. Purrington (2009, 174).
225. Huygens (1952, 554).
226. See Hooke (1969) and Purrington (2009, 175).
227. Newton said that “[f]orce is the causal principle of motion and rest” (Newton
1962, 148).
228. This, of course, suggests that Newton thought sound scientific methodol-
ogy involved the search for causes. See the case for this in Achinstein (2013,
43–83). See chapter 3: sect. 2.1 for more evidence.
229. Euler said, “[y]ou have just seen, that gravity is a general property of all the
bodies with which we are acquainted, and that it consists in the effect of an
invincible [this is not a typo] force, which presses them downward” (L. Euler
1837, 168). See also the quotations and commentary in Calinger (2016, 266,
276, 320). The abundance of such causal talk in Euler suggests that it is not
gloss, but an attempt to accurately describe matters.
230. See Cavendish (2010, 195). This suggests that Cavendish was working with a
causal understanding of scientific explanations of experimental results (cf. my
previous discussion of Kutach’s methodology).
231. For example, on page 11 of Lagrange’s Analytical Mechanics (1997, 11), he
remarked, “[i]n general, force or power is the cause, whatever it may be, which
induces or tends to impart motion to the body to which it is applied” (ibid., 11;
emphasis in the original). See also Lagrange (1997, 36), where the mathematical
formalism is explicitly interpreted causally.
232. Fourier (2007, 94). Indeed, the locution ‘external cause’ is used almost like a
technical term or clause throughout the presentation of his analytical theory
of heat.
233. Fourier (2007, 68).
234. Ibid., 109.
235. Ibid., 24. The famous remark of Fourier’s that “[f]irst causes (les causes pri-
mordiales) are unknown to us; but they are subject to simple unvarying laws
that can be discovered by observation and the study of which is the object of
natural philosophy” (as translated and quoted by R. Fox 2013, 424) is best
understood not as a dismissal of causation in physics, but as an admission that
causal relations are described by the dynamical laws of physics or else they are
beyond us.
236. Darrigol (2000, 27). Darrigol does go on to add, “But he required a clean
separation between laws and causes” (ibid.).
237. As translated and quoted by Darrigol (2000, 27). The bracketed clause was
inserted by Darrigol.
238. Whewell (vol. 1 1967, 166).
239. Whewell (vol. 2 1967, 105; emphasis in the original). Cf. the discussion of
Whewell in Lange (2017, 144).
240. As translated and quoted by Frisch (2014, 1). Compare Darrigol’s commentary
in Darrigol (2000, 215). Darrigol subsequently notes that Helmholtz would
later change his view about the comprehensibility of nature (ibid., 216).
241. Maxwell (2010, 301).
242. Maxwell (2003, 481).
243. Maxwell (1990, 378; emphasis in the original).
244. Einstein (1954, 40).
245. Bohr (1985, 104); as cited and quoted by Camilleri (2009, 114).
246. Bohm (2005, 29).
247. Ibid., 20.
248. W. Moore (1989, 246).
2 In Defense of the Causal Relation
Section 1: Introduction
Causal eliminativism is the view that there are no obtaining causal relations.
Causal realism is the view that there are obtaining causal relations. Obtain-
ing relations are not properties. They connect or tie together various sorts
of relata by holding between them. Properties inhere in substances (assum-
ing an Aristotelian conception of concrete particulars) and do not connect
them.1 When I say that causation is an obtaining relation, what I mean is
that when x causes y, x stands in an actual causal relation-instance to y.
Causal relation-instances are actual instances of some cause bringing about
some effect. They are themselves concrete states of affairs, and the relata
in the relation-instances are actual particulars (possible candidate relata
include occurring events, changes, property instances, or factsΣ (although
see chapter 7 for my view of causal relata)).2
I believe that (i) causal realism is true, and moreover, I maintain that (ii)
necessarily, all instances of causation are instances of an obtaining causal
relation. In this chapter, I play both offense and defense for the causal real-
ist team. While on offense, I show that one should affirm causal realism
because of the existence and nature of beliefs, inter alia. On the defensive
side, I object to D.H. Mellor’s argument for causal eliminativism, and I show
that E. J. Lowe’s attempt to make true causal facts with directed dispositions
and their manifestation partners alone is problematic. I then critically evalu-
ate David Lewis’s use of the void to show that (ii) fails.
Section 2: There Are Obtaining Causal Relations

Recall principle MP2 from chapter 1: sect. 1. That principle suggests (inter
alia) that any theory that asserts that there are no instances of mental cau-
sation, or that one cannot know that there are such instances, or that one
should refrain from believing that there are instances of mental causation is
a theory that cannot actually be known. This is because were a cognizer C
to believe (with belief b) any such theory (thereby also believing what the
theory asserts), C would thereby procure an actual mental state defeater for
b. That is to say, C’s forming b generates a self-defeater, a self-undermining
belief, a belief that is an actual mental state defeater for itself (my move is
reminiscent of Baker 1987, 134–148). But as chapter 1: sect. 4.3 made clear,
warrantK is a necessary condition for knowledge, and defeaters of this kind
rob cognizers of warrantK. But as was also made clear in the prolegomena of
the present work (i.e., chapter 1), metaphysical theories, theories like those
that eliminate mental causation and that cannot actually be known are sig-
nificantly deficient. Thus, my metaphysical methodology provides some
motivation for committing to the existence of instances of mental causation
and so also obtaining causal relations.
In addition, chapter 1: sect. 4.4.1 committed to the causal theory of refer-
ence. Given that theory, and that we do successfully refer to individuals that
figure in initial baptisms of the distant past (and I assume that we do success-
fully refer in that way), it follows that there are causal chains, and so also
instances of obtaining causal relations. Chapter 1: sect. 4.5.1 committed to
the existence of a basic source of epistemic justification that is sense percep-
tion. Instances of sensation and sense perception were said to involve obtain-
ing causal relations (the environment impressing itself upon the senses). Thus,
if there are instances of sensation or sense perception, there are obtaining
causal relations. There are such instances, so there are causal relations.
The causal eliminativist has a lot of work to do to overcome these three
arguments. They need to articulate theories of mental “causation,” refer-
ence, and sense perception that do without the causal relation, or else justify
getting along without such phenomena completely.
My fourth argument for the existence of obtaining causal relations is as
follows,
(1) There are beliefs.3

(2) All beliefs are formed by cognizers or by some parts of cognizers, or by
some constituents of an arrangement that can be identified as a cognizer
(where talk of cognizers, or their parts, or the appropriate constituents
is shorthand for talk of events involving cognizers, or their parts, or the
relevant constituents).
(3) If (1) and (2), then there are obtaining causal relations.
(4) Therefore, there are obtaining causal relations.
Let T2 be the eliminativist theory that there are no beliefs. MP2 suggests
that were one to believe (with belief b2) T2 (thereby also believing what the
theory asserts), one would rob b2 of warrantK,4 and thereby preclude oneself
from knowing T2. Given the metaphysicalC system of chapter 1, it looks as
if T2 cannot actually be known. That conclusion is a significant reason to
avoid T2.
There are other reasons, reasons outside a sound metaphysical methodol-
ogy, for maintaining that there are beliefs (premise (1)). It is a deliverance
of our best empirical psychology, psychiatry (specifically cognitive neu-
ropsychiatry), and cognitive neuroscience that there are delusions. Those
In Defense of the Causal Relation 91
disciplines cannot best explain behavior indicative of delusional disorders5
without entities that are delusions. Moreover, without delusions, they can-
not best explain the behaviors of patients diagnosed with severe dementia
who exhibit behavior indicative of mirrored-self misidentification delu-
sions.6 According to our best psychology, in such misidentification cases,
the behavior of said patients is best explained (and I’d maintain that it is
best explained in the abductive sense explicated in chapter 1: sect. 4.5.3) by
the fact that the involved patients have formed “the [false] belief that” their
“reflection in the mirror is a stranger. . . .”7 Importantly, the element of false
belief attribution is a part of many leading accounts of delusions generally
conceived.8 But necessarily, if there are false beliefs, then there are beliefs.
Thus, the empirical evidence for the existence of delusions from psychology,
psychiatry, and cognitive science is evidence for (1). Premise (1) is therefore
well-supported by special science.
It is also a deliverance of our best empirical psychology that (quoting
Halligan) “beliefs held by patients about their health and illness are central
to the way they present, respond to treatment and evaluate their capacity
for work.”9 Moreover, in that best empirical psychology, some beliefs of
cognizers best explain their dispositions to speak, act, and cope with stress
(Lazarus and Folkman 1984, 64–78; Halligan 2007, 358; Fodor wrote,
“commonsense belief/desire psychology explains vastly more of the facts
about behavior than any of the alternative theories available . . . there are no
alternative theories available” (Fodor 1987, x)). The list of ways in which
beliefs figure in successful special scientific argumentation and best expla-
nation in psychology is quite long.10 Our metaphysicalC system suggests an
appropriation of non-basic sources of epistemic justification like the spe-
cial science that is psychology. Because beliefs are indispensable to the best
psychological explanations, truth depends on being (cf. chapter 1), and the
relevant best explanations are true when such belief-laden explanations
are best, we should commit to the existence of those beliefs referenced in
those explanations. This is an instance of a special scientific indispensability
argument.
The eliminativists about beliefs might retort that they can explain all the
relevant data from empirical psychology with physical neural structures in
the physical brain. However, it is well-known that even the most sophis-
ticated of eliminativist programs provide only speculative and piece-meal
descriptions of the relevant replacement neural structures (see my response
to Churchland in chapter 1: note 2). For example, the eliminativist con-
nectionism of Ramsey, Stich, and Garon was advanced as a speculative
hypothesis, with very few detailed descriptions about the underlying neural
structures (Ramsey, Stich, and Garon 1990, 500).11 One of the most promi-
nent defenders of eliminativism about belief, Paul Churchland, constructed
merely (quoting William Hasker’s description) “science-fictional futures
in which our present conception of the mental and of cognitive processes
has been replaced by something quite different,”12 viz., the aforementioned
neural structures (P. Churchland, Postscript 1995, 170, where he employs a
placeholder term ‘Übersätze’ for part of the alternative eliminativist system
hoped for). The suggestion that the best special scientific explanations of the
relevant data should be couched in terms friendly to eliminative materialism
is plausible, only if we actually have some reasonably clear and established
ideas about the types of neural structures that replace the beliefs to which
our best psychological explanations appeal.13
Reductionism about beliefs will suggest that there is a type of ontologi-
cal reduction relationship between beliefs and their reductive bases. Token
instances of that reductive relation type that hold between beliefs and
their reductive bases will underwrite reductive explanations of beliefs. The
explananda in such reductive explanations will report on the existence of
beliefs, and because (even reductive) explanations are factive, that report
will be true. Thus, even if we replaced beliefs in our best psychological
explanations with their reductive bases (given reductionism about beliefs),
we would have all we needed to justify belief in the existence of beliefs.
Reducing beliefs does not eliminate them.
Premise (2) claims that all actual beliefs are formed by cognizers. That
statement is difficult to deny. It seems to be part of the very nature of beliefs
that they are possessed and formed by cognizers. For what is a belief that p
if not a cognizer’s conviction that p, or a cognizer’s mental assent to p, or a
cognizer’s propositional attitude with the content that p (see Schwitzgebel
2015, and Connors and Halligan 2015, 1–2). We truthfully speak of beliefs
as belonging to or being possessed by cognizers (i.e., we use the possessive
form of the term ‘cognizer’) precisely because cognizers retain and form
the relevant beliefs. Even those (such as Carruthers 1992) who argue that
there are innate beliefs will not deny that there are faculties of cognizers
responsible for innate belief production and that when those faculties come
into being and function appropriately, beliefs result (see ibid., 113, and the
discussion of evolution). No one, so far as I’m aware, maintains that beliefs
pop into existence ex nihilo, and yet somehow are retained by and correctly
attributable to cognizers. I therefore believe that it is not only a consequence
of our best account of the nature of beliefs that beliefs are formed, but that
it is also a deliverance of intuition that beliefs are formed. That claim seems
to me to be almost obvious.
Premise (3) states that the existence of beliefs, and the further fact that
beliefs are formed, materially implies that there are obtaining causal rela-
tions. Grant the antecedent of (3). To see that the consequent of (3) now
follows requires only the realization that formations are causal phenomena.
When a cognizer forms a thought, they relate to the thought through cau-
sation. When a cognizer forms a desire, they cause (perhaps together with
other factors) the desire. Likewise, in non-psychological contexts, when a
person forms a statue, she produces/causes the statue. When a galaxy is
formed, it is caused to have the features it does by antecedent conditions
operating in accordance with (perhaps causal) laws.14 When a crater forms
on the Moon, it does so due (causally) to fast-traveling meteorites. These
platitudes or examples are meant to generate a seeming, viz., that ‘forma-
tion’ is a causal term and that (more importantly) the phenomenon of for-
mation is a causal one. Our reasoning could be put in terms of an IBE.
The best explanation for the truth of the aforementioned claims about gal-
axy formation, crater formation, statue formation, thought formation, and
belief formation should include the admission that formations are causal
phenomena.
Conclusion (4) follows logically from the preceding premises. We have
good reasons then for affirming the existence of obtaining causal relations.
I now turn to the task of defusing arguments for causal eliminativism.
Section 3: The Causal Eliminativists and Physics

D. H. Mellor follows Mach (1920; 1923), Russell (1912–1913), and Schlick
(1979) by seeking to use physical considerations to show that there are no
obtaining causal relations (i.e., that causal eliminativism is true).15 In chap-
ter 1: sect. 4.5.5.5, I called this tradition neo-Russellianism. Mellor’s specific
case for it starts with the claim that necessarily, if causation is a relation,
then it is a factual relation.16 A factual relation or property is one that is
descriptive and therefore non-evaluative. Mellor uses the following criterion
for membership into the set of factual properties and relations,
(Mellor’s Criterion (MC)): “[T]he factual properties and relations that

exist are those that occur in laws of nature.”17
However, “no such law includes causation itself as a property or relation.”18

Therefore, the causal relation does not exist. The view that causation must
be understood as a relation-instance in the world is dismissed by Mellor as
a “formal prejudice.”19
MC is better (for Mellor’s purposes) stated in such a way that it is
restricted to the laws of physics. There are plenty of laws in the higher-level
sciences that are couched in essentially causal terms and would seem to
therefore require causal relations or properties (see, e.g., Lange 2014, 235,
who gestures at a reactive, and therefore causal law of chemistry). Mellor
is best interpreted as espousing the following argument from physics for
causal eliminativism,
(5) If the essential contents of the complete ideologies20 of our best and
most empirically successful physical theories do not include the notion
of causation, then there are no causal relations. [Premise]
(6) The essential contents of the complete ideologies of our best and most
empirically successful physical theories do not include the notion of
causation. [Premise]
(7) Therefore, there are no causal relations. [Conclusion]
The lead premise depends upon the truth of a much broader principle, viz.,
that,
(5*) For any notion n that is purportedly descriptive, if the essential contents
of the complete ideologies of our best and most empirically successful
physical theories do not include n, then the intended referent of n does
not exist.
Premise (5) is ad hoc without (5*) residing in the background. What is par-
ticularly special about causation, such that the absence of the notion of
causation from physical theory suggests its non-existence? Surely the same
thing would need to be true of other purportedly descriptive/factual rela-
tions and properties, as Mellor himself seems to indicate (Mellor 1997, 260;
1999, 185–199).
Statement (5*) bears ugly fruit. There are a great many purportedly
descriptive notions that we make use of for the purposes of referring to
ourselves, and to our mental states and mental events. However, none of
these notions show up in the ideologies of our best physical theories.21 In
other words,
(5**) The essential contents of the complete ideologies of our best and most
empirically successful physical theories do not include purportedly
descriptive notions such as ‘belief,’ ‘beliefs,’ ‘mental state(s),’ ‘mental
event(s).’
But (5*) together with (5**) implies eliminative materialism (EM), the the-
sis that there are no mental states or mental events. However, if there are no
mental states, then there are no beliefs. But I have already argued in sect.
2 that there are in fact beliefs. Thus, the motivating principle that is (5*) is
false because (5**) holds, and we are left with no good reasons to believe
premise (5).
Things are worse for the neo-Russellian who adopts Mellor’s argument
and (5*). The essential contents of the complete ideologies of our best and
most empirically successful physical theories do not include any notions that
are indicative of any of the tokens or types of replacement (for mental states
such as beliefs and the like) neural structures to which Paul Churchland
(Materialism 1995) (cf. the discussion in P. Churchland 1998) and others
allude so as to rescue EM. The same complete ideologies fail to incorporate
notions indicative of the functional components to which Daniel Dennett
(1996) reduces or identifies consciousness.22 We therefore have,
(5***) The essential contents of the complete ideologies of our best and
most empirically successful physical theories do not include any
notions indicative of physical neural structures or brain states.
Propositions (5*) and (5***) suggest there are no brain states either. That is
to say, (5*) together with (5**) and (5***) implies,
(8) Mental states, mental events, brain states, and physical neural struc-
tures do not exist.
This brand of hyper-eliminative materialism (hyper-EM) seems quite dif-

ficult to defend.
Section 3.1: Getting Causation Into Physics: A New Argument

Set aside worries about hyper-EM. The weaker position that is EM coupled
with functionalism will not aid the neo-Russellian interested in providing a
functional characterization of the mental that is somehow consistent with
the complete elimination of it. This is because a general functionalism about
the mental says “that every mental event-type can be fully characterized
by means of its typical causal connections to sensory ‘inputs,’ behavioral
‘outputs,’ and other mental event-types.”23 Functionalism about the mental
requires causation.24
David Lewis’s specific brand of functionalism will not help either. He said
that mental notions are theoretical terms of psychology (D.K. Lewis 1972).
Those terms have functional definitions, and their proper place is within
Ramsey-sentences that assign roles that should be satisfied/met/played by
appropriate physical surrogates (i.e., proxies of the referents of appropri-
ate folk psychological notions).25 The roles that are specified by Ramsey-
sentences are causal roles. This is because, for Lewis, “theoretical terms . . .
are definable functionally, by reference to causal roles.”26 The causal elimi-
nativist who hopes to save the mental cannot appropriate Lewis’s brand
of functionalism because according to causal eliminativism, there do not
exist any obtaining causal relations, and so there are no causal roles. But
the point I’m after is the more general one. Causal eliminativists cannot be
functionalists at all. The neo-Russellian willing to let the mental go and
completely embrace EM will point out that EM entails that nothing plays
the appropriate causal roles in the case of functionally explaining or defin-
ing the mental. That fact is consistent with causal eliminativism. However,
there should exist theoretical terms in physics that do receive functional role
assignments in appropriate Ramsey-sentences if the project of functional
reduction, explanation, and/or analysis is viable in physical inquiry. Thus,
recognizing that EM is a costly consequence of the argument from physics
for causal eliminativism has brought us a choice means whereby one can
inject causation into physics and thereby object to premise (6) of the argu-
ment from physics for causal eliminativism.
Are there instances of functional reduction and/or explanation in our best
physical theories? I believe so. One good case is all that is needed to defeat
premise (6). That one good case comes from classical Boltzmannian statistical
mechanics, although the story I’m about to tell would not vary too much if we
used quantum statistical mechanics.27 If one has no background in the physics
of thermodynamics and statistical mechanics, what follows may be difficult.
However, the key to understanding my discussion is to see that thermody-
namics and statistical mechanics, as standardly interpreted, crucially involve
a realization relation, and that that realization relation has a metaphysical
nature that requires the causal. So, causation enters statistical mechanics and
thermodynamics. The notion enters the ideologies of those theories.
Classical Boltzmannian statistical mechanics (SM) utilizes an abstract
mathematical space called a phase space (represented by the symbol Γγ). This
space has 6N-dimensions, where N is the number of particles in the statisti-
cal mechanical system (SY) with which one is concerned. Points of this space
represent possible microstates of SY because they give the momenta and
positions of all of the particles composing such states at respective times.
Or, more technically, X, a point that is a member of Γγ, equals in value (q1,
p1, q2, p2, q3, p3, . . . , qn, pn), where q1, and p1 (for a common example), give
the position and momentum of particle number one in the microstate of SY
at a time (Callender 2011). SM uses the dynamics of classical Hamiltonian
mechanics to give the evolutions of our choice system from t to later times,
although that dynamics is time-reversal invariant. This requires that one
use a Hamiltonian operator H(X) that represents the total energy of the
microstate of SY at a time. Possible evolutions of the possible microstates of
SY are modeled in Γγ with curves that foliate the phase space.
It is common to restrict the mathematical underpinnings of SM to the
energy hypersurface of Γγ, given that the total energy of SY is conserved.
I will assume SY is energetically isolated and total energy is conserved, and
so the entire 6N-dimensional space is not needed. We can model the dynam-
ics and the like with a 6N-1 dimensional hypersurface of Γγ, represented by
ΓE (following Callender 2011, 88). Carve up coarse-grained sub-regions of
ΓE. These regions will group together possible microstates of SY that (quot-
ing Penrose) realize physical “states that are macroscopically indistinguish-
able.”28 That is to say, the microstates of SY represented by points in one
of these sub-regions stand in a distinctive metaphysical relationship to a
distinguished and unique macrostate of SY. The macrostates of SY that each
relate in some metaphysically distinguished way to what the points in a sub-
region of the phase space represent would be (if they were actual) higher-level
concreta that exemplify phenomenological and/or derived properties that are
the object of study in thermodynamics. These properties include tempera-
ture, pressure, and entropy (see Klein and Nellis 2012, 4–6). The metaphysi-
cal relationship in question is commonly assumed to be supervenience, as
Roman Frigg remarked,
It is one of the basic posits of the Boltzmann approach that a system’s

macro-state supervenes on its fine-grained micro-state, meaning that a
change in the macro-state must be accompanied by a change in the
fine-grained micro-state (i.e. it is not possible, say, that the pressure of
a system changes while its fine-grained micro-state remains the same).29
But virtually everyone agrees that supervenience relations have explana-

tions, or what amounts to the same thing, all supervenience relations are
superdupervenience relations in the sense that they have explanations for
why they obtain.30 This is because the macrostate of SY (call it MSY) isn’t just
correlated with a certain microstate in the way suggested by supervenience
(i.e., it isn’t merely the case that if one varies MSY, one gets a variation of its
corresponding microstate). Rather, MSY depends in some robust ontologi-
cal sense on a microstate at t. There is some way the microstate of SY is at
t, such that higher-level properties instantiated in MSY become instantiated
by the higher-level macrostate by virtue of the micro-goings-on. This type
of ontological dependence is not supervenience,31 although it explains why
supervenience holds.
Can we say more? Is there a certain brand of ontological dependence on
display in Boltzmannian statistical mechanics? Yes. The relation is com-
monly said to be one of metaphysical realization.32 What is metaphysi-
cal realization? Good question. There’s debate about its precise nature.33
Roughly put, realization is a synchronic, many-one, dependence relation
between the constituents of a physical system or individual and the prop-
erties of the entire system or individual that (quoting Polger and Shapiro
2016, 22) “transmits physical legitimacy from physical realizers to what is
realized.”34 Here is Carl Gillett’s more detailed specification of the general
structure of a realization relation,
Property instances F1–Fn, in individuals s1-sm realize a property instance

G, in individual s* under background conditions $, if and only if, under
$, (a) s1-sm are members of, or are identical to, a group of individu-
als s1-sn spatially contained within s*, (b) s1-sm bear spatiotemporal,
productive, and/or powerful relations to one another, (c) s1-sn through
their joint productive role-filling together non-productively result in s*
under $, but not vice versa, (d) the powers contributed by F1–Fn to
s1-sm together through their joint productive role-filling non-produc-
tively result in the powers individuative of G, in s* under $, but not vice
versa, and (e) the processes based by F1–Fn under $ are or would jointly
non-productively result in all the processes that are or would be based
by G under $ but not vice versa.35
It is a feature of realization, as Gillett has characterized it, that the constituents

of the system directly involved in property instances F1–Fn, stand in “power-
ful relations to one another” and satisfy “joint productive role-filling” (ibid.;
emphasis mine). This talk is causal talk. Other ways of filling in the details
are likewise explicitly causal. For example, Polger and Shapiro (2016, 22–23)
regard the realization relation as one involving the possession of a function,
where a function is thought of in terms of a causal profile.36 Sydney Shoemaker
(2007, 12) said that realization involves backward- and forward- looking
causal features (see also Humphreys 2009, 640, on multiple realization).
How are realization relations, and microstates and macrostates of SM,
related to functional reductions and the case for causation in physics? Many
macroscopic thermodynamic (the higher-level science) systems reduce to
microscopic systems described by statistical mechanics. Not only that, but
the relationship between the higher-level and lower-level systems is explained
by the science of statistical mechanics.37 This can be seen in the standard
story about how entropy, a property of the macrostate of SY (or MSY), is
reductively (functionally) explained in terms of certain ways the microstate
of SY are or could be.
Add a measure over ΓE to make sense of various sizes of coarse-grained
regions, although each region contains infinitely many points. The measure
I have in mind is the Liouville measure μ, which is the “Lebesgue mea-
sure defined over canonical coordinates” (North 2011, 321, n. 21). With
that measure in hand, the coarse-grained regions introduced previously can
be understood as volumes of the energy surface (e.g., μ(MSY)). The largest,
according to the measure, volume of the space is that coarse-grained region
that corresponds to a macroscopic state of SY in equilibrium. We owe this
( )
insight to Boltzmann (1909). Entropy becomes S = klog µ Γ MSY . It follows
from this equation with appropriate interpretive assumptions that the mac-
rostate MSY that is most highly entropic (i.e., it is in equilibrium) is one
about which it is true that the coarse-grained region with the largest volume
on the energy surface of the phase space just is that coarse-grained region
in which there are the most amount of microstates/points (according to the
measure) that can realize MSY (the relation will hold even if we added explic-
itly to the discussion logarithms of volumes). It should be obvious now that
macrostates like MSY are multiply realizable. There is more than one micro-
state that can realize that self-same macrostate, and entropy is (roughly)
here being connected to the number of such states that can realize MSY.
The metaphysics of realization enters the very characterization of entropy
in thermodynamics and statistical mechanics, and the higher-level property
that is entropy is functionally reduced (via realization) to micro-goings-on.
That the reduction is functional is bolstered by the fact that the laws of func-
tional reduction use realization relations to underwrite them (see Kim 2008,
103–104). But again, realization involves causation, or causal powers, or
causal profiles. The thesis that causation does not enter physics is at odds
with a standard Boltzmannian physical story about the relationship between
thermodynamics and statistical mechanics.
Some may retort that thermodynamics and thermodynamic properties
are not fundamental. Entropy is therefore not fundamental. Realization
relations are therefore not fundamental. Mellor’s argument should have
been restricted to empirically successful fundamental physical theories. In
response, I note that even if we restricted argument (5)–(7) to the ideolo-
gies of our best fundamental physical theories, the same argument from SM
(although we would want to appropriate the quantum SM story I described
as not departing very far from the classical SM one) would run. This is
because my argument is not that the realization relation is causal, but that
just about every metaphysical theory of that relation attributes causal prop-
erties or causal profiles to the realizers of the realization relation. The real-
izers belong to the fundamental theory (in this case, quantum SM). So long
as we allow for the existence of thermodynamic systems involving mac-
roscopic states bearing entropy, we will have reason to believe in the real
existence of the realization relations, and so also the “joint productive role
filling,” or causally profiled realizers.
Section 4: Lowe’s Case Against the Causal Relation

Like the neo-Russellians, E. J. Lowe argued that there are no obtaining
causal relations. Unlike the neo-Russellians, his justification did not depend
on physical considerations. Rather, Lowe surmised that one could make true
causal statements with dispositions, powers, capacities, liabilities, and their
manifestations alone. Regarding the question, what makes the causal state-
ment <The water dissolved the salt.> true?38 Lowe remarked,
[C]ertain causal powers and liabilities, and their respective manifesta-

tions, make such a proposition true. . . . ‘The water is causing the salt
to dissolve’ can be true only if the water has a power to dissolve salt
and the salt a liability to be dissolved by water. But its truth evidently
also requires those powers to be manifested on the particular occasion
in question.39
For Lowe, powers and dispositions are universals that are monadic proper-
ties inhering in concrete particulars that are intentional in that they, like
mental states (quoting Lowe), “are ‘directed’ at other objects of various
kinds, but don’t require the existence of those objects.”40 Water’s causal
power manifestation that is dissolving salt (in the relevant actual circum-
stance) depends asymmetrically for its existence upon water’s monadic
property that is a causal power (a non-manifestation). Necessarily, if the
causal power of water to dissolve salt (non-manifestation) fails to exist, then
the manifestation of that power that is salt’s dissolving in the relevant actual
circumstance fails to exist. However, the converse of this entailment relation
does not hold (hence asymmetric existential dependence). What “connects”
water’s causal power manifestation to the manifestation involving the liabil-
ity of the salt is the fact that the manifestation of water’s causal power is a
monadic property of the salt, “namely, its dissolving on this particular occa-
sion.”41 Thus, Lowe remarks, “if the salt’s dissolving is a manifestation of
the water’s power, then it is so essentially and hence necessarily.”42
There are several problems with Lowe’s picture. Before laying out the
first problem, note that Lowe’s theory would be unnecessarily complicated
if salt’s liability manifestation (call it M-LP) and water’s causal power mani-
festation (M-CP) in this case are not identical. For if they are not identical,
then one would wonder why. M-LP just is the manifestation instance of
salt’s dissolving in water on the relevant occasion. Likewise, M-CP just is
salt’s dissolving in water on the relevant specific occasion. Why do we need
two distinct things here and not just one thing?
If M-LP is identical to M-CP, then M-CP will asymmetrically existentially
depend for its existence upon the inhering liability (power) of salt to dissolve in
water (call it LSP). At every world at which you take away LSP, you must take
away M-CP. The dependence is asymmetric because there is a possible world
at which salt has LSP, although it is never manifested (perhaps like the fragility
of a vase that never breaks). But now it seems that salt’s liability instance asym-
metrically existentially depends upon two things, viz., water’s causal power
(call it CWP) and LSP. We have shown the following (where locutions of the
type ‘∼CWP’ mean that the relevant causal power fails to exist or occur, or
obtain, and we could substitute ‘M-LP’ for the occurrences of ‘M-CP’ below),
(9)  (~ C W P → ~ M − CP )
and
(10)  (~ LS P → ~ M − CP )
Supplementing Lowe’s account with the preceding details does not save
it from significant problems. To see this, consider the following questions.
Why does M-CP occur when it does? Why does that manifestation come
into being “on this particular occasion” (as Lowe put it), or at the relevant
time? Salt does not always dissolve in water after the same interval of time,
or on the same occasions. Hot water dissolves salt more quickly than cold
water does, and various environmental conditions may affect matters as
well. Manifestations of water’s causal power at varying times and in vary-
ing environments seem to constitute differing manifestation instances of the
water’s causal power to dissolve salt because Lowe’s account asserts that
such manifestations are occasion-sensitive or circumstance-sensitive (Lowe
2016, 108). The causal power that is CWP must serve as the possible asym-
metrical existential dependency base for a great multitude of manifesta-
tions unless we are willing to multiply causal powers by a significant factor.
Assuming for now that the latter disjunct is implausible, we may ask, why
is it that CWP results in M-CP and not a different manifestation at a slightly
different time (call it M-CP*)? Lowe’s account is incomplete.
If one were to multiply causal powers and suggest that each of the many
causal powers water possesses is a possible unique asymmetric existen-
tial dependency base for a possible unique causal power manifestation,
one would be multiplying entities beyond necessity. Employing a single
fundamental causal relation is more parsimonious than positing infinitely
or nearly infinitely many causal powers to handle as simple a case as table
salt dissolving in water.
The second problem I find with the account is that it is unscientific. Salt
dissolving in water should have a perfectly respectable scientific explana-
tion. But now it seems that there is clearly more to the story than Lowe
has articulated. The truthmaker for <The water dissolved the salt.> should
report on those entities to which the correct chemical and hydrodynamical
explanations of the dissolution appeal. <The water dissolved the salt.> is
true because of the chemical composition of water H2O; the chemical com-
position of salt NaCL; the natures of the involved ionic chemical bonds;
the amount of energy in or coming into, or leaving the system; and charges
exemplified by the involved molecules, plus the chemical interactions (which
are causal) between the molecules as given by the governing or backing
chemical and hydrodynamical laws.43 Thus, Lowe’s proposed truthmaker
misses out on entities and structure indispensable to a best metaphysical
explanation of <The water dissolved the salt.>.
Here is a third problem. What does it mean to say that a disposition or
causal power is intentionally directed toward a non-existent manifestation
or entity? U. T. Place argued that the involved type of intentionality is the
same type of intentionality that is characteristic of the mental (Place 1996).
We can and do have mental states that are directed toward things that fail
to exist.44
I have three reasons for not associating causal powers and dispositions
with intentionality. First, the maneuver seems to me to involve a category
mistake. Dispositions are not the types of entities that are about anything,
and if an account of intentionality suggested as much, I would surmise that
a revision of that account was in order for fear of panpsychism and related
views (see on this complaint, Mumford 1999). Second, even if we were to
grant such entities intentionality, what would explain (leaving the water
and salt example) the fact that fragility in a vase is about or directed toward
its non-existent breaking? By virtue of what, in other words, is that vase’s
fragility about or representative of its non-existent liability-instance (i.e.,
its breaking)? This is an important question in the context of the mental,
and even in the context of the representational properties of propositions.
Plausible answers are hard to come by in the case of dispositions, although
there are some decent answers in the context of the study of propositions.
Third, Alexander Bird (2007, 118–126) has persuasively argued that dispo-
sitions do not actually satisfy conditions and criteria ordinarily indicative of
the intentional. To give just one example among many, Anscombe (1968)
argued that intentional mental states like thoughts are indeterminate. I can
think of Abraham Lincoln during the Gettysburg Address quite apart from
thinking that he has a particular number of eyelashes. Understood as the
object of my thoughts, Lincoln is an indeterminate entity, although the Lin-
coln of November 19, 1863 is not (Bird 2007, 120). Dispositions and their
manifestations are not indeterminate in this way, if we assume that deter-
minism holds. For as Bird went on to explain,
Given determinism, if the vase is dropped in a particular way or struck

in some specific manner, there is only one fully determinate way it will
in fact break. We might wish to say that the fragility of the vase points
to an infinite number of determinate manifestations, each conditional
on a specific implementation of the stimulus. But an infinity of fully
determinate objects is not the same as a single indeterminate one.45
To rescue the idea that dispositions or causal powers are in some way
directed toward their instances, one could introduce some type of non-
causal, non-intentional relation between the disposition that is fragility and
the vase’s liability-instance (i.e., it’s being broken or breaking). The problem
is that powers and dispositions can exist without manifestations. Again,
our vase is fragile, although it has not broken yet. Triggering conditions are
required for the liability-instance (the breaking) to occur. But how then can
powers or dispositions be non-causally related to their manifestations by
being connected to them if the manifestations do not exist? (This problem
is known and written about in Mumford 2009; Place 1996; and Molnar
2003). One response suggests that there are both types and tokens of dispo-
sitions, and that there are both types and tokens of their manifestations. The
disposition-types are joined to manifestation-types by a non-causal, non-
intentional necessary connection (modifying Mumford 2009). But how can
one be an immanent realist about powers or disposition and manifestation-
types on this reply to the problem? It seems one can’t. Manifestation tokens
do not always exist, so with what tokens does one associate their types? The
involved types must therefore be abstracta, and that seems like a problem
for a tradition steeped in Aristotelianism.
Stephen Mumford’s functionalist approach to dispositions in Mumford
(1999) does not help, for it is problematic at worse, and incomplete at best.46
(I will explain the view as I evaluate it.) First, that it is incomplete: Accord-
ing to Mumford’s functionalism, that dispositional properties are functional
properties is underwritten by the fact that dispositions stand in “the rela-
tion of causal mediation” (Mumford 1999, 223) to event-types appropriately
related to manifestations and triggering conditions for those manifestations.
But what is a relation of causal mediation? Moreover, how does it caus-
ally mediate between dispositions that exist and are tokened, and event-types
such as manifestation-types and triggering condition-types that have no
token instances? Does it connect the causal power token to a liability-type
by (perhaps with the triggering conditions) making/producing the liability
token? If so, it looks like we have here the causal relation itself. And like
Lowe’s response, the view seems problematic for Aristotelians who think of
all universals as immanent entities, located in or near their token instances.47
Second, that it is problematic: The claim that a disposition d is functional
is said to be partially characterized in terms of the further claim that “it is a
conceptual truth that d causally mediates from stimulus events to manifes-
tation events” (Mumford 1999, 223). This appears to be much too strong.
Does it really follow that those who fail to (a) characterize dispositions in
functional terms or that those who (b) fail to associate dispositions with
causal mediation are conceptually confused? Some argument for crown-
ing the relevant truth with conceptually necessary status seems required. In
addition, if the mediation is itself causal, and one wants to use dispositions
to account for causal relations of all sorts (in the spirit of Lowe who wants
to do without all causal relations), then how is it that one unproblematically
accounts for that causal mediation in terms of further functional/disposi-
tional states that would require yet more causal mediation?
I conclude then that the metaphysics of causal powers, dispositions, and
liabilities provides no good justification for doing without the causal rela-
tion, pace Lowe.
Section 5: Lewis and the Void

I maintain that necessarily, all instances of causation are instances of an
obtaining causal relation.48 But Lewis affirmed that the void, a veritable
absence of everything situated somewhere between points of some possible
space-time manifold, can causally produce effects,49 stating that “[t]he void
is deadly. If you were cast into a void, it would cause you to die in just a few
minutes.”50 But because the void is by nature an absence of everything, it
cannot be a causal relatum. Thus, when the void brings about some effect at
a possible world w, it does so without entailing the obtaining of a relation
at w. Therefore, it is not the case that necessarily, all instances of causation
are instances of an obtaining causal relation.
Lewis affirmed the following modal combinatorial principle,
(11) One “can generate new possibilities by patching together (copies of)
parts of other possibilities.”51 [Premise]
Lewis’s argument for the possibility of the void used (11) in the following
way,
(12) If (space-time substantivalism is possible and space-time relationalism

is possible, and (11) holds), then the void is possible. [Premise]
(13) Space-time substantivalism is possible and space-time relationalism is
possible. [Premise]
(14) Therefore, the void is possible. [Conclusion]
Lewis’s argument for (12) is as follows,
[I]f a relationist world is possible, and a world full of substantival

space-time is likewise possible, then by patching together parts of these
two worlds, we get a world that consists of substantival space-time
interrupted by occasional voids. The walls and the space-time within
them are distinct existences; ergo it is possible for either one to exist
without the other. If the walls exist without the space-time (and without
any other objects between the walls) then there is a void between the
walls.52
Even given that the void is possible, it is unclear how the void can have
properties of any kind (i.e., it is unclear how the void can be deadly for
example). Strictly speaking, the void does not exist at any metaphysically
possible world. It is by its nature not an entity to be reductively or non-
reductively reified by Lewis’s lights (D.K. Lewis, Void and Object 2004,
281–282). Therefore, the void cannot stand in any relations whatsoever
because all relations require relata. But that means the void cannot exem-
plify properties or universals, if exemplification is a relation. If it is not a
relation for reasons having to do with Bradley’s regress, note that it can-
not be a member in a nexus of exemplification, nor can it bear properties
(as some proponents of Platonic properties maintain; Moreland 2001). It
cannot be that in which universals inhere (as in the immanent universalism
of Armstrong 1997). It cannot stand in the relation of being the member
of any set (as in Lewis’s class nominalism; D.K. Lewis, Plurality 1986),
nor can it be a trope, or have a trope as a proper part, or be structured by
tropes (as in Williams 1953), for if it were so constructed it would exist
as a structure. Nor can the void resemble any particular (as in the resem-
blance nominalism of Rodriguez-Pereyra 2002) because resemblance is a
relation.53 There are no (not obviously false) accounts of properties and
their relationship to particulars that underwrite the claim that the void is
deadly. We cannot truthfully attribute any positive properties to the void,
for there appears to be no way of making that attribution-talk true. There
is nothing upon which such would-be truthful attributions would supervene
or more generally depend (in the sense of the TDB of chapter 1: sect. 1).
This is a problem for Lewis’s more general philosophical system because he
affirmed that “truths must have things as their subject matter.”54 The void
isn’t a thing. It cannot have properties. It therefore cannot have causal pow-
ers either. Thus, even if we accept the void as possible, it will be unable to
cause events, it will be unable to be deadly.
My second problem with Lewis’s reasoning is that conclusion (14) faces a
rebutting defeater, which shows that it is either false or not truth-evaluable.
Once again, grant Lewis’s view that the void’s nature is such that it pre-
cludes it from standing in any relations. It appears to follow that we can
use no pieces of language to refer to the void because reference is a relation
between referrer (i.e., some representation token) and referent. The definite
description ‘the void’ cannot be used to refer to any such absence. One
could follow Russell (1905) and the Russellian school on definite descrip-
tions by arguing that they, like ‘the void,’ are not referential. But Russell’s
view will also entail that expressions featuring descriptions like ‘The void
is possible.’ come out meaningful although false. However, Lewis wanted
statements like (14) to come out literally and non-fictionally true.55
To illustrate the fact that theories of reference preclude successful refer-
ence to the void, consider the causal theory of reference promulgated in
Kripke (1980) and taken on as an assumption in chapter 1. Assume ‘Void’ is
a word intended to name Lewis’s Void. To refer to Void in the way the causal
theory of reference suggests, we need an initial baptism whereby the term
‘Void’ becomes initially associated with Lewis’s Void. There are standardly
two ways baptisms transpire: (a) perceptual experience, or (b) description.
Given option (a), we face the difficult question: How does one actually per-
ceive (a merely possible) Void? But things are worse. How is it possible to
perceive Void? On a great many theories of perception, that phenomenon
is a relation, or else some thing or entity is required to produce that which
relates to the apparatus of the perceiver. But again, Void cannot stand in any
relations. It is difficult then to see how one can perceive it.56
Second, it is likewise difficult to see how the initial baptism can transpire
by the use of true descriptions. For as has already been argued, there is no
(decent) theory of properties according to which Void can be said to have
them, or possess them, or be an aggregate of them. Theories of properties
seem to have need of the existence of the possessor, exemplifier, or aggre-
gate. Something is required to be the member of a class (where membership
is a relation). But again, Void does not exist, and cannot stand in any rela-
tions. And again, there is nothing upon which the relevant true descriptions
can depend for their truth. If that is right, there’s no avenue to initial bap-
tism that travels through a true description. So, we have some reason then
for believing that we cannot refer, either by description or by way of a name,
to Lewis’s void. But if we cannot refer to the void or Void, then it is unclear
how sentence (14) can be true. We should therefore refrain from affirming it.
I have argued for the claim that (i) there are some obtaining causal relations,
and I have sought to defend that claim against objections. I have also tried
to defend the thesis that (ii) necessarily, all instances of causation are obtain-
ing causal relations. We have seen that considerations having to do with
causation’s absence from physical theorizing (even if granted) privilege the
deliverances of physics too much. We should not infer that some entity does
not exist because it does not figure in the laws of our best physical theo-
ries. The argument from physics for causal eliminativism does not defeat
(i). Lowe’s considerations having to do with dispositions also fail to provide
defeaters for (i). Lewis’s invocation of the void does not defeat (ii) because
the void is unable to have properties, and either success of void reference
is impossible, or else, claims about a possible void exerting causal influence
are problematic as such claims have no thing(s) upon which to depend for
their truth.
Notes
1. For a defense of the idea of Aristotelian concreta I am assuming here, see
(Loux 1978). I’m unsure of whether there are relational properties, and
I do not necessarily agree with Loux’s theory of attributes (Aristotelian sub-
stances include entities indispensable to our best physical theories (q.v., n. 4
of chapter 5)).
2. Compare this to Armstrong’s view that all relations relate particulars and that
“relations must have terms” (Armstrong, Theory of Universals 1978, 76). Or
Fodor’s affirmation “[c]ausation is, par excellence, a relation among particulars”
(Fodor 1990, 33; emphasis in the original). I settle on a theory of causal relata in
chapter 7: sect. 4.
3. This premise does not require anti-reductionism about beliefs. If beliefs reduce
to or are grounded in the non-mental, or some other suitable reductive surro-
gate, it would not follow that there are no beliefs. Reductionism about beliefs is
not eliminativism about beliefs.
4. The same point is applicable to the case in which the belief in question is in a
theory that asserts or entails that one should refrain from believing that there are
beliefs (i.e., agnosticism about beliefs).
5. There is strong empirical evidence for the thesis that delusional disorders are
not affective disorders or instances of schizophrenia. See Fear, Sharp, and Healy
(1996), who report that their data support this on p. 65. Interestingly, delusional
beliefs are often associated with neuropathological conditions or disorders such
as schizophrenia (Coltheart, Langdon, and McKay 2011).
6. See Ajuriaguerra, Strejilevitch, and Tissot (1963); Foley and Breslau (1982, “The
patient fails to recognize his or her own image in a mirror while recognizing and
properly identifying the images of other people in the mirror.” (ibid., 76)).
7. Connors and Coltheart, On the Behaviour of Senile Dementia Patients vis-à-vis
the Mirror: Ajuriaguerra, Strejilevitch, and Tissot (1963/2011), 1680.
8. See the studies of Coltheart (2010); Coltheart, Langdon, and McKay (2011);
Langdon and Coltheart (2000); and Maher (1974). The very Diagnostic and
Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) published by
the American Psychiatric Association in 2000, defines a delusion as,
A false belief on incorrect inference about external reality that is firmly sus-
tained despite what almost everyone else believes and despite what constitutes
incontrovertible and obvious proof or evidence to the contrary. The belief is
not one ordinarily accepted by other members of the person’s culture or sub-
culture (e.g. it is not an article of religious faith).
(ibid., 765; emphasis mine)
9. Halligan (2007, 358). See also Wade and Halligan (2004).
10. See Halligan (2007).
11. Interestingly, the brand of connectionism to which Ramsey, Stich, and Garon
commit themselves invokes causal systems. Causal eliminativists cannot appro-
priate it (see Smolensky 1988, 1, 15, on which Ramsey, Stich, and Garon depend
for their connectionist theory).
12. Hasker (1999, 14).
13. Compare Williamson’s remark,
To assume that the evidence can be reformulated without relevant loss in
ontologically neutral terms, in the absence of any actual such reformulation,
would be optimistic to the point of naïvety.
(Williamson 2007, 223)
14. I argue that causation does not reduce to non-causal law-governed physical his-
tory in chapter 3 and chapter 8. However, my argumentation in the main text
does not require the falsity of causal reductionism. Reductionists can affirm that
galaxy formation is causal and yet regard that causal phenomenon as ultimately
reducible to a non-causal one.
Given my argumentation in chapter 6: sect. 2.3 and sect. 3, the examples
I cite may be indicative of partial causation instead of full causation because
I do not cite or report on the entire rich causal structures responsible for the
mentioned effects. One can dodge this potential problem by simply enriching the
descriptions of the causes provided in the main text. The conclusion I’m after,
that formations are causal, will still come out true subsequent to such enriching.
Many of the examples of obtaining causal relations that I report on through-
out my discussions in this work will be somewhat shorthand representations of
the full, far richer causal structure that involves electromagnetic, gravitational,
and other influences. My examples will report mostly on just those causes one
would ordinarily require in any specification of an obvious and intuitive causal
explanation of the fact that the effect(s) occurred. But this is solely for brevity’s
sake, and if my argumentation can be affected in some way by the brevity of
presentation, I will seek to make most or all of the riches of the involved causal
structure transparent.
15. According to Mellor, there are instances of causal explanation. Mellor (2004,
322) is also open to the existence of causal powers that help necessitate the
truth of conditionals that are connected to true causal claims. However, causa-
tion does not reduce to the dependence relation those true conditionals repre-
sent. Mellor’s view is similar to that of Michael Scriven’s (1975, 11), who said
that “[a] cause is an explanatory factor (of a particular kind). Causation is the
relation between explanatory factors . . . and what they explain.” The differ-
ence between Mellor and Scriven appears to be that Scriven identifies the causal
relation with an explanatory one, whereas Mellor eliminates the causal relation
altogether, making room only for causal explanation.
16. Mellor (2004, 323, n. 9).
17. Ibid., 319, cf. ibid., 323, n. 9. Elsewhere, Mellor (1997, 260; 1999, 185–199)
argues that in order for something to be a property, it must figure in the laws.
18. Mellor (2004, 319).
19. Ibid., 318.
20. A complete ideology of a physical theory T is composed of the indispensable
notions used to represent the various members of the ontology of the full inter-
pretation of T together with the ideology of T’s partial interpretation. Q.v.,
chapter 1: sect. 4.5.5.3.
21. Compare the point made by Zimmerman (2008, 219–220) in the context of the
ability of physical theory to recognize a distinguished present moment.
22. All of the notions I reference in this paragraph are obviously purportedly descrip-
tive. I will ignore this qualification going forward.
Several authors have named Dennett an eliminativist (e.g., Block 1992; Seager
2016, 56). I think this is incorrect (see the comments in Dennett and Kinsbourne
1995, 235); cf. Schneider (2007, 315, 322–323).
23. Horgan (1984, 321; emphasis mine).
24. Block (2007, 63); Bickle (2016, sect. 1.3) testify to this point.
Notice that functionalism about the mental also requires the existence of
mental states and mental event-types. That is inconsistent with EM. The neo-
Russellian in view commits multiple errors.
25. See the nice explication in Ramsey (2013, sect. 2.2).
26. D.K. Lewis (1972, 249). See also D.K. Lewis (1970); Ramsey (2013, sect. 2.2);
Schwarz (2015).
27. Both theories are empirically successful. My explication will follow Callender
(2011); Frigg (2008); Frigg and Werndl (2011); S. Goldstein (2001); Lebowitz
(1999); North (2011); Penrose (2005; 2010).
The reader may wonder what relevance quantum theory has to my project
since I’m providing a theory of deterministic singular causation, and many inter-
pretations of quantum mechanics (QM) are indeterministic. This project assumes
that there is a plausible deterministic full interpretation of both relativistic and
non-relativistic quantum mechanics available. The Bohm-de Broglie approach is
both deterministic and empirically equivalent to orthodox QM (Dürr, Goldstein,
and Zanghì 1992). Is it relativistic? There is important progress on that ques-
tion (Dürr et al., 2014). Bohmian mechanics is not the only option available.
Recall that the many-worlds interpretation is also fully deterministic (L. Vaid-
man 2016, sect. 4.1), and it is business as usual in quantum field theory given the
many-worlds approach. As I write this, I am currently developing a new (what
I call) interactive interpretation of QM (with apologies to Richard Healey) that
is also fully deterministic, relativistic, and robustly causal. More, hopefully, on
that approach at a later time.
28. Penrose (2005, 691).
29. Frigg (2008, 104).
30. See Horgan (1993, 566, 577–582); and especially McLaughlin and Bennett
(2014, sect. 3.7). The term ‘superdupervenience’ (not necessarily the definition)
comes from W. Lycan (1986, 92). Compare Simon Blackburn’s remark, “super-
venience is usually quite uninteresting by itself. What is interesting is the reason
why it holds” (Blackburn 1984, 186). Cf. K. Bennett (2017, 14, n. 10).
31. As has been stressed by Fine (1995; 2001); Kim (1998); and Polger and Shapiro
(2016, 20).
32. Callender (2011, 88); Loewer (2012, 122); Jill North writes, “[a] macrostate cor-
responds to a region in phase space, each point of which picks out a microstate
that realizes the macrostate” (North 2011, 321). Citations could be multiplied.
33. See, e.g., the discussions in K. Bennett (2011); Gillett (2016); and Polger and
Shapiro (2016).
34. It is explicitly described as a relation employed by the sciences (see Gillett
2016, 361).
35. Gillett (2016, 89; first emphasis is original, the rest are my emphasis).
36. Stuart Glennan asks, “what exactly is a functional role?” He answered, “In gen-
eral it is a causal role!” (both quotations are from Glennan 2017, 146; emphasis
in the original).
37. Indeed, many believe that thermodynamics (the theory) reduces to the theory of
statistical mechanics (e.g., Loewer, Physics 2008). Physicists standardly put it the
other way, statistical mechanics reduces to thermodynamics when the involved
degrees of freedom are sufficient in number. However, some doubt these stronger
claims (e.g., Sklar 1993, 333–374).
38. Lowe is a truthmaker theorist about a great many claims (see Lowe 2016, 100,
for at least a sufficient condition). Lowe defends a much more complicated and
full-blooded truthmaker theory in Lowe (2006, 192–210).
39. Lowe (2016, 107; emphasis in the original).
40. Ibid., 107; emphasis in the original.
41. Ibid. 108.
42. Ibid., 108; emphasis in the original.
43. What’s puzzling is that Lowe may have been aware of this given his comments in
Lowe (2006, 171).
44. See the commentary in Armstrong (1997, 79) and Bird (2007, 118–119).
45. Bird (2007, 124).
46. I do not discuss Mumford’s more recent account of powers in Mumford and
Anjum (2011) because according to that newer account, “[p]owers are produc-
tive of their manifestations, and production is clearly . . . causal” (ibid., 8).
Production is quite obviously a relation between producer and produced (at least
in my view). The view is at odds with the relevant thesis of Lowe’s paper (i.e.,
that there are no causal relations). Mumford and Anjum (2011, 106–129, 156)
do deny that causation is a relation, but I ask what makes a causal process that
distinctive process? There’s presumably some type of connection between parts
of the process (indeed, Mumford and Anjum speak of partnered causal powers
in causal processes). That connection constitutes a relation of some kind. But
Lowe thinks there (probably) are no relations (the very title of his paper).
47. To be fair, Mumford (1999, 224–225) confesses that his view is incomplete.
48. I will soon argue that one cannot successfully refer to the void. Thus, read
occurrences of the locution ‘the void,’ or ‘void’ or ‘Lewis’s Void’ as shorthand
for what Lewis intended to refer to by his use of ‘void’ in D. K. Lewis (Void and
Object 2004).
49. D. K. Lewis (Void and Object 2004, 281).
50. Ibid., 277.
51. Ibid., 278. Throughout most of Lewis’s paper he seems to be concerned with
conceptual possibility. But Lewis maintained that conceptual possibility just
is metaphysical possibility. There are two lines of evidence for this. First, he
equated broadly logical possibility (which is metaphysical possibility) with con-
ceptual possibility (ibid., 278). Second, in Lewis’s footnote to his discussion of
the conceptual possibility of the void, he appealed to his modal combinatorial
principles for motivation and cites his book (D. K. Lewis, Plurality 1986, 86–92).
In the cited discussion, the relevant combinatorial principles are principles about
metaphysical modality.
52. D. K. Lewis (Void and Object 2004, 278).
53. I ignore conceptualist views of properties for hardly anyone affirms them, and
the consensus view seems to be that any position that affirms that properties and
the like are just ideas in the head suffers from insurmountable difficulties (see on
these Russell, 1959).
54. D.K. Lewis (1999, 206; emphasis in the original).
55. Thus, we cannot use an intentional “according to the fiction” operator to help.
Lewis thought it is literally and non-fictionally true that the void could cause
things at the relevant worlds. The context of Lewis’s remarks were not fictional
contexts, nor was Lewis attempting to say something with irony, or via indirect
speech. Thus, the pretense theory (as in Recanati 2000 and others) is of no
benefit here either. We can go further, it seems to be a general truth that ways of
dealing with talk of fictional entities that do not reify those fictional entities will
not avail Lewis. But Lewis does not reify the void either.
56. You might think I’m begging the question here. Perhaps you could perceive Void
by perceiving an instance of causation wrought by Void. Can you perceive cau-
sation? A fortiori, can you perceive causation by a non-existent? I take it that
the objections from Hume and others against perceiving causation are rendered
more robust in the case of production by a non-existent. I happen to believe you
can perceive some instances of causation, but I see no way of running a story
that allows for perception of causal production by a non-existent. Causation
looks all the more theoretical and abstract in such a case.
I should remind the reader that even if one could perceive Void, Lewis was
arguing merely for the possibility of Void, not that Void actually exists as an item
in the ontology of the actual world that we can actually perceive. But if one can’t
actually perceive Void because Void does not actually exist, then one cannot
initially baptize Void so as to refer to it (sticking to option (a)).
3 The Brute Asymmetry of Causation
Chapter 1 provided the reader with a metaphysicalC system with which
to properly interpret and evaluate my construction of a target philosophi-
cal analysis and fundamental metaphysical theory of causation. Chapter 2
began the construction of that analysis and theory by arguing for or defend-
ing the following claims,
(a) There are instances of causation (i.e., causal realism is true).

(b) Necessarily, all instances of causation are instances of an obtaining
causal relation.
I will now build upon (a) and (b) by briefly arguing that causation is a
formally asymmetric relation and that causal directionality has no physical
reductive explanation.
Section 1.1: Motivation and Thesis

Virtually everyone1 in the causation literature maintains that causation is a
formally asymmetric relation, such that,
(1) ∀x∀y (Cxy → ~ Cyx)
That is to say, necessarily, for any event x, and for any event y, if x causes y,
then it is not the case that y causes x.2 Because I can hardly find anyone who
rejects the formal asymmetry of causation (although one should see Arm-
strong (1997, 206–207); and D.K. Lewis (Postscripts 1986, 213) for two
notable exceptions (I preclude Mumford and Anjum (2011, 156, 106–129)
because they reject the hypothesis that causation is a relation)), my case for
(1) will be brief. Consider first the empirical case for (1). Certainly, most
instances of causation we behold in everyday experience and in science sup-
port it. The match strike causes the match to catch fire, the flick of the finger
causes the domino to fall, sunlight and flagpole position cause the shadow’s
The Brute Asymmetry of Causation 111
appearance. With respect to each of these, and a great many more cases, we
are compelled to add, and not vice versa.
A second argument for (1) depends upon causation’s other properties.
The asymmetry of causation is a logical consequence of its irreflexivity
■∀x∼(Cxx) and transitivity ■∀x∀y∀z ((Cxy & Cyz)→ Cxz). Most schol-
ars would ascribe both of these properties to causation (for transitivity, see
Cartwright 2007, 1923; Ehring 1987, 325; Ehring 1997, 82; Hall 2000;
Hall, Price of Transitivity 20044; Irzik 1996, 252; Koons 2000, 46; D. K.
Lewis Causation, 1973, 5635; Rosenberg 1992, 308; Schaffer, 2009, 376,
inter alios6; and for irreflexivity, see K. Bennett 2017, 68; K. Bennett 2011,
93; Cartwright 2007, 192, assumes it is irreflexive; Ehring 1997, 82; Koons
2000, 55, 65, 82; Schaffer 2009, 376; Tooley 1987, 275, inter alios. And
there are good independent arguments for those ascriptions, although I can-
not venture into those cases here (see chapter 6).
I will assume that the above considerations are enough to justify belief in
(1), and explore the following question, “Does the arrow of causation have
a reductive explanation from a non-causally interpreted physics?” I will
argue that it does not. If I am right, we will have one important piece of evi-
dence for rejecting causal reductionism, the thesis that causation reduces to,
is grounded in, or is completely determined by non-causal, law-governed,
non-causal, physical history.
Section 1.2: What Is a Physical Reductive Explanation

of Causal Direction?
Before discussing my reasons for denying the existence of a reductive expla-
nation, it will be helpful to get clear on what a physical reductive explana-
tion of causal direction (PRE-CD) looks like. Here is the idea,
(PRE-CD): A PRE-CD is a meaningful, sincere, factive, and puzzlement-

removing (for a rational cognizer) answer (α) to the question, “Why
is it true that for any event C and for any distinct event E, given that
event C caused E, E does not also cause C?” in a philosophical con-
text, where (1) α invokes only physical objects, properties, relations,
structures, processes, and/or laws; (2) α does not invoke causal direc-
tionality; and (3) α does not fundamentally depend for its truth on
any set of objects, properties, relations, structures, processes, or laws
that lie beyond the ontologies of the full interpretations of our best
physical theories.
Say that a statement S fundamentally depends for its truth on a set/arrange-

ment/sum/structure of things E, only if [E includes only physical things, and
E de re necessitates (i.e., E’s existence entails in non-de dicto fashion) the
truth of S, or E incorporates/involves only physical things and E directly
or indirectly metaphysically explains7 S’s truth].8 I should add that it is a
requirement for α’s reductive status that the ontology of our best physical
theories upon which it directly leans be empty of causal directionality. Thus,
any space-time structure, evolutions, or features of physical systems that are
both described/explained by our best physical theories, and that enter into
α’s dependency base must preclude obtaining causal relations themselves
(because therein lies causal directionality).
Section 1.3: Roadmap
Now that we are clear on what a PRE-CD looks like, we can press on with
the case against physically and reductively explaining causal direction. My
argumentation will unfold as follows. In sect. 2, I argue that theories of
causation and causal direction that seek to reductively explain it in terms
of temporal direction are PRE-CDs, and that they all fail due to plausi-
ble cases of simultaneous causation in Newtonian and relativistic physics,
and the photoelectric effect. In sect. 3, I argue that no PRE-CD that issues
forth from a theory of causation that would seek to ground causation in
underlying physical goings-on can succeed because what is fundamental
to one of our best fundamental physical theories is a causal phenomenon.
Sect. 4 responds to objections to preceding argumentation and adds addi-
tional objections to reductive theories of causation that are used to motivate
purported PRE-CDs not directly addressed (although they are indirectly
addressed) by the argumentation of sects. 2 and 3.
Section 2: Causal Priority From Temporal Priority?

Time has several important treatments in some of our best physical theories.
Its arrow is often thought to reduce to the arrow of entropic increase. Its
nature is thought to be geometric in that time is a dimension. Its flow (if it
has one) is related to the ticking of clocks, which can be slowed by gravita-
tion and motion. David Hume tried to use the physical entity that is time to
explain causal asymmetry. Indeed, many contemporary philosophers follow
Hume (1978, 1.3.2.6; 2007)9 when they opine that (1) is true because (2)
is as well,
(2) ∀x∀y (Cxy → Txy )
That is to say, necessarily, for any event x, and for any event y, if x causes y,
then x is temporally prior to y.10 The idea is that the direction of causation is
parasitic on temporal direction, or that the sense in which (2) explains (1) is
captured best by appreciating the fact that according to it, one best answers
the question, “why is it true that, for any event C and for any distinct event
E, given that event C caused E, E does not also cause C?” by (sincerely, and
in a philosophical context) answering, “because C is temporally prior to E.”
Evolutions consisting of obtaining temporally ordered events that constitute
physical history serve as the physical reductive explanatory base for causal
asymmetry.
That (2), if true, explains (1) is generally upheld (see Dowe 2000, 179, inter
alios). That it constitutes a reductive explanation of (1) seems clear because
(2) does not invoke the causal directionality it seeks to remove puzzlement
about. Moreover, the explanation suggests that it is the temporal ordering of
events that grounds causal order. Time and temporal order are physical enti-
ties that are part of the ontologies of our best physical theories (e.g., quantum
mechanics (ignoring interesting issues about time’s failing to be represented
by a linear Hermitian operator in non-relativistic QM) and general relativ-
ity). The temporal order of events C and E is both non-causal and physical.11
Thus, Hume’s view constitutes a purported or attempted PRE-CD.
Responses to the Humean position have invoked exotic causal cases
involving strange general relativistic causal structure such as closed timelike
or causal curves so as to allow for time travel and therefore backward cau-
sation (Gott 2001, 76–130), perfectly rigid seesaws (J. Carroll 1994, 141–
142), somewhat controversial readings of everyday pushes and pulls in the
world (Horwich 1987, 135–137), and other supposed instances of simul-
taneous causation (for which see R. Taylor 1966, 35–40). My approach is
different and (so far as I’m aware) completely novel in that I appropriate
counter-examples to (2), and an unnecessitated version thereof, that do not
go beyond well-understood Newtonian and relativistic gravitational phys-
ics, plus the photoelectric effect.12
But is it really novel? A.E. Dummett (1954, 29) argued that “[c]auses are
simultaneous with their immediate effects, but precede their remote effects.”
Although Dummett did appeal to Newton’s laws of motion (the first, sec-
ond, and third are typically regarded as Newton’s laws of motion) to but-
tress this claim, he did not provide a causal interpretation of the Newtonian
gravitational force (law). In Newton’s work, and classical mechanics more
generally, not all forces or fields of force resulting in motion are the same
as the (centripetal) gravitational force (on Newton, see the discussion in
Stein (2002, 283–286); on classical mechanics more generally, see e.g., the
study of oscillations in J.R. Taylor (2005, 161–207)).
Huemer and Kovitz (2003) argue that Newton’s second law of motion,
and the Lorentz force law of classical electrodynamics, causally interpreted,
entail instances of simultaneous causation that falsify—what they call—the
sequential theory of causation. The sequential theory says (quoting Huemer
and Kovitz) that “causes always temporally precede their effects” (ibid.,
556). Notice that this thesis is unnecessitated. The problem is that New-
ton’s second law of motion is false. Our space-time is relativistic. The vec-
tor quantities in the second law of motion that are force and acceleration
must receive a four-vector treatment precisely because one must account
for the nature of those quantities amidst a background four-dimensional
space-time. In special relativity, the second law becomes (expressed here
with three-vectors),
d 1
F= ( γmv) , where γ = , and where m here is the mass of the
dt 1− v
2
2
c
relativistic particle in question (Helliwell 2010, 261).
It is no longer F = ma. Thus, Huemer and Kovitz’s argumentation, even if
sound, does not provide a successful counter-example to unnecessitated (2)
(the sequential theory of causation).
Huemer and Kovitz maintain that their interpretive points about Newton’s
second law “hold for all the equations of classical physics . . .” (Huemer and
Kovitz 2003, 559), in the sense that all such equations can be understood as
detailing simultaneous causal action. This is an overstatement. The Lorentz
transformation laws are equations of classical physics. They do not explain
by causation. I know of no causal interpretation of those equations in the
literature (see the discussion in Lange (2017, 96–149)). Gauss’s law of clas-
sical electromagnetism (not to be confused with Gauss’s law simpliciter)
is expressed by the following equation of classical physics, ∇ · B = 0. It
says there are no magnetic monopoles. It asserts nothing about obtaining
instances of simultaneous causal relations because it says nothing about cau-
sation whatsoever. Perhaps the idea is that all equations of classical physics
that express dynamical laws entail instances of simultaneous causation. But
this is false as well. Mathias Frisch (2014, 129, n. 10) argues that in the case
of classical electrodynamics the retarded (causal) solution to the wave equa-
tion, plus the additional causal condition commonly attached to that solu-
tion, describes a “source” affecting “the total field only after the source is
turned on” (emphasis mine). Cf., Frisch (2009), where his summary (p. 463)
(dependent upon standard textbook sources) of the dispersion relations fea-
tures the equation for the output field. It is described as “non-local in time,”
meaning that “the displacement D at t depends on the electric field E at all
other times, both before and after t” (ibid.; emphasis in the original). With
respect to the simple harmonic oscillator and the retarded Green’s function
in classical mechanics, Sheldon R. Smith (2013, 112–113) stated that that
function “suggests that the Dirac delta function kick [representing a point
source] causes [speaking generally] harmonic oscillations after the kick is
applied to the system” (emphasis in the original). In classical electrostatics
without fields (there are such action-at-a-distance theories or approaches; see
the discussion in Lange (2002, 27–29), one restricts Coulomb’s law to static
situations precisely because the effect that is an impressed electric force on a
point charge2 at t2 is caused by point charge1 at a time t < t2. The action or
influence in such cases is at-a-distance but retarded and not instantaneous.
The moral is to build one’s argument for the (instantaneous) causal interpre-
tation of this or that force or equation in classical physics case-by-case.
No individual direct argument for a causal treatment of gravitation is devel-
oped in Huemer and Kovitz (2003), although they do discuss a case of ter-
restrial free fall (ibid., 564), which presupposes either a causal interpretation
of relativistic gravitation or the Newtonian gravitational force. The authors
never specify which effect they are interested in (GTR is classical physics).
We are also never informed about the inner workings of the interaction (e.g.,
how causal asymmetry is salvaged amidst the proposed interpretation of that
distinctive force or relativistic effect). Moreover, they attribute the causal
source of gravitation in that case to the “spatial arrangements of bodies”
(ibid., 564). If the phenomenon in play is relativistic, then the physics is
incorrect. If it is Newtonian or peculiar to classical mechanics, then their
proposal differs from the one developed here. I invoke an irreducible causal
gravitational force with a distinctive metaphysical nature (see below). Hue-
mer and Kovitz (on the Newtonian reading) do not.
Section 2.1: Newtonian Gravitation

There are naturally possible purely Newtonian worlds in which gravita-
tional interactions obtain in an instantaneous manner. This is because the
gravitational force (given by Eq. 1) has no dynamics in time, acting instan-
taneously at such worlds,13
m1m2
(Eq. 1): F12 = −G rˆ12 , where r̂12 gives the unit vector in a direction
r12 2
that extends from body one to body two (assuming SI units), G is the
gravitational constant, m1 gives the gravitational mass of body one, and
m2 gives the gravitational mass of a distinct body two, and where r is
the relative distance between the bodies.14
It will become useful for subsequent discussion to also state Newton’s law
of gravitation with the Poisson equation, which is equivalent to Eq. 1 (Lam-
bourne 2010, 125),
(Eq. 2): ∇2Φ = 4πGρ, where ∇2 is the three-dimensional Laplacian oper-

ator, Φ is the gravitational potential, G is the gravitational constant,
and ρ is mass density.
Time is completely missing from Eq. 1 and Eq. 2. The three dimensions of
the Laplacian operator ∇2 in Eq. 2 are all spatial. Choosing that operator
“instead of the four-dimensional d’Alembertian operator” , implies “that
the potential Φ responds instantaneously to changes in the density ρ at arbi-
trarily large distances away.”15
For Newton, gravitation is a causal centripetal impressed force that is an
action that is “exerted on” bodies “to change” their states “either of resting
or of moving uniformly straight forward.”16 Newton regarded virtually all
forces as causes of motion, musing that (as was quoted in chapter 1) “[f]orce
is the causal principle of motion and rest.”17 That is why it was one of the
central purposes of the Principia to search for, find, and exploit the underly-
ing causes of effects due to forces, while abiding by the rule that “[n]o more
causes of natural things should be admitted than are both true and sufficient
to explain their phenomena.”18
In both the Principia and the Opticks, Newton suggests that he is not
concerned with providing physical descriptions that explicate the physical
causes of the phenomena under study (Newton 1999, 407; 1952, 375–376,
Query 31). However, those remarks are best read as an admission of igno-
rance regarding the deep or ultimate causes of some of the forces involved
in (specifically) gravitational phenomena. And so, with respect to the gravi-
tational force he wrote,
I have not as yet been able to deduce from phenomena the reason for
these properties of gravity, and I do not feign hypotheses . . . it is enough
that gravity really exists and acts according to the laws that we have set
forth and is sufficient to explain all the motions of the heavenly bodies
and of our sea.19
Gravitation explains motion by causing it in the way explicated by the

universal law of gravitation. However, the deep cause of gravitation is left
undisclosed. This is why the editors of Acta Eruditorum are described (by
Newton himself) as having charged Newton with “making gravity a natural
or essential property of bodies, and an occult quality and miracle.”20 They
were critical of him because he confessed to not having a mechanical expla-
nation of the gravitational force itself. There were causes of gravitating bod-
ies in the gravitational force. What was unknown was the source of those
causes, that is, the source of gravitation was unknown.
The causal reading of Newton’s view of forces respects the historical
development of dynamics during the sixteenth, seventeenth, and eighteenth
centuries (cf. my discussion in chapter 1: sect. 4.5.5.5). His understanding
of forces makes sense, given the views of his intellectual predecessors and
contemporaries. For example, Kepler proffered three reasons for a causal
interpretation of the Sun’s gravitational force upon planetary bodies (Kepler
1937).21 Descartes most often spoke of forces as those entities that push
bodies and generate motion.22 He also believed forces acted instantaneously
when they imparted motion to bodies.23 Forces also act on bodies to trans-
mute their shapes in Descartes (1970, 63). In Galileo’s Dialogo, the discus-
sion between Salviati and Sagredo represents forces as causes of particular
types of motion.24 Likewise, according to Isaac Barrow (1630–1677), some
physical effects result from efficient causes.25 François De Gandt (1995, 58)
quotes from an “almost . . . random” example of various exercises that
were authored to help students learn natural philosophy around the time
Newton authored De motu corporum in gyrum (c. 1684). In these exercises,
forces are quite clearly explicated in terms of producers of work and effects.
Examples and citations could be multiplied.
If there are worlds at which gravitation is strictly and exactly as Newton
described it, and the causal interpretation of Newtonian gravitational force
is correct at those worlds, then there could be instances of simultaneous
causation. However, if possibly, there are instances of simultaneous causa-
tion, then (2) will come out false. Thus, the possibility of causal Newtonian
gravitational interaction yields an argument against the Humean view that
causes must always be temporally prior to their effects.26
The Humean will ask how the asymmetry of causation can be preserved
under the causal interpretation of instantaneous Newtonian gravitational
action. Consider the case in which we let m stand for the gravitational
mass of a planetary body orbiting the Sun (whose gravitational mass is M).
And the relation between these masses is such that M >> m so the Sun
remains at (virtual) rest or at a fixed location during gravitational interac-
tion between M and m. The motion of the planetary body is given by,
d2r GMm
(Eq. 3): m = − 2 rˆr , where m on the left-hand side is inertial
dt 2 r
r
mass, and r̂r is the unit vector equal in value to that has a direction that
r
extends from the sun to the orbiting planetary body.
I will simplify matters and avoid the Keplerian dictum that the planetary
orbit is elliptical. Instead, I will assume it is circular about M, the Sun. We
should now have by introduction of the second law of motion,
v 2 GMm (27)
(Eq. 4): m =
r r2
If in this case, the gravitational force is an instantaneous cause, then the
Sun will gravitationally causally influence the planetary body, and (in keep-
ing with Newton’s third law of motion) the planetary body will likewise
and simultaneously causally influence the Sun. In fact, in the case of two
point masses p1 and p2 exerting gravitational influence upon one another
(I will assume that each is in possession of gravitational masses m1 and m2,
respectively), dual motions instantaneously result. In both cases, we seem to
have a violation of the asymmetry of causation because p1 causes p2, and p2
causes p1.
The worry is attenuated once we more precisely describe either case.
Assume what I will later call the ESSI account of events explicated and
defended in chapter 7: sect. 4. Roughly put, that view says that events are
states of Aristotelian concrete particular substances exemplifying joint-carving
universals at ontological indices (e.g., spatial locations in a 3D space at a
world, space-time locations in a 4D space-time at a world, or a possible
world solely). The account is endowed and similar to the property exempli-
fication view defended in Kim (1976) (and see my discussion and criticism
of Kim in chapter 7: sect. 3.6). According to the ESSI account (cf. Kim 1976,
161), events are identical when they have the same substance, universal, and
index. In addition to the ESSI theory, assume that entities belonging to the
category of forces can likewise stand in causal relations, where I am leaving
open precisely what metaphysical types of entities forces are (I defend this
idea in chapter 7: sect. 5).28
With the above assumptions, we can now more accurately describe the causal
gravitational force. In the case involving the planetary body and the Sun, the
exemplification of M’s gravitational mass at its location L1 at time t0 instanta-
neously causally generates a gravitational force that instantaneously causally
produces an effect in m that is its being impressed upon by that force at t0 at its
location L2. It is that effect that either (a) causally produces subsequent motion
or (b) causally ensures time-varying dynamical behavior. As Shankar (2014,
109) put it when he interpreted Eq. 4, “the left-hand side is the effect, the right-
hand side is the cause.” It is the Sun’s exertion of a gravitational force upon
body m that yield’s the latter’s motion. The influence is asymmetric because
neither the planet’s motion nor it being impressed upon by the gravitational
force at L2 at time t0 causally produces the event that is the Sun’s exemplify-
ing its gravitational mass at its location L1 at time t0. Nor does the event of the
planetary body’s being impressed upon by the gravitational force at L2 at time
t0, produce the gravitational force exerted upon it.29 We can tell a similar story
about the dual point mass case, although a picture is worth a thousand words.
Figures 3.1 and 3.2 should be interpreted with the following guide in mind,
[Guide: C is the event involving P1 exemplifying m1 at location L1 at t0;

C2 is the event involving P2 exemplifying m2 at location L2 at t0; E is
the event that is P2 exemplifying the property of being influenced by
an impressed central force with magnitude x, in direction y; E2 is the
event that is P1 exemplifying the property of being influenced by an
1
impressed force with magnitude v, in direction w; Fgrav is the instanta-
neously exerted (at t0) gravitational force with magnitude x, and direc-
2
tion y; Fgrav is the instantaneously exerted (at t0) gravitational force with
magnitude v, and direction w; ‘⇛’ and ‘⇚’ represent causal relations.]
The understanding of Newtonian gravitation does assume a second-order

theory of magnitudes in that it presupposes that quantities like gravitational
mass are properties of objects, and not themselves concrete p articulars.30
There is an account of extensive properties like mass, with an illustrious
philosophical and formal history (and that is consistent with what I have
articulated thus far), that is available in Mundy (1987). That account
can recover the structure of quantitative properties by positing second-
order relations between them.31
What I hope Figures 3.1 and 3.2 make clear is that events C and E2 are
not identical. They feature different properties/relations in their contents.
Likewise, events E and C2 are not identical for the self-same reason. There
is therefore no violation of causal asymmetry in the Newtonian gravitation
case given a fine-grained view of events. We should dismiss the Humean
account of causal priority for well-understood physical reasons.
Tim Maudlin would take issue with both my historical exegesis of Newton,
and my understanding of Eq. 1 and Eq. 2 (Maudlin 2012, 127). He wrote,
(at time t0)

1
instantaneously exerted
P1 P2
(relative distance r)
Location L 1 Location L 2
Figure 3.1 Newtonian Gravitational Force #1
(at the same time t0)

2
instantaneously exerted
P1 P2
(relative distance r)
Location L1 Location L 2
Figure 3.2 Newtonian Gravitational Force #2
[I]t is very unlikely that Newton actually thought that the gravita-
tional force is instantaneous: he thought that the force must be medi-
ated by some sort of particle, which would have taken time to get
from, for example, the sun to the earth. Of course, it was exactly
here that Newton declared ‘Hypotheses non fingo.’ But more criti-
cally, there is nothing in the general form of Newton’s gravitational
law that suggests difficulties for a relativistic version: Coulomb’s law
of electrostatics is an inverse-square force law, just like Newton’s
law of gravity, and Maxwell’s electrodynamics returns Coulomb’s
law in the appropriate limit. But Maxwell’s electrodynamics is fully
relativistic.32
There are many problems with this paragraph. First, that a force requires
mediation does not necessarily imply that the mediating process takes time.
The mediator may travel with infinite speed and thereby escape a finite time-
dependent dynamical description.33
Second, Newton’s theory of gravitation in the Principia did involve a
commitment to an action-at-a-distance interpretation of the gravitational
force. After all, it was that characterization of gravity that precluded both
Huygens and Leibniz from endorsing larger facets of Newton’s more general
project in the Principia.34 And although it is true that at times Newton did
not like action-at-a-distance interpretations of the gravitational force (New-
ton 2007), he adopted several different analyses of that force subsequent to
the publication of the Principia. None of these accounts invoked particles as
gravitational force mediators. One of them did affirm that an agent (perhaps
even a divine one) causally produces gravitational attraction, musing that
“[g]ravity must be caused by an agent {acting} . . . consta{ntl}y according
to certain laws” (ibid.). The interpretation he seemed to settle into in late
versions of his Opticks used the luminiferous aether (Cohen 1999, 62).35
Third, it is quite likely that Newton interpreted Eq. 1 in such a way that
it described the action of an instantaneous force. Gravitation is a centripetal
force. Centripetal forces can “act with a single but great impulse”36 at dis-
tinctive times. Gravitation is a continuous force only insofar as its actions
can be described as “a series of impulses” exerted or impressed at distinc-
tive times.37 Because celestial bodies that are gravitationally interacting are
physically separated, if gravitational force is both exerted by a body b1 at
t1 and impressed upon a distant body b2 at t1, the force must be instanta-
neously impressed (and if it features no intermediary, it will likewise be
at-a-distance). Newton carefully describes how motions are generated by
continuous forces, understood as series of impressed instantaneous impulses
(Newton 1999, 444–446).38
Fourth, physicists are quite clear about why Eq. 2 (which is equivalent
to Eq. 1) should be interpreted as describing an instantaneous force. The
Laplacian operator ignores the temporal dimension. The dynamics of the
gravitational potential “is determined at each instant of time by the mass
density at” that same “time.”39 This, continued physicist Yvonne Choquet-
Bruhat, is indicative of “instantaneous ‘action at a distance,’ which surprises
our human experience.”40
Section 2.2: General Relativity and Gravitation in the Newtonian Limit

A second objection to (2) and even its unnecessitated cousin requires one
to abandon the natural modal telescope and focus on our relativistic world,
for there is a strong case to be made for a causal interpretation of the action
of the gravitational field in Einstein’s general theory of relativity (GTR).41
Let me first give reasons why GTR should be causally interpreted, and then
show why a causal GTR is problematic for proponents of both (2) and
unnecessitated (2).
By way of support for causally interpreting GTR, consider that Ein-
stein adhered to such an interpretation. There are five lines of evidence
for that attribution. First, Einstein’s preferred statement of the equivalence
principle (EP) attributes causal action to the gravitational field. Einstein’s EP
asserted that the gravitational field (understood as the metric field or geo-
metric structure of space-time itself) causally produces both gravitational
and inertial effects. That fact explains the equality of inertial and gravita-
tional mass, as well as other important facts of GTR.42 Second, and perhaps
most surprisingly, Einstein participated in a dialog about the metaphysics
of causation with Moritz Schlick (1882–1936). In that dialog, Einstein
used an explicitly causal interpretation of gravitational phenomena so as to
counter-example Schlick’s regularity theory of causation.43 Third, Einstein
interpreted the geodesic equations of motion in causal terms and was read
that way by his contemporaries.44 Fourth, Einstein closely associated, if not
reduced, the gravitational field to space-time curvature.45 So as to explain
the nature of gravitational action given that association or reduction, Ein-
stein attributed causal action to space-time itself, using the action-reaction
principle (that “a substance is the seat of actions on other substances,
and in turn subject to other actions of these other substances”46) in order
to object to the Newtonian conception of absolute space.47 Fifth, a great
many relativity scholars attribute to Einstein (at least prior to 1927) a
causal interpretation of GTR.48
There are good reasons for affirming Einstein’s causal interpretation.
Much like Newtonian gravitation theory, classical electrodynamics, and
the theories of quantum chromodynamics (QCD) and quantum electrody-
namics (QED), GTR is a theory about an interaction in so far as it is a
theory of gravitation.49 GTR says (quoting Janssen) “that inertial effects
and gravitational effects must be manifestations of one and the same field,
the metric field, gμν(x)” that is the “inertio-gravitational field.”50 The metric
structure of space is what is responsible for its curvature. But importantly
(quoting Geroch), “[t]he effect of curvature on matter in relativity, via Ein-
stein’s equation, is regarded as fundamental by itself.”51 It is the interac-
tion between curvature and matter that demands a causal interpretation,
by my lights, and any non-causal characterization of that phenomenon will
miss out on structure because, in general, physical interactions (in interac-
tive theories like Newtonian gravitation theory, GTR, and QCD) are them-
selves causal.52
Section 2.3: Physical Interactions as Causal Phenomena

The central reason for causally interpreting physical interactions in New-
tonian gravitation theory (sect. 2.1), GTR (sects. 2.2 and 2.4; chapter 8),
and QCD (sect. 3.1) depends on the truth of the emphasized (through itali-
cization) portion of my last claim. My cumulative case for it rests upon
four legs of support, the first of which consists of a type of prima facie
evidence in favor of (a) regarding gravitational activity as causal (in Newto-
nian mechanics and GTR), and (b) for regarding both the processes of gluon
absorption and particle pair production in QCD as causal.53
The Prima Facie Evidence
Consider relativistic gravitational wave propagation as described by physi-
cist Carlo Rovelli,
A strong burst of gravitational waves could come from the sky and
knock down the rock of Gibraltar, precisely as a strong burst of electro-
magnetic radiation could. Why is the . . . [second] ‘matter’ and the . . .
[first] ‘space’? Why should we regard the . . . [first] burst as ontologically
different from the second? Clearly the distinction can now be seen as
ill-founded.54
Here Rovelli is arguing for the view that the gravitational field is of the
same ontological type/category as the electromagnetic field in that both are
dynamical and matter-field-like causal entities. Irrespective of whether his
particular interpretation of the gravitational field is correct, I take it that it
is at least prima facie plausible that the interaction between the gravitational
field and the relevant matter fields he describes is causal (and notice that it
need not be irreducibly causal for my present purposes).
Now consider a case at a Newtonian world in which someone drops a
heavy boulder onto a plate glass window from a very tall building. The glass
will break. That breaking is a gravitational effect. Prima facie, the state of
affairs in view, is also a causal effect/phenomenon. The gravitational inter-
action should therefore be interpreted causally. Dare I say that if one seeks
to avoid eliminating causation altogether, one will be very hard pressed not
to accept this prima facie case as an instance of bona fide causation (again
the causation need not be irreducible for my present purposes).
In QCD,55 there is a dynamical gap equation that describes the propa-
gation of elementary particles called gluons (see Aguilar et al. 2009). The
solution to that equation implies that an individual gluon1 can be absorbed
or “eaten” by another gluon2, thereby generating the mass of gluon2 (“the
gluon mass is invisible in perturbative applications of QCD”; Roberts, Three
Lectures on Hadron Physics 2016, 19). Thus, “gluons are cannibals” (ibid.),
and I’m following Roberts (Running Masses 2016, 2–3). There’s a time at
which gluon1 ceases to exist on account of being absorbed by gluon2. That
state of affairs results, according to the dynamics of QCD, in gluon2’s new
state, viz., it’s having a (larger) mass at a time. This quite clearly constitutes
a case of destruction in fundamental microphysics. That is to say, there’s a
time at which gluon1 ceases to exist on account of being absorbed by gluon2.
In QCD, quark pairs (i.e., a quark and anti-quark) are often produced
by proton collisions. That process of production has been measured at the
CERN Large Hadron Collider by CMS.56 Here we have an interaction by
collision that results in the existence of something that was not around
before. An entity begins to exist as a result of a collision in the relevant case.
QCD therefore involves both the destruction and creation of entities as a
result of interactions. It is at least prima facie plausible that these instances
of destruction and creation are causal, and that the interactions that pro-
duce them are causal because, prima facie, processes involving absorptions
and collisions are causal processes.57
The Concept of an Interaction Is Causal: The Concept and

Phenomenon Is Treated That Way in the Work of
Psychologists Who Study Causation
Our concept of an interaction (call that concept interactionC) is a causal con-
cept. When I say that interactionC is causal, what I mean is that interactionC can
be understood as a distinctive type of mental representation that is an impor-
tant part of the representational contents of causal facts or propositions, such
that any statement expressing those causal propositions will have a privileged
set of “a priori determinable [non-trivial] causal implications.”58 First, in the
case of statements about interactions, it seems obvious that if someone truth-
fully affirmed that <The fire spark interacted with hydrogen gas.>, such an
affirmation or statement would strictly imply certain other substantive causal
facts (e.g., that the fire affected the hydrogen in some way, that it made a dif-
ference to an outcome involving the hydrogen, or that it wrought/produced a
change in the hydrogen) in a manner that one can discern a priori.
Second, that interactionC is a causal concept best explains certain empiri-
cal evidence from psychology. Human persons often interpret dynamically
interacting phenomena causally to the extent that they will (quoting Slo-
man) “impose a causal frame even when it distorts a representation.”59
Third, that interactionC is a causal concept best explains why psycholo-
gists who study and model causation conceptually analyze and treat it that
way. For example, Michael Leyton introduces “the notion of causality via”
a principle about interactions (Leyton 1992, 13). Psychologists Frederick J.
Gravetter and Larry Wallnau conceptually analyze interactions causally.60
With regard to physical interactions in Newtonian physics, Peter Kugler and
Michael Turvey have said that “[i]t is self-evident that Newtonian interactions
are causal.”61 Psychologists writing on causal analysis assume that interac-
tions are causal (see James, Mulaik, and Brett 1982, 48). Social scientist Ton
Jörg connects interactions with mutual causation (Jörg 2011, 150). Novick
and Cheng provide a causal powers–based account of precisely how interact-
ing events produce or “influence an effect” (Novick and Chen 2004, 461).
These theoreticians prove interesting mathematical results related to how one
can cogently reason to certain causal conclusions. Some of their results rest
upon the assumption that interactive influence “can be represented by a sepa-
rate conjunctive [causal] power” (ibid., 463). Examples could be multiplied.
The Term ‘Interaction’ Is a Causal Term

The term ‘interaction’ is a causal term. Its lexical meaning suggests as much.
The Oxford English Dictionary (OED; Pearsall and Hanks 2003) defines the
term interaction as “a reciprocal action or influence.” An influence is said to
be “the capacity to have an effect on the character, development, or behav-
iour of someone or something, or the effect itself” (ibid., 888), and an action
(in the sense we have in mind) is defined as “the fact or process of doing
something, typically to achieve an aim,” or “a thing done; an act” (ibid., 16).
The word interaction in scientific and physical research contexts is a causal
term. Its technical meaning suggests as much. Here is the Oxford Dictionary
of Physics on the terms technical definition: An interaction is “[a]n effect
involving a number of bodies, particles, or systems as a result of which some
physical or chemical change takes place to one or more of them.”62 Indeed,
the OED includes the following technical definition of the term “interaction”
in physics: An interaction is “a particular way in which matter, fields, and
atomic and subatomic particles affect one another, e.g. through gravitation
or electromagnetism” (ibid., 901). Quite clearly these definitions employ
causal notions. I do not apologize for quoting dictionaries. The fact that the
term interaction is both lexically and technically defined in a causal manner
serves as further evidence that interactionC is causal.
The Evidence from Consensus Among Causal Reductionists

and Causal Anti-reductionists
Theoreticians who seek PRE-CDs are almost all causal reductionists. They
maintain that causation is nothing over and above law-governed, non-
causal, physical history (as in Schaffer, Causation and Laws 2008, 82–83);
and (Hall and Paul 2013, 10, inter alios). Interestingly, causal reduction-
ists do not resist my causal understanding of interactions. Aronson indi-
viduates causality in the sciences by way of specifying “changes that result
from interactions with other objects” (Aronson 1971, 421; emphasis mine).
Heathcote reduces all causation to a more fundamental causal influence in
quantum field theory (QFT), but then interprets that causal influence rela-
tion as an interaction between the quantized fields responsible for changes
individuated by observables of QFT (Heathcote 1989). Wesley Salmon
believed that every obtaining interaction is a causal interaction, a causal
phenomenon (Salmon 1984, 179; 2010, 8, n. 14). Kutach (2013, 14), while
espousing a type of eliminativism (see chapter 1’s discussion) about causa-
tion, says of our concept of gravitation that it is generalized by our notion
of causation. Examples could be multiplied.
As further evidence that distinctively physical interactions like gravitation
are causal, consider the fact that such interactions can be understood in terms
of total causes as understood by the (anti-reductive) manipulationist or inter-
ventionist account of causation (Pearl 2009; Spirtes, Glymour, and Scheines
2000; Woodward 2003, 98–151; Woodward 2009, 243–247). Total causa-
tion enters the work of anti-reductionist James Woodward when he differenti-
ates it from contributing causation (Woodward 2003, 51ff.; 2009, 248–251).
The former idea is given by the biconditional (C is a total cause of E, just in
case, there exists an intervention that solely affects C in such a way that “for
some value of other variables besides”63 C “this intervention on” C “will
change the value of” E).64 Given satisfaction of the preceding necessary and
sufficient conditions for total causation, and satisfaction of the conditions
reported on in what Woodward opines is a “plausible” (ibid., 250) hypothesis
about causal relevance if total causation is in mind (ibid., 248): (Condition
#1): There exists a hypothetical/possible intervention I upon C, such that C’s
value is changed by I; and (Condition #2): The obtaining of I by itself results
in the correlation of C with E;65 it would follow (by material implication) that
C isn’t just a total cause of E, but also causally relevant to E. With the previ-
ously mentioned conception of total causation and total causal relevance in
hand, Woodward applies the notion of total causation to Newtonian gravita-
tion. He wrote that given (specifically) satisfaction of conditions #1 and #2,
if an ‘action at a distance’ version of Newtonian gravitational theory

had turned out to be correct, this would be a theory that described gen-
uine causal relationships on an interventionist account of causation.66
No one doubts that Newtonian gravitation is an interactive phenomenon,

and one can easily extend the interventionist account of total causation in
such a way that it accounts for interactions in GTR and QCD as well (in the
spirit of M. Frisch 2014, 77–110, especially 92–93, who applies an inter-
ventionist theory of causation to other physical interactions). We can there-
fore safely conclude that plausible manipulationist/interventionist theories
of (at least total) causation treat physical interactive phenomena causally.
Lastly, I add that anti-reductive mechanistic approaches to causation
in the work of Glennan (2002, S344) and Tabery (2004, 8) suggest that
interactions are causal (q.v., chapter 9: sect. 3). On that approach, interac-
tions are essential ingredients in the more general non-reductive mechanistic
theory of causation. Glennan wrote,
The New Mechanist account of activities and interactions suggest

[that] . . . activity (and with it cause) is an abstract concept that is used
to characterize the actual concrete activities and interactions that are
productive of changes in the world. . . . The totality of mechanisms—
including their (generally mechanism-dependent) parts, activities, and
interactions—constitutes ‘the causal structure of the world.67
So, we have three very different types of approaches to causation that all
explicate interactions causally. I hope the considerations of this section at
least afford a burden of proof shift onto the theorist who insists that physi-
cal interactions are not causal.68
Section 2.4: GTR and Instantaneous Causation

What relevance does the causal action of the gravitational field as under-
stood by GTR have on the question of whether (2) or its unnecessitated
cousin is true? Proposition (2) says that a necessary condition for obtaining
causal relations is that causes occur temporally prior to their effects. How-
ever, I will argue that in general relativity one has in space-time’s causal
action upon matter, and matter’s causal influence on space-time (in the
Newtonian limit), instances of simultaneous causation.69 A consequence of
my reasoning will be that it is not true that a necessary condition for obtain-
ing causal relations is the temporal priority of the cause. Rather, in the gen-
eral relativistic cases I’m about to discuss, causes occur simultaneously with
their effects and so do not occur temporally prior to them.
How does simultaneous causation enter the best interpretation of GTR?
Newton’s theory of gravitation, properly supplemented, is not completely
useless. Given even GTR, it is approximately correct if the following condi-
tions are satisfied: (a) the theory is supplemented with the correct ontology
from the best full interpretation of GTR; (b) stresses of the material variety
are significantly less than persisting mass-energy densities; (c) gravitation is
weak, or (equivalently) an almost flat space-time metric describes the local
geometry; and (d) relative motions are much less than that of c.70 Despite a
perceived formal barrier, GTR recovers the success of Newtonian gravita-
tion theory precisely when (a)–(d) obtain. The perceived formal barrier is
the task that is linearizing gravity. One does this by setting the Lorentzian
metric gab in Einstein’s field equation,
1
(Eq. 5) : Gab ≡ Rab − Rg ab = 8π Tab ( in a geometrized unit system)
2
equal to ηab + γab, where the first addend is the flat Minkowski metric of
STR whose derivative operator is ∂a, and where the second addend gives the
deviation from gab (the perturbed metric) from the former metric (this must
be only slight). Substitute gab = ηab + γab for the Lorentzian metric in Eq. 5.
Very roughly put (in sketch form) (and see all the details explained in Wald
(1984, 74–76)), one can use the gauge freedom of the deviation from gab, and
an analog of the Lorenz gauge condition to help simplify matters, so that
one can state a simplified linearized version of Einstein’s field equation,
(Eq. 6) : ∂c ∂cγ ab = −16π Tab
With much success, physicists use Eq. 6 when conditions (a)–(d) hold. Thus,
linearized gravitation is a good approximation of the dynamics of the gravi-
tational field in the Newtonian limit.71
In the linearized gravitation approximation for the Newtonian limit,
that conditions (b) and (d) obtain is a strict implication of the fact that the
energy-momentum tensor Tab possesses a time-time component (solely) rela-
tive to a global inertial coordinate system of the Minkowski metric (given
by Tab ≈ ρtatb).72 The sources Tab describes are involved in slow time-varying
dynamical interactions with the gravitational field or space-time geometry.
Thus, the presence of slow varying sources causes “space-time geometry to
change slowly as well” (Wald 1984, 76).
An appropriate solution of the perturbed metric (or the deviation from
the Lorentzian metric gab) in the Newtonian limit is,
1
(Eq. 7 ) : γ ab = γ ab − ηabγ = − ( 4t a tb + 2ηab ) φ
2
Critically, the potential of the gravitational field (φ in this equation) sat-
isfies the Newtonian Poisson equation (Eq. 2).73 But again, in this limit,
there remains gravitational dynamical interaction between the metric field
and matter.74 Thus, as Misner, Thorne, and Wheeler (1973, 415–416) show,
“Einstein’s field equation reduces to” Eq. 2 “for the generation of grav-
ity by mass” (ibid., 416). Again, that equation forsakes time and therefore
suggests that in the Newtonian limit, matter’s generation of gravitation is
time-independent and instantaneous even given that the world is correctly
described by GTR (ibid., 177).
There are geodesic equations of motion in linearized gravitation theory
for the Newtonian limit. For at least test particles that aren’t under the influ-
ence of any external forces, the equation of motion is,
 
(Eq. 8): a = −∇φ (with respect to global inertial coordinates of the
Minkowski metric)
But again, the dynamics of the gravitational field is independent of time

(Hawking and Ellis 1973, 71–72; Zee 2013, 303). Its action is instanta-
neous. Thus, both the field’s reaction to matter and the field’s causal action
on matter are instantaneous on GTR in the Newtonian limit.
Am I suggesting that GTR involves some type of mysterious instanta-
neous action-at-a-distance? Again, no. In fact, one reason for preferring
the orthodox interpretation of GTR according to which the gravitational
field reduces to space-time geometric structure itself is that that structure is
where the objects are. No action-at-a-distance is required. Much like Max-
well’s field theoretic approach to electrodynamics enabled physicists to put
away action-at-a-distance in classical electrostatics, so too does the gravi-
tational field enable one to jettison dynamical non-locality of the instanta-
neous variety.
The case for causally interpreting GTR was not limited to gravitational
action in the Newtonian limit. Thus, on the causal interpretation and out-
side of the Newtonian limit, the gravitational field causally influences mate-
rial bodies. Sure. But as has already been noted, matter can cause ripples in
the space-time metric or gravitational field, thereby producing gravitational
waves or radiation. Thus, matter causally influences the gravitational field,
too. If there is back-reaction of this kind, why regard the gravitational field’s
action as primary? Why think the “causal action” in GTR is asymmetric?
The objection does not appreciate the fact that there are different physical
processes to appreciate in these contexts. One involves metrical field action
on bodies that produces inertial motion (e.g., geodesic motion of massive
bodies), and one involves matter’s motion and influence upon the metric
structure of space-time yielding gravitational waves that can likewise pro-
duce effects such as the knocking down of mountains (Rovelli 1997, 193).
By my lights, both processes are best interpreted as involving asymmetric
causal influence. Each therefore attests to a causal reading of gravitational
action generally.
With respect to geodesic motion, the field’s action is primary and the
influence asymmetric (q.v., my more complete case for this in chapter 8:
sect. 6.3). The geodesic principle in GTR says that both massive and mass-
less bodies (I could focus solely on the latter), when free—that is, not under
the influence of any forces, although such entities are gravitating—take as
their trajectories in space-time, geodesics. Or, as it is sometimes put, time-
like geodesics of space-time are the possible trajectories of free (but gravi-
tating) massive bodies, just as null geodesics of space-time are the possible
trajectories of free (but gravitating) massless bodies such as photons. But
again, in GTR, this motion is due to the geometric structure to which Ein-
stein reduced the gravitational field. That is why he says the law of motion
“asserts that a gravitating particle moves in a geodesic line.”75 In the case of
the null geodesic and the photon, there’s no massive body to worry about.
In the case of the massive body that is free, the gravitational field produces
an increase in the energy-momentum of massive bodies without itself hav-
ing a local energy-momentum density (as discussed by Misner, Thorne,
and Wheeler (1973 (MTW), 466–468); repeated as point (c) in chapter 8:
sect. 6.3). If events are property exemplifications or something near enough,
the cause of the relevant action cannot be likewise causally produced by the
effect. Massive bodies cannot do to the field structure what the field structure
does to them. This is because the increase of energy-momentum is due to
a global metric structure. The direction of the increase is from that global
structure to the massive body. Indeed, the very metric tensor gab enters the
equations for the energy-momentum of massive bodies (and not vice versa).
In standard discussions of the matter (such as MTW; ibid.), the explana-
tory direction is clearly taken to be from global metric structure to the local
behavior of massive bodies.
In the second distinct physical case involving matter’s influence on the
field, matter is clearly causally primary, for the matter fields appear (indi-
rectly) in the source term of Einstein’s field equation (Eq. 5) via the energy-
momentum tensor Tab. But again, the matter fields are not primarily causally
responsible for those events that yield particle inertial motion. How could
they be? The metric field’s influence is global (in the sense that it does not
have any local-energy momentum density and in the sense that its structure
is globally defined). Matter’s influence on the field is local. Moreover, in
cases involving gravitational wave influence on matter’s motion, the radia-
tion generates trajectories that deviate from geodesics, although that devia-
tion does not itself produce the same series of wave propagation. Therefore,
the envisioned causation contains a directionality in that some events are
causally prior to others. That directionality underwrites the arrow of scien-
tific explanation you see discussed in the standard explications of the theory
(I’ve already cited MTW along these lines, but see also the discussion in
Rueger 1998).
Section 2.5: The Photoelectric Effect

The photoelectric effect was discovered by the German physicist Heinrich
Hertz (1857–1894) in 1887. The effect entails that shining light on certain
materials (usually metals) causes electrons to exit those materials. Some
years before Hertz’s discovery, Max Planck (1858–1947) would seek a
theory of blackbody radiation informed by tenets of the gas theory of Lud-
wig Boltzmann (1844–1906). That dependence is surprising because Planck
famously resisted atomism (a tenet of Boltzmann’s theory of gases). None-
theless, his editorial work on the lectures of Gustav Kirchoff (1824–1887),
specifically those on the work of Boltzmann, provided an avenue to the nec-
essary information (Brush with Segal 2015, 192), information that would
propel his theory of blackbody radiation. In 1900, Planck would propose
an equation that yielded a description, but not a dynamical explanation sit-
uated amidst a full-blooded theory, of the electromagnetic radiation spectra
for blackbodies at a great many wavelengths (Planck, Verbesserung 1900;
Theorie 1900).
Tiny microscopic resonators scattered across radiating surface materi-
als were the emitters of radiation, according to Planck, and at this point
in the history literature there exists substantial disagreement (see Darrigol
2001; Gearhart 2002; Kuhn 1978; Weinberg 2013). Planck is thought to
either have hypothesized that the energy of the emitted radiation or that
the energy of the resonators must be characterized in terms of an integral
multiple of hf (where f is the frequency at which a resonator vibrates, and h
is Planck’s constant whose value at the time was yet to be determined). That
is to say, the energy of the resonators or the energy of radiation must be
discrete or quantized. One of these disjuncts constitutes the quantum
hypothesis. By my count, it seems that most historians favor the thesis that
Planck quantized the energy of the resonators. The publication of Einstein’s
(1905) “Über einen die Erzeugung und Verwandlung des Lichtes betreffen-
den heuristischen Gesichtspunkt” advanced the explicit and unambiguous
proposal that the energy of electromagnetic radiation (light) should be dis-
cretized or quantized, such that the energy of light should be understood as
spread out into various small space-time regions in contrast to the content
of classical Maxwellian electromagnetic theory.
The photoelectric effect already made trouble for classical theory. It sug-
gested that light is corpuscular and particulate. What was observed in the
effect was that light intensity determined the amount of ejected electrons,
although the kinetic energy of those ejected remained unaffected by an increase
of light intensity. That is a puzzling discovery on classical electrodynamics
according to which cranked up intensity of incident light implies raised or
larger associated field strengths, which should yield larger kinetic energies
for ejected electrons. Moreover, the photoelectric effect made plain that once
the value of f lies below a threshold value (a minimum energy called the
work function or ϕ), absolutely no electrons would be caused to exit the
material. Even if the intensity of the light was very robust, if the frequency
was low enough, there would be no electron emission. This was left unex-
plained by classical theory.
Einstein recovered both facts, and then added a prediction. The thought
again was that light’s energy should be quantized. A single particle of light
(what would be called a photon) has a certain discrete amount of energy,
and that energy is surrendered to an exiting electron. If you crank up the
light intensity, you get more photons bombarding the material, and as a
result you get more electron ejections. Why the steady kinetic energy in the
ejecting electrons? Well, that energy is imparted by the incident photons at
a discretized value. There will be no change of that kinetic energy (I’m sim-
plifying some). Relative to any relevant metallic material, there’s a work
function, such that should the discretized energy hf of incident photons be
less than that function (i.e., should the frequency be less than a critical fre-
quency f0, where hf0 = ϕ), you will fail to see any exiting electrons (Taylor,
Zafiratos, and Dubson 2015, 128; Shankar 2016, 412–414).
Einstein’s prediction was that the kinetic energy of ejected electrons will
be hf minus the work function of the metallic material in question, or per-
haps less than that value. This was later (1916) confirmed experimentally by
the work of Robert A. Millikan (1868–1953).
Great. Why the detour? How does this relate to the philosophical issues
before us? It may be shocking to learn that it is well-known in physics and engi-
neering that the photoelectric effect is an instantaneous one (Khan and Pani-
grahi 2016, 565; de la Peña and Cetto 1996, 160, n. 20; Holton and Brush
2001, 399)! Indeed, the very abstract to Arthur H. Compton’s (1892–1962)
“On the Interaction Between Radiation and Electrons” describes the photo-
electric effect as one that involves an action “that . . . is sensibly instantaneous”
(Compton 1928, 59, and see also the remarks at 61–62).
But is the effect causal? I answer, obviously so. Everyone agrees it is an
interaction, and everyone agrees that its directionality runs from incident
light to photoelectron. In other words, there’s real asymmetry in the effect.
Let’s move on.
Section 3: Causal Priority From Physics?

My thesis is that there is no physical reductive explanation of causal asym-
metry. Causal priority is brute. Causal direction is primitive. Typically, those
who would seek to connect causal asymmetry to a type of temporal asym-
metry do so because for them, causal influence reduces to law-governed
physical history (causal reductionism), or because they believe at least
causal direction reduces in some way to an asymmetry in a physical base
(i.e., goings-on described by physics). For some, initial conditions together
with the arrow or asymmetry of entropic increase in thermodynamics and
statistical mechanics serve as the appropriate asymmetric physical base.76
I see causation in the physical base, and if that is right, then any attempt to
reduce causal direction to the arrow of entropic increase, for example, will
fail, for already within microphysical evolutions driving entropic increase
are obtaining causal relations and therefore causal direction, and all of this
quite independently of considerations such as those discussed in sect. 2.
Section 3.1: Quantum Chromodynamics

Why believe there is causation in fundamental physics (the physical base)?
The brief case for causal GTR outlined previously will not avail me because
GTR is typically thought to be non-fundamental (although see chapter 8:
sect. 6.1). Consider though the standard model of particle physics described
by QCD, QED, our best theory of the weak interaction, and at the appropri-
ate scale, electro-weak theory.77 That model incorporates forces described by
sets of equations called Yang–Mills equations (Yang and Mills 1954). Those
equations are provided by and related to various Lie groups. We do not yet
have exact solutions to many of the Yang–Mills field equations of QED and
QCD. We have, however, ascertained much of their physical content by
means of semi-classical and perturbative methods.78
What group you associate with a certain force is called that force’s gauge
group. Electrodynamics has a gauge group represented by U(1). Its equa-
tions are Maxwell’s interactive-dynamical equations. But groups can be
compounded. One does this by looking to the direct products of groups
(e.g., the direct product of group G1 and group G2 is G1 × G2). The stan-
dard model’s gauge group is the direct product of three groups, viz., U(1)
already introduced, plus SU(2) (weak force), and SU(3) (strong force) (where
the standard electro-weak group is SU(2) × U(1)). The symmetry or gauge
group of the standard model is then SU(3) × SU(2) × U(1). That group is
non-Abelian, which means that the result of a series of transformations you
perform with respect to that group depends upon the order in which the
transformations are performed. If the gauge group of a force is non-Abelian,
then the Yang–Mills equations describing it are non-linear, and as a result
one can say of the relevant force that it behaves in a non-linear fashion.79
Let’s focus on the non-linear strong force and its gauge group SU(3), or
more precisely SU(3)C (where ‘3’ is indicative of the three colors the six types
of quarks, distinguished by flavor, come in).80 As I have previously noted,
the appropriate non-Abelian gauge theory for the strong force is a quantum
field theory of Yang–Mills fields that is none other than QCD. That theory
describes and explains strong interactions between quarks and gluons that
yield hadrons such as pions, neutrons, and protons. More specifically, in
QCD there exist gauge fields that influence quarks by way of their quanta,
gluons.81 This is because quarks influence and interact with one another by
way of gluons, or by exchanging vector gluonic fields (Faddeev and Slavnov
1991, 87–96, 13; Healey 2007, 146–147; Ioffe, Fadin, and Lipatov 2014, 53).
Gluons are fundamental quanta of QCD.82 They causally interact with one
another by means of undulations of quantum fields (Heathcote 1989, 100),
and by their cannibalistic behavior (inter alia). Gluon self-interactions can
produce gluonium or glueballs. Although we have yet to discover such enti-
ties in the natural world, it is well known that SU(3)C is asymptotically free
given that there are less than 17 quark flavors with appropriate masses.83
As has already been noted, there are only six flavors of quark. The theory is
therefore free in the relevant sense. A fortiori, QCD would not be a theory
of the strong interaction without the property of asymptotic freedom (Ioffe,
Fadin, and Lipatov 2014, 53). What is more, it was proven by Coleman and
Gross (1973) that only non-Abelian gauge theories can be asymptotically free
(hence my brief summary discussion of non-Abelian gauge theory previously).
The asymptotic freedom of QCD means that the strength of the involved
strong interactions (the coupling constant g discussed below) increases as
the distance between the objects involved in the interactions increases (or at
higher and higher energies, the coupling constant strength increases). Con-
sequently, the coupling strength decreases as the relevant distance decreases
(or at lower and lower energies, the coupling constant strength decreases).
This has led to the postulation of the confinement hypothesis, that both glu-
ons and quarks are never lonesome but are bound together in hadrons. Glu-
ons and quarks that interact feebly with one another (and with themselves)
become quasi-free when they are a short distance away from one another
on account of such asymptotic freedom. Herein lies a reason for regarding
gluonic interactions as fundamental to the theory. Gluon self-interaction
is the basis of asymptotic freedom84 despite our inability to find glueballs.
It is a datum that the phenomenon of interaction is a causal one, and that
the strong force is an interactive one. I defended the first conjunct of this
claim in sect. 2.3. In defense of the second conjunct, I note how it is com-
mon (without conceptual or linguistic confusion) in the QCD literature, to
refer to the strong force as the strong interaction. I have provided reasons
why that is the case in preceding discussion. As further evidence for causally
interpreting QCD, I note that within the formalism of the theory lies g, the
coupling constant, that is representative of interactions. The dimensionless
parameter as that is associated with g is standardly understood to be “the
strength of the interaction” itself understood as “a function of the interac-
tion energy.”85 The coupling constant is immediately visible in both the local
Lagrangian density for QCD (given in Eq. 9), and the equation giving the
a
value of the gluonic field strength tensor Gµν in Eq. 10,
(Eq. 9) : QCD = qi (iγ µ

) 1 a µν
Dµ  − Mδ ij qj − Gµν
ij 4
Ga
1 a µν
( )
= qi iγ µ ∂ µ − m qi − gGµa qi γ µ Tija qj − Gµν
4
Ga
(Eq. 10) : Gµνa = ∂ µ Gνa − ∂ν Gµa + gf abcGµb Gνc
(Roberts, Three Lectures on Hadron Physics 2016, 1;
cf. Healey 2007, 147–148)
And so, not only do we have very good reasons in keeping with orthodox
understanding to interpret and understand the strong force in terms of an
interaction, but there is within the formalism of QCD a term whose value
informs us about the strength of that interaction itself. Indeed, there are
formulations of QFT that more explicitly represent and describe interac-
tions between fields. For example, it is the business of constructive field
theory to provide the mathematical means whereby one can more accurately
describe interacting quantum fields (for more on that program, see Heath-
cote (1989, 95)). Thus, it appears that causation enters the ideology of QCD
via its formulation, and via the very fundamental phenomena described and
explained by QCD given the success of my case for interpreting interactions
causally in sect. 2.3.
Section 4: Objections
Section 4.1: Time-Reversal Invariance

Some may find the causal understanding of quantum phenomena unpalat-
able because the governing microdynamics is time-reversal invariant. That
fact entails that there are naturally possible worlds at which the imagined
microdynamical causes are the effects, whereas the effects are transmuted
into the causes. But that supposition is benign. Proposition (1) does not say,
(1) ∀x∀y (Cxy →  ~ Cyx)
for that claim would suggest that if a gluon’s activity causes a quark to take
on certain properties, it is impossible for the quark’s beginning to exemplify
those properties to be the cause of the gluon’s activity. Proposition (1) only
asserts that a certain material conditional holds at every world. We should
not jump to a different possible world when seeking to discern satisfaction
of (1)’s consequent. If at an arbitrary world w, the gluon’s activity causes
a quark to take on certain properties, then (at w) it is not the case that the
quark’s taking on those properties causes the gluon’s activity.
Section 4.2: What About Reductive Theories of Causation?

One could respond to the evidence I’ve provided for a causal QCD by insist-
ing that causation receives a plausible reduction, and that all reductive theo-
ries of causation incorporate a purported PRE-CD. I have two responses to
this worry. First, recall that causal reductionists maintain that causation is
nothing over and above law-governed, non-causal, physical history. What
physical history could one use to reductively explain causal interactions in
QCD? What laws would one appropriate to connect the events constituting
that history if not the laws of QCD? I have just argued that fundamental
micro-physical history involving the strong force is pregnant with causal
interactions, and that the relevant laws of QCD should be interpreted caus-
ally because they employ the coupling constant for the purposes of describ-
ing and explaining interactive history. Sect. 2.3 went to great pains to show
that interactions are causal by even the reductionist’s lights. It seems that
the reductionist must go eliminativist with respect to the causal goings-on
of fundamental microphysics, for what more fundamental history is there
available to serve as a reductive base? QCD is a prospective fundamental
physical theory. Any deeper story could, at this stage, only be expressed as
a mere hopeful theoretical entity with no real substantial laws. Alas. Elimi-
nativism about causation in microphysics is implausible. Given scientific
realism, particles really do collide. Fields really do undulate and propagate.
Quanta really are exchanged. Fields really do interact. Denying these facts
amounts to denying both the goings-on in physical practice (there really are
large hadron colliders that collide high-energy particles at high speeds), and
the data fundamental physics describe and explain.
Second, it is not true that every reductive theory of causation incorporates
a purported PRE-CD. For example, Phil Dowe’s (2000, 89–122, 171–172)
conserved quantity theory86 has it that fundamental causal relata are physical
facts understood as objects exemplifying properties at times (ibid., 170; cf.
Kim 1976). A causal relation between fundamental physical facts Cqα and
Eqβ (where ‘qα’ and ‘qβ’ pick out conserved quantities possessed/exemplified
by the object constituents of facts C and E, respectively 87) obtains, just in case,
there exists at least one collection of causal interactions and causal processes
between facts Cqα and Eqβ such that the following conditions are satis-
fied: (Condition (i)): “any change of object from” α to β “and any change of
conserved quantity from” qα to qβ occurs “at a causal interaction involving
the following changes”: ∆qα, ∆qβ, ∆qαʹ, and ∆qβʹ; and (Condition (ii)): with
respect to any conserved quantity exchange mentioned in (Condition (i)) that
involves a plurality of conserved quantities, transmutations of such quanti-
ties are described by just one physical law.88 Causal processes are just objects
exemplifying conserved quantities on worldlines of space-time, and causal
interactions transpire when such worldlines intersect, and there is conserved
quantity exchange between the objects on the intersecting worldlines (Dowe
2000, 90).
Dowe’s theory, considered in isolation, is never directly connected to
causal direction. In fact, causal processes that include interactions can
admittedly run in two opposed directions (as Dowe 2000, 205, says, “a
causal process broadly speaking . . . can have two directions.” (speaking
about the third fork view)). Nothing about the theory fixes which direction
is privileged because nothing intrinsic to such processes fixes causal prior-
ity or directionality.89 The (third) fork theory of causal direction that Dowe
develops rests atop his reductive theory of causal processes and interactions
(the forks Dowe needs are “part-constituted by” causal processes as he has
characterized them; Dowe 2000, 204). So, although Dowe’s account does
not itself imply a theory of causal direction, his theory of causation is a nec-
essary condition for the truth of his account of causal direction (ibid., 210).
I criticize Dowe’s theory of causation in sect. 4.3.1 and in chapter 8: sect. 1,
and so I thereby indirectly criticize his account of causal direction.
Section 4.3: Reductive Accounts Once More

My case against causal reductionist attitudes about causal direction may
still be unconvincing. The reductionist will insist that there are plausible
reductive accounts of causal phenomena like the interactions appealed to in
sects. 2 and 3. These reductive theories allow causal reductionists to resist
causally interpreting phenomena in physics.90 I have three additional inter-
related replies (that go beyond my remarks in sect. 4.2). First, the possession
of a reductive theory of causal phenomena does not entail that the interac-
tions I have discussed are non-causal. One might agree with Dowe that
causal interactions reduce to conserved quantity exchange between objects
on intersecting worldlines (Dowe 2000). However, accepting that reduc-
tion does not imply that interactions are no longer causal. Reductions do
not eliminate, for reductions are asymmetric relations between the reduced
(causal interaction) and the reductive base (intersecting worldlines of objects
exchanging conserved quantities, in this case). Relations require the exis-
tence of their relata. Thus, even given a reductive theory like Dowe’s, causal
physical interactions exist, despite their being reduced to something non-
causal. Causal reductionists are not causal eliminativists.91 My argumenta-
tion for the thesis that interactions such as the high-speed collisions between
high-energy particles in a large hadron collider, or gluonic interactions that
produce glueballs in QCD, or the destruction of a building by a burst of
gravitational waves are causal is not defeated by the mere supposition that
these causal interactions are subject to a reduction as explicated by causal
reductionists. If one proffered a successful reductive theory of causation
that afforded a reductive explanation of causal priority, my argumentation
in this chapter would be defeated, although, again, it would not necessarily
follow that the interactions to which I have appealed are non-causal.
Given my first reply, the attempt by reductionists to reduce physical inter-
actions by appeal to their favorite reductive account of causation only high-
lights the fact that the phenomena under evaluation are causal. The question
then becomes whether any of the reductions they propose are successful and
(given that I am interested in defeating reductive accounts of causal direction)
whether any of the proposed reductive theories of causation afford a PRE-CD
because not all reductive theories of causation include purported PRE-CDs.92
Lastly, reductive theories of causation such as counterfactual depen-
dence accounts (D.K. Lewis, Causation 1973; Postscripts 1986), conserved
quantity and/or transference accounts (Fair 1979; cf. Dowe 2000), and
others are notoriously problematic. They are almost universally acknowl-
edged as failures (see the comments by two leading reductionists in Paul and
Hall 2013, 249; cf. the comments in Kutach 2013, 282–306; Psillos 2009,
154; Schaffer 2007, 872–874; Tooley 1987, 5). The well-known counter-
examples receive informed discussion in J. Carroll (2009) and Paul and
Hall (2013), and in many other places.93 However, philosophers are rarely
convinced by a mere documentation of the reigning paradigms of thought.
What they rightly require are arguments. And although I cannot address all
possible reductive accounts, what I present here constitutes a step toward
an inductive case for the general thesis that all extant reductive theories of
causation fail.
Section 4.3.1: The Conserved Quantity Theory of Causation

Recall that for Phil Dowe’s (2000) reductive account of the causal connec-
tion, there are only two fundamental causal phenomena, viz., causal inter-
actions and causal processes. Here is a more complete characterization of
Dowe’s reductive account of those phenomena,
(Dowe’s Process Theory of Causation): A causal process can be given the

following characterization: For any object x, x is involved in a causal
process, just in case, x has a worldline and x has a quantity that is con-
served. Causal Interactions: For any x, x is a causal interaction, just in
case, x is the intersection of two distinct worldlines and there is con-
served quantity transference from one of these worldlines to the other.
(summarized from Dowe 2000, 90)
Worldlines are sets of Minkowski points “on a space-time (Minkowski) dia-

gram that represents the history of an object” (ibid.). Conserved quantities
are physical quantities that abide by laws of conservation in “current scien-
tific theory” (ibid., 91). Objects are ordinary entities from the ontology of
common sense, or they are entities from the ontology of science. Such enti-
ties persist through time remaining identical as they persist. The objectual
worldlines or trajectories of objects are processes more generally conceived
(ibid.).
For Dowe, all causal interactions reduce to conserved quantity exchange
(ibid., 94). These exchanges amount to changes of conserved quantity val-
ues. There’s an exchange from one process x, to another process y, just in
case, x is an incoming process in the sense that the process resides upon or
in the past light cone, and y is an outgoing process in the sense that it is
in or upon the future light cone, and these processes experience conserved
quantity value transmutation or change.
Recall (from sect. 4.3) that providing a reductive causal theory of a gravi-
tational interaction does not falsify the hypothesis that that interaction is
causal. Indeed, Dowe labels physical interactions his theory reductively
explains causal interactions. The question is whether Dowe’s theory prop-
erly recovers causal phenomena in physics by providing a reductive account
of them. I think it is quite clear that the theory does not succeed along
these lines. Consider the fact that the dynamical interaction between mat-
ter and metric/gravitational fields yields gravitational radiation that can
carry energy-momentum in a vacuum or empty space (i.e., where Tab = 0).
Gravitational waves are ripples of space-time, and as I noted in sect. 2.3,
gravitational waves can be and are causally responsible for certain effects by
interacting with matter fields). Unfortunately, Dowe’s theory cannot recover
instances of causation involving gravitational waves because such entities
have no worldlines. Worldlines consist of a plurality of space-time points
in a series. Gravitational waves are ripples of space-time itself, and there-
fore do not have worldlines in space-time (thanks to theoretical Tom Banks
for help here). By consequence of Dowe’s account, gravitational waves or
radiation cannot be said to causally interact with anything. Moreover, Dowe
must deny that they are engaged in any causal processes.94
Section 4.3.2: The Counterfactual Dependence Account of Causation

The counterfactual dependency accounts of deterministic causation in the
work of David Lewis (Causation 1986; Postscripts 1986; 2000; 2004) have
received much attention.95 I will focus on the counterfactual theory of Lewis
(Postscripts 1986), which is an attempted amelioration of Lewis (Causation
1973). According to that account, event c is a deterministic cause of event e,
just in case, c and e are non-identical, c and e are both mereologically and
logically distinct or unrelated, and e stands in the relation of counterfac-
tual dependence to c, or else there is step-wise counterfactual dependence
between c and e.96 Event e counterfactually depends on c, just in case, were
c not to have occurred, e would not have occurred. Step-wise counterfactual
dependence is just the ancestral of the counterfactual dependence relation.
The ‘would’-counterfactual conditional <If it were the case that (a) c failed
to occur, then it would be the case that (b) e failed to occur.> non-trivially
holds, just in case, there is an (a)&(b)-world (i.e, a world at which (a) and
(b) holds) that is closer to the actual world than any (a)&~(b)-world (i.e.,
than any world at which (a), but not (b) holds).97 The metric for assessing
closeness is given by degreed similarity relations assessed on the basis of
shared physical laws and physical history between possible, albeit concrete
worlds.98 For some (e.g., Field 2003, 453–454, and others; although q.v.,
my challenge to this interpretation below), Lewis maintained that contingent
physical facts about physical world histories together with his proposed met-
ric provide a purported PRE-CD because it purportedly reductively explains
an asymmetry or arrow of counterfactual dependence to which causal direc-
tion reduces. Notice that in order for this account of causal direction to
be a PRE-CD, it cannot smuggle in obtaining causal relations (with their
directionality) into the reductive base.99 Let’s see if this is accomplished (and
here I follow D. K. Lewis 1979, 465–475; and the interpretations of him in
the critiques by Elga 2001, S315–S317, and Horwich 1987, 171–173).
Consider the following true causal claim,
Causal Claim (CC): The contracting orbit of PSR B1913+16 (the Hulse-
Taylor pulsar) at times t501, . . . , tn causally produced gravitational
radiation at a time tn+1.100
I will call the contracting orbit of the pulsar (the cause in this case), C,
and the production of gravitational waves or radiation (the effect), E. I will
assume, as D.K. Lewis (1979) does, that the actual world is deterministic,
and that the causation involved is therefore deterministic. I will call one of
the nearby deterministic possible worlds at which C and E do not obtain
(where ‘~C’ means that it is not the case that C occurs, and where ‘~E’
says that it is not the case that E occurs), w. And let ‘@’ pick out the actual
world (following Elga 2001). Lewis’s reductive strategy recommends that
we reduce CC to the counterfactual conditional,
(1) If it were the case that ~C, then it would be the case that ~E.
Conditional (1) comes out non-trivially true, given that w is closer to the
actual world than any ~C&E-world. Consider now the following false
causal claim,
Neutron101 Degeneration (ND): The contracting orbit of PSR B1913+16

(the Hulse-Taylor pulsar) at times t501, . . . , tn causally produced the
star’s own process of neutron degeneration plus its taking on a mass
much less than three solar masses at times tm, . . . , t499 [I’m denoting that
which is chiefly responsible for its neutron star or degenerate (pulsar)
status].
Call the imagined effect (i.e., the taking on of a certain mass much less than
three solar masses, and the process of neutron degeneration at the relevant
times) in this case E*, and notice that E* occurs before C. Let ‘~E*’ mean it
is not the case that E* occurs, and call a nearby (to the actual world) deter-
ministic ~C&~E*-world, w*. Lewis’s theory of causation reduces ND to,
(2) If it were the case that ~C, then it would be the case that ~E*.
Lewis would ensure that the later temporal event E is counterfactually

dependent upon the earlier event C and that the even earlier event E* is
not counterfactually dependent upon the temporally subsequent event C—
thereby ensuring that ND is false, although CC is true—by arguing that w
is closer to the actual world than w*. This is because while both w and w*
have the same laws as @, w’s history up until t500 (i.e., shortly before the
occurrence of C) matches the history of @ precisely. However, at t500, a small
miracle occurs (i.e., an event that violates a law of nature occurs), resulting
in the prevention of the contracting orbit of the Hulse-Taylor pulsar. No
further laws are violated, although w unfolds/evolves in a way dramatically
unlike @ subsequent to the prevention of the contracting orbit (e.g., E fails
to occur) because we are assuming, as Lewis did, that determinism holds at
all of the relevant worlds.102 The history of w therefore diverges from the
history of @, and that divergence is achieved via a small localized miracle
that precludes C from occurring.
With respect to w*, things are quite different. To ensure reasonable prox-
imity to @, at w* we have not divergence from @’s history, but convergence
to it just after tn+1 (the time of E at @). But for w* to match @’s history after
tn+1, a significantly large, widespread and diverse miracle is required. This
is because w* matches @ after tn+1, but because @ is a C-world, and w* is
a ~C-world, there will be traces or a type of shadowy residue of C in w*
even though C does not occur at w*.103 However, before tn+1 or before the
occurrence of the convergence miracle at w*, there are no traces of C in
w*. The miracle at tn+1 must be so significant that it makes a deterministic
~C-world evolve in such a way that it jumps/converges from a world with-
out any traces of C’s occurrence at the interval t501, . . . , tn, to a world with
significant shadowy C-like residue. Or as Elga put it in a similar context,
“the miracle in” w* “has to take care of making all of those (misleading)
traces,” and that “doctoring” requires “a big, widespread, diverse mira-
cle.”104 A picture of this is provided in Figure 3.3 (inspired by Elga (2001,
S316)).
When assessing closeness in this case, one must realize that (quoting Lewis)
“[d]ivergence from a world such as” @ “is easier than perfect convergence
to it.”105 The explanation for this lies in an asymmetry of overdetermina-
tion. Lewis thought that in order for a history such as w’s to diverge (as it
does) from @’s history, one needs only a small local miracle so as to “break
the links whereby certain past conditions determine that” the Hulse-Taylor
pulsar does indeed enjoy a contracted orbit.106 However, in order for w*’s
history “to converge” to @, “a world where” the Hulse-Taylor pulsar does
not enjoy a contracted orbit, “must break the links whereby a varied multi-
tude of future conditions vastly overdetermine that” the pulsar does indeed
enjoy a contracted orbit.107 Because Lewis’s metric for assessing similar-
ity disfavors large miracles, his account buttresses the judgment that w
is closer to the actual world than w*. This entails that it is not the case
that the ~C&~E*-world is closer to the actual world than any ~C&E*-
world because at w, E* occurs (the Hulse-Taylor pulsar became a pulsar
via the relevant processes), and yet at w, the relevant pulsar was, by a small
miracle, kept from enjoying a contracting orbit at times t501, . . . , tn. The
asymmetry of counterfactual dependence is explained via an asymmetry of
overdetermination.
Figure 3.3 Illustration of David Lewis’s Similarity Metric Applied
The preceding theory has been heavily criticized (see particularly J. Ben-
nett 1984; Elga 2001; Field 2003, 453–459; Horwich 1987, 171–173; and
Rescher 2007, 165–169). I have two altogether different worries than those
that have already been voiced. First, recall that we are working with an
interpretation of Lewis’s account, according to which it is a PRE-CD, and
that it must therefore proffer an explanation of the arrow of causation, and
therefore also (for Lewis) an explanation of the arrow of counterfactual
dependence that does not invoke causation. The problem is that Lewis’s
theory invites causation into the reductive base. The occurrence of the small
miracle at w that precludes the Hulse-Taylor pulsar from contracting its
orbit (C) is preventing C in a law violating manner. When an event occurs
that brings about an omission, there exists an instance of negative causation
by prevention. Lewis embraced causal prevention cases, and understood
omissions, as things that come into being when there is “the nonoccurrence
of any event of a certain sort.”108 Lewis attempted to account for causal
prevention cases by appeal to effects that feature in instances of positive
causation between events. For example, with regard to the fact of Xan-
thippe becoming a widow, he wrote,
I think there is no genuine event that can be called Xanthippe’s becom-

ing a widow. But the causes of her marriage together with the causes of
Socrates’s death may nevertheless be said to have caused her to become
a widow: they caused genuine events that comprised a pattern on which
the fact that she became a widow supervened.109
In the pulsar case, the occurrence of the small miracle caused events (call
them E1–En) to occur in C’s stead. Events E1–En constitute a pattern that
serves as a subvenient base for the omission of C. That Lewis thought of
miracles as causes of certain events seems clear. In one place he likens a
tiny localized miracle to “a few extra neurons” that “fire in some corner of
Nixon’s brain.”110 He then states that “as a result” of such firing, “Nixon
presses the button.”111 Surely, the sense of “result” here is causal. Surely,
the firing of Nixon’s neurons made him (in the sense that involves causa-
tion) do something.112 In fact, for Lewis, neuron firings are the quintes-
sential causal events (recall the use of neuron diagrams to illustrate causal
relations throughout the Lewisian corpus on causation). But if that is the
most plausible reading of the ways in which miracles produce positive
events that serve as the subvenient bases for omissions, then it looks as if
causation is being used in Lewis’s explanation of the asymmetry of counter-
factual dependence, and the asymmetry of causation. That is problematic
because Lewis’s account is purportedly reductive. A picture of this problem
is captured by Figure 3.4, where the arrows represent explanations, and
although one should read the top of each individual node first, the correct
ordering runs from (bottom) most fundamental explanans to most deriva-
tive explanandum.
The comparisons of histories in the reductive base involve obtaining
causal relations in so far as the nearest worlds involve miracles like neuron
firings that are preventions.
There is an out (and thanks to Barry Loewer here). PRE-CDs must ulti-
mately depend on physical history. Neuron firings and other higher-level
phenomena are not fundamental physical history. It is well-known that
Lewis defended a more general reductive hypothesis called Humeanism.
Non-fundamental features of the world such as obtaining causal relations
globally supervene upon categorical, qualitative, intrinsic, and natural fea-
tures of fundamental microphysical entities (perhaps space-time points, or
particles, or fields, or strings) together with the spatiotemporal relations in
which such microphysical entities stand (see D. K. Lewis, Plurality 1986,
14–15). Thus, what should ultimately enter the world-history comparisons
are the evolutions of these fundamental physical features of @, w, and w*.
Those evolutions will not involve instances of positive causation, and as a
But (at least some) miracles are causes and all causal
relations involve causal asymmetry. Ergo, we have
all over again at this node (at least with respect to
some miracles),
Figure 3.4 Illustration of Circularity
result, they will not directly underwrite supervening instances of negative

causation either. But this reply only works if causation does not enter fun-
damental microphysics. Yet, sect. 3.1 argued that causation does enter what
is fundamental to one of our best fundamental microphysical theories of the
strong force, viz., gluonic interactions in QCD. Thus, even if one restricts
world-comparisons to the fundamental microphysical features of possible
concrete worlds, both Lewis’s account of causation and his account of
causal direction fail to earn their reductive stripes. The fundamental physi-
cal history to which causation reduces is pregnant with causation.
Second, the explanandum that is the target of Lewis’s theory is the time-
asymmetry of a restricted class of instances of causation. Lewis is interested
in removing puzzlement about why it is that, in many (although not all)
cases, effects are temporally subsequent to their causes. The theory of the
asymmetry of counterfactual dependence he provides does not explain the
formal asymmetry of the instances of simultaneous causation discussed in
sect. 2. Lewis’s theory does not provide an explanation of (1) or an unne-
cessitated version thereof. In addition, if there were instances of backward
causation, Lewis’s theory would, for the same reasons, fail to explain the
formal asymmetry and priority on display in such cases as well.113 Thus,
given my first worry, if we were to grant that Lewis’s theory explains (1) or
an unnecessitated version of (1), it would not do so reductively. Given my
second objection, Lewis’s theory was never intended as an explanation of
(1) or its cousin. The Lewisian approach leaves (1) and its cousin without
a PRE-CD precisely because there are instances of simultaneous causation
such as those discussed in sect. 2.
Causation is asymmetric. There is privileged causal direction. I’ve argued
that this is not because causal direction is parasitic on temporal direction
because there are plausible cases of simultaneous causation in Newtonian
physics, relativistic physics, and photoelectric effects. In addition, causal
direction does not reduce to some direction in a non-causally interpreted
physics because what’s fundamental in one of our best quantum theories
should be interpreted causally. Very plausibly then, causal direction is brute. In
chapter 4, I will argue that primitivism about causal directionality does not
breed epistemological problems.
Notes
1. K. Bennett (2017, 68); De Muijnck (2003, 29–30); Eells (1991, 57); Hausman
(1998, on page 80 he writes “In chapter 12, I shall argue that asymmetry should
be built into the theory of causation.” Chapter 12 and 12* appear in 239–262);
Papineau (2013, 127, that causation is “asymmetric in time” (ibid.) entails that
it is formally asymmetric, for if a cause c always temporally precedes its effect
e, then given that c causes e, e cannot cause c because c precedes e (asymmetry).
In addition, c cannot cause itself because c is simultaneous with c (irreflexivity));
Paul and Hall (2013, 60–61, ask questions that presuppose causal asymmetry);
Tooley (1987, 179, 287, inter alios). Those who deny irreflexivity because of
the metaphysical possibility of causal loops must also deny the asymmetry of
causation. This is because asymmetry entails irreflexivity. I consider that a cost
of embracing the possibility of causal loops. I defend irreflexivity from the pos-
sibility of causal loops in chapter 6: sect. 2.3.
2. The necessity operators flanking the various formal claims about the causal rela-
tion are meant to be metaphysical necessity operators. The idea is that these claims
reflect the deep metaphysical nature of causation. In many places of the present
work, I represent various formal claims about the causal relation with a two-place
relational predicate, forsaking the use of universal plural quantifiers and variably
polyadic predicates. I do this for brevity and ease of comprehension. Chapter 7:
sect 2 maintains that causation is multigrade (see MacBride 2005, 578–588). Thus,
in addition to what’s stated in the text (q.v., the conjunction in the account of chap-
ter 9: sect. 4), I affirm versions of my various formal claims that are each appro-
priately characterized with a universal plural-quantifier and a variably polyadic
relational predicate for causation. The arguments I give for the explicitly stated
formal claims can be easily used to justify the formal plural-quantifier expressions.
3. She states that the relation is “functionally transitive” (Cartwright 2007, 192).
4. There are nuances with Hall. He thinks there are two different types of causa-
tion. One of these types, the central kind (Hall, Price of Transitivity 2004, 182),
is transitive. The other is not.
5. Although I should point out that Lewis differentiates between causal dependence
and the causal relation. He thinks that while the causal relation is transitive,
causal dependence is not.
6. “Many believe, however, that singular causation is transitive” (Hitchcock 1995,
276). “That causation is, necessarily, a transitive relation on events seems to
many a bedrock datum, one of the few indisputable a priori insights we have
into the workings of the concept” (Hall, Price of Transitivity 2004, 181; empha-
sis in the original).
7. On the motif of metaphysical explanation, I have in mind theories of ground-
ing (see the discussions in Dasgupta 2014; Fine 2001; Rosen 2010; and Schaffer
2009). I am assuming, however, that the grounding relations involved here are
non-causal, although I will argue in chapter 9: sect. 5.1 that instances of what
I call natural causation are instances of Schafferian grounding although not
merely that. Cf. K. Bennett (2017, 67–101). One might need to adjust these theo-
ries so as to allow for grounding between collections of entities and the truth of
propositions or statements. One might be able to do this by characterizing depen-
dence for truth in terms of truthmaking, and then by defining truthmaking in
terms of grounding. See the interesting discussion of the latter step in Koons and
Pickavance (2017, 55–58).
8. What are physical things? Tough question. The physical things I have in mind
(mostly) are things that belong to the ontologies of the full interpretations of our
best physical theories. There is more to say, but I will leave the rest to intuition
plus an acquaintance with our best physical theories.
9. Recall that Hume gave a reductio argument against the thesis that causes are not
temporally prior to their effects (see Ryan’s commentary in Ryan (2003)).
10. Proponents of (2) or an unnecessitated version thereof include Aronson (1971,
422); Frisch (2014, 234, he says that according to “two core properties of
causal structures,” arguably “causal structures are asymmetric and, in particu-
lar, [they] are time-asymmetric”); Heathcote (1989, 83); Kutach (2007, 328,
332); Reichenbach (1956); and Suppes (1970, 12, for prima facie causation).
Mackie (1974, 190) also attempted to account for causal direction by means of
temporal direction because he accounted for causal direction via fixity, and then
explained fixity by way of temporal priority. Many of these scholars espouse
reductive theories of causation that require a commitment to (2) or unnecessi-
tated (2). Dowe (2000, 187) says that “[t]he Humean position is a common one
in philosophy.”
The principle that ■∀x∀y (Cxy → (Sxy ∨ Txy))—where ‘Sxy’ means that x
is simultaneous with y—does not explain (1) or unnecessitated (1) (its cousin).
The directionality of causation is left unexplained in cases involving simultane-
ous causation.
11. I will grant, for the purposes of deliberation in this chapter, that events that are
causal relata are physical events.
12. An unnecessitated version of (2) might be able to explain an unnecessitated ver-
sion of (1). If one followed Dowe (2000) and others by proffering an empirical
analysis of causation and causal direction that purports to be, at best, a con-
tingent truth, then one’s research program might look to only explain ∀x∀y
(Cxy → ∼Cyx). A successful objection to unnecessitated (2) will therefore be a
substantial result.
13. Wald (1984, 8); Zee (2013, 146).
14. Newton (1999, 423) affirmed a principle of relativity in Corollary 5 of the Prin-
cipia. Roughly put, the idea is that the dynamical laws of motion—including
the universal law of gravitation—hold relative to inertial frames (although the
notion of an inertial frame came later), where the law of inertia is used to define
an inertial frame. Thus, I will assume that accelerations that result from forces
impressed are appreciated from the perspective of an inertial frame (a frame of
reference that is not itself accelerating, but is enjoying inertial motion by remain-
ing at rest or by maintaining a constant velocity in a straight line). Newton’s
dynamics takes place against a background that is Newtonian absolute space
with an absolute time parameter. I assume all that equipment is available at the
relevant possible worlds.
15. Misner, Thorne, and Wheeler (1973, 177).
16. Newton (1999, 405; italics removed). This appears to be the standard way to
read Newton on the matter as one can see in the histories: Jammer (1957, 121)
and Westfall (1971, 323). van Fraassen likewise attests to the causal reading of
Newton (van Fraassen 1989, 282).
17. Newton (1962, 148).
18. Newton (1999, 794; emphasis removed). See the discussion in De Gandt (1995,
271) and Smith (2002, 160). Westfall (1971, 380) tells us that Newton affirmed
“that forces are the ultimate causal agents in nature” and that that view explic-
itly emerges in his work on chemistry.
19. Newton (1999, 943). I am deliberately avoiding the debate about Newton’s
appeal to the aether in gravitational dynamics (see query 21 in Newton 1952).
20. Newton (2004, 125).
21. When I say that an interpretation is causal, I will mean that the causal relation shows
up somewhere in the ontology of the full interpretation of the theory in question.
22. There are other uses of ‘force’ and the related term ‘fort’ in Descartes’s corpus.
See on all of this Westfall (1971, 529–530 and the following discussion), who
says, “[b]y far the most common usage of the word [force in Descartes] was in
phrases such as ‘the force of its [a ball’s] motion’” (ibid.).
23. You see this in Descartes’s Le Monde or Descartes (1998); cf. (Schuster 2013,
71–72).
24. Galileo (Opere 1890–1909; 1967). In another place in the Discourses (see Gali-
leo, Dialogues 1914, 286), Sagredo carefully differentiates between the speed
of a body and force (forze), by regarding forces as that which causally gener-
ates momentum. Salviati seems to think that one important part of science is to
obtain “knowledge of a single fact . . . through a discovery of its causes” so as
to prepare “the mind to understand and ascertain other facts without need of
recourse to experiment” (ibid., p. 276).
25. Barrow (2013, 43). I had some help from Latin dictionaries.
26. See my comments above on previous work.
27. “In the radially inward direction, F = ma gives” (Eq. 4), according to Shankar
(2014, 109). I’m following Shankar’s (2014, 104–112) discussion closely includ-
ing his mathematical moves.
28. It will be important that we allow forces to have directions and magnitudes.
After all, they are represented by vector quantities. Directions and magnitudes
may be understood as structures in the entities that are forces, or as properties
those entities bear (qq.v., notes 29–31 for more on the metaphysics of quanti-
ties). Allowing for directionality helps us distinguish, for example, the force M
exerts on m, and the force (by Newton’s third law) m exerts on M. Newton
himself distinguished between two forces exerted in a system of two gravitating
bodies, although the forces are of the same type.
Newton did not believe that gravitation was somehow intrinsic to massive
bodies (see Newton 2004, 100–101, the correspondence with Bentley).
29. We can easily include other causally relevant factors by adding in events involv-
ing the direction and speed of the process of motion the planetary body is
involved in at time t0. That, together with its being impressed upon by a force
with a particular direction and magnitude, will more determinately affect the
motion the planet enjoys for times t > t0.
30. See on this distinction Mundy (1987, 30–32).
31. I greet the Platonism Mundy’s view invites with open arms. My account of the
causal interaction involved is also consistent with David Armstrong’s structural
universals view of quantitative properties, for which see Armstrong (Theory of
Universals 1978, 116–131; 1989, 101–107). Cf. the discussion in Eddon (2013).
32. Maudlin (2012, 127).
33. For a long time now, physicists have discussed and defended interpretations of
(Eq. 1) and (Eq. 2) that introduce infinite velocities. See Birkhoff (1944, 49);
Blair et al. (2012, 4); and Ferronsky (2016, 3, 286); among many others.
34. Smith writes,
To his contemporaries, what seemed most confusing about Newton’s way of

talking about forces was his willingness to put forward a theory of gravita-
tional ‘attraction’ without regard to the causal mechanism effecting it. They
generally concluded that he had to be committed to action at a distance as a
causal mechanism in its own right.
(G. Smith 2016, 203)
35. Newton voiced an aether-laden theory of the causal source of gravitation prob-
ably as far back as 1675 (Swenson 1972, 10). For more on Newton’s views of
the aether, gravitation, and light, see Westfall (1971, 394–395) and Whittaker
(1987, 19ff). Smeenk and Schliesser discuss a relational view of gravitation in
the work of Newton (Smeenk and Schliesser 2013, 160).
36. Newton (1999, 444).
37. Pourciau (2016, 110).
38. This reading isn’t new.
Newton and every other philosopher had taken it for granted that the force
of gravity was propagated instantaneously from bodies, and not in time like
the rays of light.
(Brewster 2009, 344)
39. Choquet-Bruhat (2009, 38).
40. Ibid. See also Wudka (2006, 132). Many more citations could be included.
41. And here my argumentation is meant to cut against even the view of Tooley
(1987, 208), who said that “it seems likely that our world does not, in fact,
contain any cases in which a cause and its effect are simultaneous.” Moreover,
a great many philosophers have stated that typically causes temporally precede
their effects. This will also come out false given the view of the Newtonian limit
in GTR defended here.
42. See Einstein (2002, 33). John Norton’s careful and rigorous 1989 study of the EP
in the Einstein corpus concludes,
Einstein’s principle of equivalence asserted that the properties of space that
manifest themselves in inertial effects are really the properties of a field struc-
ture in space; moreover this same structure also governs gravitational effects.
(Norton 1989, 40–41; emphasis in the original)
The plain reading of the previous statement suggests that space-time’s quali-
ties cause certain gravitational and inertial effects.
43. Einstein (2006, 186). See Ben-Menahem (1993, 294–295) for commentary. Ein-
stein goes on to accept Schlick’s argument that regularity is a necessary condition
for causation, but he does not think it is sufficient for causation (see ibid.).
44. See Einstein (2002, 339), where he said that the geodesic equation of motion
describes “the motion of a material particle under the action only of inertia and
gravitation” (emphasis mine). A little later in the same work Einstein says of
the geodesic equation of motion that it expresses “the influence of inertia and
gravitation upon the material particle” (341; emphasis mine). See also Einstein
(1997, 238) and Lehmkuhl (2014, 323) for commentary.
45. See Rovelli (2004, 77). Weinberg (1972, vii, although he seems dissatisfied with
that view).
46. Brown and Lehmkuhl (2013, 3).
47. Pooley (2013, 541).
48. Brown (2005, 161, at least prior to 1927); Brown and Lehmkuhl (2013, 2);
Hoefer 2009 (701–702 and especially n. 10 at 702); Pooley (2013, 541). Contra
the implicit point in Brown’s remarks, even after 1927, Einstein affirmed that, in
general, physics should be interpreted causally (see the quotation of Einstein in
Pais (1982, 465); cf. chapter 8: sect. 4.1 for more on this point).
49. “Gravity is one of the four fundamental interactions” (Hartle 2003, 3). See
chapter 8: sect. 4.
50. Janssen (2012, 160).
51. Geroch (1978, 180; emphasis mine).
52. They may be causal in two different senses. (Sense #1): Interactions are causal
producers of effects. (Sense #2): Interactions are a class of phenomena that
involve obtaining causal relations. Either (inclusive disjunction) sense #1 or
sense #2 works for my purposes.
53. Before continuing, I ask that the reader assume with me that in the present con-
text notions like ‘effect,’ ‘affect,’ ‘produce,’ ‘create,’ ‘collide,’ ‘absorb,’ ‘emit,’
‘destroy,’ and ‘make’ are causal notions in that both the relevant terms and the
concepts standing behind them are causal (more on what a causal concept is
below). I will not attempt to say precisely what a causal notion is. I think it is
somewhat obvious that certain notions are causal notions, and I believe we can
recognize good evidence for categorizing certain notions as causal notions.
54. Rovelli (1997, 193). Reminiscent of Hermann Weyl (1952, 220).
55. See note 27 in chapter 2 for a comment about my assumptions regarding deter-
minism, QM, and QFT.
56. See Khachatryan et al. (2016).
57. We could add to the prima facie case for causation in QCD by describing the
process of hadron production. That process involves the emission or radiation
of gluons by moving quarks (Ali and Kramer 2011, 250–251). Moreover, there
are processes described by QCD that involve the coupling of a collection of four
gluons such that (a) two members of the collection transform into two distinct
gluons, or (b) one member transmutes into three, or (c) three members of the col-
lection change into one member (thanks to physicist Tom Banks for help here).
58. Roessler (2011, 81); following Snowdon (2011, 125) on the notion of causal
concept1.
59. Sloman (2005, 71), who gives an example and cites studies by Andersson (1986)
and Driver, Guesne, and Tiberghien (1985, inter alios).
60. Gravetter and Wallnau (2013). “In the previous section, we introduced the
concept of an interaction as the unique effect produced by two factors work-
ing together” (ibid., 453). They give three characterizations of an interaction.
The second definition (ibid., 453) is explicitly causal and is an attempt to ana-
lyze the concept that represents the phenomenon of an interaction in the world.
Although the third definition (ibid., 454) is non-causal, it is also operational in
that it is an attempt to define a useful principle for the detection of interactions
modeled by graphs. The first definition is interpreted causally via the quote that
begins this note.
61. Kugler and Turvey (1987, 95). They continued, “indeed, in the interactions
addressed by particle mechanics are to be found the historical roots of the causal
principle” (ibid.).
62. Daintith (2005, 246).
63. Manipulationists/interventionists commonly assume that causal relata are the
values of variables, but one can easily understand events as relata, and the values
and variables as entities that indicate that an event occurred or failed to occur.
64. Quotations are from Woodward (2009, 250). He adds “or the probability distri-
bution of” E (Woodward 2003, 51).
65. Paraphrasing Woodward (2009, 248).
66. Ibid., 249.
67. Glennan (2017, 148).
68. We could add a further leg of support. I have argued that Newton and Einstein
interpreted their theories of gravitation causally. The fact that gravitation is an
interaction and that interactions are causal would explain very well the interpre-
tive judgments of those two eminent physicists.
69. In some anonymous comments, someone noted that successful causal action by
matter on space-time is enough to yield a problem for (2) because the entity
influenced is space-time itself. There can be no external time parameter accord-
ing to which there exists a finite time-difference between the occurrence of cause
and effect because of the nature of the effect itself. Restricting my reasoning to
GTR in the Newtonian limit is therefore unnecessary. I have two responses.
First, some relativity scholars reject the geometrization of gravitation, and so
do not believe that the metric or gravitational field reduces to space-time curva-
ture or space-time geometry (see e.g., Anderson 1996; Brown 2005; Brown and
Lehmkuhl 2013; Rovelli 1997; cf. Lehmkuhl 2014, on this motif (believed to be)
in Einstein’s own work). With respect to its dynamical behavior and interaction,
the gravitational/metric field given by gab is very much like matter fields, including
the electromagnetic field, although it couples with all other fields, never vanishes,
and exhibits other distinguishing properties. On this interpretation, space-time
is neither effect nor cause. However, my objection to (2) from GTR in the New-
tonian limit goes through even given the non-geometric interpretation. I argue
that the gravitational field’s causal action should be judged as simultaneous with
its effect for reasons that are independent of the geometrization of gravitation.
And although the main text speaks as if the gravitational field reduces to space-
time geometry, that manner of speaking is inessential to my argumentation.
(Although that reduction is contemporary orthodoxy. See S. M. Carroll 2004,
50; Choquet-Bruhat 2009, 39; Geroch 2013, 65; Hartle 2003, 13, inter alios)
Second, I am unsure of the cogency of the reasoning in full general relativistic
contexts. The propagation of gravitational radiation in a vacuum takes time. One
instance of time-varying change is the cause of the next. Importantly, on the geomet-
ric interpretation, gravitational waves are ripples of space-time curvature. So, if ever
there was a case of causally affecting space-time itself, it would be in such instances
of propagation. Nevertheless, the undulation processes take time, and like waves of
water, or electromagnetic waves, the undulations transpire one after the other with a
finite time-difference between them (see the helpful illustration of this at LIGO 2017).
70. Wald (1984, 76).
71. I am following the standard treatments in Wald (1984, 74–78); Misner, Thorne,
and Wheeler (1973, 412–416), although I have left out some formal details
explored in the sources just cited.
Note that GTR in the Newtonian limit does not include just static field cases.
Linearized gravitation treatments of the Newtonian limit allow for time-varying
changes in the gravitational field (S. M. Carroll 2004, 274).
72. Wald (1984, 76).
73. Both (Eq. 7) and (Eq. 8) are from Wald (1984, 77).
74. After all, Einstein said that a “‘a kinematic, nondynamic interpretation of gravi-
tation’ is not possible” (Einstein 1997, 238).
75. Einstein (1950, 109; emphasis mine).
76. Sometimes the reduction includes a mediating step. Causal direction reduces to
the direction of a certain type of counterfactual dependence, and that depen-
dence reduces to an appropriate asymmetry studied by physical science (entropic
increase with a special initial condition and a statistical postulate). See D.K.
Lewis (Causation 1986; Postscripts 1986; 1979) on the mediating step, and see
Albert (2000); Loewer (2007); and Kutach (2002, 2007) for good discussion of
other aspects of the general picture. See Frisch (2007); Kutach (2013, 250–252);
and Weslake (2014) for important criticism.
77. Compare now my argumentation to Heathcote (1989), with whom I agree in part.
Why have I chosen QCD? Because evidence suggests that it is probably not an
effective field theory but a real quantum field theory that is non-perturbatively
well-defined, and because physicists have suggested that if we seek improve-
ments to the standard model by seeking additions to it, those additions “will
be based on the paradigm established by QCD” (Roberts, Three Lectures on
Hadron Physics 2016, 2).
78. Jackiw (1980, 661). Perturbative methods are particularly relevant for QED,
where a great many experimentally important issues can be addressed via those
methods. As was noted in chapter 1, regarding phenomena peculiar to the high
momentum regime, perturbative methods are accurate and essential in QCD.
However, for the low momentum regime, neither semi-classical nor perturba-
tive methods are appropriated. Instead, lattice gauge theory is used. Thanks to
physicist Tom Banks for help here.
79. I have leaned heavily upon Baez and Muniain (1994, 167–168).
80. Weinberg (2005, 152). “[T]he SU(3) theory of the strong force is still regarded
as correct, and is part of the standard model” (Baez and Muniain 1994, 220).
“QCD works: there is no confirmed breakdown over an enormous energy
domain: 0 < E < 8TEV” (Roberts, Three Lectures on Hadron Physics 2016, 1;
emphasis in the original).
The color of a quark is a particular type of three-fold charge.
81. As A. Pais (1986, 588) puts it, “[t]hese forces act on quarks.”
82. On the evidence for the existence of gluons, see Riordan (1987, 335–354).
83. Gross and Wilczek (1973); Politzer (1973); Weinberg (2005, 153).
84. As Leader and Predazzi put it in their two-volume introduction to gauge theo-
ries, “[i]t is in fact these [self] interactions which are responsible for the property
of asymptotic freedom in QCD” (Leader and Predazzi 1996, 108).
85. Healey (2007, 148).
86. The account is standardly classified as a reductive one (see, e.g., Tooley
2003, 389).
87. Call the object constituent in fact C, α, and call the object constituent in fact E, β.
88. Quoting and following Dowe (2000, 171–172).
89. Dowe confesses that his third fork theory falls prey to Tooley’s (1987, 237) objec-
tion that “the direction of a causal process is made true by things extrinsic to the
process” (Dowe 2000, 205; emphasis in the original). I am grateful to Phil Dowe,
who in email correspondence 7/13/2017 helped me with Dowe-interpretation on
this matter.
90. This move is made by Schaffer (Causation and Laws 2008, 92), who writes “[o]nly
the reductionist can render causation fit to play a role in the foundations of special
relativity,” already having said “that causation disappears from sophisticated
physics.” If causation disappears from physics, then causation plays no role in
special relativity. As I will go on to argue in the main text, reducing causation in
physics does not eliminate it.
91. Given my reasoning in the main text, the demand that I address causal elimina-
tivist positions (as in Earman 1976; Mellor 2004, although he allows for causal
explanation; and Russell 1912–1913) is a red herring. There is no causal direc-
tionality to reduce if there are no causal relations.
92. Besides the Humean explanation of causal direction discussed in sect. 2, there
is at least one other stand-alone reductive explanation of causal direction called
the kaon theory of causal direction (as discussed in Dowe 1992, 189–195). That
theory suggests that the direction of causal processes is parasitic on the direction
of evolutions involving K° meson decay, a process that is not time-reversal invari-
ant given the C-P-T theorem (see Sachs 1987, 188–233). This process of decay
is described by our best theory of weak interactions. No one has shown that
the directionality of this process has anything to do with interactions described
by QCD, GTR, or the great many instances of macroscopic causation in the
special sciences. The consensus view is that meson decay is not a pervasive
enough phenomenon to underwrite the directionality of time, or causation (see
Mehlberg 1980, 183; North 2011, 315; Skyrms 1980, 120; Whitrow 1980).
Dowe has since rejected the view (Dowe 2000, 204).
93. See Beebee, Hitchcock, and Menzies (2009, 1); J. Carroll (2009, 285).
94. Cf. the argumentation in Rueger (1998, 34).
95. E.g., the studies of Lewis’s theories in Paul (2009) and Hitchcock (2015).
I should add that Lewis presented an account of indeterministic/chancy causa-
tion as well in (Postscripts 1986, 175–184). He rejected this account in Lewis,
Postscripts (2004).
96. Lewis allowed for causation by omission as well.
97. This is the account of the non-trivial truth-conditions for the ‘would’-counter-
factual defended in D.K. Lewis (Counterfactuals 1973).
98. The features of the metric for judging similarity (in order of importance) are as
follows:
Feature #1: Avoid large miracles or violations of law.
Feature #2: Ensure exact matches between histories.
Feature #3: Avoid small miracles.
Feature #4: “It is of little or no importance to secure approximate similarity
of particular fact, even in matters that concern us greatly” (D.K. Lewis
1979, 472).
99. Compare Field’s (2003, 454) remark, “[i]f causation is [reductively] explained
in terms of counterfactual dependence, counterfactual dependence must be
explained without use of the (full) notion of cause.” This means that Lewis’s
account of similarity between worlds cannot likewise smuggle in causation.
100. See Press Release (2017). This is an instance of matter’s primary causal influ-
ence on the gravitational field resulting in the production or emission of gravi-
tational radiation. Again, given that gravitation is a type of interaction, and
that gravitational radiation emission is a type of production (it is described
that way in many places, see e.g., Zee 2013, 567), the process referenced here
is causal (as I previously argued). Reductionists like Lewis will need to reduce
this causal fact to a non-causal fact.
With respect to this example, I am being imprecise with exact time spans or
intervals in a way that simplifies matters.
101. A pulsar is a rotating, dense, radiation-emitting neutron star that enjoys a high
degree of magnetization.
102. Determinism is the thesis that the exhaustive state of the world at a time t,
together with the laws, completely fixes how the world is at any other time (see
Loewer, Determinism 2008, 327; Mele 2009, 561; van Inwagen 1983, 2). If
two miracleless deterministic worlds with the same set of strictly deterministic
laws fail to agree on what transpires at a time t, they will fail to agree on what
transpires at every other time (agreeing with Elga 2001, S315).
103. Given determinism, and that causation is transitive, the residue in play could
be some of the evidence used to infer the existence of gravitational wave emis-
sion by the Hulse-Taylor pulsar. For example, at a later time, the pulsar would
enter a periastron stage with its partner star in the binary system to which
the pulsar belongs somewhat earlier if it emitted gravitational waves. At w*,
the pulsar enters that stage at the earlier time, but it does this without having
undergone the type of motion responsible for gravitational wave emission.
104. Elga (2001, S317). Elga’s layout of all of this is similar. I follow his example.
105. D.K. Lewis (1979, 473).
106. Quoted portions from D.K. Lewis (1979, 475).
107. Quoted portions from ibid.; emphasis mine.
108. D.K. Lewis (Postscripts 1986, 189).
109. Ibid.
110. D.K. Lewis (1979, 468).
111. Ibid.
112. Compare Woodward’s judgment that Lewis’s “‘small localized miracles’ . . .
function in broadly the same way as the notion of an intervention” (Wood-
ward 2003, 135). Interventions are causal.
113. This is admitted by Lewis (1979, 456–458). There he says that his story “might
break down in the different conditions that might obtain in a time machine, or
at the edge of a black hole, or before the Big Bang etc.” (ibid., 458). He also
states that the immediate past may be a problem in some cases (ibid.). This
suggests that simultaneous causation will be particularly troublesome.
4 On the Epistemological Isolation
Objection to Causal Hyperrealism
In chapter 3, I asked, why is causation formally asymmetric? Why is it true
that necessarily, for any event x and for any event y, if x causes y, then it is
not the case that y causes x? My answer is the causal hyperrealist one.
(Hyperrealism): Causal direction is brute. There is no explanation of the

asymmetry of causation because the causal relation cannot be reduced
to anything non-causal.1
There are at least three other potential answers to the question before us.
For example, the Humean answer you will remember states that,
(The Humean Answer): Causation is, by definition, temporally asym-

metric. Thus, necessarily, for any x and for any y, x causes y, only if x
is temporally prior to y. The temporal asymmetry of causation explains
its formal asymmetry.
Chapter 3 provided arguments from Newtonian mechanics, GTR, and the

photoelectric effect for the falsity of the Humean answer. One could add
the further problem that there are general relativistic space-times that allow
for past causal influence and/or backward causation.2 But here is another
potential answer,
(Perspectivalism): Causal direction (like causation more generally) is

perspectival, such that the world independent of contextually embed-
ded cognitive activity does not feature obtaining asymmetric causal
relations. Causation—but more specifically its directionality—arises
from the asymmetric nature of our deliberative practices. Thus, “the
asymmetry of causation is anthropocentric in origin.”3
I believe perspectivalism faces difficulties because it does not allow for

actual asymmetric causal pushing and pulling in the mind-independent
On the Epistemological Isolation Objection 153
world free of deliberating agents.4 The view therefore abandons the causal
realism defended in chapter 2. Apart from my discussion (in sect. 2.1) of a
connection between the agency-based theory of causation underlying per-
spectivalism and a causal principle, I will largely ignore the view in the
chapter before us.
Lastly, we have,
(Reductionism): Causal direction, like most other asymmetries, is ulti-

mately grounded in a “third arrow” that is an asymmetry of physics
(e.g., the entropy gradient coupled with the initial low-entropy macro-
state of the universe (the past hypothesis) with appropriate statistical
mechanical probability distributions for obtaining macroscopic states
of systems given the low entropy initial condition or past hypothesis).
Physicalism is a common motivator of this brand of reductionism.5 Reduc-

tionism is the view of David Albert, Barry Loewer, and others.6 The posi-
tion is closely allied with causal reductionism, which you will remember
is the view that causation (in general) reduces to non-causal law-governed
and non-causal physical history (again, see Schaffer, Causation and Laws
2008, 82–83; and Hall and Paul 2013, 10). Call the conjunction of reduc-
tionism about causal direction explicated above and causal reductionism,
strong reductionism. Both Albert and Loewer are strong reductionists.
I criticized proposals of the reductionist and strong reductionist varieties in
chapter 3, but for now I’d like to set those criticisms aside and ask whether
the debate between hyperrealists and reductionists about causal direction
can be tipped in favor of reductionism on epistemological grounds. Huw
Price and Brad Weslake (henceforth, P&W) (although not reductionists)
provide an argument that some will and do use to justify the thesis that it
can. They maintain that hyperrealism threatens “to make” causal direction
“epistemologically inaccessible,”7 musing that,
if . . . causal direction is detached from physics, then presumably the

world could have had the same physics, with an oppositely directed
causal arrow—in which case, apparently, we have no way of know-
ing whether our ordinary ascriptions of the terms ‘cause’ and ‘effect’
are correct or back to front. Perhaps the past actually depends on the
future. How could we tell?8
Elsewhere, Price stated,
The main difficulty with . . . [hyperrealism] is that in putting causation

and dependence beyond physics, it threatens to make them inaccessible
and irrelevant. The inaccessibility in question is epistemological: it seems
to be a consequence of this view that we simply have no way of know-
ing whether our ordinary ascriptions of the terms cause and effect are
154 On the Epistemological Isolation Objection
correct or back to front. Perhaps the past actually depends on the future,
and not vice versa—but how could we tell? . . . The only temporal asym-
metry that could possibly constitute evidence one way or the other is the
macroscopic thermodynamic asymmetry, and here, as we have seen, the
best explanation seems to lie not in causal connections between particu-
lar events . . . but in cosmological constraints on the boundary condi-
tions of the universe.
(Price 1996, 154; emphasis in the original)
This is the Price-Weslake objection (henceforth, P&W-O). Far from a state-

ment made in passing, the objection is (a) repeated in Price (2007), (b) used
to criticize Tim Maudlin’s view of time (ibid., 273, n. 18), and (c) featured
in an objection/dilemma to/for Cartwright’s (1979) position in Price (2007,
286). The argument therefore does considerable work for Price.9
To repeat for emphasis: Although P&W do not adopt reductionism,
they maintain that hyperrealism is unsatisfactory because of the P&W-O.10
Despite P&W’s rejection of reductionism and appropriation of perspectiv-
alism, it is quite natural for a reductionist to proffer the P&W-O as an
objection to hyperrealism, and so the P&W-O is a type of ammunition for
reductionists as well.11 In fact, something close to the worry is proffered by
reductionist David Z. Albert. Consider his view (a view he describes as obvi-
ously true) of how we acquire knowledge of what will transpire,
Everything we can know about the future (for example, the sun will rise
tomorrow, the ice in the glass on the table in front of me will soon be
melted . . . can in principle be deduced from nothing over and above the
dynamical equations of motion and the probability-distribution which
is uniform, on the standard measure, over the world’s present directly
surveyable condition.
(Albert 2000, 114)
If causation is ordinarily temporally asymmetric in addition to being for-

mally asymmetric, and causation does not reduce to law-governed (non-
causal) physical history (as hyperrealists maintain), then any causal fact that
reports on the causal production of some future effect by presently occurring
events will not be deducible “from . . . the dynamical equations of motion
and the probability-distribution . . . over the world’s present . . . condition.”
All such facts will therefore lie beyond the sphere of the knowable (well,
at least this will be true of higher-level causality, for I input causation into
physics (see chapters 1, 2, 3, 7, and 8)). Hence, Albert’s relegation of obtain-
ing causal relations that do not have their directedness reductively explained
by the asymmetry of the third arrow to the sphere of the “paltry, useless,
rare, uncontrollable,” and “undetectable.” (Albert, 2015, 60, in the mouth
of Huckleberry, although clearly representing his view; emphasis mine).12
The P&W-O seems to be in line with a venerable empiricist tradi-
tion that may stretch back at least to Hume. There is a growing bevy of
Humean scholars who argue that Hume should be read as espousing the
view that causation is not to be identified with regularity. Rather, causa-
tion exists as a real obtaining relation in the world beyond the immedi-
ately sensible. Causation is therefore not identical or reducible to that
which presents itself in our immediate experiences. As a consequence,
(quoting Strawson’s discussion) “we can never hope to form any empiri-
cally respectable grasp of its intrinsic nature.”13 Causation is a type of
elusive (quoting Hume) “power of force” responsible for regularity but
ultimately and “entirely concealed from us.”14 Hume added, “nature has
kept us at a great distance from all her secrets, . . . [and] conceals from
us those powers and principles upon which the influence of those objects
entirely depends.”15
The P&W-O therefore constitutes a substantial and important challenge
to hyperrealism. In what follows, I seek to confute five different versions
of the objection. Some ways of putting the argument (particularly those
statements of it in sect. 3 and sect. 4) are not attempted best interpretations
of the specific work of P&W. However, there is value in seeking to rebut
several different ways of formulating a general epistemological worry. The
hope is to put to rest objections to hyperrealism that are of the epistemologi-
cal isolation variety (generally) regardless of whether every argument of that
variety is a faithful interpretation of P&W.
In sect. 2, I advance three different underdetermination versions of the
P&W-O, subsequently showing that each version fails. Sect. 3 rebuts the
claim that hyperrealism implies a general causal skepticism, and sect. 4
wrestles with a P&W-like debunking argument for the thesis that without
the truth of reductionism one cannot acquire an explanation of our knowl-
edge of causal direction.
Section 2: Underdetermination Considerations
Section 2.1: UnderdeterminationK
P&W can be plausibly read as espousing an underdetermination argument.
When interpreting their own remarks, P&W stated, “hyperrealism entails
that causal facts [in general] are underdetermined by all available non-
causal evidence.”16 A reasonable characterization of underdetermination in
this context would amount to the following:
(UnderdeterminationK): Set S1 underdeterminesK a distinct set of propo-

sitions S2, just in case, knowledge of all members of S1 does not yield
knowledge of any one individual member of S2.17
And now the argument against hyperrealism can be formulated as follows,
(1) If hyperrealism is true, then every member of the set of all causal facts
is underdeterminedK by the set of “all available non-causal evidence.”18
Some may think that this premise is too strong to characterize P&W’s
argumentation. However, P&W spoke of the underdetermination of causal
facts in general when they interpreted their own worry (ibid., 418, n. 5). (I
address a P&W-O with a weaker version of this premise in sect. 2.2.)
(2) If every member of the set of all causal facts is underdeterminedK by the
set of “all available non-causal evidence,” then we cannot [actual inabil-
ity] come to inferentially know about the world’s causal structure.19
(3) Therefore, if hyperrealism is true, then we cannot come to inferentially
know about the world’s causal structure.
But does hyperrealism carry with it the imagined material implication

that is expressed by premise (1)? I believe it does not. Presumably, in this
context, non-causal evidence consists of non-causal facts, facts not directly
about some obtaining causal structure. But one can derive from the claim
that all purely contingent events (i.e., events that feature only contingent
substances) could be caused (itself a non-causal fact), the further claim
that all purely contingent events have causes (a causal fact) (q.v., chap-
ter 5).20 Here is the way you do it. Suppose one could treat causation like
a factive and distributive operator: “it is caused that . . .”. Suppose further
that causal relata are facts that can be prejacents of the causal operator
(neither assumption is essential; you could provide the same proof with a
relational predicate tying events together). Given these assumptions and
the thesis that all purely contingent facts could be caused, you can prove
that every purely contingent fact has a cause using reasoning analogous
to the Church–Fitch knowability result (see Church 2009; Fitch 1963;
Salerno, Introduction 2009; Knowability 2009), that, if every fact could
be known, then every fact is known. In the case of the knowability result,
the conclusion is thought to be paradoxical, but the result is acknowl-
edged as such (i.e., a result). The idea was that one ought to give up
the thesis that all facts could be known to save one from the embarrass-
ment that is the omniscience principle (i.e., that all truths are known). But
in the case of causation, is there a paradox? I do not believe so, especially
if the result is restricted to types of facts or events, such as purely contin-
gent facts or events.
One might disagree. That all purely contingent events or facts have causes
is paradoxical because there are spontaneous collapse interpretations of
quantum mechanics (e.g., the GRW interpretation, for which see Ghirardi,
Rimini, and Weber 1986; Ghirardi 2016; Goldstein, Tumulka, and Zanghí
2012, with continuous matter fields (mass densities) or flashes) according
to which the governing fundamental dynamics is not always and only given
∂ψ
by the linear and deterministic Schrödinger equation i  = Ĥψ . When
∂t
there are random collapses of the wave function, the governing dynamics is
stroboscopic, non-linear, and chancy (Bell 2004, 202–203). The chances of
a collapse are connected to the complexities of physical systems. For a sim-
ple quantum physical system (QPS) featuring but one particle, there exists a
probability per unit time for that individual particle to be hit by a collaps-
ing wave function (Allori et al. 2008). These collapses help ensure that
QPSes jump from superposed states to determinate states.21 These jumps
are indeterministic. It seems then that my conclusion (that all purely contin-
gent events/facts are caused—call this the universal causal principle or UCP)
breeds a paradox if we have strong empirical reasons for embracing GRW.
I have three responses. First, spontaneous collapse quantum theories are
not empirically equivalent to orthodox quantum mechanics. In fact, some
spontaneous collapse theories may already be empirically disconfirmed.22 As
Lev Vaidman has said, “the original GRW proposal was ruled out” by obser-
vation.23 The empirical case against such theories is mounting. Moreover
(quoting Richard Healey), “[n]o current ‘collapse’ theory can be considered
a serious alternative to the interacting quantum field theories of the Stan-
dard Model.”24 Second, even if GRW and other indeterministic approaches
to QM are not empirically sub-par, my result could be viewed as a type of
a priori evidence for a deterministic interpretation of QM, if the causation
involved is exclusively deterministic. Bohm-de Broglie mechanics is what
I have in mind (although q.v., n. 27 of chapter 2).25 Third, that every con-
tingent event has a cause does not by itself entail determinism. Some causal
chains might be composed of indeterministic causal connections. My rea-
soning, in this subsection, need not commit us to any specific deterministic
theory of causation. It may in fact leave room for the truth of indeterministic
physical theories.
If my causal result is correct (q.v., also chapter 5), then necessarily, if (CP
for causability principle) every purely contingent event could be caused, then
(UCP) all purely contingent events are caused. Claim CP, if known, does not
leave the UCP underdeterminedK because it entails it. But do we know CP?
Lest I be charged with relying upon controversial assumptions, I note here
that on a great many contemporary modal epistemologies, one can come
to know CP without relying upon inferential epistemic justification rooted
in causal facts. I have already sketched such a path to knowledge of CP on
(Road 1) the empiricist modal epistemology of Jenkins (2010), (Road 2) the
somewhat rationalist modal epistemology of Chalmers (2002), and (Road
3) the continuity based approach of Rasmussen (2014).26 I will not venture
into an exploration of yet another pathway that travels via modal episte-
mology. Instead, I ask the reader to consider the fact that P&W’s theory of
causation implies CP, and so any consideration in favor of adopting that
account will serve as indirect evidence for the CP.
According to Price (2007) and Price and Weslake (2009), every instance
of causation involves an intervention, and every instance of an interven-
tion involves hypothetical manipulation by a deliberating free acting
agent.27 More specifically, let event e1 be the cause of e2, and let event e2 be
the cause of event e3. An intervention involving manipulation (for P&W)
on e3 is (at least in part) characterized as a phenomenon in which a hypo-
thetical agent surgically influences e2 in a way that enables us to correctly
judge whether e3 “wiggles” as a result. Anything “that changes or ‘wiggles’
in these imagined cases, when” e2 “is ‘wiggled’, is regarded as an effect
of” e2.28 Most advocates of intervention-based theories of causation treat
interventions as causal processes or mechanisms. P&W should follow suit
and regard interventions involving manipulations as causal because the
imagined agents conducting the manipulation bring about events like e2.29
This has caused some to charge the AP-A with circularity. I will grant the
rejoinder (made by Price 2017, and Menzies and Price 1993, 193–195)
that we can acquire our concept of an acting agent or intervention inde-
pendent of our concept of causation. I am happy to allow for the direct
acquisition of the concept of an acting/intervening agent based on our
experience of behaving as agents do. As Menzies and Price described the
justification-contributor,
from an early age, we all have direct experience of acting as agents.

That is, we have direct experience not merely of the Humean succession
of events in the external world, but of a very special class of such suc-
cessions . . . we all have direct personal experience of doing one thing
and thence achieving another.
(Menzies and Price 1993, 194)
But notice that this acquisition story, which is consistent with hyperreal-
ism, does not imply that interventions or manipulations are non-causal.
Indeed, Menzies and Price go on to note how it specifies the sense in which
“the notion of causation . . . arises . . . from our experience of success
in . . . achieving our ends by acting in one way rather than another” (ibid.;
emphasis mine). The rejoinder therefore highlights the fact that interven-
tions of the manipulation variety are causal.30
What I will not grant is that the AP-A is a genuinely reductive conceptual
analysis of causation (Price, in 2017, claims that it was intended that way
in Menzies and Price 1993; cf. Woodward 2003, on the point I am about to
make). How can the concept of causation be reductively analyzed (at least
in part) in terms of the actions of agents when it is admitted that gleaning
the concept of an agent-centered intervention from our experience of acting
as agents in the world involves the acquisition of “the notion of causation”?
Clearly, the analysans contains a causal notion. The analysis is therefore
non-reductive and hyperrealist in the sense that the bringing about imagined
in manipulations is not conceptually reduced to the non-causal. Whether the
AP-A is uninformatively circular because of this is a separate matter.
The AP-A has consequences for the epistemology of causation (as noted
by Menzies and Price 1993, 192–193; cf. Price 2017, 75). If you want to
know whether an event c caused an event e, then you should ask if an agent
can intervene upon c in such a way that influencing it produces e. In your
effort to acquire causal knowledge, you would proceed this way even if it
were physically impossible for an agent to manipulate c. This is because
P&W’s AP-A is best thought of as providing counterfactuals that connect
“causal claims to claims about what would happen if certain manipula-
tions were performed.”31 Establish the relevant counterfactual and thereby
ensure causal knowledge or knowledge that some event has a cause. But
notice what this implication of the account entails. On the standard Lewis-
ian semantics for the ‘would-counterfactual’ (D.K. Lewis, Counterfactuals
1973), were it metaphysically impossible to manipulate c, the counter-
factual <If “certain [surgical] manipulations were performed” on c and e
“wiggles” as a result, it would be the case that c caused e.> comes out trivi-
ally true because its antecedent is impossible. That’s bad. Events that are
metaphysically impossible to manipulate should not be counted as causes
of e or any other event on the AP-A. The way out is to (a) restrict the AP-A
to relata that are purely contingent events, and (b) affirm that every purely
contingent event could (metaphysical possibility) be intervened upon.32 But
as I argued previously, the AP-A (and so also the AP-AM) says that inter-
ventions are causal. Thus, the AP-A/AP-AM implies that every purely con-
tingent event could be caused. At this point, the hyperrealist can argue that
either (a) the AP-AM is really a disguised hyperrealist analysis (after all,
the analysis is non-reductive; I argued for this above) whose characteriza-
tion of interventions and accompanying causal epistemology is correct, and
that we know it, (or) they can show (b) that they can acquire knowledge
of CP via (Road 1), (Road 2), or (Road 3). Knowing CP, and that that fact
entails UCP, is enough for one to come to know UCP. But is the anteced-
ent a real non-causal fact? It seems to me that it is. Its content is about the
modal nature of events, not some obtaining causal structure. Premise (1) is
therefore false.
One might insist that my assumed understanding of non-causal facts is far
too permissive. A fact’s failure to be about obtaining causal structure does
not make that fact non-causal. Rather, what makes a fact non-causal, is that
one can express that fact “in some language without using causal terminol-
ogy” (Field 2003, 443). This view of non-causal facts has two significant
problems in this context. First, the promulgator of the P&W-O in defense
of reductionism cannot appropriate it. On the assumption of either hyper-
realism or reductionism, it seems that to acquire inferential knowledge of
causal facts based on non-causal facts, one must have some understanding
of our causal concept.33 But then a precondition for inferential knowledge
of a causal fact is knowledge of our causal concept because understanding
p entails knowing p’s meaning. The object of that knowledge will inevitably
and inescapably be a causal fact in Field’s sense of the term because there
is no successful conceptual reduction of our causal concept to some other
non-causal concepts (as two reductionists Paul and Hall 2013, 249, and a
host of others have affirmed).34 Let me fill in the details of the argument I’m
sketching.
Assume that Tyler Burge is right (Burge 2009, 2010, 540; q.v., chapter 1:
sect. 4.4.1).35 Concepts are part of the representational contents of propo-
sitions. For me to know that p, I must have grasped the concepts consti-
tutive of part of the representational content of p (if knowing p involves
understanding p’s content). When one cogently or soundly, inductively or
deductively, infers that p based on some other truths in a way that involves
one’s coming to know that p, one exercises a skill involving the exploitation
of a capacity to reason and think about p. Thus, cases involving inferential
knowledge of a causal fact based on non-causal facts will very plausibly
include competent use (in thought) of concepts that are part of p’s content.
I therefore recommend the following,
(Conceptual Condition for Inferential Causal Knowledge): For any

human cognizer C, if C inferentially knows a causal fact p (e.g.) of
the form (CF) <Concrete particular O is an obtaining causal relation.>,
on the basis of some non-causal facts, then (a) C grasped the causal
concept, (b) has some understanding of the causal concept, and (c) can
competently use the causal concept in thought.
But the fact that C’s relation to and use of the concept of causation satisfies
conditions (a)–(c) is strong evidence that C knows how to use the concept of
causation for the purposes of conceptual attribution. Thus, a precondition
for C’s reasoning to a causal fact such as (CF) on the basis of non-causal
facts is knowledge of how to use that concept. But on intellectualism about
knowing-how (see Stanley 2011), knowing how to do something involves
propositional knowledge or knowledge-that. Knowing how to employ and
use our causal concept at least involves knowing that there’s an avenue that
is a possible process of thought involving one’s competently using the con-
cept of causation. The involved propositional knowledge includes knowl-
edge of correct answers to questions concerning how to employ the concept
of causation in thought.
The assumption that is intellectualism is, like the two that have come
before it, inessential. If you do not like it, then it will be enough for my
purposes to recognize that a precondition for inferentially knowing a causal
fact is having some understanding of our causal concept. Understanding
requires knowing, although understanding p does not require knowing p.
If the above reasoning is correct, then if P&W appropriate Field’s under-
standing of non-causal facts so as to explicate the meaning of non-causal
facts/evidence in (1)–(3), causal facts will remain underdeterminedK by non-
causal evidence even if we affirm reductionism. And that is because moving
in a knowledge-conducive inferential way from non-causal facts to causal
facts requires a prior knowledge of our causal concept, and that prior con-
ceptual knowledge, on Field’s view of causal facts, entails knowledge of a
causal fact. Thus, one will not be able to inferentially know some causal
fact on the basis of purely non-causal facts if Field is right. That means that
even given reductionism, the set of all non-causal facts underdeterminesK the
set of causal facts. Given Field’s view of non-causal facts, you cannot move
from knowledge of non-causal facts alone to knowledge of causal facts.
One will always have to appropriate knowledge of causal facts (particularly
knowledge of the causal concept).
Second, Field’s characterization is somewhat vague. What are the bound-
aries around the set of all “causal terminology”? Presumably terms like
‘push,’ ‘pull,’ ‘produce,’ ‘create,’ ‘make,’ ‘interact,’ ‘influence,’ ‘act upon,’
‘decide,’ ‘form,’ ‘build,’ ‘ignite,’ etc., are causal notions at least in a great
many (even scientific) contexts. But what about notions such as ‘propagate,’
‘undulate,’ ‘evolve,’ ‘annihilate,’ ‘reproduce,’ ‘collapse,’ ‘collide,’ ‘raise,’
‘lower,’ ‘change,’ etc.? These also appear to be causal notions. But no one
would challenge that one important modus operandi of physics is to pro-
vide descriptions and explanations of evolutions of physical systems involv-
ing undulating fields (e.g., Maxwell’s equations), collapsing entities such
as wave functions (e.g., the GRW interpretation of quantum mechanics),
propagating waves (e.g., gravitational radiation equations), and/or interact-
ing fields and particles, etc. (e.g., Einstein’s field equations and the geodesic
equations of motion). In other words, evolutions seem to involve causa-
tion. As Stephen Hawking observed, “if state A evolved into state B, one
could say that A caused B.”36 But recall that causal reductionism states that
obtaining causal relations reduce to law-governed non-causal physical his-
tory, where physical history is impregnated by evolutions of physical sys-
tems, systems involving phenomena not unlike that which has already been
mentioned (i.e., propagation, radiation, interaction). It seems then, that on
Field’s way of bifurcating the causal and non-causal, it is no surprise that the
set of non-causal facts underdeterminesK the set of causal facts because even
given causal reductionism, that underdetermination remains in place. This
is because knowledge of the underlying physical facts involves knowledge
of causal facts. The fundamental physical facts are facts about law-governed
evolutions featuring undulating fields, propagating waves, and interacting
particles. Thus, if the underdeterminationK version of the P&W-O is some-
thing to fear, reductionists, causal reductionists, and hyperrealists all share
the burden of avoiding the object of that fear.37
The objection that the formalisms of physical theories do not themselves
include notions like ‘undulating,’ ‘propagating,’ ‘interacting,’ etc., is both
false and irrelevant. It is false because some of our most fundamental physi-
cal theories do indeed include notions that stand for interactions.38 But the
objection is also irrelevant because as was pointed out in chapter 1: sect.
4.5.5, physical theories aren’t just their formalisms.
Suppose that a non-causal fact is a fact that one can understand without
employing any causal concepts in thought (following the recommendation
of an anonymous individual). Call this way of understanding causal and
non-causal facts the Conceptual View (CV). Given CV, the proposition that
<All purely contingent events could be caused.> is a causal fact because
to understand it, one would need to employ the concept of causation in
thought. But this way of bifurcating between causal facts and non-causal
facts will fall prey to the first objection I lodged against Field’s choice way
of distinguishing causal facts from non-causal facts. For suppose we let a
choice causal fact (understood in the CV sense) be CF, and let a choice set of
non-causal facts be NF. I have argued that to acquire inferential knowledge
of CF from NF, one will need prior knowledge of the causal concept. That
means that the psychologically real inference or justification supports rela-
tions extending to CF will run, not from NF alone to CF, but instead from NF
plus various facts about our causal concept to CF. However, facts about our
causal concept are causal facts according to CV. For example, to understand
the following claims (assume they are true) about our causal concept, one
will need to employ causal concepts:
(a) The concept of causation is a theoretical concept.
You will need some understanding of the concept of causation to relate it to

the class of concepts that are theoretical.
(b) Causation is asymmetric.
You will need the concept of causation to relate the relation it represents
to the formal property that is asymmetry. Thus, to understand (b), one will
need the concept of causation.
(c) If event E counterfactually depends on event C, then C causes E.
You will need the concept of causation to understand counterfactual depen-

dence as a sufficient condition for causation (the relation the concept repre-
sents). It seems clear that facts about our causal concept are causal facts on
CV. Does this mean we can never acquire our causal concept without first
having a causal concept? No. I am currently concerned with gaining under-
standing of causal facts. I am not concerned with causal concept acquisition.
We may very well be able to acquire the causal concept through having cer-
tain experiences (see my previous discussion of Menzies and Price) alone, but
the question of whether we can come to initially possess (acquire) the causal
concept independent of considering certain causal facts is quite different
from the question of whether we can understand causal facts about the con-
cept of causation without already having a causal concept in mental hand.
Thus, if my response to Field is correct, then if P&W appropriate CV so
as to explicate the meaning of non-causal facts/evidence in (1)–(3), causal
facts will remain underdeterminedK by non-causal evidence even if we affirm
reductionism. And that is because moving in a knowledge-conducive infer-
ential way from non-causal facts to causal facts requires a prior knowledge
of our causal concept, and that prior conceptual knowledge, on CV’s view
of causal facts, entails knowledge of a causal fact. Therefore, one will not be
able to inferentially know some causal fact on the basis of purely non-causal
facts, if CV is right. That means that even given reductionism, the set of all
non-causal facts underdeterminesK the set of causal facts.
Section 2.2: A Restricted UnderdeterminationK Argument

Perhaps the underdeterminationK argument should be restricted,
(1*) If hyperrealism is true, then every member of the set of all causal direc-
tion facts is underdeterminedK by the set of “all available non-causal
evidence.”
(2*) If every member of the set of all causal direction facts is underdeter-
minedK by the set of “all available non-causal evidence,” then we can-
not [actual inability] come to inferentially know about the world’s
causal structure.
(3*) Therefore, if hyperrealism is true, then we cannot come to inferentially
Say that a causal direction fact is one of the form: Event c caused event e,
and e did not cause c, that is, c is causally prior to e.
Premise (1*) is, like (1), false. There are instances of knowledge provided
by a source for it called testimony (q.v., D16 from chapter 1: sect. 4.5.4).
None of us would have been able to acquire our natural languages, or our
ability to speak those languages, without that source (Audi 2011, 150; cf.
Coady 1992). And I want to confess at the start that my objection to (1*)
requires no specific theory of testimony as a source of knowledge. It can
run on a variegated assortment of theories of testimony and belief acquisi-
tion.39 It can likewise run on multifarious theories of precisely how it is
that testimony provides cognizers with epistemic justification, warrant, or
knowledge.
Suppose that a large group of witnesses or people (i) you know to be
credible testify to you that James (or James’s match strike act) caused a fire
spark. You (ii) regard those acts of testimony as sincere factual attempts to
represent the world. You form the belief that James caused the fire spark
on the basis of their testimony, and on the basis of (i) and (ii). Using your
knowledge of the causal concept (more specifically, that causation is for-
mally asymmetric, and perhaps also transitive), you infer that,
(A) <James’s match strike caused the fire spark, and the fire spark did not
cause James’s act of match striking.> (a causal direction fact)
I am not being sly. I argued before that the hyperrealist must be allowed to
incorporate into the set of “all available non-causal evidence” knowledge
of our causal concept. But does this inference transfer knowledge-conducive
epistemic justification to your belief that (A) is the case? Assume that (A) is
true. Assume also that you know (i) and (ii). I take it that the inference to
(A) is now a quite strong knowledge-conducive warrantK preserving infer-
ence. The reductionist will ask, but how is it that the testifiers came to know
that James caused the fire spark? The best reply in this context is that they
came to know that fact based on their perceptual experience, and therefore
in one of the ways I lay out in detail in sect. 3. They saw James strike the
match and witnessed the resulting spark. They subsequently formed a per-
ceptual causal belief on the basis of their experience.
Is this case an instance of inferential knowledge acquisition of a causal
direction fact? Consider what facts played a role in the inference to (A).
They were (i), (ii), and that causation is asymmetric (and perhaps also tran-
sitive). None of these are causal facts. None of these are true propositions
directly about obtaining causal structure. The story I have told is consistent
with hyperrealism. Why believe that? Well, it seems clear that no feature or
proposition that is part of the preceding testimonial story is incompatible
with the definition of hyperrealism. To undermine my claim about compat-
ibility, the reductionist would need to advance some proposition or set of
propositions that show that the conjunction that is <Hyperrealism is true
and the above testimonial story of inferential knowledge of causation is
true.> is not possible. No reductionist (or perspectivalist for that matter) has
accomplished such a feat. The story therefore provides the hyperrealist with
a reason to reject (1*).40
Section 2.3: Contrastive Underdetermination

We have seen that on my view of the causal/non-causal fact distinction, non-
causal facts do not leave causal facts underdeterminedK due to the causal
analog of the Church–Fitch knowability result. If one bans the CP from the
domain of non-causal facts by using a different causal/non-causal fact dis-
tinction (such as Field’s or CV’s), both reductionists and causal reductionists
will be unable to overcome a respective underdeterminationK formulation of
the P&W-O. That is a bad consequence because the P&W-O was supposed
to be an objection-type uniquely applicable to hyperrealism. In addition,
I argued that the hyperrealist can actually acquire inferential knowledge
of causal direction facts due, in large part, to the veracity of testimony as
a legitimate source of knowledge. Those direction facts are not underdeter-
minedK by all “available non-causal evidence,” given hyperrealism.
The reductionist can press on in several ways. She might take issue with
how the underdetermination thesis was stated. She can urge that we use
a different conception of underdetermination. Perhaps a Duhemian theory
of contrastive underdetermination will do the trick (see Duhem 1954 and
Bonk 2008). On that understanding, there are two hypotheses, H1 and H2.
H1 says that an event c caused event e. H2 says that event e caused event c.
With Duhemian contrastive underdetermination, the P&W-O will say that
if hyperrealism holds, then no amount of available non-causal evidence E
confirms H1 over H2, where H1 is confirmed by E over H2, given that Pr(E/
H1) > Pr(E/H2). The argument is now,
(4) If hyperrealism is true, then with respect to any obtaining causal rela-
tion involving an event c, and another event e, the hypothesis that c
caused e (H1) and the hypothesis that e caused c (H2) are contrastively
underdetermined by “all available non-causal evidence.”
(5) If, with respect to any obtaining causal relation involving an event c,
and an event e, the hypothesis that c caused e (H1) and the hypothesis
that e caused c (H2) are contrastively underdetermined by “all available
non-causal evidence,” then we cannot [actual inability] come to infer-
entially know about the world’s causal structure.
(6) Therefore, if hyperrealism is true, then we cannot come to inferentially
Is (4) true? Grant P&W and both the reductionist causal reductionist
what they desire. Physics is best interpreted non-causally.41 Both general
relativity and the standard Λ-CDM (cosmological constant plus cold dark
matter) model it affords are extremely successful physical theories (call their
conjunction standard cosmology or SC). Insert those theories into the set of
“all available non-causal evidence.” Again, given the cogency of the reason-
ing used to respond to the P&W-O in sect. 2.1, the hyperrealist must be
allowed to incorporate some knowledge of our causal concept into the set
of “all available non-causal evidence.” That knowledge should involve the
understanding that causation is formally asymmetric and ordinarily tem-
porally asymmetric.42 We can now break contrastive underdetermination.
According to SC, the universe originated in a primordial big bang
singularity—understood as a point of infinite matter density, infinite
space-time curvature, and arbitrarily high energy density—indicative of
geodesic incompleteness.43 This big bang constitutes a space-time bound-
ary in that “[i]t represents the creation of the universe from a singular
state, not an explosion of matter into a pre-existing space-time.”44
Just after the big bang, the cosmos was in thermal equilibrium and elec-
trons, neutrinos (i.e., those endowed with a certain chemical potential),
and photons possessed generalized blackbody distribution.45 As the tem-
perature of the cosmos changed, there was neutrino decoupling (Lyth and
Liddle 2009, 56–59). After that decoupling and more temperature change,
the thermal history of the cosmos unfolded in the way that is described by
big bang nucleosynthesis (BBN), a process that decorated the cosmos with
important nuclei for various elements (see Weinberg 1993, 101–121). BBN
began only 10 seconds after the big bang. It ended about 21 minutes later.
To capture subsequent evolution, SC has it that space-time expands
from the big bang and continues to (at an accelerated rate) expand after
BBN. Roughly put, when Einstein’s field equations are applied to the
Robertson-Walker metric (i.e., put the metric into the EFEs) and when an
appropriate energy-momentum tensor is likewise applied to the EFEs, the
results are the two independent Friedmann equations that give the dynamics
of cosmic expansion,
 a 4π G Λ
(Eq. 1):   = −
 a
( ρ + 3p) + governing the universe’s accelerated
3 3
expansion.
2
 a  8π Gρ k Λ
(Eq. 2): H 2 =   = − 2 + , which provides the value for the
 a 3 a 3
Hubble constant (H), the rate of the universe’s expansion. H is here
8π Gρ k
related to matter density , space-time curvature 2 , and dark
Λ 3 a
energy .46
3
SC populates the cosmos with cold dark matter and a time-independent
dark energy density sometimes suggested as that which determines the accel-
erated cosmic expansion. The preceding equipment, together with the total
matter density, the spectrum and spectral index of the CMBR, and the
baryon density is all that is needed to account for an immense amount of
relevant cosmological data. As Lyth and Liddle remarked,
With its six parameters, the simplest version of the ΛCDM model [one
part of SC] seems able at the time of writing [2009] to explain all rel-
evant types of observation, some of which have order 1% accuracy.
(Lyth and Liddle 2009, 60)
SC and facts of stellar evolution imply that starlight emission involves a

radiation dissipation process that produces an entropic increase. That pro-
cess is not time-reversal invariant. A fortiori, the phenomenon contributes a
tiny (usually ignored) temporal directedness/asymmetry to the macroscopic
approximation that yields Eq. 1 and Eq. 2. The stress-energy tensor (Tμv)
is, for this reason, not time-reversal invariant, and so neither are the Fried-
mann equations because they depend in part on matter (thanks to theoreti-
cal physicist Don Page for help here). One must preclude this phenomenon
and add delicate assumptions about the form of the stress-energy tensor to
make them invariant under time-reversal.
Throw out the argument just provided if you do not like it. In my presen-
tation of SC, I described the expansion as accelerated expansion. Virtually
all contemporary cosmologists affirm this tenet of the model, and the 2011
Nobel prize was given for work related to it (see Riess et al. 1998; Riess
et al. 2004; Perlmutter et al. 1999). If our universe expands at an accelerated
rate, and its matter density is as observations indicate, then it will not col-
lapse into a big crunch, but will instead continue to expand forever into the
future, approaching de Sitter space-time. The conclusion we can now infer is
that our universe has a cosmological asymmetry that is the arrow of expan-
sion. This holds even given the assumption that the Friedmann equations
are time-reversal invariant together with invariance under time-reversal for
an appropriate equation of state (thanks to theoretical physicists Tom Banks
and Don Page for helpful discussion on these points).
Although SC lays over our space-time manifold a continuous timelike
vector field τa, that condition merely implies that space-time is time ori-
entable. However, given that there are two such fields τa and τ a’ (as SC
suggests) related by τ aτ a’ > 0 , at every space-time point, the nonvanishing
continuous timelike vector fields represented by τa and τ a’ will be co-
oriented. That relation holding between those fields is an equivalence
relation. With it are two equivalence classes. The choice of predicating
to one such class ‘being the future’ and the other ‘being the past’ time-
orients our general relativistic space-time (Malament, Topics in the Foun-
dations of GR 2012, 131–132). SC does this. It privileges one class as
future and the other as past. That choice suggests that the cosmological
asymmetry previously discussed entails a future expansion approaching
de Sitter space-time, and a past that approaches the big bang singularity.
We therefore have strong reasons—reasons that in the absence of defeat-
ers look like knowledge-conducive epistemic justification or warrant—for
believing that our space-time is time-oriented, and not just time-orientable.
The justification issues forth from the empirical successes of SC.47
According to SC then, the universe evolved and will continue to evolve from
early initial conditions of the big bang to a future expanded state that is the
heat death of the cosmos. With knowledge of our causal concept in hand, con-
firming one causal hypothesis over another should be straightforward. On the
non-causally interpreted SC, the current distribution of the cosmic microwave
background radiation (CMBR) is non-causally due to the dynamics of expan-
sion, and the big bang (inter alia). The hyperrealist is well within her epistemic
rights in claiming that the CMBR distribution was caused by earlier cosmic
evolution as described by SC because hyperrealists commonly connect causa-
tion to laws of nature, and remember that the laws in view are either not time-
reversal invariant or else their solutions are not time-reversal invariant.48
One way, among many ways to forge the connection between laws and
causation, is by adopting the hypothesis that causation is an intrinsic relation
(I am now adding to what we know about the causal concept, and I defend
an addition like this in chapter 9: sect. 4) such that at the actual world @
there exists a causal structure S that stretches back into time (although not
indefinitely), and at some nearby possible world w there is a structure very
similar to S, S*. In S, c causally produces e, and in S* intrinsic duplicates of c
and e (viz., c* and e*, respectively) stand in a causal relation as well, if S* is
an intrinsic duplicate of S, and @ and w have the same laws (Hall, Intrinsic
2004, 264). The reductionist should not refuse the intrinsicness thesis because
causal reductionism is often used to motivate it (that is in fact how Hall
motivates it). The position is, however, consistent with hyperrealism, a fact
Hall admits (ibid., 257–258).
Now let hypothesis H1 be the claim that the current distribution of the
CMBR is causally due to the primordial big bang event, and let H2 be the
hypothesis that the current distribution (space-time location) of the CMBR
(by itself) is causally responsible for the big bang event. But again, because
we know that SC is approximately true or that it can at least be understood
as non-causal evidence, we know or can appropriate as non-causal evidence
that the dynamics of expansion has a preferred temporal direction (past to
future). And because our knowledge of the causal concept suggests that cau-
sation is ordinarily temporally asymmetric, ceteris paribus, E will be more
probable on H1 than H2.
We can break the contrastive underdetermination in a second way with
non-causally interpreted physics and knowledge of our causal concept. There
are no known laws of physics that could give you an evolution that begins
with the mere distribution of electromagnetic radiation (solely) and ends with
the big bang singularity. It is, in fact, physically impossible to evolve from the
mere space-time location of electromagnetic radiation, and that alone, to the
big bang singularity. Any such evolution would require a great deal more
(gravitation, etc.), and statistical mechanics tells us that any such evolution
that takes you from a highly entropic state through a very low-entropic one
(our universe began in a low-entropic state) is exceedingly unlikely (recall
that SC rules out the Gold universe that features symmetric entropic increase
on either side of the special low density state of the cosmos). Again, E is more
probable on H1 over against H2. Contrastive underdetermination is broken
by available non-causal evidence and knowledge of our causal concept.
Section 3: Causal Skepticism

I have argued that three different underdetermination versions of the
P&W-O fail. However, the reductionist could revert to an epistemological
objection (which is in the spirit, although not the letter of the P&W-O) to
hyperrealism, which states that without a successful reduction of causation
in the background, we cannot come to know the causal structure of the
world.49 Simply put, the idea is that hyperrealism implies causal skepticism.
But I think that conditional is false. There are many epistemologies of causa-
tion that are consistent with hyperrealism. Indeed, Price himself has detailed
a means whereby cognizers like us acquire causal knowledge independent of
conscious reflection and inference that is consistent with hyperrealism. But
let me explicate a more detailed and hyperrealist-consistent theory of how
we attain at least some of our causal knowledge.
Suppose that one day while waiting for a train, some stranger surprises
you by gently pushing your forehead back with his or her hand (an often-
discussed case). Assume there is no resulting injury or harm involved. Con-
sider now the following claim,
(1) My [referring to you the reader] forehead is being pushed backward (i.e.,
caused to move backward, where the directionality is in the pushing).
Nothing about hyperrealism implies the falsity of Mooreanism with respect

to one’s knowledge of (1). The Moorean defense affirms that any argument
for the falsity of (1) is going to rest on premises that are less plausible than a
denial of that self-same argument’s conclusion. The hyperrealist can affirm
that we do in fact know propositions like (1), and that we come to know
causal truths like it from the mother’s knee.50 As David Danks wrote,
We are clearly ‘causal cognizers’, as we easily and automatically (try to)

learn the causal structure of the world, use causal knowledge to make
decisions and predictions, generate explanations using our beliefs about
the causal structure of the world, and use causal knowledge in many
other ways.51
James Woodward noted,
[S]mall children and nonhuman animals distinguish between causal

relationships and noncausal correlational relationships. It is also uncon-
troversial that adult humans in preliterate, scientifically unsophisticated
societies share many of our causal beliefs, have a notion of causation
that seems much like ours, and engage in counterfactual reasoning.52
How precisely we acquire causal knowledge is a matter that can be settled by

an appropriate theory of the epistemology of causation.53 That epistemology
might recommend that cognizers like us perceptually behold some obtain-
ing causal relations (following a gaggle of scholars, including Armstrong
1968; 1988, 225; 1997, 213–215; Cartwright 1993, 426–427; Ducasse
1968, 25–28; Fales 1990, 15–46; Locke 1975; Mumford and Anjum 2011,
196–202; Reid 1983; and Siegel 2009).54
It seems that in the circumstances surrounding one’s apprehension of (1),
one perceptually beholds an obtaining causal relation, a relation-instance,
a concrete state of affairs, and an interaction. One thereby comes to pos-
sess a perceptual belief in the sense, and in the way already adumbrated in
chapter 1: sect. 4.5.1 (leaning also on the theory of warrant articulated in
chapter 1: sect. 4.3),55 although our understanding of how such beliefs get
warranted or justified need not be distinctively externalist.56
If I were to insert myself into the relevant environment featuring the
pushing, the detractor of perceptually beholding obtaining causal relations
might ask why I believe that my perceptual experience presents me with an
obtaining causal relation. But I should note that although I wouldn’t neces-
sarily automatically form the perceptual causal belief about my experience,
the belief I do form (we are supposing in this case) has (1) for its content,
not a more sophisticated belief about what concrete relation instances the
experience presents me with. I can be acquainted with objects in my experi-
ences and even form the belief that such objects exist based on that acquain-
tance and yet not form more exacting beliefs about features of those objects,
or how those objects relate to my experience. I believe this is one lesson to
draw from Ernest Sosa’s speckled hen case (Sosa 2003). Perhaps the question
is a request for reasons why I believe my experience has the content it does.
But again, there is no guarantee that during the pushing or causing, I would
form the belief that my experience has certain content.57 Moreover, given
my views on warrant, it is not necessary for my being warranted that I be
aware of reasons for such a belief were I to have formed it in the relevant
circumstances.
If I come to acquire causal knowledge by forming a perceptual belief
with (1) as its content, then presumably my belief must enjoy warrant of
some kind. So how does it acquire its warrant? There is no reason why the
hyperrealist cannot at this juncture proffer any one of several responses in
the epistemology literature. For example (and I’ve already gestured at this
option), the hyperrealist can affirm that their belief is warranted in a foun-
dational way, a way that involves non-inferential warrant.58 I simply find
myself believing, as a result of properly functioning cognitive faculities in
the right kind of environment, that the push caused my head to fling back-
ward. I believe that proposition based on my perceptual experience, and the
warrant-contributor just is the perceptual experience itself.
There are many other ways of achieving basic perceptual causal knowl-
edge that are consistent with hyperrealism. Proper functionalism is not
required. Here is one for good measure. You can be justified in believing
that a certain causal relation obtains in the virtue foundational way (Sosa
2007, 44–69; 2009, 154–177; 2011, 74–90).59 A belief b enjoys virtue foun-
dational justification when its foundational justification issues forth from b
being formed in a manner that manifests “a certain epistemic competence”
(in this case a perceptual epistemic competence) “one that is not consti-
tuted by” basing b “on some other conscious state/reason for which” one
formed b.60 An epistemic competence is an ability, and an ability is a dis-
position. These dispositions should be (or at least can be) linked to certain
sets or collections of conditionals with “triggering antecedents and mani-
festing consequents.”61 But more specifically, perceptual epistemic compe-
tencies consist of skillful abilities to discriminate between falsehood and
truth.62 Sosa’s account is famously externalist, and so cognizers need not
demonstrate or be strongly aware of the fact that their beliefs were formed
in the relevant competence manifesting ways in order for those beliefs to
enjoy virtue foundational justification. On hyperrealism, I see no reason
why we cannot believe based on perceptual experience that some causal
relation obtains and claim to be manifesting the aforementioned perceptual
epistemic competence. The mere report that we could be in some skeptical
scenario in which causal structure is radically different (e.g., the effect is the
cause, and the cause the effect) does nothing to defeat that claim and such a
scenario itself can be met by the virtue epistemologist’s response to skepti-
cism in general.63
Our replies need not be distinctively externalist either. There are many
internalist responses to the same problems that are all completely consistent
with hyperrealism about causal direction (see e.g., BonJour 2003; McGrew
and McGrew 2007; etc.). My point here is therefore quite general. It does
not rest on any specific controversial assumptions. Hyperrealists do not
have a distinctive epistemological problem any more than anyone else does.
We can (actual ability) and do come to know about causal structure irre-
spective of causation’s detachment from physics. Skeptical scenarios can be
dealt with by using sound traditional epistemology, whether externalist or
internalist.64
Section 4: A Debunking Argument

The last and perhaps most interesting way for a reductionist to run an epis-
temological worry for hyperrealism in the spirit (although not in the letter)
of the P&W-O is as follows. Without a reduction of causal directionality
to some non-causal priority and directionality in physics, one would not
have an explanation for our knowledge of causal direction. In other words,
grant that we have knowledge of causal direction, but note that a necessary
condition for such knowledge is the truth of the reduction of causal priority
to some non-causal physical priority. The objection resembles a debunking
argument against hyperrealism.
Debunking arguments are those that rob cognizers of their rationality.
They do this by showing that a cognizer’s belief b fails somehow to be prop-
erly related to that which it is about. The connection in view is thought by
some to be an alethic explanatory one (Korman 2014, 2). Here is something
like the idea in play,
(1) A cognizer C’s belief b has an alethic explanation, just in case, b is

true, and some factsΣ F1–Fn stand in relation R to b’s content, and F1–Fn
explain the true proposition that <C formed b>.65
Perhaps (1) is too strong. Sometimes we form beliefs based on reasons we

have. Let us therefore attenuate (1) some,
(2) A cognizer C’s belief b has an alethic explanation, just in case, b is true,
and some factsΣ F1–Fn stand in relation R to b’s content, and F1–Fn par-
tially explain the true proposition that <C formed b>.
The relation that is R may be nothing over and above an aboutness relation.
That is to say, b’s content is about F1–Fn. Or perhaps the requisite factsΣ
make b’s content true in the sense that those facts de re necessitate the truth
of b’s content, where an entity O de re necessitates a proposition p if and
only if O’s existence entails (in non-de dicto fashion) p. Or perhaps these
facts ground or otherwise metaphysically explain b’s truth or the truth of
b’s content. There seem to me to be many ways one could cash out the idea
here. Nothing terribly substantial will hang on which precise option one
chooses.
With (2) in hand, consider the following epistemological principle,
(3) A cognizer rationally forms their belief b, or rationally retains their

belief b, only if b has an alethic explanation.66
The P&W-O-like debunking objection can now be outfitted with the above
equipment. Their problem with divorcing causal asymmetry from an asym-
metry in non-causal physics is that our beliefs about causal direction will
not have an appropriate alethic explanation, and so by consequence, one
will be irrational for forming or retaining one’s beliefs about causal priority
or causal direction. But is this right? If causal direction/priority/asymmetry
is not itself reduced to or grounded in non-causal facts about the arrow
of entropic increase or some other suitable reductive physical base, does it
really follow that our beliefs about causal directionality are not alethically
explained?
Brandon throws a rock at a window (event c). The window breaks (event
e). I (cognizer C) perceptually behold this in my rock-throwing competition
with Brandon (see the picture painted above about how C could acquire
epistemic justification or warrant for perceptual beliefs about causal facts).
I form the belief that c caused e, and perhaps also that e did not cause c. The
causal relation runs from c to e. My belief’s content is therefore <c caused e
and c is causally prior to e.> (call this proposition D). Notice that my belief
that D is true, and there really are some factsΣ F1–Fn that b’s content (D) is
about, viz., there being the events c and e, their being causally related to one
another, and the causation running from c to e, and not from e to c. These
are ways the world is or was. They at least partially explain the fact that
I formed a belief with D’s content (if that idea makes sense at all; see my
complaint below). But maybe the proponent of the P&W-O-like debunking
argument will want to demand that there ought to exist a fundamental or
micro-physical partial explanation of the event of my forming the belief that
D. Proposition (2) should therefore be read as,

true, and some fundamental physical factsΣ F1–Fn stand in relation R
to b’s content, and F1–Fn partially explain the true proposition that <C
formed b.>.
Given the assumption that there is no causation in physics, it is difficult to

see how there could exist fundamental physical factsΣ F1–Fn that stand in any
one of the ways of cashing out relation R, to my belief that D, and that these
facts related existential true propositions about them also partially explain
my forming b. Moreover, for instances of causation in the special sciences,
hyperrealism implies that there are no fundamental physical facts of the
relevant kind that explain the directionality of those instances.
But there is a problem with this account (i.e., the conjunction that is (3)
and (4)) of the proposed necessary condition for rational belief formation or
retainment. Explanation is a factive relation. For any p and q, p explains q,
entails q. But it is also true that for any p and q, p explains q, entails p. How
could a falsehood or something that is not truth-evaluable at all, explain
a truth (q.v., chapter 1: sect. 3 and especially n. 77)? The explainer must
be a truth, and so it must be truth-evaluable. Partial explanation relates
propositions, just as full explanation does. That is why a sufficient amount
of partial explanations of q yield the explanation of q, and so the sufficient
partial explanations will be true, just as that which is explained is true.
But if the partial explanations are true, then they must be truth-evaluable.
Proposition (4) asserts that concrete states of affairs (factsΣ) partially explain
a proposition. Concrete states cannot be bearers of truth. They may make
certain propositions true, but they are not themselves truth-bearers.
What then of this substitute principle?

true, and some fundamental physical factsΣ F1–Fn stand in relation R
to b’s content, and F1–Fn de re necessitate the true proposition that <C
formed b>.
The mere existence of the concrete states necessitate the truth of the propo-
sition <C formed b.>.
Given (5), proposition (3) seems implausible. Are all instances of men-
tal causation involving belief formation as an effect necessitated? Couldn’t
we be rational and free in the libertarian sense with respect to some of
the beliefs we form? Moreover, couldn’t we be rational in circumstances in
which the relationship between the relevant concrete states and the proposi-
tion <C formed belief b.> is such that those facts render the proposition <C
formed belief b.> highly likely, although not 1? These considerations suggest
that we can do better than (5).
(6) A cognizer C’s belief b has an alethic explanation, just in case, b is true,
and some fundamental physical factsΣ F1–Fn stand in relation R (where
R is now something other than the truthmaking relation) to b’s content,
and F1–Fn de re necessitate that b is true.
This account of alethic explanation, together with our initial principle about
rationality (proposition (3)), entails that a necessary condition for rational
belief formation or belief retainment is that the belief have fundamental
physical factsΣ that make that belief true (in the truthmaker sense) because
many understand truthmakers to just be worldly facts that de re necessitate
proposition-facts (see e.g., the discussion of these matters in Merricks 2007,
14, although Merricks is no truthmaker theorist). But now the proponent of
this account of the debunking P&W-O-like objection (i.e., the one with (3)
and (6) underwriting it) has become a truthmaker theorist with respect to
all of our true rational beliefs. They will therefore inherit almost all of the
well-known difficulties of truthmaker theory, including the following:
• What would make true negative existentials like <Unicorns do not

exist.> true?
• What would make true counterfactual conditional claims like <If I trav-
eled at the speed of light, my velocity would be approximately 186,000
miles per second.>?
• In this context, true beliefs about modal claims, like <Necessarily, 2 is
even.>, or <Necessarily, 2 + 2 = 4.> would seem to be problematic because
these would all have to have truthmakers that reside in the physical.
Grant that there are plausible responses to all of these questions (after
all, I allow for the appropriation of a truthmaker theory of some kind in
chapter 1). Things are considerably worse for the debunker. The cost is
not just a truthmaker theory for all rational beliefs, it is a microphysicalist
truthmaker theory for all rational beliefs. All rational beliefs must be made
true by underlying fundamental or microphysical factsΣ. Given that we seem
to have lots of rational beliefs about a great many extremely complicated
macroscopic physical systems (e.g., baseball games and Olympic swimming
events), and even some systems that may not be physical at all (e.g., men-
tal goings-on, or certain facts of moral obligation), the debunker has come
awfully close to what many typically mean by physicalism.67 The debate has
therefore shifted to the question of whether physicalism, or something near
enough (call it approximate physicalism), is true. I take it that that question
is still very much an open one.
If one established the truth of approximate physicalism and that phys-
ics is to be non-causally interpreted, one would have very strong reasons
for rejecting hyperrealism. P&W grant that we have rational beliefs about
causal structure that constitute knowledge. They are here being interpreted
as challenging the hyperrealist to provide an explanation of that causal
and rational knowledge. Thus, if one showed that the content of all of our
rational causal beliefs (where that content consists of causal factsP) is made
true by non-causal microphysical factsΣ, or that causation is grounded in or
reduced to non-causal microphysical entities and evolutions, then one has
provided strong reasons for the falsity of hyperrealism.
Perhaps principle (3) should be revised to read,
(7) Given that a cognizer C forms belief b or retains belief b and C is con-
fronted with the challenge of providing an alethic explanation for b,
and C is unable to proffer an alethic explanation of b, then C irratio-
nally formed their belief b, or irrationally retains their belief b.
But given that alethic explanations are understood in terms of (6) above,
(7) becomes overly burdensome. If to be challenged in relevant respects
amounts to simply being questioned in the appropriate way, then it would
be quite easy for a great many of us to be rendered irrational by our col-
leagues who know our beliefs (and quite quickly!). I doubt that any of us
are able to articulate complete metaphysical (and microphysicalist) explana-
tions for most of the contents of our beliefs. I can and do form, quite eas-
ily, complicated beliefs about extremely detailed goings-on at my favorite
restaurant. I’m asked by my extremely intelligent wife for an alethic expla-
nation of those beliefs. I’m thereby, if the aforementioned story is correct,
rendered irrational. The goings-on are far too complicated for me to provide
the actual microphysical facts that make the content of my beliefs true by
de re necessitating them (i.e., their content). Something seems wrong with
this picture. After all (we will grant), in the relevant scenario, my cognitive
faculties were functioning properly in a very amicable environment for their
proper function. And it seems right to say that I was rational, despite being
unable to report on the precise and actual dynamics of the involved funda-
mental quantum fields and the like.
I have argued that hyperrealism about causal direction faces no substantial
or particularly worrying epistemological isolation objections of a variety
weakly or strongly related to (Price 1996, 2007, and Price and Weslake
2009). In sect. 2, we saw how arguments from underdetermination can be
defeated by (a) appropriating knowledge of important logical results that
connect non-causal facts to causal facts, (b) appealing to the reliable tes-
timony of cognizers who perceptually behold obtaining causal relations in
their experiences, or (c) by noting how hyperrealists can connect causation
to laws of nature without embracing reductionism and then use knowledge
of non-causally interpreted physics to break contrastive underdetermination
of causal facts by non-causal facts. Sect. 3 established that no P&W-O-like
argument that seeks to connect hyperrealism to causal skepticism will work
because a great many causal epistemologies are consistent with hyperreal-
ism. And lastly, sect. 4 showed that a P&W-O-like worry, understood as a
debunking argument, ends up begging the question against hyperrealism
because it requires the truth of physicalism, or else it is overly demanding.
Reductionists should not appropriate epistemological objections to causal
hyperrealism.
Notes
1. The term ‘hyperrealism’ comes to us from Price (1996, 154), although he associ-
ates it with the position that microphysics is acausal. I do not intend to invoke
that association in my discussion of the view. There are, nonetheless, obtain-
ing causal relations with causal direction that floats free of the micro-physical
world (this is because there are instances of causation in the special sciences and
those instances feature brute causal directionality). Interestingly, Price identi-
fies Tooley’s (1987) position as a hyperrealist one (in the sense he has in mind).
I believe this is a mistake. Causation, for Tooley, “is that theoretical relation,”
neither properly analyzable in non-causal terms nor reducible to the non-causal,
that fixes the arrow “of the logical transmission of probabilities” (ibid., 251;
emphasis in the original). Tooley’s view (the “third alternative”) requires that
causal relations always be backed by laws. These will be microphysical laws
when the states of affairs (causal relata proper for Tooley) are those in the micro-
physical domain. It is therefore difficult to see how causation disappears from,
e.g., quantum microphysics on Tooley’s view. Hartry Field (2003, 443) classifies
Nancy Cartwright’s (1979) position as a hyperrealist one (in Price’s sense). How-
ever, Field attributes to Cartwright the view that there’s no causation in physics
because (on Field’s understanding of Cartwright) the laws of physics are strictly
non-causal laws reporting on functional dependencies or associations. “[P]hys-
ics” (continuing with Field’s reading of Cartwright) “leaves out,” for example,
causal statements about forces producing motion (Field 2003, 443). This read-
ing is inaccurate. There are, for Cartwright, objective (i.e., mind-independent)
causal laws of physics that are required for accurately discerning effective strate-
gies, and that “cannot be done away with” (Cartwright 1979, 419). This reading
fits with Cartwright’s (1983) views in How the Laws of Physics Lie, articulated
not many years later. As one interpreter of her work summarized matters, “Cart-
wright’s arguments go to show that only causal laws, and some high-level phe-
nomenological laws in physics, can be held to be literally true, even in a restricted
domain of application” (Hoefer 2008, 3; emphasis mine).
Some have labeled Tim Maudlin a hyperrealist about causation. This is incor-
rect. Maudlin explicitly disowns causal hyperrealism in Maudlin (2007, 143–
169). Thanks to Tim Maudlin for confirming my reading on this matter.
2. E.g., the space-times described by the Gödel metric (see Earman 1995, 160–202,
specifically p. 164, where backward causation of the kind that doesn’t involve
energy-momentum transference or exchange is said to be admissible although
not required by instances of Gödelian time travel). These space-times may bother
the reader. My paper assumes the asymmetry of causation. If causation is asym-
metric, it will also be irreflexive. However, some general relativistic space-times
with closed-time-like curves (CTCs), or closed-causal-curves (CCCs) (like those
described by the Gödel metric) appear to allow for causal loops that violate irre-
flexivity. Nonetheless, one can secure my criticism of Hume with no more than
general relativistic space-times that include causal structure that violates what
Robert Wald has called the strong causality condition. Space-times of this type
can allow for genuine backward causal influence, although there are no causal
loops because there are no CTCs or CCCs in such space-times (see Wald 1984,
196–197, and see Fig. 8.8 there). It is not at all unscientific to preclude from
the realm of genuine nomological possibility space-times captured by legitimate
solutions to Einstein’s field equations. Physicists commonly do that to avoid par-
adoxes (see e.g., Friedman et al. 1990; cf. Hawking 1992). Cf. my discussion in
chapter 8: sect. 6.2.
3. Price (1996, 10). See also Healey (1983); Price (1996, 132–161; 2007; 2017);
Price and Weslake (2009); Ramsey (1931, 237–255). Cf. the critiques in Kutach
(2013, 252–254) and Frisch (2014, 228–233).
4. “The perspectival solution to this problem is to say that the asymmetry of causa-
tion is a kind of projection of some internal asymmetry in us, rather than a real
asymmetry in the world” (Price 1996, 158).
5. Indeed, Price and Weslake (2009, 416) motivate their objection to hyperrealism
by using what they call “the physicalist constraint” (ibid.; q.v., n. 11 regarding
reductionism and perspectivalism).
6. See Albert (2000, 2015) and Loewer (2007, 2012) for the best statements and
defenses. See Frisch (2007; 2014, 201–228); Kutach (2013, 250–252); and
Weslake (2014) for important criticism.
7. Price and Weslake (2009, 417).
8. Price and Weslake (2009, 417–418). One might read P&W as scholars interested
in explaining only the temporal asymmetry of causation. That reading is incor-
rect. P&W are clearly after an explanation of the formal asymmetry of causation
as well. They remarked, “it is our perspective as deliberators that underpins the
distinction between cause and effect” (ibid., 419; second emphasis mine).
9. The worry is a serious one. In fact, Tooley’s (1987, 296–303) book-length
defense of hyperrealism (as I understand the term) anticipated epistemological
objections of the kind to which the P&W-O belongs. However, his responses
require the truth of his theory of causation. In subsequent discussion I state that
there are P&W-Os for reductionists and causal reductionists. What this means is
that there are versions of the P&W-O-type that are about those positions rather
than hyperrealism.
10. P&W add that hyperrealism cannot explain the time-asymmetry of deliberation.
11. Kutach (2013, 254) argues that perspectivalism is best understood as a type of
reductionism. And because Price (2007, 289) seems to think of the more funda-
mental aspects of reality as aspects that correspond to a “bare Humean world”
empty of causation, Price (2007), and even Price and Weslake (2009), may bet-
ter be understood as reductionists with an additional story to tell about how to
account for the directedness of deliberative practices. There is, however, some
tension between P&W’s underlying agency/perspectivalist analysis of causation
and reductionism.
12. In addition, Albert (2015) (and Loewer 2007) clearly believe that those causal
relations that we can come to know about have an asymmetry that is reductively
explained by a temporal asymmetry, and that that temporal asymmetry is reduc-
tively explained by the third arrow.
13. Strawson (2015, 97). Cf. the literature cited in ibid. for more on this reading.
14. Hume (1987, 472).
15. Hume (1975, 33; emphasis mine).
16. Price and Weslake (2009, 418, n. 5).
17. Paraphrasing Douven (2014, 336), specifically the characterization of <know,
know> underdetermination articulated there.
18. Price and Weslake (2009, 418, n. 5).
19. Why the qualification “inferentially”? In sect. 3, I sketch a path to non-inferen-
tial knowledge of obtaining causal relations, where non-inferential knowledge
involves knowledge-conducive non-inferential warrantK. If that path travels all
the way to genuine causal knowledge, then the hyperrealist could grant that
the set of all causal facts is underdeterminedK by the set of “all available non-
causal evidence,” and yet show (via sect. 3) that cognizers like us know about
the world’s causal structure nonetheless. Premise (2), absent the “inferentially”
qualification, would thereby become defeated. So as not to render the underde-
termination versions of the P&W-O susceptible to the same objection I lodge
against the causal skeptical version of the P&W-O in sect. 3, I have inserted
‘inferentially’ into statements (2), (3), (2*), (3*), (5), and (6).
20. Although I criticize one small feature of the derivation in Weaver (2013) in chap-
ter 5, I should note here that I still believe that the basic argument of that work
is correct, and any potential problems are completely remedied by my arguments
in chapter 5 of the present work.
21. Albert (2000, 149); cf. North (2011, 333). Regarding more complicated systems,
such as macroscopic bodies, one might ask, do the collapses hit every composing
particle at the same time resulting in (for example) a determinate localization
of the entire body? No. But because there exists strong correlation between the
composing particles, a spontaneous hit on just some of the constituent particles
of a single macroscopic body b will result in the localization of b (see the discus-
sions of this in P. Lewis 2016, 52, and Vaidman 2014, 14).
22. Some believe that GRW may violate energy conservation (Pearle 1984; cf. the
discussion in Adler 2007).
23. Vaidman (2014, 17). See Collett et al. (1995, inter alios).
24. Healey (2014, 364). Elsewhere, Healey has said, “one can use the terminology
of measurement in describing a hypothetical experiment whose outcome, as pre-
dicted by quantum theory, differs from the prediction of a well-defined collapse
theory.” He continued in the footnote, “[o]ne such experiment would involve
interference of objects much larger then [sic.] C60 molecules” (both quotations in
this note are from Healey 2017, 109).
25. See Albert (1992, 134–179); Allori et al. (2008, 355–356); Dürr, Goldstein, and
Zanghí (1992). As was hinted at previously, there are Bohmian versions of QFT
(see the discussion in Dürr et al. 2014).
26. See Weaver (2016); and Rasmussen and Weaver (forthcoming).
27. For more on the agency/perspectival analysis of causation (henceforth AP-A),
see Healey (1983); Price (1996, 132–161; 2007); Price and Weslake (2009, 429–
439); and Ramsey (1931, 237–255).
28. Price (2007, 268).
29. That interventions are causal is the majority view in the literature. See Meek and
Glymour (1994, 1008), where some interventions are understood as volitional
decisions, e.g., to refrain from smoking; Pearl (2009, 29), where local interven-
tions are said to be actions like fixing prices and turning on sprinklers; Woodward
(2016, sect. 6; 2003, 94); and Woodward and Hitchcock (2003, 9, inter alios),
who treat interventions as causal processes. See chapter 9: sect. 2.
30. Price does not reject the rejoinder, but instead endorses it, writing, “I think this
reply . . . stands up on its own terms” (Price 2017, 76).
Price’s (2017) additional response to the circularity worry says that his analy-
sis of causation can be regarded as a piece of philosophical anthropology. But
even if we were to grant that fact, it would not falsify my claim that interven-
tions on the AP-A are causal.
31. Woodward (2009, 236; emphasis in the original). I quote Woodward here
because Price (2017, 89) himself does so approvingly in response to an objection
from Woodward.
32. I recently discovered that Woodward uses the preceding type of consideration
as a reason to abandon Lewis’s semantics for counterfactuals (Woodward 2016,
sect. 5). Q.v., my discussion in chapter 9: sect. 2.
We restrict the AP-A to purely contingent events so as to avoid certain
unwanted paradoxes. Call this modified AP-A, AP-AM. AP-AM would seem to
capture all that P&W wanted the AP-A to account for, including the asymmetry
of our deliberative practices, because the entire natural world is contingent.
An anonymous person asked me, how could it be metaphysically possible for
one to intervene or manipulate the collection of all agents? I have two responses.
First, is the mere existence of such a collection an event? On some accounts it is
not (see, e.g., Lombard (1986) and Mumford and Anjum (2011)). Second, if the
collection were an event restricted to agents that are contingent substances, then
if there could be a necessary agent, then there could be an intervention upon the
existence of all such contingent agents. One might think this has taken us too
close to theism, but perhaps this is a cost (or benefit?) of saving the view from
the absurdity.
An anonymous person asked me, how could it be metaphysically possible for
one to intervene or manipulate the beginning of the universe? There are scientific
accounts of how our universe was produced by a mother multiverse (see the
discussion in Carroll and Chen 2004). Such multiverse models seem at least
metaphysically possible. Second, if theism is at least coherent, then there could
be a cause of such a beginning.
33. As was made clear in chapter 1: sect. 4.4.1 of this project, I assume a mental
representation theory of concepts as part of a representational theory of mind
(see Fodor 1987; 2003; Pinker 1995). These assumptions are not necessary.
There are plenty of other theories of concepts and the mind that afford my rea-
soning. I add them to help facilitate understanding only.
34. Recall that not even the AP-A(M) is a truly reductive analysis of causation. And
so if the hyperrealist used it to get knowledge of CP, the point in the main text
would still stand. One might now counter that using the AP-A(M) to secure
knowledge of CP is cheating because the challenge of the P&W-O is to move
from non-causal evidence/facts to knowledge of causal facts. But as I am about
to go on to argue, the reductionist must allow into the set of “all available non-
causal evidence,” knowledge of our causal concept. The AP-A(M) purports to be
an analysis of our causal concept. If it is correct, it should be an element of the
aforementioned set. Ergo, using it to gain knowledge of CP is not cheating.
35. This particular assumption is not essential. I can sketch the same argument on
other views about the relationship between concepts and propositions. I use
Burge’s account because (a) it is one of the most well-developed in the literature,
(b) I think it is true, and (c) it helps facilitate understanding of my argument best.
36. Hawking (1994, 346). He would, of course, go on to say that it is possible
to evolve the state in the other direction, but that possibility does not rob the
quoted point of substance.
37. Q.v., my discussion in chapter 3: sect. 2.
38. Q.v., chapter 3.
39. Here are just two examples. First, one could maintain that beliefs formed on the
basis of testimony are inferentially acquired in a manner involving supplemen-
tary premises (as I will assume holds for at least the case I provided previously).
One could appropriate the view that testimony is a more direct source requiring
no additional premises. Nothing essential hangs on one’s choice.
40. The perceptive reader will notice that my reasons for rejecting (1*) also consti-
tute reasons for rejecting (1). If you did not like my response to (1), regard my
response to (1*) as a defeater for (1).
I should add that the claims about hyperrealism and compatibility should
be read into all future assertions of the form “epistemology x . . . is consistent
with hyperrealism.” That is to say, when I go on to claim that the stories of
sect. 3 (for example) are compatible or consistent with hyperrealism, I’m intend-
ing to suggest (additionally) that no proposition or feature of those stories has
been shown to be inconsistent with hyperrealism as defined previously. This
philosophical maneuver is respectable. It is on display in philosophy of religion
debates over the soundness of the logical problem of evil. Virtually all philoso-
phers of religion (including atheists and agnostics) reject the logical problem of
evil precisely because no one has been able to provide a proposition or set of
propositions that show that the conjunction <Theism is true and evil exists.> is
impossible (see Howard-Snyder 1996, xiii, on this point).
41. With this, Price (2007, 286) agrees.
42. Most scholars affirm that causation is formally asymmetric (q.v., chapter 3: sect.
1.1). Most also agree that causation is ordinarily temporally asymmetric (see
Paul 2014, 194; citations can be happily multiplied).
43. It is common to just define a space-time singularity in terms of geodesic incom-
pleteness (see the discussion of this issue in Earman 1995, 28–31). The Banks–
Fischler quantum cosmological theory (see Banks 2015; cf. 2017) includes a
model of our cosmos that is geodesically incomplete but that does not begin with
a singularity. Instead, there exists a trapped surface “where” you would expect
the initial singularity.
44. S.M. Carroll (2004, 340).
45. The fact that the universe was in thermal equilibrium this early on does not con-
tradict the past hypothesis (the thesis that the universe started in a low-entropy
state). See Wald (2006).
46. See Peebles (1993, 75–76); Sciences (2011, 5-6).
47. My appropriation of SC refutes Price’s (2007, 272–274, 279) appeal to the Gold
model of the universe as “a live possibility.” That model requires a cosmologi-
cal symmetry. The cosmos expands and along the way increases entropy until
a state of low-density. It then evolves and collapses into a big crunch, decreas-
ing in entropy toward the crunch. Price wrote, “[a]t present . . . it remains a live
empirical possibility that the universe contains regions in which the thermody-
namic gradient is reversed” (ibid., 273). He goes on to argue that for observers
situated in such regions with a reversed gradient, “the causal arrow runs directly
counter to the way it runs for us” (ibid.). This response fails because the Gold
universe is not a live empirical possibility. The hyperrealist can ground the arrow
of time in the arrow of entropic increase (hyperrealism does not entail anti-
reductionism about time) and not have to worry about the Gold universe.
48. There are many ways to forge the connection (see Tooley 1987, 37–105; or J.
Carroll 1994, inter alios). My next move toward Hall’s intrinsicality thesis is
non-essential.
49. I am not suggesting that P&W are causal skeptics. Nor am I claiming that they
would endorse the hypothesis that hyperrealism implies causal skepticism.
Rather, sect. 3 aims to engage a reductionist who would argue (in the spirit
of something like a P&W-O) that hyperrealism is epistemologically deficient
because it implies causal skepticism.
50. Strevens (2013) argues that from a very young age we are able to read off of the
causal structure of the world reliable probability judgments through equidyn-
amic thinking.
51. Danks (2009, 447).
52. Woodward (2003, 137).
53. And notice that I am now going beyond the Moorean response by underwriting
it with an explanation of how causal knowledge is gleaned from experience.
Those who dismiss Moorean replies to skepticism in epistemology often require
just such underwriting (see the discussion in Steup 2016, sect. 5.4).
54. Hume’s objections to experiencing obtaining causal relations have been refuted
by Fales (1990, 15–46). There is some question as to whether one can under-
stand causation as a theoretical relation if one can perceive the relation (see
Tooley 1987, 297–298). Even if causation is a theoretical relation that is
therefore abstract or removed from the sensible in some way, we may nonethe-
less be able to perceive it (as Prinz (2006) has argued more generally with respect
to the abstract that we can reference).
55. As in Plantinga (Proper Function 1993, 92; 1988, 38).
56. For discussions of fit in the internalist tradition, see BonJour (2003). The hyper-
realist can go internalist or externalist on these issues. I’m not relying upon any
one controversial epistemology.
57. And here I follow Plantinga (Proper Function 1993, 94–95).
58. See M. Bergmann (2006, 109–177) and Plantinga (Proper Function 1993, 93–98).
The argument here may also rely on the supposition that perceptual experiences
of the external world do not involve propositional content. But this supposi-
tion has been defended at length with great sophistication in Burge (2010) (see
specifically the comments at 537–547). I will proffer a second option that leans
on propositional content in the external world. Again, the options for the hyper-
realist are embarrassingly many.
59. Sosa’s theory of the manifestation of perceptual epistemic competence assumes
that perceptual experiences of the external world essentially include or involve
propositional contents (Sosa 2011, 75).
60. Sosa (2007, 51).
61. Sosa (2011, 80).
62. Ibid., 82.
63. I take the previous considerations to be such that they rule out the type of error-
theoretic responses to perceiving causation discussed in Beebee (2003).
64. Schaffer’s (Causation and Laws 2008, 90) objection to the use of causal skepti-
cism for the purposes of arguing for reductionism is both incomplete, and in
two ways, fundamentally different. It is different because he argues that (i) there
can be abductive reasons for causal beliefs, that (ii) one could acquire evidence
that renders one’s causal beliefs likely, and that (iii) we might be able to directly
perceive causation. I don’t discuss (i) or (ii), and although Schaffer and I both
discuss (iii), Schaffer fails to answer the question: How is (iii) related to knowl-
edge? Indeed, he fails to answer this question for (i) and (ii) as well. My discus-
sion completes the rebuttal by escorting at least (iii) to knowledge.
65. I’m adopting (in transmuted form) the notion of alethic explanation in Korman
(2014, 3).
66. This departs somewhat from Korman (2014, 3–5). For a principle more in line
with his view, see line (7). The type of debunking argument I am now construct-
ing is close to what Roger White has called a blocking debunking argument
(White 2010, 575).
67. I have in mind the brand of physicalism that says that all factsP are microphysi-
cal factsP, or else they are made true by microphysical factsΣ, and all entities are
microphysical entities, or else they are grounded in/determined by microphysical
entities.
5 Universal Causal Determination
Present events are connected with preceding ones by a tie based upon the
evident principle that a thing cannot occur without a cause which produces it.
(Laplace 1917, 3)
Not a few philosophers and scientists in the history of thought have affirmed
that singular causation is universal, that every event (or contingent event)
has a cause (e.g., Aristotle (Complete Works vol. 1 1984, 432; Complete
Works vol. 2 1984, 1555) save the first cause(s); Donald Davidson (accord-
ing to correspondence cited by Brand (1977, 332 and n. 8)); Kant (1998,
303–316, at least for objects we can experience); Koons (2000, 107–119);
Laplace (1917, 3–4); Lewis (Events 1986, 242)1; Pruss (2006)). I believe that
when causation relates, what I call, purely contingent events, that relation
is universal (i.e., every purely contingent event has a cause). I will try to
provide two derivations of the universality of causation, or something near
enough, that proceed from very weak principles about explanation and/or
what can (metaphysical modality) be caused.
But what exactly is a purely contingent event?
(1) Necessarily, for any x, x is a purely contingent event, just in case, x is an

event that occurs,2 and every substance and/or arrangement of two or
more substances that is a constituent of x is contingent.3
There are also merely contingent events. Purely contingent events only fea-
ture contingent substances as constituents. Merely contingent events can have
constituents that are necessary substances, and yet be contingent in that they
could have failed to occur. What is a substance again? I continue to assume an
Aristotelian theory of concrete particulars, according to which fundamental
bearers of properties, or that in which some (not all) universals inhere just are
Aristotelian substances belonging to various other universals called kinds.4 If
this bothers the reader, then exchange all of my talk of substances with talk of
individuals, entities, or collections or aggregates of tropes, etc. Nothing I argue
for here depends too much upon my assumed theory of concrete particulars.
Universal Causal Determination 183
How does one show that all purely contingent events have causes? Both
Pruss (2006) and I (Weaver 2012) tried to demonstrate that all contingent
events (Pruss) or wholly contingent events5 are caused on the basis of coun-
terfactual reasoning. Unfortunately, Pruss’s argument is demonstrably invalid
(as Weaver 2012 shows), and my earlier argument falters when it attempts to
move from (where the domain is restricted to events, where Wx means that x is
a wholly contingent event, and where Cx means that x has a cause):
(# 1) : ∼ ∃x(Wx & ∼ Cx) or ∀x(Wx → Cx)

to
( # 2) : ∼ ∃x(Wx & ∼ Cx) or ∀x(Wx → Cx) (see ibid., 312) .
Perhaps some wholly contingent events are not necessarily wholly contin-
gent. The move is obviously non-sequitur. Let us assume, as I did in (2012)
and chapter 1, that the best modal logic is an S5 quantified modal logic (as
in Plantinga 1974 and Konyndyk 1986, 103–117). Here is a tableaux proof
in Priest’s (2008, 45–46) S5 tableaux system for my claim,
(∎Wx → Cx), 0
~ (Wx → Cx), 0
∃x~(Wx → Cx), 0
~(Wa → Ca), 0
(∎Wa → Ca), 0
Wa, 0
~Ca, 0
∎Wa, 0 Ca, 0
♦~Wa, 0 X
~Wa, 1
Because the left branch is open, the lead node is not incompatible with the
falsity of the second node, i.e., ∀x(■ Wx → Cx) can be true and yet ∀x(Wx
→ Cx), false. The latter does not follow from the former.
Section 2: A Church–Fitch Argument

From Causal Explanation
We need a better route to universal causal determination theses. Here is one
such route:
(2) All successful causal explanations are backed by obtaining causal rela-
tions such that for any contingently true proposition that merely reports
on the occurrence of a purely contingent event x, there is a true proposi-
tion that reports on the occurrence of at least one distinct event y and
the fact that y is a or the cause of x. [Premise]
(3) All contingent truths that merely report on the occurrence of purely
contingent events could be causally explained. [Premise]
(4) If (3), then all contingent truths that merely report on the occurrence of
purely contingent events are causally explained. [Premise]
(5) If (2) and all contingent truths that merely report on the occurrence of
purely contingent events are causally explained, then all purely contin-
gent events have causes. [Premise]
(6) Therefore, all purely contingent events have causes. [Conclusion]
Premise (2) suggests that obtaining causal relations back causal expla-
nations and that supposition follows on not a few theories of causal
and scientific explanation,6 although the work of David Lewis (Expla-
nation 1986) and Bradford Skow (2014, who follows Lewis) suggests
otherwise.
For Lewis (Explanation 1986), causally explaining that some event e
occurred amounts to articulating causal information about e or e’s etiol-
ogy.7 Call the collection of those details and information e’s causal story.
Although that story need not include disclosure of laws of nature that may
back relations between constituents of e’s causal history (Lewis, Explana-
tion 1986, 238–240), it should amount to at least one of the following:
(1*) A causal story that reports on causes of e (ibid., 219)

(2*) A causal story built from mere existential claims about event-types suit-
ably related to event-tokens in e’s actual causal history (ibid., 219–220)
(3*) A causal story outfitted with information about events that are not in
e’s causal history (ibid., 220)
Lewis added that one can causally explain e by providing information about
e lacking a cause because “[n]egative information is still information,” and
causally explaining e amounts to the provision of information about e’s
etiology (ibid., 222). That e lacks a cause is therefore the right kind of infor-
mation per causally explaining e.
Suppose Lewis is right about causal explanation. It follows immediately
that every event that lacks a cause (call these acausal events) has one and
the same causal explanation, viz., the report that it is acausal. But now
give attention to an event en that is grounded by some other entity or enti-
ties g1, . . . , gn. Take up once again the theory of grounding assumed in
chapter 1: sect. 4.5.3. On that view, an event can be grounded. However,
grounding is not causation. Causation is strictly a relation between events,
whereas grounding is not.8 There is therefore nothing illicit about a scenario
in which an acausal event en is grounded by g1, . . . , gn. Notice, however,
that it will now follow that en is explanatorily overdetermined. Because en
has both metaphysical and causal explanations. Is such overdetermination
malignant? It seems rather intuitive that in the case before us, the presence
of explanatory overdetermination at least reveals that the Lewis-style causal
explanation is superfluous. It also suggests that in a great many contexts of
communicative information exchange, providing that type of causal expla-
nation will not be salient. If I asked someone why en occurred, the report
that g1, . . . , gn grounded it, seems sufficient by way of an answer, although
that may only be part of the story if grounding is transitive. If that report
added in information about en’s acausality, I believe it would then clearly
violate two Gricean maxims of the pragmatics of communicative activity,
viz. “be brief,” and be relevant.9
We could tell a similar story about non-causal nomological determination
(I am skeptical about whether there is such a thing (q.v., chapter 8: sect.
4.4)). An event e may occur, and it may come to be as a result of antecedent
conditions and evolutions governed by laws of physics that specify non-
causal functional dependency relations. If there are such non-causal laws
of evolution, then once the right antecedent conditions obtain, events will
occur and be non-causally nomologically determined.10 If someone were to
ask why e occurred, reporting on the aforementioned conditions together
with the laws would be all one needed by way of an explanation. Adding
that e lacks a cause would seem both superfluous and irrelevant.
I only highlight these points about the pragmatics of explanation because
Lewis clearly believed that reports on how acausal events lack causes stand
the test of abiding by appropriate pragmatic principles of explanation. But
as I have argued, that does not appear to be the case when an event is
grounded, or nomologically determined in some non-causal fashion.
Perhaps the cases I have used are cases in which negative “explanatory
information is” unworthy of the “honorific name ‘explanation’” (Lewis,
Explanation 1986, 221), but then the Lewisian would need to bear the burden
of proof regarding the existence of events that are not otherwise explained,
and yet that are acausal. I believe that is a tremendous burden to bear.
There is another problem. I see no reason why causal explanation should
be considered unique among the various types of explanation. If Lewis was
right, that reporting that some event fails to be caused is a sufficient causal
explanation of that event (given that it really lacks a cause), then why isn’t
reporting that some event lacks a mathematical derivation from mathemati-
cal axioms (or theorems, or from the empty set of propositional param-
eters) a sufficient mathematical explanation of that event? Notice that if
you thought it permissible to judge that a negative causal explanation of an
event is salient, then one would be hard pressed to preclude the saliency or
relevance of negative mathematical explanatory judgments as well.
Proposition (3) receives inductive support from the empirical sciences. It is
the business of science to explain (Strevens 2008, 3), and a great many suc-
cessful explanations in science are causal (Koons 2000, 109).11 In addition,
some characterizations of both scientific realism and scientific essentialism
plainly assert that the sciences are in the business of discovering the world’s
causal structure (Ellis 2002, 23–24, 159). Moreover, as Rasmussen and I
(Rasmussen and Weaver forthcoming) argue, various modal epistemologies
can be utilized for the purposes of justifying claims like (3), as can a plau-
sible view about modal continuity (see Rasmussen, Continuity 2014). Premise
(3) also has a certain intuitive pull (in the sense of chapter 1: sect. 4.5.2).
If one were a physicalist (everything is physical or else metaphysically
necessitated by the physical) and a mereological nihilist12 (there exist but
mereologically simple objects), then there should be available very strong
justification for the following version of proposition (3),
(3**): For any contingent truth p that merely reports on the occurrence
of purely contingent events, it is physically possible that there is at least
one proposition q, such that q causally explains p.
Justification for (3**) falls out of considerations having to do with quan-

tum field theories (QFTs). To appropriate that justification, however, we
need several metaphysical assumptions. First, we will need to assume that
origin(s) essentialism13 about events is false; that the way an event comes
into existence is not essential to it. Second, we will need to suppose that
all contingent physical substances or individuals or objects that exist are
simples that serve as the object of study in fundamental physics. Arguably,
QFTs deal in the fundamental. Arguably, theories and/or interpretations
formulated under the umbrella term ‘QFT’ are attempts at proffering fun-
damental physical theories.14
Now let there be a set of all physical events involving the aforementioned
simple physical substances that are the object of study in our best physi-
cal theories (i.e., QFTs), and let’s agree with Richard Swinburne’s way of
phrasing matters (for the purposes of deliberation), that all there is to the
world’s history is “the occurrence of all the events of that set.”15 Call this
history H.
A quantum system in a state of lowest energy, a state described by anti-
particle and particle number operators with eigenvalues that are zero, is a
quantum system in a vacuum state (Redhead 1995, 123; 1994, 79–80).16
The Reeh–Schlieder theorem says that “for a scalar Wightman field in
Minkowski space-time any state in the Hilbert space can be approximated
arbitrarily well by acting on the vacuum with operations performed in any
prescribed open region.”17 The axiomatic QFT literature agrees that one
physical implication of the quantum field theoretic treatment of the vacuum
state and the Reeh–Schlieder theorem (Reeh and Schlieder 1961) is that any
and every local quantum event is a physically possible effect of the quantum
vacuum state of a quantum system (Redhead 1995, 128; Summers 2011,
322) “[i]n the vacuum, any local event can occur!” (ibid.).
Now, let our metaphysical assumptions meet up with the deliverances of
QFT. That meeting will imply that any and every local quantum event is
causable by the quantum vacuum state. It now seems to follow that all of
the members of H could have been caused by some other respective event,
whether those members could be caused because they could have been
products of the vacuum itself, or whether those members could be caused
because there could be some suitable causal dependency base that was itself
the result of the physical quantum vacuum state. To deny this admission is
to deny the preceding result, that according to QFT any and every quantum
event can occur. Again, all we need do is coherently posit that the vacuum
brought forth a quasi-plausible candidate for the causal dependency base of
any would be obtaining event that is a member of H, or else it brought about
that event itself (q.v., my comments on determinism and quantum theory in
chapter 2: note 27). What the results I’m pointing to suggest is that, given
the relevant metaphysical assumptions, no (quantum field theoretic) event
lacks a cause at every physically possible world, and so, given our assump-
tions, no purely contingent event that is part of the appropriate domain of
physical goings-on will lack a cause at every metaphysically possible world,
given that all physically possible worlds are metaphysically possible.
But what about gravitational effects and simples? I argue that gravita-
tional phenomena are causal phenomena in chapter 8: sect. 4. So, given
our assumptions, I believe we’ve left out nothing substantial. (3**) seems
well-supported on the appropriate set of assumptions.
In point of fact, I do not actually embrace the assumptions needed to
generate the justification for claim (3**) articulated above. The point of
the discussion was to help convince those scientifically minded philosophers
who doubt whether there is even such a thing as metaphysical possibility
(e.g., Maudlin 2007, 184–191, seems to come close to this viewpoint).
Why think that (3) materially implies the consequent of (4)? There is
a proof that gets us close enough to that claim (and here I’m indebted
to Church 2009; F. Fitch 1963; Gale and Pruss 2002, 90; Kvanvig 2006,
12–14; Oppy 2000, 347–348; Salerno, Introduction 2009; Knowabil-
ity 2009; Weaver 2013; 2016).18 Let bracketed constants, variables, and
pseudo-names represent what I call reporters. Reporters are proposi-
tions19 that report (truthfully or falsely) that some purely contingent
event (without logical redundancy) occurs, although we will let the
expression-type ‘[x]’ mean that x is a reporter that is true. Let ‘Exy’
mean that x causally explains y. Let ‘v’ and ‘w’ be pseudo-names,20 and
let us restrict our domain to propositions. Let’s also require the follow-
ing three restrictions on the nature of causal explanation, or the rela-
tional predicate E,
(Principle #1): The relation E is distributive, such that necessarily, if

some proposition causally explains ([p] & q), then that proposition
causally explains [p], and that proposition causally explains q.21
(Principle #2): The relation E is factive, such that necessarily, if some
proposition causally explains p, then p is true.
(Principle #3): Necessarily, all complex expressions of the form ~[[x] &
y] hold, if and only if, ~[x] v ~y holds (that is to say, it is not the case
that x is a true reporter, or it is not the case that y).
With these three principles in hand we can show that from line (8), it fol-
lows that every purely contingent event has a causal explanation. Consider,
(8)∀x([ x ] (→♦∃ 8) ∀xy(([Eyx x ] →♦∃ ) ) y (Eyx) ) [Premise]

8)∀[xw ([]x ]&(→♦∃ ([Ez ][& zy((Ez ()Ey [Premise UI (8)  ] [ Premise ]
 [ ]
(9 (98~ ))∀∃x[zyw (x]Eyx →♦∃
w]) )~)∃→♦∃ Eyx [yw]) )  (→ w & ~ ∃z (Ez [ w ])
( Eyx (10 )
(9)) ∃[yw((8]Ey
) )∀ &(x[~ ( 9 () 8
(w[ x ∃][z]&
)
 ∀ wx(→♦∃ (
Ez]♦∃[ x&
~[w]) ] →♦∃
~
∃yyz(Ez ((Ey ∃
Eyx z
→♦∃([[w
( y
Ez
w ( Eyx
) )]])&
[ w])
y(Ey )
 )
~ (∃[zw(E
→♦∃ [ ]&
Premise
z [w
y (~ ]∃)))
Ey ( ]
z[ (Ez
w ] [ w ]) [Assumpt
& ~ ∃ z (Ez [ w ]()8[)Premise
UI  ion [Premise
UI ] (]]8) ]
[ ] ( 8 ) ∀ x ( [
( x
([y& ] →♦∃
[w∃(Ey )
]~]&∃(∃z[[~z(Ez y ( Eyx ) ) [
 [ ~ ]∃z (Ez [ w ]Assumpt )EI[ (10
]] UI[ Premise
8 )] (8) ion]
w~][)w ])  [[yw(Ey (8])&
( 8 )∀ x ( x →♦∃ y( 9 ) Eyx w ) & ~ ∃ z ( Ez [ w])  →♦∃ y (Ey ( w & ~ ∃Premise
z (Ez w ) UI
Ez[w]) (11 10) Ev ∃→♦∃
y((9[Ey ) (y(10
w ]([([w
10 Ey
&))]]∃
w ∃~& ~
z(Ez ([Ez
w w ∃]z[))
] ] &
(Ez
[
w]) ∃→♦∃
∃ z(Ez ]]UI) ([w
))) [   )  i[on(Assumpt
(9~) ∃([z10 ]Ev
&(9~) ((∃ [z(Ez
w] &
11)) Ev
12 Ev[(10([[w
w ]))(](11
w ∃]11 y&
w
&
(Ez
))(Ey
)~∃[&
Ev
w])
Ev((∃([[~
(~Ey
w∃]]z(&
y z(Ez
w
∃→♦∃ z( (w
&
Ez[yw])
[[Ewz~~]][))w ∃∃&
(Ey
]
z(Ez
z(Ez)) [[w
∃[ w
~ (→♦∃ z(Ez ] &y[(w
w]]))))) [
~Ey ]∃)(z[(Ez
Assumpt
w ]& [ wion
~])∃z (Ez
]  UI [ w ]()8)  UI (EI
EI (10
Principle [Assumpt
) #1, [8Assumpt)
((10 ion
11 ))] ion]
[ ] ∃ [ ]
[ ∃] ~z(]& [ ] ∃ [
]))E[z [[w ] ] ]))EI [ [ ]  ) ((12
]
11))#1, (11) 
( 10 ) ( Ey ( w & ~ z(Ez w ) Assumpt i on
12[ w ]))(11 y
) (10 ) 
(∃10 ) ∃ y (Ey ( w & ()[~ ~ ∃ z(Ez w ) Assumpt i on
Ev∃][[(z(Ez ~][)w
z(Ez (11 ) Ev ( w & ~ ∃ z(Ez w )) EIle (#2,  EI 10
((13 )) Ev
Ev [(11 w)(](12
w Ev12 & & ) w
Ev
Ev &
w
~
w ] [
& ∃Ez(Ez
w ]
Ev
zEv [
) w((~ ∃z(Ez
)∃ z( w ] ) Princip
 Principle 10
Principle
#1,
z(E([zw ] ]& [ [Ev ]& [wwEv ]))(∃&~z(Ez [~wz∃[]z( EI (10 ) #1, (11
 
v((11~ )∃(Ev [(w12
Ev([12
) Ev
))) Ev
w)(](13
~(12 ∃(w )w
z(Ez
~ ]∃[[z(Ez
Ev [& ] ]&
~
[ ~])∃ Ev
∃z( (E
[w
))Ez w ))
w]]])))) [ Principle ] #1, (11 EI ) (Simp. 10 )  Principle
Principle (12#1, l)e #2, #1,) 12
( (11))
(13 14)) Ev
[ ]
w 13 & )) Ev
[ ]
w
w ] &
[
w~
]
∃
∃
z(Ez
z(Ez [[ w
] 

Princip  (l13

e #2,)
Princip
(  ) ( )
( 12
∃z(Ez ) Ev w
[ w ]Ev &
) (13w) (Ev
Ev(13 )[ w
( ~
)Evw
] [&
Ev ∃ z( & Ev
[ w]~] ∃&z(Ez
E z w ( ~ ) ) z(
~ ∃[z(Ez E z w ) )
w ]) [ w ]) Principle #2, (12  Principle
 )  #1,Principle
Princip 11
Princip 
le #2, le(12 11
#2, 
 )  (12 ) 
((14 15)) ~ [z(Ez
∃(13 (]14 14 [ )w Ev] ) [ ww ] Simp. Simp. (13 13Simp.)  ( 13 )
(13) Ev [ w(]14 &))Ev ~Ev (14 ∃[[z(Ez
w )wEv ]] z(Ez [[ww~]]) ∃z(Ez [ w ])
&
Simp. (13)  Princip  le#2,
Simp.
Princip (12  (13
Simp. )le )#2,13 (13)(12) )
((1516)) ~ ∃z(Ez
∀z(~Ez ((115 5[ w)w )~ ~]))∃ ∃z(Ez [[w w]])) Simp.
QN ( 15 ( 13 ) )
Simp.  ( 
(14) Ev [ w(]15 (14))~Ev ∃z(Ez [)w~] ∃[z(Ez w ]) [ w ] ) Simp.  (13   Simp.
)Simp.
Simp. ((13
13)) 13 
16)) ∀
((17 ~ z(~Ez [ ( (1616 (15
[] w ) )∀ ∀] z(~Ez[[w
)z(~Ez w]])) Simp. (13)  QN ( ( 15) ) QN  (15() )
(15) ~ ∃z(Ez [ w ]()16) ∀z(~Ez
( Ev
16
( 15 )) w∀ ~ z(~Ez
∃ z(Ez [ w w ]
[ ] [ w ]) ) )
QN (15)  Simp. (13)  QN  UI 16  QNSimp. ( 15 )
(  ()15
13  
)
[ ] [ [ ] ] ( )   (
()))  )
([16 [ w))]~]∀()17 (  ()16
17
((18 8)) Ev~ (Ev (
w 17 ) ~ Ev w  UI 16  UI 16
(16) ∀z(~Ez 17 w ( 17 & Ev)
z(~Ez
~ [) ~wEv
~Ev Ev ][ w ]w)
ww

 Conj.
 (15)UI QN14  ,
( 15 17 
[ ]∃&z(] ~Ev  UI (16 )   QN  Conj.  UI (1614
17))(,)(17
  , 17 
(v17 )
((19 18
~
8)) Ev
Ev
~ (∃18
[ w
[y8w()Ey
(17 ]
(](18
)Ev
18 &
~(18()[Ev
8 )w Ev~Ev ][ w[& & w [ w]~~Ev
] [Ewz []w[ w]))] CP (10 ) - (18

Conj.
( ) )and
(UI
14
Conj. Re )16
( ((14
,ductio
)  
) )( )
[ w((19 ]20)) ~(∃~19 8
(∃)[)yw
) Ev
]w∃&
([Ey [
y]((& w
[Ey ] ](z( &
[&∃wEz(]w
~Ev
z~E[& []w w ]
~E))∃z]z([wEz]))[ w]Rule Conj. ( 14 ) ,
CP (10 ( 17 )
 UI

)-)((CP
16
10))(-and

(18

) -)(Re
(19  Conj.
) follows
ductio  ( 14 ) , (17 )
y∃()Eyy((1~19(Ev9
Ey
[ (~w
]
w~∃~
[
∃zwz(
] [ z(Ez
)) )) of Nec.
))) (19
CP 18
because 10 18
and ) (
and
Reductio
)
Re from
(
ductio
)  no  premise
∃z()[Ez(]~[20 [ ~y(∃]Eyy((Eyw([& ( ) ( )
(&188)~Ev w ( 18& 8 ) ~Ev w & ~Ev ww]~CP   Conj. 14 , 17
w∃)]y()) (19 )∃ [ ] [ ] & ~~(E ∃10 z( [)
Ew-z(z][18 [ ww)]))and
] Re ductio Conj.
of Nec. (CP )
14
(( 10 ,
)
17
 ) - (18) and Re ductio
( ) ( )   from
((20
21) ~♦∃ y()20 
Ey 20
( Ey ~ ([y
) ( 
w [(Ey
w ~] &]∃( y~w(Ey
∃ & y ( Ey
~ ∃z(Ez ∃ ( z(
[ E [
]z w&[ ]E))~]z∃w
∃ w z( ))∃zz( E ))

[
w
 Rule]
Duality)) ofRule
 of  the(Modal

Nec. of
Rule Nec.
19 ) 10Operators
because 19 because
19 ( and
because
19 ) and 19
follows DNfollows
19 (
from follows
20 )   pre
from
no no pr
mise n
( 19 ) ~ ∃ y (Ey ( 19 ( [ w ~ ] )
∃ & ~~ ∃ z([ E ]
z [ &( w
] ~ ∃ & z( [ z(Ez ] )) ))) ( ) ( CP ) ( ) - ( 18 )  Re ductio 
[ wof~[] &
w ))
[~∃w]z(Ez [])) ] Duality
w ] &(21 zy[)(w ]& CP 10 - 18 and Re ductio
( noOperators)preOperators (and all)and
~]))∃(z(Ez w ]))) (19 )the
~ ∃z((E21 ( 20 ) ~ ∃[~Ey
w∃](z(Ez
y ( Ey ( w & ~ [19
∃ z( E z w 
))  Rule  of Nec. 19 because  19 follows (20)(20 )
from
(22)) ~~♦∃ 20[ w)(]Ey ~21 ♦∃ ))() ~
[yw ~ ♦∃
(Ey ]∃(& y(((Rule
zEy Ez [w
&
w])
Nec. [because
w Duality follows
Duality
of of
Modal the offrom Modal
the MModal
Operators
Tbecause ( 9follows
) , (and
misesDN
2119 )from
at
( 20  ) DN
DN
(] &
20)~∃~z(Ez ∃y((22
([Ey ( [ w ~ ]
))[ w
∃
& y
~]♦∃ ~ ∃ ( [
z( E ]
([z(w
z &[w]∃&
([Ez w ]
]of )) ∃ z( E z [ w
 [ w
 Rule ] ))
[ w]))) ]))
of Nec. Rule ( 19of ) Nec.
because ( 19 )
( 19 ) (( 9T) ,((921) follows
) no pre from
mises no atp
( 20
w)(]]~22 and DNof(the )

(22) ~ [w
(2))(121
& ~~♦∃
)~
∃ [ &
z
y(yDuality
w ( Ez
Ey
](Ey
~ &
[ ∃w]) ~  z[&w])
( Ez ~∃
~the [ ∃
z(Ez
w]) z(Ez
Modal Operators Duality
 20
M T  9M,TM
Modal
( ) ( Operators
21 ) ) , 21
( and
)  DN
(21
∃ z(Ez) ~♦∃ [w])y(Ey([ w
(21) ~(♦∃ 22 ] &)y~(~Ey∃[(z(Ez w [ w] ]&& [ w~~]))∃∃zz(Ez (Ez[w]) [ wDuality
])) MTofDuality
  (the ( 21of
9 ) , Modal ) the Operators 
Modal Operators
andDN  20
( and) DN ( 20 )
MT ( 9 ) , ( 21) 
(22) ~ [ w ](22 &()22 ~ )∃[z~w (Ez ][[w & w]) ] ~& ∃z~(Ez ∃z[(w]) Ez[w]) MT ( 9 ) , (M 21T)  ( 9 ) , ( 21) 
But (22) is just the claim that it is not the case that the conjunction [w] &
~Ǝz(Ez[w]) is a true reporter, and so by Principle #3 we can infer,
(23) ~ [ w ] v ~~∃z ( Ez [ w ]) ~Principle

(23) ∃z ((Ez
[w ] v ~~#3 22[)w ]) Principle #3 ( 22) 
(24) [ w ] → ∃z ( Ez [ w ]) (24) [ w ∃z ( Ez
] → and
Impl. DN[w ]()23) Impl. and DN ( 23) 
But proposition (24) will allow us to infer that all true reporters have causal
explanations. The derivation therefore clearly backs premise (4). Here’s how
to make the supports relation I’m invoking even more explicit. Recall that
premise (4) states that,
(Premise 4): If all contingent truths that merely report on the occur-
rence of purely contingent events could be causally explained, then
all contingent truths that merely report on the occurrence of purely
contingent events are causally explained.
The derivation that begins with line (8) shows that
(Derivation): If every true reporter could be causally explained, then all

true reporters have causal explanations.
But a reporter (in general), you will remember, is a proposition that reports
truthfully or falsely that some purely contingent event occurs. So (Deriva-
tion) can be restated as,
(Derivation*): If every true proposition that reports on the occurrence

of some purely contingent events could be causally explained, then
all true propositions that report on the occurrence of some purely
contingent events are causally explained.
The explicit argument that moves from (Derivation*) to (Premise 4) is as

follows:
(a) If (Derivation*) is true, then (Premise 4) is true.

(b) (Derivation*) is true.
(c) Therefore, (Premise 4) is true.
Again, justification for (b) comes from the proof that begins with line (8).
What about (a)? Grant that its antecedent holds. Now to see that
(Premise 4) follows by material implication, note that (set i) the set of
all true reporters (where, again, a true reporter is just a true proposition
that reports on the occurrence of some purely contingent event) incor-
porates (set ii) the set of all contingent truths that merely report on the
occurrence of purely contingent events. Here are some examples to help
support the idea.
Example #1: The contingent truth that particle p spins at t is a true

proposition that merely reports on the occurrence of some purely
contingent event. It is therefore a member of (set ii). However, it is
also a member of (set i) because it is a true proposition that reports
on the occurrence of some purely contingent event.
Example #2: The contingent truth that Corey Clement scored a touch-
down at t is a true proposition that merely reports on the occurrence
of some purely contingent event. It is therefore a member of (set ii).
However, it is also a member of (set i) because it is a true proposition
that reports on the occurrence of some purely contingent event.
Example #3: The contingent truth (we can suppose) that a quark emit-
ted a gluon at t is a true proposition that merely reports on the oc-
currence of some purely contingent event. It is therefore a member
of (set ii). However, it is also a member of (set i) because it is a true
proposition that reports on the occurrence of some purely contingent
event.
We could continue on with more examples, but it seems clear that there
is substantial evidence for the thesis that (Premise 4) really is materially
implied by (Derivation*). Examples like the above support this.
Some of my previous attempts at showing related results by way of
Church–Fitch-like reasoning in Weaver (2013 (not the main argument that
I believe survives); 2016) may come up short. The worry is that some of
those past attempts involved a universally quantified sentence of the form
∀[x]([x] → ♦Ǝy(Ey[x])) at the start of the derivations. The problem is that
starting with such a statement restricts the domain of quantification to
purely contingent facts or events (for ibid.). That means that the initial uni-
versal instantiation step in my preceding derivations must involve actually
existing or occurring purely contingent facts or events. However, the deri-
vations go on to show (as you can see in the correlative case in derivation
(8)–(24)) that the instantiated purely contingent fact/event is not a fact or
event at all (see my line (22) here, where the correlative lines in my previous
work are lines (16) in Weaver 2016, and (9) in Weaver 2013, appendix 2;
I do think there is a way out of this worry). The preceding derivation can be
viewed therefore as an improvement, and the argument from causal expla-
nation is an addition that advances the discussion in novel ways.
Premise (5) seems plainly true, and so it follows that all purely contingent
events have causes (i.e., (6) holds).
Section 3: Yet Another Counterfactual Argument

There may be another way to show that causation is universal. This other
way is reminiscent of counterfactual arguments in Pruss (2006) and Weaver
(2012). It is admittedly highly speculative and should not be interpreted
as something that I assert in accord with the knowledge norm of assertion
explicated in the opening section of chapter 1. The following reasoning
is merely exploratory and serves as a launching pad to encourage more
thought.
I begin by asking, what if abstract objects could be caused? More spe-
cifically, what if the abstract essences of contingent things or substances
could be causally produced? Assume that the essence of a contingent con-
crete substance is (i) a haecceitistic property that substance exemplifies at all
worlds at which it exists, (ii) a property that no other entity could exemplify,
and (iii) a necessarily existent abstract entity that can exist uninstantiated.22
Haecceitistic properties are themselves types of entities and their existing at
an ontological index (space-time point, or world) can be regarded as a type
of platonic event. Call such events involving the essences of contingent
substances purely platonic events (although see the analysis in PPEs), and
notice that if purely platonic events could be caused to obtain, it would
not follow that a necessary being exists, for varying eternally persisting
obtaining states or events featuring contingent entities could very well be
that which is causally sufficient for the existence of such events at varying
worlds.23
A number of recent authors have argued that abstract objects can stand in
causal relations.24 Andrew Kania tells us that in the context of debates about
the ontology of works of music such as Beethoven’s 5th Symphony, “the
consensus has favoured realism—the view that musical works are abstract
objects” (Kania, Piece 2008, 77, although Kania criticizes the view in other
work (Review 2008)).25 However, some theorists have argued for the intui-
tive view that such entities are created by efforts of great musical genius,
and so by consequence abstracta can be causally influenced after all.26 If one
can stomach the idea of causally affected abstracta, and if one is willing to
suppose that there are purely platonic events, the proof I articulate provides
some evidence for the view that all purely platonic events have causes (the
principle of platonic causality (PPC)).27
Understand the term v as a pseudo-name, and consider the following claim:
[Interpretation: Px: x is a purely platonic event; Ox: x occurs; Cxy: x

causes y; Domain: unrestricted28]
(Key Claim): (∀x)[(Ǝy)(Cyx) ■→ (∼(Ǝz)(Czx) ■→ ∼Ox)] [Principle #129]
That is to say, for any entity x, if it were the case that there is at least one
entity y such that y caused x, then it would be the case that, (embedded)
were it not the case that there is at least one entity z, such that z caused x,
then it would be the case that x fails to occur.
In Lewis-speak (for the non-trivial truth-conditions for the ‘would’-
counterfactual), Principle #1 says that for any entity x, there is a world at
which both (a) <x is caused by at least one entity y> and (b) <∼Ǝz(Czx ■→
∼Ox)> hold that is closer to the actual world than any world at which (a)
holds, and yet (b) fails to hold.
Consider now the following four premises (one of which is a theorem),30
(1) ∀x[∃y(Cyx)  → (∼ ∃y(Cyx)  → ∼ Px)] [Premise]

(2) ∀x(Px →♦∃y(Cyx)) [Premise]
(3)  q → [~p  → (p♦→ q)] [Theorem]
(4) ∀x(Px →  Px) [Premise]
On the assumption that events are states, I would recommend the following
analysis of purely platonic events:
(PPEs): Necessarily, for any x, x is a purely platonic event, just in case,
x is a necessarily obtaining or occurring state that is the essence of a
contingent substance existing at a world or time.31
Proposition (1) is just (Key Claim) except its consequent says that were there
no cause of x, x would not be a purely platonic event. (Key Claim) entails
(1). Such entailment holds because an event failing to occur entails that it is
not a purely platonic event.
Line (2) follows from the supposition I invited readers to entertain for the
purposes of deliberation (viz., that purely platonic events could be caused to
obtain), although there are good arguments for (2).
Line (3) is a trivial truth in that it is a theorem of our assumed logic,
although I have included a statement of it for emphasis (i.e., theorems are
not ordinarily placed alongside premises). Given that q is a necessary truth,
~p-worlds at which (p ♦→ q) holds will be closer to the actual world than
any ~p-worlds at which (p ♦→ q) is false because given that q is a necessary
truth, there are no worlds at which ~p and ~(p ♦→ q) holds. On S5, neces-
sarily true propositions hold at every possible world. Thus, the following
proof will run at any possible world,
(Informal Proof): Recall that for Lewis (p ♦→ q) is equivalent to ~(p

■→ ~q). Thus, ~(p ♦→ q) is equivalent to ~~(p ■→ ~q), and that double
negated sentence amounts to (p ■→ ~q). But (p ■→ ~q) non-trivially
holds, just in case, there is at least one p-world at which ~q holds that
is closer to the actual world than any world at which (p & q) holds.
However, we were assuming that q is a necessary truth. Thus, there is
no p-world at which ~q holds. The ‘would’-counterfactual ~~(p ■→ ~q)
is therefore false, and its falsity yields (p ♦→ q). Thus, ~p-worlds at
which (p ♦→ q) is true are closer to the actual world than any ~p-world
at which ~(p ♦→ q) holds because there are no relevant ~p-worlds.
Line (4) is the claim that every purely platonic event is necessarily purely
platonic. This is a particularly plausible claim. If events understood as states
are like property exemplifications and purely platonic events are understood
in terms of (PPEs), then the constituents of a purely platonic event belong
to that event necessarily (or, on some views essentially). But if that’s right
then if a purely platonic event has certain constituents at one world, at every
other world at which it occurs it will feature those self-same constituents,
and so that event will remain purely platonic.
Let me now show that given (1)–(4), every purely platonic event has a
cause,
(5) ∃x( Px & ∼ ∃y(Cyx)) [Assumption]

(6)  Pv & ∼ ∃y(Cyv) [EI (5)]
(7) Pv →♦∃y(Cyv)  UI ( 2) 
(8)  Pv [Simp. (6)]
(9) Pv [Nec. Elim (8)]
(10) ♦∃y(Cyv) [MP (7), (9)]
(11)  Pv → [~~ ∃y(Cyv)  → (~ ∃y(Cyv) ♦→ Pv)] [Instance of (3)]
(12) ∼~ ∃y(Cyv)  → (~ ∃y(Cyv) ♦→ Pv) [MP (8), (11)]
(13) ∃y(Cyv)  → (~ ∃y(Cyv) ♦→ Pv) [DN (12)]
(14) ∃y(Cyv)  → (~ ∃y(Cyv)  → ~ Pv) [UI (1)]
However, (~Ǝy(Cyv) ■→ ~Pv) is equivalent to ~(~Ǝy(Cyv) ♦→ Pv), on Lew-

is’s semantics. Thus,
(15) ∃y(Cyv)→~ (∼ ∃y(Cyv)♦→ Pv) Sem. for CFs (14) 
But now our derivation suggests that ∃y(Cyv) counterfactually implies

both (~∃y(Cyv) ♦→ Pv) and ~(∼∃y(Cyv) ♦→ Pv). However, Lewis’s seman-
tics yield the validity of the following sequent: p ■→ q, p ■→ ~q ˫ ~♦p.
Therefore, ˫ [(p ■→ q) & (p ■→ ~q)] → ~♦p. Thus,
((16 16)) [[((∃ ∃16 yy(())Cy [[∃ vy)( Cy →

→vv))(( ~→
~ ∃
∃yy((~ Cy∃
~Cy ∃vvyy))((♦→Cy P♦→
v)P vv)) ] P
P& vv)) ][ ∃
∃& [[ ∃∃∃vvyyy))(((
yy((Cy Cy →
→vv)))~ ((∼
 →
→∃ ∃~ ~yy((∼Cy∃
∼Cy ∃vvyy))((♦→ Cy v)P P♦→ vv))]] P Pvv))]]
[∃ v) ] P&v) ][ ∃&
16 Cy ∃vvyy))(( Cy  → ♦→
Cy
Cyvv))P ♦→ & & Cy Cy
 ~ ∼ ♦→
Cy
Cyvv))P ♦→
(16) [(∃16 y()Cy Cy →v)( ~→ ∃y((~ Cy∃vy)(♦→ ♦→ y(Cy v) Cy→v~ (∼→∃~y((∼ Cy∃vy)(♦→ ♦→ v)] Pv)]
]♦→ [ v∃)y][(Cy Conj. Conj. (13 P)v)v,,))((]P♦→
, 13 (15 15)) ),,, ((15 15) 
15
Conj. v)]) Pv)]) 
[∃y) ([Cy Conj. 13 13
(16)(16 (∃16 y(v)Cy )[ ∃vy→ )( Cy (→ ~v∃)( y~(→ Cy
∃y(v~Cy )♦→
∃vy)(♦→ CyPvv))P v&) ] P& ∃& y(vCy[ ∃vy→)(Cy
) ~→ (v∼)~∃(y∼ →(Cy
∃~Conj. y(v∼Cy)♦→ ∃Conj. vy)((♦→ 13
Cy 13
(15
Conj. (13 )(13
, (15
)for , (13 )(15 ,  15 
) ) ( 16 ))
((17
17)) ~ ~(((♦∃ 17
♦∃ ))yy~ ((♦∃Cy
♦∃ vvyy))((Cy vv)) Sem. Conj. Sem. Conj.
for CFs (CFs16) ((16
16
Sem. for CFs (16)  ) 
17
17 ) ~
~ Cy
♦∃ y ( Cy
Cy v )  Sem. 
 Sem.
for
Sem. CFs for
for CFs
CFs 16 
(17) ~♦∃y(Cyv) 
((1 8)
) ♦∃ ♦∃ ( 1 8 y )
))(♦∃ ( ♦∃
Cy vy ) ( Cy & v
& vv~))♦∃ ~ ) ♦∃ & y ~ ( ♦∃
Cy
♦∃ y
v ) ( Cy v )  
Conj. Conj. 
Conj. Conj. ( 10 ) , ( 10( 17 ) ,) (17
,,) (17
) 
17 ) 
, (10 )16
1~8♦∃ ((y1
1(8 8yyCy Cy
♦∃ )vvyy))((vyCy &
&yy~ ~((Cy
Cyvyy)((Cy
♦∃ v) 10 (,16 10 17 )(CFs
Cy v) Cyv) Sem. Conj. Conj. (-10 )for -((18
(17)((1 (v♦∃ for CFs
((19178))♦∃
)) ∃ ∃~(((19 ♦∃
17
x (( )))y∃
(Cy
∃
~
Px Cy
x (( &
&)(Cy
Px ∼∃
∼
~♦∃
∃& vy)(∼C
∼ ∃
yx
∃ y(())C →
→ yx ))♦∃
[[))
♦∃ →yy(([[♦∃
→ ♦∃
Cy vvyy))((Cy
Cy
& vv~)♦∃&
♦∃ yy~ ♦∃
((♦∃
Cy vvyy)](Cy
Cyvv)] )]CP Sem. ( 5 Sem.
for
CP ) ( (
18 CFs
5 ) 17
18 ) ) (16) 
(18)((19 19
♦∃8))y♦∃∃((19
19
x
xCy
 )v∃Px Px)x&x (  &
)((~
Px
Px ∼v~∃&
♦∃ &
)y♦∃
y
y(&
(
(Cy
C
∼Cyx yx y
∃vy) ))))C
(C→ yx
yx[)) →y([Cy Cy
♦∃vy)(Cy & ~
& v~)♦∃
) &
&y~(Cy ~ Cy
♦∃vy)] )]
(
(Cy Conj. v)]CP ((10 CP 5CP
5CP
-
))(-10 ((517
18
,(18 5
)(19 ),)(-10
-
()(17
18,) 17  
((1 20
20 ) ~
~ ((1 20
20
(8y
∃
∃ x
x
))((♦∃
) (
Cy~

~
 ∃Px
∃ Px
∃Px
vyx&
x (
Cy&

 &
& Px
Px ∼
∼ ∃
&
∃
& yyy((~(∼ ♦∃
Cy
∼Cyx
Cyx ∃yyvy(())
∃ ))
Cy♦∃
)(Cyx
Cyx
)
v))
)) Conj.
 Reductio
Reductio
(5) -((5CP
Conj. Reductio
Reductio
18-) 18 19 ))((19)(  )
19)) 
19
(19)((20 ∃x)(
19 ~∃((19 20
x Px
∃ ( x

)(~
) &

∃Px x ( ∼x(∃
&

Px& y∼Px(C∼∃&yyx ∃& ( yC))
∼ (∼ ∃
yx →
Cyx
y
∃y[))
(
))
C
(♦∃
→
Cyx
yx y♦∃
[
)) ())
Cy
→ y v♦∃
(
[Cy)& y
v )(~
Cy
&♦∃ v ~y♦∃
) (&Cy y ~(v♦∃
Cy )] vy)] ( Cy CP v )] Reductio CP
Reductio
) ( ( 5( 19) -  )
 )  ) 
( 18 
(21 )) ∀ ∀ ((21 x
x~ )) ∀ ∀ (( x Px~ (( & Px ∼
∼(Cyx ∃
∃& yy((∼ Cyx
∼ ∃
∃ yy(())))Cyx )) QN QN (QN 20 )(()20
(20()((21  ))(19) 
21 ∃ ( 21
21 ~) ∀  x
x Px~
~ ∼ ( 
&∃ Px
Px &
& Cyx
∼ ∃ y (Cyx
Cyx ))
))   QN
QN 20 ( 20
20
(DeM
~
20))) ∀ x~((22 (x∃x
20 ~)(~ Px (∃Px & Px & y ∼∼v∃ ∃yy(∼ ))
(Cyx ∃ )) Reductio
 QN  20)  (19
19
((x~ ( vv&PxPx & ((Cyx yy()) Cyx )) Reductio Reductio
( ) ( 21
21)) 
((22 ∀
∀ x (())~ ∀

∀ x Px ∼∼
∼∼ ∃
∃ yy∼∼
∼∼ Cyx ∃
∃ (())Cyx )) DeM 21 
(21)((22 22
∀x) ∀
21
) ~(21 (
( 22
22x
x( (~ )
~
) 
∀
Px

∀ ( 
x
x
x
Px
Px ( &
Px~
~
~ 
 v(∼ &
Px
Px
∼∼ ∃Px y∼∃ v
v
(Cyx ∃
&
∼∼
yy(Cyx
Cyx
( ∼
Cyx
∃ y
y ())Cyx
Cyx
))∃y))())Cyx)) ))
)) QN DeM  DeM
QN (20()QN 
 DeM
DeM 21
( 21) )(20
20
 (21 ) 
and ((DN 22) ( 2222) 
(23 )) ∀∀(((~23 x
x(( )) ∀ ∀ Px x (→ → Px ∃
∃yyy→ → (Cyx ∃
∃ y(())Cyx)) Im Imp pl. l.Im and p l. DN and  DN 22
(22()((23 22) ( 22) 
23 23  Px x
xv(( Px (((Cyx
Cyx ∃yy)) ())Cyx
Cyx)) Im and p
pl. DNand
∀x
22 ))(∀ ∀
23
(P24x
22 x (( )Px
)~(∀
∀Px
 x Px (→

~∼∼→ 
Px∃∃
v)Px∼∼ y→ v∃Cyx y∼∼ (Cyx)) ∃y())Cyx))
)) DeM Im DeM p( 21 l. Im DeM )(21
and l. DN
)  ( 21 ) DN

((24 ) ( ( v
v ))→ )→ P v
vv → P v P v )
ImpUI
 UI (4UI   4 UI )DN
 ( 4)(22)
4   
 p(l.4)DN
24
∀x)) (((∀ ((P24 24 ((P P→  →P∃vy) P
Pvv)))) UI
(23)((24 23  P
(2523vPx
x ()→ )v(∀ Px 
xP( P → vPx )
(Cyx∃Py→ (Cyx ∃y())Cyx)) Im p((l.
l. and 4UIIm)3and  (DN
and  ) ( 22) 
((25 25 ) (
P)v)) (((
(
((→ 
2525
P
P

))v(( 

→
→ P
P
P v
Pvvv→ v →
→
)
) P P
v
vv)))

  
 UIUI
( )  2UI 2 UI
UI 3 (( 2
)  2
2
3
3
3 )) 22
(24()(25 (24 Pv)→ P
→ vvP ) vv)) Pv)   UI (24UI
UI 4  3) ) ( 4)  
(P24
vvv→ )) (P P∃v ((→ PCy ( 24
∃vvvyy)→ Cy∃ HS)),,(( 24 (25 25)) ,,,) (( 25 25) 
(26 )P P(((v26 → → ∃ vy)(Cyvv)) HS HS , 24
(25)(((26  25) 
26))PP 26
26 → )P PP∃ → ∃vvyy))((CyCyv)  HS
(
HS ) 24
HS
24 24
(25 ( ( v
25 P → ) v ( →
 y P ( P vCy v →) P v )  UI  UI2 3
( (  2 UI 3 )  ( 2 ( 3 25 )  )
(27 )∀ ∀ ((27 x
x(()))Px ∀
∀ x→
x→
((Px ∃
∃yy→ → ((Cyx ∃
∃ y())Cyx))
HSU
 U G (U
G  26 G )()26 )
 ), 25 
(26()((27 ),((24
27
Pv)) → (27 Px ∀x (vPx Cyx ∃yy(())Cyx Cyx)) U( 24 U 26 G 26
26 ∀
P(26
27
vx∃(→ )yP(Cy
Px ∃v→ y→
Px
(Cy )∃ y→
∃ v(yCyx
)(Cyv)))
))
 HS G )(U ,26
24
HS ()25
G
(26 25 ) )( )
(27)(27 ∀x) (∀ (Px x(→ ∀∃ y(Px (Cyx ))∃y())Cyx))  UG U(G 26)U  G  26 
27 )Px x→ ∃y→ (Cyx  ( 26 )( )
Line (27) says that all purely platonic events have at least one cause. That
is the PPC.
But how does PPC relate to the universality of causation for purely con-
tingent events? PPC says that every PPE has a cause. Consider an arbitrary
purely contingent event. That event is one that on the assumed ontology,
began to exist just as soon as the constituent contingent substances began
to exemplify the properties or stand in the relations (at the relevant indices)
that build up, and make that event an event. Notice, however, that because
purely contingent events feature only contingent substances as constituents,
and because PPEs involve the essences of contingent substances existing
at some time and/or at some world, every purely contingent event will be
such that it would not exist were the corresponding purely platonic event or
events to fail to occur.
If we let ‘Px’ mean that x is a purely platonic event, let ‘Ly’ mean that x
is a purely contingent event, and let ‘Ryx’ mean that y has the essence of
the contingent concrete substance that is a constituent of x, as a constituent,
then we can affirm the following:
(28) ∀x∀y[(Lx & Py & Ryx) → (~ Oy ■ → ~ Ox)]
That is to say, for any x and for any y, if x is a purely contingent event,
and y is a purely platonic event, and y stands in relation R to x, then x’s
occurrence is counterfactually dependent upon y’s occurrence. Of course,
counterfactual dependence is not itself indicative of causation (as cases of
symmetric overdetermination have taught us). But still, proposition (28) is
awfully close to a universality of causation thesis. It (i.e., (28)) would seem
to be obvious evidence for the universality of causal determination thesis
(with respect to purely contingent events) at the very least. That is to say, it
seems obvious and intuitive that the probability of the universality thesis,
given (i) is greater than the probability of the universality thesis alone.
Let me now attempt to connect theses related to premise (1) to a contras-
tive theory of causation. Consider Cei Maslen’s (2004, 342–343) contras-
tive counterfactual account of causation. That account has need of contrast
situations and contrast events. A contrast situation is “the complex of a
contrast event and the event in which the absence of the cause [or the effect]
consisted.”32 A contrast event is one that is not identical to the effect, or the
cause, or the cause’s omission, although it is compatible with that omission
(ibid., 342).
According to Maslen, relative to a set of contrast situations {α*} an event
α is the cause of an event β relative to a set of contrast situations {β*}, just in
case, if it were the case that some arbitrary member of {α*} occurred, then it
would be the case that some member from the set {β*} occurred.33 Necessar-
ily, if a member of {β*} occurs, then event β will fail to occur. Thus, if Maslen’s
account is true, then, as Steglich-Petersen (2012, 118, n. 4) has pointed out,
the following equivalence holds: α rather than a member from {α*} caused β,
just in case, if it were the case that a member from {α*} occurred (rather than
α), then it would be the case that β failed to occur. Given that α rather than
a member of {α*} caused β, it follows that there’s a cause of β. And given
that there’s a cause of β (the cause just referenced), and Maslen’s contrastive
theory of causation holds, it will follow by entailment, and not just coun-
terfactual implication, that α failing to occur, plus the occurrence of some
member of {α*} counterfactually, implies that β failed to occur.
Let ‘R’ represent the ternary contrastive causal relation, and let α be the
actual cause, β the effect, and {α*} is, with respect to α, a contrast situation.
Read ‘Rα{α*}β’ as follows: ‘α rather than contrast situation {α*} caused β.’
And our interpretation is as before,
(Implication of Maslen’s Contrastive Account): Rα{α*}β ⇒ (O{α*}■→ ~Oβ)
That is to say, α rather than contrast situation {α*} caused β entails that
were (a member of) {α*} to occur, β would have failed to occur. But notice
that this will imply the would-counterfactual statement: (1*) Rα{α*}β ■→
(O{α*}■→ ~Oβ) because strict implication implies variably strict implica-
tion. Claim (1*) looks like the contrastive causal version of premise (1) in
preceding discussion (in sect. 3). With a correlative contrastive version of
premise (2), call it (2*), and the following two additional premises—(3*) ■q
→ [∼p ■→ (r ♦→ q)] and (4) ∀y(Py → ■Py)—we can derive the universality
of causation with respect to purely platonic events just as we did without
the contrastive analysis of causation. Notice, however, that our lead prem-
ise (1*) enjoys (indirectly) all of the justification that Maslen’s contrastive
theory of causation enjoys because it follows from that account. Similar
reasoning could be used to connect other contrastive theories to versions of
premise (1).
If premise (8) in sect. 2 holds, then (a) every true reporter has a causal expla-
nation. I believe I’ve shown how claim (a) can be used to argue that every
purely contingent event has a cause (see argument (2)–(6) in sect. 2). And
although the truth of premises (1), (2), and (4) in sect. 3 yield a conclusion
that helps evidentially support the universality of causation thesis I’m after,
at least premises (1) and (2) (of the same section) are at best speculative.
Thus, my case for causation’s universality ultimately rests upon the argu-
mentation in sect. 2.
Notes
1. Lewis said,
Any event has a causal history: a vast branching structure consisting of that
event and all the events which cause it, together with all the relations of
causal dependence among these events.
(D.K. Lewis, Events 1986, 242; emphasis mine)
Although he seems to deny this in Lewis (Postscripts 1986, 213), musing that
the entire history of the cosmos is a complex event that has no cause.
2. For my account of events, see chapter 7: sect. 4.
3. Cf. Chisholm (1990, 419, see definition D11) and Koons (2000).
4. A kind is that which determines the natures of the particulars that belong to it.
A kind can be said to be exemplified by those particulars that belong to or are
members of that kind. For a complete presentation of the theory of substances
and kinds that I am assuming, see Loux (1978; 2006, 107–117), on which I lean
for my partial explication. See also van Inwagen (1990); Hoffman and Rosen-
krantz (1994).
It is important to highlight the fact that Aristotelian substances include those
entities that are part of the ontologies of our fundamental physical theories.
Loux wrote,
Concrete particulars, then, are substances; or at least some are. . . Aristotle
himself was particularly stingy in his allocation of the term; he restricted the
set of substances to individual living beings—plants, animals, and persons—
and, perhaps, to the elementary items physics tells us enter into the composi-
tion of everything that is material. For Aristotle, the latter include the four
elements, fire, earth, air, and water; for a contemporary Aristotelian, they
would include the basic entities posited by contemporary physical theory.
(Loux 2006, 114; emphasis in the original)
5. These entities are similar to purely contingent events.
6. See W. Salmon (1984); Strevens (2008, 77); van Fraassen (1980, 124–125); and
Woodward (2003, 209–220), although Strevens seems open to non-causal scien-
tific explanations.
7. D. K. Lewis (Explanation 1986, 217).
8. Schaffer himself distinguishes causation from grounding in Schaffer (2010, 345).
However, I do maintain that all instances of what I will call (in chapter 9: sect.
5.1) natural causation are also instances of grounding. Nonetheless, there are
instances of grounding that are not instances of causation.
9. See Grice (1989, 27–28).
10. I happen to believe that many of the laws of physics are causal. I voice the objec-
tion in the main text for those who may disagree with my causal approach to
many of the dynamical laws of physics. I should add that I do believe there are
some laws that specify mere functional dependency relations (q.v., my discussion
of Boyle’s ideal gas law in chapter 1: sect. 4.5.5.4).
11. Indeed, not a few theories of scientific explanation build in causation. See W.
Salmon (1977, 162; 1984, 19); Strevens (2008, 24–27); and Maudlin (2007),
who uses production talk. Q.v., n. 6 above.
12. See Cameron (2010); Dorr (2002); Sider (2013).
13. Kripke (1980, 114, n. 56). For criticism see Robertson (1998).
14. Although see the discussion in Ruetsche (2011, 340–356).
15. Swinburne (2007, 149). I’m using his wording to make a different point.
16. The reader may retort: But vacuum states are idealizations, and the theorem
appealed to in the main text applies to Minkowski space-times. So what. Sanders
(2009) retrieves Reeh-Schlieder states in curved space-times, and my argumenta-
tion only requires a set of physically possible worlds that collectively ensure that
all members of H could be caused by a quantum vacuum state.
17. Sanders (2009, 271).
18. The background logical system is an actualist S5 quantified modal logic (as in
Plantinga 1974 and Konyndyk 1986, 103–117) with supplementary principles
about certain relational predicates assumed as necessary truths (i.e., (Principle
#1), (Principle #2), and (Principle #3) are appropriated as necessary truths).
I rest my assumed modal logic on a classical first-order logic with pseudo-names
(q.v., n. 20) as articulated in Gustason and Ulrich (1989, 213–254). This is com-
monly done (see, e.g., Maydole 2012).
19. If you do not like quantification over propositions, then replace all of my talk
of propositions with talk of sentences, or statements, or whatever one considers
bearers of truth to be.
20. Pseudo-names are used for the purposes of referring “to a thing of a certain sort,”
although we “do not know or care which thing of that sort it is” (Gustason and
Ulrich 1989, 215).
21. We actually do not need (Principle #1). There is a way to run the argument
(following Williamson 2000, 318–319, in the correlative epistemic context of
Church–Fitch-like arguments) without (Principle #1).
22. Plantinga (1974, 70–77).
23. Notice that by condition (iii), platonic events occur at every metaphysically pos-
sible world.
24. Leftow (1989, 135–139; 2012, 61); McCann (2012, 201–202, or at least the
account discussed there). David Lewis seemed open to the idea when he identified
events with sets of space-time regions (D. K. Lewis, Plurality 1986, 83–84). Also,
there may be some reason for believing that Gottlob Frege thought abstracta
could be causally potent because, as Tennant (1997, 299, n. 30) suggests, he
thought that “abstract objects could indirectly influence the motions of masses
because of things we thinkers are sometimes moved to do in grasping them and
holding them to be true” (ibid.). Hoffmann and Rosenkrantz (2003, 50) provide
several objections to the view that abstracta should be understood as acausal
entities.
25. See Dodd (2007); Kivy (1993, 35–58); and Levinson (1980).
26. See Howell (2002) for a defense of the idea in the main text.
27. In what follows, I also appropriate S5 quantified modal logic, and quantified
counterfactual logic. You see examples of the latter type of logic in Yablo (2004,
125–126; 136, n. 2) and D. K. Lewis (Counterfactuals 1973).
28. Let me stipulate for my exploratory discussion that the causal relation is for-
mally asymmetric, and transitive.
29. This is akin to my counterfactual dependence principle in Weaver (2012, 307)
and Pruss’s proposition (113) in Pruss (2006, 240). Notice that Principle #1 is
weaker than Pruss’s proposition (113) because his (113) entails Principle #1,
although Principle #1 does not entail his (113).
30. Both Pruss and I used a premise almost identical to (2), except the earlier sources
were interested in (wholly) contingent events, and I’m interested in purely pla-
tonic events. This will make an important difference in the use of theorem (3).
31. The existence of the essence does not entail that the essence is instantiated. Thus,
(PPEs) does not lead me to endorse the absurd view that contingent entities exist
at all worlds.
32. Maslen (2004, 342).
33. Ibid., 343.
6 On the Irreflexivity, Transitivity, and
Well-Foundedness of Causation
I have argued that causation is an asymmetric obtaining relation that is
universal with respect to purely contingent events. In this chapter, I demon-
strate that causation is also irreflexive, transitive, and (at least with respect
to purely contingent events) well-founded.
Section 2: Irreflexivity
When I say that causation is irreflexive, what I mean is that ■∀x (~Cxx)
(where ‘Cxx’ means that x causes x) (although q.v., n. 2 in chapter 3 on
these types of formal claims). The asymmetry of causation entails that it is
irreflexive. I argued for the asymmetry of causation in chapter 3. That is
the first argument for irreflexivity. The second is somewhat of an argumen-
tum ad populum. Almost everyone in the causation literature maintains
that causation is irreflexive.1 The third argument (sect. 2.1) issues forth
from the irreflexive nature of explanation, and the fourth is what I call the
master argument presented in sect. 2.2. I will now unpack the latter two
more substantial arguments for the thesis that causation is irreflexive.
Section 2.1: Irreflexive Causation From Irreflexive Explanation2

Obtaining causal relations back successful causal explanations.3 However,
causal explanation is irreflexive because (non-mathematical) explana-
tion (in general) is irreflexive. In support of this, recall that in chapter 1,
I provided Desiderata for a Potential Explanation as part of my account
of inference to the best explanation IBE. The Desiderata for a Potential
Explanation help constitute an outline for general explanation in that when
one (generally but non-mathematically) explains that p, one does at least
four things in a meaningful, coherent, sincere, conjunction distributive, and
factive manner.
(i) One proffers a deductively sound, inductively cogent, or abductively

cogent (in the sense and in the way specified by chapter 1: sect. 4.5.3)
argument that fulfills conditions (ii)–(iv).
Irreflexivity, Transitivity, Well-Foundedness 199
Explanations are arguments. This is the third dogma of empiricism (Salmon
1998, 98–107). The dogma has an illustrious philosophical history. It can be
traced back to Aristotle’s Posterior Analytics (Bk. 1, Chapt. 2), and has been
defended in more modern times by Braithwaite (1968, 22), Hempel (1965),
Hempel and Oppenheim (1948, 137–138), Nagel (1961, 29–46), K. R. Pop-
per (1961), and Strevens (2008, 77; 2012, 448–449).
(ii) At least when p is a contingent truth, one or one’s argument reduces

surprise about the truth of p.
Explanantia render more likely their explananda such that the probability
(understood in the subjective Bayesian manner discussed in chapter 1: sect.
4.5.3) of p given the explanans, q, is greater than the probability of p. Satis-
faction of this condition does not imply that explanations render what they
explain likely (above some objective threshold). Blocking such an implica-
tion is a benefit and not a cost. This is because there are successful explana-
tions of events that are highly unlikely, as Wesley Salmon noted,
[A] uranium nucleus may have a probability as low as 10–38 of decay-

ing by spontaneously ejecting an alpha-particle at a particular moment.
When decay does occur, we explain it in terms of the tunnel effect,
which assigns a low probability to that event.4
Again, I escape Salmon’s worry by favoring a mere incremental evidential

support view of the probabilification of the explananda by the explanantia,
and by rejecting the need for probabilification that raises the likelihood of
the explananda above some objective threshold, thereby always rendering
the explananda highly likely.
(iii) In an appropriate context of communicative exchange, one or one’s

argument removes puzzlement, i.e., the puzzlement of a properly func-
tioning cognizer (in the sense at work in the characterization of war-
rantK in chapter 1: sect. 4.3) who is the target individual to whom the
explanation is being proffered, about why p holds.
Condition (iii) captures the sense in which explanations are answers to why-
questions, and so when they are communicated, they abide by an appro-
priate pragmatics for why-question answering. In addition, condition (iii)
asserts that communicating a successful explanation of p involves removing
the puzzlement of a properly functioning cognizer who seeks knowledge
of why p is the case, or why p holds. Successful explanation provision can
impart knowledge-why, insofar as explanation provision gives the properly
functioning inquirer warrant for a belief about why p holds.
(iv) One or one’s argument reduces a properly functioning cognizer’s cogni-

tive discomfort, when it is present, about why p holds.
200 Irreflexivity, Transitivity, Well-Foundedness
There is a certain appropriateness of fit or coherence (understood as more
than mere consistency)5 with respect to p and that which explains it (again
call that explanans q). Such fit generates understanding of the truth of q in
such a way that when cognitive discomfort about p is present, that discomfort
is reduced or nullified.6 This may be because q reveals a relationship between
what p is about and the environs of what is represented by p. Cognitive dis-
comfort involves the presence of a type of cognitive dissonance in the sense
employed by our best psychology (see specifically Festinger 1957; Richter
2012). According to that psychology, human persons sometimes experience
cognitive dissonance, understood as a negative interpersonal “psychological
state in which an individuals’ cognitions” (particularly their beliefs and other
attitudes) “are at odds” with one another (the quotations are from Egan,
Santos, and Bloom 2007, 978; cf. Festinger 1957, 31, who thinks of disso-
nance as a phenomenon involving “‘nonfitting’ relations among cognitive ele-
ments.”).7 I will take no stand on that which generates cognitive discomfort.
One knows from your experiences that lemons taste sour. However, sup-
pose on one occasion you eat a miracle fruit or red West African berry (call
this event b), subsequently consuming a lemon (call this event e where this
event comes after eating the miracle fruit and so involves some backward
looking texturedness). However, on this occasion (and roll the following details
into e), the lemon tastes sweet instead of sour. Puzzlement ensues, as does
some cognitive discomfort. One begins to wonder if there is something amiss
with the lemon or with one’s taste receptors. One asks, “Why did the lemon
taste sweet on this occasion?”. And one asks, “Why did e occur?”. If someone
responded to the latter explanation-seeking why-question with a scientifically
detailed description/report on the behavior of your taste receptors and ion
channels, plus the chemical composition of the lemon in event e together with
the governing laws and the like with regard to your tasting the lemon, one
would still want to know why the phenomenology of the taste experience did
not fit cognitions you have arrived at on the basis of past experiences involv-
ing lemon tasting. That is to say, cognitive discomfort would not be reduced,
neither would one’s puzzlement be removed after receiving the description.
The detailed report does nothing but describe all over again (and we can even
assume the phenomenological what-it-is-likeness of the experience reduces to
the underlying physical goings-on) the event of tasting a “sweet lemon.” In
other words, because of the nature of explanation (i.e., because it involves
removal of puzzlement and cognitive discomfort), one cannot explain e by
appeal to anything involving a description of e. Explanation is irreflexive.8
The explanation of e, in this case, involves b (specifically the miraculin in
the berry) together with the laws and details about one’s receptors and the
like. Only after learning how miracle fruit alters one’s sense of taste by coat-
ing your tongue in such a way that different receptors are stimulated than
those that are ordinarily stimulated when one eats a lemon by its lonesome
(Koizumi et al. 2011) is one’s puzzlement removed, and one’s cognitive dis-
comfort reduced. That explanation moves from b (inter alia) to e, not from
anything involving a description of e to e.
Non-mathematical explanations, generally, are irreflexive. Causal expla-
nations are therefore irreflexive because causal explanation is a species of
non-mathematical explanation (generally conceived). As was noted previ-
ously, causal relations back causal explanations, and these backing relations,
too, will never be reflexive. How could one provide a causal explanation of
a proposition p that is irreflexive and yet backed by an obtaining reflexive
causal relation? Of course, it does not follow from the thesis that causal
explanations are irreflexive, and also only ever feature irreflexive causal rela-
tions to back them, that every instance of causation is irreflexive such that
■ ∀x~(Cxx). Perhaps there are some events that do not have irreflexive causal
explanations, although they cause themselves. But recall that chapter 5:
sect. 2 argued that at least all purely contingent events have causal explana-
tions, and because all causal explanations are irreflexive, it seems we have
good evidence for the claim that causation is more generally irreflexive,
although, admittedly, I’ve said nothing about goings-on at distant meta-
physically possible worlds. But consider the fact that proposition (1),
(1) Every purely contingent event has a causal explanation featuring an

obtaining irreflexive causal relation to back it.
evidentially supports the further thesis that,
(2) Every causal relation is irreflexive.
by rendering it more likely. And very clearly, proposition (2), likewise evi-
dentially supports,
(3) Necessarily, every causal relation is irreflexive (i.e., ■ ∀x(~Cxx)).
However, because the above supports-relations are risky or inductive, we

should look for further substantiation of (3).
Section 2.2: The Master Argument for Irreflexivity

Here is the master argument for irreflexivity:9
(4) If there can be instances of reflexive causation such that ◆(∃x)(Cxx) is

true, then every event causes itself. [Premise]
(5) It is not the case that every event causes itself. [Premise]
Before passing on to the rest of the argument, I would like to ask, is the
negation of premise (5) really so bad? After all, Douglas Kutach writes that
given his distinctive methodology,
[I]t is perfectly acceptable for a metaphysics of causation to ensure

that every event causes itself. One can dismiss the importance of this
counterintuitive result merely by recognizing that self-causation is a
trivial form of causation.10
Suppose one has a complex event (D) that is built from simple events A and
B. If every event causes itself, then D causes D. However, if D causes D,
then D causes A and D causes B. But now we can ask, how can there exist
an external cause (F) of A that isn’t preempted by D? For suppose F is an
external causal producer of D by directly externally causing its component
parts (A and B). However, F cannot causally produce A in such a way that
it overdetermines A with D. This is because it is a condition on symmetric
overdetermination (see Paul 2007) that overdetermining causes be causally
distinct. Because F causally produces D, it is not distinct in the relevant way
from D. So, in what way are both D and F causally related to A? It seems
that F is going to preempt D when it reaches out and causally touches A.
But if F preempts D when it causes A, then it is hard to imagine how D
causes D. One could evade my reasoning by arguing that when D causes D,
it doesn’t thereby cause one of its component event parts, viz., A. But that
seems very implausible. On what account of causation will it follow that
the cause of a conjunctive event doesn’t cause the conjuncts? None that
I’m aware of. So, we should get rid of the thesis that every event causes
itself. It will prohibit us from affirming that complex events like A+B have
external causes like F. The negation of (5) isn’t just counter-intuitive, it is
demonstrably false.
(6) Therefore, it is not the case that there can be instances of reflexive cau-
sation such that ◆(∃x)(Cxx) is true. [Conclusion 1]
(7) If it is not the case that there can be instances of reflexive causation such
that ◆(∃x)(Cxx) is true, then necessarily, causation is irreflexive such
that ■ ∀x~(Cxx) holds. [Premise]
(8) Therefore, necessarily, causation is irreflexive such that ■ ∀x~(Cxx)
holds. [Conclusion 2]
Line (6) follows from lines (4) and (5). Line (8) follows logically from lines
(6) and (7). The material conditional that is line (7) seems obvious because
it is not possible that ~◆(∃x)(Cxx) holds and yet ■∀x~(Cxx) fails to hold
(i.e., the latter proposition is entailed by the former). The key premises are,
therefore, lines (4) and (5). I have already argued that line (5) holds. That
leaves us with the task of defending line (4).
As it turns out, not a few analyses or theories of causation yield (4). Let
me proffer two instructive examples. Give attention to what Ned Hall calls
the “[c]rude sufficient condition account”11 of causation:
(Crude Sufficient Condition Account of Causation or CSC): For any

world w, and for any event c, and event e, c is a bona fide cause of e at a
possible world w, just in case, both c and e occur in w, and for any pos-
sible world w* with the same laws as w and in which c occurs, e occurs.12
If causation can be reflexive, then given CSC, every event causes itself.
Assume that c is identical to e. And assume c, and therefore also e, occur at
w. The group of worlds w* that have the same laws as w and that feature c
will also be a collection of worlds at which e occurs. This is because, again, c
is identical to e. Thus, CSC will entail that c causes e, and because c is identi-
cal to e, c causes itself. But c was chosen arbitrarily, and so our reasoning
can be extended to any event that occurs at w. World w can be the actual
world because world w need only be a world at which events occur and laws
hold. Thus, every event at the actual world causes itself, given the CSC.
Consider now the probabilistic account of (Eells and Sober 1983, 37).
According to that account,
(Eells-Sober Probabilistic Analysis): For any event C, and for any event
E, C causes E if, and only if, Pr(E/Ki & C) > Pr(E/Ki & ~C), for every
i that has clear defined conditional probabilities (where ‘Ki’ represents
“maximal specifications of causally relevant background factors”13).
The maximal specifications in the preceding account are “conjunctions in

which, for every factor F (other than C) that is causally relevant to E, exactly
one of F and ~F . . . is a conjunct.”14
Note that this account does not include a ban on instances of reflexive
causation and that it does not entail any such ban. It will therefore allow
for possible instances of reflexive causation. To see how the account entails
that every event causes itself, note the obvious and intuitive truths that the
probability of an actually occurring event C given Ki & C is going to be very
high, if not 1, and that the probability of C (the same event) conditional on
the background causally relevant factors for C and ~C is going to be very
low. Thus, C will be a cause of itself. However, C was chosen arbitrarily, so
every event will cause itself.
Considerations like the above have led many theorists to write into their
accounts a ban on self-causation ensured either by fiat or by means of the
nature of the causal connection itself. But not all of them have done so.
Indeed, there are at least two ways to resist premise (4) without banning
outright all instances of self-causation. The first is through David M. Arm-
strong’s anti-reductive account of causation.15 According to it, for any con-
crete state of affairs C and for any concrete state of affairs E, the factP that
<C causes E.> entails that “if not-C, then in the absence of a back-up cause,
or E coming to exist uncaused, then not-E,”16 and (as in Heathcote and
Armstrong 1991) the (intrinsic) relation or connection between C and E is
nothing over and above the instantiation of a deterministic or indeterminis-
tic fundamental17 causal law understood as an entity that connects, with a
brand of nomological necessity, property instances understood as instanti-
ated Armstrongian imminent universals (these just are concrete states of
affairs as in Armstrong 1997) at the deep joints of nature. Pick out an arbi-
trary instantiation of an imminent Armstrongian universal, perhaps an elec-
tron sitting somewhere in space-time with negative charge. We know from
quantum electrodynamics, or QED, that electrically charged bits of matter
like electrons emit photons (in the sense and in the way described in Feyn-
man (1985), leaving aside his interesting interpretation of positrons). Let us
call the event of this electron’s photon emission, E. If the fully interpreted (in
the sense of chapter 1: sect. 4.5.5.3) causal laws of QED governing the pho-
ton’s emission are irreflexive causal laws barring E from lawful self-causal
determination, then the instantiation of the relevant laws will not entail
nomological self-causal determination of E. Likewise, note that if there are
relevant causal self-determination laws of QED, then any instantiation of
them will causally relate or connect (in Armstrongian necessitated glory) a
target event like E to itself. But notice that Armstrong’s account does not
support premise (4) given such laws. Only instances of causation that are
instantiations of self-causal laws yield instances of self-causation. If we call
the list of possible fundamental self-causals LN, then we have a set of pos-
sible instantiations of LN, and those alone that entail possible instances of
self-causation.
Armstrong’s account of causation is independently implausible. There
are bona fide instances of causation in the special sciences. Chapter 2 went
through great pains to show that our best psychology has need of beliefs
and that beliefs are causally formed. What fundamental laws are instanti-
ated in such cases? Psychology may be a very important special science with
some empirical success to boot (even Churchland (Materialism 1995, 154)
said folk psychology “does enjoy a substantial amount of explanatory and
predictive success”), but that does not at all provide what is needed to build
a strong case for truthfully predicating to it or its laws fundamental status.
Even if we allowed psychology into the set of fundamental scientific theo-
ries, its “laws” would more than likely be ceteris paribus laws (as Fodor
1974; 1990 has argued). But chapter 1: sect. 4.5.5.2 maintained that laws of
nature are, in general, exceptionless.
David K. Lewis distinguished between counterfactual dependence (defined
in chapter 1: sect. 4.4.1) and causation. Although both are sufficient for cau-
sation (trivial for the latter), the latter, but not the former is transitive, and
the former, but not the latter is irreflexive. He would add that “no event
can be self-caused unless it is caused by some event distinct from it” (D.K.
Lewis, Postscripts 1986, 213). But as we have seen (chapter 3: sect. 4.3.2),
Lewis’s theory is independently implausible. And so, although there may
be accounts of causation that do not entail premise (4), those accounts are
independently problematic. We are left with the irreflexivity of causation.
Section 2.3: Time Travel and Causal Loops

There are solutions to Einstein’s field equations that allow for time travel
into the past and so also past causal influence. It is thought that the causal
structure induced by solutions featuring the Gödel metric, for example,
allow for causal loops. I provide a more detailed response to these cases in
(chapter 8: sect. 6.2), but we would do well to note that all cases of time
travel that supposedly involve causal loops generate not just a violation
of irreflexivity, but also asymmetry. But again, that causation is directed is
bedrock, and arguments for it were given in chapter 3.
There are other more extravagant physical interpretations of various the-
ories that would appear to allow for time travel of some kind (e.g., tachyon
theories, and the Feynman interpretation of positrons in QED) as well, but
if those fully interpreted theories likewise allow for causal loops (what I will
call loopy-theories), they will also generate violations of formal asymmetry.
That is, I believe, too costly a price to pay. In addition, I note that we should
set aside loopy-theories because we can preclude instances of time travel
along causal loops from the domain of the nomologically possible by adopt-
ing only those solutions or interpretations that facilitate such preclusion.18
We lose nothing if we so proceed, and we help ensure that our physical
theories do not entail metaphysically impossible states of affairs. That latter
conjunct is especially good, for (again) as I argued in chapter 1: sect. 4, one
has done something unscientific if one has proffered an interpreted “for-
mulation” of a physical theory that entails a contradiction or metaphysical
absurdity.19
Some may find my reasoning completely unsatisfactory. They will insist
that I am strong arming when I appeal to the asymmetry of causation for
the purposes of prohibiting causal loops. Something else, or at least, some-
thing more convincing is needed. Here then is my best argument against the
possibility of causal loops.
When we imagine causal loops (see Figure 6.1), we ordinarily do so by
giving mental attention to something like the structure in Figure 6.1. In that
structure, event A causes B, which causes C, which causes D, which causes
E, which causes F, which then closes the loop by causing A. Even if we
were to imagine something like this structure as a structure embedded in a
space-time with closed timelike curves or closed causal curves allowing for
the loop, we would reveal a great naivety if we were to actually believe that
causal loop structures can obtain so pristinely in general relativistic space-
times. I argued in chapter 3: sect. 2, and I will argue more exhaustively
in chapter 8, that general relativity is best fully interpreted causally. Thus,
aside from A occurring and causally influencing B, we can and often do
also have a gravitational causal influence on B.20 All deposits of matter or
energy-momentum are connected to the metric “by Einstein’s equation,”21
in such a way that
in all physical equations one now sees the direct influence of the gravita-
tional field . . . any measurement of length, area or volume is, in reality,
a measurement of features of the gravitational field.22
Indeed, the metric tensor representative of the gravitational field enters the
energy-momentum tensors of matter (see chapter 8: sect. 6.3 for more on
B
A C
F
D
E
Figure 6.1 Causal Loop
this point). The influence of the gravitational field does not vanish in cases
in which forces such as electromagnetic forces act on massive and massless
bodies. This can be seen in the best partial interpretations of formalisms
of our best theories of electromagnetism. To take just one simple classical
example, the following equation,
(Eq. 1): Aα ; µµ = Rαµ Aµ (Thorne and Blandford 2017, 1219)
represents (in the Lorenz, not the Lorentz, gauge) the vector 4-potential
for the vacuum wave equation. The Ricci tensor is given by Rαμ. To see the
importance of this, note that the Ricci tensor is but a contraction of the
Riemann curvature tensor. That latter tensor is a quantitative measure of
the curvature of space-time.23 (Eq. 1) suggests what is called a “curvature-
coupling modification.”24 Such modification underwrites the truth that the
curvature of space-time contributes an influence to dynamical evolutions as
adequately described by classical electromagnetism in the appropriate limit.
Other influences from forces described by interactive theories like QCD or
QED will be causal influences as well (qq.v., chapters 3 and 7). The causal
structure of all of the relevant cases is more complicated than what a mere
loop-like representation would recommend. For while a cause of event B is A,
the cause of B will incorporate the gravitational influence due to space-time
curvature upon B, as well as any other influences due to forces or interac-
tions between substances or individuals involved in or operating near events
A–F.25 The picture of influences on B is therefore more accurately given by
Figure 6.2 (suppressing extra events and just illustrating influences, black
arrows indicate gravitational influences, and the disconnected grey arrow
represents other influences from things like electromagnetic forces, etc.):
Gravitational influence from

space-time geometry
A C
F
D
E
Figure 6.2 Complexity of Causal Structure
Quite clearly, we no longer have a causal loop (a fortiori when we con-

sider additional influences upon other event nodes in the structure). Some
will insist that it is metaphysically possible for there to be cases in which
events form a pristine causal loop. But how exactly are such merely meta-
physically possible cases supposed to go? I’m not sure I can make sense of
the idea of a particle persisting without space-time structure. If there is such
structure, then one wonders how it interacts with the substances involved in
the imagined case. But perhaps the world in question is one in which space-
time relationalism holds. Which relationalist space-time is it? Some choices
include Machian, Maxwellian, or Leibnizian space-times (Earman 1989).
Unfortunately, none of these appear to be able to afford rich enough causal
structure to allow for time travel and the like.
I will leave the task of describing merely metaphysically possible cases of
causal loop structures to the proponents of the possibility of causal loops.
I turn now to transitivity.
Section 3: Transitivity
Causation is transitive such that ■∀x∀y∀z[(Rxy & Ryz) → Rxz] (where ‘R’
denotes the causal relation and where the variables x, y, and z range over
events). As I have noted in chapter 3: sect. 1 (and see the sources cited there),
most theorists writing on causation consider that relation to be formally tran-
sitive. In fact, it seems that only until rather recently has the transitivity of
causation come under heavy fire.26 Such recent detraction from transitivity is
mostly due to a number of difficult cases thought to be counter-examples to
transitivity. I, however, agree with Hall (Price of Transitivity 2004, 182–183,
for “the central kind of” causation (ibid., 182)), that at the end of the day,
the difficult cases do not show that causation fails to be formally transitive. In
what follows, I examine the difficult cases. I articulate precisely why they fail
to serve as worries for the proponent of the transitivity of causation.
Before critically evaluating the relevant cases, I should note that the thesis
that causation is transitive best explains (in the abductive sense of chapter 1:
sect. 4.5.3) empirical data from empirical psychology regarding our concept
of causation. Generally, cognizers reason transitively about obtaining causal
relations, and they do this “even if no information on the distal event[s]” is
provided.27 There is some indirect social scientific evidence for transitivity
from both Bayes net and causal model theoretic approaches to causation
that materially imply the transitivity of causation.28 For example, according
to Judea Pearl, the transitivity (at least for binary variables) of causal influ-
ence follows from a property of all causal models, viz., compositionality
(Pearl 2009, 229, 237).
What are causal model approaches? Roughly put, they are theories of cer-
tain facets of causation that exploit models of causal structure to elucidate
and describe those facets. The models are paired with rigorous mathemati-
cal formulae that feature endogenous and exogenous variables abiding by
certain rules.29 This type of modeling has been very successful in the context
of the social and special sciences. Its successes are here appropriated as indi-
rect support for the thesis that causation is transitive (q.v., n. 29).
Section 3.1: The Tough Cases

Here are a group of some of the best (potential) counter-examples to the
transitivity of causation I could find in the literature,
(Case #1): “[A] boulder is dislodged, and begins rolling ominously

toward Hiker. Before it reaches him, Hiker sees the boulder and
ducks. The boulder sails harmlessly over his head with nary a centi-
meter to spare. Hiker survives his ordeal.”30 The idea is that the boul-
der’s dislodgement causes Hiker to duck. Hiker’s ducking causes the
prevention of his being struck by the boulder, which in turn causes
his survival. If causation is transitive, then the boulder’s dislodge-
ment causes Hiker’s survival. That seems odd.31
(Case #2): James places an armed fused bomb outside of Julie’s apart-
ment, but as Henry passes by, he defuses the bomb, and as a re-
sult, Julie survives. The idea is plainly that James’s bomb placement
causes Henry’s diffusing, which in turn causes the prevention of an
explosion, which in turn causes Julie’s survival. But if causation is
transitive, then James’s bomb placement causes Julie’s survival, and
that seems like an unintuitive result.32
There are other similar cases in the literature (e.g., Kvart 1991; Hall 2000,
183–184), but they all seem to have similar underlying structure. How
should the proponent of the transitivity of causation respond?
I believe one can defend the transitivity of causation against such cases
by appropriating two types of strategies. The first, I call the stinginess of
causal relata. As we will see in chapter 7: sect. 6, every metaphysical the-
ory of omissions understood as absences that are causal relata fails, and
attempts by metaphysicians to argue that they are somehow indispensable
to the sciences (as in Schaffer 2004) likewise fail. I will argue that there
are no good reasons for affirming the existence of absences or omissions.
However, as Paul and Hall (2013, 231) have noted, many of the putative
counter-examples to the transitivity of causation involve negative causation
and so also causal relata that are absences or omissions (e.g., in Case #1, we
have Hiker’s survival, i.e., the prevention of his or her death, and Case #2
features Julie’s survival, i.e., the prevention of her death). Thus, if my argu-
mentation in chapter 7 is cogent, then we will have no reason to suspect that
what’s involved in Case #1 and Case #2 are bona fide instances of causal
chains, and that independent of considerations having to do with transitiv-
ity or intransitivity.
Paul and Hall wrote, “there are other putative counterexamples to tran-
sitivity that make no use of omission-involving causation.”33 Here is their
potentially problematic case for transitivity,
(Case #3): “A train rushes along a track towards the point where the
track forks. Suzy flips a switch, directing the train along the left-hand
side of the fork. If Suzy hadn’t flipped the switch, the train would
have taken the right side of the fork. A few miles later, the left and
right tracks converge, and a mile after this point Billy robs the train.
Assume that taking the left side of the fork rather than the right
makes no difference to the train’s manner or time of arrival. Does
Suzy’s flip cause Billy’s robbery?”34
This case is under-described. Billy’s robbery is caused by, among other

things, Billy’s decision to rob the train, together with all manner of jointly
acting forces (we can ignore the involved forces if you’d like). Billy’s deci-
sion with other joint (partial) causes are not caused by Suzy’s flip. Here is
the final difficult case I will entertain,
(Case #4): “My dog bites off my right forefinger. Next day I have
occasion to detonate a bomb. I do it the only way I can, by pressing
the button with my left forefinger; if the dog-bite had not occurred,
I would have pressed the button with my right forefinger. The bomb
duly explodes. It seems clear that my pressing the button with my
left forefinger was caused by the dog-bite, and that it caused the
explosion; yet the dog-bite was not a cause of the explosion.”35
But again, the dog bite does not cause individual X to press the button with
their left forefinger. At best that is a partial cause in this case. X’s decision,
together with forces and the like, are joint causes of the detonation/explo-
sion. There is nothing to fear here.
There is one retort that troubles me. Causal structures cannot be as com-
plicated as I maintain because if they were, we’d have good reason to refrain
from believing that there are actual transitive causal chains. The objection is
overly aggressive. According to standard cosmology, the cosmos is a physi-
cal system that began to exist around 13.7 billion years ago. Its beginning
to exist causally produced both subsequent causal structures that are physi-
cal systems such as galaxies and states involving the operation of forces and
gravitational influences between constituents of galaxies (we could enlarge
the causal system involved and make the same point). Those structures and
influences gave rise to specific events, e.g., the formation of planets. The
physical evolution that starts with the big bang and ends with the formation
of planets constitutes a transitive causal chain, even if an immensely com-
plicated one. The cosmos caused the cause of the formation of a planet, and
thereby caused the formation of the planet.
Section 4: The Well-Foundedness of Causation36

In this section, I follow the important book Realism Regained by Rob-
ert Koons. In that work, Koons proffered a very interesting argument for
the well-foundedness of the causal relation (see Koons 2000, 113), which
depended upon the universality of causation. A close cousin of that argu-
ment proceeds as follows.
(Premise 1): All purely contingent events (defined in chapter 5) have

causes and there is a complex arrangement, absent any redundancies,
of every distinct and simple or atomic (non-complex) purely contin-
gent event Σ that is itself a purely contingent event.37
I argued for the first conjunct of this premise in chapter 5: sect. 2. The second
conjunct should be embraced under the banner of the Realism and default
setting discussed in chapter 1: sect. 3, and the theory of events developed in
chapter 7: sect. 4. I should add that the ontology of our best cosmology (what
I called standard cosmology in chapter 4: sect. 2.3) has need of the cosmos as
a whole, where one can understand the cosmos to just be an arrangement of
all atomic natural events (i.e., events that do not involve supernatural entities
if there are or can be such events).38 Setting aside worries about the super-
natural, it would seem safe to infer that the set of all distinct atomic natural
events is the same as the set of every distinct atomic purely contingent event.
(Premise 2): If all purely contingent facts have causes and there is a Σ
that is a purely contingent event (i.e., premise 1 holds), then Σ has a
cause, call it c (i.e., c causes Σ).
This seems quite obvious.
(Premise 3): If (Σ has a cause c, and there is an infinitely long causal

chain Γ that incorporates only events that are constituents of Σ), then
for any obtaining causal relation that is in Γ, the cause in such a rela-
tion is preempted by c.
Suppose c causes Σ, and suppose that there’s an infinitely long causal chain
(with no initial and/or first event) with only events that are in Σ serving as
pieces or parts in that causal chain (i.e., suppose there’s a Γ). Every event in
Γ will be caused in some way by c. Now pick out any obtaining causal rela-
tion that is a constituent Γ. Let’s say the relation is one involving C* and
E*, where C* is the would-be cause of E* (the relation is an arbitrary chain
link of Γ). C*, if preempted by c, will be barred from causally producing E*.
Because this would hold for any obtaining causal relation, there simply could
not be a Γ given that a cause c brings about Σ because every would-be cause
featured in causal relations building up Γ would fail to actually bring about
the relevant effects.39
One might object that perhaps c causes E*, and any other link in the
chain that is Γ, by way of the transitivity of causation. Unfortunately, there
appears to be no way for c to enter the chain as a first cause because our
original supposition was that Γ is an infinitely long causal chain with no
beginning, or first, or initial event. Thus, premise 3 seems well in hand. But
notice also what else we have shown,
(Premise 4): If for any obtaining causal relation that is in Γ, the cause
in such a relation is preempted by c, then it is not the case that there
is an infinitely long causal chain Γ that incorporates only events that
are constituents of Σ.
We showed this by noting how the preempting cause c will preclude causes
that are would-be links in the chain that is Γ from bringing about their
effects. But that result is incompatible with there being a Γ. Thus, given
the antecedent of premise 4, the relevant consequent follows. We can now
conclude that,
(Conclusion): Therefore, it is not the case that there is an infinitely long

causal chain Γ that incorporates only events that are constituents of Σ.
Of course, Γ was an arbitrary infinitely long causal chain of atomic purely

contingent events. Our results will therefore generalize appropriately. Cau-
sation is well-founded for all such events.
Notes
1. See the citations supporting this claim in chapter 3: sect. 1.1.
2. Something like the argument I’m about to present in this section was in the mind
of Michael Scriven (1975, 13). He, too, argues for the irreflexive nature of causa-
tion on the basis of the irreflexivity of explanation. However, both his understand-
ing of general explanation, and the explanatory factors to which he appeals to run
his argument differ substantially from my account of general explanation and the
factors of that account I use here. What is more, I do not agree with Scriven that
the sole reason why one should think of causation as an irreflexive relation is
because explanation is irreflexive, and I do not identify, as Scriven does, causation
with an explanatory relation. Such a view will entail a denial of the causal realism
defended in chapter 2. In addition, Scriven’s account is incoherent (on the reading
of him in Kistler 2006, 141, at least). According to Kistler, Scriven affirms that
causation is asymmetric. However, Scriven also says that “causation is not anti-
symmetrical” (Scriven 1975, 13). But necessarily any relation that is asymmetric
must also be antisymmetric, and a relation is asymmetric, just in case, it is both
irreflexive and antisymmetric. Scriven accepts irreflexivity. These facts are either
reasons to reject Kistler’s reading as uncharitable, or else a reason to abandon the
account. If the former disjunct is correct, one would wonder what Scriven meant
by “the distinction between cause and effect” (Scriven 1975, 13).
3. I argued in chapter 5: sect. 2 that theories of causal explanation that do not
include this backing idea (as in the work of David Lewis and Bradford Skow) are
problematic.
4. Salmon (1998, 97).
5. Robert Audi connects coherence (understood as more than consistency) with
explanation (Audi 2011, 221).
6. Audi proposes that explanations generate understanding in ibid.
7. There are some elements of Festinger’s original theory that have been rejected by
contemporary social psychologists. However, Cooper (2012) could write as late
as 2012 that “[a]s Festinger surmised, dissonance is an arousing, uncomfortable
tension state that motivates change” (ibid., 381). The general nature and qual-
ity of the phenomenon seems to have been captured by the original pioneering
work. Elliot and Devine have noted that the empirical case for the thesis that
cognitive dissonance is a type of cognitive discomfort understood “as a negative
interpersonal state” has been sufficiently established at least “in part” (Elliot and
Devine 1994, 382).
8. Compare the similar point made in Scriven (1975, 13).
9. I do not put forward this argument as something completely novel. Many theo-
rists have been worried for a long time about espousing accounts that entail
that every event causes itself. I simply use that counter-intuitive result against
theories that imply it. I’ve witnessed others do this in many seminars and con-
versations about causation. I do not know who was the first to put forward the
basic worry.
10. Kutach (2013, 8).
11. Hall (Intrinsic 2004, 266, italics removed). This account is not a mere toy theory.
Something close to it is the object of substantive criticism in Sosa (1975, 1–3).
12. Weaver (2016, 15). Paraphrased from Hall (Intrinsic 2004, 266).
13. Eells and Sober (1983, 37).
14. Ibid.
15. It is puzzling to note how Armstrong said that “[c]ausation cannot be reflexive”
(Armstrong 1997, 208). If one can show, as I do in sect. 3, that causation is
transitive, then from Armstrong’s claim that causation (not his other relation,
causation*) can be circular (ibid.), it will follow that causation can be reflexive
after all.
16. Armstrong (1999, 181).
17. Armstrong restricts the instantiated causal laws for his theory of causation to
fundamental causal laws (Armstrong 1999, 184).
18. For example, in the case of QED and the backward in time traveling positron,
we may instead adopt Paul Dirac’s hole theory, for which see Greiner (1997,
291–323). The holes are not omissions, rather positrons are nothing over and
above space-time locations in the Dirac sea at which a negatively charged elec-
tron comes into being carrying the negative energy imputed to it by the Dirac sea
(Greiner and Reinhardt 2009, 40).
19. Cf. the discussion in Segal (2015).
20. Zee (2013, 258–259).
21. Geroch and Horowitz (1979, 212).
22. Rovelli (1999, 7). I elaborate on these points of general relativistic physics in
chapter 8.
23. S. M. Carroll (2004, 128); Thorne and Blandford (2017, 1213–1215).
24. Thorne and Blandford (2017, 1219).
25. On this distinction, see D. K. Lewis (Causation 1973, 558–559).
26. Paul (2009, 159).
27. Hebbelmann and von Sydow (2014, 2339). For the data, see Baetu and Baker
(2009). Even Halpern (who denies transitivity) agrees that we “are surprised
when” transitivity fails (Halpern 2016, 43).
28. As Hebbelmann and von Sydow (2014, 2339) likewise judge. Some who lean
heavily upon causal models and the like reject the transitivity of causation (see
e.g., Halpern 2016, 41–46, who leans on cases like those I address later).
29. There are helpful introductions to causal models and structural equations for
both causation and explanation. See Halpern (2016, 9–70); Halpern and Pearl
(Part I 2005; Part II 2005); and Schaffer (Grounding in the Image of Causation
2016, 61–63). Some approaches to causal models marry them to intervention/
manipulation theories of causation. That marriage is not necessary, however.
Indeed, it seems best to think of causal models as helpful mathematical tools or
devices useful for making explicit certain features of obtaining causal relations,
and for helping with various tasks of causal epistemology (e.g., facilitating infer-
ences). That is to say, I believe those who would appropriate causal models as
entities that put us in touch with the deep metaphysical nature of causation face
the worry expressed by Paul and Hall (2013, 18–19) when they wrote,
It is an excellent question, inadequately addressed in the literature, precisely
what principles should guide the construction of a causal model. One could
be forgiven for suspecting that these principles really require one to figure out
what causes what, in the situation to be modeled, and then to select variables
and function relationships among them to fit.
(ibid.)
However, if my sentiments are correct, one might wonder why I am here not-
ing that causal model theoretic approaches to causation imply that causation
is transitive if such models do not teach us deep direct metaphysical lessons.
I answer that there can be indirect lessons learned about the nature of causa-
tion from successful modeling and causal search success. More specifically, one
can reason abductively as follows: The best explanation for why causal model
approaches are so useful to researchers in the special and social sciences (not
even Paul and Hall doubt this (ibid., 18: “[c]ausal modeling has been proven to
have enormous value, especially when it is deployed in the social sciences”) is
that they have successfully latched on to some (at least formal) features of causa-
tion. Transitivity appears to be one such formal feature.
30. See Hitchcock (2001, 276), who attributes the case to an early draft of Hall’s
(2000) paper, “Causation and the Price of Transitivity.”
31. The case is illustrated by a neuron diagram in Paul and Hall (2013, 229).
32. This case is due to Hartry Field, according to the report in Paul and Hall (2013,
215).
33. Paul and Hall (2013, 231).
34. Ibid., 232.
35. McDermott (1995, 531).
36. This section has been adopted from Weaver (2017, 107–109).
37. Premise 1 may require four-dimensionalism or eternalism. Eternalism is the the-
sis that the past, present, and future of the cosmos enjoy positive ontological
status. None are ontologically privileged.
38. It may be understood as a substance, in which case we could reformulate the
argument of this section in explicitly space-time substantivalist garb. My ESSI
account of events (chapter 7: sect. 4) allows for the possibility of an event involv-
ing all of space-time or the entire cosmos.
39. The involved causes are not involved in a case of overdetermination (cf. my dis-
cussion in sect. 2.2; cf. the explication of overdetermination in Paul (2007)).
7 Causal Relata
I have shown that causation is an obtaining, formally asymmetric, irreflexive,
and transitive relation. I have likewise shown that it is a universal and well-
founded relation when it relates purely contingent events. I will now argue
(in sect. 2) that causation is a multigrade relation. Its arity is not fixed. For
example, the relation is such that many relata, understood as joint causes are
sometimes related by causation to one relatum, and sometimes one relatum
is related by causation to many relata understood as effects. In sects. 3, 4,
and 5, I demonstrate that the only types of stuff that can enter the causal
relation are events suitably understood, and physical forces suitably under-
stood. I end the chapter with a critical discussion of negative causation.
Before I proceed with my discussion of the metaphysics of causal relata,
it will be important to disclose to the reader a certain bias in methodology.
I am interested in a philosophical analysis (in the sense of chapter 1: sect.
4.4.1) of causal relata that cuts at the deep metaphysical nature of those
relata. I do not believe that how we talk about events and causation reveals
anything metaphysically deep about the causal relation or causal relata
themselves. In what follows, I try to abide by the dictum that one should
not “substitute for questions about entities questions about sentences about
entities.”1 I will therefore ignore the massive piles of literature that have
sought to establish this or that theory of causal relata or causation by way
of some theory about how we speak about events or causation (in English!).
A number of philosophers would applaud this approach. Consider these
comments (which provide reasons for my approach) from distinguished
metaphysicians, Roderick Chisholm and David Lewis,
Many contemporary philosophers have developed theories about the

nature of events on the basis of theories about the nature of what would
be an adequate semantics for describing events. . . . I find it very difficult
to see how such a linguistic approach could throw any light upon the
nature of nonlinguistic things—unless the linguistic or semantic theory
that is proposed is itself derived from prior considerations about the
kinds of things there are.2
216 Causal Relata
There is no guarantee that events made for semantics are the same as
the events that are causes and effects.3
With the above admission out of the way, I can proceed to my discussion of
the metaphysics of causal relata and the like without having to worry about
semantic considerations.
Section 2: Causation Is a Multigrade Relation

Causation is a multigrade relation. Its arity is not fixed. Consider the fact
that our best applied nuclear physics and its history delivers to us the follow-
ing data (keeping in mind that causation is transitive; see chapter 6: sect. 34),
(Datum 1): In a gun barrel assembly (of the kind used in the first fission
atomic bomb design), firing a sub-critical piece of uranium into a
distinct sub-critical piece of uranium causes/generates a critical mass,
and that critical mass causes/generates an explosion.5
(Datum 2): On July 16, 1945, with respect to the Trinity device, a nu-
clear fission process involving plutonium caused an implosion, which
itself caused critical mass. Critical mass, in turn, caused a massive
nuclear explosion.6
(Datum 3): On August 9, 1945, through the same type of process that
caused the Trinity device to reach critical mass and cause a nuclear
explosion, the Fat Man atomic bomb was detonated above the Japa-
nese city of Nagasaki.7
With respect to (Datum 2) and (Datum 3), it is a deliverance of unfortu-

nate perceptual experience, knowledge-conducive testimony, and science
that the relevant nuclear explosions (events) caused a great many more (in
number) events that are destructive effects.8 With respect to (Datum 1), it is
a prediction of science that were such causal activity to take place, numer-
ous destructive effects would result. The above data constitute conclusive
empirical evidence for the thesis that actual instances of causation some-
times relate a smaller amount of occurrences or individual events (e.g., a
nuclear explosion, or the event of a material reaching critical mass) to vary-
ing large numbers of events understood as effects.9
Our best biological theory of the unicellular green algae known as Chlam-
ydomonas reinhardtii paints a very interesting picture of their life cycle.
Two gametes (one mt-, the other mt+) can release their cellular walls and
begin to mate (this stage of the process is described as mating structure acti-
vation). That activation produces a fusion of the mating structure belonging
to mt- and the fertilization tubule of mt+. That fusion causes a cell fusion
from the inside out.10 Here we have a case of joint causation. Two gametes
causally produce or create one new biological organism (again, adding more
richness, e.g., gravitational effects and the like, to the causal structure only
Causal Relata 217
helps my case). We therefore have knowledge-conducive empirical evidence
for the thesis that there are instances of joint causation (i.e., an obtaining
relation in which many events give rise to one, viz., in this case, the new
existence of a single biological organism).
There are strategies for unifying the causes/effects in the two previously
discussed cases. One could embrace mereological universalism and maintain
that the disastrous effects of the nuclear explosion are a single individual sum.
Such a response is flawed in that it does not successfully rule out the instance
of causation to which I have appealed for the purposes of motivating the
causation-is-a-multigrade-relation thesis. This is because, although mereological
universalists acknowledge the existence of such a sum, they also acknowledge
the existence of the disparate entities or events to which I have appealed in
the preceding case. I’m aware of no view of the metaphysics of concrete par-
ticulars or events, according to which the horrible effects of nuclear explo-
sions come out as unified entities or objects at the expense of the preclusion
of the relevant pluralities from the realm of existence. Likewise, I’m aware of
no view of concrete particulars or events that would view the behavior of the
two gametes as the behavior of but one entity, individual, or object at the cost
of precluding the two separate gametes from the realm of existence. Our case
for the multigrade nature of causation seems well in hand.
Section 3: The Competing Proposals11

I have argued that causation is a multigrade relation. But what are the types
of things that can be so related? In this section, I examine the many pro
posals in the metaphysics literature before turning to an explication and
defense of a specific metaphysics of events and forces in sects. 4 and 5.
Section 3.1: Making a Good Theory of Causal Relata Good Enough

Recall that the present work on causation rests atop a metaphysicalC system
and prolegomena detailed in chapter 1. That system assumed a default setting
of Realism (chapter 1: sect. 3) about entities that extends as far as is rationally
permissible. That system also appropriated a robust scientific realism (chap-
ter 1: sect. 4.5.5.5). Given the aforementioned two assumptions, it is highly
intuitive (at least it seems to me (in the sense I discussed in chapter 1: sect.
4.5.2)), and in many cases, it seems to be in accord with our best science, that
the following represent undeniable facts (some modal) about events proper,
such that any plausible theory of events ought to be able to countenance them,
or else suffer great theoretical cost,
(Datum 4): There are coincident events.
One’s theory of events ought to allow for two of them (in the simple case) to
occur at the same space-time location. I will assume the Minkowskian full
218 Causal Relata
interpretation of special relativity (SR) for the appropriate limit presented,
in part, in the sources cited in note 12.12 I will also assume the causal full
interpretation of general relativity whose designative formulation is outlined
in Wald (1984), and whose ontology is partially presented in chapter 3: sect.
2 and chapter 8. Given these assumptions, one can say that one’s theory of
events ought to allow for coincident events that occur at the same invariant
space-time location. An implication of this for at least the special relativistic
limit (local space-time geometry described by an approximately flat metric)
is that when rational observers situated in different inertial frames of refer-
ence execute their experiments so as to determine the spatial location and the
time of the coincident events, they will all agree that the events in question
occurred coincidently, or at the same space-time location, although the exper-
imental notes of the observers may disagree about the locations and times of
the events (Mermin 1989, 24–25). What I am describing is a special relativis-
tic phenomenon captured by the special relativistic principle of the invariance
of coincidences (as I was taught by Mermin 1989, 24–25), which says that
“[w]hen one observer [correctly] says two events coincide in space and time,
so will all other observers” who report accurately on such events (ibid., 24).
Here is N. David Mermin’s reason for invoking this principle of SR,
The reason we take this extra principle to be true is that when two
things arrive at the same place at the same time, they can have a direct
and immediate effect on each other. (Consider, for example, two cars,
one moving north and the one moving south, arriving at the same place
at the same time.) As a result of such conjunctions noticeable things
happen. It would be absurd if there were one observer who said the
two things were never at the same place at the same time and another
observer who said there were definite modifications in the things (e.g.,
dents in the cars) that could only have arisen from their having been in
the same place at the same time.13
Here is Einstein expressing a similar idea (dripping with some verificationist

flavor)14 through what has come to be called the point coincidence argument
(a motivation for the general covariance of general relativity (GR)),
[T]his requirement of general co-variance . . . is a natural one. . . . All our

space-time verifications invariably amount to a determination of space-
time coincidences. If, for example, events consisted merely in the motion
of material points, then ultimately nothing would be observable but the
meetings of two or more of these points . . . the results of our measur-
ings are nothing but verifications of such meetings of the material points
of our measuring instruments with other material points, coincidences
between the hands of a clock and points on the clock dial, and observed
point-events happening at the same place at the same time. The introduc-
tion of a system of reference serves no other purpose than to facilitate the
Causal Relata 219
description of the totality of such coincidences. . . . As all our physical
experience can be ultimately reduced to such coincidences, there is no
immediate reason for preferring certain systems of co-ordinates to others,
that is to say, we arrive at the requirement of general co-variance.15
There are two points I’d like to draw from Einstein and Mermin. First,
whatever one’s attitude is about the proffered motivating reasons for this
or that principle, it seems clear that SR and GR have need of coincident
events.16 Both Einstein and Mermin maintain that one can experimentally
verify such coincidences in relativistic contexts. They are right. I know of no
challenges to this in the relativity literature.
Second, and perhaps most interestingly, I note how some have insisted
that we read Einstein’s remarks in such a way that we take them as advocat-
ing a way of understanding the observables understood as invariant quanti-
ties (following the characterization of general relativistic observables in P.G.
Bergmann 1961) of GR (Rickles Hole Problem, 2008, 117). But that is not
correct. To understand what an invariant quantity is, you need the notion of
an inertial frame. For example, consider that in the special relativistic limit, it
is common in relativity physics to define Lorentz scalars (invariant quantities
of a particular kind) as quantities that are invariant, but as in all other relativ-
istic cases of the relevant kind, what that means is that they are invariant under
the Lorentz transformations. But the Lorentz transformations are, in the most
common of examples, repeated in not a few textbooks, transformations for
two inertial frames in the standard configuration17 described by unprimed and
primed systems of coordinates. The latter system in that configuration repre-
sents a “moving” system, whereas the former represents a “stationary” one.
(Lorentz Transformations):
x − vt
x′ =
1 − v2 / c2
y′ = y
z′ = z
t − vx / c 2 v2
t′ = (where the factor 1 / 1 − is occasionally abbrevi-
1 − v2 / c2 c2
ated as γ, and given that v is, with respect to the two systems, relative
velocity)
Again, the coordinate systems describe features of inertial frames (see Savitt
2011, 548). The notion of a frame of reference is part of the meaning of
a (local) Lorentz invariant quantity. In full GTR, frames of reference or
coordinate systems are part of the meanings of the various types of invari-
ance of quantities as well, although there are Lorentz scalars in GTR, too.
The quotation from Einstein just provided says that the meetings (call them
220 Causal Relata
Einstein-meetings, which are understood as coincident relativistic events
involving moving particles) motivate the introduction of frames. But if meet-
ings just are observables understood as invariant quantities, and invariant
quantities are partly defined or understood in terms of frames, then Einstein
would be recommending that (at least in part) frames of reference moti-
vate the use of frames of reference for descriptive purposes. That is clearly
wrongheaded. Einstein is judging that frames of reference help facilitate
verification and description of Einstein-meetings that are not partly defined
by means of frames of reference. That is to say, and this is the crux of the
matter, Einstein meetings are something more fundamental than invariant
quantities. They can be accurately described as such, but they are, in terms
of their metaphysical natures, something much more fundamental than that.
I think Einstein’s point coincidence argument is best understood amidst
a phenomenon-first natural philosophy, an approach already noted as one
at home with Einstein’s more general scientific methodology (see chapter 1:
sect. 4.5.5.5). And fortuitously, my approach to the matter coheres nicely
with Mermin’s previously quoted causal interpretation of the principle of
coincidence. Here then is the main thrust of my reasoning/argument,
(A) The Einstein-meetings are interactions involving coincident relativistic

events.
Because Einstein focused on the motions of particles in his comments about

Einstein-meetings, one might judge that the interactions of principle (A) are
strictly reducible to intersecting worldlines (cf. Dowe 2000). But that view is
mistaken (as was argued in chapter 3: sect. 4.3.1; cf. chapter 8: sect. 1). As
has been noted by Earman (2002, 12), Einstein must incorporate Einstein-
meetings between fields. However, Earman did not go far enough. Einstein
must also incorporate into Einstein-meetings the propagation of gravita-
tional waves understood as undulations of the gravitational field or space-
time itself without worldlines in a vacuum. These waves interact with matter.
The main reason we use frames is to help with the description and expla-
nation of these types of interactions between these types of entities. They
(that is, the frames) are therefore non-essential pieces of at least GR and can
be dispensed with. Let’s go right after the interactions themselves. What is
indispensable to SR and GR are the aforementioned interactions that often
leave detectable evidence in their wake. And so (Datum 4), introduced for
the study of causal relata has led us to another argument for inserting causa-
tion into physics because interactions are causal (q.v., chapter 3: sect. 2.3).
Am I suggesting that Einstein-meetings are not invariant? No. They are.
But what that really means is that those features of the designative formula-
tion or formulation of GR that represent them possess a certain mathemati-
cal property. The Einstein-meetings themselves have a metaphysical nature
deeper than that. That metaphysical nature metaphysically explains why the
formulation and/or designative formulation features representations of the
meetings that are invariant.
Causal Relata 221
(Datum 5): It is metaphysically possible that there exists but one of Max
Black’s18 two pure iron spheres that persists over time19 in a Leibniz-
ian space-time without undergoing any change of position,20 spin,
internal vibrations, etc. (i.e., the sphere just sits at a spatial location
and persists over time there).21
Leibnizian space-times have rich enough topological and metric structure

to afford spatial locations of iron spheres on three-dimensional spacelike
hypersurfaces constituting privileged surfaces of absolute simultaneity.
I see no reason for rejecting Datum 5. The mathematics and physics of the
assumed space-time is well-understood, and the persistence of a pure iron
ball presents no metaphysical or physical problems. It would be a very bad
consequence of one’s modal epistemology if it failed to provide epistemic
justification for beliefs with Datum 5 as their contents.
Now consider the more substantial metaphysical thesis,
(Datum 5.5): In the metaphysically possible case reported on in Datum

5, the temporal persistence of Max Black’s sphere involves causation
or causal continuity such that the relevant fact “about persistence” is
“grounded in facts about causation.”22
I cannot properly defend Datum 5.5 here. I note only that something close
to it is or would be affirmed by many philosophers from varying philosophi-
cal traditions (e.g., Armstrong 1980; Collier 1999, 216; D. K. Lewis S urvival
and Identity, 1983; Perry 1976; Shoemaker 1997; Wasserman 2005).
(Datum 6): It is metaphysically possible that a relatum or some relata of

causation x or xx causally produced our cosmos.
This datum may seem completely indefensible to atheistic and/or agnostic

philosophers, but I do not believe things are so dire. Sean M. Carroll and
Jennifer Chen have articulated a cosmogonic model according to which
our space-time is produced by the dynamical activity of a background
de Sitter mother universe together through a quantum tunneling process
(see Carroll and Chen 2004; S.M. Carroll 2010; cf. the critical discus-
sion in Weaver 2017). Here are partial descriptions of the model from
Carroll,
de Sitter space, the solution of Einstein’s equation in the presence of a

positive cosmological constant, is unstable; there must be some way for
it to undergo a transition into a state with even more entropy. Chen
and I imagined that the mechanism was the quantum creation of baby
universes, as suggested by Farhi, Guth, and Guven. . . .23
What we see is simultaneous fluctuation of the inflaton field, creating

a bubble of false vacuum, and of space itself, creating a region that
222 Causal Relata
pinches off from the rest of the universe. The tiny throat that connects
the two is a wormhole. . . . But this wormhole is unstable and will
quickly collapse to nothing, leaving us with two disconnected space-
times: the original parent universe and the tiny baby.24
Although I believe the model suffers from serious problems, I do not main-
tain that it cannot be revised so as to render it metaphysically possible.25 At
a world w at which the universe is created in such a way that it becomes
causally isolated from the background processes that generated it, there
exists, at w, a causal process (whether the involved causation can be reduced
or not) responsible for it. It seems then that there is a perfectly naturalisti-
cally respectable means whereby one can justify belief in Datum 6.
(Datum 7): Complex arrangements of events are causally efficacious.
This datum follows from our previous discussion of biology and nuclear
physics in sect. 2. There I illustrated how many events can causally produce
a smaller collections of events.
(Datum 8): At least some complex arrangements of events are them-

selves causally efficacious (complex) events.
In quantum statistical mechanics (QSM), there is a ratio called the mean

occupation number of a quantum state, η, given by,
 h3
(Eq. 1): η = =  , where gs provides the multiplicity of a quan-
 states g s
tum particle (having to do with spin), h is Planck’s constant, Ɲ is the
relativistic distribution function (which is reference frame independent),
and Ɲstates gives the number density of states featured in the appropri-
ated phase space being used to describe the system of QSM.26
The above equation provides the ratio (quoting Thorne and Blandford) “of
the number density of particles to the number density of quantum states.”27
According to Thorne and Blandford (ibid.), when the mean occupation
number is such that η >>1 for bosonic particles, those particles dynamically
evolve in nomologically correlated fashion like a wave of classical mechan-
ics. When the large collection of bosonic particles are photons, they evolve in
accordance with the wave equations of Maxwellian classical electrodynam-
ics. Electromagnetic waves propagate in a way that is indicative of a causal
process (one time-varying change is the cause of the next, q.v., chapter 8).
These waves can causally affect the motions of particles.28 One can think of
the nomologically correlated (quantum mechanical) photons like a complex
object or event (given our Realism of chapter 1: sect. 3). When the wave undu-
lates (a higher-level (complex) event) it does so on account of the ways the
Causal Relata 223
underlying microconstituents are at the relevant times. These ways they are
constitute micro-events. That the lower-level individuals are certain ways
is captured well by the sense in which they are nomologically correlated.
There is some way they are arranged, and there are some ways they are
interacting such that they are nomologically correlated so as to serve/func-
tion as the quantum mechanical reductive, grounding, or realization base for
the higher-level entity. We can say, more accurately then, that the arrange-
ment of events realizes, grounds, or otherwise metaphysically relates to and
explains the time-varying change of the wave. The time-varying changes of
the wave are, intuitively, higher-level (complex) events. The causal relations
between such events (the ways the wave are at times) yields the causal pro-
cess of propagation.
I will now evaluate various theories of causal relata and argue that many
of those theories fail because either they do not properly recover data points
4–8 or they fail for independent reasons.
Section 3.2: Are Causal Relata Concrete Particular Objects?

For W.V.O. Quine, causal relata are events, and events are physical objects.
He said, “[p]hysical objects . . . are not to be distinguished from events,
or . . . processes. Each comprises simply the content, however heteroge-
neous, of some portion of space-time, however disconnected or gerryman-
dered.”29 E. J. Lemmon held a similar view. He said that “we may invoke a
version of the identity of indiscernibles and identify events with space-time
zones.”30 For Lemmon then, two events are identical, just in case, they occur
at the same time and at the same place.31
Because events are physical objects for Quine, if one makes manifest a flaw
in his theory of physical objects, one would also reveal a flaw in his theory
of events. Given space-time substantivalism,32 Quine’s underlying theory of
physical objects entails that space-time itself is not a physical object because
space-time does not reside in a region of space-time. However, space-time
is a physical object on space-time substantivalism (i.e., space-time substan-
tivalism entails that space-time is a physical object). Therefore, Quine’s
theory of physical objects, and therefore events, is incompatible with space-
time substantivalism. That is a cost (q.v., chapter 8: note 78).
Here is a second problem for Quine, and a distinctive problem for Lem-
mon’s account. Datum 4 teaches us not to preclude from our ontologies
coincident events. Donald Davidson articulated a famous instance of coin-
cident events, viz., a sphere or metal ball that heats up while spinning. The
event that is the spinning of a ball does not appear to be identical to the event
of the ball’s heating up—although in actuality—the two events coincide with
respect to space-time location (or what is sometimes called a world-point).33
And as Quine himself pointed out, the event that is the ball’s heating up
transpires slowly, whereas the event that is the ball’s rotating or spinning
transpires quickly (Quine 1985, 167). If the two events are really identical,
224 Causal Relata
how can that one event transpire both slowly and quickly?34 This question
constitutes a clear problem for Lemmon because his theory entails that the
coincident events are one and the same. The question is also problematic for
Quine because on his view, a physical object x, just is “the material content”
of some region of space-time (Quine 1985, 167). If there is but one ball serv-
ing as the material content of a space-time region, then despite the various
activities of the ball, there will remain at the relevant region but one event
that is the ball itself. Quine’s view does not countenance Datum 4.
While still insisting that events are physical objects, Myles Brand attempted
to ameliorate Lemmon’s individuation conditions for events by positing that
two events e1 and e2 are identical, just in case, necessarily, for any space-time
region r, e1 occurs within r if, and only if, e2 occurs within r.35 Notice that
Brand’s account of the individuation conditions for events is not susceptible
to Davidson’s spinning ball case because the ball may (possibility) spin and
yet fail to heat up at the relevant region. Unfortunately, Brand’s adjusted
theory (if we are restricted to it as the sole type of entity that can stand in
causal relations) does not account for Datum 6. The causal production of our
orphaned cosmos by a metaphysically possible Carroll–Chen background
mother universe features a causal relatum in the effect slot that does not occur
within a space-time region because it is our space-time or cosmos in toto.36
Another view in the tradition of analyzing events in terms of physical
objects is the view of Donald Davidson. He argued that causal relata are
events that “are concrete occurrences,” which are themselves particulars
that are located in space-time.37 One individuates an event e by appeal to
its effects and causes. Thus, e1 and e2 are identical, just in case, e1 and e2
produce the same effects and have the same causes. Davidson dismissed the
idea that e1 and e2 could be without a cause because he was an advocate
of the universality of causation. Like Brand’s account, Davidson’s theory
avoids the spinning/heating ball counter-example; however, it too cannot
countenance Datum 6 and is therefore problematic in the same way Brand’s
theory is problematic.
Section 3.3: Are Causal Relata Changes?

Lawrence Lombard has said that changes are movements of physical objects
“at an interval of time in a quality space” and that causal relata (under-
stood as events) just are such changes (Lombard 1986, 178). The view faces
an important problem. It does not leave room for the conjunction that is
Datum 5 and Datum 5.5.38 Such data entail the possibility of a persist-
ing iron sphere that does not enjoy movement “in a quality space” and so
does not undergo the changes indicative of the presence of a causal rela-
tum. However, Datum 5.5 says that in the relevant possible case, there is
causation entailed by the persistence of the sphere. However, if Lombard’s
account is a true philosophical analysis of causal relata, it will hold even at
Causal Relata 225
the world described by Datum 5 and 5.5. It doesn’t, so it fails as a philo-
sophical analysis of causal relata.
Carol Cleland tries to revive the change theory of causal relata by enrich-
ing it with the metaphysics of existential conditions.39 Existential condi-
tions are those entities that undergo changes. Existential conditions are
built from phases and states. Phases are determinable properties. States
are the values of those determinable properties (determinate properties).40
Phases that are concrete (concrete phases) are existential conditions that
are essential to events (causal relata). Events are property instances or
tropes of some kind.41 The determinate properties of these tropes or prop-
erty instances are changes. Although the underlying tropes are not repeat-
able, sometimes their determinable values (the changes) are repeatable.42
And although Cleland’s treatment of generic changes is interesting, I will
focus my attention exclusively on concrete changes because she defines
events in terms of them.43
(Concrete Changes): “A concrete change R is a pair {x, y} such that x is

the exemplification of a state s by a concrete phase CP at a time t and y
is the exemplification of a state s’ by a concrete phase CP’ at a time t’,
where (i) t is earlier than t’, (ii) CP is the same concrete phase as CP’,
and (iii) s is not the same state as s’.”44
Events are described well by pairs {[concrete phase CP, initial state s, time t],
[concrete phase CP', terminal state s’, time t’]}, where it is understood that
the square brackets pick out exemplification, CP = CP', t’ > t, and s ≠ s’.45
Cleland’s account is truly brilliant, and its benefits are legion. However,
it suffers from the same problem as Lombard’s theory. It does not recover
Datum 5 and Datum 5.5. The concrete phase involving the iron sphere at
time t exemplifies the same state as the concrete phase involving the same
iron sphere at t’ (where t’ > t). But again, Datum 5.5 tells us we have a case
of causation (and remember that causation is irreflexive) due to persistence
over time of the iron sphere. The relata in that causal relation are not as
envisioned by Cleland’s account. So, it fails as a philosophical analysis of
causal relata.
Section 3.4: Are Causal Relata Sets of Space-time Regions?

According to David K. Lewis, causal relata are events, and events are
properties of regions of space-time. But Lewis is a class nominalist about
properties, and so events are sets of space-time regions (see D. K. Lewis,
Plurality 1986; Events 1986, 255).46 Complex events for Lewis are mereo-
logical sums that can have sub-regions as mereological proper parts.
Thus, some events are mereological proper parts of others. In addition, an
event C can imply another event E because that C occurs at region R can
226 Causal Relata
entail that E occurs at the same location (ibid.). No event ever recurs in
two separated expanses of the actual world, and every event “occurs if
and where and when there is a region that is a member of it.”47 Lewis
also maintained that events have essences, and that these are conditions
that some expanses of the actual world satisfy to ensure that the relevant
events occur there.48
There are three problems with Lewis’s account. First (and again), it
requires class nominalism, but class nominalism is false. Suppose that prop-
erty instances (and by that I just mean instances of property exemplification)
are collections or sets of individuals. On that supposition, a radical essen-
tialism is born. Sets have their members essentially (see van Cleve 1985). My
mug being red amounts to it being a member of a collection of red things.
But my mug could have failed to exist, and at the world at which every other
red thing exists save my mug, nothing would really have redness, for that
unique set would fail to exist. Reductio ad absurdum.49
Second, consider the fact that if property instances are sets of particulars,
then the following principle holds,
(B) β is φ, just in case, β is a member of the class or set of φ concrete particu-

lars (following the summary discussion in Moreland (2001)).
The class or set to which β belongs must be an abstract object of some sort,
for it cannot be the mereological sum of the members of the respective class
(as Armstrong, Nominalism and Realism 1978, 29–30 noted). If you do not
believe in abstract objects, then the ontological commitment here is a cost.
But more importantly, two classes are identical, just in case, their members
are identical (i.e., they have the same members). But if that’s right, then
class nominalism faces a well-known problem of companionship, for it is
perfectly metaphysically possible that there be two distinct properties that
are necessarily coextensive, a fortiori, there actually are distinct necessar-
ily coextensive properties. Consider the example from Parfit (2011, 297)
involving the property of being the positive square root of four, and the
property of being the only even prime number. These two properties are
clearly necessarily coextensive although they are also clearly not identical.50
If one were to fiercely insist that necessary coextensionality of properties
yields the identity of those properties, then (as Shafer-Landau (2003, 91)
pointed out) the property of being identical and the property of being nec-
essarily coextensive would come out identical. But that’s clearly absurd. As
Shafer-Landau stated, “[i]t seems that we are referring to different features
when we assert the existence of an identity relation, as opposed to one of
necessary coextension.”51 An insistence on such a relationship between nec-
essary coextensionality and identity entails that if there were properties that
everything necessarily had, then those properties would be identical (this is the
point made by Majors 2005, 488). Thus, the property of being self-identical
and the property of being a member of one’s own singleton set would be
Causal Relata 227
identical. The property having one’s own haecceity and the property of
existing would be identical as well. These results count as costs.
That events end up being sets is problematic for a third reason. Sets are abstract
objects, and typically, theorists maintain that abstract objects do not stand in
causal relations.52 It seems that Lewis agreed, but he may have been willing to
give up the almost obvious truth that sets are abstracta. I believe more details
are required. One needs a story about why it is that sets aren’t abstract. If one
is unwilling to bear that burden, then one at least needs a story about how it is
that abstract objects can stand in causal relations given that they are immutable,
and non-spatiotemporal necessary (this may have been the option Lewis pre-
ferred, I’m not sure). Chapter 5: sect. 3 opened the door to understanding some
abstracta as causal effects, but only for the purposes of exploratory speculation.
If one would like to advance a theory of causal relata in a way that obeys prin-
ciple K-A (chapter 1), then one would do well to ask about the underlying theory
of abstracta and precisely how it is those types of objects are apt for being effects.
Section 3.5: Are Tropes Causal Relata?

Jonathan Bennett’s theory of causal relata and events is very complicated as is
his theory of causation. He says that event c causes event e, just in case, c has a
companion fact, some part of which is an NS condition for or of e (J. Bennett
1991, 627). A companion fact of an event c is (roughly) a factP that reports on
the obtaining of the property instantiation at a time and location that is c with
further details about the goings-on where the event happens. The types of
entities that can be NS conditions and the things that can have NS conditions
are factsP (Bennett 1988, 45, 49). A fact F1 is an NS condition for a fact F2, just
in case, F1 is “a Necessary part of a Sufficient condition for” F2.53
Essential to Bennett’s characterization of causation is the notion of an
event understood as a property instantiation. These instantiations are tropes
(particularized properties) at zones or regions of space-time (J. Bennett
1988, 88, cf. 90, 128, 156).54
There are two problems with Bennet’s account. First, the theory of causa-
tion to which it is wed requires events and their companion facts or else it
will suffer from the following difficulty. As stated, NS conditions for effects
need only be “some part of” (1991, 627) the companion fact of an event.
Part of the companion of c (the purported cause) is the time of the rel-
evant instantiation. How can the mere time of c be an NS condition for e?
Bennett may reply that there is indeterminacy involved in what part plays
the relevant role, but my challenge consists of pushing him to provide a
metaphysics that keeps us from merely appealing to a time. Moreover, NS
conditions relate factsP, not parts of companion facts or property instances.
So, there is some inconsistency at work. If we fix matters by strengthening
the account of causation so that it says, c causes event e, just in case, c has
a companion fact that is an NS condition for or of e’s companion fact, we
will still face the challenge of recovering Datum 6. The beginning of our
228 Causal Relata
cosmos cannot be a Bennett-style event because it is not an instantiation of a
property at a zone or region of space-time. Our space-time itself comes into
being. But if there can be no appropriate property instantiation, then there
can be no appropriate companion fact for it (so the beginning of our cosmos
as reported on in Datum 6 can neither be involved in a cause or effect).
Douglas Ehring maintains that causal relata are property instances under-
stood as persisting tropes (see Ehring 2009, 389–390, 406–407; 1997,
71–115). He said, “causal relatum . . . consist in a ‘trope at a time’, the
existence or presence of an enduring trope at a time and location.”55 Ehring
motivates his account of causal relata by way of his persistence theory of
causation. There appear to me to be two problems with Ehring’s theory.
First, if the theory holds, then tropes endure by being wholly present at
the present moment of time, and so the theory is true, only if, perdurantist
theories of persistence are false.56 If one’s theory of causal relata could avoid
a commitment to either endurantism or perdurantism, that would be theo-
retically preferable because both positions are very controversial. Second,
Ehring’s view does not make room for Datum 8, viz., that there are causal
relata that are causally efficacious complex events that have other events as
their (non-mereological) parts.57 This is because a trope is a particularized
property, it does not allow for rich enough structure so as to accommodate
Datum 8, as Helen Beebee noted in her review,
Ehring’s view that tropes are causal relata runs counter to the strong
intuition that causes and effects are generally, or at least often, multi-
featured events like parties, wars and chess championships.58
Section 3.6: Are Causal Relata Property Exemplifications?

A number of philosophers have argued that events are property exemplifica-
tions at times.59 As Jaegwon Kim famously put the position,
An event or state can be explained as a particular (substance) having

a certain property, or more generally a certain number of particulars
standing in a certain relation to one another.60
We think of an event as a concrete object (or n-tuple of objects) exem-

plifying a property (or n-adic relation) at a time. In this sense of ‘event’,
events include states, conditions, and the like, and not only events nar-
rowly conceived as involving changes.61
Events, therefore, turn out to be complexes of objects and properties,

and also time points and segments, and they have something like a
propositional structure; the event that consists in the exemplification of
property P by an object x at time t bears a structural similarity to the
sentence ‘x has P at t’.62
Causal Relata 229
Kim’s account of events includes discernible identity conditions for them.
According to those conditions, the event of substance O’s being P at t, or the
triple {O, P, t} is identical to another event O*’s being P*, at t*, or the triple
{O*, P*, t*}, just in case, O = O*, P = P*, and t = t*.63 Now let the event of
O’s being P at t, be E. It follows from Kim’s identity conditions that E cannot
have constituents other than being P, O, and t, for if it did, E would not be
identical to E. Thus, no matter what world you move to, if E occurs at that
world E must feature O, being P, and t (on the uncontroversial assumption
that identity is a necessary relation) as constituents. Thus, Kim’s structured
complex view of events entails that events have their constituents essentially.
Kim’s account has one flaw, and it is not the well-worn point that it pro-
duces too many events.64 Rather, the flaw can be identified by looking, as
we did in chapter 4: sect. 2, to the GRW approach to quantum mechanics.65
Although I do not consider that approach to be empirically on par with
other quantum theories (e.g., the Bohmian and/or interactive approach),66
I do believe that the picture it paints of the kinematics and dynamics of some
quantum systems is metaphysically possible, and I rest upon anyone who
would challenge such a thesis the burden of proving the contrary. Thus, if
Kim’s account of events is a true philosophical analysis of their nature, then
it should hold at a GRW-world.67
The metaphysically possible world I ask the reader to visit is one at which
physical systems are accurately described by the GRW model in Ghirardi
et al. (1988) or something near enough.68 The picture one can paint of this
world is not unlike what has already been said about what is represented
by the GRW approach broadly construed (see, again, chapter 4: sect. 2).
Quantum physical systems evolve in accordance with the deterministic lin-
ear Schrödinger equation interrupted by stroboscopic evolutions governed
by “a stochastic modification” of that equation, “consisting in the assump-
tion that each constituent particle of any physical system is subjected at
random times to a random localization process,”69 i.e., a process involving
the random or spontaneous collapse of the wave function ψ. The measure-
ment problem (for which see Ney 2013 for a clear presentation) is solved by
using the stroboscopic evolutions caused by a collapsing wave function to
save determinate measurement outcomes. The wave function is not only the
chief means whereby one assures shifts out of superpositions, it is, on this
approach, understood to just be the states of quantum physical systems.70
This view of the ontology of the wave function amidst GRW’s partial inter-
pretation is at home in what is called GRWF or GRW with flashes.71 It was
embraced (at least for exploratory purposes) by Adrian Kent who wrote,
[W]e first clarify our interpretation. We regard the collapse centres as

physical events in Galilean space-time. All other quantities are taken
to be merely part of the mathematical formalism, having no physical
referent. In particular, we take the wave function as just a convenient
mathematical representation of the system’s physical state at time t. The
fundamental description of this state is the historical record of collapse,
230 Causal Relata
i.e. the collection of times and centres for all collapses prior to t, distin-
guished by particle type.72
The best exegesis of Ghirardi, Nicrosini, Rimini, Weber, and Kent suggests
that they espouse the view that the only concrete particular object (apart
from neo-Newtonian or Galilean space-time for Kent’s approach) that is
part of the partial interpretation of GRWF is the wave function (identical to
quantum states themselves). That one object collapses, and its collapses are
flashes that drive the dynamics of the theory, thereby building its history.
Some friends of GRW approaches will object to the assumed partial inter-
pretation. For example, Valia Allori maintains that regarding the wave func-
tion as “‘all that is needed,’” is “a mistake”73 because,
given a wave function, ψ different ways of defining the matter density

and flashes are possible, even if only one is representing physical objects.
To determine which one it is, we need more than just the wave function.74
But this worry strikes me as an epistemological one that need not concern us
because our metaphysically possible theory can be epistemologically dubious
in that it pushes facts about quantum systems and their evolutions out past the
domain of what can be known by cognizers like us situated at that world. On
GRW with flashes, the wave function does indeed define the flashes (ibid., 70),
although what definition it provides at w can be inserted by hand. We are fix-
ing the contours of w by stipulation (as Kripke 1980 taught us). What matters
is whether the partially interpreted theory in question is metaphysically pos-
sible. It does not matter if it cannot be known at w by cognizers like us situated
within it. Roderich Tumulka’s objection is a little more serious (relative to my
purposes in this context). He states, “[t]he more fundamental problem . . . is
that while the wave function may govern the behavior of matter, it is not mat-
ter; instead, matter corresponds to variables in space and time.”75 It seems that
Tumulka would like to differentiate flashes of the wave function (quantum
states) from the wave function itself. I think this is right. I would therefore like
to add to the partial interpretation of GRWF at w, unsure if Tumulka would
agree, that the flashes are events involving the wave function, and “a piece of
matter” (as Bell summarized the view) “is a galaxy of such events.”76
What does all of this have to do with Kim’s view of events? It has been
emphasized by several expositors of this distinctive GRW outlook (see spe-
cifically Kent 1989, 1841, who is summarizing Ghirardi et al. 1988) that for
a system exhibiting particles of the same type, it is possible for the sponta-
neous wave function collapses to transpire at one and the same time (i.e.,
simultaneously) in a way governed by the following equation,77
 1
( )
2
(Eq. 2): ψ ( x; s ; t ) → C ∑ π exp  − 2 ∑ i =1 xi − x ′π (i )  ψ ( x; s ; t ) , (Kent
N
 2a 
1989, eq. (5)), where N is the number of particles, C is a constant, x is
the location of the center of wave function collapse, and t is the time of
wave function collapse.
Causal Relata 231
The incorporation of collective collapses of this kind suggests that the
approach implies a built-in resistance to a Lorentz-invariant relativistic
extension (Tumulka, Spontaneous Wave Function 2006, 1906). That ought
to trouble the theoretician looking for a realistic quantum field theory, but
I am metaphysical space traveling in a way that is informed by physical
theorizing.78 So, I am not bothered. Consider then two flashes governed by
Eq. 2, and therefore occurring at the same time (but on different centers).
Both involve the same concrete physical object, viz., the wave function. Both
involve the same time, and both involve the same property viz., being a
quantum state.79 However, Kim’s analysis will demand that we treat these
events as one and the same despite the fact that they occur at different cen-
ters, or spatial locations.
If we add spatial locations to Kim’s view, we will no longer be able to
make sense of Datum 6. Our cosmos does not achieve its separation from
the mother de Sitter universe of the Carroll–Chen model at some spatial
location. Kim’s view seems problematic.
Section 3.7: Are Substances and Agents Causal Relata?

Some have argued that some concrete particulars that are agents and/
or substances (by themselves) in which properties inhere stand in causal
relations.80 The view seems initially plausible. Brandon threw a brick that
caused the window to break. It seems right to maintain that Brandon was
the indirect cause of the broken window.81 When Brandon decides (mental
event) to raise his hand in physics class, it seems right to say that it is Bran-
don the agent or individual or substance who causally produces the requisite
decision, not some event involving Brandon.
Unfortunately, the view is constitutionally unable to explain why effects
occur at the indices they do (Broad 1952, 215; Aune 1977, 5–6). If a sub-
stance or individual brings about causal effects, then why don’t all of the
effects that causal substance is responsible for occur exactly when that sub-
stance comes into existence? There should exist some change or some way
the substance is at an index that explains why the substance involved brings
about the relevant effect when it occurs. But if that’s right, then fundamen-
tal causal relata are more event or state of affairs-like than individual or
substance-like in cases in which substances are thought to produce causes.
So, substances alone do not produce causes.
Timothy O’Connor’s account of free agential action suggests that an
agent being in possession of certain reasons at a certain time structures “the
agent-causal capacity, in the sense of including or altering, in the agent, an
objective propensity or likelihood . . . to cause an appropriately matching
decision to act.”82 Thus, what explains why the agent freely acts when they
do is given by the agent being in possession of certain reasons at the appro-
priate times.
In response, note first that O’Connor’s view does not save the idea that
substances that are not agents are causal relata from the preceding objection.
232 Causal Relata
Only agents can have their causal activity structured by reason possession.
Second, I ask the reader to examine the case in which there is a first inten-
tional free action of an agent α, viz., ϕ, performed by α at time t in the way
O’Connor has in mind. In that case, α should be in possession of reasons
{R1, . . . , Rn} at t that structure the capacity to intentionally and voluntarily
perform φ freely. But now, ex hypothesi, α acquiring {R1, . . . , Rn} and/or
reasoning or thinking through them is an event or set of actions that are
not brought about by the agent in the intentional/voluntary and free way φ
was produced. Indeed, it appears to be a consequence of O’Connor’s view
that the first free, intentional/voluntary action(s) of α cannot be a decision
to acquire her only reasons, or the mental activity of thinking through her
only reasons, because any such activity will require prior structuring causes
that involve reasons. This strikes me as an odd consequence especially because
reasons may come in the form of the contents of beliefs, and belief forma-
tion is activity or behavior proper. Third, O’Connor’s theory (especially in
O’Connor 2008) seems to presuppose some brand of the propensity interpre-
tation of probability.83 That view has been heavily criticized (see Gillies 2016,
414). For example, Alan Hájek has, I think rightly, complained that such
accounts “do not provide an admissible interpretation of the (full) probability
calculus, for the same reasons that relative frequencies do not.”84 By the same
token, appealing to propensities to structure agential causes for free action
entails that the type of triggering free agent causation involved is indetermin-
istic (i.e., there are alternative possible courses of action with non-zero prob-
abilities). But as the preface of the current work made clear, I am interested
in the metaphysics of deterministic causation, and so also and only the relata
of deterministic causation (although one should q.v., n. 67 on one’s general
theory of causal relata). O’Connor’s way out of the timing objection forces us
to invoke indeterministic causation about which I will say no more.
Section 4: A New Account of Events That Are

Causal Relata
A proper account of the relata of causation must make room for the fact
that causation is a multigrade relation, and the following (already intro-
duced) pieces of data,
(Datum 4): There are coincident events.

(Datum 5): It is metaphysically possible that there exists but one of Max
Black’s two pure iron spheres that persists over time in a Leibnizian
space-time without undergoing any change of position, spin, internal
vibrations, etc. (i.e., the sphere just sits at a spatial location and per-
sists over time there).
(Datum 5.5): In the metaphysically possible case reported on in Datum
5, the temporal persistence of Max Black’s sphere involves causation
or causal continuity such that the relevant fact “about persistence” is
“grounded in facts about causation.”85
Causal Relata 233
(Datum 6): It is metaphysically possible that a relatum or some relata of
causation x or xx causally produced our cosmos.
(Datum 7): Complex arrangements of events are causally efficacious.
(Datum 8): At least some complex arrangements of events are them-
selves causally efficacious (complex) events.
Here then is my philosophical analysis of events:
(Events as States of Substances at Indices (ESSI)): Necessarily, for any x,

x is an event, just in case, x is the obtaining state of a concrete particular
substance86 or arrangement (without redundancy or identity) of two or
more concrete particular substances contingently exemplifying a joint-
carving universal or contingently standing in a joint-carving relation
(themselves joint-carving universals) at an ontological index (specifi-
cally spatial locations in a 3D space at a world, or a space-time location
in a 4D space-time at a world, or a world index solely)87 (and there can
be causally efficacious complex events which are themselves arrange-
ments of events as specified above).
The type of contingency employed by ESSI is metaphysical contingency.

Joint-carving universals are the universals expressed by the non-grue-like
predicates mentioned in D10 of chapter 1 and are often enough part of the
full interpretations of our best physical theories.88
I do not maintain that ESSI captures the deep metaphysical nature of all
causal relata. My philosophical analysis of causal relata is more general (see
sect. 5). Rather, I affirm that events have the metaphysical natures ESSI indi
cates, and that necessarily, events that stand in causal relations have the meta
physical natures specified by ESSI. The nature of an event e resides in the
obtaining contingent tie/unity holding between the constituent substance(s),
joint-carving universal(s), and ontological index or indices involved in it.
How does ESSI fare with respect to Datum 4 through Datum 8? First,
ESSI allows for coincident events, occurring at the same space-time location.
Davidson’s sphere can both spin and heat up because the events in question
are distinct on ESSI. Both involve distinct or different joint-carving univer-
sals despite their involving the same substance and the same space-time loca-
tion and world. The metaphysical possibilities of Datum 5 and 5.5 are not
precluded by ESSI. The iron sphere will at a time t be located on a surface
of absolute simultaneity in the Leibnizian space-time that is distinct from
the temporally subsequent surface at (for example) a later time t + 1. Thus,
the event of the sphere’s being iron at space-time location (x, y, z, t) at w is
distinct from the event of the sphere’s being iron at space-time location (x,
y, z, t + 1) at w. The former event can therefore causally produce the latter,
thereby helping ensure the persistence of the iron sphere over time without
violating the irreflexivity of causation established in chapter 6: sect. 2.
Likewise, ESSI does not preclude the possibility of a process causally pro-
ducing the cosmos. The event of the cosmos achieving separation from a
234 Causal Relata
mother de Sitter space-time can be indexed to a metaphysically possible
world if not a spatial location at a time, or space-time location.
And lastly, I note that ESSI allows for complex arrangements of events to
be causally efficacious. Complex events are written into the account via the
last conjunct in parentheses.
Section 5: Sometimes Forces Are Causes

There are four fundamental types of interactions between fundamental enti-
ties in our best physical theories, viz., the strong, weak, electromagnetic,
and gravitational interactions (one could include the electro-weak interac-
tions at the appropriate scale and in the appropriate limits). The first three
are forces, whereas the latter (as in chapter 3 and chapter 8: sect. 2) reduces
to space-time geometry or curvature. No one (so far as I’m aware) in the
physics literature denies that all four types of physical phenomena are inter-
active phenomena. As was discussed in chapter 3: sect. 3.1, the strong force
or causal interaction involves the exchange of vector gluonic fields or glu-
ons. These exchanges, together with other interactions involving fermions
(restricted to quarks in QCD) and gluons or gluonic fields, are sometimes
violent.
Let us revisit part of the standard quantum field theoretic model of parti-
cle physics, specifically QCD,89 according to which there exist down quarks
1
(d-quarks). These are spin first-generation particles that are extremely
2
1
light and endowed with − electric charge. The ontology of the standard
3 1
model also includes up quarks (u-quarks). These are also spin first gen-
2
eration particles, although they are lighter than d-quarks and endowed with
2
+ electric charge. Quarks of both types can have colors indicative of color
3
charges or quantum number (blue, red, and green).90 In QCD, a specific
d-quark with blue charge (db) can scatter a u-quark with red color (ur),
and as a result of the scattering, db will transmute into a down red quark
(dr), and ur will transmute into an up blue quark (ub) (i.e., ur + db → ub +
dr).91 This interaction can be straightforwardly understood in terms of cau-
sation between events as specified by the ESSI account. The relevant quarks
(understood as concrete particulars that are Aristotelian substances) have
properties (colors or color charges) at space-time locations. But the story is
only partially complete. There is an interaction between the quarks that pro-
duces their transmutations. Chapter 3: sects. 2.3 and 3.1 argued that these
interactions are causal. The interpretation should not change when shift-
ing one’s view to the weak and electromagnetic interactions. Both are com-
monly causally interpreted.92 To see how the interaction between our two
quarks fits with the ESSI, we should first realize that the interaction in ques-
tion just is the strong force itself, and that a chief means whereby physicists
Causal Relata 235
Gluonic Exchange
(le to right)
Figure 7.1 Quark Interaction

Note: Usually there are dots on both ends of the gluonic propagator representation in the dia-
gram (the vertices). These represent interactions. I left out the dots because of artistic failure
on my part. Sorry!
model quark interactions themselves in QCD (as with electron and photon
interactions in QED) is through Feynman diagrams like the one in Figure 7.1.
Here we have a standard illustration of strong quark interactions. The
instate is given by the lower half of the diagram. The scattering is mediated
by the fundamental quanta of QCD, the gluon (again see chapter 3: sect. 3.1
for more on gluons). The interactions are represented at the vertex points
(where the arrowed lines meet the squiggly line). Importantly, the vertex
points are located at space-time locations. That is to say, they occur or
transpire at space-time positions that are Lorentz invariant, such that with
respect to any arbitrary Lorentz frame, in accordance with the time coordi-
nate in that frame, every pair of vertex points (but measure zero) enjoys a
time-separation of a finite variety.93
Feynman diagrams were invented by Richard P. Feynman (1918–1988) in
the late 1940s and subsequently refined and set atop first-principles by Free-
man J. Dyson to help with calculations in QED.94 At least one important sense
in which many quantum field theorists used these diagrams as calculational
helps was as bookkeeping devices that encode the mathematical formalism
needed to represent the interactions illustrated by the diagrams.95 That is to say,
amidst the early diagrams-as-calculation-helps era (i.e., before their applica-
tion to nuclear physics96) (quoting Kaiser), “Feynman introduced his diagrams
to keep track of all the different ways that electrons and photons (. . . light
quantum) could interact.”97 Subsequent use of these diagrams, specifically in
the context of strong interactions, involved their utilization for the purposes of
236 Causal Relata
representing, not just physical possibilities, but actual physical interactive pro-
cesses. As David Kaiser noted, “[f]aced with the influx of new and unexpected
particles and interactions, some theoretical physicists began to use Feynman
diagrams as pictures of physical processes.”98 Amidst the S-matrix theoretic
approach to quantum physics led by Lev Landau and Geoffrey Chew, Feyn-
man diagrams were used to model particle interactions (Kaiser 1999, 343).
No matter the era or mode of use, everyone would agree that Feynman
diagrams are so compellingly representative of that which they model that
(ordinarily in contexts like QED, QCD, and S-matrix theory) each part of
a Feynman diagram corresponds to a specific chunk of the mathematical
designative formulation used to describe the (perhaps merely possible) inter-
active processes.99 At the same time, in the hands of Feynman, the diagrams
were used as intuitive pictures of physical processes that enjoyed a type of
theoretical priority over calculation much like the phenomenon-first natural
philosophy discussed in chapter 1: sect. 4.5.5.5 would recommend. Kaiser
elaborated on the case of Feynman as follows:
From the very beginning, Feynman and Dyson held different ideas
about how the diagrams should be drawn, interpreted, and used. For
Feynman, doodling simple spacetime pictures preceded any attempts
to derive or justify his new calculational scheme. . . . To Feynman, his
new diagrams provided pictures of actual physical processes, and hence
added an intuitive dimension beyond furnishing a simple pneumonic
calculational device.100
The approach is not idiosyncratic. Some leading physicists take up an explic-

itly phenomenon-first attitude about the use and role of Feynman diagrams.
Consider these remarks from Nobel laureates Gerard ‘t Hooft and Martinus
J. G. Veltman:
Few physicists object nowadays to the idea that [Feynman] diagrams

contain more truth than the underlying formalism. . . . The situation
must be reversed: diagrams form the basis from which everything must
be derived. They define the operational rules, and tell us when to worry
about Schwinger terms, subtractions, and whatever other mythological
objects need to be introduced. . . . Using diagrams as a starting point
seems therefore to be a capitulation in the struggle to go beyond pertur-
bation theory. It is unthinkable to accept as a final goal a perturbation
theory, and it is not our purpose to forward such a notion.101
Let me now ask an important question. What constitutes a causal model?

According to Judea Pearl (2009, 203), causal models are triples <U,V,F>,
where ‘F’ stands for a collection of structural equations that are functions
of the mathematical values associated with the two types of variables in
sets U and V. Members of the set of variables U are exogenous variables
Causal Relata 237
that have mathematical values fixed by factors external to the system mod-
eled. Members that are variables in set V are endogenous variables whose
values are fixed by internal features of the modeled system.102 These models
have “associated . . . directed graphs . . . in which each node corresponds
to a variable” (ibid., 203). Although Sloman’s characterization of a causal
model is somewhat different, he too builds causal models out of three parts,
although his parts only include the target of the model (i.e., the causal sys-
tem itself), a graph that “[d]epicts causal relations amongst events,” and a
probability distribution used to give the probabilities that the events in the
model occur with one another (Sloman 2005, 38).
Notice that on both depictions of what a causal model amounts to, Feynman
diagrams with associated formalism come out as causal models.103 Feynman
diagrams and underlying math include parts that explicitly represent interac-
tions (Lancaster and Blundell 2014, 181, note how “[v]ertices where lines join
together . . . represent interactions”).104 They (the underlying formalism) are
given by graphs whose targets are causal interactions, and they include exoge-
nous and endogenous variables related by mathematical functions, which in the
Feynman diagram cases can be read directly off of the diagrams or graphs.105
If we treat Feynman diagrams much like Feynman did, that is to say, if
we treat them as physical pictures of physical processes, and we regard
them as causal models, we can explain well why they are so useful to the
physicist in making predictions about how the world works. Recall my
use of the empirical data from psychology as reported in the work of Ste-
ven Sloman in chapter 1: sect. 4.5.5.5. The quotation used there is worth
repeating:
[M]any studies have shown that students use general knowledge about
the way the world works, sometimes causal knowledge, to solve word
problems in physics and mathematics. Instead of thinking about prob-
lems in a purely abstract way, manipulating symbols until arriving at
the correct answer, people solve problems guided by an understanding
of the situation. If the situation is causal, then we use a causal model.106
I’ve argued that in the context of the fundamental quantum interactions, the
situation is causal, and so it is no surprise that one of the most beneficial
ways of solving problems is delivered via causal models that are Feynman
diagrams. That the diagrams are beneficial to student learning and prob-
lem solving is well-evidenced. Indeed, early textbook introductions to the
diagrams and their use in physics stressed or (quoting Mattuck’s textbook)
“concentrated exclusively on giving the reader a feeling for the diagrams
themselves [and] their physical significance.”107 According to Kaiser, schol-
ars like “Mattuck upheld” their promise:
[T]hroughout the entire book, discussion was formulated directly

in terms of diagrams, with the corresponding integrals simply left
238 Causal Relata
unevaluated. Mattuck believed the line drawings were the key both to
conceptual clarity and pedagogical effectiveness.108
If you refocus your attention on the causal model that is the Feynman
diagram of Figure 7.1, you’ll find a representation of the interaction respon-
sible for the scattering and transmutations of our quarks in the causal pro-
cess given by the middle portion of the diagram (i.e., the gluonic exchange).
That process can be understood as one involving an emission of a gluon
from db at a space-time point. As already noted, the emission is more faith-
fully represented in more complete Feynman diagrams with a vertex point/
dot. The diagram depicts a series of ESSI events indicative of the persistence
of the gluon or the undulation of a gluonic field resulting in an absorption
interaction by the incoming ur. That absorption causes a recoil and change
of motion plus transmutation of ur into ub. Thus, the strong interaction
here is a causal process between an emission (an irreducible causally potent
interaction), a process of persistence involving ESSI-events over time, and
an absorption (an irreducible causally potent interaction). Notice that the
interactive process is in some important ways fundamental and irreducible
in that I am interpreting emissions and absorptions in fundamental physics
as physical quantities that are causal interactive happenings. These interac-
tive happenings generate the beginnings of quantity ESSI-like events involv-
ing quantity changes and physical evolutions including motions.
I explain precisely how gravitation in Newtonian mechanics and general
relativity are causal and in what way they might involve events of an ESSI
variety in chapter 3: sect. 2.1, and chapter 8. The important lesson to draw
from my current discussion is this: Fundamental forces are fundamental
interactive processes, processes involving irreducible causes or effects that
are themselves interactions (emissions/absorptions, or in the case of New-
tonian gravitation, at a purely Newtonian world, the gravitational force
being a certain way itself, etc.). I call these irreducible interactions funda-
mental events. These fundamental interactive processes are causally potent,
although they are sometimes arrangements of events (in the ESSI and fun-
damental sense). Moreover, a complex fundamental interactive process can
itself be labeled a complex event. Indeed, that fact recovers the common
way physicists talk (e.g., when they call a quark scattering process, a quark
scattering event) and illustrates how ESSI recovers data points 7 and 8.
Section 6: Against Negative Causation

I maintain that causation is a relation between events in the ESSI sense, and
sometimes between forces, understood as processes involving fundamen-
tal events (as in our best quantum theory of weak interactions, QED, and
QCD), and events. Some have argued that omissions or absences can stand
in causal relations. When causation involves absences, there exists negative
causation. When an absence produces another absence, there is thought to
Causal Relata 239
exist an instance of prevention by omission. When an absence produces
an event, there exists causation by omission. When an absence is the effect
caused by a positive event, there exists causation by prevention.109 Notice
that according to this standard way of characterizing matters, omissions
are absences. This is in keeping with the literature on omissions that often
stipulates (as Sara Bernstein did) from “the outset that ‘omission’ refers to
any action or event that doesn’t occur.”110
My attitude about supposed cases of negative causation is the same atti-
tude that David M. Armstrong held about them. He wrote that,
when we reflect a little on such cases [of supposed negative causation],

we are very ready to admit that the actual causal processes involved
proceed solely in virtue of the (positive) properties of the situation. To
say that the lack of water caused his death reflects not a metaphysic of
the causal efficacy of absences but merely ignorance. Certain (positive)
processes were going on in his body, processes which, in the absence of
water, resulted in a physiological condition in virtue of which the predi-
cate ‘dead’ applied to his body.111
I justify my Armstrongian dismissal of negative causation through the sup-

port of two theses,
(Thesis #1): There are no plausible metaphysical theories of omissions

understood as absences that are causal relata.
(Thesis #2): Virtually all supposed cases of negative causation can be
faithfully/accurately redescribed without omissions/absences.
I will support (Thesis #1) in sect. 6.1, and I will support (Thesis #2) in
sect. 6.2.
Section 6.1: Against Metaphysical Theories of Omissions

In Schaffer (Causation by Disconnection 2000), Schaffer takes omissions
to be absences and provides no substantive analysis or theory of what
absences are, although he says there is a choice way of describing them via
the denial that some event occurs. In Schaffer’s extended case for negative
causation (Schaffer 2004), he does much the same. He proceeds to make his
case for negative causation all the while equating omissions with absences,
but never placing them at the center of a searching metaphysical theory or
philosophical analysis. He flags only how it is highly intuitive that there
are instances of negative causation involving absences and that many com-
monly speak that way in special scientific and even physical practice.112 In
later discussion of Michael S. Moore’s (2009) work (see Schaffer, Discon-
nection 2012, where Moore also rejects omissions as causal relata), Schaf-
fer entertains the idea that perhaps the way to save negative causation is
240 Causal Relata
by identifying causal relata with abstract factsP.113 Omissions or absences
could then be certain true negative existential propositions. This suggestion
is strained. In fact, I can find no actual advocates of the view that abstract
true propositions literally stand in obtaining causal relations. The citations
of J. Bennett and Mellor along these lines are, in my opinion, based on
incorrect readings (q.v., n. 113). Any theorist who believes that abstract
propositions can stand in causal relations will face the same predicament
Lewis faced. They will need to provide a sufficient explanation of how it
can be that abstract objects stand in causal relations when the very abstract/
concrete divide is usually cut by appeal to the fact that concreta are causally
potent and abstracta are not.114
Randolph Clarke is a pluralist about omissions. He identifies some of them
(the intentional ones) with acts of refraining, like when one gives one’s keys
to a friend to refrain from using them to drive (call this view Clarke-1).115
He identifies others with certain positive actions (Clarke-2), and still oth-
ers (Clarke-3) with “an absence of action of a certain type by an agent at
some time.”116 Clarke-1 and Clarke-2 are incompatible with the idea that
omissions are absences involving actions or events that fail to occur. If my
act of giving my keys to my friend just is the omission, then it is an action
that occurs. But then in what sense is it an absence? The action happens or
transpires, there’s no absence or failure of occurrence. Indeed, my action
can be readily recovered under the ESSI approach to events. I exemplify a
number of properties over an interval of time thereby producing an event
series or process that is my handing my keys to my friend (assuming agency
can be captured by event-event causation approaches). We seem to have left
omissions behind altogether if we follow Clarke-1 and Clarke-2. Clarke-3
appears to be the most promising view of omissions understood as absences.
About it, Clarke wrote,
I favor the view that absences aren’t beings of any sort: there’s no entity
that is an absence. If this is correct, then in many cases in which some-
one omits to act, there’s nothing in the world that is the omission.117
Notice what this view entails for instances of negative causation. If one
maintains that absences understood in terms of Clarke-3 cause things, then
it will come out false that necessarily, causation is an obtaining relation. As
I noted in chapter 2: sect. 5, relations require relata. But if omissions are
absences, and absences “aren’t beings of any sort,” omissions or absences
do not exist. If they do not exist, they aren’t any thing or stuff that can be
related to anything else. But I argued in chapter 2: sect. 5 that necessarily, all
instances of causation are instances of an obtaining causal relation. Thus, if
we take Clarke-3 seriously, omissions understood as absences cannot stand
in causal relations, and so as a result we will be left without instances of
negative causation on Clarke-3.
There’s the view that omissions are instantiations of negative properties
by objects or individuals. This view too departs from the idea that omissions
Causal Relata 241
are absences of some kind. Moreover, if an individual has a negative prop-
erty at a space-time location, and that negative property is a joint-carving
universal (I doubt there are such universals, see Armstrong, Theory of Uni-
versals 1978, 23–29, and Mellor 1999, 196–199, for objections), then we
have an event as pictured by the ESSI account. Again, we have left omis-
sions/absences behind.118
Sara Bernstein has recently argued that omissions are events that are such
that had they occurred, they “would have caused another event to occur.”119
Omissions are therefore three-part entities involving (i) a close (to the actual
world) unactualized possible event that stands in a (ii) counterpart relation
to (iii) an actual event.120 Ingredients (i)–(ii) demand that we regard events
as worldbound entities or existents. I see no reason for adopting such a
position, and it is precluded by the actualism that has been invoked as an
assumption of the current project (see chapter 5). Ingredient (ii) requires a
counterpart theoretic semantics for de re modality. That semantics has been
heavily criticized by Fara and Williamson (2005). Third, Bernstein (2014)
does not appear to be in the business of providing an account of omissions
properly understood as causal relata. Rather, her account, even if successful,
only provides us with a metaphysics of omissions that render them causally
salient for causal discourse (see ibid., 12). Although Bernstein does empha-
size that the sense of saliency here helps with providing a correct theory of
how omissions figure into true causal claims and explanations, she extends
her account of omissions to include a theory of them as causally salient in a
metaphysically distinctive sense, a sense that entails that,
causally relevant omissions occupy a place between total non-being and

full-fledged being. Taking omissions to be possibilities strikes the right
ontological balance between nonexistence and existence.121
The metametaphysicC and metaphysicalC system I articulated and defended

in chapter 1: sect. 4.2 and n. 97 followed P. van Inwagen (2009, 492–499)
and Sider (2009, 397–402) by denying that there’s space between being and
non-being. In other words, being, like existence, is univocal. There is then
an abundance of reasons for the natural philosopher who follows my lead
to abandon Bernstein’s theory of omissions, and even if we did appropriate
it, it would not give us omissions fit to be causal relata for instances of nega-
tive causation.
Section 6.2: Supposed Cases of Negative Causation in the Sciences

Recall the theses I’m interested in defending:
(Thesis #1): There are no plausible metaphysical theories of omissions

understood as absences that are causal relata.
(Thesis #2): Virtually all supposed cases of negative causation can be
faithfully/accurately redescribed without omissions/absences.
242 Causal Relata
I believe my case for (Thesis #1) provides one with some reason to seek
redescriptions of supposed cases of negative causation. We must be care-
ful, however. If we do not know even remotely what an absence is, then it
will be difficult to discern that which requires redescription. I will therefore
take my clue as to what precisely is being appealed to as an absence in the
supposed cases of negative causation below by discerning authorial intent.
What, in particular, Jonathan Schaffer (2004) intends his term ‘absence’ to
refer to will be my target of redescription in each of the two cases discussed
below. I limit my revisionary exercise to two cases for two reasons. First,
I have limited time and space. And second, I believe my efforts will suggest
that there’s a cogent inductive case against supposed instances of negative
causation.
We begin with a visit to solid state physics for a supposed example of
negative causation in the case of electron–hole pair production. In the
relevant physics, there are systems built from bands (or we could focus
on objects like crystals). These can be populated by electrons moving on
surfaces of constant energy. When an electron leaves a vacancy, it leaves
a hole. Schaffer interprets these holes as literal absences or omissions of
electrons.122 Because these holes can affect voltages and the like, Schaffer
believes he’s found a case of negative causation in physics. Unfortunately,
the interpretation fails if we forsake a particle ontology of quantum field
theory and think of the holes and electrons as excitations of quantum fields
(ways the quantum fields are at certain locations), real concrete particular
substances.123 But focusing for now on a QFT with a fundamental particle
ontology, I note that “holes” appear to me to be particles. In fact, they
are sometimes called quasi-particles. Here’s my case for that claim. First,
when our band system is affected by a magnetic field, the electrons and
holes enjoy orbits opposite in direction. This suggests holes and electrons
have opposite charges. Indeed, there’s conclusive evidence that holes must
have positive charges. Holes can have velocities and accelerations. In fact,
under sway of an electric field, they will accelerate in the same direction of
that field. Aside from charges, holes have other physical properties, such
as effective mass, energy, probability amplitudes related to their motions,
and a force law governs their dynamical behaviors (see Feynman, Leighton,
and Sands 2010; Kittel 1966, 272–280).124 In light of the above facts, it
seems best to interpret holes not as absences of electrons, but as quasi-
particles that are created by the dynamical motions of electrons. But again,
if we insist on a field ontology for QFT, we aren’t dealing with absences
here at all, but states of fields.
Let us now turn to a case of biomedical science. Schaffer maintains that
“[w]hat causes scurvy is an absence of vitamin C.”125 I counter that expres-
sions like these are but gloss. A more accurate description of the situation
need not (and does not) invoke absences as causally efficacious at all. To
see this, one should explore with me some of the history and science of
the disease that is scurvy. First, I note how Homo sapiens are part of a
Causal Relata 243
unique group of mammals unable to generate the L-ascorbic acid that is
vitamin C. By far, the majority of mammals (and a great many plants) can
produce that acid by their lonesome (Burns 1959; Stone 1965). There’s a
four-step process that takes place in the liver of vertebrate mammals who
have the ability to create their own L-ascorbic acid. “[E]ach of the four”
steps is “controlled by a different enzyme” (Stone 1965, 83). The relevant
process takes in glucose and spits out L-ascorbic acid. Homo sapiens do not
have the gene responsible for ensuring the presence of the enzyme needed
to govern the final step of glucose to L-ascorbic acid production in the liver
(ibid., 84). The general evolutionary etiology of our genetic makeup with
respect to this matter is one involving conditional lethal mutation (as Stone
notes in ibid., cf. Gluecksohn-Waelsch 1963, 1270). Listen now with me to
Gluecksohn-Waelsch,
This [conditional lethal mutation] is the case, for example, when a

mutation results in inability of the organism to synthesize an essential
amino acid. If the organism is grown on a basic minimal medium, such
a mutation will have a lethal effect; if however, the missing amino acid
is supplied from the outside—that is, if a nongenetic modification of the
environment is produced—the same mutant becomes ‘viable.’126
The Cavia porcellus or guinea pig suffered an evolutionary fate similar to

that of Homo sapiens. They too are organisms unable to produce L-ascorbic
acid. Following Carpenter’s discussion, I note how during the late 1800s
and early 1900s, scientists were discouraged from using rats as specimens
because they were judged to be dangerous disease carriers. However, guinea
pigs were pets in Europe, and so some scientists found guinea pigs to be of
important experimental use in their research instead. This historical contin-
gency helped with scurvy research given what I have already asserted about
their evolutionary heritage.127 In the follow-up (part II) of what Kenneth
Carpenter (1986, 173) has described as “the most important single paper
in the whole history of” the subject of scurvy, viz., Axel Holst and Theodor
Frölich’s “Experimental Studies Relating to Ship-Beri-Beri and Scurvy”
(Introduction, 1907),”128 Holst and Frölich used various trials involving
guinea pigs to show (in their own words),
that scurvy cannot be caused in guinea-pigs either by simple starvation

or by diets of any kind; on the contrary, the disease originates in these
animals as well as in man as a result only of certain special diet.129
There’s no question that Holst and Frölich’s papers were quite important,
but so too was Jack Cecil Drummond’s, “Note on the Rôle of the Anti-
scorbutic Factor in Nutrition” (Drummond 1919). Drummond concluded
that the necessary dietary needs of “higher animals” should involve water-
soluble B, fat-soluble A, “a satisfactorily balanced ration of protein, fats,
244 Causal Relata
carbohydrate and mineral salts,” and “[w]ater-soluble C, or antiscorbutic
factor.”130
Given this discussion, my hypothesis is that the cause of scurvy in Homo
sapiens is the evolutionary heritage of the Homo sapiens that generated its
genetic makeup, together with that genetic makeup, plus a special diet (one
about which it is true that there does not exist within it adequate vitamin C)
the Homo sapiens actively consumes. These positive factors/happenings/states
of affairs/events (perhaps together with certain others left suppressed for
brevity) cause the Homo sapiens to enter a process of biological malfunction,
itself composed of several differing biological malfunctioning sub-processes
that cause the Homo sapiens to enter the process of dying. Our evolutionary
heritage has so endowed us that special diets can kill us. But strictly and most
accurately speaking, our deaths are not caused by any absence. Our deaths
are caused by a rather long and ancient history of evolution, together with
how we are built, and perhaps our dietary choices together with the resulting
biological malfunction. That’s what kills us. The goal is to avoid biological
malfunction by actively performing certain acts of food consumption. The
goal is to function properly, given how we have been endowed by evolution.
We can help cause our bodies to properly function by enjoying the type of
diets (at least) partially described by Drummond.
There are other supposed cases of negative causation to which Schaffer
and others appeal. I remain unimpressed by them. I believe that in every
instance a redescription of the relevant cases can be provided, which does
not entail that absences are causes. My hope is that in my evaluation of a
physics and biomedical case, I have advanced some inductive evidence for
(Thesis #2).
Notes
1. R. Chisholm (1990, 422), who borrows wording from Davidson (Events 2001,
163).
2. R. Chisholm (1990, 422; emphasis mine).
3. D.K. Lewis (Events 1986, 241).
4. In chapter 6: sect. 3, I defended the transitivity of causation. In my defense,
I dismissed some difficult cases on account of the fact that they ignored the
rich complex nature of the causal structure of the actual and relevant possible
worlds. That richness and complexity does nothing but substantiate the doctrine
that causation is a multigrade relation.
The examples I use in the main text often rest upon transitivity. Although the
descriptions I provide are often piecemeal, I believe they can all be enriched in
such a way that transitivity will not fail given such enrichment.
5. “The first bomb design considered for both uranium and plutonium was the gun
barrel assembly, in which one sub-critical piece of uranium or plutonium is fired
into a second, creating a critical mass and a nuclear explosion” (Hayes 2017, 5).
I’m assuming that it is not the case that there is a pre-initiation process involved
in the nuclear process. Thus, decay processes do not preempt the reaction.
6. See Hayes (2017); Peierls (1997).
7. See Hayes (2017); Peierls (1997).
Causal Relata 245
8. One should accept the data listed in the main text even if one is a reductionist
about causation. A great many reductive theories of causation will deliver the
verdict that the preceding cases are causal. Again, causal reductionists are not
causal eliminativists. I have already shown (in chapter 2: sect. 2) that causal
eliminativism is false. Thus, causal eliminativism cannot be employed in this
context to undermine my point.
9. Again, even if the imagined full cause is much more complicated, involving
many more events, that would only help my case for the multigrade nature of
causation.
10. Harris (2009, 142–148). Interestingly, the relevant process (the phenomenon)
is not only accurately described in ibid. with causal terms like ‘activation’ and
‘deactivation,’ but also ‘interaction’ (ibid., 146).
11. I’m indebted to the following discussions and introductions to the literature on
causal relata throughout my discussion: Ehring (2009); Simons (2003). These
sources provided a valuable introduction to the causal relata literature.
12. For which see Minkowski (1952) for a taste of the groundbreaking geometric
approach. See Thorne and Blandford (2017, 37–89) for a definitive statement of
much of the geometric designative formulation and partial interpretation of SR
that I will assume.
13. Mermin (1989, 25; emphasis mine).
14. Although that flavor may not be rich enough to support verificationist theories
of meaning, or logical empiricist approaches to general relativity more generally
(see on this matter, Giovanelli 2013, 116, and the literature cited therein).
15. Einstein (1952, 117; emphasis mine). There is a literature on the argumenta-
tion here. See Bergmann (1961); J. D. Norton (1984); Rickles (Hole Problem,
2008); Stachel (2002, 301–337). This last source is a reprint from a paper that
appeared in 1980. For interesting historical background related to Kretschmann’s
point coincidence ideas, see Giovanelli (2013).
16. Indeed, some would call the coincidence of two events, itself an event (a so-
called Komar event as in Earman’s discussion; see Earman 2002, 13).
17. This notion’s definition requires an appeal to frames of reference.
18. From Black (1952, 156).
19. I will not commit to any particular view about what would metaphysically
explain the relevant persistence fact, nor do I commit to any specific ontology
of that persistence. One should not identify the explanatory account with the
ontological account of persistence (Wasserman 2016, 244–245).
20. Although if there were motions in this space-time they would be strictly relative
motions (Earman 1989, 31).
21. For the details on a Leibnizian space-time, see Earman (1989, 27–31); Ehlers
(1973, 73–79); and Stein (1977).
22. Wasserman (2016, 249), subsequently noting that this causal theory is consistent
with perdurantist and endurantist views of persistence (suitably understood).
Wasserman does note how mereological universalism (the thesis that composi-
tion is unrestricted) poses an initial problem for the causal view. But I reject
mereological universalism on independent grounds.
23. S.M. Carroll (2008, 8).
24. S.M. Carroll (2010, 357–358).
25. In (Weaver 2017), I argue that the model as stated is metaphysically impossible
because of the proposed properties of the mother universe. However, one can
adjust these properties so as to avoid my objection.
26. Thorne and Blandford (2017, 101–110). These authors start off describing the
classical relativistic case and then expand their discussion to include quantum
statistical mechanics, appropriating some of the same terminology and formal-
ism for the quantum case.
246 Causal Relata
27. Thorne and Blandford (2017, 110).
28. For a defense of a causal interpretation of classical electrodynamics, see M.
Frisch (2005). Frisch points out numerous times that the standard way of inter-
preting several of Maxwell’s equations is in causal terms.
29. Quine (1960, 171).
30. Lemmon (1967, 99; emphasis in the original).
31. Ibid., 98–99.
32. I am assuming that substantivalism is the view that space-time exists as a sub-
stance or object “independently of the existence of any ordinary material objects,
where the latter phrase is taken to include even such extraordinary material
objects as rays of light, physical fields . . . etc.” (Sklar 1976, 161).
33. Davidson (Events 2001, 178–179).
34. See also Cleland (1991, 230) for a similar counter-example.
35. Brand (1977, 334) does add that locutions used to pick out events e1 and e2 be
rigid.
36. Even if you maintain that the proto-inflationary patch that issues forth out of
the background space-time due to fluctuations is not an independent space-time,
there is a relatum showing up later in the causal process that is our space-time
acquiring independence (its orphaned status) from the mother cosmos. That
relatum does not occupy a space-time region.
37. Ehring (2009, 388), summarizing Davidson’s view. See Davidson (Events 2001,
163–180).
38. The objection follows Ehring (1997, 87).
39. Cleland (1991, 232–242).
40. Ibid., 233. Cleland assumes that “determinable properties are not” properly
reducible “to determinate properties.” Ibid.
41. Ibid, 235.
42. Ibid., 238.
43. Ibid., 245.
44. Ibid., 238. I have changed Cleland’s formatting.
45. Ibid., 245.
46. He said, “I propose to identify an event with the set of spacetime regions where
it occurs” (D. K. Lewis, Plurality 1986, 84; cf. 95).
47. D. K. Lewis (Events 1986, 245).
48. Ibid., 247.
49. This objection is not new. I’ve heard it repeated in many seminars.
50. He said, “[b]eing the only even prime number cannot be the same as being—or
be what it is to be—the positive square root of 4” (Parfit 2011, 297; emphasis in
the original).
51. Shafer-Landau (2003, 91).
52. As Gideon Rosen said, “[t]o strike a theme that will recur, it is widely supposed
that sets and classes are abstract entities—even the impure sets whose urelements
are concrete objects” (Rosen 2017, sect. 3; emphasis in the original). And as
Chris Swoyer maintained, “the philosophical important features of the para-
digm examples of abstracta . . . are pretty clear. They are atemporal, non-spatial,
and acausal . . . nothing can affect them, and they are incapable of change”
(Swoyer 2008, 13–14).
53. J. Bennett (1988, 44).
54. He seems to also (see J. Bennett 1988, 117) think that property instances them-
selves are best understood as a zone possessing a property. Thus, substances are
not the sole entities that bear properties, zones do, too.
I should add here that Ehring (1997, 85) is reluctant to judge Bennett’s account
as a truly trope theoretic one because the particularized properties themselves do
Causal Relata 247
not do the causing, instances of such properties do. If, however, tropes just are
such instances, then it seems to me that the tropes do the causing. The idea that
tropes are causal relata shows up in Keith Campbell’s work. He said, “the terms
of the causal relation are always tropes” (Campbell 1990, 22; emphasis in the
original). Cleland (1991) could also be understood as a trope theory of a modi-
fied kind. Some think Bennett maintains that the facts do the causing. Matters
are complicated!
55. Ehring (2009, 407).
56. See Ehring (1997, 14, 100).
57. Talk of non-mereological parts may strike some analytic metaphysicians as
strange, but if you recall, my explication of Datum 8 appealed to (inter alia)
micro-based determination and/or realization as in Gillett (2016). Gillett calls
this type of composition, scientific composition (ibid., 65–71). It may be an
instance of grounding, or it may underwrite a reduction. I leave these further
questions open (on related themes and motifs, see the excellent collection of
essays in Aizawa and Gillett 2016).
58. Beebee (1998, 183).
59. See Goldman (1970); Martin (1969); N. Wilson (1974).
60. Kim (1966, 231). Cf. Kim (Events as Property Exemplifications 1973).
61. Kim (Nomic Subsumption 1973, 222).
62. Ibid.
63. Kim (1976, 161); cf. the discussion of the account in Simons (2003, 365, cf. 375).
64. Kim anticipated this worry by restricting the types of properties that figure in
events proper (Kim 1976, 162–163).
65. In addition to the sources cited in chapter 4, I now add Goldstein (Part 1 1998;
Part 2 1998); Kent (1989); and Tumulka (Relativistic Version of GRW 2006;
Spontaneous Wave Function 2006). There is another type of collapse theory in
the work of Daniel J. Bedingham (2011).
66. A point also made by Healey (2017, 109–110, inter alios).
67. The reader might ask, but aren’t you concerned in this project with deterministic
causation? Yes, I am. But one’s general theory of causal relata should work for
both brands of causation, deterministic and indeterministic. Thus, if Kim’s account
fails at a GRW-world, it isn’t a correct philosophical analysis of causal relata for
that reason.
68. The adroit reader will recall the fact that in chapter 1: sect. 4, I noted several
ways in which metaphysics enters physics. One might wonder then how it is
that physical theory can be used to provide evidence of the metaphysical pos-
sibility of this or that state of affairs when one needs metaphysics to interpret
designative formulations of physical theories themselves. Here’s how it might
in this special case. Outfit GRWF with the designative formulation supplied
by (Ghirardi et al. 1988), or the summary of it in (Kent 1989), then give that
designative formulation whatever partial interpretation is needed to ensure
consistency with well-established (independently motivated) metaphysical laws
together with the laws of classical logic. We should not include the dictums
of Kim’s analysis of events as laws of metaphysics because I am objecting to
that analysis, and Kim never provides us with an independent argument for his
analysis. We can now use GRW to help paint a picture of a metaphysically pos-
sible world without begging any questions about the truth or falsity of Kim’s
analysis. The idea is that GRW (with flashes, see the main text) appropriates a
conception of events that seems perfectly possible and scientifically respectable
in that it is precisely described, fitting into the mathematics cleanly as an inter-
pretation of that mathematics.
69. Ghirardi et al. (1988, 384).
248 Causal Relata
70. “[T]he wave function can be interpreted as the state of a physical system” (Ghi-
rardi et al. 1988, 385).
71. For an excellent study of the metaphysics of the wave function and the spaces
in which it may live, see Chen (2017). I will assume Chen’s view that the wave
function for GRWF at our possible world lives in three-dimensional space

(3D-fundamentalism), and not a 3N-dimensional space (N being the number of
particles at the relevant world understood as a quantum physical system).
72. Kent (1989, 1842; emphasis mine). This excerpt may suggest that we eliminate
the wave function in favor of just the state, but I think it is best understood as
the view that the wave function is the state (identity).
73. Allori (2013, 71).
74. Ibid.
75. Tumulka (Spontaneous Wave Function 2006, 1905; emphasis in the original).
76. Bell (2004, 205).
77. See also Ghirardi et al. (1988, 385).
78. I take it that those scientifically minded folks who have expressed doubts about
the very existence of metaphysical possibility (e.g., Maudlin 2007, 184–191)
would find my metaphysical space exploration more palatable. In this, at least,
I can hope.
79. If you do not like this property, then use being a local beable, or being a quantum
observable, or being a physical quantity. All of these are scientifically respectable
enough to constitute constituents of Kim’s generic events. I should add here that
Kent (1989) resists the take on collective collapses in Ghirardi et al. (1988).
80. There are some scholars who maintain that every cause must be a substance
(Nelkin 2011 (there’s a hint of it there at least at 88–89); Swinburne 1997).
Most of my criticisms of the weaker view that some substances are causes will
therefore constitute objections for this much stronger position that all causes are
substances.
81. See Byerly (1979); Clarke (1993; 1996; 2003); O’Connor (2000); R.M. Chisholm
(1966); Reid (1969); and R. Taylor (1966); cf. the discussion (without endorse-
ment) of the view in Ehring (2009, 391), and the helpful overview in O’Connor
(2013). I should add that the view discussed in Clarke (2003) suggests that when
an agent causes an action, it does so while being in the possession of certain
reasons it has to act. Clarke did not endorse the view explicated (as he notes
in Clarke forthcoming, 15, n. 1). I will not engage that view here as it has been
criticized by O’Connor and Churchill (2004).
Some may argue that libertarian accounts of free will require that agents be
able to serve as causal relata. This is not necessarily the case. There are theo-
ries of libertarianism that do without agent causation (see Kane 1996; see also
Clarke 2002, for an overview).
82. O’Connor (2013, 136). Cf. O’Connor (2008).
83. On this approach to probability, see Giere (2011) and Popper (2011). Cf. the
very helpful study and overview in Gillies (2016).
84. Hájek (2012, sect. 3.5). See ibid. for elaboration of the objection.
85. Wasserman (2016, 249).
86. As in chapter 5: note 4, I am assuming an Aristotelian view of substances, for
which see Loux (1978; 2006, 107–117). Again, these substances can be physical
entities like particles and fields.
87. I am open to other indices (spaces with just two dimensions, etc.). The above
is indebted to property exemplification (e.g., Kim, Events as Property Exem-
plifications 1973) and property instance (e.g., Paul 2000) theories of events.
See also R.M. Chisholm (1990, 419, def. D11) and R.C. Koons (2000) on
situations.
Causal Relata 249
88. They may also be indispensable in some way to our best special scientific theo-
ries, as well as any other collection of theories that correspond to reality. These
theories need not be scientific.
89. I will purposely suppress all kinds of finer details needed to both accommodate
the preceding summary description and to more fully detail the case of scatter-
ing quarks.
90. Sometimes yellow is used instead of blue, but it appears to be more orthodox to
use blue for one of three colors of quarks and yellow for the anti-color residing
in anti-quarks that are anti-blue (i.e., yellow) as in Goldberg (2017, 230). A.
Zee uses yellow as a color of quarks instead of blue in his example (Zee 2016,
532). The framework for much of this was given to us by the great physicist
Murray Gell-Mann, who, as it turns out, is a realist about quarks despite what
some histories report.
91. Zee (2016, 532).
92. “QED explains the force of electromagnetism—the physical force that causes
like charges to repel each other and opposite charges to attract—at the quan-
tum-mechanical level” (Kaiser, Physics and Feynman’s Diagrams 2005, 157).
93. Thanks to physicist Tom Banks for help here.
94. Dyson (S Matrix 1949; Radiation Theories 1949); Feynman (Theory of Positrons
1949); Feynman (Space-Time Approach to Quantum Electrodynamics 1949).
95. Feynman explicitly acknowledges that the diagrams illustrate interactions
(see Feynman, Space-Time Approach to Quantum Electrodynamics 1949,
787–788).
96. See the history in Kaiser (Drawing Theories Apart 2005).
97. Kaiser (2009, 236). And see also Kaiser (Physics and Feynman’s Diagrams 2005,
159–160). Physicist Tom Banks (personal correspondence 11/26/2017) tells me
that Feynman diagrams are used “to encode the predictions of” quantum theo-
ries suitably understood. There is textual evidence for these claims in Lancaster
and Blundell (2014, 181ff.), and in Zee (2016, 308), who says,
“According to Yukawa, the scattering is caused by the exchange of a pion
between the two nucleons, as shown in the . . . diagrams (surely you’ve heard
of Feynman diagrams!). . . . For our purposes here, you could think of the
diagrams as showing what’s happening in spacetime” (ibid.; emphasis mine).
98. Kaiser (Physics and Feynman’s Diagrams 2005, 162), citing Marshak (1952).
Kaiser (Physics and Feynman’s Diagrams 2005, 164) tells us that “physicists
most often [during the relevant era (circa 1950s) at least] used the diagrams to
study nuclear particles and interactions rather than the familiar electrodynamic
interactions between electrons and photons.”
99. Kaiser, Ito, and Hall (2004, 885). Rules connecting such diagrams to equations
are given in Lancaster and Blundell (2014, 182) for a φ 4 theory. See the rules
for QCD in Goldberg (2017, 234–235).
100. Kaiser (Drawing Theories Apart 2005, 175; but cf. also 177 and 187).
101. ‘t Hooft and Veltman (1994, 29).
102. Pearl includes more strictures on causal models but acknowledges that others
loosen those strictures.
103. On the assumption that Sloman will allow us to replace his “The World” ingre-
dient with some merely naturally possible system.
104. The same truth applies in more realistic models/diagrams of more realistic
nomological possibilities. See Lancaster and Blundell (2014, 298).
105. There is some hint of the idea that Feynman diagrams are causal models in
Hoover (2013, 36). However, this is all contra Handfield (2010, 123–126),
who explicitly rejects the idea that Feynman diagrams are causal models for
reasons that remain unclear to me.
250 Causal Relata
106. Sloman (2005, 70).
107. Mattuck (1967, 268). Following Kaiser’s source trail.
108. Kaiser (Drawing Theories Apart 2005, 268).
109. This helpful way of characterizing matters follows Dowe (2000, 123).
110. Bernstein (2015, 208). Bernstein does not actually maintain that omissions are
absences. Cf. my discussion of her view below.
111. Armstrong (Theory of Universals 1978, 44).
112. On the one hand, Schaffer (Causation and Laws 2008, 92) dismisses causation
in physics, whereas on the other hand, he is ready to insert not just causation,
but negative causation, into physics (Schaffer 2004, 202, 203; Disconnection
2012, 409–410). If there’s negative causation in physics, there’s causation in
physics. If there’s causation in physics, then his argument from physics for
causal reductionism (to be criticized in chapter 8) is unsound.
113. The discussion in Schaffer (Disconnection 2012) involves a misreading of Mel-
lor (1999). Schaffer thinks Mellor believes factsP are the relata of the causal
relation. The problem is that Mellor does not think of causation as a relation
at all (q.v., my criticism of him in chapter 2: sect. 3). For Mellor, there is but
causal explanation. And so it makes perfect sense for Mellor to appropriate
factsP because he’s giving a theory of causal explanation not causation. Mellor
is an eliminativist about the causal relation. Consider, “even if there are univer-
sals, there is no such universal as a relation of causation holding between par-
ticulars” (Mellor 1999, 161). Later, he rejects the existence of a causal relation
between ‘facta’ (understood as the entities that make factsP true; ibid., 163).
Then he denies that causation is a relation holding between factsP or statements
(ibid., 161, 168).
Schaffer likewise cites Jonathan Bennett (1988) as a proponent of the factsP
view of causal relata. I think this is likewise mistaken, although things are a lot
more complicated in the case of Bennett exegesis. I do not have time or space to
unwrap my reasons for doubting the interpretation, however.
114. I have refrained from discussing the details of Schaffer (2005) in the main
text for two reasons. First, no metaphysical theory or analysis, or account of
omissions, is provided therein. Indeed, the term ‘omission’ only occurs once in
the entire paper, and that is in a footnote 9. Absences are once again placed
at the center of a discussion about negative causation, but we are never told
what they are. Certain negative nominals are said to pick out positive events
that occur in their stead, but that presupposes that absences are distinct from
such positive events. So, there appears to be no actual theory of absences in
Schaffer’s entire corpus (see my discussion of his other papers in the main text).
Second, Schaffer’s (2005) attempt to save negative causation requires his dis-
tinctive contrastive theory of causation. That theory assumes a coarse-grained
theory of events that bounds them to individual worlds (ibid., 346). It says
of events that they are “Lewisian individuals . . . with counter-part theoretic
modal profiles” (ibid., 357). This modified Davidsonian view of events should
be confronted with the same series of difficulties I present for Sara Bernstein’s
appropriation of unactualized events that stand in counterpart relations below.
115. Clarke (2014, 21–28). Clarke does not actually commit to the thesis that omis-
sions are causal relata. He deliberately refrains from committing to a position
on that issue because he believes that in order to settle the matter one will need
a theory of causation first. He has “no such theory to offer” (ibid., 58). And
so, even if my criticisms fail, we are still left without a theory of omissions that
makes them fit to be causal relata.
116. Clarke (2014, 10).
117. Ibid., 35; emphasis mine.
Causal Relata 251
118. This view is discussed by Bernstein (2015) and others, but I can’t actually find
anyone who endorses it.
119. See Bernstein (2014, 6). This is the natural conception upon which Bernstein
builds her account.
120. See Bernstein (2014; 2015).
121. Bernstein (2014, 12–13).
122. Schaffer (2004, 202–203).
123. Thanks to physicists Tom Banks and Michael B. Weissman for help here. These
scholars were not commenting on Schaffer’s claims in the communications that
helped me develop my thoughts on these matters. Any misapplication of their
help is my unintentional fault.
124. Thanks to physicists Tom Banks, Don Page, and Michael B. Weissman for valu-
able correspondence on these issues.
125. Schaffer (2004, 202). In the evaluation that follows, I was greatly helped by the
discussion in chapter 4 of Brickley and Ives (2008). The sources cited therein
were a great boon to me. The historical discussion of Carpenter (1986) was
also quite helpful.
For modern studies on scurvy, see Brickley and Ives (2006); Hughes (1990);
Nishikimi and Udenfriend (1977); Pimentel (2003); and Stone (1965).
126. Gluecksohn-Waelsch (1963, 1270).
127. See on these matters Carpenter (1986, 173–175).
128. The follow-up essay is Holst and Frölich (Part II, 1907).
129. Holst and Frölich (Part II, 1907, 656; emphasis in the original). Compare Wil-
son’s summary,
They [Holst and Frölich] had shown that scurvy could be produced by diet
and could be cured by diet. Of the three theories of the cause of scurvy then
existing, infection, toxification, and faulty diet, only the last was supported
by their results.
(L. G. Wilson 1975, 51)
130. Drummond (1919, 80).
8 On the Argument From Physics
and General Relativity
I have probed some features of the causal relation. I have argued that it is an
obtaining, multigrade, asymmetric, transitive, and irreflexive relation that
relates events (understood in terms of the ESSI of chapter 7: sect. 4) and
sometimes forces and/or fundamental events (as understood in terms of the
characterizations in chapter 3: sect. 2 and chapter 7: sect. 5). I also argued
that causation is universal (for purely contingent events), and well-founded
(for purely contingent events). A theory of causation should say more. More
specifically, it should say whether its metaphysical nature is reducible to
the non-causal. That is to say, one’s theory of causation should be distin-
guished as reductionist or anti-reductionist, where you will recall that causal
reductionism is the doctrine that obtaining causal relations are grounded
in, reduced to, or completely determined by non-causal natural nomicity
coupled with the world’s unfolding non-causal history. And where causal
anti-reductionism or fundamentalism is the view that the causal relation is
not grounded in, reduced to, or completely determined by some non-causal
phenomenon or phenomena even if in combination with non-causal laws.
There are two commonly traveled direct paths to causal reductionism.
The first tries to show that causal reductionism is true by demonstrating that
a distinctive reductive theory of the causal relation is correct. The accounts
of David Fair (1979) and the previously discussed theory of David Lewis
(Postscripts 1986) constitute reductive theories. Fair (1979, 236) argued that
causation is nothing over and above the transfer of momentum or energy
(a conserved quantity), whereas Lewis maintained that causation reduces
to counterfactual dependence or the ancestral of that relation.1 Recall that
Lewis’s complete story characterized counterfactual dependence in terms of
the truth of particular counterfactual conditionals whose truth-conditions
are strongly related to obtaining degreed similarity relations between pos-
sible, albeit, concrete physical worlds.2 Non-causal laws of nature and the
overall non-causal history of the involved concrete physical worlds fix the
similarity relations (see my discussion of this in chapter 3: sect. 4.3.2). Thus,
for Lewis, causation reduces to non-causal physical history and non-causal
natural laws.
Argument From Physics and General Relativity 253
Many consider cases of symmetric overdetermination, asymmetric over-
determination, causation by prevention, and causation by omission to be
counter-examples to reductive theories.3 To illustrate the point by way of
just two examples, consider the fact that Fair’s view cannot countenance
instances of negative causation. For how can an absence transfer energy-
momentum? The worry is a difficult one.4 However, Fair and others who
affirm similar views (including Dowe 2000) can embrace my criticisms
of negative causation in chapter 7: sect. 6. Unfortunately for transfer and
conserved quantity theories more generally, there is a substantial empirical
objection to overcome (and now I follow the lead of Rueger 1998, 33–36).
Rehearsing some material from chapter 3: sect. 2, and chapter 4: sect.
2.3, I note how in special and general relativity (SR and GR), energy and
momentum are married, just as space and time are. The energy-momentum
of a body is given by its energy-momentum tensor (Tab). An associated
energy-density (or mass-energy) per unit volume (as measured by a hypo-
thetical observer) is given by Tabvavb, and our observer can be described with
a four-velocity va (a four-dimensional vector whose magnitude is given by
what relation it stands in to the metric tensor). Recall that Einstein’s field
equations are dynamical-interactive equations that relate the distribution of
matter (given by the energy-momentum tensor Tab) to the metric (gab) that
represents the inertio-gravitational field,
1
(Eq. 1): Gab ≡ Rab − Rg ab = 8π Tab (in a geometrized unit system)
2
Or in SI units and more modern notation with a cosmological constant,
1 8π G
Rµν − Rg µν + Λg µν = 4 Tµν
2 c
On the standard approach to GTR, our space-time can be represented by
the triple (M, gab, Tab), where ‘M’ is the four-dimensional differentiable
manifold that is smooth and without boundary over which (or at each point of
which) the metric tensor is defined. As is well-known, the energy-momentum
tensor Tab does not include the contribution of energy-momentum from the
gravitational field even though that field contributes energy-momentum to
physical systems. That contribution can happen quite independent of matter
(e.g., when gravitational waves propagate without the presence of matter,
or where Tab = 0). Because of the global nature of the inertia-gravitational
field’s contribution, it is difficult to acquire a global conservation law for
energy-momentum in GR. This may seem strange because relativity scholars
do use something like a conservation law, viz.,
(Eq. 2): ∇ aT ab = 0
which is called the covariant divergence law. However, that law pertains
only to local energy-momentum contribution from the matter fields of
254 Argument From Physics and General Relativity
space-time. According to Roger Penrose, because of the raised “extra” index
‘b’, in Eq. 2 one will be unable to obtain a global integrated conservation
law for energy-momentum from Eq. 2.5 Einstein knew of these problems
and tried to mollify them by introducing a pseudo-tensor that one could
associate with Tab. However, as Roger Penrose points out, that “quantity has
no local coordinate-coordinate independent meaning” (Penrose 1986, 136).
It therefore violated the spirit of general covariance. In addition, Einstein’s
attitude toward it was instrumentalist in that he used it to help with calcula-
tions (ibid.). Interestingly, even the instrumental value of that pseudo-tensor
was severely limited.
I may have moved too quickly. There are two options that may help to
defend the energy transfer or conserved quantity approach.
(Option A): There exists an avenue to a global integrated conservation

law if our space-time includes a Killing vector field ξ a (see Wald
1984, 286, n. 3; Dowe 2000, 97–98).
(Option B): We can also rescue the idea of the total energy of an isolated
physical system on the assumption that our space-time is asymptoti-
cally flat (i.e., at large-scale distances away from a single source, a
star or what have you, where curvature is negligible) (see Wald 1984,
269–297; cf. the discussion of asymptotically simple space-times in
Hawking and Ellis 1973, 221–225).
Both of the above options attempt to solve the problem of energy-

momentum in GR by using features of space-time structure. There are prob-
lems. (Option B) is a non-starter. Our space-time is not asymptotically flat.
The standard cosmological model SC (a model that is (or approximately
is) FLRW), discussed in chapter 4: sect. 2.3, does not describe a space-time
that is asymptotically flat. In addition, I’m aware of no plausible way of
recovering the notion of “the total energy of an isolated system”6 by means
of (Option B) that does not also use a Killing vector field.7 So, the success of
(Option B) depends upon an essential part (Option A).
(Option A) implies that some features of space-time geometry or structure
are not described by the metric in Einstein’s field equations. But as Wald
observed, that runs “completely counter to the spirit of general relativity,
which views the space-time metric as fully describing all aspects of space-
time structure and the gravitational field.”8 Adding in structure to rescue
an empirical analysis of causation seems like illicit natural philosophy. The
idea was to look at the world and come away with a sound theory of causal
phenomena. Requiring that we alter one of the most empirically successful
physical theories to achieve that goal seems dubious if we are without any
independent motivation for that modification.
Things are actually much worse for (Option A). Relativity scholars are
in agreement. Our space-time does not possess the requisite symmetries
indicative of the presence of a Killing vector field, and space-times with such
symmetries or fields are not generic.9 It appears that energy transfer and
conserved quantity theories have been empirically refuted.
In spite of individual problems with reductionist accounts of causation,
reductionists can remain steadfast in their adherence to causal reductionism
because there is thought to be a second independent path to that position,
which travels by way of the argument from physics for causal reductionism,
(1) If all approximately true physical theories require only laws of nature
and physical history and do not have need of causation, then causal
reductionism is true.
(2) All approximately true physical theories require only laws of nature and
physical history and do not have need of causation.
(3) Therefore, causal reductionism is true.10
In chapter 2: sect. 3, we saw a similar argument from physics, although

that argument attempted to show that the absence of causation from our
best physical theories demonstrates the elimination of causation. Argument
(1)–(3) is more modest in that it only moves from the absence of causation
in physical theory to causal reductionism.
Chapter 2: sect. 3, chapter 3: sects. 2–4, chapter 4: sect. 2, and chapter 7:
sect. 5 all argued that causation enters physics.11 However, those arguments
either explicitly or implicitly relied upon the propositional view of physi-
cal theory structure laid down in chapter 1: sect. 4.5.5, and they left the
notion of causation at work in the argumentation at an intuitive level. In
this chapter, beginning in sects. 2 and 3, I explain how best to defeat prem-
ise (2) of the argument from physics with only very minimal assumptions
about the nature of physical theory structure (i.e., without having to endorse
any posit about theory structure that is incompatible with the propositional,
semantic, or syntactic views of theory structure). In sect. 4, I employ the
strategies outlined in sects. 2 and 3 to argue that what I call minimal fun-
damentalist causation (explained in sect. 3.2) enters the best interpretation
(broadly understood) of general relativity. That minimal theory of causation
foreshadows the full account or analysis provided in chapter 9. In sect. 5,
I argue that standard cosmology or SC uses domains of causal influence that
help constitute causal pasts and causal futures. I maintain that that causal
structure cannot be reduced to or identified with non-causal structure such
as light cone structure despite the constant confusion of the two. Sect. 6
includes my responses to objections to causal GTR. These responses will
protect not only the case for causal GTR presented in this chapter, but also
the shorter, more assumption-laden case for causal GTR in chapter 3: sect. 2.
Section 2: The Argument

In keeping with remarks previously expressed about the first path to
causal reductionism that travels via reductive theories of causation and the
independence of the argument from physics, I ask the reader to at least
assume that,
(Premise 1*): All extant reductive theories or philosophical analyses of

causation fail.
Premise (1*) is generally upheld in our contemporary philosophical milieu,12

and I have already sampled some of the reasons why reductive theories are
regarded as problematic in numerous places of the current project (i.e.,
I have just refuted David Fair’s and Phil Dowe’s view in sect. 1 of this chap-
ter; I criticized the theories of Dowe (again), Lewis, and reductive theories
more generally in chapter 3: sect. 4; I criticized Menzies and Price’s purport-
edly reductive analysis in chapter 4: sect. 2; and I criticized the reductive
probabilistic theory of Eells and Sober in chapter 6: sect. 2.2).
Assuming that extant reductive theories and analyses fail does not by
itself privilege causal fundamentalism. All contemporary reductive theories
and analyses of causation could fail, and yet causal reductionism come out
true. The latter position is not identical to any one reductive analysis or
theory. Moreover, that all contemporary theories fail does not guarantee
that there will not be a future successful account.
The strategy behind my use of (1*) is simple. If your choice reductive the-
ory or analysis were correct, one could always respond to any consideration
in favor of admitting causation into any scientific theory whatsoever by
reductively analyzing away the involved causal notions or causal structure.
I therefore ask the reductionist to check their reductive analyses and theories
at the door. The question is whether they can justify causal reductionism
by way of the argument from physics without using reductive theories or
analyses. The question is whether the argument from physics can stand on
its own.
(Premise 2*): If the argument from physics is sound, then all approxi-
mately true physical theories require only laws of nature and physical
history and do not have need of causation.
This premise follows from the formulation of the argument from physics in
sect. 1.
(Premise 3*): The general theory of relativity is an approximately true

physical theory whose formalism requires a causal interpretation.
I take for granted the fact that GTR is an approximately true physical theory,
just as I take robust scientific realism for granted (q.v., chapter 1: sect. 4.5.5.5).
(Premise 4*): If (1*) and (3*), then it is not the case that all approxi-
mately true physical theories require only laws of nature and physical
history and do not have need of causation.
If my demonstration of (3*) is successful, and there really is good justifica-
tion for (1*), then there is an approximately true physical theory that does
not require only non-causal laws of nature and non-causal physical history
because it does in fact have need of causation. One is not entitled to inter-
pret the examples of causal structure and causal phenomena in GTR in
reductive terms, for that maneuver is precluded by the truth of (1*). If one
needed the truth of a reductive theory of causation so as to escape objec-
tions to the argument from physics, it would follow that the soundness of
the argument from physics is derivative in that it is dependent upon the
success of a reductive theory or analysis of causation. But again, the argu-
ment from physics is supposed to be independent motivation for causal
reductionism.
(Conclusion): Therefore, the argument from physics is unsound.
The conclusion follows logically from the premises.
Section 3: On Pushing Causation Out of Physics and on

How to Reintroduce It
Section 3.1: How to Defeat the Argument From Physics

With Minimal Assumptions
Let me now articulate a very general approach to physical theory structure
upon which my case for causal GTR in this chapter will be built. Let an
interpretation of a physical theory T be a part of T that specifies how its
formalism describes the world (Cartwright 1993, 426). An interpretation of
T’s formalism will employ T’s primitive ideology, or those notions that are
indispensable to that interpretation (e.g., the notion of a force in Newtonian
mechanics, or energy in classical Lagrangian mechanics, or gravitational
field in GTR).13 An interpretation will also make explicit use of the ontology
of T. The ontology of T is that list of entities (including structures, processes,
properties, and relations) that are required to properly connect—by way of
a representation relation—the formalism of T to some proper or improper
part of the world (or some suitable possible world). The ontology of T will
also include entities (or entity-types) that are indispensable to the direct
metaphysical explanation of the truth of the interpreted fundamental laws
of T. As was noted in chapter 1, one popular way to think about metaphysi-
cal explanation is in terms of truthmaking, where a truthmaker M makes
some proposition true by de re necessitating its truth. That is to say, M’s
existence necessitates the truth of the relevant proposition in some non-de
dicto fashion.
Let me illustrate what I have in mind by means of an example in classi-
cal physics. Consider one of the Lorentz–Heaviside–Maxwell (differential,
microscopic, vacuum, three-vector form) equations of classical electrody-
namics in SI units,
∂b
(Faraday’s Law (FL)): ∇ × e + =0
∂t
The curl of the microscopic electric field plus the partial derivative of the
microscopic magnetic field with respect to time is equal to zero. If I had
space, I would argue that FL should be interpreted in such a way that that
interpretation at least involves the following,
(Interpreted FL (I-FL)): Necessarily(natural), for any microscopic magnetic

field m, if m exhibits dynamical behavior (time-varying change involv-
ing field undulation), then there is a microscopic electric field e such that
m’s time-varying undulations are causally responsible for e’s dynamical
behavior (its undulations).
I have included the necessity operator because laws must hold in some sense
that is stronger than the modal force of accidental truths, but weaker than
the modal force of logically and metaphysically necessary truths. The inter-
pretation will need to go further by informing us about the metaphysical
nature of microscopic magnetic and electric fields, as well as undulation
processes, and the causal relation between such processes (keeping in mind
that causation is asymmetric and FL is time-reversal invariant). The ontol-
ogy of classical electrodynamics will need to include, inter alia, microscopic
magnetic and electric fields, as well as any other entities that would be
responsible for directly making I-FL true.
All of this breeds a somewhat narrow conception of theories in that dif-
ferent interpretations of the same formalism produce different physical
theories. But even after taking on board the narrow conception, one can
speak loosely of the general theory of relativity, or non-relativistic quantum
theory. A more general use of such locutions may serve historical or peda-
gogical purposes. In other words, one might, via the locution ‘general theory
of relativity’ intend to pick out that thing which Einstein promulgated in
1915, or a class of theories that predict and explain virtually all of the same
phenomena (e.g., Einstein’s GTR and the ADM-formulation of 1962).14
These theories* (broad uses of the term will now include asterisks) involve
formalisms and partial interpretations of those formalisms. Regarding par-
tial interpretations (for the purposes of this chapter), Laura Ruetsche wrote,
[T]he vast majority of the theories philosophers talk about are already
partially interpreted. Otherwise they wouldn’t be theories of physics.
These theories typically come under philosophical scrutiny already hav-
ing been equipped, by tradition and by lore, with an interpretive core
almost universally acknowledged as uncontroversial.15
I have already introduced some of the partial interpretation of GTR*

in sect. 1 and in chapter 3: sect. 2. But to be clear, at the cost of some
repetitiveness, the current chapter will assume at least a partial interpreta-
tion of GTR*, which affirms that the Lorentz metric (i.e., the metric tensor
gab of Lorentz signature) represents the gravitational field.16 This seems to be
the standard view in contemporary relativity physics,17 and one could make
the case that Einstein’s view was close enough to it, despite his remarks
about the Christoffel connection.18 I will also continue to assume that the
gravitational field reduces to space-time geometry. That reduction is also
fairly standard in the relativity physics literature (as I noted in chapter 3).19
Thus, if the gravitational field is a causal entity, space-time curvature should
be understood as causally efficacious. As the Class of Physics of the Royal
Swedish Academy of Sciences put it, “[t]he new insight [of GTR] was that
gravity is really geometric in nature and that the curving of space and time,
space-time, makes bodies move as if they were affected by a force.”20
Completing the partial interpretation of GTR* should involve at least the
additional claims that the energy-momentum tensor describes matter fields,
and that solutions to Einstein’s field equations afford a description of mat-
ter, energy, and space-time geometry.21 And so, Einstein and others working
in the domain of relativity physics already associated certain interpretive
postulates and principles with the relevant formalism. That received “inter-
pretive core” is a part of the extension of more general uses of the locution
‘general theory of relativity.’ When teaching relativity physics, references to
GTR* are often meant to pick out the formalism and accompanying partial
interpretation in the early pioneering and contemporary work.
Contra Ruetsche, I do not affirm that the full or complete interpreta-
tion of a physical theory includes a physically significant specification of
those physically possible worlds consistent with its laws.22 Such a concep-
tion is far too contextualist. The physical significance of the specification
is, by Ruetsche’s lights, a disclosure of physical possibilities that are impor-
tant for the progress of physics, and that helps explain why the theory
is successful. What may or may not be important for progress in physics
depends on the context of physical inquiry. Precisely what is significant for
physical inquiry is contingent upon the context of that inquiry, the par-
ticular researchers involved, their pragmatic interests and commitments,
their funding, and their goals. Ruetsche seems to invite this reading in the
context of her discussion of quantum theories of infinite physical systems
(what she calls “QM∞”),
In slogan form, my contention is that, when it comes to QM∞ at least,

there is a pragmatic dimension to theory articulation. What set of pos-
sible worlds we associate with a theory of QM∞ can depend on what
we’d like to do with that theory: what explanations, involving which
magnitudes and guided by what laws, we aspire to; what phenomeno-
logical models we need to construct; what projects of theory develop-
ment we’d like to sponsor.23
But the worlds that one associates with T are essential to the interpretation
of T. Thus, the contextual and pragmatic aspect of interpretation is written
directly into Ruetsche’s view of physical interpretation. Ruetsche states,
“[t]o interpret a physical theory is to delimit what that theory recognizes as
physically significant . . . [and] . . . the delimitation” should “respect how the
theory matters to the present and future of physics.”24 The set of physically
possible worlds wrapped up in the interpretation of T must be worlds that
matter to the pursuits of “present and future” physics.25 Again, that makes
physical interpretation dependent upon pragmatic and highly contextual
matters. The pragmatic and context dependency of her account makes it
difficult to remain committed to the robust scientific realism assumed in the
current work. For if it is the case that how one should understand a theory
depends in part on pragmatic concerns, then whether one should commit
to the mind-independent existence of unobservable entities postulated by
that theory will likewise partially depend on one’s pragmatic concerns,
and that seems contrary to the spirit of scientific realism. Thus, in con-
trast to Ruetsche’s conception of theoretical interpretation, my view (that
the general structure of physical theories is best understood in terms of the
formalism/interpretation bifurcation) is thoroughly consistent with realism.
That is a plus.
As I noted in chapter 1, on the semantic view of theory structure (for
which see Giere 1988; Suppe 1989; van Fraassen 1970; 1991, 4–8; 2008,
309–311), physical theories are given by formalisms that specify models.
Models can be used to depict or represent an appropriate domain of goings-
on in the world. Bas C. van Fraassen maintains that a necessary condition
for rightly deeming a model’s representation of some phenomenon accurate
is that that model includes a substructure that is isomorphic to the phe-
nomenon that one is attempting to represent.26 It is natural to understand
interpretations on this approach to be postulates and principles that connect
models of theories to the mind-independent world (Giere 1988, 85). Hence,
what I’ve said so far about theoretical interpretation fits snuggly within the
semantic view of scientific theory structure. On that conception, my thesis
would be that causation enters the postulates and principles (or what van
Fraassen calls “theoretical hypotheses”) that join GTR’s models—those that
are determined by its (perhaps partially interpreted) formalism—with the
world.
My approach in this chapter is also completely consonant with the syn-
tactic conception of theory structure.27 With respect to GTR and a syntactic
understanding of its structure, my suggestion would be that the proper set of
interpretational postulates includes, in its constituent sentences laden with
observational terms, the causal influence relation.
Return now to the question under consideration: How does one demon-
strate that a physical theory like GTR requires causation? Answer: Show
that the causal relation is a non-excludable member of that theory’s ontol-
ogy. Evidence for the thesis that the causal relation is essential to GTR’s
ontology can issue forth from the fact that the notion of causation shows
up indispensably in the best interpretation (its primitive ideology) of the
underlying formalism of that theory. But what do I mean by causation?
What is the causal relation that I believe is indispensable to GTR like? I will
now sketch a general characterization of the type of causation that I believe
best suits causal phenomena generally, but general relativistic effects par-
ticularly. The theory I present is incomplete, and it merely foreshadows my
full account in chapter 9.
Section 3.2: Which Causation? The Minimal Fundamentalist Account

Because we are suppressing, for the purposes of deliberation, the truth of
extant reductive analyses and theories of causation, the type of causation at
the heart of the causal interpretation of GTR (or C-GTR) is not a relation
that is reducible to counterfactual dependence, probabilistic dependence,
the transfer of energy or momentum, or some other reductive surrogate rela-
tion or process. Rather, if C-GTR is the best interpretation of GTR’s formal-
ism, there will exist in the primitive ideology (the indispensable notions of
the ideology) of the theory, a causal notion that refers to a causal relation.
That causal relation will be a constituent in the metaphysical explanation
of the truth of the interpretation. It will likewise be multigrade, asymmetric
(although not always temporally asymmetric), transitive, and irreflexive (in
the spirit of chapters 3, 6, and 7, and foreshadowing the account to be pre-
sented in chapter 9).
Events will be understood in the ESSI sense of chapter 7. Moreover, cau-
sation is a dependence relation. When event x causes event y, y depends for
its existence and contingent content on x. Event y’s contingent content con-
sists of the contingent nexus of exemplification connecting the substance-
constituent of y with the properties it exemplifies at the relevant ontological
index. If the substance constituent of y is σ, and we want to know why σ has
Fness at space-time location L, we can respond with a report on that which
caused y, viz., x (given that x satisfies conditions of causal relevance for
which see the plausible account of Woodward (2003, 187–238, 351–353)).
It is event x’s occurrence that is causally responsible for y’s occurring and for
y’s having the contingent content it does (σ’s exemplifying Fness) at the rel-
evant index. Of course, not all causes are causally relevant, but I articulate
a case that assumes x is relevant so as to highlight the fact that causation, as
I am understanding it, hooks up with causal explanation. Moreover, mini-
mal fundamentalist causation is consistent with multifarious theories of sci-
entific and causal explanation (e.g., Strevens 2008; Woodward 2003).
Is the above a reductive minimal theory of causation? No. The dependence
in play is causal dependence. I therefore agree with Lewis (Causation 1986)
that causation should be understood in terms of causal dependence. How-
ever, I resist Lewis’s additional step of reducing causal dependence to coun-
terfactual dependence. And although this bare-bones theory of causation
is anti-reductionist/fundamentalist, it is not primitivist. Causal primitivism
is the view that causation cannot be accurately described by any theory or
analysis that includes entities (including properties and relations), or con-
cepts that are more fundamental than causation itself. The dependence rela-
tion at the heart of the phenomenon is itself irreducibly causal. But minimal
fundamentalist causation does provide some understanding of the causal
relation in terms of something more fundamental, insofar as it seeks to cash
out the nature of causation in terms of dependence. What we have here then
is an anti-reductionist account of causation that can allow for primitivism
about causal dependence although it precludes primitivism about causation.
Primitivism about causation entails anti-reductionism about causation,
although the converse does not hold (J. Carroll 2009, 281).
I call the preceding account, the minimal fundamentalist account. It is fun-
damentalist because either the involved causal dependence is irreducible and
primitive, or else there is some further informative causal characterization
of that dependence (i.e., it is not primitive). The latter disjunct is why the
account is minimal (and here is where I depart from the contents of chap-
ter 9, and render my objection to the argument from physics more amiable
to almost all anti-reductionist theories). It is consistent with more detailed
anti-reductionist theories. The fact that y causally depends on x (in the pre-
ceding sense) may itself be grounded in a further fact about manipulability
(as in Woodward, 2003, 25–93), or a further fact about how a hypothetical
free deliberating agent acting so as to secure the occurrence of y by way of
an effective strategy, would cause x in order to bring about y (as in Men-
zies and Price 1993, which I argued is anti-reductive in chapter 3), or the
further fact that some causal law involving a causal necessitation relation
between states-of-affair types begins to be instantiated (as in Armstrong
1997, 202–219). And yet, my theory is incompatible with some powers-
based approaches to causation that reject the tenet that causation is a rela-
tion (as in Lowe 2016).
I do not actually believe that the causal dependence relation in the minimal
account can be adequately metaphysically explained by causal powers, or
instantiations of causal laws, or facts about manipulability by a hypotheti-
cal agent or causal mechanism (for reasons why, see previous discussions in
this work, and see my criticisms of the interventionist account in chapter 9).
However, for the purposes of this chapter, I tolerate the subsumption of the
minimal theory by other anti-reductive accounts so as to provide a more
persuasive strategy for combating the reductionist argument from physics.
Before passing on to my case for C-GTR, I should highlight that sub-
stantiation of the soundness of the main argument that is (1*)–(Conclu-
sion), does not require that I show that minimal fundamentalist causation
is instantiated in every single general relativistic effect or dynamical action.
Nor does it require that I show that every feature of minimal fundamentalist
causation is transparently and expressly manifest in every general relativistic
effect or dynamical evolution. That burden is overly cumbersome for any
proponent of any theory of causation in science. Rather, it is enough to
provide an abductive case. The best hypothesis that we can proffer about
the type of relation or interaction that gravitation is, is the causal hypoth-
esis. I will show that no other relation is fit for the job.
Section 4: The Gravitational Field as Cause

Now that one knows that I will inject causation into physics by arguing
for C-GTR in a manner that is consistent with the syntactic and semantic
approaches to theory structure, and now that one knows what type of cau-
sation I am attempting to put in GR, it is time for me to build on my case
for causally interpreting gravitational action in instances of inertial motion
(the very case initially sketched in chapter 3: sect. 2).
Section 4.1: Geodesic Motion in Contemporary Physics

Recall that Einstein’s general theory of relativity is based upon the strong
equivalence principle, which asserts that the metric field structure of space-
time is responsible for both inertial and gravitational effects. It implies that
gravitation is strongly related to space-time geometry.28 Einstein’s field equa-
tions (EFEs) given in (Eq. 1) detail the precise nature of the relationship.29
GTR adds to the preceding formalism (perhaps, merely as theorems), geo-
desic equations of motion for free and test particles as well as varying types
of matter. The easiest case is the perfect fluid whose dynamical geodesic
equation of motion is reminiscent of (Eq. 2),
(Eq. 3): ∇ aTab = 0
Whereas ‘▽a’ is the covariant derivative operator that satisfies the usual
conditions, Tab now equals ρuaub + P(gab + uaub) in that it gives the per-
fect fluid’s energy-momentum tensor. ‘ρ’ represents the fluid’s density, ‘P’ its
pressure, ‘u’ its four-velocity, and ‘gab’ is the Lorentz metric representing the
gravitational field (Wald 1984, 69). It is uncontroversial that (Eq. 3) follows
from (Eq. 1) (given certain interpretational postulates), and as was noted
in chapter 3: sect. 2.4, there is an associated general principle of geodesic
motion (GP) commonly discussed in contexts like these. The GP (for mas-
sive bodies, the massless case is given by Lambourne below) says that it is
naturally necessary that due to gravitational influence, free massive bodies
tread down timelike geodesics understood as the straightest possible curves
of the space-time metric.30 Or it is naturally necessary that timelike geode-
sics represent the possible trajectories of free massive bodies that are under
the influence of gravitation. The language of influence is no foreign intruder.
Physicists typically understand GP that way. Robert J. A. Lambourne wrote,
In general relativity, the time-like geodesics of a spacetime represent the

possible world-lines of massive particles falling freely under the influ-
ence of gravity alone. And, similarly, the null geodesics of a spacetime
represent the possible world-lines of massless particles moving under
the influence of gravity alone.31
As I argued in chapter 3: sect. 2.2, the GP has traditionally been inter-

preted causally because Einstein and others affirmed that how such objects
find themselves in the aforementioned paths is through the determining
causal influence of the gravitational field (qq.v., notes 44 and 48 of chap-
ter 3 for references).
Arguably, one can derive the geodesic equations of motion for massive
particles from the field equations (given certain interpretational postu-
lates).32 Causal interpretations of the relevant formalism may therefore be
unproblematic for the reductionist, for they can greatly mitigate the evi-
dential importance of those interpretations for the causal reductionism vs.
causal anti-reductionism debate by insisting that the theoremhood of those
equations/laws suggests that even given a causal interpretation, the involved
causal phenomena are not fundamental to GTR or GTR*.
Notwithstanding the easy case of the perfect fluid discussed before,
I worry about the relevance and plausibility of many of the derivations for
free massive particles in the real world.33 Many of them assume an overly
idealized or implausible picture of particles. Einstein and Grommer rejected
attempts to derive the equations of motion from the EFEs that appealed to
an energy-momentum tensor field Tab description of matter (Einstein and
Grommer 1927).34 To help with attempts at deriving the geodesic equa-
tions of motion, they chose to describe matter in terms of singularities. That
characterization breeds rather absurd consequences because on it, geodesics
of massive bodies do not reside in space-time (Earman 1995, 12; Tamir
2012, 142).35 Other ways of understanding matter in the relevant equations
exist, as do other types of attempted proofs that help one skirt around some
of the issues. There are, however, potential problems with these alterna-
tive approaches. In fact, the problem of articulating exact solutions for two
gravitating masses is particularly disturbing. So much so that in 1989, Peter
Havas could write,
No . . . exact solutions can be hoped for in the case of two or more bod-
ies of comparable masses interacting gravitationally . . . the only exact
result to date is a negative one, Weyl’s proof of the nonexistence of a
static two-body solution.36
Suppose there were exact solutions and uncontroversial derivations of

the geodesic equations of motion for real particles. What interesting conclu-
sions could we draw? The reductionist could follow Harvey Brown (2005,
161–163). He maintains that plausibility arguments like those in Misner,
Thorne, and Wheeler (1973, 471–480; henceforth, MTW) recommend that
the geodesic equations of motion are not axiomatic, and that therefore the
geodesic principle should not be understood in causal terms.37 I have two
responses to Brown’s reasoning.
First, I have asserted repeatedly that gravitation is an interaction, and
that with this everyone (or virtually everyone) agrees. But let me now make
a case for that claim. Observe that if there are successful derivations, they
would connect field equations with particle behavior. That is initially bizarre.
Puzzlement is removed through an appreciation of what the entailment and
geodesic equations say. They are not directly and solely about the motions
of particles. Rather, one gains insight into the dynamical behavior of the
particles “by looking at the geometry outside the object[s]. That geometry
provides all the handle one needs to follow the motion[s].”38 The motions
are affected by space-time geometry, by the behavior of the gravitational
field because matter “and the metric are tied together by Einstein’s equa-
tion.”39 Gravitational interaction determines geodesic motion. Moreover,
departure from geodesics can arise because of features of space-time geom-
etry. Or, to put it another way, “anomalies in the motion go hand in hand
with anomalies in the geometry.”40 These facts only highlight and empha-
size the interactive connection between field and particle motion.41 In other
words, even if one can derive (exactly) the geodesic equations of motion
from the EFEs, that entailment underwrites a coupling fact regarding the
relationship between the gravitational field and matter fields (and thereby
particles). However, that coupling fact is best metaphysically explained by
the presence of an interactive connection between the aforementioned fields.
Second, Brown’s reliance on MTW is, from my standpoint, fortuitous.
Misner, Thorne, and Wheeler (1973, 476–477) use causal locutions to inter-
pret the geodesic equations of motion (e.g., “zone of influence,” 477). This
supports my contention that, theorem or not, the correct interpretation of
the relevant dynamical coupling is interactive. That interpretation is not
threatened by recognition of theoremhood. Here then is my interpretive
proposal. The correct interpretation of the geodesic principle in GR ought
to incorporate at least the following,
(The GP Dictum for Massive Bodies): Necessarily(natural), the gravita-

tional field understood as space-time curvature causally generates the
inertial and time-like geodesic motion of free massive bodies.
It is the truth of (The GP Dictum) that I believe explains the truth of MTW’s
assertions about space-time geometry and free massive particle motion.
While attempting to counter the C-GTR, Brown seems to (implicitly
at least) suggest that once Einstein was able to see the dynamical connec-
tion between the field equations and the geodesic equations of motion he
abandoned the causal interpretation of the geodesic principle (H.R. Brown,
2005, 161, note his use of the “during this period” clause). If Brown were
right about this, we would have an important piece of data that proponents
of C-GTR would be hard-pressed to explain. Fortunately, even after the
publication of Einstein’s first derivation of the equations of motion from the
vacuum field equations in 1927 (see Einstein and Grommer 1927), Einstein
continued to affirm a causal interpretation of both GP and classical physics
more generally. For example, one sees in Einstein’s discussion of the rela-
tionship between quantum and classical theory in his essay in Forschungen
und Fortschritte a clear commitment to causal structure in the natural world
that is described by classical physics (which includes relativity), and that will
be further revealed by a complete causal unified field theory (Einstein 1929).
Einstein remarked,
Natural phenomena seem to be determined to such an extent that not

only the temporal sequence but also the initial state is fixed to a large
extent by law. It seemed to me that I should express this idea by search-
ing for overdetermined systems of differential equations. . . . I strongly
believe that we will not end up with a Subkausalität but that, in the
indicated sense, we will arrive at an Überkausalität.42
Pais translates ‘Überkausalität’ as “supercausality.” Einstein’s 1929 discus-

sion presupposed a causal understanding of classical (see the discussion in Pais
(1982, 464–465; cf. Pais 1991, 191–192; 230–231)) and quantum theory (see
Weinert 2005, 96).
Like Einstein, contemporary relativity scholars have not refrained from
causally interpreting the action of the gravitational field even after many
attempted derivations and plausibility arguments were published. When dis-
cussing how precisely to interpret the GP, given issues about the argument
from physics and causal reductionism in the background, Robert M. Wald
remarked that generally “[t]he metric and matter fields are coupled and
undergo causal interactions,” and both metric and matter “influence each
other causally.”43
Brown may have been assuming that what is fundamental to a physical
theory is that which can be closely read off of the axiomatic formalism of
that theory (in this case, the field equations). But if one goes in for such a
view of fundamentality and physical theorizing it ought to be essential to
one’s understanding of what’s fundamental to that theory that everything
that is derivative fall out of true descriptions of the world’s fundamental
structure according to those axiomatic equations and their solutions. God
needed only to ensure that a general relativistic space-time satisfied general
relativistic axiomatic equations and/or their solutions, and that certain ini-
tial conditions obtained. All other relativistic structure falls out by conse-
quence of the creation of both fundamental structure and initial conditions.
This view of the structure of a physical theory does not suit GTR well even
though there exists a well-posed initial value formulation of the theory (see
Wald 1984, 243–267).44 The geodesic equations of motion for a free particle
follow from the field equations only if certain interpretational postulates
are assumed to hold (see on this Malament 2009 and Malament, Remark
about Geodesic Principle 2012). For example, the distributional proofs of
the geodesic equation use Einstein’s generalized EFEs. However, the covari-
ant divergence law ▽aTab = 0 does not follow from those equations because
the Bianchi identities are not true for all solutions to a generalized form
of the field equations because those equations use distributional tensors
(these points are from Tamir 2012, 144). However, the divergence law is
an extremely important principle that does powerful explanatory work in
GTR. One must assume it in the context of distributional proofs and even
then, there are problems to overcome.
Add that the famous limit operation proof of Geroch and Jang (1975)
must use an interpretive postulate in the antecedent of the theorem. As origi-
nally stated, the theorem did not use what’s called the weak energy condition
(i.e., that the energy density of the relevant matter fields is non-negative).45
But rather, it made use of (although perhaps inessentially (see the discussion
of this in Weatherall (2011)) the strengthened dominant energy condition
(although Geroch and Jang called it “the (strong) energy condition” (1975,
66)). That condition subsumes the weak energy condition, but adds that the
four-momentum density of the relevant matter fields is a vector that is both
timelike and future-directed (cf. Malament, Topics in the Foundations of
GR 2012, 144, or more technically, “[g]iven any timelike vector ξ a, at any
point in M, Tab ξ aξ b ≥ 0 and, if Tab ≠ 0 there, then Tbaξ b is timelike” (ibid.)).
These energy conditions amount to restrictions on the energy-momentum
tensor in the field equations that do not follow from the field equations
themselves (see the comments of Curiel 2014, 41 on the EFEs and the energy
conditions). They are therefore interpretational postulates not part of the
axiomatic equations. They are used to help one understand the nature of the
objects described by the axiomatic formalism of GTR.
Perhaps one could incorporate the relevant postulates into the theory as
axioms in order to escape the causal interpretation? There are many reasons
why that is implausible. Here are two. First, it unnecessarily restricts the
domain of nomologically possible relativistic space-times to those that are
described by the field equations (and perhaps their solutions or simplifica-
tions) with an energy-momentum tensor that satisfies the relevant promoted
energy conditions. Second, the energy-momentum tensor depends on, inter
alia, the metric gab.46 One cannot specify the distribution of matter with-
out determining the metric. The two are intimately related, and one must
account for that relationship when seeking to interpret and solve for Tab.47
But as I’ve suggested before, and as I will attempt to conclusively argue
below, the entity represented by gab is a causal entity, relating causally to
matter distribution. That is the best account. Let me fuel this idea some.
Section 4.2: The Interaction Must Be Causal: Closing

the Argumentative Loophole
In preceding discussion, I argued that Einstein and a great many relativity
scholars interpreted/interpret the geodesic principle causally. I have likewise
argued that that interpretation is not defeated by the mere fact that geodesic
equations of motion can be derived from the field equations. The existence
of a derivation of the geodesic equations of motion from the EFEs under-
writes, and does not count against, the judgment that the gravitational field
interacts with matter fields and particles. I will now argue that the causal
interpretation is the best interpretation.
Section 4.3: Interactions as Causal Phenomena Again

Gravitation is an interaction. Chapter 3: sect. 2.3 argued that physical
interactions in general are causal. Support for causally interpreting physi-
cal interactions consisted of evidence from the nature of our concept of an
interaction, lexical evidence that the term is a causal one (and I’m assuming
that physicists are neither conceptually nor linguistically confused or incom-
petent when describing the relevant phenomena with the term ‘interaction’),
evidence from the deliverances of reductive and anti-reductive theories of
causation that entail that interactions are causal, and prima facie evidence
from specific cases of physical interactions that appear to be plainly causal
in nature. There is yet another argument for understanding physical inter-
actions causally. I call it the historical argument. It turns on the fact that
interactive theories have been traditionally outfitted with causal interpreta-
tions. For example, in classical Maxwellian electrodynamics, the Lorentz
force law F = q(E + v × B) specifies an interactive relationship between
external fields and charged particles. Frisch adds, “the association between
external fields and the acceleration of a charge is usually interpreted caus-
ally: External fields cause charges to accelerate” (Frisch 2005, 30). More-
over, the acceleration of a point charge interacts with electromagnetic fields.
That is likewise standardly understood in causal terms (Frisch 2005, 30).
That there exists a firm tradition of causally interpreting the strong, weak,
quantum electrodynamical, and even the unified electro-weak interactions
is well-supported by the fact that physicists have used Feynman diagrams
as representations of such interactions. I argued in chapter 7: sect. 5 that
these diagrams (with associated formalism) are causal models used to help
elucidate the deep metaphysical nature of the phenomena under study. The
best explanation of the fact that the fundamental interactions (including
gravitation) in physics are usually causally interpreted is the further fact that
the phenomena involved, the interactions, are causal.
Section 4.4: Gravitational Interactions Are Not Instances

of Non-Causal Determination
Given that gravitation is an interaction, and that interactions are causal, we
can mount a strong case for the claim that the causal interpretation of gravi-
tational interaction is indeed forced upon us. We should not, for example,
regard the determining influence of the gravitational field as mere non-causal
nomological determination, where that relation involves one state of a phys-
ical system necessitating another in a way that is backed by dynamical laws
(i.e., the dynamical law in question guarantees that the necessitation rela-
tion obtains). The “effects abstractly result from certain constellations of
temporally prior conditions without necessarily being actively produced by
them.”48 There are two reasons for this. First, an appeal to nomological
determination begs the question against my argumentation. What ensures
that the necessitation relation obtains are the backing laws, and if those
laws are causal as I have argued, they will, with a brand of necessity, guar-
antee that causes relate to effects. In contrast, the objection on offer has
need of non-causal nomological determination relations, and so the laws in
play should be non-causal dynamical laws. The backing laws in general rela-
tivistic cases will be the laws of GTR. And so, the question just is whether
GTR’s laws should be interpreted causally.49
The most substantial benefit of the causal nomological determination
approach is that it fits more appropriately with the datum that gravitation
is an interaction. If a physical phenomenon is interactive, then the laws that
describe it should be causal. If they are not, then the scientific explanations of
general occurrences of that phenomenon that those laws afford will miss out
on structure and consequently fail to be complete. As I have reminded the
reader many times, in the spirit of chapter 1: sect. 4, my approach is a world/
phenomenon-first natural philosophical approach. We do not learn what the
world is like by just looking at disconnected math. There would be no rea-
son for deeming the math approximately correct if it did not track goings-on
in the world. Einstein and others advanced GR for the purposes of repre-
senting the world. It was intended to explain the perihelion of Mercury,
and it was intended to explain why accelerating in a vacuum and “resting”
in a uniform gravitational field cannot be discriminated by an observer on
operational grounds. Phenomena come first. Math comes later. That meth-
odological and/or procedural order motivates classes of interpretations of
the formalism, in this case, the causal interpretation of GTR.
There is, of course, a debate in the philosophy of physics literature about
whether laws of nature should in general be understood causally. I will not
take up the banner of any one side on the more general issue. The spe-
cific case of GTR’s laws of geodesic motion quite clearly supports causal
approaches to laws of nature in general relativistic contexts. My judgment
is supported by the fact that the leading non-causal approach to laws fails
to adequately explain the phenomenon that is the focus of this section.
Let’s see why by focusing on Robert DiSalle’s best-systems treatment of the
phenomenon.
Section 4.5: A Mill–Ramsey–Lewis Best-Systems Treatment?

DiSalle maintained that space-time structure ought not be regarded as a
cause of inertial motion (DiSalle 1995, 323–328). Rather, there exist
non-causal laws of the geometric structure of space-time, which state flatly
that free particles follow geodesics. He said,
When we say that a free particle follows, while a particle experiencing a

force deviates from, a geodesic of spacetime, we are not explaining the
cause of the difference between the two states or explaining “relative
to what” such a difference holds. Instead, we are giving the physical
definition of a spacetime geodesic. To say that spacetime has the affine
structure thus defined is not to postulate some hidden entity to explain
the appearances, but rather to say that empirical facts support a system
of physical laws that incorporates such a definition.50
DiSalle’s take on geodesic motion in relativity is comfortably situated amidst

a Mill–Ramsey–Lewis best-systems account (BSA) of laws of nature.51 This
is because what matters for DiSalle is whether the physical theory in ques-
tion accounts for observations. Puzzlement is removed once one realizes the
truth of the definition of geodesic motion and the fact that the system in
which that definition is incorporated is sufficiently explanatorily powerful,
although not at the cost of simplicity. Recall that the BSA affirms that laws
of nature constitute a system of propositions that best explains true empiri-
cal generalizations in a way that is balanced with simplicity. Ergo, DiSalle’s
account seems to assume something like the BSA, and so the fate of that
account is tied to the plausibility of the BSA or something near enough. This
is bad. Independent problems for the BSA abound,52 and if that account is
for reasons laying outside of GTR false, it should not be invoked here.
Let us set to one side the independent objections to the BSA and ask
what it is specifically about GTR that spells trouble for the BSA. There is an
interactive-dynamical explanation of “the appearances” with which DiSalle
is concerned. The inertial motion to be explained by the laws, and described
by the GP, is not merely “a free particle” following “a geodesic of space-
time.” Rather it is a state involving a gravitating free particle following
a geodesic. DiSalle’s account fails because it mischaracterizes the geodesic
principle, and geodesic motion. It is interesting that so many contemporary
statements of the principle remove the language original to it. That language
connects geodesic motion to the gravitational field. As I noted in chapter 3,
Einstein’s original statement of the principle (what he called the “law of
motion”) included just such a qualification, for he said that the principle
“asserts that a gravitating particle moves in a geodesic line.”53 Thus, DiS-
alle’s best-systems treatment of the laws about the phenomenon of geodesic
motion fails to explain what is at issue, the interaction of field and particle.
Section 4.6: Other Treatments?

Perhaps the relation between gravitational field and matter fields or particles
is one of supervenience. The matter fields supervene upon the gravitational
field. Of course, more needs to be said, as there are plenty of supervenience
relations to go around (e.g., weak, strong, global, mereological, nomologi-
cal). The most natural choice in this context is nomological supervenience.
J property instances nomologically supervene on K property instances, just
in case, in the galaxy of naturally possible worlds, J property instances are
nomologically necessitated by K property instances. At any world in the
relevant galaxy, at the relevant times and locations, a distribution of K prop-
erty instances (if they exist) ensures a distribution of J property instances
in a manner that holds with a kind of natural nomological necessity. Or,
to put it another way, K-facts “about a situation naturally necessitate the”
J-facts.54 The idea is that wherever you look out and see in the actual world
or (with a modal telescope) in some naturally possible world a certain set
arrangement of K property instances you will see the self-same partnering
arrangement of J property instances.
Given that nomological supervenience is one’s choice relation, the objec-
tion from supervenience will beg the question in precisely the same way the
invocation of nomological determination did, for as was just explicated,
the standard way to characterize nomological supervenience is in terms of
nomological determination or necessitation.55
One might forsake supervenience and instead recommend one of the
other metaphysically significant relations such as composition, constitu-
tion, or realization. Unfortunately, these are, for formal reasons, the wrong
relations in this context. Composition, for one thing, is a many-one rela-
tion. It is for that, and other obvious reasons disqualified. Constitution is
a relation between collocated massive bodies. The gravitational field is not
a (traditional finitely) massive body and is not collocated with any massive
bodies in cases involving photons (photons are massless). Realization of
the kind I favor (see Gillett 2003) is, like composition, a many-one relation
between property-instances “of the parts of a whole, and the properties [or
property-instances] of the whole” (modifying K. Bennett’s 2017, 11, sum-
mary discussion of Gillett), and that disqualifies it in the general relativistic
context too.
We are left with grounding. As I did in chapter 1: sect. 4.5.3, I help myself
to Jonathan Schaffer’s conception (Schaffer 2009). According to that view, an
entity x grounds another entity y, just in case, y depends for its positive onto-
logical status (its existence) and nature on x (where the dependence is asym-
metric, transitive, and well-founded) (Schaffer seems to take back transitivity
in (Schaffer, Transitivity 2012, 138), although something like it (differen-
tial transitivity) may be preserved by a contrastive approach to grounding).
I confess that I do not see a reason for distinguishing between causation
and grounding in this context. For although Schaffer contrasts grounding
priority with causal priority, musing that when x is (in the grounding sense)
prior to y, it is not causally prior to y (Schaffer 2010, 345), if that sugges-
tion is interpreted as one which claims that causation and grounding can
never coincide (which can mean to correspond in nature), the point is made
in haste. Surely when (leaning on examples discussed in Schaffer, Trumping
Preemption 2000, 165–166) Merlin casts a spell that transforms a prince into
a toad at t, he causes a new entity to come into being at t. We can say of the
toad that it depends for its nature and positive ontological status upon Mer-
lin’s spell-casting event. It seems then that causal priority can coincide with
grounding priority. Call such coincident instances of causation and grounding
causal grounding (for more on this motif, see chapter 9: sect. 5.1).56
Return to the case involving the gravitational field and matter fields or
particles. Just as in the case of Merlin and spell casting, here, too, I believe
there is causal grounding. If the field’s properties cause a particle to have
a property (change its trajectory, etc.) at some time, or at some space-time
location, then the latter event will depend for its existence and content on
the former event as predicted by minimal fundamentalist causation. Schaf-
fer allows for entities from any ontological category to stand in grounding
relations.
Section 5: Domains of Causal Influence in Cosmology:

Causal Pasts and Causal Futures
I have argued that a proper understanding of the gravitational field in GTR
implies that causation is indispensable to the best interpretation of GTR’s
formalism. There is, however, another sense in which causation enters suc-
cessful physics and that is by way of a proper understanding of the for-
malism of the standard cosmological model (which of course subsumes the
formalism of GTR; see Weinberg 2008, 1–100; 511–529, for the details).
It is well-known that on SC (standard cosmology), space-time points
induce double light cone structure as they do in the Minkowski space-time of
STR. But there is some confusion about that type of structure in GTR. Some
physicists use the locution ‘light cone’ to refer to a set of vectors associated
with a space-time point p, where those vectors live in tangent space Vp. Call
these structures null cones. There is another “light cone” structure surround-
ing space-time point p that is induced by “null geodesics from p” (Wald 1984,
189, n. 1). I will call this latter type of structure “light cone structure.” Neither
can be identified with causal structure in GTR, where that structure is what
(given our previous assumptions) the standard cosmological model associates
with space-time points. That type of structure is sometimes called “domains
of influence.” Future and past domains of causal influence are represented by
the locutions ‘J+(p)’ and ‘J-(p)’, where p is a space-time point that is a member
of M. J+(p) represents p’s causal future, whereas J-(p) represents p’s causal
past. J+(p) is that “region of space-time which can be causally affected by
events in” p.57 “Physically,” (quoting Robert Geroch), J-(p) “represents the
collection of all events of space-time which can affect what happens at p.”58
Quoting scholars who use causal locutions is, of course, not enough for
my purposes. But appreciate the further fact, again, that GTR does not
reduce or identify domain of influence structure to/with null or light cone
structure (e.g., J. Norton 2015, 211, denies such identification). First, it is
in fact well-known that gravitational lensing precludes the identification of
causal structure with light cones.59 Second, null cone structure induced by
vertex space-time point p lives in tangent space Vp, whereas domains of
influence J+(p), and J-(p) are both subsets of the manifold M itself. They are
not in Vp. How can null cones be identical to causal futures and pasts if the
two entities live in different spaces? Third, against identification of causal
structure and light cone structure, Hawking and Ellis wrote, “in Minkowski
space, the . . . [light] cone of p forms the boundary of the causal and chrono-
logical futures of p. However, in more complicated space-times this is not
necessarily the case (e.g. see Figure 34)” (Hawking and Ellis 1973, 184;
emphasis mine). Figure 34 references a Minkowski space-time in which a
space-time point has been removed.60
What my argumentation shows is that relativity scholars use explicitly
causal notions in the ideology of GTR. They give attention to questions
about what can causally influence what when using GTR to theorize about
the structure of the cosmos. It is therefore not true that all approximately
true physical theories have no need of causation because relativity scholars
need causation to carry on with the business of relativity scholarship.
Section 6: Objections to the Causal Interpretation of GTR

Causation enters physics by way of the gravitational field and by way of
domains of influence. The argument from physics is therefore unsound if
the reasoning of sects. 4 and 5 is cogent. I will now fend off objections to
my case for causal GTR.
Section 6.1: GTR Is Not Fundamental

Carl Hoefer argues that if GTR implies that there are certain obtaining causal
relations, or if its best interpretation makes use of causal notions, the causal
reductionist should not be worried, for GTR is not itself a fundamental phys-
ical theory (Hoefer 2009). GTR’s picture of the world is not the quantum
mechanical picture of the world. GTR will have to yield to QM in ways that
would rub out any attempt to understand the causal activity of the gravita-
tional field as fundamental physical activity. Look back to my characteriza-
tion of the argument from physics for causal reductionism. Notice that one
of the premises of the argument states that physics can proceed without cau-
sation making use instead of non-causal natural nomicity and history solely.
That premise is not restricted to fundamental physics. GTR is an approxi-
mately true and extremely successful physical theory, and that is precisely
why any quantum theory of gravity must recover its predictive success. Thus,
Hoefer’s complaint should not worry the fundamentalist about causation.
It is true that both string theory and loop quantum gravity proponents
maintain that GTR should be recoverable within the framework of such
quantum theories of gravity in the classical limit. However, there are theo-
ries of quantum gravity (QG) that do not seek for such subsumption. The
correct theory of QG may be one that is more fundamental than both
QM and GTR. Lucian Hardy’s causaloid approach to quantum gravity is
like this. According to that approach, GTR and QM are “special cases”
(Hardy 2007, 3082). Important to Hardy’s theory however is indetermi-
nate fundamental dynamical causal structure.61 What is more, there are
other takes on QG that promote causal structure to fundamental status.
For example, causal set approaches to quantum gravity are approaches
that, according to Dean Rickles, treat “the causal structure of space-time
as fundamental.”62 Furthermore, Aron C. Wall has recently proposed an
explicitly causal theory of quantum gravity (Wall 2013). Hoefer’s objec-
tion may therefore rest upon more than one false assumption. The correct
framework for a truly quantum theory of gravity is far from settled. The
current status of QG studies suggests that any case for the claim “quantum
gravitational physics does not need causation” is at best uncertain and
incomplete.
Section 6.2: Time-Reversal Invariance and Closed Timelike Curves

The causal reductionist may still object: The dynamical laws of GTR are
time-reversal invariant. Therefore, any causal reading of those equations
will imply the negation of the principle that necessarily, causes always tem-
porally precede their effects. But surely that principle is true!
The principle that necessarily, causes always temporally precede their
effects is in fact false. I have already shown why that is the case in chap-
ter 3: sect. 2. But the reductionist will continue to push by noting that GR
does not preclude space-times with closed timelike curves (CTCs) or closed
causal curves (CCCs). If GR holds, then space-times with CTCs are natu-
rally nomologically possible. But it is well known that if space-times with
CTCs are nomologically possible, then time travel is nomologically possible.
If, however, time travel is nomologically possible, then (given the natural
necessary falsehood of causal eliminativism) backward causation is natu-
rally possible. Minimal fundamentalist causation is transitive and asymmet-
ric. That entails that it is also irreflexive. But as was noted in chapter 6: sect.
2.3, backward causation of the kind that involves a body acting at a time
t, exerting an influence on itself via a causal loop that exploits the exotic
causal structure of a space-time violates irreflexivity (given transitivity). The
action and influence in GTR must therefore be non-causal.
I have a two-part response. First, the overlapping and intersecting causal
paths provided by CTCs and CCCs are the real cause for worry, not nec-
essarily the phenomenon of backward causation, for there is a defensible
position on the matter of CTCs in GTR that preserves the metaphysical
possibility of the latter peculiarity while unproblematically abandoning the
metaphysical possibility of the former. Here is the idea (previously sketched
in chapter 4, n. 2):
Backward Causation Without Violation of Irreflexivity: The trajecto-
ries of particles tracing out parts of exotic causal structure can be near
loops. One need only invoke a space-time in which a space-time point
is removed, as in space-times that violate the strong causality condi-
tion (for which see Wald 1984, 196–197). In such a physically possible
world, “there” are “no closed causal curves . . . but there are causal
curves through” space-time point p, say, “which come ‘arbitrarily close’
to intersecting themselves” (ibid., 197). These space-times represent
nomological possibilities in which there is genuine backward causa-
tion but no violation of irreflexivity. This is because the involved causal
paths do not completely loop back onto themselves thereby creating
CCCs or CTCs, although the relevant causal structure does allow for
past influence.
The reductionist will ask, is it not unscientific to reject as out of hand

or impossible those space-times that incorporate CTCs on the basis of an
unscientific principle about causation’s formal properties? After all, there
are solutions to the EFEs that feature such structure.
Morris, Thorne, and Yurtsever (MTY) argued that it is nomologically
possible that there are wormhole exploiting time machines (MTY 1988).
Although their study concerned not just general relativity, but also quan-
tum field theory, that fact will not rob my illustrative point of all power.
So, set it to one side. The history of reactions to MTY is important for
acquiring insight into how relativity scholars were thinking about the laws
of physics. That history is told well by Kip Thorne, who notes how Joseph
Polchinski objected to MTY’s claim that time travel through wormholes is
not paradox-free by envisioning a scenario in which a billiard ball b enters
a wormhole w1 at some time t, exits a suitably connected wormhole w2 at
an earlier time t-3 in such a way that it prevents b from ever entering w1
by colliding with it (Thorne 1994, 508–521). The situation is manifestly
paradoxical. If b is prevented from ever entering w1, it cannot venture back
in time so as to preclude itself from entering w1, although b did enter w1.
What was inferred from this scenario and from the nature of CTCs and
time travel in semi-classical contexts (i.e., contexts involving quantized or
classical fields, and classical gravitation as in GTR) was that the laws of
physics ought to be supplemented with self-consistency principles like the
following:
(Self-Consistency Principle (SCP)): If the universe is described by a local

solution of the mathematical formalism of the physical laws, then that
description is extendable in such a way that it can be a constituent “of
a global solution one which is well defined throughout the (nonsingular
regions of) classical spacetime.”63
The SCP is a principle that qualifies the laws of physics. Something like
it was introduced (by Friedman et al. 1990) to ensure non-paradoxical
evolutions, although those evolutions may very well be consistent with solu-
tions to the formalism of the underlying physical laws. Thus, it seems that
SCP’s modal strength should at least be as strong as that of nomological
necessity so as to ensure that the laws of the theory that we use to correctly
judge some evolution as nomologically possible or impossible do not let in
paradoxes. And although John Earman does not think much of time travel
paradoxes, he (Earman 1995, 161, 176–177) argues that consistency con-
straints ought to be understood as bona fide laws of nature and that physi-
cal possibilities are those possibilities consistent with what are ordinarily
considered laws of physics plus consistency principles.64
I conclude then that it is not at all unscientific to introduce a nomologi-
cally necessary ancillary principle that restricts the space of nomological
possibilities to those that are also metaphysically and/or logically possible.
Notice that my case for this conclusion remains cogent even if the supposed
paradoxes can be resolved.65 My conclusion here is that there is nothing
unscientific about the strategy, and my justification is that physicists and
very scientifically informed philosophers of physics have employed just such
a strategy in scientific and philosophical practice.
The question remains. In general relativity, does one encounter paradoxes or
impossibilities if one allows for CCCs and CTCs? Because I’m arguing that one
should motivate principles that preclude CTCs by appeal to paradoxes implied
by them, it will be important to articulate an actual paradox. So here it is.
My articulation of the paradox will assume that at any world at which
there exist entities that evolve in accordance with the laws of general rela-
tivity, causal eliminativism is false. But I also assume that causal reduction-
ism is true, and that my use of the locution ‘causeR’ picks out reductionist
causation. Thus, my argument for the paradox begs no questions against
causal reductionism. I will also need the posit that causation is transitive
(see chapter 6: sect. 3).
(1) If GTR without constraining auxiliary principles is true, then it is physi-

cally possible that there are closed timelike curves that allow for causalR
loops.
(2) If it is physically possible that there are closed timelike curves that allow
for causalR loops, then an event can causeR itself.
CTCs that allow for causalR loops, allow for a causalR chain that loops
back onto itself. If an event A causesR B, and B causesR C, and C causesR A,
by transitivity (which we are assuming), A will causeR itself. Thus, if there
can be CTCs that allow for causalR loops, then an event (A) can causeR itself.
(3) If an event can causeR itself, then any event can causeR itself.
I argued in chapter 6: sect. 2.2 that once we preclude asymmetry or irre-

flexivity, and allow for the possibility of some self-causation, it becomes
exceedingly difficult to craft a not obviously ad hoc theory of causation
according to which it is not the case that every event (actually) causes itself.
That every event causes itself strikes me as utterly paradoxical and unto
ward. Causation, whether causationR or not, cannot be a relation that holds
between every event and itself (I object to the possibility of self-causation
in chapter 6: sect. 2.2). Thus, we ought to slim down the space of general
relativistic nomological possibilities by adding to the theory, a metaphysi-
cal prohibition on self-causal influence even if that causation is reductively
understood. So,
(4) It is not the case that any event can causeR itself.
(5) Therefore, it is not physically possible that there are closed timelike
curves that allow for causalR loops, and it is not the case that GTR
without constraining auxiliary principles is true.66
Thus, the proponent of C-GTR should not worry about CTCs.
Section 6.3: Back-Reaction and Causal Asymmetry

John A. Wheeler remarked, “space-time tells matter how to move; mat-
ter tells space-time how to curve.”67 If GTR is interpreted causally, then
Wheeler’s remark will recommend that the gravitational field’s behaving
y-ly causes a material body to behave x-ly, whereas the material body’s
behaving x-ly causes the field to behave y-ly. But causation is formally asym-
metric. We should enforce a ban on instances of symmetric causation and
so on causal GTR.
The gravitational field’s dynamical action is primary and causally prior
to the inertial motion of massive bodies. There are four reasons that consti-
tute a cumulative case for such a thesis. (a) It is the gravitational field that
can enjoy positive ontological status in the absence of matter, and not vice
versa.68 (b) It is the gravitational field that is fundamental to GTR and not
the matter fields (q.v., note 47). (c) It is the gravitational field that can bring
about an increase or decrease in the energy-momentum of massive bodies
without itself having any localized energy-momentum density, and not con-
versely.69 And (d) it is the gravitational field’s direct coupling with matter
fields that “results in the latter having an energy tensor” and not contrari-
wise.70 The best explanation of facts (a)–(d) is that the primary causal mover
in the context of the inertial motion of massive bodies is the gravitational
field and not the matter fields of GTR.
Dennis Lehmkuhl would challenge my appropriation of (d) (and perhaps
also the quotation of him for present purposes) (Lehmkuhl 2011, 469).71
A dynamical and causal understanding of (d) is, perhaps, for Lehmkuhl,
incompatible with the fact that tensors of the energy-momentum variety
are defined in special relativistic contexts, too. In STR, the metric ημv isn’t a
dynamical entity.
On the contrary, there are good reasons for understanding the geometric
structure of Minkowski space-time in STR as dynamical. That is to say,
there are good reasons for regarding the geometry of Minkowski space-
time as an indispensable entity in explanations of both the forms of forces
acting on objects (their Lorentz invariance), and the behavior of those
objects.72 Indeed, Einstein went further by insisting that Minkowski space-
time acts.73
But a proper response to Lehmkuhl need not go so far. We can instead
invoke (non-causal) grounding between facts. Causal facts reporting on the
relevant obtaining causal relations are the base, the direct coupling facts
(which themselves ground the definitional dependences) are the grounded
entities.74 We need not believe that the grounds of an entity are essential to
it. Thus, the metaphysical explanation of definitional dependence of energy-
momentum tensor on metric field is, by transitivity, grounded in the fact
that the gravitational field causally interacts with the relevant matter fields.
At purely STR-worlds, some other ground exists (e.g., conformal features
of Minkowski space-time) for the direct coupling and definitional depen-
dency facts.
Why invoke causation if the direct coupling relations are enough to
ground definitional dependency between metric and energy-momentum ten-
sor? The reason lies in what direct coupling amounts to in Lagrangian theo-
ries. We say that two (quoting Lehmkuhl) “fields . . . directly couple if they
are factors of the same product term in the Lagrangian, which then gives
coupled differential equations as field equations via the Euler–Lagrange
equations.”75 The condition for coupling is couched in terms of a math-
ematical fact. A Lagrangian is a scalar quantity. Product terms and factors
are mathematical entities. What do such mathematical facts suggest about
the world? Here I believe is where the causal interaction of the fields has
some indispensable grounding role to play.76 Interestingly, the metaphysical
explanation of the direct coupling facts enters the interpretation of those
facts. Thus, if we impugn attributing action to Minkowski space-time, we
can still interpret the Lagrangian formulation of GTR causally, and hold on
to fact (d) in the above inference to the best explanation argument.
Section 6.4: Causal GTR Presupposes Substantivalism

One might judge that my argumentation essentially requires the truth of
space-time substantivalism.77 I do not think such a position is an essential
presupposition of my argumentation, but even if it were, I would not be
worried, for substantivalism is the default position in the context of GTR.78
Still, space-time relationalism will not provide an escape from my reasoning.
Instead of reducing the gravitational field to space-time structure, thereby
attributing dynamical activity to space-time itself, one could do without
an ontological commitment to space-time understood as an independent
substance and treat the metric or gravitational field like a matter field and
(perhaps) account for all of the empirical success ordinarily associated with
a more orthodox GTR.79 I have already mentioned this option in the context
of discussing the work of H. R. Brown (2005; 2009) and Rovelli (1997) in
chapter 3: note 69. Belief in the causal action of a matter-like metric or grav-
itational field does not necessarily commit one to space-time substantivalism
as I have defined it. Although Rovelli (2004, 78) maintains that the debate
between substantivalists and relationalists in the context of GTR is a matter
of “semantics,”80 I believe there remains a substantial dispute. Does there
exist an independent substance that is space-time itself? Does the gravita-
tional field reduce to space-time geometry? It seems to me that both H. R.
Brown (2005, 156) and Rovelli (2004, 77) answer no to both questions. For
them, there is just the gravitational or metric field, and my point here is that
their position is at least consistent with relationalism. Ergo, substantivalism
is no essential presupposition of my reasoning.
In point of fact, the matter field-like approach to gravitation only strength-
ens the case for the causal interpretation of GTR. As we’ve noted before,
the gravitational field can come to possess ripples understood as gravita-
tional waves or gravitational radiation. Gravitational waves propagate, and
they are similar to electromagnetic waves in that they exhibit similar causal
behavior on the matter field-like approach.81 We can therefore avoid having
to commit to space-time substantivalism and actually bolster the case for
C-GTR if we appropriate one of the best relationalist-friendly approaches to
the metric field or gravitational field in general relativistic contexts.
Section 6.5: The Causal Talk Is Gloss

The reductionist may suggest that all of the causal talk I’ve referenced can
be removed without loss of explanatory power. Explicitly causal interpre-
tations of all of the above are therefore problematic. I reply: Domains of
causal influence help determine the global causal structure of space-time.
Without such causal structure one cannot derive the space-time singularity
theorems that are necessary for describing and explaining features of the
beginning of the universe (see Wald 1984, 188; 237–242, who calls such
domains or causal structure a “crucial ingredient” in the proofs of the sin-
gularity theorems (ibid., 188)).
Perhaps the reason why a reductionist would insist that the causal talk of
physicists in general relativistic contexts is gloss is because one can provide
a local reduction of causation and causal structure in cosmology and GTR.
After all, the problems with many reductive accounts of causation rely on
very unique and artificial cases of preemption, overdetermination, and the
like. Do such cases arise in GTR? The reductionist will bet that they do
not. Thus, counterfactual dependence (or some similar reductive surrogate
notion) may serve as a worthy proxy for general relativistic causation and
causal structure even if it cannot serve as a worthy proxy for causation
wherever it is found in the actual world or in broadly logical space.
Local reduction strategies will not work. One can, on paper or with the
mind’s eye, craft general relativistic worlds at which cases of overdetermi-
nation, preemption, and the like occur, although these cases involve mat-
ter fields and the gravitational field, or gravitational waves, or certain tidal
forces (ignoring for now the complexities of causal structure highlighted in
chapter 6). It is particularly easy to get such cases because our world is a
general relativistic world. Drop two rocks that are virtually identical at the
same time from the exact same height on the Earth over a rectangular sheet
of glass placed just one meter off the ground, and set up the glass in such a
way that one of its longest and widest sides is parallel to the ground. If the
rocks are dropped high enough, and if the glass is thin enough, each rock
will, in a way involving causal overdetermination, bring about the shattering
of the glass (hold fixed what you need to in the imagined case). The example
involves the gravitational field induced by the Earth, and the appropriate
matter fields. The example is general relativistic in that it is said to be actual,
and the case is one of symmetric overdetermination. This is the same brand
of overdetermination that is problematic for counterfactual dependence
theories of causation. Thus, general relativistic nomological possibilities
and actualities recommend an incompatibility between the local reductive
theory of causation and GTR itself. A fortiori, the problems with reductive
theories of causation are not all revealed in counter-examples or difficult
cases. There are many other problems with reductive accounts.82
Still causal reductionists will insist that my appeals to GTR and cos-
mology are cheap and shallow. They will urge that the authorities I have
invoked are merely describing matters with a particular gloss. Replace the
occurrence of the term ‘cause’ or ‘causal’ with some other notion, and all
will be well for the argument from physics. But what notion do we replace
it with? The phenomenon described is interactive, the relation that obtains
between matter fields and gravitational field is one of asymmetric influ-
ence that is present when gravitational waves knock down a mountain (to
repeat Rovelli’s illustrative example discussed in chapter 3). The goings-on
described are causal by even reductionists lights. It seems therefore safe
to conclude that in the absence of a truly successful reductive analysis or
theory of causation, we have no reason to dispense with causal notions in
the primitive ideology of GTR.
Section 7: Concluding Musings

Minimal fundamental causation is indispensable to the best interpreta-
tion of GTR. No other distinctive metaphysical relation is suitable for the
interpretive work of standing in for the interactive relationship between the
gravitational field and matter fields in GR, and non-causal approaches to
the laws of GR miss out on structure. We have therefore refuted the second
premise of the argument from physics for causal reductionism. I now turn
to a more complete explication of what I believe is the correct philosophical
analysis of fundamentalist causation.
Notes
1. On transference theories generally, see the discussion in Dowe (2000, 41–61).
On counterfactual theories generally, see Paul (2009).
2. See my presentation of the view in chapter 1: sect. 4.4.1 proposition (CAC).
3. For more discussion of these types of cases, see Paul and Hall (2013) and my
handling of many of these cases in chapter 9: sect. 5.
4. See the exchange between Dowe (2004) and Schaffer (2004) for a debate about
the problem.
5. I’m following Penrose (1986). And so relativity scholars have said,
. . . the gravitational energy is necessarily a non-local one.
(Penrose 1986, 137)
In general relativity there is no known meaningful notion of local energy
density of the gravitational field. The basic reason for this is closely related to
the fact that the spacetime metric, gab, describes both the background space
time structure and the dynamical aspects of the gravitational field, but no
natural way is known to decompose it into its ‘background’ and “dynamical”
parts . . . for an isolated system, the total energy can be defined by examining
the gravitational field at large distances from the system.
(Wald 1984, 84)
. . . there is no meaningful notion of the local stress-energy of the gravita-
tional field in general relativity.
(Wald 1984, 85, see also page 286)
6. Wald (1984, 287; emphasis in the original).
7. For an explication of a path to the relevant notion, see Wald (1984, 287–288).
Notice that it uses a Killing vector field.
8. Wald (1984, 286). Kriele (1999, 205) remarked,
Clearly, only very special pseudo-Riemannian manifolds can have non-zero
killing vector fields. A simple example is given by a metric which does not
depend on one of the coordinates.
(ibid.)
It seems that Dowe’s specific conserved quantity account is actually better
suited for special and not general relativity. Indeed, Dowe’s definition of world-
lines appropriates space-time points on the special relativistic space-time that
is Minkowskian. In Minkowski space-time and STR, you don’t have to worry
about gravitational waves, and one automatically gets a Killing vector so that one
can have a global energy conservation law. As Narlikar and Padmanabhan wrote,
“[t]he Minkowski time coordinate assures the existence of a global timelike kill-
ing vector field in flat spacetime” (Narlikar and Padmanabhan 1986, 284).
9. See S.M. Carroll (Energy is Not Conserved 2010) for a very helpful and acces-
sible discussion of the matter. Thanks to Tom Banks, Sean Carroll, and Don Page
for confirming the non-sociological point in correspondence.
10. Schaffer (Causation and Laws 2008, 92). The argument Schaffer develops is
explicitly one for causal reductionism and not causal eliminativism (cf. ibid.,
82–83; 86; 92). Recall that causal eliminativism is the view that there are no
obtaining causal relations.
Schaffer’s original argument appealed to scientific practice in general. It is
clear, however, that he privileges physics as that science that puts us in touch
with the fundamental. Thus, if physics has no need of causation, causal reduc-
tionism is well-evidenced.
For other arguments in favor of causal reductionism, see Norton (Causation
as Folk Science 2007; Causal Principles 2007). Cf. the recent critical discussion
in Frisch (2014, 1–21).
11. I should highlight some differences between the content of this chapter and my
much shorter case for causal GTR in chapter 3: sect. 2. Whereas chapter 3: sects.
2.2 and 2.3 provided a number of reasons for interpreting GTR causally, this
chapter provides a different set of reasons for that causal interpretation. More
specifically, it spends more time laying bare why we label GTR an interactive
theory, and then makes much of the historical truth that interactive physical
theories in physics are ordinarily given causal interpretations. It then uses that
historical fact as a datum in an (adumbrated) IBE argument for a causal inter-
pretation of GTR. I then provide additional support for causal GTR by spelling
out how no other distinctively metaphysical relation is fit for the job of connect-
ing the inertio-gravitational field to the geodesic motions of relativistic particles.
This is not done in chapter 3.
12. See Beebee, Hitchcock, and Menzies (2009, 1); J. Carroll (2009); Paul and Hall
(2013, 249); Psillos (2009, 154); Schaffer (2007, 872–874); and Tooley (1987,
5, inter alios). Kutach (2013, 282–306) tries to show that there probably is no
“complete and consistent systematization of cause-effect relations ‘out there in
reality’ that correspond to our folk concept of causation” (ibid., 282), gesturing
at the well-known problems and counter-examples to unified theories.
13. Following some related ideas on ideologies of theories in Hellman and Thomp-
son (1977); Post (1987); and Quine (1951, 14–15).
14. See Arnowitt, Deser, and Misner (2008) for the republication of the original
1962 paper.
15. Ruetsche (2011, 7; emphasis in the original). Cf. (Saunders 2003, 290–291).
16. That close association has been challenged in the philosophy of physics literature
(see the discussion of the debate in Lehmkuhl 2008 and Rey 2013, and the lit-
erature cited therein).
The view I adopt is similar to Lehmkuhl’s “Candidate 3: the metric—and all
that it determines” (Lehmkuhl 2008, 97).
17. See Norton (1989, 40), who said that “[i]f we are to call any structure ‘gravi-
tational field’ in relativity theory, then it should be the metric”; Rindler (2006,
179); Rovelli (1997, 194); Cf. Torretti (1983, 245–246).
18. Rovelli’s remarks assume that the partial interpretation is part of Einstein’s
metrical approach to GTR studies even if that’s not the case in quantum grav-
ity studies (Rovelli 2004, 33–34). Norton (1989, 40) says that Einstein “calls
the metric tensor the gravitational field.” See Einstein (1949, 685; 1952, 144;
2002, 33).
19. To repeat some citations from note 69 in chapter 3: S.M. Carroll (2004, 50);
Choquet-Bruhat (2009, 39); and Geroch (2013, 65). Hartle (2003, 13) says
“the central idea of general relativity is that gravity arises from the curvature
of spacetime—the four-dimensional union of space and time. Gravity is geom-
etry” (emphasis in the original); Wald (1984, 9, cf. 68). Weinberg (1972, vii)
clearly suggests that both (a) the geometrization of gravitation and (b) the
metric tensor represents the gravitational field are standard views (although he
seems dissatisfied with at least (a)). Weinberg would go on to say that Einstein
adhered to both (a) and (b). See also Rovelli (2004, 77). Cf. Lehmkuhl (2014).
20. Sciences (2011, 2).
21. I’m leaning slightly on Ruetsche (2011, 7).
22. Disagreeing with Ruetsche (2011, 5–12). There is more to Ruetsche’s account.
Interpretations also involve the dynamics, kinematics, and observables of the
theory (ibid.).
23. Ruetsche (2011, 352).
24. Ibid., 12. The reason why it should show such respect is because Ruetsche’s view
of the criteria for interpretive adequacy suggests as much.
25. Ibid.
26. van Fraassen (2008, 309). van Fraassen notes that “[a] theory may . . . be taken
to represent its domain as thus or so in the sense that the models it makes avail-
able for the representation of phenomena in that domain are thus and so” (ibid.).
27. See my discussion of the view in chapter 1: sect 4.5.5.1.
28. S. M. Carroll (2004, 50); Janssen (2012, 160); Norton (1989, 40–41); Penrose
(2005, 459). See also Einstein (2002, 33).
29. See S.M. Carroll (2004, 155–159) and Wald (1984, 66–74) for clear expositions
of the meaning of Eq. 1.
30. Einstein (2002, 339). See the discussions of geodesic equations of motion in S.M.
Carroll (2004, 106–113); Weinberg (1972, 70–73; 121–129); and (Zee 2013,
301–311).
31. Lambourne (2010, 137). See also Foster and Nightingale (2005, 4); Rindler (2006,
178), who says that GTR predicts “the paths of light in [a] vacuum under the influ-
ence of gravity”; and Romero, Fonseca-Neto, and Pucheu (2011, 31). Compare
Bergmann (1961, 510), “The classical field theorist’s way of stating the law” is “that
under the influence of gravity all material bodies suffer the same acceleration” (ibid.).
32. See Eddington (2014, 125–127; 149–170); Einstein and Gromer (1927); Ein-
stein, Infeld, and Hoffmann (1938); Einstein and Infeld (1940); Einstein (1949);
Fock (1959, 215–218); Geroch and Jang (1975); and Infeld and Schild (1949).
The history of attempts to derive the law of motion for particles from the EFEs
receives careful attention in Havas (1989; 1993).
33. For worries about the success and importance of such derivations, see Ehlers
(1987, 63–65); Tavakol and Zalaletdinov (1998, 312–314, 323, 325); and
Tamir (2012).
34. See Einstein (1954, 311) and the excerpts from Einstein and Grommer (1927)
quoted in Havas (1989, 240–241) and Tamir (2012, 141).
35. Einstein admits in a letter to Ludwik Silberstein that “a really complete theory
would exist only if the ‘matter’ could be represented in it by fields and without
singularities” (as quoted in Havas 1993, 102; emphasis in the original).
36. Havas (1989, 254). Cf. Torretti (1983, 177), who says,
While an exact general solution to the problem of motion in General Relativ-
ity is not yet known and may even be impossible, many authors have dealt
with it by approximation methods in a variety of cases.
(ibid.)
37. “The fact that geodesic motion is a theorem and not a postulate has striking con-
sequences that cannot be overemphasized. Earlier . . . I argued that it . . . casts
doubt on the widespread view that space-time structure, in and of itself, can act
directly on test bodies” (Brown 2005, 162).
38. Misner, Thorne, and Wheeler (1973, 476).
39. Geroch and Horowitz (1979, 212). In the perfect fluid case one clearly sees this
in the presence of the Lorentz metric tensor in the energy-momentum tensor of
the perfect fluid. Cf., (Rovelli 1999, 7).
40. Misner, Thorne, and Wheeler (1973, 476)
41. See also Nerlich (1976, 264).
42. As quoted and translated by Pais (1982, 465). With respect to Einstein and the
ultimate unified field theory, Pais goes on to remark,
He [Einstein] demanded that the theory shall be strictly causal, that it shall
unify gravitation and electromagnetism, that the particles of physics shall
emerge as special solutions of the general field equations.
(ibid.; emphasis mine)
Also see especially his discussion (with Infeld) of the elevator thought experi-
ment (Einstein and Infeld 1938, 226–235).
43. Robert M. Wald (personal correspondence, 12/18/2014). See also R. Geroch
(1978, 180); Geroch (2013, 2, 65, 68); cf. Misner, Thorne, and Wheeler
(1973, 476–477); and Nerlich (1994), although not clearly committed to
a causal interpretation of space-time’s action says, “[b]ut GR surely makes
spacetime something not easily distinguished from a real concrete entity with
causal powers” (ibid., 183). And Carl Hoefer says the causal interpretation
is commonly accepted (Hoefer 2009, 702). Brown describes the position that
spacetime in GTR acts as a view that is “widespread” (Brown 2005, 162). So
far as I’m aware, neither Hoefer nor Brown endorse the causal interpretation
I’m defending.
44. I’m ignoring the problem of specifying “choice” coordinates or a gauge.
45. More technically, for every time-like vector ξ µ associated with any point in the
differentiable manifold representing space-time, the following relation holds:
Tµvξ µξ v ≥ 0. It is assumed that the energy-momentum tensor is appropriately
related to the matter fields a hypothetical observer gives attention to. See the
discussion in Malament (2009, 6–7; Topics in Foundations of GR 2012, 144)
and Weatherall (2011).
46. Geroch (2013, 67); Lehmkuhl (2011, 470–474).
47. As Geroch noted, “[e]verywhere, we see the metric, directly or indirectly, in the
stress-energy. . . . It appears that it is simply impossible to make any reasonable
description of matter without the notions of space and time provided by the
metric” (Geroch 2013, 67). See also Hawking and Ellis (1973, 61); Malament
(Topics in the Foundations of GR 2012, 160); and Pooley (2013, 541, n. 38).
The dependence holds even for Lagrangian formulations (see Lehmkuhl 2011,
464–474). Why is this true? The answer lies in the interpretation of GTR on
offer.
48. Birnbacher and Hommen (2013, 144; emphasis in the original). The authors are
there concerned with a reductionist (the laws are non-causal) account of causa-
tion whose backbone is essentially the relation of nomological determination.
49. Glennan (2011, 811) points out that nomological determination is at home in a
causal approach to laws of nature (mentioning the well-known causal account
of David Armstrong inter alia).
50. DiSalle (1995, 327).
51. See D.K. Lewis (Counterfactuals 1973, 72–77; New Work, 1983, 365–368;
1994, 478–482) and Loewer (2012) for clear discussions or presentations of the
view.
52. See particularly the objections in Belot (2011); Hall (2015); and Maudlin (2007).
53. Einstein (1950, 109; emphasis mine). See also Geroch and Jang (1975, 65).
54. Chalmers (1996, 37), although Chalmers calls it natural supervenience.
55. Again see the characterization in Chalmers (1996, 37).
56. This point is similar to one that has recently been argued by Karen Bennett
(2017, 67–101, specifically the remarks at the top of 94). Bennett tries to show
that there are instances of non-causal building relations (asymmetric, genera-
tive, necessitation relations) that hold by virtue of certain causal relations. Ben-
nett maintains that obtaining causal relations are themselves building relations
too (see summary remark in ibid., 99). My point here is that causal relations
can coincide with, say, instances of grounding (where grounding is a building
relation). At the end of chapter 9, I will argue that these coincident instances
should just be understood as a special kind of causation. This point does fol-
low the spirit of Bennett’s thesis that causal relations are building relations.
I should add that Bennett does not seem interested in demonstrating that cau-
sation is a building relation. She seems interested in only showing that “it is
interesting that causation is a building relation” (ibid., 69; emphasis in the
original). I have the less modest goal in mind in chapter 9, although I’m inter-
ested in grounding and not the more general idea of building.
57. Hawking and Ellis (1973, 183; emphasis mine).
58. Geroch (2013, 123; emphasis mine). Geroch uses ‘I-(p)’ to represent p’s past
domain of influence. I am widening Geroch’s notion of a domain of influence a bit
because even he admits that the term can be misleading because I-(p) (having to
do with the narrower conception of a past domain of influence) “does not include
its boundary, whereas points on the boundary can, in general, also influence p”
(ibid., 123). I am deliberately intending to include the boundary. In so doing, I fol-
low Bhattacharyya et al. (2016) in my identification of domains of influence with
causal pasts and futures. Cf. Manchak (2013, 590); Ellis and Stoeger (2009).
59. Gravitational lensing involving matter or associated caustics can produce a col-
lapse or folding in of the light cone induced by a vertex point p in the manifold
(Ellis, Bassett, and Dunsby (1998, 2346–2347)); Schutz says “. . . even a small
amount of matter in spacetime will distort light-cones enough to make them fold
over on themselves” (2003, 336). That folding can entail that part of the (past
or future) cone bends inside of the causal past or future of p (Perlick (2004);
Tavakol and Ellis (1999, 41), who include further references on this point).
The past/future light cone and the causal past/future of p cannot therefore be
identical.
See the excellent introduction to gravitational lensing in Dodelson (2017).
60. The mistake of identifying causal structure in relativity with light cone structure
is often committed by philosophers (see e.g., Frisch 2014, 16–17; Field 2003,
436, who comes close to suggesting such identification). Causal or influence
structure is standardly regarded as more fundamental than light cone structure
in GTR (see Geroch 2013, 125).
61. Hardy (2007, 3084–3085).
62. Rickles (Quantum Gravity, 2008, 347, n. 124; emphasis mine). The quote’s
immediate context is about Robb’s formulation of STR, but it’s clear from con-
text that Rickles is also intending to characterize causal set approaches to quan-
tum gravity.
63. Friedman et al. (1990, 1915).
64. Cf. Hawking (1992).
65. Thus, one can endorse the conclusions and resolutions of time travel paradoxes
in Earman, Smeenk, and Wüthrich (2009, 93–100) and yet hold on to the sup-
position that the introduction of principles like SCP is not unscientific.
66. The same argument will run with minimal fundamentalist causation in mind.
67. As quoted by Wheeler (1998, 235).
68. Cf. Einstein (1952, 144).
69. See Misner, Thorne, and Wheeler (1973, 466–468) and Rueger (1998, 34).
70. Lehmkuhl (2011, 469). Lehmkuhl is explicitly concerned there with the Lagran
gian formulation (for which see S.M. Carroll 2004, 159–165).
71. Lehmkuhl does not deny that there is interaction between metric field and matter
in GTR. He said that “the matter fields and the metric field gμν are interacting in
GR” (Lehmkuhl, 2011, 469; emphasis in the original).
72. See the argumentation in Balashov and Janssen (2003, 339–342). Contra Brown
(2005); Brown and Pooley (2006); and Craig (2001).
73. See the quotation of Einstein in Brown and Pooley (2006, 68), and their
commentary.
74. Continuing to use the theory in Schaffer (2009).
75. Lehmkuhl (2011, 467; emphasis in the original).
76. It is important to point out that Lehmkuhl believes that the causal interaction of
fields is a sufficient condition for direct coupling (see Lehmkuhl 2011, 469). He,
too, sees room for a relationship between interaction and coupling.
77. See chapter 7: sect. 3.2 for a definition of space-time substantivalism.
On the basis of parsimony motivations, Brown (2005, 24–25) argues against
the view that space-time structure plays an explanatory role in accounting for
the motions of bodies. Cf. the discussions in DiSalle (1994, 276; 1995, 327).
78. See the comments in Arntzenius (2012, 17) and Nerlich (2003, 281). In general,
a realist approach to our most successful theories of physics suggests space-time
substantivalism (see the comments in Pooley 2013, 539 and Sklar 1976, 214;
cf. Weinstein (2001) who argues that substantivalism falls out of considerations
having to do with quantum physics).
79. I do have reservations about such an interpretation.
80. Although later in Rovelli (2004), he seems to think of his choice position as
relationalist. He said, “Thus, both GR and QM are characterized by a form of
relationalism” (ibid., 220).
81. Rovelli (1997, 193). See also Weyl (1952, 220). In Weyl’s ontology of physics,
he made room for a guidance field that appears to be causal (see Coleman and
Korté 2001, 198, and the sources cited therein).
82. See the discussions in Tooley (2003) and Rueger (1998, 33–36).
9 Fundamental Causation
Section 1: Causation After Reductionism

Throughout the current project, we have found that reductive theories and
analyses of the causal relation are problematic. Many theorists writing on
causation would agree. For example, reductionists L. A. Paul and Ned Hall
have recently concluded,
After surveying the literature in some depth, we conclude that, as yet,

there is no reasonably successful reduction of the causal relation. And
correspondingly, there is no reasonably successful conceptual analysis
of a philosophical causal concept. No extant approach seems able
to incorporate all of our desiderata for the causal relation, nor to
capture the wide range of our causal judgments and applications of
our causal concept. Barring a fundamental change in approach, the
prospects of a relatively simple, elegant and intuitively attractive, uni-
fied theory of causation, whether ontological reduction or conceptual
analysis, are dim.1
In chapter 8, I argued that one of the most popular ways of justifying causal
reductionism was through the argument from physics. That argument was
weighed and found wanting. Indeed, I showed how a very minimal anti-
reductive brand of causation is indispensable to the best interpretation of
one of our most successful theories of physics, viz., general relativity.
Given the soundness and cogency of the argumentation in preceding
chapters, I believe it is safe to conclude that if there is a correct philosophical
analysis of causation, it is an anti-reductive one. In this chapter (specifically
sects. 2 and 3), I give attention to anti-reductive analyses.2 I argue that two
of the most prominent anti-reductive theories of causation fail. In sect. 4,
I supplant them with what I believe is a better anti-reductive theory of cau-
sation, the full fundamentalist account (abbreviated ‘The Account’). I then
show how that account can overcome the so-called difficult cases involving
symmetric overdetermination, preemption, and other scenarios by entering
through David M. Armstrong’s open door. I conclude with some reflections
on how my fundamentalist/anti-reductive account of causation relates to
Jonathan Schaffer’s theory of grounding.
Section 2: Against the Interventionist

Manipulability Account
The interventionist manipulability account of token singular (or actual)
causation3 (IMA) in the work of James Woodward is hailed as one of the
best anti-reductive theories in the literature (see Woodward 2003, 25–151;
2015; 2016, sect. 5, building on work in Pearl (2000, 2009)). But my discus-
sion of IMA comes with a warning. IMA was not put forward by Wood-
ward with the intent to abide by the K-A of chapter 1: opening section. His
(quoting Woodward) “discussion” was “meant to be illustrative and sug-
gestive, rather than comprehensive and definitive.”4 That said, we can still
evaluate the theory as if it were proposed as a definitive account and note
its shortcomings.
In Woodward (2003), we are provided with two theories of token cau-
sation, a first-approximation account, and an end-game theory meant to
handle the difficult case of symmetric overdetermination. The first-approx-
imation account is prefaced with a directive to write down a causal graph
connecting causal relata X and Y in a way that respects a principle of direct
causation called (DC),
(DC): Relative to a set of variables V, X is a direct cause of Y, just in

case, at some possible world w, there’s an “intervention on X” that
wrought a change in Y “(or the probability distribution of Y) when all
other variables in V besides X and Y are held fixed at some value by
interventions.”5
According to Woodward (ibid., 42), the conditional <If X is a direct cause

of Y, then X is a cause (simpliciter) of Y.> (quoting the antecedent) holds,
although <It is not the case that, (if X is a cause (simpliciter) of Y, then X
is a direct cause of Y).> likewise holds. The end-game account does not
have this preface. I am not sure if this is an oversight or if the omission was
intended. Setting that matter aside for now, Woodward’s end-game theory
is as follows:
(IMA): The actual values of variables X and Y (x and y, respectively)

stand in a token singular causal relation directed from variable X to
variable Y, just in case, (a) X has x as its value, Y has y as its value, and
(b) *one determines that* there is a possible world w at which (bi) there
exists an avenue α (a directed pathway that extends through X to Y),
and (bii) there exists interventions on all direct causes of Y that are not
on α that fix the values of those direct causes of Y “at” one distinctive
“combination of values” that fall “within” a “redundancy range,”6 and
(biii) there exists an “intervention on X that will change the value of Y.”7
What is an intervention? As was noted in our discussion of the AP-A in
chapter 4: sect. 2, with respect to a causal relation involving X (the cause)
and Y (the effect), we would say that an intervention I (imagined as some-
thing that actually obtains) is surgical, in that it is the sole cause of X. I caus-
ally influences Y in such a way that no causes of it influence Y other than
those causes on a causal chain or pathway stretching through X. Interven-
tion I is precluded from influencing any causal relata that affect Y but that
exist outside of the avenue that extends from I through X to Y. I is statisti-
cally independent “of any variable Z that causes Y and that is on a directed
path [avenue] that does not go through X.”8 And lastly, intervention I must
causally influence Y via X.9
Although IMA accurately reports on the essential content of Woodward’s
theory, I believe it may not be the best characterization. This is because
condition (b) includes the starred phrase ‘ . . . one determines that. . . . ’10
The presence of that clause entails that whether there exists an instance of a
causal relation between the values of X and Y depends (at least in part) upon
whether one determines that a certain modal claim is true. That is to say,
one’s determining/verifying/coming to know a certain modal claim is a neces-
sary condition for the truth of the statement <x caused y.>.11 IMA clearly con-
fuses modal causal epistemology with causal ontology or metaphysics. There
is no guarantee that the causal structure of the world is perfectly fit for our
cognitive gloves. To assume otherwise is to special plead. Why couldn’t there
actually exist causal structure that we cannot know about because the struc-
ture in question is beyond our ken? I therefore propose the following tweak:
(IMA*-P): The actual values of variables X and Y (x and y, respectively)

stand in a token singular causal relation directed from variable X to
variable Y, just in case, (a) X has x as its value, Y has y as its value, and
(b) there is a possible world w at which (bi) there exists an avenue α (a
directed pathway that extends through X to Y), and (bii) there exists
interventions on all direct causes of Y that are not on α that fix the
values of those direct causes of Y “at” one distinctive “combination of
values” that fall “within” a “redundancy range,” and (biii) there exists
an “intervention on X that will change the value of Y.”12
I’ve included the ‘P’ in ‘IMA*-P’ so as to make explicit the fact that Wood-
ward’s theory presupposes the possibility constraining interpretation of inter-
ventions. According to that interpretation, when seeking to assess whether
the value of variable X caused the value of variable Y, there should be a meta-
physically coherent story to tell about an intervention of the relevant kind on
X (Woodward 2016, sect. 5).13 This is because IMA*-P says that a necessary
condition for the truth of any causal claim is that the cause reported on in
that claim be the value of a variable that can be intervened upon.
A competing interpretation is called the settings intervention approach
(henceforth IMA*-S). With respect to the X causes Y case (or respective
values of those variables), IMA*-S says that we should,
just think of X as set to some new value in the arrow-breaking or equa-
tion replacement manner . . . with no further restrictions on when such
a setting operation is possible (or when it is permissible or legitimate to
invoke it).14
It is unclear what place interventions have in IMA*-S. It is unclear in

what way IMA*-S incorporates hypothetical manipulations of values. To
manipulate X is to exercise some type of causal control over X. To simply
imagine a difference in X, and not a true productive changing of X, is
to entertain in one’s mind’s eye a non-causal state or situation. Values of
variables are imagined to just change without the assistance of any causal
mechanism. Such a theory is not worthy of the name “intervention manip-
ulability theory.” I will therefore largely ignore it, evincing only indirect
objections to it.
Interventions on X play an indispensable role in the IMA*-P approach to
the causal relation connecting X and Y (i.e., their actual values). Woodward
connects the effort of determining how interventions on X yield changes of
Y to the activity of assessing certain counterfactuals (viz., interventionist
counterfactuals). This is because (as with the AP-A criticized in chapter 4)
IMA*-P is perhaps best thought of as providing counterfactuals that connect
“causal claims to claims about what would happen if certain manipulations
were performed.”15 Interventionist counterfactuals are conditionals of the
form, “If it were the case that an intervention I on variable X transpired,
thereby setting X to value x, then it would be the case that Y would. . . .”16
Thus, according to Woodward, given the truth of something like IMA*-P,
it is best to appropriate a semantics for counterfactuals that differs from the
standard Lewisian approach of D.K. Lewis (Counterfactuals 1973). This
is because of the trivial truth-condition for the ‘would’-counterfactual in
Lewis (1973). According to that condition, if it is impossible for there to
be an intervention on X, then any and every interventionist counterfactual
like the previous sample (it involves X in the antecedent and Y in the conse-
quent) will come out trivially true and as a result we will be forced to judge
that X is indeed a cause of Y.17 But it is a tenet of the metaphysicalC system
of this project (q.v., chapter 1) that the correct semantics for counterfac-
tual conditionals is that of Lewis (Counterfactuals 1973). I have therefore
secured my first criticism of IMA*-P. I turn now to a second worry.
If one is a proponent of IMA*-P, and one abandons Lewis’s semantics
for counterfactuals, one might wonder which semantics is appropriate? The
discussion of Woodward (2003, 133–145) showcases how divergent judg-
ments about certain causal claims result from adjusting Lewis’s comparative
metric for similarity relations between worlds (presented and discussed in
chapter 3: sect. 4.3.2). In one of Woodward’s discussions, the more specific
thought seems to be that one can produce an IMA-friendly semantics by
assessing interventionist counterfactuals relevant to causal claims in such a
way that the semantics for them rejects strong centering (Woodward 2016,
sect. 5; sect. 9). Strong centering is the idea that, given that w is a q-world
(a world at which q holds), then the possible world that is both a q-world
and most similar to w, is none other than w. Rejecting strong centering
amounts to a reductio for any possible world and similarity-based seman-
tics for counterfactuals. Necessarily, everything resembles or is more similar
(according to any reasonable metric of similarity) to itself than anything else
is. Metrics of similarity become unreasonable if they entail that something
is not most uniquely similar to itself. I therefore find strong centering to be
highly intuitive. Thus, if one sides with Woodward’s (2003; 2016) discus-
sions (the appropriate sections), one will be forced to abandon what appears
to be an obvious truth. I turn now to my third criticism.
There are more attempts to supply altogether different semantics for
interventionist counterfactuals that depart even further from Lewis’s theory
(see, e.g., Hiddleston 2005).18 Most of these accounts are incomplete (Briggs
2012, 139). The semantics articulated in the work of Rachael Briggs ( 2012),
a summary statement of which would require too much space, is brilliant
and technically erudite. Unfortunately, it suffers from two problems. First,
it invalidates modus ponens (Briggs 2012, 150–152), and second, it is at
odds with a formal assumption of this project, that Lewis’s semantics is
the correct semantics for the ‘would’-counterfactual. These two complaints
are intimately related. Briggs is forced to reject instances of modus ponens
for counterfactuals with conditional consequents because she rejects weak
centering, or the thesis that, given that w is a q-world, then the set of worlds
that are most similar to w and that are q-worlds includes or incorporates
w itself.19 But as Briggs points out, weak centering follows from strong
centering. But again, that doctrine is above reproach. I therefore conclude
my second criticism with the judgment that there appears to be no reason-
ably complete semantics for interventionist counterfactuals, and those more
complete proposals that have been articulated reject obvious and intuitive
truths.
My last objection has the following formulation:
(1) If IMA*-P holds, then (if variables X and Y fail (negation) to satisfy
condition (bii) of IMA*-P, then it is not the case that x causes (simplic-
iter) y).20
(2) If x directly causes y, then x causes (simpliciter) y.
(3) It is not the case that (if x directly causes y, then X and Y satisfy condi-
tion (bii) of IMA*-P).
(4) Therefore, it is not the case that IMA*-P holds.
Premise (1) follows from Woodward’s characterization of IMA*-P. Premise

(2) is expressly affirmed by Woodward (2003, 42). The conditional that (if
x directly causes y, then X and Y satisfy condition (bii) of IMA*-P) is false.
If one looks at the conditions on direct causation specified by DC, one will
notice the absence of any content equivalent to condition (bii).21 Thus, it is
not true that <X and Y satisfy condition (bii) of IMA*-P.> is a necessary
condition for the truth of <x directly causes y.>. Conclusion (4) follows logi-
cally from (1)–(3).
Section 3: Against the Mechanistic Account

Stuart Glennan has recently proposed a mechanistic anti-reductive theory
of causation (Glennan 2017, 144–169).22 According to it, causation consists
of a connection between effects and causes that is mediated by mechanisms
such that the following biconditionals hold:
(The New Mechanistic Account (MA)): (a^) A causal statement such as

<C causes E> holds if, and only if, by way of at least one mechanism,
C “contributes to the production of” E,23 and (b^) “one event causes
another just in case there exists a mechanism connecting them.”24
Claim (a^) in MA is not a purported conceptual analysis of causation.25 Nor

is it a purported philosophical analysis of the causal relation. Rather, (a^) is
a report on what makes causal statements true.26 Nonetheless, (a^) together
with (b^), and Glennan’s theories of mechanism and production shed some
light on the deep metaphysical nature of causation. Below is my best exeget-
ical effort at reconstructing Glennan’s metaphysics of that relation.
According to Glennan, the connection27 mediated by a mechanism is sin-
gularist in the sense that causal relations can obtain even if they are “not
instances of causal regularities or laws.”28 Causation is also intrinsic such
that the connection between a cause C and an effect E is determined by the
intrinsic properties of the mechanistic process or processes whereby C con-
tributes to E’s production, and the intrinsic properties of C and E.
Mechanisms are collections of concrete particulars or “organized collec-
tions of entities whose activities and interactions take place at some particu-
lar place and time” (ibid., 153).29 The plurality of interactions and activities
in nature is determinative of the multifarious types of productive causation.
Glennan therefore endorses a very weak brand of causal pluralism. There
are different types of causation because there are different types of mecha-
nisms. There are different types of mechanisms because there are different
types of activities and interactions. What are activities and interactions?
The two are sometimes, although not always, identical (ibid., 31). Activities
and interactions come apart in cases in which, due to the manifestation of
an entity’s causal powers, that single entity acts to produce a change in itself
solely. Interactions can never be monadic in this way. They always involve,
at least, a plurality of entities such that at least one such entity creates a
change in the other.
Activities and interactions share some affinities. Both involve entities, and
also manifested capacities or causal powers. Both are parts of mechanisms.
Both “are temporally extended processes,”30 and both are causal in that
they are productive.31
I have four objections to MA. Here is the first:
(1) If MA holds, then every instance of causation involves a mechanism

whereby the cause gives way32 to the production of its effect.
(2) Every mechanism is a temporally extended causal process.
(3) If (every instance of causation involves a mechanism whereby the cause
gives way to the production of its effect and (2) holds), then (every
instance of causation is temporally extended such that either (a) the
cause occurs temporally prior to its effect, or (b) the cause occurs tem-
porally subsequent to its effect).
(4) It is not the case that every instance of causation is temporally extended
such that either (a) the cause occurs temporally prior to its effect, or (b)
the cause occurs temporally subsequent to its effect.
(5) Therefore, MA is false.
As I have already noted, Glennan endorses premises (1) and (2) (if all
mechanisms involve activities and/or interactions (ibid., 57), and all such
activities or interactions are temporally extended, then all mechanisms are
temporally extended). To see that premise (3) holds, grant its antecedent
and notice how, given that supposition, any instance of causation one points
to involves a temporally extended mechanistic process. Because that process
is the means whereby the cause gives way to production, and that process
is temporally extended, the creation or production must take time. Because
MA says nothing directly about temporal direction, I will grant that the
direction of the temporally extended process can run past to future, or
future to past (hence (b) in (3)’s consequent).
The consequent of (3) is false for reasons already articulated in chap-
ter 3: sect. 2. There both could be, and actually are, instances of causation
involving physical interactions that are instances of simultaneous causation.
Instances of simultaneous causation fail to satisfy both (a) and (b). MA is
therefore false.
My second argument precludes one from regarding MA as a true philo-
sophical analysis of causation.33 (The modal operators in this argument are
indicative of metaphysical modality.)
(6) If MA is a true philosophical analysis of causation, then necessarily,

every instance of causation involves a mechanism whereby the cause in
that relation gives way to the production of its effect.
(7) Necessarily, every mechanism incorporates at least one interaction
among its parts.
(8) If ((7) holds and necessarily every instance of causation involves a mech-
anism whereby the cause in that relation gives way to the production
of its effect), then necessarily every instance of causation includes one
entity that produces a change in another.
(9) It is not the case that necessarily every instance of causation includes
one entity that produces a change in another.
(10) Therefore, it is not the case that MA is a true philosophical analysis of
causation.
Premise (6) follows from the content of MA, given that the appropriate way
to think about MA is as a philosophical analysis of causation, and the fact
that philosophical analyses are necessarily true, if true (both of these are
assumptions in play, see chapter 1: sect. 4.4.1). An unnecessitated version
of premise (7) is accepted by Glennan.34 He should accept (7), if his theory
of mechanism is a philosophical analysis, and so a theory about the deep
metaphysical nature of mechanism and mechanistic processes that holds at
all metaphysically possible worlds. Glennan provides no reasons to think
that his account is anything but a metaphysical theory or philosophical
analysis. But even if my reading is incorrect, it is natural to regard Glen-
nan’s characterization of mechanism as a metaphysical hypothesis that is
necessarily true, if true, given that MA is a proposed philosophical analysis
of causation. Similarly (and giving attention now to premise (8)), it is quite
natural to regard Glennan’s definition of interaction as one that holds at all
metaphysically possible worlds, given that MA is a proposed philosophi-
cal analysis. But recall that Glennan defines interactions as processes that
involve multiple entities (interactors). Interactions are processes in which
at least one interactor produces a change in another (ibid., 31). This is one
way in which interactions differ from activities (ibid.). But now it seems to
follow that given (7), and the consequent of (6), necessarily, every instance
of causation includes one entity producing a change in another. That is to
say, premise (8) holds.
The consequent of premise (8) is false for reasons that are closely related
to my discussion in chapter 7: sect. 3.1 (specifically, Datum 5 and Datum
5.5). Instead of envisioning Max Black’s persisting sphere at a Leibnizian
space-time, insert a possible simple object such as an electron e*, and stipu-
late (as we did with Black’s iron sphere) that e* is an example of an object
that persists over time, but does so in a way that is grounded by causal
relations.35 e* bearing its joint-carving universals at one time at a world
causes e* to bear the same universals at the next at the same world. There
are no changes in e* as it persists. It does not move, nor does it bear differ-
ent substantive relations or intrinsic properties over the times that it persists.
The scenario I’ve set up appears to be a possible case of causation without
change. Given certain views of persistence—ones which identify e* at t with
e* at a later time—we appear to also have a possible case of causal produc-
tion without one interactor affecting a distinct entity.
Conclusion (10) follows logically from premises (6)–(9).
Here is the third objection. The theory of activities and/or interactions
at the heart of MA is circular. Consider first how Glennan’s theory of (a)
interactions and activities appropriates the notion of (b) production (ibid.,
18, 148, 150; Glennan 2002, S344, mechanisms are said to produce via
interactions). Glennan’s theories of every type of production make use of
the notion of an (c) event (Glennan 2017, 179). But Glennan defines events
in terms of things involved in (a) interactions or activities (ibid., 177). It
seems illicit to explicate a notion in terms of (a), and then explicate (a) in
terms of (b), (b) in terms of (c), and (c) in terms of (a) in this way, unless
one regards interactions and/or activities as primitive entities.36 Although
the new anti-reductive theory proffered in sect. 4 requires primitivism about
causal dependence, it avoids circularity of this kind.
My last objection is aimed at the underlying ontology of the new mecha-
nistic approach to systems upon which MA is built. Consider the scenario
in which gluon1 causes gluon2 to change its trajectory. According to my
argumentation in chapter 3: sect. 3.1, this interaction involves an instance of
causation. It should be recovered as such by MA and its underlying mecha-
nistic ontology. Unfortunately, MA and its underlying ontology cannot
explain instances of causation involving gluons. This is because gluons are
simple entities and
[i]f the world is in fact made of mechanisms, it has consequences for the
philosophical understanding of the ontology of the natural world. New
mechanist ontology is an ontology of compound systems. It suggests
that the properties and activities of things must be explained by refer-
ence to the activities and organization of their parts.37
But what parts of gluon1 does one appeal to to explain its activity? More-
over, what parts of gluon2 does one look to to explain gluon2’s activity?
Again, these entities are simple. So, something is awry with MA coupled
with its underlying (unqualified) new mechanistic ontology, given a causal
approach to QCD.
Section 4: A New Proposal

Causal reductionism has failed us. Two leading anti-reductive theories of
causation have likewise failed us. I have a new anti-reductive proposal built
on the work of the preceding chapters of the current project. Here it is:
(The Account): (Necessarily, for any event x, and for event y, x deter-
ministically causes y, just in case, y causally depends for its occurrence
and contingent content upon x’s occurrence,38 where the causal depen-
dence relation is a multigrade, obtaining, formally asymmetric, formally
transitive, formally irreflexive relation, that is, with respect to purely
contingent events, universal and well-founded, and where the result-
ing causal dependency structure is intrinsic (in the sense of IT below)
when it connects purely contingent events) and (Necessarily, for any
plurality of events xx, and for any plurality of events yy, the xx cause
the yy, just in case, the yy causally depend for their occurrence and
contingent content upon the occurrence of the xx,39 where the causal
dependence relation is a multigrade, obtaining, formally asymmetric,
formally transitive, formally irreflexive relation, that is, with respect to
pluralities of purely contingent events, universal and well-founded, and
where the resulting causal dependency structure is intrinsic (in the sense
of IT below) when it connects pluralities of purely contingent events).
The contingent content of an event is the contingently exemplified joint-

carving universal or the relation contingently born by x (or relations or
properties born by the xx). With respect to an irreducible force F, F’s magni-
tude and direction constitute its contingent content, or in the case of forces
that are complex interactions, the strength of the interaction energy consti-
tutes the relevant contingent content (these are properties of the involved
irreducible force and therefore fundamental events are born out of the rel-
evant property exemplifications at indices).
In a scenario involving an obtaining causal relation between event c and e
(without pluralities and with a single joint-carving monadic universal), that the
substance constituent in e has the joint-carving universal it does at the relevant
index is (causally) due to the occurrence of the cause. That substances involved
in a plurality of events that are effects have the joint-carving universals they do
at the relevant index or indices, is (causally) due to the occurrence of the plural-
ity of events that are their causes. Causal structure “is, at bottom, dependency
structure.”40 What is more, the dependency involved is primitive causal depen-
dency. The idea is that effects existentially depend on their causes in a way that
underwrites production claims. Event e is produced by event c by virtue of the
factΣ that e causally depends in the way suggested by (The Account) upon c.
One might think that (The Account) needs to be revised. One might argue
that for any given effect e brought about by causes, (The Account) implies
that any event or force that made a mere difference to how e occurred is
the cause of e. (The Account) is not susceptible to this objection. This is
because, according to it, e must depend for its occurrence (its existence) on
its cause. Mere difference makers (in the above sense) do not serve as such
causal dependency bases by themselves. They may help fix the contingent
content of events by partially influencing them in some way, but they fail to
do much else. Whether they do even that much is a matter that should be
settled by sound causal epistemology. It is here that I believe manipulation-
ist/interventionist/structural equation approaches to causation can help (in
the epistemology of causation, not the metaphysics).
I am worried about a charge of triviality. (The Account) appears to almost
say that the cause of an event is what causes the event. How can one push
a proper understanding of (The Account) far away from such triviality?
I believe (The Account) is saved from triviality by way of the proposed
theory of causal dependence articulated in (The Account). More specifically,
I believe the account is far from trivial in large part due to the unique con-
nection it posits between the causal relation and laws of nature via the IT
thesis (see below).
I affirm the intrinsicness thesis.41 The idea (leaning heavily upon N. Hall,
Intrinsic 2004) is that when causation is a natural relation, causal structures
are intrinsic in a way that entails the following thesis:
(Intrinsicness Thesis (IT)): “Let S be a structure of [purely contingent]

events consisting of [purely contingent] event e, together with all of its
causes back to some earlier time t. Let S* be a structure of [purely con-
tingent] events that intrinsically matches S in relevant respects, and that
exists in a world with the same laws. Let e* be the [purely contingent]
event in S* that corresponds to e in S. Let c be some [purely contingent]
event in S distinct from e [that causes e], and let c* be the [purely con-
tingent] event in S* that corresponds to c (i.e., c* is an intrinsic dupli-
cate of c). Then c* is a cause of e*.”42
There are several points of clarification to make about IT. First, by ‘S*
intrinsically matches S in relevant respects’ I mean “S* is an intrinsic dupli-
cate of S,” and by ‘S* is an intrinsic duplicate of S’, I mean, structures S and
S* can be represented by an ordered pair {{S}, {S*}} such that each of the
constituent events of the members of that pair have substance constituents
that bear the self-same non-grue-like (in the sense of chapter 1: sect. 4.3)
intrinsic, qualitative, categorical properties (and also the self-same joint-carving
universals).43
An intrinsic property of a substance o that is Fness is a property that o
bears wholly by virtue of what o is.44 For me, how or what something is,
has to do with its nature or essence. I’m a primitivist about the notion of
a nature or essence. I have no, even, informative analysis of a nature or
essence generally conceived. I believe the notion can be cashed out, at least
partly, by looking to highly intuitive examples (e.g., having four sides is
in the nature of a square, being normative is in the nature of a true moral
principle, being abstract is in the nature of the number 2, etc., but I resist the
idea that such properties are always necessarily possessed by the substances
that exemplify them).
A qualitative property is one whose instantiation does not entail the exis-
tence of any individual. Following Armstrong (1997, 80–83) and Bird (2007,
66-67), I’ll define a categorical property as one that is non-dispositional and
non-modal. It is not a causal power or capacity and affords no necessary
connections with anything in nature.
In his statement of IT, Hall included all of the causes of e stretching back
to an antecedent time because a great many causes bring about their effects
with other causes and through causal processes featuring intermediates
(ibid., 264–265, 270–274).
Hall uses causal reductionism to motivate IT. That motivation can be
easily jettisoned. In fact, IT fits better with an anti-reductionist approach to
causation rather than a reductionist one. Hall,45 Armstrong, and Menzies
have all suggested as much (see Armstrong 1999; N. Hall, Intrinsic 2004,
257–258, cf. 261, n. 9; and Menzies 1999, 314–317, 319–320; I noted this
in chapter 4).
IT helps (The Account) escape the charge of triviality by providing the
means whereby a proponent of (The Account) can clarify the claim that it
is a purported theory or philosophical analysis of deterministic causation.
This is because with respect to instances of deterministic causation, the laws
in IT must be deterministic laws, like the laws of motion in Newtonian
mechanics, or the dynamical laws of general relativity (all appropriately
interpreted).
Section 5: Entering Through Armstrong’s Door

Consider Figure 9.1:
A
C
Figure 9.1 Late Preemption (adapted from Collins, Hall, and Paul 2004, 23)
Figure 9.1 is a neuron diagram. Following closely (ibid., 17; Paul and Hall
2013, 9–10), I note how circles that are shaded in give us firing neurons at
space-time locations (events in the ESSI sense from chapter 7: sect. 4), the
arrows are causal stimulatory connections or signals, and the times at which
these events and connections are established are read off of the illustration
from left to right. Italicized letters will represent the event of the firing of the
labeled neuron, whereas bold letters will represent individual neurons. Thus,
neurons A and B fired at a time t0, while C fired at a time t1.46 What Fig-
ure 9.1 says, therefore, is that A’s firing caused C’s firing, and while B fired at
the same time A fired, its stimulatory signal failed to reach C at t1, although
had A failed to fire, C would have still fired because the stimulatory signal
sent out from the event that is B’s firing would have brought about C’s firing.
Figure 9.1 represents a causal structure that is highly idealized. Real
world causal structures are incredibly more complicated than any structure
like the above would suggest. Indeed, I wonder if it is even metaphysically
possible to find such pristine causal scenarios like Figure 9.1 and those that
follow. However, it is common to treat these idealized scenarios as real dif-
ficult cases that should be overcome by any plausible theory of causation.
I will therefore assume the metaphysical possibility of the scenarios dis-
cussed later, and detail how (The Account) easily handles them by entering
through David Armstrong’s door.
So what exactly is Armstrong’s door? Here is Armstrong’s description,
provided with models like the neuron diagram that is Figure 9.1 in mind,
Where there is an arrow in a diagram showing that one neuron brings it

about that another neuron fires, take it that here there is a genuine two-
term relation of singular causation holding between cause and effect.
Where there is no such arrow, deny that there is any such relation. This
is the open door.47
Following Armstrong’s strategy, (The Account) coupled with IT handles the

situation depicted by Figure 9.1 by way of the following description.
Description afforded by (The Account)48: an effect that is C at t1 is

wrought by a stimulatory signal, itself caused by A (at t0). B occurred
at t0, but its resulting stimulatory signal traveled slower and a shorter
distance than that of A’s. It fails to cause C because A’s signal produces
C first. There is no difficulty here.
Now consider Figure 9.2:
A B
E
C D
Figure 9.2 Early Preemption

Note: I draw upon Collins, Hall, and Paul (2004, 18) and Paul (2009, 168). The arrowless line
is an inhibiting signal sent from A to D.
In this case (as it is typically understood), A causes B, and that brings

about E’s firing. But a distinct and independent causal process is initiated
by C’s sending a stimulatory signal to D, although because A’s firing also
sends an inhibitory signal to D, D fails to fire (the inhibitory signal keeps C’s
stimulatory signal from causing D to fire). Were D to have fired, it would
have sent a stimulatory signal to E, thereby causing E’s firing. A is therefore
an early preemptor of C because its firing inhibits D halting the causal pro-
cess that would have stretched from C to E. The last lower black arrow is a
mere hypothetical stimulatory signal (see Paul 2009, 168).
Early preemption cases such as this give both regularity and simpler coun-
terfactual analyses of causation a very hard time,49 but for my specific view,
the preceding case is unproblematic, although it will turn out that there are
more delicate causal facts to appreciate on that view.
Description afforded by (The Account): Via stimulatory signals, A causes

B, which causes E. In addition, A partially causes D to exhibit the joint-
carving features it does at time t1, and C causes a stimulatory signal, which
helps (with A) to fix the contingent content of D at t1. Because C influences
D, it serves as a partial cause of D (so the arrow from C to D isn’t quite a
full causal arrow). The arrow from D to E remains hypothetical.
What of symmetric overdetermination cases such as Figure 9.3?
A
C
Figure 9.3 Symmetric Overdetermination (adapted from Paul 2009, 178)
In Figure 9.3, both A and B causally produce C by sending their own

(independently) causally sufficient stimulatory signals to C. Each signal
reaches C at the same time, and both A and B fire at the same time. This
example is also trouble for several different theories of causation, but it
should not trouble the proponent of (The Account) plus IT for natural cau-
sation (i.e., causation involving only distinct purely contingent events).
Description afforded by (The Account): A and B are both causes of C.

Period. There’s simply nothing more to do except close Armstrong’s
door behind you.
Finally, consider a trumping preemption case illustrated by Figure 9.4:
A
C
Figure 9.4 Trumping Preemption (adapted from Collins, Hall, and Paul 2004, 28
fig. 1.7)
In Figure 9.4, both A and B send stimulatory signals to C, except that

A’s stimulatory signal is more intense than B’s. Because A and its stimula-
tory signal are so overpowering, A’s firing is the cause of C. Moreover, one
should not understand B’s firing as an additional cause of C because the
intensity of A’s firing trumps B (hence the skinny gray arrow). It is important
to add (as ibid. do) that if A did not transpire, C would have still obtained
because of the ever so slight signal sent out from B. The idea is that there’s
something interesting about the relationship between the signals, such that
the one overpowers the other when both are present, although in the absence
of the more intense signal, the weaker signal from B is enough to do the
causal job. If all this is unclear, perhaps the quoted pericope below will help:
The sergeant and the major are shouting orders at the soldiers. The
soldiers know that in case of conflict, they must obey the superior offi-
cer. But as it happens, there is no conflict. Sergeant and major simul-
taneously shout ‘Advance!’; the soldiers hear them both; the soldiers
advance . . . since the soldiers obey the superior officer, they advance
because the major orders them to, not because the sergeant does. The
major preempts the sergeant in causing them to advance. The major’s
order trumps the sergeant’s.50
The case as described by this illustrative story is very helpful because it high-
lights precisely why the trumping preemption phenomenon is unproblem-
atic for (The Account). If it is true that the soldier advances only because of
the Major’s order, what fixes the contingent content of the event that is the
soldier’s advancing is the major’s order, not the sergeant’s. (The Account),
therefore, provides the right result since it counts as the cause only that
which fixes such content.
Although there are other types of difficult cases in the literature (see,
e.g., the brilliant discussion of such cases in Paul and Hall 2013), this sec-
tion provides some reason to suspect that (The Account) can handle them.
That it at least handles those explicitly addressed here constitutes a reason
to believe that it outperforms those theories for which the cases are clear
counter-examples.
Section 5.1: The Account and Schafferian Grounding

In both chapter 1: sect. 4.5.3 and chapter 8: sect. 4.6, I adopted Jonathan
Schaffer’s theory of grounding (Schaffer 2009). That theory says that entity x,
grounds entity y, just in case, y depends for its positive ontological status and
nature on x, where the dependence relation is asymmetric, transitive, and well-
founded. I argued in chapter 8: sect. 4.6 that grounding and causation are rela-
tions that can at least coincide (and by that I meant correspond in character). In
this sub-section, I argue that obtaining causal relations that hold between dis-
tinct purely contingent events—what I have call “natural causal relations”—are
instances of Schafferian grounding between distinct purely contingent events
characterized, in part, by the ESSI (q.v., chapter 7: sect. 4). When I say that they
are instances of grounding, what I mean is that the nature of the natural causal
relation is such that it satisfies the conditions for being a grounding relation
although its nature involves more than that which is required for such satisfac-
tion. These details about causation’s relation to grounding underwrite causa-
tion’s fundamental status, in the sense that the notion of causal dependence that
is central to natural causation is “an unanalyzable but needed notion—it is”
one of “the primitive structuring conception[s] of metaphysics.”51
When event c naturally causes event e in the way suggested by (The
Account), event e stands in a causal dependence relation upon c, such that
e’s existence is due to c. As I noted previously, obtaining causal dependency
relations back production factsP. Event c produces event e, by virtue of the
factΣ that e causally depends on c. That is to say, the kind of dependence at
work here at least involves existential dependence. Production claims are
causal claims by virtue of their being backed by causal dependence. But
causal dependence is richer still. When e causally depends on c, event e caus-
ally depends for its nature upon c. Recall from chapter 7 that the nature of
an event e resides in the obtaining contingent tie/unity holding between the
constituent substance(s), joint-carving universal(s), and ontological index
or indices involved in it. That tie or connection obtains/occurs because of
c’s occurrence. The tie or connection is due to c’s occurrence. Thus, accord-
ing to (The Account) an obtaining natural causal relation between c and e
(directed from c to e) entails that e causally depends for its existence and
nature upon c. That dependence relation is well-founded (q.v., my defense
of this in chapter 6: sect. 4), transitive (q.v., my defense of this in chapter 6:
sect. 3), and asymmetric (q.v., my defense of this in chapter 3: sect. 1.1), just
like the dependence relation involved in Schafferian grounding. Instances of
natural causation are instances of Schafferian grounding between distinct
purely contingent events. I hope I have provided a clearer picture of how it
is that natural causation and grounding coincide or correspond in character.
My account of natural causation is not Schaffer’s theory of ground-
ing. This is because natural causation has a richer nature than Schaffer-
ian grounding. Consider the following facts. (The Account), not Schaffer’s
theory of grounding (call it G), maintains that natural causation relates a
specific type of entity for which I have a specific philosophical analysis in
chapter 7. (The Account), not G, maintains that the relevant dependency
relation is multigrade. (The Account), not G, maintains that the involved
dependency is intrinsic. (The Account), not G, maintains that the involved
dependency is deterministic by virtue of the way the relation is intrinsic.
That is to say, when causation is deterministic, the applicable version of IT
incorporates deterministic laws. (The Account), not G, maintains that the
involved dependency is, via the intrinsicness thesis, backed by or otherwise
related to deterministic laws of nature and not just laws of metaphysics.
And lastly, (The Account), not G, maintains that the kind of dependence at
the heart of the phenomenon is primitive causal dependence.
In recent work, Schaffer has argued that although there are several analo-
gies between causation and grounding, there are also disanalogies that should
keep us from accepting the thesis that grounding and causation are unified,
such that the latter “drives the world through time,” and the former “drives
the world up levels” (this is the grounding-causation unity thesis).52 His
argument is based on three points:
(a) Causation is not well-founded, but grounding is.

(b) There could be instances of indeterministic causation, but there cannot
be instances of indeterministic grounding.
(c) Necessarily, causation brings together distinct relata, and it is not the
case that necessarily grounding relates distinct relata. In fact, ground-
ing is an internal relation such that it only “connects indistinct entities”
(leaning on Schaffer, Grounding in the Image of Causation 2016, 94–95,
the quotation in (c) is from ibid., 94; although it is unclear what Schaf-
fer means precisely by “distinct” and “indistinct” (I’ll leave matters at
an intuitive level, at first)).
These considerations do not help provide a strong objection to my view of

the relationship between grounding and causation. I’ve argued that natu-
ral causation is well-founded. And second, point (b) is irrelevant because
the instances of causation I claim are instances of grounding are all
instances of deterministic causation. Third, and I hope not to vex my for-
mer teacher, according to Schaffer (2010), everything is internally related,
thus no two things are really distinct. If this reading is right, and Schaffer’s
argumentation in (2010) is sound, then causation will relate indistinct (in
the sense that seems relevant to issues about the metaphysics of grounding
anyway) things after all. But I’m no priority monist, so I note in closing
that in Schaffer’s discussion of the paradigm cases of grounding (Ground-
ing in the Image of Causation 2016, 52–54) we have examples like the
following:
(d) Newton’s physical state realizes, and thereby grounds, a relevant mental
state (where physicalism is an assumption in the background).
(e) Newton makes true, and thereby grounds, the proposition <Newton
exists.>
(f) {Newton} depends (for its existence and nature) upon, and is thereby
grounded in, Newton.
If by the postulate that grounding brings together indistinct entities, Schaf-

fer meant entities that are identical, (d)–(f) all present problems. Realization
is not a relation between something and itself. Newton is not the propo-
sition <Newton exists.>, and Newton is not identical to the singleton set
{Newton}.
If by the postulate that grounding brings together indistinct entities, Schaf-
fer meant that the grounding entity has the grounded entity as a mereo-
logical part (or vice versa), (f) presents a problem. Newton does not have
{Newton} as a part, and neither should we say that {Newton} has Newton
as a mereological part. Newton is the sole member of {Newton}, but (stan-
dardly) set membership is non-mereological (“[m]embership could not be
(a special case of) the part-whole relation because of a difference in formal
character” (Lewis 1991, 5)).
If by the postulate that grounding brings together indistinct entities Schaf-
fer meant that there’s a metaphysically necessary connection between the
grounder and grounded things, then (d) presents a problem. It is at least
highly controversial that realization relations are relations involving meta-
physically necessary connections between realizers and realized things.
Indeed, according to one sophisticated account of realization, realization
relations essentially involve natural necessitation (Gillett 2016, 82–90). But
natural necessitation is natural law governed or backed. On Gillet’s view, the
physical realizers could have failed to realize the higher-level mental stuff at
metaphysically possible worlds with radically different laws or no laws at all.
It appears to be difficult to find a precisification of ‘indistinct’ according
to which Schaffer’s claims in (c) come out true, the paradigm instances of
grounding remain such instances, and yet instances of natural causation are
precluded from being instances of grounding.
Notes
1. Paul and Hall (2013, 249). Q.v., chapter 8: note 12.
2. Recall that in earlier discussion, I also criticized David Armstrong (chapter 6:
sect. 2.2) and E. J. Lowe’s (chapter 2: sect. 4) anti-reductive theories.
3. Although Woodward spends much time developing a theory of general or type-
level causation, here I focus on his theory of “actual causation” or what I’ve
called token singular causation. In this chapter, the term ‘causation’ will pick out
token singular causation exclusively unless otherwise noted.
4. Woodward (2003, 75).
5. Ibid., 55; bold emphasis removed. Use of V is meant to capture the fact “that any
description of causal relationships reflects a choice of level of analysis” (ibid.).
6. Here is Woodward on a redundancy range:
Consider a particular directed path P from X to Y and those variables
V1 . . . Vn that are not on P. Consider . . . a set of values v1 . . . vn, one for
each of the variables Vi. The values v1 . . . vn are in what Hitchcock calls the
redundancy range for the variables Vi with respect to the path P if, given the
actual value of X, there is no intervention that in setting the values of Vi to
v1 . . . vn, will change the (actual) value of Y.
(ibid., 83)
7. Quotations in the statement of IMA are from Woodward (2003, 84). By “avenue,”
I have in mind Woodward’s notion of a directed path, for which see ibid., 42.
8. Woodward (2003, 98).
9. See Woodward (2003, 98); Woodward and Hitchcock (2003); and Woodward
(2016, sect. 5).
10. In Woodward’s original text, he uses the clause “Then determine whether, for
each path” and adds that the condition he specifies is “satisfied if the answer to”
the ‘whether’-question “is ‘yes’ for at least one route and possible combination
of values within the redundancy range of the” direct causes of Y not on avenue
α (Woodward 2003, 84).
11. If one were to revert to interventionist counterfactuals (introduced later in the
main text) in one’s statement of IMA, my point would still hold water. That is
to say, it would remain true that whether there’s a causal relation between the
values of X and Y depends upon whether one figures out if a particular interven-
tionist counterfactual (modal statement) is true.
12. Quotations in the statement of IMA*-P are from Woodward (2003, 84). Perhaps
one could adjust even this version further by ensuring that w is, in some sense,
close to the actual world (q.v., n. 13 however).
13. In Woodward (2003, 128), it is claimed that the interventions need only be “log-
ically possible and well-defined.” Elsewhere (ibid., 114–117), a theory of causal
claims is provided that incorporates the idea of hypothetical experimentation.
14. Woodward (2016, sect. 5). I’ve modified the X variable slightly by taking away
the subscript.
15. Woodward (2009, 236; emphasis in the original).
16. I am not quoting word-for-word, but I lean on Woodward (2016, sect. 5).
17. This was a problem we noted for the AP-A theory criticized in chapter 4: sect.
2.1. Please note my citation of Woodward in note 32.
18. In the spirit of documenting approaches to the semantics for counterfactuals that
are friendly to the IMA, Woodward (2016, sect. 5) cites Kit Fine’s (2012) recent
attempt to articulate a semantics for the counterfactual conditional that forsakes
possible worlds. Unfortunately, like the study of Hiddleston, the application of
that semantics to the IMA is admittedly incomplete. Fine remarked, “certain
details in how exactly the semantics is to be applied will still need to be decided”
(ibid., 243, n. 24). I should add that Fine’s semantics validates modus ponens
(ibid., 239–240), while Briggs’ does not. It is unclear which semantics we are to
try to adopt given Woodward’s citation of both papers.
19. Or, if p is true at a world (or model) w, then w is a member of f(p, w) itself a set
of worlds or models (Briggs 2012, 152). The variable ‘f’ represents a selection
function, which takes in as its inputs both a possible world w, and a proposition
q. It then spits out every q-world that is most similar to w.
20. I have in mind the values of the relevant variables, but I’ll refrain from inserting
this qualification all over the place. That would be annoying.
21. This isn’t just because there’s no talk of redundancy ranges in DC. It is also
because DC is stated with a “Relative to a set of variables V . . .” clause. Nor are
there any conditions that entail condition (bii) in Woodward’s statement of DC.
22. I thank Stuart Glennan for valuable correspondence that helped me better under-
stand his book. Any problems with interpretation are my unintentional fault.
The study of mechanical philosophy and mechanisms is enjoying somewhat of a
resurgence in contemporary analytic philosophy (see, e.g., the collection of essays
in Glennan and Illari (2018) for a small taste of some of the recent literature).
23. Paraphrased from Glennan (2017, 156). Quotation from ibid.
24. Glennan (2017, 185). There he is characterizing his view as part of the summary
of a circularity criticism. However, he clearly seems to regard the characteriza-
tion as accurate, eluding the criticism not by rejecting its characterization of his
approach but by other means.
25. Glennan disowns conceptual analysis as a choice methodology for the study of
causation and mechanism (Glennan 2017, 11–12).
26. Glennan does not seem to have a theory of truthmaking in mind when he says
that goings-on in the world involving mechanisms, events, and productions
make true certain causal claims.
27. There are places in Glennan’s recent work where he seems to step away from the
idea that causation is a relation. Yet, there are many other places in that same work
where Glennan represents his position as one that is committed to the idea that
causation is a relation (Glennan 2017, 151–152, 155, 157, and especially 185).
28. Glennan (2017, 151). According to Glennan, laws of nature are true descrip-
tions of mechanistic processes; they do not explain such processes (they are “just
descriptions of the behavior of mechanisms”; ibid., 57; emphasis mine). Rather,
those processes explain the relevant laws (ibid., 153). This theory of laws con-
flicts with the partial theory articulated in chapter 1: sect. 4.5.5.2.
29. For Glennan’s complete account of mechanism, see Glennan (2017, 17–58, spe-
cifically 57–58).
30. Ibid., 32. The immediate context of the quote is about activities, but Glennan
adds that “[t]his stipulation applies both to monadic activities and to interac-
tions,” and that “[i]t will be crucial to our account of interactions . . . that these
points of intersection [Salmon’s interactions] are not extensionless points, but
take time” (ibid., 32; emphasis mine).
31. Ibid., 31.
32. By ‘gives way’ I mean contributes.
33. Even if a proponent of MA does not regard it as a thesis that is necessarily true, if
true, argument (6)–(10) would still have some value. This is because (6)–(10), if
sound, restrict the proponent of MA by precluding them from truthfully claim-
ing that MA exhausts the deep metaphysical nature of causation. The argument
thereby ensures that MA is not a real competitor with the account presented in
sect. 4 because MA cannot be a philosophical analysis.
34. Glennan stated,
But why then can we not dispense with the term ‘interaction’? The reason,
as Tabery (2004) has argued, is that proper mechanisms require that at least
some of the parts interact . . . there is no production without interaction.
(Glennan 2017, 21–22)
And see Tabery (2004).
35. In chapter 7, I noted how a great many philosophers from a great many different
philosophical traditions accept the above persistence case as a causal one. The
objection I’m now proffering resembles one voiced by Ehring (1997, 87).
36. L. R. Franklin-Hall notes how new mechanistic theories of causation seem com-
mitted to the bruteness of activity (Franklin-Hall 2016).
37. Glennan (2017, 57). Although cf. his comments at ibid., 185, where he seems to
leave open the possibility of a fundamental type of basic interaction.
38. The events can be the fundamental events as explicated in chapter 7: sect. 4.
39. The events can be a plurality of the fundamental events as explicated in chap-
ter 7: sect. 4.
40. Hall (2011, 101; emphasis removed). For Hall, the quoted statement holds for
one of two different types of causation. For me, it holds for all instances of token
causation.
41. The thesis is taken mostly from Hall (Intrinsic 2004, 264).
42. Quoting Hall (Intrinsic 2004, 264; emphasis mine). I’ve replaced his apostro-
phes with asterisks. Hall assumes (i) causal reductionism, (ii) Maudlin’s (2007)
theory of laws, and (iii) that the fundamental laws are deterministic (ibid., 261).
As the main text will go on to point out, neither assumptions (i) nor (ii) are nec-
essary. Assumption (iii) is unproblematic in this context because I’m providing
an account of deterministic token causation.
43. This departs significantly from Hall’s own understanding of (IT).
44. Following Marshall (2009, 646–647).
45. “This position [an intrinsic view of causation] is most naturally developed as
part of a certain kind of non-reductionist position about causation, according to
which facts about what causes what are metaphysically primitive” (Hall Intrin-
sic 2004, 258).
46. See Collins, Hall, and Paul (2004, 17).
47. Armstrong (1999, 176).
48. All causal terminology in my descriptions of the “difficult cases” are notions that
denote token causation with the nature (The Account) describes.
49. As has been pointed out by Collins, Hall, and Paul (2004, 18) and Paul (2009,
168).
50. D.K. Lewis (2000, 183; emphasis in the original, who credits the example to Bas
van Fraassen in ibid., 183, n. 3; cf. Schaffer 2000).
51. Schaffer (2009, 364; emphasis removed). I have borrowed Schaffer’s wording to
make a slightly different point. I ask the reader to grant me some artistic license
with respect to my use of the term ‘fundamental’ in the main text. In chapter 1:
sect. 4.5.3, I said that fundamental things are things that aren’t grounded. Here
I use the term fundamental to classify a notion. So, my use is somewhat different
here.
52. Schaffer (Grounding in the Image of Causation 2016, 94).
Bibliography
Achinstein, Peter. 2013. Evidence and Method: Scientific Strategies of Isaac Newton
and James Clerk Maxwell. New York: Oxford University Press.
Adams, Robert M. 1981. “Actualism and Thisness.” Synthese 49 (1): 3–41.
Adler, S.L. 2007. “Lower and Upper Bounds on CSL Parameters From Latent
Image Formation and IGM Heating.” Journal of Physics A: Mathematical and
Theoretical 40: 2935–2957.
Aguilar, A.C., D. Binosi, J. Papavassiliou, and J. Rodríguez-Quintero. 2009.
“Nonperturbative Comparison of QCD Effective Charges.” Physical Review D
80: 085018 (pagination is 1–18).
Aizawa, Kenneth, and Carl Gillett (editors). 2016. Scientific Composition and
Metaphysical Ground (New Directions in the Philosophy of Science). London:
Palgrave Macmillan.
Ajuriaguerra, J. de, M. Strejilevitch, and R. Tissot. 1963. “A propos de quelques
conduites devant le miroir de sujets atteints de syndromes démentiels du grand
âge.” [That is to say, “On Some Behaviors in Front of the Mirror of Subjects with
Dementia Syndromes of Old Age”] Neuropsychologia 1 (1): 59–73.
Albert, David Z. 1992. Quantum Mechanics and Experience. Cambridge, MA:
Harvard University Press.
Albert, David Z. 2000. Time and Chance. Cambridge, MA: Harvard University
Press.
Albert, David Z. 2015. After Physics. Cambridge, MA: Harvard University Press.
Ali, A., and G. Kramer. 2011. “JETS and QCD: A Historical Review of the
Discovery of the Quark and Gluon Jets and Its Impact on QCD.” The European
Physical Journal H 36: 245–326.
Alkon, Daniel L. 1989. “Memory Storage and Neural Systems.” Scientific American
261 (1): 42–50.
Allori, Valia. 2013. “Primitive Ontology and the Structure of Fundamental Physical
Theories.” In The Wave Function: Essays on the Metaphysics of Quantum
Mechanics, edited by Alyssa Ney and David Z. Albert, 58–75. New York: Oxford
University Press.
Allori, Valia, Sheldon Goldstein, Roderich Tumulka, and Nino Zanghì. 2008. “On
the Common Structure of Bohmian Mechanics and the Ghirardi-Rimini-Weber
Theory.” The British Journal for the Philosophy of Science 59 (3): 353–389.
Alston, William P. 1996. A Realist Conception of Truth. Ithaca, NY: Cornell
University Press.
Alston, William P. 2000. Illocutionary Acts and Sentence Meaning. Ithaca, NY:
Cornell University Press.
310 Bibliography
Alston, William P. 2001. A Sensible Metaphysical Realism. Milwaukee, WI:
Marquette University Press.
Anderson, J.L. 1996. “Answer to Question #15 [‘What Space Scales Participate in
Cosmic Expansion?,’ Frank Munley.” American Journal of Physics 64: 527–528.
Andersson, B. 1986. “The Experiential Gestalt of Causation: A Common Core to
Pupils’ Preconceptions in Science.” European Journal of Science Education 8 (2):
155–171.
Anscombe, G.E.M. 1968. “The Intentionality of Sensation: A Grammatical
Feature.” In Analytic Philosophy, Second Series, edited by R.J. Butler, 158–180.
Oxford: Basil Blackwell.
Aristotle. 1984. Categories. Vol. 1, in The Complete Works of Aristotle: The Revised
Oxford Translation. Edited by Jonathan Barnes, translated by J.L. Ackrill, 3–24.
Princeton, NJ: Princeton University Press. Cited as “Categories”
Aristotle. 1984. The Complete Works of Aristotle. Revised Oxford Translation.
Edited by Jonathan Barnes. Vol. 1. Princeton, NJ: Princeton University Press.
Cited as “Complete Works vol. 1”
Aristotle. 1984. The Complete Works of Aristotle. Revised Oxford Translation.
Edited by Jonathan Barnes. Vol. 2. Princeton, NJ: Princeton University Press.
Cited as “Complete Works vol. 2”
Armstrong, David M. 1968. A Materialist Theory of Mind. New York: Humanities
Press.
Armstrong, David M. 1978. Nominalism & Realism: Universals & Scientific
Realism. Vol. 1. New York: Cambridge University Press. Cited as “Nominalism &
Realism”
Armstrong, David M. 1978. A Theory of Universals: Universals & Scientific
Realism. Vol. 2. New York: Cambridge University Press. Cited as “Theory of
Universals”
Armstrong, David M. 1980. “Identity Through Time.” In Time and Cause: Essays
Presented to Richard Taylor, edited by Peter van Inwagen, 67–78. Dordrecht: D.
Reidel.
Armstrong, David M. 1988. “Reply to van Fraassen.” Australasian Journal of
Philosophy 66 (2): 224–229.
Armstrong, David M. 1989. Universals: An Opinionated Introduction. Boulder,
CO: Westview Press.
Armstrong, David M. 1997. A World of States of Affairs. Cambridge: Cambridge
University Press.
Armstrong, David M. 1999. “The Open Door: Counterfactual Versus Singularist
Theories of Causation.” In Causation and Laws of Nature, edited by Howard
Sankey, 175–185. Dordrecht: Kluwer Academic Publishers.
Armstrong, David M. 2004. Truth and Truthmakers. Cambridge: Cambridge
University Press.
Arnowitt, R., S. Deser, and C.W. Misner. 2008. “Republication of: The Dynamics
of General Relativity.” General Relativity and Gravitation 40 (9): 1997–2027.
Arntzenius, Frank. 2012. Space, Time, and Stuff. New York: Oxford University
Press.
Aronson, Jerrold L. 1971. “On the Grammar of ‘Cause’.” Synthese 22 (3–4):
414–430.
Aronson, Jerrold L. 1982. “Untangling Ontology From Epistemology in Causation.”
Erkenntnis 18 (3): 293–305.
Bibliography 311
Audi, Robert. 2002. “The Sources of Knowledge.” In The Oxford Handbook of
Epistemology, edited by Paul K. Moser, 71–94. New York: Oxford University
Press.
Audi, Robert. 2011. Epistemology: A Contemporary Introduction to the Theory of
Knowledge. 3rd Edition. New York: Routledge.
Aune, Bruce. 1977. Reason and Action. Boston, MA: D. Reidel.
Baetu, Irina, and A.G. Baker. 2009. “Human Judgments of Positive and Negative
Causal Chains.” Journal of Experimental Psychology: Animal Behaviour
Processes 35 (2): 153–168.
Baez, J., and J.P. Muniain. 1994. Gauge Fields, Knots and Gravity. River Edge, NJ:
World Scientific.
Baker, Lynne Rudder. 1987. Saving Belief: A Critique of Physicalism. Princeton, NJ:
Princeton University Press.
Balashov, Y., and M. Janssen, 2003. “Review: Presentism and Relativity.” The
British Journal for the Philosophy of Science 54, 327–346.
Banks, Thomas. 2008. Modern Quantum Field Theory: A Concise Introduction.
New York: Cambridge University Press.
Banks, Thomas. 2015. “Holographic Inflation and the Low Entropy of the
Early Universe.” arXiv.org. Accessed January 12, 2015. https://arxiv.org/
abs/1501.02681.
Banks, Thomas. 2017. “Holographic Inflation Revised.” In The Philosophy of
Cosmology, edited by Khalil Chamcham, Joseph Silk, John D. Barrow, and Simon
Saunders, 241–262. New York: Cambridge University Press.
Barbour, Julian B. 2001. The Discovery of Dynamics: A Study From a Machian
Point of View of the Discovery and the Structure of Dynamical Theories. New
York: Oxford University Press.
Barrow, Isaac. 2013. The Mathematical Works of Isaac Barrow. Edited for Trinity
College by William Whewell. Cambridge: Cambridge University Press.
Barwise, Jon., and Robin Cooper. 1981. “Generalized Quantifiers and Natural
Language.” Linguistics and Philosophy 4 (2): 159–219.
Bealer, George. 2002. “Modal Epistemology and the Rationalist Renaissance.”
In Conceivability and Possibility, edited by Tamar Szabó Gendler and John
Hawthorne, 71–125. New York: Oxford University Press.
Beall, J.C., and G. Restall. 2006. Logical Pluralism. Oxford: Clarendon Press.
Beaney, Michael. 2016. “Conceptions of Analysis in Analytic Philosophy: Supplement
to Analysis” In The Stanford Encyclopedia of Philosophy, edited by Edward N.
Zalta. June 21. https://plato.stanford.edu/archives/sum2016/entries/analysis/.
Beck, Jacob Scott 2013. “Atomism About Concept.” In Encyclopedia of the Mind,
edited by Harold Pashler. Thousand Oaks, CA: Sage Publications. http://dx.doi.
org/10.4135/9781452257044.n20
Bedingham, Daniel. 2011. “Relativistic State Reduction Dynamics.” Foundations of
Physics 41 (4): 686–704.
Beebee, Helen. 1998. “Review: Causation and Persistence: A Theory of Causation
by Douglas Ehring.” The British Journal for the Philosophy of Science 49 (1):
181–184.
Beebee, Helen. 2003. “Seeing Causing.” Proceedings of the Aristotelian Society 103
(3): 257–280.
Beebee, Helen, and Julian Dodd (editors). 2005. Truthmakers: The Contemporary
Debate. Oxford: Clarendon Press.
312 Bibliography
Beebee, H., C. Hitchcock, and P. Menzies. 2009. “Introduction.” In The Oxford
Handbook of Causation, edited by H. Beebee, C. Hitchcock, and P. Menzies,
1–18. New York: Oxford University Press.
Bell, J.S. 1964. “On the Einstein Podolsky Rosen Paradox.” Physics 1 (3): 195–200.
Bell, J.S. 2004. Speakable and Unspeakable in Quantum Mechanics. 2nd Edition
With a new introduction by Alain Aspect. New York: Cambridge University Press.
Belot, Gordon. 2011. Geometric Possibility. New York: Oxford University Press.
Ben-Menahem, Yemima. 1993. “Struggling With Causality: Einstein’s Case.”
Science in Context 6 (1): 291–310.
Bennett, Jonathan. 1984. “Counterfactuals and Temporal Direction.” The
Philosophical Review 93 (1): 57–91.
Bennett, Jonathan. 1988. Events and Their Names. Indianapolis, IN: Hackett
Publishing.
Bennett, Jonathan. 1991. “Précis of Events and Their Names.” Philosophy and
Phenomenological Research 51 (3): 625–628.
Bennett, Karen. 2003. “Why the Exclusion Problem Seems Intractable, and How,
Just Maybe, to Tract It.” Noûs 37 (3): 471–497.
Bennett, Karen. 2011. “Construction Area (No Hard Hat Required).” Philosophical
Studies 154 (1): 79–104.
Bennett, Karen. 2017. Making Things Up. Oxford: Oxford University Press.
Bergmann, Michael. 2006. Justification Without Awareness: A Defense of Epistemic
Externalism. New York: Oxford University Press.
Bergmann, Peter G. 1961. “Observables in General Relativity.” Reviews of Modern
Physics 33 (4): 510–514.
Bernstein, Sara. 2014. “Omissions as Possibilities.” Philosophical Studies 167 (1):
1–23.
Bernstein, Sara. 2015. “The Metaphysics of Omissions.” Philosophy Compass 10
(3): 208–218.
Bhattacharyya, J., M. Colombo, and T.P. Sotiriou. 2016. “Causality and Black
Holes in Spacetimes with a Preferred Foliation.” Classical and Quantum Gravity
33: 235003.
Bickle, J. 2016. “Multiple Realizability.” In The Stanford Encyclopedia of
Philosophy, edited by Edward N. Zalta. Stanford, CA. https://plato.stanford.edu/
archives/spr2016/entries/multiple-realizability/.
Biggs, Stephen, and Jessica M. Wilson. 2017. “The A Priority of Abduction.”
Philosophical Studies 174 (3): 735–758.
Bird, Alexander. 2007. Nature’s Metaphysics: Laws and Properties. New York:
Oxford University Press.
Birkhoff, G.D. 1944. “Newtonian and Other Forms of Gravitational Theory I.” The
Scientific Monthly 58 (1): 49–57.
Birnbacher, D., and D. Hommen. 2013. “Omissions as Causes-Genuine, Quasi,
or Not at All?” In Critical Essays on “Causation and Responsibility,” edited by
Benedikt Kahmen and Markus Stepanians, 133–156. Walter De Gruyter and Co.
Black, Max. 1952. “The Identity of Indiscernibles.” Mind 61 (242): 153–164.
Blackburn, Simon. 1984. Spreading the Word: Groundings in the Philosophy of
Language. Oxford: Clarendon Press.
Blair, D.G., L. Ju, C. Zhao, and E.J. Howell. 2012. “Gravitational Waves.” In
Advanced Gravitational Wave Detectors, edited by D.G. Blair, E.J. Howell, L. Ju,
and C. Zhao, 3–15. New York: Cambridge University Press.
Bibliography 313
Blanchard, Thomas. 2015. “Review of Mathias Frisch, Causal Reasoning in Physics.”
Notre Dame Philosophical Reviews: An Electronic Journal. June 12. Accessed
December 1, 2017. http://ndpr.nd.edu/news/causal-reasoning-in-physics/.
Block, Ned. 1992. “Begging the Question Against Phenomenal Consciousness.”
Behavioral and Brain Sciences 15: 205–206.
Block, Ned. 2007. “Troubles with Functionalism.” In Consciousness, Function, and
Representation: Collected Papers, Volume 1, by Ned Block, 63–101. Cambridge,
MA: MIT Press.
Bohm, David. 1957. Causality and Chance in Modern Physics. Foreword by Louis
de Broglie. Princeton, NJ: D. Van Nostrand Company, Inc.
Bohm, David. 2005. Causality and Chance in Modern Physics. Published in
Taylor & Francis e-Library. London: Routledge.
Bohr, Niels. 1985. Collected Works. Edited by J. Kalckar. Vol. 6. Amsterdam:
North-Holland.
Boltzmann, Ludwig. 1909. Wissenschaftliche Abhandlungen. Vol. 2. Leipzig: Verlag
von Johann Ambrosius Barth.
Bolzano, Bernard. 2014. Theory of Science: Volume 1. Translated by Paul Rusnock
and Rolf George. Vol. 1. 4 vols. Oxford: Oxford University Press. Cited as
“Theory of Science: Volume 1”
Bolzano, Bernard. 2014. Theory of Science: Volume 2. Translated by Paul Rusnock
and Rolf George. Vol. 2. 4 vols. Oxford: Oxford University Press. Cited as
“Theory of Science: Volume 2”
BonJour, Laurence. 1998. In Defense of Pure Reason: A Rationalist Account of
A Priori Justification. New York: Cambridge University Press.
BonJour, Laurence. 2003. “A Version of Internalist Foundationalism.” In Epistemic
Justification: Internalism vs. Externalism, Foundations vs. Virtues, by L. BonJour
and Ernest Sosa, 3–96. Malden, MA: Blackwell.
BonJour, Laurence. 2005. “Is There a Priori Knowledge?: In Defense of the a pri-
ori” In Contemporary Debates in Epistemology, edited by Matthias Steup and
Ernest Sosa, 98–105. Malden, MA: Blackwell.
Bonk, Thomas. 2008. Underdetermination: An Essay on Evidence and the Limits of
Natural Knowledge. Dordrecht: Springer.
Boyce, Kenny, and Andrew Moon. 2016. “In Defense of Proper Functionalism:
Cognitive Science Takes on Swampman.” Synthese 193 (9): 2987–3001.
Boyd, Richard N. 1984. “The Current Status of Scientific Realism.” In Scientific
Realism, edited by Jarrett Leplin, 41–82. Berkeley: University of California Press.
Boyer, Jennifer L., Stephenie Harrison, Tony Ro, and edited by Lawrence Weiskrantz.
2005. “Unconscious Processing of Orientation and Color without Primary Visual
Cortex.” Proceedings of the National Academy of Sciences of the United States
of America. 102 (46): 16875–16879.
Braithwaite, R.B. 1968. Scientific Explanation: A Study of the Function of Theory,
Probability and Law in Science. London: Cambridge University Press.
Brand, Myles. 1977. “Identity Conditions for Events.” American Philosophical
Quarterly 14 (4): 329–337.
Brewster, D. 2009. “Memoirs of the Life, Writings, and Discoveries of Sir Isaac
Newton Volume 1 (Normalized Text).” Newton Project. September. www.new
tonproject.ox.ac.uk/view/texts/normalized/OTHE00129.
Brickley, Megan, and Rachel Ives. 2006. “Skeletal Manifestations of Infantile
Scurvy.” American Journal of Physical Anthropology 129 (2): 163–172.
314 Bibliography
Brickley, Megan, and Rachel Ives. 2008. The Bioarchaeology of Metabolic Bone
Disease. Oxford: Academic Press.
Briggs, Rachael. 2012. “Interventionist Counterfactuals.” Philosophical Studies
160: 139–166.
Broad, C.D. 1952. Ethics and the History of Philosophy: Selected Essays. London:
Routledge & Kegan Paul Ltd.
Brown, Harvey R. 2005. Physical Relativity: Space-Time Structure From a
Dynamical Perspective. New York: Oxford University Press.
Brown, Harvey R. 2009. “The Behaviour of Rods and Clocks in General Relativity,
and the Meaning of the Metric Field”, arXiv:0911.4440v1 [gr-qc] 23 Nov 2009.
November 23. Accessed December 11, 2017. https://arxiv.org/abs/0911.4440.
Brown, Harvey R., and Dennis Lehmkuhl. 2013. “Einstein, the Reality of Space,
and the Action-Reaction Principle,” arXiv:1306.4902v1 [physics.hist-ph] 20 Jun
2013. Accessed December 11, 2017. https://arxiv.org/abs/1306.4902v1.
Brown, Harvey R., and Oliver Pooley. 2006. “Minkowski Space-Time: A Glorious
Non-Entity.” In The Ontology of Spacetime, edited by D. Dieks, 67–89.
Amsterdam: Elsevier.
Brown, James R. 1986. “Thought Experiments Since the Scientific Revolution.”
International Studies in the Philosophy of Science 1 (1): 1–15.
Brown, James R. 1991. Laboratory of the Mind: Thought Experiments in the
Natural Sciences. London: Routledge.
Brown, James R. 2004. “Why Thought Experiments Transcend Empiricism.”
In Contemporary Debates in Philosophy of Science, edited by Christopher
Hitchcock, 23–43. Malden, MA: Blackwell.
Brush, Stephen G., with Ariel Segal. 2015. Making 20th Century Science: How
Theories Became Knowledge. New York: Oxford University Press.
Burge, Tyler. 2009. “Five Theses on De Re States and Attitudes.” In The Philosophy
of David Kaplan, edited by Joseph Almog and Paolo Leonardi, 246–316. New
Burge, Tyler. 2010. Origins of Objectivity. New York: Oxford University Press.
Burns, J.J. 1959. “Biosynthesis of L-Ascorbic Acid; Basic Defect in Scurvy.”
American Journal of Medicine 26 (5): 740–748.
Byerly, H. 1979. “Substantial Causes and Nomic Determination.” Philosophy of
Science 46 (1): 57–81.
Calinger, Ronald S. 2016. Leonhard Euler: Mathematical Genius in the
Enlightenment. Princeton, NJ: Princeton University Press.
Callender, Craig. 2011. “The Past Histories of Molecules.” In Probabilities in
Physics, edited by Claus Beisbart and Stephan Hartmann, 83–113. New York:
Cameron, R.P. 2010. “How to Have a Radically Minimal Ontology.” Philosophical
Studies 151 (2): 249–264.
Camilleri, Kristian. 2009. Heisenberg and the Interpretation of Quantum Mechanics:
The Physicist as Philosopher. New York: Cambridge University Press.
Campbell, Keith. 1990. Abstract Particulars. Oxford: Basil Blackwell.
Carnap, Rudolf. 1939. Foundations of Logic and Mathematics. Vol. 1. n. 3. In
International Encyclopedia of Unified Science. Chicago: University of Chicago
Press.
Carnap, Rudolf. 1956. Meaning and Necessity: A Study in Semantics and Modal
Logic. Enlarged Edition. Chicago: University of Chicago Press.
Bibliography 315
Carpenter, Kenneth J. 1986. The History of Scurvy and Vitamin C. New York:
Cambridge University Press.
Carroll, J.W. 1994. Laws of Nature. New York: Cambridge University Press.
Carroll, J.W. 2009. “Anti-Reductionism.” In The Oxford Handbook of Causation,
edited by Helen Beebee, Christopher Hitchcock, and Peter Menzies, 279–298.
New York: Oxford University Press.
Carroll, Sean M. 2004. Spacetime and Geometry: An Introduction to General
Relativity. New York: Addison Wesley.
Carroll, Sean M. 2008. “What If Time Really Exists?” arXiv.org. November 23.
https://arxiv.org/pdf/0811.3772.pdf.
Carroll, Sean M. 2010. From Eternity to Here: The Quest for the Ultimate Theory
of Time. New York: Plume. Cited as “From Eternity to Here”
Carroll, Sean M. 2010. Energy Is Not Conserved. February 22. Accessed
December 1, 2017. www.preposterousuniverse.com/blog/2010/02/22/energy-is-
not-conserved/. Cited as “Energy Is Not Conserved”
Carroll, Sean M., and Jennifer Chen. 2004. “Spontaneous Inflation and the Origin
of the Arrow of Time.” Vers. 1. arXiv.org. October 27. https://arxiv.org/pdf/
hep-th/0410270.pdf.
Carruthers, P. 1992. Human Knowledge and Human Nature: A New Introduction
to an Ancient Debate. Oxford: Oxford University Press.
Cartwright, Nancy. 1979. “Causal Laws and Effective Strategies.” Noûs 13 (4):
419–437.
Cartwright, Nancy. 1983. How the Laws of Physics Lie. New York: Oxford
University Press.
Cartwright, Nancy. 1993. “In Defence of ‘This Worldly’ Causality: Comments
on van Fraassen’s Laws and Symmetry.” Philosophy and Phenomenological
Research 53 (2): 423–429.
Cartwright, Nancy. 1999. The Dappled World: A Study of the Boundaries of
Science. Cambridge: Cambridge University Press.
Cartwright, Nancy. 2007. “Counterfactuals in Economics: A Commentary.” In
Causation and Explanation, edited by J.K. Campbell, M. O’Rourke, and H.
Silverstein, 191–216. Cambridge, MA: MIT Press.
Casullo, Albert. 2003. A Priori Justification. New York: Oxford University Press.
Cavendish, Henry. 2010. The Scientific Papers of the Honourable Henry Cavendish,
F.R. S: Chemical and Dynamical. Edited by Sir Edward Thorpe. Vol. 2. Cambridge:
Cei, Angelo. 2005. “Structural Distinctions: Entities, Structures, and Changes in
Science.” Philosophy of Science 72 (5): 1385–1396.
Chakravartty, Anjan. 2007. A Metaphysics for Scientific Realism: Knowing the
Unobservable. Cambridge: Cambridge University Press.
Chalmers, David J. 1996. The Conscious Mind: In Search of a Fundamental Theory.
Chalmers, David J. 2002. “Does Conceivability entail Possibility?” In Conceivability
and Possibility, edited by T.S. Gendler and J. Hawthorne, 145–200. New York:
Chalmers, David J., and Frank Jackson. 2001. “Conceptual Analysis and Reductive
Explanation.” The Philosophical Review 110 (3): 315–360.
Chase, Carl Trueblood. 1932. A History of Experimental Physics. New York: D.
Van Nostrand Company, Inc.
316 Bibliography
Chen, Eddy Keming. 2017. “Our Fundamental Physical Space: An Essay on the
Metaphysics of the Wave Function.” The Journal of Philosophy CXIV (7):
333–365.
Chihara, C.S., and Jerry A. Fodor. 1965. “Operationalism and Ordinary Language:
A Critique of Wittgenstein.” American Philosophical Quarterly 2 (4): 281–295.
Chisholm, Roderick M. 1966. “Freedom and Action.” In Freedom and Determinism,
edited by Keith Lehrer, 11–44. New York: Random House.
Chisholm, Roderick M. 1990. “Events Without Times: An Essay on Ontology.”
Noûs 24 (3): 413–427.
Choquet-Bruhat, Yvonne. 2009. General Relativity and the Einstein Equations.
Church, Alonzo. 1996. Introduction to Mathematical Logic. Princeton, NJ:
Church, Alonzo. 2009. “Referee Reports on Fitch’s ‘A Definition of Value’.” In
New Essays on the Knowability Paradox, edited by Joe Salerno, 13–20. New
Churchland, Paul M. 1979. Scientific Realism and the Plasticity of Mind. Cambridge:
Churchland, Paul M. 1995. “Eliminative Materialism and Propositional Attitudes.”
In Contemporary Materialism: A Reader, edited by Paul K. Moser and J.D. Trout,
150–168. London: Routledge. Cited as “Materialism”
Churchland, Paul M. 1995. “Postscript: Evaluating Our Self Conception.” In
Contemporary Materialism: A Reader, edited by Paul K. Moser and J.D. Trout,
168–179. Cited as “Postscript”
Churchland, Paul M. 1998. “Folk Psychology.” In On the Contrary: Critical
Essays, 1987–1997, by Paul M. Churchland and Patricia S. Churchland, 3–15.
Cambridge, MA: MIT Press.
Churchland, Paul M. 2007. “The Evolving Fortunes of Eliminative Materialism.”
In Contemporary Debates in Philosophy of Mind, edited by Brian P. McLaughlin
and Jonathan Cohen, 160–181. Malden, MA: Blackwell.
Churchland, Paul M. 2013. Matter and Consciousness. 3rd Edition. Cambridge,
MA: MIT Press.
Clapham, Christopher. 1996. The Concise Oxford Dictionary of Mathematics. 2nd
Edition. New York: Oxford University Press.
Clarke, Randolph. 1993. “Toward a Credible Agent-Causal Account of Free Will.”
Noûs 27 (2): 191–203.
Clarke, Randolph. 1996. “Agent Causation and Event Causation.” Philosophical
Topics 24 (2): 19–48.
Clarke, Randolph. 2002. “Libertarianism Views: Critical Survey of Noncausal and
Event-Causal Accounts of Free Agency.” In The Oxford Handbook of Free Will,
edited by Robert Kane, 356–385. New York: Oxford University Press.
Clarke, Randolph. 2003. Libertarian Accounts of Free Will. New York: Oxford
University Press.
Clarke, Randolph. 2014. Omissions: Agency, Metaphysics, and Responsibility.
Clarke, Randolph. forthcoming. “Free Will, Agent Causation, and ‘Disappearing
Agents’.” Noûs. doi: 10.1111/nous.12206: 1–21.
Cleland, C. 1991. “On the Individuation of Events.” Synthese 86 (2): 229–254.
Climenhaga, Nevin. 2017. “Inference to the Best Explanation Made Incoherent.”
The Journal of Philosophy 114 (5): 251–273.
Bibliography 317
Climenhaga, Nevin. forthcoming. “Intuitions Are Used as Evidence in Philosophy.”
Mind, 1–36.
Coady, C.A.J. 1992. Testimony: A Philosophical Study. Oxford: Clarendon Press.
Cohen, I.B. 1999. “A Guide to Newton’s Principia.” In The Principia: Mathematical
Principles of Natural Philosophy, edited by Isaac Newton, translated by I.
Bernard Cohen, Anne Whitman, and J. Budenz, 1–370. Berkeley and Los Angeles:
University of California Press.
Coleman, Robert A., and Herbert Korté. 2001. “Hermann Weyl: Mathematician,
Physicist, Philosopher.” In Hermann Weyl’s Raum-Zeit-Materie and a General
Introduction to His Scientific Work, edited by E. Scholz, 161–386. Basel, AG:
Springer.
Coleman, S., and D.J. Gross. 1973. “Price of Asymptotic Freedom.” Physical
Review Letters 31: 851–854.
Collett, B., P. Pearle, F. Avignone, and S. Nussinov. 1995. “Constraint on Collapse
Models by Limit on Spontaneous X-Ray Emission in Ge.” Foundations of Physics
25 (10): 1399–1412.
Collier, J. 1999. “Causation Is the Transfer of Information.” In Causation, Natural
Laws and Explanation, edited by Howard Sankey, 279–331. Dordrecht: Kluwer.
Collins, J., N. Hall, and L.A. Paul. 2004. “Counterfactuals and Causation: History,
Problems, and Prospects.” In Causation and Counterfactuals, edited by J. Collins,
N. Hall, and L.A. Paul, 1–57. Cambridge, MA: MIT Press.
Coltheart, Max. 2010. “The Neuropsychology of Delusions.” Annals of the New
York Academy of Sciences 1191 (1): 16–26.
Coltheart, Max, Robyn Langdon, and Ryan McKay. 2011. “Delusional Belief.”
Annual Review of Psychology 62: 271–298.
Compton, Arthur H. 1928. “On the Interaction Between Radiation and Electrons.”
Physical Review 31: 59–65.
Connors, Michael H., and Max Coltheart. 2011. “On the Behaviour of Senile
Dementia Patients vis-à-vis the Mirror: Ajuriaguerra, Strejilevitch and Tissot
(1963).” Neuropsychologia 49: 1679–1692.
Connors, Michael H., and Peter W. Halligan. 2015. “A Cognitive Account of Belief:
A Tentative Road Map.” Frontiers in Psychology 5: 1588.
Cooper, Joel. 2012. “Cognitive Dissonance Theory.” In Handbook of Theories of
Social Psychology, Vol. 1. Edited by Paul A.M. Van Lange, Arie W. Kruglanske,
and E. Tory Higgins, 377–397. London: Sage.
Craig, William L. 2001. Time and the Metaphysics of Relativity. Dordrecht: Kluwer
Academic Publishers.
Craver, C.F. 2008. “Structures of Scientific Theories.” In The Blackwell Guide to
the Philosophy of Science, edited by Peter Machamer and Michael Silberstein,
55–79. Malden, MA: Blackwell.
Curiel, Erik. 2014. “A Primer on Energy Conditions”, arXiv:1405.0403v1 [physics.
hist-ph] 30 Apr 2014.
Cushing, James T. 1998. Philosophical Concepts in Physics: The Historical Relation
Between Philosophy and Scientific Theories. New York: Cambridge University
Press.
Dürr, D., S. Goldstein, and N. Zanghí. 1992. “Quantum Equilibrium and the
Origin of Absolute Uncertainty.” Journal of Statistical Physics 67 (5/6): 843–907.
Dürr, D., S. Goldstein, T. Norsen, W. Struyve, and N. Zanghí. 2014. “Can
Bohmian Mechanics Be Made Relativistic?” Proceedings of the Royal Society A
470: 20130699.
318 Bibliography
Daintith, John (editor). 2005. A Oxford Dictionary of Physics. 5th Edition. New
Danks, David. 2009. “The Psychology of Causal Perception and Reasoning.” In The
Oxford Handbook of Causation, edited by Helen Beebee, Christopher Hitchcock,
and Peter Menzies, 447–470. New York: Oxford University Press.
Darrigol, Olivier. 2000. Electrodynamics From Ampère to Einstein. New York:
Darrigol, Olivier. 2001. “The Historians' Disagreements over the Meaning of
Planck's Quantum.” Centaurus 43: 219–239.
Darrigol, Olivier. 2014. Physics and Necessity: Rationalist Pursuits From the
Cartesian Past to the Quantum Present. New York: Oxford University Press.
Dasgupta, Shamik. 2014. “On the Plurality of Grounds.” Philosophers’ Imprint 14
(20): 1–28.
Davidson, Donald. 1967. “Truth and Meaning.” Synthese 17 (1): 304–323. Cited
as “Truth and Meaning”
Davidson, Donald. 1967. “Causal Relations.” The Journal of Philosophy 64 (21):
691–703.
Davidson, Donald. 1984. “Semantics for Natural Languages.” In Inquiries Into
Truth and Interpretation, 2nd Edition. by Donald Davidson, 55–64. New York:
Davidson, Donald. 1987. “Knowing One’s Own Mind.” Proceedings and Addresses
of the American Philosophical Association 60 (3): 443–444.
Davidson, Donald. 1993. “Thinking Causes.” In Mental Causation, edited by John
Heil and Alfred Mele, 3–17. New York: Oxford University Press.
Davidson, Donald. 2001. Essays on Actions and Events. 2nd Edition. New York:
Davidson, Donald. 2001. “The Individuation of Events.” In Essays on Actions and
Events. 2nd Edition, 163–180. New York: Oxford University Press. Cited as
“Events”
Davidson, Donald. 2001. “Mental Events.” In Essays on Actions and Events, 2nd
Edition, edited by Donald Davidson, 207–225. New York: Oxford University
Press. Cited as “Mental Events”
De Gandt, F. 1995. Force and Geometry in Newton’s Principia. Translated by
Curtis Wilson. Princeton, NJ: Princeton University Press.
de la Peña, Luis, and Ana María Cetto. 1996. The Quantum Dice: An Introduction
to Stochastic Electrodynamics. Dordrecht: Springer.
De Muijnck, W. 2003. Dependences, Connections, and Other Relations: A Theory
of Mental Causation. Dordrecht: Kluwer Academic Publishers.
Demopoulos, William, and Michael Friedman. 1985. “Bertrand Russell’s The
Analysis of Matter: Its Historical Context and Contemporary Interest.” Philosophy
of Science 52 (4): 621–639.
Dennett, Daniel. 1996. “Facing Backwards on the Problem of Consciousness.”
Journal of Consciousness Studies 3 (1): 4–6.
Dennett, Daniel, and M. Kinsbourne. 1995. “Time and the Observer: The Where
and When of Consciousness in the Brain.” Behavioral and Brain Sciences 15:
183–247.
Descartes, R. 1970. Philosophical Letters. Translated and edited by Anthony Kenny.
Oxford: Clarendon Press.
Bibliography 319
Descartes, R. 1998. The World and Other Writings. Edited by Stephen Gaukroger.
Translated by Stephen Gaukroger. Cambridge: Cambridge University Press.
d'Espagnat, Bernard. 1979. “The Quantum Theory and Reality.” Scientific
American November. 158–181.
Devitt, Michael. 1991. Realism and Truth. Oxford: Blackwell.
Devitt, Michael. 2005. “Scientific Realism.” In The Oxford Handbook of
Contemporary Philosophy, edited by Frank Jackson and Michael Smith, 767–
790. New York: Oxford University Press.
Devitt, Michael. 2014. “Realism/Anti-Realism”, In The Routledge Companion to
Philosophy of Science. 2nd Edition, edited by Martin Curd and Stathis Psillos,
256–267. New York: Routledge.
Dirac, P.A.M. 1927. “The Quantum Theory of Dispersion.” Proceedings of the
Royal Society 114: 710–728. Cited as “Dispersion”
Dirac, P.A.M. 1927. “The Quantum Theory of the Emission and Absorption of
Radiation.” Proceedings of the Royal Society 114: 243–265. Cited as “Emission
and Absorption”
DiSalle, Robert. 1994. “On Dynamics, Indiscernibility, and Spacetime Ontology,”
The British Journal for the Philosophy of Science 45 (1): 265–287.
DiSalle, Robert. 1995. “Spacetime Theory as Physical Geometry.” Erkenntnis 42:
317–337.
Dodd, Julian. 2007. Works of Music: An Essay in Ontology. Oxford: Oxford
University Press.
Dodelson, Scott. 2017. Gravitational Lensing. New York: Cambridge University Press.
Dorr, Cian. 2002. The Simplicity of Everything. PhD Thesis, Princeton University.
Douven, I. 2014. “Underdetermination.” In The Routledge Companion to
Philosophy of Science, 2nd Edition, edited by Martin Curd and Stathis Psillos,
336–345. New York: Routledge.
Dowe, Phil. 1992. “Process Causality and Asymmetry.” Erkenntnis 37: 179–196.
Dowe, Phil. 2000. Physical Causation. Cambridge: Cambridge University Press.
Dowe, Phil. 2004. “Causes Are Physically Connected to Their Effects: Why Preventers
and Omissions Are Not Causes.” In Contemporary Debates in Philosophy of
Science, edited by Christopher Hitchcock, 189–196. Malden, MA: Blackwell.
Driver, R., E. Guesne, and A. Tiberghien. 1985. “Some Features of Children’s Ideas
and Their Implications for Teaching.” In Childrens Ideas in Science, edited by
Rosalind Driver, Edith Guesne, and Andrée Tiberghien, 193–201. Buckingham:
Open University Press.
Drummond, Jack Cecil. 1919. “Note on the Rôle of the Antiscorbutic Factor in
Nutrition.” Biochemical Journal 13 (1): 77–80.
Ducasse, C.J. 1968. Truth, Knowledge and Causation. London: Routledge & Kegan
Paul.
Duhem, P. 1954. The Aim and Structure of Physical Theory. 2nd Edition. Translated
by P.W. Wiener. Princeton, NJ: Princeton University Press.
Dummett, A.E. 1954. “Can An Effect Precede Its Cause?” Proceedings of the
Aristotelian Society Suppl 28: 27–44.
Dummett, Michael. 1978. Truth and Other Enigmas. 3rd Edition. Cambridge, MA:
Dyson, Freeman J. 1949. “The Radiation Theories of Tomonaga, Schwinger, and
Feynman.” Physical Review 75 (3): 486–502. Cited as “Radiation Theories”
320 Bibliography
Dyson, Freeman J. 1949. “The S Matrix in Quantum Electrodynamics.” Physical
Review 75 (11): 1736–1755. Cited as “S Matrix”
Earman, John. 1976. “Causation: A Matter of Life and Death.” The Journal of
Philosophy 73: 5–25.
Earman, John. 1989. World Enough and Spacetime: Absolute Versus Relational
Theories of Space and Time. Cambridge, MA: MIT Press.
Earman, John. 1992. Bayes or Bust? A Critical Examination of Bayesian
Confirmation Theory. Cambridge, MA: MIT Press.
Earman, John. 1995. Bangs, Crunches, Whimpers, and Shrieks: Singularities and
Acausalities in Relativistic Spacetimes. New York: Oxford University Press.
Earman, John. 2002. “Thoroughly Modern McTaggart: Or What McTaggart
Would Have Said If He Had Read the General Theory of Relativity.” Philosophers
Imprint 2 (3): 1–28.
Earman, John, and Michael Friedman. 1973. “The Meaning and Status of Newton’s
Law of Inertia and the Nature of Gravitational Forces.” Philosophy of Science
40 (3): 329–359.
Earman, John, Chris Smeenk, and Christian Wüthrich. 2009. “Do the Laws of
Physics Forbid the Operation of Time Machines?” Synthese 169 (1): 91–124.
Eddington, A.S. 2014. The Mathematical Theory of Relativity. Lexington, KY:
Bibliolife.
Eddon, M. 2013. “Quantitative Properties.” Philosophy Compass 8 (7): 633–645.
Eells, Ellery. 1991. Probabilistic Causality. Cambridge: Cambridge University Press.
Eells, Ellery, and Elliott Sober. 1983. “Probabilistic Causality and the Question of
Transitivity.” Philosophy of Science 50 (1): 35–57.
Egan, Louisa C., Laurie R. Santos, and Paul Bloom. 2007. “The Origins of Cognitive
Dissonance: Evidence From Children and Monkeys.” Psychological Science 18
(11): 978–983.
Ehlers, Jürgen. 1973. “The Nature and Structure of Spacetime”, In The Physicist’s
Conception of Nature, edited by Jagdish Mehra, 71–91. Dordrecht: Reidel.
Ehlers, Jürgen. 1987. “Folklore in Relativity and What Is Really Known.” In General
Relativity and Gravitation: Proceedings of the 11th International Conference on
General Relativity and Gravitation, 61–71. New York: Cambridge University
Press.
Ehring, Douglas. 1987. “Causal Relata.” Synthese 73 (2): 319–328.
Ehring, Douglas. 1997. Causation and Persistence: A Theory of Causation. New
Ehring, Douglas. 2009. “Causal Relata.” In The Oxford Handbook of Causation,
edited by Helen Beebee, Christopher Hitchcock, and Peter Menzies. 387–413.
Einstein, Albert. 1905. “Zur Elektrodynamik bewegter Körper.” Annalen Der
Physik 322 (10): 891–921. Cited by title.
Einstein, Albert. 1905. “Über einen die Erzeugung und Verwandlung des Lichtes
betreffenden heuristischen Gesichtspunkt.” Annalen der Physik 322 (6): 132–
148. Cited by title.
Einstein, Albert. 1929. “Ansprache von Prof. Einstein an Prof. Planck.”
Sitzungsberichte Wissenschaftlicher Körperschafen. Forschungen und Fortschritte.
Nachrichtenblatt Der Deutschen Wissenschaft und Technik, 248–249.
Einstein, Albert. 1949. “Remarks Concerning the Essays Brought Together in This
Co-operative Volume.” In Albert Einstein: Philosopher-Scientist, edited by Paul
Arthur Schilpp, 665–688. Evanston, IL: The Library of Living Philosophers, Inc.
Bibliography 321
Einstein, Albert. 1950. The Meaning of Relativity. 3rd Edition including The
Generalized Theory of Gravitation. Translated by Edwin P. Adams, Appendix
I translated by Ernst G. Straus and Appendix II translated by Sonja Bargmann.
Princeton, NJ: Princeton University Press.
Einstein, Albert. 1952. “The Foundation of the General Theory of Relativity.” In
The Principle of Relativity: A Collection of Original Papers on the Special and
General Theory of Relativity. Notes by A. Sommerfeld, edited by H.A. Lorentz,
Hermann Weyl, and Hermann Minkowski, translated by W. Perrett and G.B.
Jeffery, 111–164. New York: Dover.
Einstein, Albert. 1954. Ideas and Opinions. New Translations and Revisions by
Sonja Bargmann. Based on Mein Weltbild edited by Carl Seelig and other sources.
New York: Bonzanza Books.
Einstein, Albert. 1997. The Collected Papers of Albert Einstein: Volume 6 (English): The
Berlin Years: Writings, 1914–1917. (English Translation Supplement) Translated
by Alfred Engel. Engelbert Schucking, Consultant. Princeton, NJ: Princeton
University Press.
Einstein, Albert. 2002. The Collected Papers of Albert Einstein: Volume 7, The
Berlin Years: Writings, 1918–1921. (English Translation of Selected Texts).
English translation by Alfred Engel. Engelbert Schucking, Consultant. Princeton,
NJ: Princeton University Press.
Einstein, Albert. 2006. “Letter to Moritz Schlick [Berlin,] June 1920”, Document
47. In The Collected Papers of Albert Einstein: Volume 10 The Berlin Years:
Correspondence, May-December 1920 and Supplementary Correspondence, 1909-
1920, edited by Diana Kormos Buchwald, Tilman Sauer, Ze’ev Rosenkrantz, József
Illy, and Virginia Iris Holmes. Translated by Ann Hentschel. Klaus Hentschel,
Consultant. Princeton, NJ: Princeton University Press, 186–187.
Einstein, Albert, and J. Grommer. 1927. “Allgemeine Relativitätstheorie und
Bewegungsgesetz.” Sitzungsberichte der Preussischen Akademie der Wissenschaften,
Physikalisch-Mathematische Klasse. 2–13.
Einstein, Albert, and Leopold Infeld. 1938. The Evolution of Physics: The Growth of
Ideas From Early Concepts to Relativity and Quanta. New York: Simon & Schuster.
Einstein, Albert, and L. Infeld. 1940. “The Gravitational Equations and the Problem
of Motion II.” Annals of Mathematics. Second Series 41 (2): 455–464.
Einstein, Albert, and L. Infeld. 1949. “On the Motion of Particles in General
Relativity Theory.” Canadian Journal of Mathematics 1: 209–241.
Einstein, Albert, L. Infeld, and B. Hoffmann. 1938. “The Gravitational Equations and
the Problem of Motion.” Annals of Mathematics. Second Series 39 (1): 65–100.
Elga, Adam. 2001. “Statistical Mechanics and the Asymmetry of Counterfactual
Dependence.” Philosophy of Science 68 (S1): S313–S324.
Elliot, Andrew J., and Patricia G. Devine. 1994. “On the Motivational Nature
of Cognitive Dissonance: Dissonance as Psychological Discomfort.” Journal of
Personality and Social Psychology 67 (3): 382–394.
Ellis, Brian. 2002. The Philosophy of Nature: A Guide to the New Essentialism.
Montreal & Kingston: McGill-Queen’s University Press.
Ellis, G.F.R., B.A.C.C. Bassett, and P.K.S. Dunsby. 1998. “Lensing and Caustic
Effects on Cosmological Distances.” Classical and Quantum Gravity 15:
2345–2361.
Ellis, G.F.R., and W.R. Stoeger. 2009. “The Evolution of Our Local Cosmic
Domain: Effective Causal Limits.” Monthly Notices of the Royal Astronomical
Society 398: 1527–1536.
322 Bibliography
Elster, Jon. 1983. Explaining Technical Change: A Case Study in the Philosophy of
Science. New York: Cambridge University Press.
Euler, Leonard. 1837. Letters of Euler on Different Subjects in Natural Philosophy.
Vol. 1. New York: Harper & Brothers.
Faddeev, L.D., and A.A. Slavnov. 1991. Gauge Fields: Introduction to Quantum
Theory. 2nd Edition. Translated from the Russian Edition by G.B. Pontecorvo.
Reading, MA: Benjamin/Cummings Publishing Company.
Fair, David. 1979. “Causation and the Flow of Energy.” Erkenntnis 14 (3): 219–250.
Fales, Evan. 1990. Causation and Universals. New York: Routledge.
Fara, Michael, and Timothy Williamson. 2005. “Counterparts and Actuality.”
Mind 114 (453): 1–30.
Fear, Christopher, Helen Sharp, and David Healy. 1996. “Cognitive Processes in
Delusional Disorders.” The British Journal of Psychiatry 168 (1): 61–67.
Feigl, Herbert. 1970. “The ‘Orthodox’ View of Theories: Remarks in Defense as Well
as Critique.” In Analyses of Theories and Methods of Physics and Psychology,
edited by Michael Radner and Stephen Winokur. 3–16. Minneapolis: University
of Minnesota Press.
Ferronsky, V.I. 2016. Gravitation, Inertia and Weightlessness: Centrifugal and
Gyroscopic Effects of the n-Body System’s Interaction Energy. Switzerland:
Springer.
Festinger, Leon. 1957. A Theory of Cognitive Dissonance. Stanford, CA: Stanford
University Press.
Feynman, Richard P. 1949. “Space-Time Approach to Quantum Electrodynamics.”
Physical Review 76 (6): 769–789. Cited as “Space-Time Approach to Quantum
Electrodynamics”
Feynman, Richard P. 1949. “The Theory of Positrons.” Physical Review 76 (6):
749–759. Cited as “Theory of Positrons”
Feynman, Richard. 1985. QED: The Strange Theory of Light and Matter. Princeton,
NJ: Princeton University Press.
Feynman, Richard P., Robert B. Leighton, and Matthew Sands. 2010. The
Feynman Lectures on Physics: Volume III Quantum Mechanics. New York:
Basic Books.
Field, Hartry. 2003. “Causation in a Physical World.” In The Oxford Handbook
of Metaphysics, edited by Michael J. Loux and Dean W. Zimmerman, 435–460.
Fine, Kit. 1995. “Ontological Dependence.” Proceedings of the Aristotelian Society
95: 269–290.
Fine, Kit. 2001. “The Question of Realism.” Philosophers’ Imprint 1 (1): 1–30.
Fine, Kit. 2012. “Counterfactuals Without Possible Worlds.” The Journal of
Philosophy CIX (3): 221–246.
Fitch, F.B. 1963. “A Logical Analysis of Some Value Concepts.” The Journal of
Symbolic Logic 28 (2): 135–142.
Fitch, G.W. 1994. “Singular Propositions in Time.” Philosophical Studies 73 (2/3):
181–187.
Fock, V. 1959. The Theory of Space, Time and Gravitation, translated by N.
Kemmer. London: Pergamon Press Ltd.
Fodor, Jerry A. 1974. “Special Sciences (Or: The Disunity of Science as a Working
Hypothesis).” Synthese 28 (2): 97–115.
Fodor, Jerry A. 1975. The Language of Thought. Cambridge, MA: Harvard
University Press.
Bibliography 323
Fodor, Jerry A. 1987. Psychosemantics: The Problem of Meaning in the Philosophy
of Mind. Cambridge, MA: MIT Press.
Fodor, Jerry A. 1990. A Theory of Content and Other Essays. Cambridge, MA:
MIT Press.
Fodor, Jerry A. 1998. Concepts: Where Cognitive Science Went Wrong. New York:
Fodor, Jerry A. 2003. Hume Variations. Oxford: Oxford University Press.
Fodor, Jerry A. 2007. “The Revenge of the Given.” In Contemporary Debates in
Philosophy of Mind, edited by Brian P. McLaughlin and Jonathan Cohen, 105–
116. Malden, MA: Blackwell.
Fodor, Jerry A. 2008. LOT2: The Language of Thought Revisited. New York:
Fodor, Jerry A., and Zenon W. Pylyshyn. 2015. Minds Without Meanings: An Essay
on the Content of Concepts. Cambridge, MA: MIT Press.
Foley, J.H., and L. Breslau. 1982. “A New Syndrome of Delusional Misidentification.”
Annals of Neurology 12 (1): 76. (Abstract A12)
Forbes, Graeme. 1989. Languages of Possibility: An Essay in Philosophical Logic.
New York: Basil Blackwell.
Forster, Malcolm. 2006. “Counterexamples to a Likelihood Theory of Evidence.”
Minds and Machines 16 (3): 319–338.
Forster, Malcolm. 2014. “Prediction.” In The Routledge Companion to Philosophy
of Science, 2nd Edition, edited by Martin Curd and Stathis Psillos, 449–457. New
York: Routledge.
Forster, M., and E. Sober. 2004. “Why Likelihood?” In The Nature of Scientific
Evidence: Statistical, Philosophical, and Empirical Considerations, edited by M.
Taper and S. Lee, 153–164. Chicago: University of Chicago Press.
Foster, J., and J.D. Nightingale. 2005. A Short Course in General Relativity. 3rd Edition.
New York: Springer.
Fourier, Joseph. 2007. The Analytical Theory of Heat. New York: Cosimo Classics.
Fox, John F. 1987. “Truthmaker.” Australasian Journal of Philosophy 65 (2):
188–207.
Fox, Robert. 2013. “Laplace and the Physics of Short-Range Forces.” In The Oxford
Handbook of the History of Physics, edited by Jed Z. Buchwald and Robert Fox,
Franklin-Hall, L.R. 2016. “New Mechanistic Explanation and the Need for
Explanatory Constraints.” In Scientific Composition and Metaphysical Ground,
edited by K. Aizawa and C. Gillett, 41–74. London: Palgrave Macmillan.
Frege, Gottlob. 1980. “On Sense and Meaning.” Translated by Max Black. In
Translations from the Philosophical Writings of Gottlob Frege, edited by Peter
Geach and Max Black. Index prepared by E.D. Klemke, 3rd Edition, 56–78.
Oxford: Basil Blackwell.
Frege, Gottlob. 1997. “Thought.” In The Frege Reader, edited by Michael Beaney,
325–345. Translated by Peter Geach and R.H. Stoothoff. Malden, MA: Blackwell.
Cited as “Thought”
Frege, Gottlob. 1997. “A Brief Survey of My Logical Doctrines.” In The Frege
Reader, edited by Michael Beaney, 299–300. Translated by Peter Long and Roger
White. Malden, MA: Blackwell. Cited as “Brief Survey”
Frege, Gottlob. 1997. “Logic in Mathematics.” In The Frege Reader, edited by
Michael Beaney, 308–318. Translated by Peter Long and Roger White. Malden,
MA: Blackwell. Cited as “Logic in Mathematics”
324 Bibliography
Frege, Gottlob. 1997. “The Foundations of Arithmetic: A Logico-Mathematical
Investigation into the Concept of Number.” In The Frege Reader, edited by
Michael Beaney, 84–129. Translated by Michael Beaney. Malden, MA: Blackwell.
Cited as “The Foundations of Arithmetic”
Frege, Gottlob. 1997. “Thought.” In The Frege Reader, edited by Michael Beaney,
325–345. Translated by Peter Geach and R.H. Stoothoff. Malden, MA: Blackwell.
Cited as “Thought”
French, Steven. 2014. “The Structure of Theories.” In The Routledge Companion
to Philosophy of Science, 2nd Edition, edited by Martin Curd and Stathis Psillos,
301–312. New York: Routledge Press. Cited as “Structure of Theories”
French, Steven. 2014. The Structure of the World: Metaphysics and Representation.
Oxford: Oxford University Press. Cited as “Structure”
Friedman, John, Michael S. Morris, Igor D. Novikov, Fernando Encheverria, Gunnar
Klinkhammer, Kip S. Thorne, and Ulvi Yurtsever. 1990. “Cauchy Problem in
Spacetimes with Closed Timelike Curves.” Physical Review D 42 (6): 1915–1930.
Frigg, Roman. 2008. “A Field Guide to Recent Work on the Foundations of
Statistical Mechanics.” In The Ashgate Companion to Contemporary Philosophy
of Physics, edited by Dean Rickles, 99–196. Burlington, VT: Ashgate.
Frigg, Roman, and Charlotte Werndl. 2011. “Entropy: A Guide to the Perplexed.”
In Probabilities in Physics, edited by Claus Beisbart and Stephan Hartmann, 115–
142. New York: Oxford University Press.
Frisch, Mathias. 2005. Inconsistency, Asymmetry, and Non-Locality:
A Philosophical Investigation of Classical Electrodynamics. New York: Oxford
University Press.
Frisch, Mathias. 2007. “Causation, Counterfactuals, and Entropy.” In Causation,
Physics, and the Constitution of Reality: Russell’s Republic Revisited, edited by
Huw Price and Richard Corry, 351–395. New York: Oxford University Press.
Frisch, Mathias. 2009. “ ‘The Most Sacred Tenet’? Causal Reasoning in Physics.”
The British Journal for the Philosophy 60 (3): 459–474.
Frisch, Mathias. 2014. Causal Reasoning in Physics. Cambridge: Cambridge
University Press.
Gale, R., and A. Pruss. 2002. “A Response to Oppy and to Davey and Clifton.”
Religious Studies 38 (1): 89–99.
Galileo, Galilei. 1890–1909. Opere. Edizione Nazionale. Vol. 7. Florence: Barbèra.
Galileo, Galilei. 1914. Dialogues Concerning Two New Sciences. Translated by
Henry Crew and Alfonso de Salvio. New York: Dover.
Galileo, Galilei. 1967. Dialogue Concerning the Two Chief World Systems—
Ptolemaic & Copernican. 2nd Edition Revised Edition. Translated by Stillman
Drake. Foreword by Albert Einstein. Berkeley: University of California Press.
Galison, Peter. 1997. Image & Logic: A Material Culture of Microphysics. Chicago:
University of Chicago Press.
Galison, Peter. 2003. Einstein’s Clocks, Poincaré’s Maps: Empires of Time. New
York: W.W. Norton & Company.
Gamut, L.T.F. 1991. Logic, Language, and Meaning. Volume 2: Intensional Logic
and Logical Grammar. Chicago: University of Chicago Press.
Ganson, Dorit. 2008. “Evidentialism and Pragmatic Constraints on Outright
Belief.” Philosophical Studies 139 (3): 441–458.
Gearhart, Clayton A. 2002. “Planck, the Quantum, and the Historians.” Physics in
Perspective 4: 170–215.
Bibliography 325
Gendler, Tamar Szabó. 1998. “Galileo and the Indispensability of Scientific Thought
Experiment.” The British Journal for the Philosophy of Science 49 (3): 397–424.
Geroch, Robert. 1978. General Relativity From A to B. Chicago: University of
Chicago Press.
Geroch, Robert. 2013. General Relativity: 1972 Lecture Notes. Montreal and
Quebec: Minkowski Institute.
Geroch, Robert, and G.T. Horowitz. 1979. “Global Structure of Spacetimes.” In
General Relativity: An Einstein Centenary Survey, edited by S.W. Hawking and
W. Israel, 212–293. Cambridge: Cambridge University Press.
Geroch, Robert, and P.S. Jang. 1975. “Motion of a Body in General Relativity.”
Journal of Mathematical Physics 16: 65–67.
Gettier, Edmund L. 1963. “Is Justified True Belief Knowledge?” Analysis 23 (6):
121–123.
Ghirardi, G. 2016. “Collapse Theories.” In The Stanford Encyclopedia of Philosophy,
edited by Edward N. Zalta. Spring. http://plato.stanford.edu/archives/spr2016/
entries/qm-collapse/.
Ghirardi, G., A. Rimini, and T. Weber. 1986. “Unified Dynamics for Microscopic
and Macroscopic Systems.” Physical Review D 34: 470–491.
Ghirardi, G.C., O. Nicrosini, A. Rimini, and T. Weber. 1988. “Spontaneous Localization
of a System of Identical Particles.” IL Nuovo Cimento B 102 (4): 383–396.
Giere, Ronald N. 1988. Explaining Science: A Cognitive Approach. Chicago:
University of Chicago Press.
Giere, Ronald N. 2011. “Objective Single-Case Probabilities and the Foundations
of Statistics.” In The Philosophy of Probability: Contemporary Readings, edited
by Antony Eagle, 498–510. New York: Routledge.
Gillett, Carl. 2003. “The Metaphysics of Realization, Multiple Realizability, and
the Special Sciences.” The Journal of Philosophy 100 (11): 591–603.
Gillett, Carl. 2016. Reduction and Emergence in Science and Philosophy. Cambridge:
Gillies, Donald. 2016. “The Propensity Interpretation.” In The Oxford Handbook
of Probability and Philosophy, edited by Alan Hájek and Christopher Hitchcock,
Gillmor, C. Stewart. 1971. Coulomb and the Evolution of Physics and Engineering
in Eighteenth-Century France. Princeton, NJ: Princeton University Press.
Giovanelli, Marco. 2013. “Erich Kretschmann as a Proto-Logical-Empiricist:
Adventures and Misadventures of the Point-Coincidence Argument.” Studies in
History and Philosophy of Modern Physics 44 (2): 115–134.
Glennan, Stuart. 2002. “Rethinking Mechanistic Explanation.” Philosophy of
Science 69 (S3): S342–S353.
Glennan, Stuart. 2011. “Singular and General Causal Relations: A Mechanist
Perspective,” In Causality in the Sciences, edited by Phyllis McKay Illari, Federica
Russo, and Jon Williamson, 789–817. New York: Oxford University Press.
Glennan, Stuart. 2017. The New Mechanical Philosophy. New York: Oxford
University Press.
Glennan, Stuart, and Phyllis Illari. 2018. The Routledge Handbook of Mechanisms
and Mechanical Philosophy. New York: Routledge.
Glock, Hans-Johann. 2017. “Impure Conceptual Analysis.” In The Cambridge
Companion to Philosophical Methodology, edited by Giuseppina D’Oro and
Søren Overgaard, 77–100. Cambridge: Cambridge University Press.
326 Bibliography
Gluecksohn-Waelsch, Salome. 1963. “Lethal Genes and Analysis of Differentiation.”
Science 142 (3597): 1269–1276.
Glymour, Bruce. 2007. “In Defense of Explanatory Deductivism.” In Causation and
Explanation, edited by J.K. Campbell, Michael O’Rourke, and Harry Silverstein,
133–153. Cambridge, MA: MIT Press.
Goldberg, Dave. 2017. The Standard Model in a Nutshell. Princeton, NJ: Princeton
University Press.
Goldman, Alvin I. 1970. A Theory of Human Action. Princeton, NJ: Princeton
University Press.
Goldstein, Sheldon. 1998. “Quantum Theory Without Observers.” Physics Today
(Part One): 42–46. Cited as “Part 1”
Goldstein, Sheldon. 1998. “Quantum Theory Without Observers.” Physics Today
(Part Two): 38–42. Cited as “Part 2”
Goldstein, Sheldon. 2001. “Boltzmann’s Approach to Statistical Mechanics.” In
Chance in Physics, edited by J. Bricmont, D. Dürr, M.C. Galavotti, G. Ghirardi,
F. Petruccione, and N. Zanghì, 39–54. Berlin-Heidelberg: Springer-Verlag.
Goldstein, S., R. Tumulka, and N. Zanghì. 2012. “The Quantum Formalism and
the GRW Formalism.” Journal of Statistical Physics 149 (1): 142–201.
Gomes, Anil. 2015. “Testimony and Other Minds.” Erkenntnis 80 (1): 173–183.
Goodman, Nelson. 1955. Fact, Fiction and Forecast. Cambridge, MA: Harvard
University Press.
Goodman, Nelson. 1972. Problems and Projects. Indianapolis, IN: Bobbs-Merrill.
Gott III, J. Richard. 2001. Time Travel in Einstein’s Universe: The Physical
Possibilities of Travel Through Time. New York: Houghton Mifflin.
Gower, Barry. 2000. “Cassirer, Schlick and ‘Structural’ Realism: The Philosophy
of the Exact Sciences in the Background to Early Logical Empiricism.” British
Journal for the History of Philosophy 8 (1): 71–106.
Gowers, Timothy. 2008. “Part I Introduction.” In The Princeton Companion to
Mathematics, edited by Timothy Gowers, associate editors June Barrow-Green
and Imre Leader, 1–76. Princeton, NJ: Princeton University Press.
Grandy, R. 2002. “Many-Valued, Free, and Intuitionistic Logics.” In A Companion
to Philosophical Logic, edited by Dale Jacquette, 531–544. Malden, MA:
Blackwell.
Gravetter, F.J., and L.B. Wallnau. 2013. Statistics for the Behavioral Sciences. 9th
Edition. Belmont, CA: Wadsworth Cengage Learning.
Greiner, Walter. 1997. Relativistic Quantum Mechanics: Wave Equations. 2nd
Revised Edition. Berlin: Springer.
Greiner, Walter, and Joachim Reinhardt. 2009. Quantum Electrodynamics. 4th
Edition. Berlin: Springer.
Grice, H.P. 1989. Studies in the Way of Words. Cambridge, MA: Harvard University
Press.
Gross, D.J., and F. Wilczek. 1973. “Ultraviolet Behavior of Non-Abelian Gauge
Theories.” Physical Review Letters 30 (26): 1343–1346.
Guicciardini, Niccolò. 2013. “Mathematics and the New Sciences.” In The Oxford
Gupta, Rajan. 1998. “Introduction to Lattice QCD.” arXiv:hep-lat/9807028v1 Sat,
11 Jul 1998.
Bibliography 327
Gustason, W., and D.E. Ulrich. 1989. Elementary Symbolic Logic. 2nd Edition.
Prospect Heights, IL: Waveland Press Inc.
Hájek, Alan. 2012. “Interpretations of Probability.” In The Stanford Encyclopedia
of Philosophy, edited by Edward N. Zalta. Winter. https://plato.stanford.edu/
archives/win2012/entries/probability-interpret/.
Hall, Ned. 2000. “Causation and the Price of Transitivity.” The Journal of
Philosophy 97 (4): 198–222.
Hall, Ned. 2004. “Causation and the Price of Transitivity.” In Causation and
Counterfactuals, edited by John Collins, Ned Hall, and L.A. Paul, 181–203.
Cambridge, MA: MIT Press. Cited as “Price of Transitivity”
Hall, Ned. 2004. “The Intrinsic Character of Causation.” In Oxford Studies in
Metaphysics Volume 1, edited by Dean W. Zimmerman, 255–300. New York:
Oxford University Press. Cited as “Intrinsic”
Hall, Ned. 2011. “Causation and the Sciences.” In The Continuum Companion
to the Philosophy of Science, edited by Steven French and Juha Saatsi, 96–119.
London: Continuum International Publishing Group.
Hall, Ned. 2015. “Humean Reductionism about Laws of Nature”. In A Companion
to David Lewis, edited by Barry Loewer and Jonathan Schaffer, 262–277. West
Sussex: John Wiley & Sons Ltd.
Hall, Ned, and L.A. Paul. 2013. “Metaphysically Reductive Causation.” Erkenntnis
78 (Supplement 1): 9–41.
Halligan, Peter. 2007. “Belief and Illness.” The Psychologist 20 (6): 358–361.
Halpern, Joseph Y. 2016. Actual Causality. Cambridge, MA: MIT Press.
Halpern, Joseph Y., and Judea Pearl. 2005. “Causes and Explanations: A Structural-
Model Approach—Part I: Causes.” The British Journal for the Philosophy of
Science 56: 843–887. Cited as “Part I”
Halpern, Joseph Y., and Judea Pearl. 2005. “Causes and Explanations: A Structural-
Model Approach—Part II: Explanations.” The British Journal for the Philosophy
of Science 56 (4): 889–911. Cited as “Part II”
Halvorson, Hans. 2014. “Scientific Theories.” In The Oxford Handbook of
Philosophy of Science, edited by Paul Humphreys, 585–608. New York: Oxford
University Press.
Handfield, Toby. 2010. “Dispositions, Manifestations, and Causal Structure.” In
The Metaphysics of Powers: Their Grounding and Their Manifestations, edited
by Anna Marmodoro, 106–132. New York: Routledge.
Hardy, Lucien. 2007. “Towards Quantum Gravity: A Framework for Probabilistic
Theories with Non-fixed Causal Structure.” Journal of Physics A: Mathematical
and Theoretical 40: 3081–3099.
Harman, Gilbert. 1965. “The Inference to the Best Explanation.” The Philosophical
Review 74 (1): 88–95.
Harman, Gilbert. 1986. Change in View: Principles of Reasoning. Cambridge: MIT
Press.
Harman, Gilbert. 1987. “(Nonsolipsistic) Conceptual Role Semantics.” In New
Directions in Semantics, edited by Ernest LePore, 55–81. Orlando, FL: Academic
Press.
Harris, Elizabeth H. 2009. The Chlamydomonas Sourcebook: Introduction to
Chlamydomonas and Its Laboratory Use. 2nd Edition. Vol. 1. San Diego, CA:
Academic Press.
328 Bibliography
Hartle, James. B. 2003. Gravity: An Introduction to Einstein’s General Relativity.
San Francisco, CA: Addison Wesley.
Hasker, William. 1999. The Emergent Self. Ithaca, NY: Cornell University Press.
Hausman, Daniel M. 1998. Causal Asymmetries. Cambridge: Cambridge University
Press.
Havas, Peter. 1989. “The Early History of the ‘Problem of Motion’ in General
Relativity.” In Einstein and the History of General Relativity, edited by D.
Howard and J. Stachel, 234–276. Boston, MA: Birkhäuser.
Havas, Peter. 1993. “The General-Relativistic Two-Body Problem and the Einstein-
Silberstein Controversy.” In The Attraction of Gravitation: New Studies in the
History of General Relativity vol. 5, edited by J. Earman, M. Janssen, and J.D.
Norton, 88–125. Boston, MA: Birkhäuser.
Hawking, Stephen W. 1992. “Chronology Protection Conjecture.” Physical Review
D 46 (2): 603–611.
Hawking, Stephen W. 1994. “The No Boundary Condition and the Arrow of Time.”
In Physical Origins of Time Asymmetry, edited by Jonathan J. Halliwell, Juan
Pérez-Mercader, and Wojciech Hubert Zurek, 346–357. Cambridge: Cambridge
University Press.
Hawking, Stephen W., and G.F.R. Ellis. 1973. The Large Scale Structure of Space-
Time. New York: Cambridge University Press.
Hayes, A.C. 2017. “Applications of Nuclear Physics.” Reports on Progress in
Physics 80: 026301.
Healey, Richard. 1983. “Temporal and Causal Asymmetry.” In Space, Time and
Causality, edited by Richard Swinburne, 79–103. Dordrecht: D. Reidel.
Healey, Richard. 2007. Gauging What’s Real: The Conceptual Foundations of
Contemporary Gauge Theories. New York: Oxford University Press.
Healey, Richard. 2014. “Metaphysics in Science.” In The Oxford Handbook of
Philosophy of Science, edited by Paul Humphreys, 356–377. New York: Oxford
University Press.
Healey, Richard. 2017. The Quantum Revolution in Philosophy. Oxford: Oxford
University Press.
Heathcote, Adrian. 1989. “A Theory of Causality: Causality=Interaction (as Defined
by a Suitable Quantum Field Theory).” Erkenntnis 31 (1): 77–108.
Heathcote, Adrian, and David M. Armstrong. 1991. “Causes and Laws.” Noûs 25
(1): 63–73.
Hebbelmann, Dennis, and Momme von Sydow. 2014. “Betting on Transitivity in an
Economic Setting.” Proceedings of the Annual Meeting of the Cognitive Science
Society 36: 2339–2344.
Heil, John. forthcoming. “Truthmaking and Fundamentality.” Synthese. doi:
10.1007/s11229-016-1292-9.
Heilbron, J.L. 2010. Galileo. New York: Oxford University Press.
Heisenberg, Werner. 1967. “Quantum Theory and Its Interpretation.” In Niels
Bohr: His Life and Work as Seen by His Friends, edited by S. Rozental, 94–108.
Amsterdam: North-Holland.
Helliwell, T.M. 2010. Special Relativity. Mill Valley, CA: University Science Books.
Hellman, Geoffrey. 1993. “Constructive Mathematics and Quantum Mechanics:
Unbounded Operators and the Spectral Theorem.” Journal of Philosophical
Logic 22 (3): 221–248.
Hellman, Geoffrey. 1998. “Mathematical Constructivism in Spacetime.” The British
Journal for the Philosophy of Science 49 (3): 425–450.
Bibliography 329
Hellman, Geoffrey, and Frank W. Thompson. 1977. “Physicalist Materialism.”
Noûs 11 (4): 309–345.
Hempel, Carl G. 1965. Aspects of Scientific Explanation: And Other Essays in the
Philosophy of Science. New York: The Free Press.
Hempel, Carl G., and Paul Oppenheim. 1948. “Studies in the Logic of Explanation.”
Philosophy of Science 15 (2): 135–175.
Herbrand, Jacques 1971. “Investigations in Proof Theory.” In Logical Writings,
edited by Warren D. Goldfarb, with a translation of Écrits logiques edited by Jean
Van Heijenoort, 44–202. Dordrecht-Holland: D. Reidel Publishing Company.
Hetherington, Stephen. 2016. Knowledge and the Gettier Problem. Cambridge:
Hiddleston, Eric. 2005. “A Causal Theory of Counterfactuals.” Noûs 39 (4):
632–657.
Hirsch, Eli. 2008. “Language, Ontology, and Structure.” Noûs 42 (3): 509–528.
Cited as “Structure”
Hirsch, Eli. 2008. “Ontological Arguments: Interpretive Charity and Quantifier
Variance.” In Contemporary Debates in Metaphysics, edited by Theodore Sider,
John Hawthorne, and Dean W. Zimmerman, 367–381. Malden, MA: Blackwell.
Cited as “Ontological Arguments”
Hirsch, Eli. 2009. “Ontology and Alternative Languages.” In Metametaphysics:
New Essays on the Foundations of Ontology, edited by David J. Chalmers, David
Manley, and Ryan Wasserman, 231–259. New York: Oxford University Press.
Hitchcock, Christopher. 1995. “The Mishap at Reichenbach Fall: Singular vs.
General Causation.” Philosophical Studies 78: 257–291.
Hitchcock, Christopher. 2001. “The Intransitivity of Causation Revealed in
Equations and Graphs.” The Journal of Philosophy 98 (6): 273–299.
Hitchcock, Christopher. 2007. “What Russell Got Right.” In Causation, Physics,
and the Constitution of Reality: Russell’s Republic Revisited, edited by Huw Price
and Richard Corry, 45–65. New York: Oxford University Press.
Hitchcock, Christopher. 2015. “Lewis on Causation.” In A Companion to David
Lewis, edited by Barry Loewer and Jonathan Schaffer, 295–311. Oxford: John
Wiley & Sons Ltd.
Hodges, Wilfred. 2001. “Classical Logic I-First -Order Logic.” In The Blackwell Guide
to Philosophical Logic, edited by Lou Goble, 9–32. Malden, MA: Blackwell.
Hoefer, Carl. 2008. “Introducing Nancy Cartwright’s Philosophy of Science”, In
Nancy Cartwright’s Philosophy of Science, edited by Stephan Hartmann, Carl
Hoefer, and Luc Bovens, 1–15. New York: Routledge.
Hoefer, Carl. 2009. “Causation in Spacetime Theories.” In The Oxford Handbook
of Causation, edited by Helen Beebee, Christopher Hitchcock, and Peter Menzies,
Hoffman, Joshua, and Gary S. Rosenkrantz. 1994. Substance among Other
Categories. Cambridge: Cambridge University Press.
Hoffmann, Joshua, and Gary S. Rosenkrantz. 2003. “Platonic Theories of
Universals.” In The Oxford Handbook of Metaphysics, edited by Michael J. Loux
and Dean W. Zimmerman, 46–74. New York: Oxford University Press.
Hofweber, Thomas. 2016. Ontology and the Ambitions of Metaphysics. New York:
Holst, A., and T. Frölich. 1907. “Experimental Studies Relating to ‘Ship-Beri-
Beri’ and Scurvy,” Epidemiology and Infection 7 (5): 619–633. doi: 10.1017/
S0022172400033611. Cited as “Introduction”
330 Bibliography
Holst, A., and T. Frölich. 1907. “Experimental Studies Relating to Ship-Beri-
Beri and Scurvy.” Epidemiology and Infection 7 (5): 634–671. doi:10.1017/
S0022172400033623. Cited as “Part II”
Holton, Gerald, and Stephen G. Brush. 2001. Physics, the Human Adventure: From
Copernicus to Einstein and Beyond. 3rd Edition. New Brunswick, NJ: Rutgers
University Press.
Hooke, Robert. 1969. The Posthumous Works of Robert Hooke, The Sources
of Science, n. 73. With new Introduction by Richard S. Westfall. New York:
Johnson Reprint Corporation.
Hoover, Kevin D. 2013. “Identity, Structure, and Causal Representation in Scientific
Models.” In Mechanism and Causality in Biology and Economics, edited by
Hsiang-Ke Chao, Szu-Ting Chen, and Roberta L. Millstein, 35–57. Dordrecht:
Springer.
Horgan, Terence. 1984. “Functionalism and Token Physicalism.” Synthese 59 (3):
321–338.
Horgan, Terence. 1987. “Supervenient Qualia.” The Philosophical Review 96 (4):
491–520.
Horgan, Terence. 1989. “Mental Quausation.” Philosophical Perspectives, Volume
3, Philosophy of Mind and Action Theory 3: 47–76.
Horgan, Terence. 1993. “From Supervenience to Superdupervenience: Meeting the
Demands of a Material World.” Mind 102 (408): 555–586.
Horwich, Paul. 1982. Probability and Evidence. Cambridge: Cambridge University
Press.
Horwich, Paul. 1987. Asymmetries in Time: Problems in the Philosophy of Science.
Howard-Snyder, D. 1996. “Introduction: The Evidential Argument From Evil.” In
The Evidential Argument From Evil, edited by Daniel Howard-Snyder, xi–xx.
Bloomington and Indianapolis: Indiana University Press.
Howell, R. 2002. “Types Indicated, and Initiated.” The British Journal of Aesthetics
42 (2): 105–127.
Howson, Colin. 2014. “Bayesianism.” In The Routledge Companion to Philosophy
of Science, 2nd Edition, edited by Martin Curd and Stathis Psillos, 123–134. New
York: Routledge.
Huemer, M., and B. Kovitz. 2003. “Causation as Simultaneous and Continuous.”
The Philosophical Quarterly 53 (213): 556–565.
Hughes, G.E., and M.J. Cresswell. 1996. A New Introduction to Model Logic. New
York: Routledge.
Hughes, R.E. 1990. “The Rise and Fall of the ‘Antiscorbutics’: Some Notes on the
Traditional Cures for ‘Land Scurvy.’ ” Medical History 34 (1): 52–64.
Hume, David. 1975. Enquiries Concerning Human Understanding and Concerning
the Principles of Morals. 3rd Edition. Edited by L.A. Selby-Bigge and P.H.
Nidditch (revised by). Oxford: Clarendon Press.
Hume, David. 1978. A Treatise of Human Nature. Edited by L.A. Selby-Bigge.
Hume, David. 1987. “An Enquiry Concerning Human Understanding.” Vol. 35.
In Great Books of the Western World, edited by Robert Maynard Hutchins,
451–509. Chicago: Encyclopedia Britannica, Inc.
Hume, David. 2007. A Treatise of Human Nature: A Critical Edition, edited by David
Fate Norton and Mary J. Norton. Volume 1: Texts. Oxford: Clarendon Press.
Bibliography 331
Humphreys, P. 2009. “Causation and Reduction.” In The Oxford Handbook of
Causation, edited by Helen Beebee, Christopher Hitchcock, and Peter Menzies,
Huygens, Christiaan. 1952. Treatise on Light. Vol. 34. Edited by Robert Maynard
Hutchins (Editor in Chief), Mortimer Adler (Associate Editor), translated by
Silvanus P. Thompson, 546–619. Chicago: University of Chicago Encyclopædia
Britannica, Inc.
Infeld, L., and A. Schild. 1949. “On the Motion of Test Particles in General
Relativity.” Reviews of Modern Physics 21 (3): 408–413.
Ioffe, B.L., V.S. Fadin, and N. Lipatov. 2014. Quantum Chromodynamics:
Perturbative and Nonperturbative Aspects. New York: Cambridge University
Press.
Irzik, G. 1996. “Can Causes Be Reduced to Correlations.” The British Journal for
the Philosophy of Science 47: 249–270.
Ismael, J.T. 2016. How Physics Makes Us Free. New York: Oxford University Press.
Jackiw, R. 1980. “Introduction to the Yang-Mills Quantum Theory.” Review of
Modern Physics 52 (4): 661–673.
Jackson, Frank. 1998. From Metaphysics to Ethics: A Defence of Conceptual
Analysis. Oxford: Oxford University Press.
Jackson, Frank, and Philip Pettit. 1992. “In Defense of Explanatory Ecumenism.”
Economics and Philosophy 8 (1): 1–21.
Jackson, John David. 1999. Classical Electrodynamics. Hoboken, NJ: John Wiley &
Sons, Inc.
James, L.R., S.A. Mulaik, and J.M. Brett. 1982. Causal Analysis: Assumptions,
Models, and Data. Beverly Hills, CA: Sage.
Jammer, Max. 1957. Concepts of Force: A Study in the Foundations of Dynamics.
Cambridge, MA: Harvard University Press.
Janssen, Michel. 2012. “The Twins and the Bucket; How Einstein Made Gravity
Rather Than Motion Relative in General Relativity.” Studies in History and
Philosophy of Modern Physics 43 (3): 159–175.
Jeffrey, Richard. 1992. Probability and the Art of Judgment. Cambridge: Cambridge
University Press.
Jenkins, C.S. 2010. “Concepts, Experience and Modal Knowledge.” In Philosophical
Perspectives 24, edited by John Hawthorne and Jason Turner, 255–279. Malden,
MA: Wiley Periodicals, Inc.
Jörg, T. 2011. New Thinking in Complexity for the Social Sciences and Humanities:
A Generative, Transdisciplinary Approach. Dordrecht: Springer.
Joyce, James M. 1998. “A Non-Pragmatic Vindication of Probabilism.” Philosophy
of Science 65: 575–603.
Joyce, James M. 2004. “Bayesianism.” In The Oxford Handbook of Rationality, edited
by Alfred R. Mele and Piers Rawlings, 132–155. New York: Oxford University
Press.
Joyce, James M. 2010. “A Defence of Imprecise Credences in Inference and Decision
Making.” Philosophical Perspectives 24: 281–323.
Kahneman, Daniel, Paul Slovic, and Amos Tversky. 1982. Judgment Under
Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press.
Kahneman, Daniel, and Amos Tversky. 1982. “On the Psychology of Prediction.” In
Judgment Under Uncertainty: Heuristics and Biases, edited by Daniel Kahneman, Paul
Slovic, and Amos Tversky, 48–68. Cambridge: Cambridge University Press.
332 Bibliography
Kaiser, David. 1999. “Do Feynman Diagrams Endorse a Particle Ontology? The
Roles of Feynman Diagrams in S-Matrix Theory,” In Conceptual Foundations of
Quantum Field Theory, edited by Tian Yu Cao, 343–356. New York: Cambridge
University Press.
Kaiser, David. 2005. Drawing Theories Apart: The Dispersion of Feynman
Diagrams in Postwar Physics. Chicago: University of Chicago Press. Cited as
“Drawing Theories Apart”
Kaiser, David. 2005. “Physics and Feynman’s Diagrams: In the Hands of a Postwar
Generation, a Tool Intended to Lead Quantum Electrodynamics Out of a Decades-
Long Morass Helped Transform Physics.” American Scientist 93 (March-April)
(2): 156–165. Cited as “Physics and Feynman’s Diagrams”
Kaiser, David. 2009. “Feynman Diagrams,” In Compendium of Quantum Physics:
Concepts, Experiments, History, and Philosophy, edited by Daniel Greenberger,
Klaus Hentschel, and Friedel Weinert, 235–239. New York: Springer.
Kaiser, David, Kinji Ito, and Karl Hall. 2004. “Spreading the Tools of Theory:
Feynman Diagrams in the United States, Japan, and the Soviet Union.” Social
Studies of Science 34: 879–922.
Kane, Robert. 1996. The Significance of Free Will. New York: Oxford University
Press.
Kania, Andrew. 2008. “Piece for the End of Time: In Defence of Musical Ontology.”
British Journal of Aesthetics 48 (1): 65–79. Cited as “Piece”
Kania, Andrew. 2008. “Review of Works of Music: An Essay in Ontology by Julian
Dodd.” The Journal of Aesthetics and Art Criticism 66 (2): 201–203. Cited as
“Review”
Kant, Immanuel. 1992. Theoretical Philosophy, 1755–1770. Edited by David
Walford and Ralf Meerbote (Collaborator). Translated by David Walford and
Ralf Meerbote (Collaborator). Cambridge: Cambridge University Press.
Kant, Immanuel. 1998. Critique of Pure Reason: The Cambridge Edition of the
Works of Immanuel Kant. Translated and edited by Paul Guyer and Allen W.
Wood. New York: Cambridge University Press.
Kaplan, David. 1979. “On the Logic of Demonstratives.” Journal of Philosophical
Logic 8 (1): 81–98.
Kaplan, Mark. 1996. Decision Theory as Philosophy. Cambridge: Cambridge
University Press.
Kent, Adrian. 1989. “ ‘Quantum Jumps’ and Indistinguishability.” Modern Physics
Letters 4 (19): 1839–1845.
Kentridge, R.W., C.A. Heywood, and L. Weiskrantz. 1999. “Attention with-
out Awareness in Blindsight.” Proceedings: Biological Sciences 266 (1430):
1805–1811.
Kepler, J. 1937. Gesammelte Werke. 22 vols. Munich: C.H. Beck.
Khachatryan, V. et al. 2016. “Measurement of the Top Quark Pair Production
Cross Section in Proton-Proton Collisions at s = 13TeV ", Physical Review
Letters 116 (5): 052002. (pagination is 1–18) The complete author list for this
work is over six pages long. You can find the complete list of authors on pages
7–14 of the work. I include a link to the doi webpage for good measure. Accessed
August 25, 2017. https://doi.org/10.1103/PhysRevLett.116.052002.
Khan, M.D.N., and S. Panigrahi. 2016. Principles of Engineering Physics 1. New
York: Cambridge University Press.
Bibliography 333
Kim, Jaegwon. 1966. “On the Psycho-Physical Identity Theory.” American
Philosophical Quarterly 3: 227–235.
Kim, Jaegwon. 1973. “Causation, Nomic Subsumption, and the Concept of Event.”
The Journal of Philosophy 70: 217–236. Cited as “Nomic Subsumption”
Kim, Jaegwon. 1973. “Events as Property Exemplifications.” In Supervenience and
Mind: Selected Philosophical Essays, edited by Jaegwon Kim, 33–52. New York:
Cambridge University Press. Cited as “Events as Property Exemplifications”
Kim, Jaegwon. 1976. “Events as Property Exemplifications.” In Action Theory,
edited by Myles Brand and D. Walton, 159–177. Dordrecht: D. Reidel.
Kim, Jaegwon. 1993. “Mechanism, Purpose, and Explanatory Exclusion.” In
Supervenience and Mind: Selected Philosophical Essays, by Jaegwon Kim, 237–
264. New York: Cambridge University Press.
Kim, Jaegwon. 1998. Mind in a Physical World: An Essay on the Mind-Body
Problem and Mental Causation. Cambridge, MA: MIT Press.
Kim, Jaegwon. 2007. “Causation and Mental Causation.” In Contemporary
Debates in Philosophy of Mind, edited by Brian P. McLaughlin and Jonathan
Cohen, 227–242. Malden, MA: Blackwell.
Kim, Jaegwon. 2008. “Reduction and Reductive Explanation: Is One Possible
Without the Other?” In Being Reduced: New Essays on Reduction, Explanation,
and Causation, edited by Jakob Hohwy and Jesper Kallestrup, 93–114. New
King, Jeffrey C. 1998. “What Is a Philosophical Analysis?” Philosophical Studies
90 (2): 155–179.
King, Jeffrey C. 2007. The Nature and Structure of Content. New York: Oxford
University Press.
King, Jeffrey C. 2009. “Questions of Unity.” Proceedings of the Aristotelian Society
109: 257–277.
King, Jeffrey C. 2014. “Naturalized Propositions.” In New Thinking About
Propositions, by Jeffrey C. King, Scott Soames, and Jeff Speaks, 47–70. New
York: Oxford University Press. Cited as “Naturalized”
King, Jeffrey C. 2014. “Criticisms of Soames and Speaks.” In New Thinking about
York: Oxford University Press. Cited as “Criticisms”
King, Jeffrey C. 2016. “Philosophical and Conceptual Analysis.” In The Oxford
Handbook of Philosophical Methodology, edited by Herman Cappelen, Tamar
Szabó Gendler, and John Hawthorne, 249–261. Oxford: Oxford University
Press.
King, Jeffrey C. 2017. “Structured Propositions.” In The Stanford Encyclopedia
of Philosophy, edited by Edward N. Zalta. https://plato.stanford.edu/archives/
fall2017/entries/propositions-structured/
Kistler, Max. 2006. Causation and Laws of Nature. New York: Routledge.
Kittel, Charles. 1966. Introduction to Solid State Physics. 3rd Edition. New York:
John Wiley & Sons, Inc.
Kivy, P. 1993. The Fine Art of Repetition: Essays in the Philosophy of Music. New
Klein, Sanford, and Gregory Nellis. 2012. Thermodynamics. New York: Cambridge
University Press.
Kline, Morris. 1985. Mathematics for the Nonmathematician. New York: Dover.
334 Bibliography
Koizumi, Ayako, Asami Tsuchiya, Ken-ichiro Nakajima, Keisuke Ito, Tohru Terada,
Akiko Shimizu-Ibuka, Loïc Briand, Tomiko Asakura, Takumi Misaka, and Keiko
Abe. 2011. “Human Sweet Taste Receptor Mediates Acid-Induced Sweetness of
Miraculin.” Proceedings of the National Academy of Sciences of the United States
of America 108 (40): 16819–16824.
Konyndyk, K. 1986. Introductory Modal Logic. South Bend, IN: University of
Notre Dame Press.
Koons, Robert C. 2000. Realism Regained: An Exact Theory of Causation,
Teleology, and the Mind. New York: Oxford University Press.
Koons, Robert C., and Timothy H. Pickavance. 2017. The Atlas of Reality:
A Comprehensive Guide to Metaphysics. Malden, MA: John Wiley & Sons, Ltd.
Korman, Daniel Z. 2014. “Debunking Perceptual Beliefs About Ordinary Objects.”
Philosophers’ Imprint 14 (13): 1–21.
Kotzen, Matthew. 2016. “Probability in Epistemology.” In The Oxford Handbook
of Probability and Philosophy, edited by Alan Hájek and Christopher Hitchcock,
Kriele, Marcus. 1999. Spacetime: Foundations of General Relativity and Differential
Geometry. Berlin: Springer.
Kripke, Saul. 1959. “A Completeness Theorem in Modal Logic.” The Journal of
Symbolic Logic 24: 1–14.
Kripke, Saul. 1976. “Is There a Problem About Substitutional Quantification?”
In Truth and Meaning: Essays in Semantics, edited by Gareth Evans and John
McDowell, 325–419. New York: Oxford University Press.
Kripke, Saul. 1980. Naming and Necessity. Cambridge, MA: Harvard University
Press.
Kugler, P.N., and M.T. Turvey. 1987. Information, Natural Law, and the Self-
Assembly of Rhythmic Movement. Mahwah, NJ: L. Erlbaum Associates.
Kuhn, Thomas S. 1978. Black-Body Theory and the Quantum Discontinuity 1894-
1912. Oxford: Oxford University Press.
Kutach, Douglas. 2002. “The Entropy Theory of Counterfactuals.” Philosophy of
Science 69 (1): 82–104.
Kutach, Douglas. 2007. “The Physical Foundations of Causation.” In Causation,
Physics, and the Constitution of Reality: Russell’s Republic Revisited, edited by
Huw Price and Richard Corry, 327–350. New York: Oxford University Press.
Kutach, Douglas. 2013. Causation and Its Basis in Fundamental Physics. New
Kvanvig, J.L. 2006. The Knowability Paradox. New York: Oxford University Press.
Kvart, I. 1991. “Transitivity and Preemption of Causal Relevance.” Philosophical
Studies 64 (2): 125–160.
Kwong, Jack M.C. 2007. “Is Conceptual Atomism a Plausible Theory of Concepts?”
The Southern Journal of Philosophy 45 (3): 413–434.
Lackey, Jennifer, and Ernest Sosa (editors). 2006. The Epistemology of Testimony.
Oxford: Oxford University Press.
Ladyman, James. 1998. “What Is Structural Realism?” Studies in History and
Ladyman, James. 2016. “Structural Realism.” In The Stanford Encyclopedia of
Philosophy (Winter 2016 Edition), edited by Edward N. Zalta. https://plato.stan
ford.edu/archives/win2016/entries/structural-realism/.
Bibliography 335
Lagrange, Joseph-Louis. 1997. Analytical Mechanics. Translated by Auguste
Boissonnade and Victor N. Vagliente. Dordrecht: Kluwer.
Lambourne, R.J.A. 2010. Relativity, Gravitation and Cosmology. New York:
Lancaster, Tom, and Stephen J. Blundell. 2014. Quantum Field Theory for the
Gifted Amateur. Oxford: Oxford University Press.
Langdon, Robyn, and Max Coltheart. 2000. “The Cognitive Neuropsychology of
Delusions.” Mind and Language. 15 (1): 184–218.
Lange, Marc. 2002. An Introduction to the Philosophy of Physics: Locality, Fields,
Energy, and Mass. New York: Routledge.
Lange, Marc. 2014. “Laws of Nature.” In The Routledge Companion to Philosophy
of Science, edited by Martin Curd and Stathis Psillos, 235–244. New York:
Routledge Publishers.
Lange, Marc. 2017. Because Without Cause: Non-Causal Explanations in Science
and Mathematics. New York: Oxford University of Press.
Laplace, Pierre-Simon. 1917. A Philosophical Essay on Probabilities. 2nd Edition
and Revised. Edited by Frederick Wilson and Frederick L. Emory. New York:
John Wiley & Sons, Inc.
Lazarus, Richard. S., and Susan Folkman. 1984. Stress, Appraisal, and Coping.
New York: Springer.
Leader, E., and E. Predazzi. 1996. An Introduction to Gauge Theories and Modern
Particle Physics, Volume 2: CP-Violation, QCD and Hard Processes. New York:
Lebowitz, Joel L. 1999. “Statistical Mechanics: A Selective Review of Two Central
Issues.” Reviews of Modern Physics 71 (2): S346–S357.
Leftow, Brian. 1989. “A Leibnizian Cosmological Argument.” Philosophical Studies
57 (2): 135–155.
Leftow, Brian. 2012. God and Necessity. New York: Oxford University Press.
Lehmkuhl, Dennis. 2008. “Is Spacetime a Gravitational Field?” In The Ontology of
Spacetime II, edited by Dennis Dieks, 83–110. Amsterdam: Elsevier B.V.
Lehmkuhl, Dennis. 2011. “Mass-Energy-Momentum: Only There Because of
Spacetime?” The British Journal for the Philosophy of Science 62 (3): 453–488.
Lehmkuhl, Dennis. 2014. “Why Einstein Did Not Believe That General Relativity
Geometrizes Gravity.” Studies in History and Philosophy of Modern Physics 46
(2): 316–326.
Lemmon, J. 1967. “ ‘Comments’ on D. Davidson’s ‘The Logical Form of Action
Sentences’.” In The Logic of Decision and Action, edited by Nicholas Rescher,
96–103. Pittsburgh, PA: University of Pittsburgh Press.
Lepore, Ernie, and Kirk Ludwig. 2011. “Truth and Meaning Redux.” Philosophical
Studies 154 (2): 251–277.
Levinson, J. 1980. “What a Musical Work Is.” The Journal of Philosophy 77 (1):
5–28.
Lévy-Leblond, Jean-Marc. 1976. “One More Derivation of the Lorentz
Transformations.” American Journal of Physics 44: 271–277.
Lewis, David K. 1970. “How to Define Theoretical Terms.” The Journal of
Philosophy 67 (13): 427–446.
Lewis, David K. 1972. “Psychophysical and Theoretical Identifications.” The
Australasian Journal of Philosophy 50 (3): 249–258.
336 Bibliography
Lewis, David K. 1973. “Causation.” The Journal of Philosophy 70 (17): 556–567.
Cited as “Causation”
Lewis, David K. 1973. Counterfactuals. Malden, MA: Blackwell. Cited as “Counter-
factuals”
Lewis, David K. 1979. “Counterfactual Dependence and Time’s Arrow.” Noûs 13
(4): 455–476.
Lewis, David K. 1983. “Survival and Identity.” Vol. 1. In Philosophical Papers, by David
Lewis, 55–77. New York: Oxford University Press. Cited as “Survival and Identity”
Lewis, David K. 1983. “New Work for a Theory of Universals”, Australasian
Journal of Philosophy 61 (4): 343–377. Cited as “New Work”
Lewis, David K. 1986. “Causal Explanation.” In Philosophical Papers: Volume II, by
David Lewis, 214–240. New York: Oxford University Press. Cited as “Explanation”
Lewis, David K. 1986. “Causation.” In Philosophical Papers Volume II, by David
Lewis, 159–172. New York: Oxford University Press. Cited as “Causation”
Lewis, David K. 1986. “Events,” In Philosophical Papers: Volume II, by David
Lewis, 241–269. New York: Oxford University Press. Cited as “Events”
Lewis, David K. 1986. On the Plurality of Worlds. Malden, MA: Blackwell. Cited
as “Plurality”
Lewis, David K. 1986. “Postscripts to ‘Causation’.” In Philosophical Papers Volume
II, by David K. Lewis, 172–213. New York: Oxford University Press. Cited as
“Postscripts”
Lewis, David K. 1991. Parts of Classes. Oxford: Wiley-Blackwell.
Lewis, David K. 1994. “Humean Supervenience Debugged.” Mind 103 (412):
473–490.
Lewis, David K. 1999. Papers in Metaphysics and Epistemology. Cambridge:
Lewis, David K. 2000. “Causation as Influence.” The Journal of Philosophy 97 (4)
Special Issue: Causation: 182–197.
Lewis, David K. 2004. “Causation as Influence.” In Causation and Counterfactuals,
edited by John Collins, Ned Hall, and L.A. Paul, 75–106. Cambridge, MA: MIT
Press. Cited as “Causation as Influence”
Lewis, David K. 2004. “Void and Object.” In Causation and Counterfactuals,
edited by John Collins, Ned Hall, and L.A. Paul, 277–290. Cambridge, MA: MIT
Press. Cited as “Void and Object”
Lewis, P.J. 2016. Quantum Ontology: A Guide to the Metaphysics of Quantum
Mechanics. New York: Oxford University Press.
Leyton, Michael. 1992. Symmetry, Causality, Mind. Cambridge, MA: MIT Press.
LIGO. 2017. “Laser Interferometer Gravitational-Wave Observatory Supported
by the National Science Foundation. Operated by Caltech and MIT.” LIGO.
Accessed July 31, 2017. www.ligo.caltech.edu/page/what-are-gw.
Lipton, Peter. 2004. Inference to the Best Explanation. 2nd Edition. New York:
Routledge.
Locke, John. 1975. An Essay Concerning Human Understanding. Edited with a
foreword by Peter H. Nidditch. Oxford: Clarendon Press.
Loewer, Barry. 2007. “Counterfactuals and the Second Law.” In Causation, Physics,
and the Constitution of Reality: Russell’s Republic Revisited, edited by Huw Price
and Richard Corry, 293–326. New York: Oxford University Press.
Loewer, Barry. 2008. “Determinism.” In The Routledge Companion to Philosophy
of Science, 1st Edition, edited by Stathis Psillos and Martin Curd, 327–336. New
York: Routledge. Cited as “Determinism”
Bibliography 337
Loewer, Barry. 2008. “Why There Is Anything Except Physics,” In Being Reduced:
New Essays on Reduction, Explanation, and Causation, edited by Jakob Hohwy
and Jesper Kallestrup. New York: Oxford University Press. Cited as “Physics”
Loewer, Barry. 2012. “Two Accounts of Laws and Time.” Philosophical Studies
160 (1): 115–137.
Lombard, Lawrence B. 1986. Events: A Metaphysical Study. London: Routledge &
Kegan Paul.
Lormand, Eric. 2006. “Phenomenal Impressions.” In Perceptual Experience, edited
by Tamar Szabó Gendler and John Hawthorne, 316–353. New York: Oxford
University Press.
Loux, Michael J. 1978. Substance and Attribute: A Study in Ontology. Dordrecht:
D. Reidel.
Loux, Michael J. 2001. “Realism and Anti-Realism.” In Metaphysics: Contemporary
Readings, edited by Michael J. Loux, 449–458. New York: Routledge.
Loux, Michael J. 2006. Metaphysics: A Contemporary Introduction. 3rd Edition.
New York: Routledge.
Lowe, E.J. 2006. The Four-Category Ontology: A Metaphysical Foundation for
Natural Science. New York: Oxford University Press.
Lowe, E.J. 2016. “There Are (Probably) No Relations.” In The Metaphysics of
Relations, edited by Anna Marmodoro and David Yates, 100–112. New York:
Lycan, William. G. 1986. “Moral Facts and Moral Knowledge.” Edited by N. Gillespie.
Southern Journal of Philosophy (Spindel Conference Supplement) 24 (1): 79–94.
Lycan, William. G. 2000. Philosophy of Language. New York: Routledge.
Lycan, William G. 2002. “Explanation and Epistemology.” In The Oxford
Handbook of Epistemology, edited by Paul K. Moser, 408–433. New York:
Lyth, D.H., and A.R. Liddle. 2009. The Primordial Density Perturbation: Cosmology,
Inflation and the Origin of Structure. New York: Cambridge University Press.
MacBride, Fraser. 2005. “The Particular-Universal Distinction: A Dogma of
Metaphysics?” Mind 114 (455): 565–614.
Mach, Ernst. 1920. Erkenntnis und irrtum: Skizzen zur Psychologie der Forschung.
Leipzig: Verlag von Johann Ambrosius Barth.
Mach, Ernst. 1923. Die Principien der Wärmelehre: Historisch-Kritisch Entwickelt.
Leipzig: Verlag von Johann Ambrosius Barth.
MacIntyre, A.C. 1957. “Determinism.” Mind 66 (261): 28–41.
Mackie, J.L. 1974. The Cement of the Universe: A Study of Causation. Oxford:
Mahan, Gerald D. 2009. Quantum Mechanics in a Nutshell. Princeton, NJ:
Maher, Brendan A. 1974. “Delusional Thinking and Perceptual Disorder.” Journal
of Individual Psychology 30 (1): 98–113.
Majors, B. 2005. “Moral Discourse and Descriptive Properties.” The Philosophical
Quarterly 55 (220): 475–494.
Malament, D.B. 2009. “On the Status of the Geodesic Law.” http://philsci-archive.
pitt.edu/4908/1/GeodesicLaw.pdf. Downloaded December 29, 2015.
Malament, D.B. 2012. “A Remark About the ‘Geodesic Principle’ in General
Relativity.” In Analysis and Interpretation in Exact Sciences: Essays in Honour of
William Demopoulos, edited by Mélanie Frappier, Derek H. Brown, and Robert
DiSalle, 245–252. Dordrecht: Springer. Cited as “Remark about Geodesic Principle”
338 Bibliography
Malament, D.B. 2012. Topics in the Foundations of General Relativity and
Newtonian Gravitation Theory. Chicago: University of Chicago Press. Cited as
“Topics in the Foundations of GR”
Malcom, Norman. 1968. “The Conceivability of Mechanism.” The Philosophical
Review 77 (1): 45–72.
Manchak, J.B. 2013. “Global Spacetime Structure.” In The Oxford Handbook of
Philosophy of Physics, edited by Robert Batterman, 587–606. New York: Oxford
University Press.
Margolis, Eric, and Stephen Laurence. 2007. “The Ontology of Concepts—Abstract
Objects or Mental Representations?” Noûs 41 (4): 561–593.
Marshak, Robert E. 1952. Meson Physics. New York: McGraw-Hill.
Marshall, D. 2009. “Can ‘Intrinsic’ Be Defined Using Only Broadly Logical
Notions?” Philosophy and Phenomenological Research 78 (3): 646–672.
Martin, R.M. 1969. “On Events and Event-Descriptions.” In Fact and Existence:
Proceedings of the University of Western Ontario Philosophy Colloquium 1966,
edited by Joseph Margolis, 63–74. Toronto: University of Toronto Press.
Maslen, C. 2004. “Causes, Contrasts, and the Nontransitivity of Causation.” In
Causation and Counterfactuals, edited by John Collins, Ned Hall, and L.A. Paul,
341–357. Cambridge: MIT Press.
Maslen, Cei, Terry Horgan, and Helen Daly. 2009. “Mental Causation,” In The
Mattuck, Richard D. 1967. A Guide to Feynman Diagrams in the Many-Body
Problem. London: McGraw-Hill.
Maudlin, Tim. 2007. The Metaphysics Within Physics. New York: Oxford
University Press.
Maudlin, Tim. 2012. Philosophy of Physics: Space and Time. Princeton, NJ:
Maudlin, Tim. 2015. “Confessions of a Hardcore, Unsophisticated Metaphysical
Realist.” In The Philosophy of Hilary Putnam, edited by Randall E. Auxier,
Douglas R. Anderson, and Lewis Edwin Hahn, 487–501. Chicago: Open Court.
Maxwell, James Clerk. 1990. The Scientific Letters and Papers of James Clerk
Maxwell. Vol. 1. 1846–1862. New York: Cambridge University Press.
Maxwell, James Clerk. 2003. The Scientific Papers of James Clerk Maxwell. Edited
by W.D. Niven. Vol. 1. Mineola, NY: Dover.
Maxwell, James Clerk. 2010. A Treatise on Electricity and Magnetism. Vol. 1. 2
vols. Cambridge: Cambridge University Press.
Maydole, R.E. 2012. “The Ontological Argument.” In The Blackwell Companion
to Natural Theology, edited by William Lane Craig and J.P. Moreland, 553–592.
Malden, MA: Blackwell.
McCann, H.J. 2012. Creation and the Sovereignty of God. Bloomington: Indiana
University Press.
McDermott, M. 1995. “Redundant Causation.” The British Journal for the
McGrath, Matthew. 2014. “Propositions.” Vers. Spring 2014. The Stanford
Encyclopedia of Philosophy. Edited by Edward N. Zalta. Spring. https://plato.
stanford.edu/archives/spr2014/entries/propositions/.
McGrew, T., and L. McGrew. 2007. Internalism and Epistemology: The Architecture
of Reason. New York: Routledge.
Bibliography 339
McKinsey, J.C.C., A.C. Sugar, and Patrick Suppes. 1953. “Axiomatic Foundations
of Classical Particle Mechanics.” Journal of Rational Mechanics and Analysis 2:
253–272.
McLaughlin, Brian P. 1989. “Type Epiphenomenalism, Type Dualism, and the
Causal Priority of the Physical.” Philosophical Perspectives, Vol. 3, Philosophy
of Mind and Action Theory 3: 109–135.
McLaughlin, Brian P. 1997. “Supervenience, Vagueness, and Determination.”
Philosophical Perspectives: Mind, Causation, and World 11: 209–230.
McLaughlin, Brian, and Karen Bennett. 2014. “Supervenience.” In The Stanford
Encyclopedia of Philosophy, edited by Edward N. Zalta. Stanford, CA. https://
plato.stanford.edu/archives/spr2014/entries/supervenience/.
Meek, C., and C. Glymour. 1994. “Conditioning and Intervening.” The British
Journal for the Philosophy of Science 45 (4): 1001–1021.
Mehlberg, H. 1980. Time, Causality, and the Quantum Theory: Studies in the
Philosophy of Science Volume Two. Time in a Quantized Universe. Edited by
Robert S. Cohen. Dordrecht: D. Reidel.
Mele, A.R. 2009. “Causation, Action, and Free Will.” In The Oxford Handbook of
Mellor, D.H. 1997. “Properties and Predicates.” In Properties, edited by D.H.
Mellor and Alex Oliver, 255–267. New York: Oxford University Press.
Mellor, D.H. 1999. The Facts of Causation. New York: Routledge.
Mellor, D.H. 2004. “For Facts as Causes and Effects.” In Causation and
Counterfactuals, edited by John Collins, Ned Hall, and L.A. Paul, 309–323.
Melnyk, Andrew. 1994. “Inference to the Best Explanation and Other Minds.”
Australasian Journal of Philosophy 72 (4): 482–491.
Menzies, Peter. 1999. “Intrinsic Versus Extrinsic Conceptions of Causation.” In
Causation and Laws of Nature, edited by Howard Sankey, 313–329. Dordrecht:
Kluwer.
Menzies, Peter, and Huw Price. 1993. “Causation as a Secondary Quality.” The
British Journal for the Philosophy of Science 44 (2): 187–203.
Mermin, N. David. 1989. Space and Time in Special Relativity. Long Grove, IL:
Waveland Press.
Merricks, Trenton. 1995. “Warrant Entails Truth.” Philosophy and Phenomenological
Research 55 (4): 841–855.
Merricks, Trenton. 2007. Truth and Ontology. New York: Oxford University Press.
Merricks, Trenton. 2015. Propositions. Oxford: Oxford University Press.
Miller, Arthur I. 1981. Albert Einstein’s Special Theory of Relativity: Emergence
(1905) and Early Interpretation (1905–1911). Reading, MA: Addison-Wesley.
Miller, Richard. 1987. Fact and Method: Explanation, Confirmation and Reality in
the Natural and Social Sciences. Princeton, NJ: Princeton University Press.
Millikan, Ruth. 2000. On Clear and Confused Ideas: An Essay about Substance
Concepts. Cambridge: Cambridge University Press.
Minkowski, Hermann. 1952. “Space and Time.” In The Principle of Relativity:
A Collection of Original Papers on the Special and General Theory of Relativity.
Notes by A. Sommerfeld, edited by H.A. Lorentz, Albert Einstein, Hermann
Minkowski, and Hermann Weyl, translated by W. Perrett and G.B. Jeffrey, 75–96.
New York: Dover.
340 Bibliography
Misner, Charles W., Kip S. Thorne, and John A. Wheeler. 1973. Gravitation. San
Francisco, CA: W.H. Freeman and Company.
Molnar, George. 2003. Powers: A Study in Metaphysics. Edited with an introduc-
tion by Stephen Mumford and a foreword by D.M. Armstrong. Oxford: Oxford
University Press.
Mook, D.E., and T. Vargish. 1987. Inside Relativity. Princeton, NJ: Princeton
University Press.
Moore, Michael S. 2009. Causation and Responsibility: An Essay in Law, Morals,
and Metaphysics. New York: Oxford University Press.
Moore, Walter. 1989. Schrödinger: Life and Thought. New York: Cambridge
University Press.
Moreland, J.P. 2001. Universals. Montreal: McGill-Queen’s University Press.
Morris, Michael S., Kip S. Thorne, and Ulvi Yurtsever. 1988. “Wormholes, Time
Machines, and the Weak Energy Condition.” Physical Review Letters 61:
1446–1449.
Moser, Paul K. 1989. Knowledge and Evidence. New York: Cambridge University
Press.
Moser, Paul K., Dwayne H. Mulder, and J.D. Trout. 1998. The Theory of
Knowledge: A Thematic Introduction. New York: Oxford University Press.
Moss, Sarah. 2013. “Epistemology Formalized.” The Philosophical Review 122
(1): 1–43.
Mumford, Stephen. 1999. “Intentionality and the Physical: A New Theory of
Disposition Ascription.” The Philosophical Quarterly 49 (195): 215–225.
Mumford, Stephen. 2009. “Causal Powers and Capacities.” In The Oxford
Handbook of Causation, edited by Christopher Hitchcock and Peter Menzies,
Mumford, Stephen, and Rani Lill Anjum. 2011. Getting Causes From Powers. New
Mundy, B. 1987. “The Metaphysics of Quantity.” Philosophical Studies 51 (1):
29–54.
Nagel, Ernest. 1961. The Structure of Science: Problems in the Logic of Scientific
Explanation. New York: Harcourt, Brace & World, Inc.
Narlikar, Jayant V., and T. Padmanabhan. 1986. Gravity, Gauge Theories and
Quantum Cosmology. Dordrecht: D. Riedel.
Nelkin, Dana Kay. 2011. Making Sense of Freedom and Responsibility. Oxford:
Nemiroff, Robert, and Jerry Bonnell. 2017. The Square Root of Two =. August 15.
Accessed August 15, 2017. https://apod.nasa.gov/htmltest/gifcity/sqrt2.10mil.
The computation was made by Robert Nemiroff. Checking (up to 6 million dig-
its) was performed by Jerry Bonnell.
Nerlich, G. 1976. The Shape of Space. New York: Cambridge University Press.
Nerlich, G. 1994. What Spacetime Explains: Metaphysical Essays on Space and
Time. New York: Cambridge University Press.
Nerlich, G. 2003. “Space-Time Substantivalism.” In The Oxford Handbook of
Metaphysics, edited by Michael J. Loux and Dean W. Zimmerman, 281–314.
Newman, M.H.A. 1928. “Mr. Russell’s ‘Causal Theory of Perception’.” Mind 37
(146): 137–148.
Bibliography 341
Newton, Isaac. 1952. Opticks: or, A Treatise of the Reflections, Refractions,
Inflections and Colours of Light. Based on the Fourth Edition London, 1730.
Preface by I.B. Cohen; Introduction by E.T. Whittaker. With a Foreword by
Albert Einstein. New York: Dover.
Newton, Isaac. 1962. Unpublished Scientific Papers of Isaac Newton. Edited and
translated by A.R. Hall and M.B. Hall. Cambridge: Cambridge University Press.
Newton, Isaac. 1999. The Principia: Mathematical Principles of Natural Philosophy.
Translated by I. Bernard Cohen and Anne Whitman. Assisted by Julia Budenz.
Preceded by A Guide to Newton’s Principia by I. Bernard Cohen. Berkeley and
Los Angeles: University of California Press.
Newton, Isaac. 2004. Philosophical Writings. Edited by Andrew Janiak. Cambridge:
Newton, Isaac. 2007. “Original Letter From Isaac Newton to Richard Bentley
189.R.4.47, ff. 7–8.” Newton Project. Cambridge, UK Trinity College Library.
October. www.newtonproject.ox.ac.uk/view/texts/normalized/THEM00258.
Ney, Alyssa. 2013. “Introduction.” In The Wave Function: Essays on the Metaphysics
of Quantum Mechanics, edited by Alyssa Ney and David Z. Albert, 1–51. New
Nisbett, Richard, and Lee Ross. 1980. Human Inference: Strategies and Shortcomings
of Social Judgment. Englewood Cliffs, NJ: Prentice-Hall.
Nishikimi, Morimitsu, and Sidney Udenfriend. 1977. “Scurvy as an Inborn Error
of Ascorbic Acid Biosynthesis.” Trends in Biochemical Science 2 (5): 111–113.
North, Jill. 2011. “Time in Thermodynamics.” In The Oxford Handbook of
Philosophy of Time, edited by Craig Callender, 312–350. New York: Oxford
University Press.
Norton, John D. 1984. “How Einstein Found His Field Equations: 1912–1915.”
Historical Studies in the Physical Sciences 14 (2): 253–316.
Norton, John D. 1989. “What Was Einstein’s Principle of Equivalence.” In Einstein
and the History of General Relativity: Einstein Studies Volume 1, edited by D.
Howard and J. Stachel, 5–47. Boston, MA: Birkhäuser.
Norton, John D. 2004. “Why Thought Experiments Do Not Transcend Empiricism.”
In Contemporary Debates in Philosophy of Science, edited by Christopher
Hitchcock, 44–66. Malden, MA: Blackwell.
Norton, John D. 2007. “Causation as Folk Science.” In Causation, Physics, and the
Constitution of Reality: Russell’s Republic Revisited. Edited by Huw Price and
Richard Corry, 11–44. New York: Oxford University Press. Cited as “Causation
as Folk Science”
Norton, John D. 2007. “Do the Causal Principles of Modern Physics Contradict
Causal Anti-Fundamentalism?” In Thinking about Causes: From Greek Philosophy
to Modern Physics, edited by Peter Machamer and Gereon Wolters, 222–234.
Pittsburgh, PA: University of Pittsburgh Press. Cited as “Causal Principles”
Norton, John D. 2015. “What Can We Learn About the Ontology of Space and Time
From the Theory of Relativity?” In Physical Theory: Method and Interpretation,
edited by Lawrence Sklar, 185–228. New York: Oxford University Press.
Novick, L.R., and P.W. Chen. 2004. “Assessing Interactive Causal Influence.”
Psychological Review 111 (2): 455–485.
O’Connor, Timothy. 2000. Persons and Causes: The Metaphysics of Free Will. New
342 Bibliography
O’Connor, Timothy. 2008. “Agent-Causal Power,” In Dispositions and Causes,
edited by Toby Handfield, 189–214. Oxford: Clarendon Press.
O’Connor, Timothy. 2013. “Reasons and Causes.” In A Companion to the
Philosophy of Action, edited by Timothy O’Connor and Constantine Sandis,
129–138. Malden, MA: Wiley-Blackwell.
O’Connor, Timothy, and J. Churchill. 2004. “Reasons Explanation and Agent
Control: In Search of an Integrated Account.” Philosophical Topics 32 (1/2):
241–253.
Oppy, Graham. 2000. “On ‘A New Cosmological Argument’.” Religious Studies
36 (3): 345–353.
Pais, Abraham. 1982. Subtle Is the Lord: The Science and the Life of Albert Einstein.
Pais, Abraham. 1986. Inward Bound: Of Matter and Forces in the Physical World.
Pais, Abraham. 1991. Niels Bohr’s Times, In Physics, Philosophy and Polity.
Papineau, David. 2013. “Causation Is Macroscopic but Not Irreducible.” In Mental
Causation and Ontology, edited by S.C. Gibb, E.J. Lowe, and R.D. Ingthorsson,
126–152. Oxford: Oxford University Press.
Parfit, Derek. 2011. On What Matters. Vol. 2. New York: Oxford University Press.
Pargetter, Robert. 1984. “The Scientific Inference to Other Minds.” Australasian
Journal of Philosophy 62 (2): 158–163.
Paul, L.A. 2000. “Aspect Causation.” The Journal of Philosophy 97 (4): 235–256.
Paul, L.A. 2007. “Constitutive Overdetermination.” In Causation and Explanation,
edited by Joseph Keim Campbell, Michael O’Rourke, and Harry Silverstein, 265–
290. Cambridge, MA: MIT Press.
Paul, L.A. 2009. “Counterfactual Theories.” In The Oxford Handbook of
Paul, L.A. 2014. “Experience and the Arrow.” In Chance and Temporal Asymmetry,
edited by Alastair Wilson, 175–193. Oxford: Oxford University Press.
Paul, L.A., and Ned Hall. 2013. Causation: A User’s Guide. New York: Oxford
University Press.
Peacocke, Christopher. 1979. Holistic Explanation: Action, Space, Interpretation.
Peacocke, Christopher. 2004. The Realm of Reason. Oxford: Oxford University
Press.
Pearl, Judea. 2000. Causality: Models, Reasoning, and Inference. 1st Edition. New
Pearl, Judea. 2009. Causality. 2nd Edition. New York: Cambridge University Press.
Pearle, P. 1984. “Experimental Tests of Dynamical State-Vector Reduction.”
Physical Review D 29: 235–240.
Pearsall, Judy, and Patrick Hanks (editors). 2003. Oxford Dictionary of English.
2nd Edition. Oxford: Oxford University Press.
Peebles, P.J.E. 1993. Principles of Physical Cosmology. Princeton, NJ: Princeton
University Press.
Peierls, Rudolf. 1997. Atomic Histories. Woodbury, NY: American Institute of
Physics.
Bibliography 343
Penrose, Roger. 1986. “Energy and Its Definition in General Relativity.” Annals of
the New York Academy of Sciences: Twelfth Texas Symposium on Relativistic
Astrophysics 470 (1): 136–146.
Penrose, Roger. 2005. The Road to Reality: A Complete Guide to the Laws of the
Universe. New York: Vintage Books.
Penrose, Roger. 2010. Cycles of Time: An Extraordinary New View of the Universe.
New York: Vintage Books.
Perlick, V. 2004. “Gravitational Lensing from a Spacetime Perspective,” Living
Reviews in Relativity 7 (1): 9.
Perlmutter, S., G. Aldering, G. Goldhaber, R.A. Knop, P. Nugent, P.G. Castro,
S. Deustua, S. Fabbro, A. Goobar, D.E. Groom, I.M. Hook, A.G. Kim, M.Y.
Kim, J.C. Lee, N.J. Nunes, R. Pain, C.R. Pennypacker, R. Quimby, C. Lidman,
R.S. Ellis, M. Irwin, R.G. McMahon, P. Ruiz-Lapuente, N. Walton, B. Schaefer,
B.J. Boyle, A.V. Filippenko, T. Matheson, A.S. Fruchter, N. Panagia, H.J.M.
Newberg, and W.J. Couch, 1999. “Measurement of Ω and Λ from 42 High-
redshift Supernovae.” The Astrophysical Journal 517: 565–586.
Perry, John. 1976. “The Importance of Being Identical.” In The Identities of Persons,
edited by A.O. Rorty, 67–90. Berkeley: University of California Press.
Perry, John. 1979. “The Problem of the Essential Indexical.” Noûs 13 (1): 3–21.
Phillips, Lawrence D., and Ward Edwards. 1966. “Conservatism in a Simple
Probability Inference Task.” Journal of Experimental Psychology 72 (3):
346–354.
Pimentel, Laura. 2003. “Scurvy: Historical Review and Current Diagnostic
Approach.” American Journal of Emergency Medicine 21 (4): 328–332.
Pincock, Christopher. 2012. Mathematics and Scientific Representation. New York:
Pinker, S. 1995. The Language Instinct: How the Mind Creates Language. New
York: HarperPerennial.
Place, U.T. 1996. “Intentionality as the Mark of the Dispositional.” Dialectica 50
(2): 91–120.
Planck, Max. 1900. “Über eine Verbesserung der Wienschen Spektralgleichung
(Deutsche physikalische Gesellschaft).” Verhandlungen 2: 202–204. Cited as
“Verbesserung”
Planck, Max. 1900. “Zur Theorie des Gesetzes der Energievertheilung im
Normalspektrum (Deutsche physikalische Gesellschaft).” Verhandlungen 2: 237–
245. Cited as “Theorie”
Plantinga, Alvin. 1967. God and Other Minds: A Study of the Rational Justification
of Belief in God. Ithaca, NY: Cornell University Press.
Plantinga, Alvin. 1974. The Nature of Necessity. New York: Oxford University
Press.
Plantinga, Alvin. 1988. “Positive Epistemic Status and Proper Function.” Edited by
James Tomberlin. Philosophical Perspectives 2, 1–50.
Plantinga, Alvin. 1993. Warrant and Proper Function. New York: Oxford University
Press. Cited as “Proper Function”
Plantinga, Alvin. 1993. Warrant: The Current Debate. New York: Oxford University
Press. Cited as “Warrant: The Current Debate”
Plantinga, Alvin. 1993. “Why We Need Proper Function.” Noûs 27 (1): 66–82.
Cited as “Why We Need Proper Function”
344 Bibliography
Plantinga, Alvin. 2000. Warranted Christian Belief. New York: Oxford University
Press.
Plantinga, Alvin. 2010. “Reformed Epistemology.” In A Companion to Philosophy
of Religion, 2nd Edition, edited by Charles Taliaferro, Paul Draper, and Philip L.
Quinn, 674–680. Malden, MA: Blackwell.
Poincaré, Henri. 1905. Science and Hypothesis. Introduction by Josiah Royce with a
special preface by M. Poincaré. Translated by George Bruce Halsted. New York:
The Science Press.
Poincaré, Henri. 1952. Science and Hypothesis. With a preface by J. Larmor.
Translated by W.J.G. New York: Dover.
Polger, T.W., and L.A. Shapiro. 2016. The Multiple Realization Book. Oxford:
Politzer, H.D. 1973. “Reliable Perturbative Results for Strong Interactions?”
Physical Review Letters 30 (26): 1346–1349.
Pooley, Oliver. 2013. “Substantivalist and Relationalist Approaches to Spacetime.”
In The Oxford Handbook of Philosophy of Physics, edited by Robert Batterman,
Popper, Karl. R. 1961. The Logic of Scientific Discovery. New York: Science
Editions, Inc.
Popper, Karl. 2011. “The Propensity Interpretation of Probability.” In Philosophy
of Probability: Contemporary Readings, edited by Antony Eagle, 488–497. New
York: Routledge.
Post, John F. 1987. The Faces of Existence: An Essay in Nonreductive Metaphysics.
Ithaca, NY: Cornell University Press.
Pourciau, Bruce. 2016. “Instantaneous Impulse and Continuous Force: The
Foundations of Newton’s Principia.” In The Cambridge Companion to Newton,
2nd Edition, edited by Robert Iliffe and George E. Smith, 93–186. Cambridge:
Press Release. 2017. The 1993 Nobel Prize in Physics. Nobelprize.org Nobel Media AB
2014. July 18. www.nobelprize.org/nobel_prizes/physics/laureates/1993/press.html.
Price, Huw. 1996. Time’s Arrow and Archimedes’ Point: New Directions for the
Physics of Time. New York: Oxford University Press.
Price, Huw. 2007. “Causal Perspectivalism.” In Causation, Physics, and the
Constitution of Reality: Russell’s Republic Revisited, edited by Huw Price and
Richard Corry, 250–292. New York: Oxford University Press.
Price, Huw. 2017. “Causation, Intervention, and Agency: Woodward on Menzies
and Price.” In Making a Difference: Essays on the Philosophy of Causation,
edited by Helen Beebee, Christopher Hitchcock, and Huw Price, 73–98. Oxford:
Price, Huw, and Brad Weslake. 2009. “The Time-Asymmetry of Causation.” In The
Priest, Graham. 2008. An Introduction to Non-Classical Logic: From If to Is. 2nd
Edition. Cambridge: Cambridge University Press.
Prinz, Jesse J. 2002. Furnishing the Mind: Concepts and Their Perceptual Basis.
Prinz, Jesse J. 2006. “Beyond Appearances: The Content of Sensation and
Perception.” In Perceptual Experience, edited by Tamar Szabó Gendler and John
Hawthorne, 434–460. New York: Oxford University Press.
Bibliography 345
Prior, A.N. 1968. Papers on Time and Tense. Oxford: Clarendon Press.
Pruss, Alexander R. 2006. The Principle of Sufficient Reason: A Reassessment. New
Psillos, Stathis. 1999. Scientific Realism: How Science Tracks Truth. New York:
Routledge.
Psillos, Stathis. 2009. “Regularity Theories.” In The Oxford Handbook of
Purrington, Robert D. 2009. The First Professional Scientist: Robert Hooke and the
Royal Society of London. Basel: Birkhäuser Verlag AG.
Putnam, Hilary. 1962. “What Theories Are Not.” In Logic, Methodology and
Philosophy of Science: Proceedings of the 1960 International Congress, edited
by Ernest Nagel, Patrick Suppes, and Alfred Tarski, 240–251. Stanford, CA:
Stanford University Press.
Putnam, Hilary. 1975. “The Meaning of ‘Meaning’.” In Mind, Language and
Reality: Philosophical Papers, Vol. 2, edited by Hilary Putnam, 215–271.
Cambridge: Cambridge University Press.
Putnam, Hilary. 1979. Mathematics, Matter, and Method: Philosophical Papers.
Volume 1. 2nd Edition. Vol. 1. New York: Cambridge University Press.
Putnam, Hilary. 1983. Realism and Reason: Philosophical Papers Volume 3.
Cambridge: Cambridge University Press.
Putnam, Hilary. 1987. The Many Faces of Realism: The Paul Carus Lectures.
LaSalle, IL: Open Court.
Putnam, Hilary. 1990. Realism with a Human Face. Edited by James Conant.
Putnam, Hilary. 1994. Words and Life. Edited by James Conant. Cambridge, MA:
Putnam, Hilary. 2004. Ethics Without Ontology. Cambridge, MA: Harvard
University Press.
Putnam, Hilary. 2010. Meaning and the Moral Sciences. New York: Routledge.
Quine, W.V.O. 1948/1949. “On What There Is.” The Review of Metaphysics 2:
21–38.
Quine, W.V.O. 1951. “Ontology and Ideology.” Philosophical Studies 2 (1): 11–15.
Quine, W.V.O. 1960. Word & Object. Cambridge, MA: MIT Press.
Quine, W.V.O. 1980. From a Logical Point of View: 9 Logico-Philosophical Essays.
2nd Edition. Revised. Cambridge, MA: Harvard University Press.
Quine, W.V.O. 1985. “Events and Reification.” In Actions and Events: Perspectives
on the Philosophy of Donald Davidson, edited by Ernest LePore and Brian P.
McLaughlin, 162–176. New York: Oxford University Press.
Raftopoulos, A. 2016. “Abduction, Inference to the Best Explanation, and Scientific
Practise: The Case of Newton’s Optics.” In Model-Based Reasoning in Science and
Technology. Studies in Applied Philosophy, Epistemology and Rational Ethics.
Volume 27, edited by L. Magnani and C. Casadio, 259–277. Berlin: Springer.
Ramsey, F.P. 1931. “General Propositions and Causality.” In The Foundations of
Mathematics and Other Logical Essays, edited by Richard B. Braithwaite, 237–
255. With a preface by G.E. Moore. London: Kegan Paul.
Ramsey, William. 2013. “Eliminative Materialism.” In The Stanford Encyclopedia
of Philosophy, Winter 2016, edited by Edward N. Zalta. https://plato.stanford.
edu/archives/win2016/entries/materialism-eliminative/.
346 Bibliography
Ramsey, W., S. Stich, and J. Garon. 1990. “Connectionism, Eliminativism and the
Future of Folk Psychology.” Philosophical Perspectives, 4, Action Theory and
Philosophy of Mind 4: 499–533.
Rasmussen, Joshua. 2014. “Continuity as a Guide to Possibility.” Australasian
Journal of Philosophy 92 (3): 525–538. Cited as “Continuity”
Rasmussen, Joshua. 2014. Defending the Correspondence Theory of Truth.
Cambridge: Cambridge University Press. Cited as “Correspondence”
Rasmussen, Joshua, and C.G. Weaver. forthcoming. “Why Is There Anything?” In
Two Dozen (or so) Arguments for God: The Plantinga Project, edited by Jerry
Walls and Trent Dougherty. Oxford University Press.
Recanati, François. 2000. Oratio Obliqua, Oratio Recta: An Essay on
Metarepresentation. Cambridge, MA: MIT Press.
Redhead, M. 1990. “Explanation.” Edited by Dudley Knowles. Royal Institute of
Philosophy Supplement 27: 135–154.
Redhead, M. 1994. “The Vacuum in Relativistic Quantum Field Theory.” PSA:
Proceedings of the Biennial Meeting of the Philosophy of Science Association. 2
Symposia and Invited Papers (1994): 77–87.
Redhead, M. 1995. “More Ado About Nothing.” Foundations of Physics 25 (1):
123–137.
Reeh, H., and S. Schlieder. 1961. “Bemerkungen zur unitäräquivalenz von
Lorentzinvairanten Feldern.” Nuovo Cimento 22: 1051.
Reichenbach, H. 1956. The Direction of Time. Mineola, NY: Dover.
Reid, T. 1969. Essays on the Active Powers of the Human Mind. Cambridge, MA:
MIT Press.
Reid, T. 1983. Inquiry and Essays. Edited by R.E. Beanblossom and K. Lehrer.
Indianapolis, IN: Hackett.
Rescher, N. 2007. Conditionals. Cambridge, MA: MIT Press.
Rey, D. 2013. “Similarity Assessments, Spacetime, and the Gravitational Field:
What Does the Metric Tensor Represent in General Relativity?”, http://philsci-
archive.pitt.edu/id/eprint/9615 Foundations of Physics 2013: The 17th UK and
European Meeting on the Foundations of Physics (Munich; 29-31 July 2013).
Reyes, Gonzalo E. 2011. “A Derivation of Einstein’s Vacuum Field Equations.”
In Models, Logics, and Higher-Dimensional Categories, edited by Bradd Hart,
Thomas G. Kucera, Anand Pillay, Philip J. Scott, and Robert A.G. Seely, 245–
261. Providence, RI: American Mathematical Society.
Richards, D.G. 2000. “Lattice Gauge Theory—QCD From Quarks to Hadrons,”
arXiv:nucl-th/0006020v2, June 13. Accessed October 4, 2017. https://arxiv.org/
pdf/nucl-th/0006020.pdf.
Richter, Maurice N. 2012. “The Concept of Cognitive Dissonance.” The Journal of
Psychology 60 (2): 291–294.
Rickles, Dean. 2008. “Quantum Gravity: A Primer for Philosophers.” In The
Ashgate Companion to Contemporary Philosophy of Physics, edited by Dean
Rickles, 262–383. Aldershot and Burlington: Ashgate Publishing Limited and
Ashgate Publishing Company. Cited as “Quantum Gravity”
Rickles, Dean. 2008. “Responding to the Hole Problem.” In Symmetry, Structure,
and Spacetime, edited by Dean Rickles, 89–125. Amsterdam, Netherlands:
Elsevier. Cited as “Hole Problem”
Riess, A.G., Alexei V. Filippenko, Peter Challis, Alejandro Clocchiatti, Alan Diercks,
Peter M. Garnavich, Ron L. Gilliland, Craig J. Hogan, Saurabh Jha, Robert P.
Bibliography 347
Kirshner, B. Leibundgut, M.M. Phillips, David Reiss, Brian P. Schmidt, Robert A.
Schommer, R. Chris Smith, J. Spyromilio, Christopher Stubbs, Nicholas B. Suntzeff,
and John Tonry 1998. “Observational Evidence From Supernovae for an Accelerating
Universe and a Cosmological Constant,” Astronomical Journal 116: 1009–1038.
Riess, A.G., Louis-Gregory Strolger, John Tonry, Stefano Casertano, Henry C.
Ferguson, Bahram Mobasher, Peter Challis, Alexei V. Filippenko, Saurabh Jha,
and Weidong Li. 2004. “Type Ia Supernova Discoveries at z >1 From the Hubble
Space Telescope: Evidence for Past Deceleration and Constraints on Dark Energy
Evolution,” The Astrophysical Journal 607: 665–687. For full author list see:
http://iopscience.iop.org/article/10.1086/383612/meta
Rindler, W. 2006. Relativity: Special, General and Cosmological. 2nd Edition. New
Riordan, Michael. 1987. The Hunting of the Quark: A True Story of Modern
Physics. New York: Simon & Schuster.
Roberts, C.D. 2016. “Running Masses in the Nucleon and its Resonances.” JPS
Conference Proceedings: Proceedings of the 10th International Workshop on the
Physics of Excited Nucleons (NSTAR2015) 10: 010012 (pagination is 1–10).
Cited as “Running Masses”
Roberts, C.D. 2016. “Three Lectures on Hadron Physics.” Journal of Physic:
Conference Series 706: 022003 (pagination is 1–53). Cited as “Three Lectures
on Hadron Physics”
Robertson, T. 1998. “Possibilities and the Arguments for Origin Essentialism.”
Mind 107 (428): 729–749.
Rodriguez-Pereyra, Gonzalo. 2002. Resemblance Nominalism: A Solution to the
Problem of Universals. New York: Oxford University Press.
Rodriguez-Pereyra, Gonzalo. 2006. “Truthmakers.” Philosophy Compass 1/2: 186–
200. Cited as “Truthmakers”
Rodriguez-Pereyra, Gonzalo. 2006. “Truthmaking, Entailment, and the Conjunction
Thesis.” Mind 115 (460): 957–982. Cited as “Entailment”
Roessler, J. 2011. “Perceptual Causality, Counterfactuals, and Special Causal
Concepts.” In Understanding Counterfactuals, Understanding Causation: Issues
in Philosophy and Psychology, edited by Christoph Hoerl, Teresa McCormack,
and Sarah R. Beck, 75–89. New York: Oxford University Press.
Romero, C., J.B. Fonseca-Neto, and M.L. Pucheu. 2011. “General Relativity and Weyl
Frames.” International Journal of Modern Physics: Conference Series 3, 27–35.
Rosen, Gideon. 2010. “Metaphysical Dependence: Grounding and Reduction.” In
Modality: Metaphysics, Logic, and Epistemology, edited by Bob Hale and Aviv
Hoffmann, 109–135. New York: Oxford University Press.
Rosen, Gideon. 2017. “Abstract Objects.” In The Stanford Encyclopedia of
Philosophy (Winter 2017 Edition), edited by Edward N. Zalta. https://plato.stan
ford.edu/archives/win2017/entries/abstract-objects/.
Rosenberg, Alex. 1992. “Causation, Probability, and the Monarchy.” American
Philosophical Quarterly 29 (4): 305–318.
Ross, Don, James Ladyman, and David Spurrett. 2007. “Causation in a Structural
World.” In Everything Must Go: Metaphysics Naturalized, James Ladyman,
Don Ross, with David Spurrett and John Collier, 258–297. New York: Oxford
University Press. Cited as “Causation”
Ross, Don, James Ladyman, and David Spurrett. 2007. “In Defense of Scientism.”
In Everything Must Go: Metaphysics Naturalized, James Ladyman, Don Ross,
348 Bibliography
with David Spurrett and John Collier, 1–65. New York: Oxford University Press.
Cited as “Scientism”
Ross, Jacob, and Mark Schroeder. 2014. “Belief, Credence, and Pragmatic
Encroachment.” Philosophy and Phenomenological Research 88 (2): 259–288.
Rovelli, Carlo. 1997. “Halfway Through the Woods: Contemporary Research on
Space and Time.” In The Cosmos Science: Essays of Exploration, edited by John
Earman and John Norton, 180–223. Pittsburgh, PA: University of Pittsburgh Press.
Rovelli, Carlo. 1999. “Quantum Spacetime: What Do We Know?” arXiv.org.
March 12. https://arxiv.org/abs/gr-qc/9903045v1.
Rovelli, Carlo. 2004. Quantum Gravity. New York: Cambridge University Press.
Rueger, Alexander. 1998. “Local Theories of Causation and the a Posteriori
Identification of the Causal Relation.” Erkenntnis 48 (1): 25–38.
Ruetsche, Laura. 2011. Interpreting Quantum Theory: The Art of the Possible.
Rumfitt, Ian. 2015. The Boundary Stones of Thought: An Essay in the Philosophy
of Logic. Oxford: Oxford University Press.
Russell, Bertrand. 1903. The Principles of Mathematics. Vol. 1. Cambridge:
Russell, Bertrand. 1905. “On Denoting.” Mind 14 (4): 479–493.
Russell, Bertrand. 1912–1913. “On the Notion of Cause.” Proceedings of the
Aristotelian Society 13: 1–26.
Russell, Bertrand. 1927. The Analysis of Matter. London: Kegan Paul, Trench,
Trubner & Co. LTD.
Russell, Bertrand. 1959. The Problems of Philosophy. New York: Oxford University
Press.
Ryan, Todd. 2003. “Hume’s Argument for the Temporal Priority of Causes.” Hume
Studies 29 (1): 29–41.
Sachs, R.G. 1987. The Physics of Time Reversal. Chicago: University of Chicago
Press.
Salerno, Joe. 2009. “Introduction.” In New Essays on the Knowability Paradox,
edited by Joe Salerno, 1–10. New York: Oxford University Press. Cited as
“Introduction”
Salerno, Joe. 2009. “Knowability Noir: 1945–1963.” In New Essays on the
Knowability Paradox, edited by Joe Salerno, 29–48. New York: Oxford University
Press. Cited as “Knowability”
Salmon, Nathan. 1989. “Tense and Singular Propositions.” In Themes From Kaplan,
edited by Joseph Almog, Howard Wettstein, and John Perry with the assistance of
Ingrid Deiwiks and Edward N. Zalta, 331–392. Oxford: Oxford University Press.
Salmon, W.C. 1977. “A Third Dogma of Empiricism.” In Basic Problems in
Methodology and Linguistics, edited by Robert E. Butts and Jaakko Hintikka,
149–166. Dordrecht: D. Reidel.
Salmon, W.C. 1984. Scientific Explanation and the Causal Structure of the World.
Princeton, NJ: Princeton University Press.
Salmon, Wesley, C. 1998. Causality and Explanation. New York: Oxford University
Press.
Salmon, W.C. 2010. “The Causal Structure of the World.” Metatheoria 1: 1–13.
Sanders, K. 2009. “On the Reeh-Schlieder Property in Curved Spacetime.”
Communications in Mathematical Physics 288 (1): 271–285.
Bibliography 349
Saunders, Simon. 2003. “Physics and Leibniz’s Principles.” In Symmetries in Physics:
Philosophical Reflections, edited by Katherine Brading and Elena Castellani,
289–307. Cambridge: Cambridge University Press.
Savage, C.W. 1990. “Preface.” In Scientific Theories: Minnesota Studies in the
Philosophy of Science, Vol. XIV, edited by C.W. Savage, vii–ix. Minneapolis:
University of Minnesota Press.
Savitt, Steven. 2011. “Time in the Special Theory of Relativity.” In The Oxford
Handbook of Philosophy of Time, edited by Craig Callender, 546–570. New
Schaffer, Jonathan. 2000. “Causation by Disconnection.” Philosophy of Science 67
(2): 285–300. Cited as “Causation by Disconnection”
Schaffer, Jonathan. 2000. “Trumping Preemption.” The Journal of Philosophy 97
(4): 165–181. Cited as “Trumping Preemption”
Schaffer, Jonathan. 2004. “Causes Need Not Be Physically Connected to Their
Effects: The Case for Negative Causation.” In Contemporary Debates in
Philosophy of Science, 197–216. Malden, MA: Blackwell.
Schaffer, Jonathan. 2005. “Contrastive Causation.” The Philosophical Review 114
(3): 327–358.
Schaffer, Jonathan. 2007. “Review: Dowe and Noordhof, Cause and Chance.” The
British Journal for the Philosophy of Science 58 (4): 869–874.
Schaffer, Jonathan. 2008. “Causation and Laws of Nature: Reductionism.” In
Contemporary Debates in Metaphysics, edited by Theodore Sider, John Hawthorne,
and Dean W. Zimmerman, 82–107. Malden, MA: Blackwell. Cited as “Causation
and Laws”
Schaffer, Jonathan. 2008. “Truth and Fundamentality: On Merricks’s Truth and
Ontology.” Philosophical Books 49 (4): 302–316. Cited as “Truth”
Schaffer, Jonathan. 2009. “On What Grounds What.” In Metametaphysics: New
Essays on the Foundations of Ontology, edited by David J. Chalmers, David
Manley and Ryan, and Wasserman, 347–383. New York: Oxford University Press.
Schaffer, Jonathan. 2010. “The Internal Relatedness of All Things.” Mind 119
(474): 341–376.
Schaffer, Jonathan. 2012. “Disconnection and Responsibility.” Legal Theory 18
(4): 399–435. Cited as “Disconnection”
Schaffer, Jonathan. 2012. “Grounding, Transitivity, and Contrastivity”, In
Metaphysical Grounding: Understanding the Structure of Reality, edited by
Fabrice Correia and Benjamin Schnieder, 122–138. Cambridge: Cambridge
University Press. Cited as “Transitivity”
Schaffer, Jonathan. 2014. “Review of Theodore Sider, Writing the Book of the
World.” The Philosophical Review 123 (1): 125–129.
Schaffer, Jonathan. 2015. “What Not to Multiply Beyond Necessity.” The Australasian
Schaffer, Jonathan. 2016. “Grounding in the Image of Causation.” Philosophical
Studies 173 (1): 49–100. Cited as “Grounding in the Image of Causation”
Schaffer, Jonathan. 2016. “The Metaphysics of Causation.” In The Stanford
Encyclopedia of Philosophy, edited by Edward N. Zalta. Fall. https://plato.stan-
ford.edu/archives/fall2016/entries/causation-metaphysics/. Cited as “Metaphysics
of Causation”
Schiffer, Stephen. 1987. Remnants of Meaning. Cambridge, MA: MIT Press.
350 Bibliography
Schlick, Moritz. 1979. Philosophical Papers: Volume I (1909–1922). Edited by
Henk L. Mulder and F.B. Van De Velde-Schlick. Translated by Peter Heath. Vol.
1. Dordrecht: D. Reidel.
Schneider, Susan. 2007. “Daniel Dennett on the Nature of Consciousness.” In The
Blackwell Companion to Consciousness, edited by Max Velmans and Susan
Schneider, 313–324. Malden, MA: Blackwell.
Schuster, J.A. 2013. “Cartesian Physics.” In The Oxford Handbook of the History
of Physics, edited by Jed Z. Buchwald and Robert Fox, 56–95. New York: Oxford
University Press.
Schutz, B. 2003. Gravity from the Ground Up: An Introductory Guide to Gravity
and General Relativity. New York: Cambridge University Press.
Schwarz, Wolfgang. 2015. “Analytic Functionalism.” In A Companion to David
Lewis, edited by Barry Loewer and Jonathan Schaffer, 504–518. Malden, MA:
Blackwell.
Schwitzgebel, Eric. 2011. Perplexities of Consciousness. Cambridge, MA: MIT
Press.
Schwitzgebel, Eric. 2015. “Belief.” The Stanford Encyclopedia of Philosophy.
Summer 2015. Edited by Edward N. Zalta. Stanford, CA. https://plato.stanford.
edu/archives/sum2015/entries/belief/, June 21.
Sciences: Class for Physics of the Royal Swedish Academy of Sciences. 2011.
“Scientific Background on the Nobel Prize in Physics 2011: The Accelerating
Universe.” Nobel Prize. October 4. Accessed December 11, 2017. www.nobel-
prize.org/nobel_prizes/physics/laureates/2011/advanced-physicsprize2011.pdf.
Scriven, Michael. 1975. “Causation as Explanation.” Noûs 9 (1): 3–16.
Seager, William. 1991. Metaphysics of Consciousness. London: Routledge.
Seager, William. 2016. Theories of Consciousness: An Introduction and Assessment.
2nd Edition. New York: Routledge.
Searle, John. 1958. “Proper Names.” Mind 67 (266): 166–173.
Segal, Aaron. 2015. “Half-hearted Humeanism.” In Oxford Studies in Metaphysics.
Vol. 9, edited by Karen Bennett and Dean W. Zimmerman, 262–305. New York:
Shafer-Landau, Russ. 2003. Moral Realism: A Defence. New York: Oxford
University Press.
Shankar, R. 2014. Fundamental Physics: Mechanics, Relativity, and Thermodynamics.
New Haven, CT: Yale University Press.
Shankar, R. 2016. Fundamentals of Physics II: Electromagnetism, Optics, and
Quantum Mechanics. New Haven, CT: Yale University Press.
Sherwood, Michael. 1969. The Logic of Explanation in Psychoanalysis. New York
and London: Academic Press.
Shoemaker, Sydney. 1997. “Self and Substance.” In Philosophical Perspectives:
Mind, Causation, and World, edited by J.E. Tomberlin, 283–304. Atascadero,
CA: Ridgeview.
Shoemaker, Sydney. 2007. Physical Realization. New York: Oxford University Press.
Sider, Theodore. 2001. Four Dimensionalism: An Ontology of Persistence and
Time. New York: Oxford University Press.
Sider, Theodore. 2009. “Ontological Realism.” In Metametaphysics: New Essays
on the Foundations of Ontology, edited by David J. Chalmers, David Manley,
and Ryan Wasserman, 384–423. New York: Oxford University Press.
Sider, Theodore. 2011. Writing the Book of the World. New York: Oxford
University Press.
Bibliography 351
Sider, Theodore. 2013. “Against Parthood.” Vol. 8. In Oxford Studies in
Metaphysics, edited by Karen Bennett and Dean W. Zimmerman, 237–293. New
Sider, Theodore. 2014. “Hirsch’s Attack on Ontologese.” Noûs 48 (3): 565–572.
Siegel, Susanna. 2006. “Which Properties Are Represented?” In Perceptual
Experience, edited by Tamar Szabó Gendler and John Hawthorne, 481–503.
Siegel, Susanna. 2009. “The Visual Experience of Causation.” The Philosophical
Quarterly 59 (236): 519–540.
Siegel, Susanna. 2010. The Contents of Visual Experience. New York: Oxford
University Press.
Simons, Peter. 2003. “Events.” In The Oxford Handbook of Metaphysics, edited
by Michael J. Loux and Dean W. Zimmerman, 357–385. New York: Oxford
University Press.
Sklar, Lawrence. 1976. Space, Time, and Spacetime. Berkeley: University of California
Press.
Sklar, Lawrence. 1993. Physics and Chance: Philosophical Issues in the Foundations
of Statistical Mechanics. New York: Cambridge University Press.
Skow, Bradford. 2014. “Are There Non-Causal Explanations (of Particular
Events)?” The British Journal for the Philosophy of Science 65 (3): 445–467.
Skyrms, Brian. 1980. Causal Necessity: A Pragmatic Investigation of the Necessity
of Laws. New Haven and London: Yale University Press.
Sloman, Steven. 2005. Causal Models: How People Think About the World and Its
Alternatives. New York: Oxford University Press.
Smeenk, C., and E. Schliesser. 2013. “Newton's Principia.” In The Oxford
Smith, G.E. 2016. “The Methodology of the Principia.” In The Cambridge
Companion to Newton, 2nd Edition, edited by Rob Iliffe and George E. Smith,
187–228. Cambridge: Cambridge University Press.
Smith, Sheldon R. 2013. “Causation in Classical Physics.” In The Oxford Handbook
of Philosophy of Physics, edited by Robert Batterman, 107–140. New York:
Smithson, Michael. 2016. “Human Understandings of Probability.” In The Oxford
Handbook of Probability and Philosophy, edited by Alan Hájek and Christopher
Hitchcock, 477–496. New York: Oxford University Press.
Smolensky, Paul. 1988. “On the Proper Treatment of Connectionism.” Behavioral
and Brain Sciences 11 (1): 1–23.
Snowdon, Paul. 2011. “Perceptual Concepts as Non-Causal Concepts.” In
Perception, Causation, and Objectivity, edited by J. Roessler, N. Eilan, and H.
Lerman, 121–138. Oxford: Oxford University Press.
Soames, Scott. 1987. “Direct Reference, Propositional Attitudes, and Semantic
Content.” Philosophical Topics 15 (1): 47–87.
Soames, Scott. 1989. “Semantics and Semantic Competence.” Philosophical
Perspectives: 3, Philosophy of Mind and Action (Ridgeview Publishing) 3:
575–596.
Soames, Scott. 2005. Reference and Description: The Case Against Two-
Dimensionalism. Princeton, NJ: Princeton University Press.
Soames, Scott. 2008. “Truth and Meaning: In Perspective.” Midwest Studies in
Philosophy 32: (1): 1–19.
352 Bibliography
Soames, Scott. 2010. Philosophy of Language. Princeton, NJ: Princeton University
Press. Cited as “Language”
Soames, Scott. 2010. What Is Meaning? Princeton, NJ: Princeton University Press.
Cited as “Meaning”
Soames, Scott. 2014. “Cognitive Propositions.” In New Thinking About
York: Oxford University Press. Cited as “Cognitive Propositions”
Soames, Scott. 2014. “Propositions vs Properties and Facts.” In New Thinking
About Propositions, edited by Jeffrey C. King, Scott Soames, and Jeff Speaks,
166–181. New York: Oxford University Press. Cited as “Propositions”
Sober, Elliott. 1984. The Nature of Selection: Evolutionary Theory in Philosophical
Focus. Chicago: University of Chicago Press.
Sober, Elliott. 2002. “Bayesianism-Its Scope and Limits.” In Bayes’s Theorem,
edited by Richard Swinburne, 21–38. Oxford: Oxford University Press.
Sosa, Ernest. 1975. “Introduction.” In Causation and Conditionals, edited by Ernest
Sosa. London: Oxford University Press.
Sosa, Ernest. 1983. “Classical Analysis.” The Journal of Philosophy 80 (11):
695–710.
Sosa, Ernest. 1993. “Review: Proper Functionalism and Virtue Epistemology.”
Noûs 27 (1): 51–65.
Sosa, Ernest. 2003. “Privileged Access.” In Consciousness: New Philosophical
Essays, edited by Quentin Smith and Aleksandar Jokic, 273–292. Oxford: Oxford
University Press.
Sosa, Ernest. 2007. A Virtue Epistemology: Apt Belief and Reflective Knowledge.
Vol. 1. New York: Oxford University Press.
Sosa, Ernest. 2009. Reflective Knowledge: Apt Belief and Reflective Knowledge.
Vol. 2. New York: Oxford University Press.
Sosa, Ernest. 2011. Knowing Full Well. Princeton, NJ: Princeton University Press.
Sosa, Ernest. 2015. Judgment and Agency. New York: Oxford University Press.
Speaks, Jeff. 2014. “Propositions Are Properties of Everything or Nothing.” In New
Thinking About Propositions, by Jeffrey C. King, Scott Soames, and Jeff Speaks,
Speaks, Jeff. 2015. The Phenomenal and the Representational. Oxford: Oxford
University Press.
Sperber, Dan. 1995. “Introduction.” In Causal Cognition: A Multidisciplinary
Debate, edited by Dan Sperber, xv–xx. New York: Oxford University Press.
Spirtes, P., C. Glymour, and R. Scheines. 2000. Causation, Prediction and Search.
With additional material by David Heckerman, Christopher Meek, Gregory F.
Cooper, and Thomas Richardson. 2nd Edition. Cambridge, MA: MIT Press.
Stachel, John. 2002. Einstein From ‘B’ to ‘Z’. Boston, MA: Birkhäuser.
Stalnaker, Robert. 1984. Inquiry. Cambridge, MA: MIT Press.
Stanley, J. 2011. Know How. New York: Oxford University Press.
Stefánsson, H. Orri. 2017. “What Is ‘Real’ in Probabilism?” The Australasian
Steglich-Petersen, A. 2012. “Against the Contrastive Account of Singular Causation.”
The British Journal for the Philosophy of Science 63 (1): 115–143.
Stein, Howard. 1977. “On Space-Time Ontology: Extracts from a Letter to Adolf
Grünbaum.” In Minnesota Studies in Philosophy of Science Volume 8, edited by
Bibliography 353
John S. Earman, Clark N. Glymour, and John J. Stachel, 374–402. Minneapolis:
University of Minnesota Press.
Stein, Howard. 1989. “Yes, but . . . Some Skeptical Remarks on Realism and Anti-
Realism.” Dialectica 43 (1/2): 47–65.
Stein, Howard. 2002. “Newton’s Metaphysics,” In The Cambridge Companion to
Newton, edited by I. Bernard Cohen and George E. Smith, 1st Edition, 256–307.
New York: Cambridge University Press.
Sterman, George, John Smith, John C. Collins, James Whitmore, Raymond Brock,
Joey Huston, Jon Pumplin, Wu-Ki Tung, Hendrik Weerts, and Chien-Peng Yuan
et al. 1995. “Handbook of Perturbative QCD.” Reviews of Modern Physics 67
(1): 157–248.
Steup, M. 2016. “Epistemology.” In The Stanford Encyclopedia of Philosophy,
edited by Edward N. Zalta. Fall. https://plato.stanford.edu/archives/fall2016/
entries/epistemology/.
Stone, I. 1965. “Studies of a Mammalian Enzyme System for Producing Evolutionary
Evidence in Man.” American Journal of Physical Anthropology 23: 83–86.
Strawson, Galen. 2015. “ ‘Humeanism’.” Journal of the American Philosophical
Association 1 (1): 96–102.
Strevens, Michael. 2008. Depth: An Account of Scientific Explanation. Cambridge,
MA: Harvard University Press.
Strevens, Michael. 2012. “Précis of Depth.” Philosophy and Phenomenological
Research 84 (2): 447–460.
Strevens, Michael. 2013. Tychomancy: Inferring Probability From Causal Structure.
Styer, Daniel F., Miranda S. Balkin, Kathryn M. Becker, Matthew R. Burns,
Christopher E. Dudley, Scott, T. Forth, Jeremy S. Gaumer, et al. 2002. “Nine
Formulations of Quantum Mechanics.” American Journal of Physics 70 (3):
288–297.
Summers, S.J. 2011. “Yet More Ado About Nothing: The Remarkable Vacuum
State.” In Deep Beauty: Understanding the Quantum World Through
Mathematical Innovation, edited by Hans Halverson, 317–342. New York:
Suppe, Frederick. 1974. “The Search for Philosophic Understanding of Scientific
Theories.” In The Structure of Scientific Theories, edited by Frederick Suppe,
3–232. Urbana: University of Illinois Press.
Suppe, Frederick. 1989. The Semantic Conception of Theories and Scientific
Realism. Urbana and Chicago: University of Illinois Press.
Suppes, Patrick. 1970. A Probabilistic Theory of Causality. Amsterdam: North-
Holland.
Swenson Jr., L.S. 1972. The Ethereal Aether: A History of the Michelson-Morley-
Miller Aether-Drift Experiments, 1880–1930. Austin: University of Texas Press.
Swerdlow, N.M. 2013. “Galileo’s Mechanics of Natural Motion and Projectiles.”
In The Oxford Handbook of the History of Physics, edited by Jed Z. Buchwald
and Robert Fox, 25–55. New York: Oxford University Press.
Swinburne, Richard. 1997. “The Irreducibility of Causation.” Dialectica 51 (1):
79–92.
Swinburne, Richard. 2001. Epistemic Justification. Oxford: Oxford University
Press.
354 Bibliography
Swinburne, Richard. 2007. “From Mental/Physical Identity to Substance Dualism.”
In Persons: Human and Divine, edited by Peter van Inwagen and Dean
Zimmerman, 142–165. New York: Oxford University Press.
Swoyer, Chris. 2008. “Abstract Entities.” In Contemporary Debates in Metaphysics,
edited by Theodore Sider, John Hawthorne, and Dean W. Zimmerman, 11–31.
Malden, MA: Blackwell.
‘t Hooft, Gerard, and Martinus J.G. Veltman. 1994. “Diagrammar.” Vol. 19. In
Under the Spell of the Gauge Principle (Advanced Series in Mathematical Physics),
by Gerard ‘t Hooft, 28–173. World Scientific.
Tabery, James G. 2004. “Synthesizing Activities and Interactions in the Concept of
a Mechanism.” Philosophy of Science 71 (1): 1–15.
Tamir, M. 2012. “Proving the Principle: Taking Geodesic Dynamics Too Seriously
in Einstein’s Theory.” Studies in History and Philosophy of Modern Physics 43
(2): 137–154.
Tavakol, R., and G. Ellis. 1999. “Holography and Cosmology.” Physics Letters B
469: 37–45.
Tavakol, R., and R. Zalaletdinov. 1998. “On the Domain of Applicability of
General Relativity.” Foundations of Physics 28 (2): 307–331.
Taylor, Richard. 1966. Action and Purpose. Englewood Cliffs, NJ: Prentice Hall.
Taylor, James E. 1991. “Plantinga’s Proper Functioning Analysis of Epistemic
Warrant.” Philosophical Studies 64 (2): 185–202.
Taylor, John R. 2005. Classical Mechanics. Mill Valley, CA: University Science
Books.
Taylor, John R., Chris D. Zafiratos, and Michael A. Dubson. 2015. Modern Physics:
For Scientists and Engineers. Mill Valley, CA: University Science Books.
Tennant, Neil. 1997. The Taming of the True. Oxford: Clarendon Press.
Tennant, Neil. 2015. Introducing Philosophy: God, Mind, World, and Logic. New
York: Routledge.
Ter Harr, D. 1971. Elements of Hamiltonian Mechanics. 2nd Edition. New York:
Pergamon Press.
Thorne, K.S. 1994. Black Holes and Time Warps: Einstein’s Outrageous Legacy.
New York: W.W. Norton & Company.
Thorne, Kip S., and Roger D. Blandford. 2017. Modern Classical Physics: Optics,
Fluids, Plasmas, Elasticity, Relativity, and Statistical Physics. Princeton, NJ:
Tooley, Michael. 1987. Causation: A Realist Approach. New York: Oxford
University Press.
Tooley, Michael. 1990. “The Nature of Causation: A Singularist Account.” The
Canadian Journal of Philosophy 20 Issue sup1: Supplementary Volume 16
Canadian Philosophers: Celebrating Twenty Years of the CJP: 271–322.
Tooley, Michael. 2003. “Causation and Supervenience.” In The Oxford Handbook
of Metaphysics, edited by Michael J. Loux and Dean W. Zimmerman, 386–434.
Torretti, R. 1983. Relativity and Geometry. New York: Pergamon Press.
Tumulka, Roderich. 2006. “A Relativistic Version of the Ghirardi-Rimini-Weber
Model.” Journal of Statistical Physics 125: 821–840. Cited as “Relativistic
Version of GRW”
Tumulka, Roderich. 2006. “On Spontaneous Wave Function Collapse and Quantum
Field Theory.” Proceedings of the Royal Society A 462: 1897–1908. Cited as
“Spontaneous Wave Function”
Bibliography 355
Turri, John. 2015. “Knowledge and the Norm of Assertion: A Simple Test.”
Synthese 192 (2): 385–392.
Tversky, Amos, and Daniel Kahneman. 1983. “Extensional Versus Intuitive
Reasoning: The Conjunction Fallacy in Probability Judgement.” Psychological
Review 90: 293–315.
Vaidman, L. 2014. “Quantum Theory and Determinism.” Quantum Studies:
Mathematics and Foundations 1 (1–2): 5–38.
Vaidman, Lev. 2016. “Many-Worlds Interpretation of Quantum Mechanics.” In
The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta. Fall.
Accessed November 2017. https://plato.stanford.edu/archives/fall2016/entries/
qm-manyworlds/.
Vallinder, Aron. forthcoming. “Imprecise Bayesianism and Global Belief Inertia.”
The British Journal for the Philosophy of Science 1–26.
van Cleve, James. 1985. “Why a Set Contains Its Members Essentially.” Noûs 19
(4): 585–602.
van Fraassen, Bas C. 1970. “On the Extension of Beth’s Semantics of Physical
Theories.” Philosophy of Science 37 (3): 325–339.
van Fraassen, Bas C. 1980. The Scientific Image. Oxford: Clarendon Press.
van Fraassen, Bas C. 1989. Laws and Symmetry. Oxford: Oxford University Press.
van Fraassen, Bas C. 1991. Quantum Mechanics: An Empiricist View. Oxford:
van Fraassen, Bas C. 2008. Scientific Representation: Paradoxes of Perspective.
van Inwagen, Peter. 1983. An Essay on Free Will. New York: Oxford University
Press.
van Inwagen, Peter. 1990. Material Beings. Ithaca, NY: Cornell University Press.
van Inwagen, Peter. 2001. “Objectivity.” In Metaphysics: Contemporary Readings,
edited by Michael J. Loux, 524–539. New York: Routledge.
van Inwagen, Peter. 2009. “Being, Existence, and Ontological Commitment.” In
Metametaphysics: New Essays on the Foundations of Ontology, edited by David
J. Chalmers, David Manley, and Ryan Wasserman, 472–506. New York: Oxford
University Press.
Wade, D.T., and P.W. Halligan. 2004. “Do Biomedical Models of Illness Make for
Good Healthcare Systems?” The British Medical Journal 329: 1398–1401.
Wald, Robert M. 1984. General Relativity. Chicago: University of Chicago Press.
Wald, Robert M. 2006. “The Arrow of Time and the Initial Conditions of
the Universe.” Studies in History and Philosophy of Modern Physics 37 (3):
394–398.
Wall, Aron C. 2013. “A Discrete, Unitary, Causal Theory of Quantum Gravity.”
Classical and Quantum Gravity 30: 115002.
Wasserman, Ryan. 2005. “Humean Supervenience and Personal Identity.” The
Philosophical Quarterly 55 (221): 582–593.
Wasserman, Ryan. 2016. “Theories of Persistence.” Philosophical Studies 173 (1):
243–250.
Weatherall, J.O. 2011. “A Brief Remark on Energy Conditions and the Geroch-Jang
Theorem.” https://arxiv.org/abs/1106.2336.
Weatherson, Brian. 2005. “Can We Do Without Pragmatic Encroachment?”
Philosophical Perspectives, 19, Epistemology 19: 417–443.
Weaver, C.G. 2012. “What Could Be Caused Must Actually Be Caused.” Synthese
184 (3): 299–317.
356 Bibliography
Weaver, C.G. 2013. “A Church-Fitch Proof for the Universality of Causation.”
Synthese 190 (14): 2749–2772.
Weaver, C.G. 2016. “Yet Another New Cosmological Argument.” International
Journal for Philosophy of Religion 80 (1): 11–31.
Weaver, C.G. 2017. “On the Carroll-Chen Model.” Journal for General Philosophy
of Science 48 (1): 97–124.
Weaver, C.G. forthcoming. “An Objection to Naturalism and Atheism From Logic.”
In The Blackwell Companion to Atheism and Philosophy, edited by Graham
Oppy. Malden, MA: Blackwell.
Weinberg, Steven. 1972. Gravitation and Cosmology: Principles and Applications
of the General Theory of Relativity. New York: John Wiley & Sons, Inc.
Weinberg, Steven. 1993. The First Three Minutes: A Modern View of the Origin of
the Universe. Updated Edition. New York: Basic Books.
Weinberg, Steven. 2005. The Quantum Theory of Fields: Volume II Modern
Applications. New York: Cambridge University Press.
Weinberg, Steven. 2008. Cosmology. New York: Oxford University Press.
Weinberg, Steven. 2013. Lectures on Quantum Mechanics. 1st Edition. New York:
Weinert, Friedel. 2005. The Scientist as Philosopher: Philosophical Consequences of
Great Scientific Discoveries. Berlin: Springer-Verlag.
Weinstein, S. 2001. “Absolute Quantum Mechanics.” The British Journal for
Weintraub, Ruth. 2013. “Induction and Inference to the Best Explanation.”
Philosophical Studies 166 (1): 203–216.
Weiskrantz, Lawrence. 1980. “Varieties of Residual Experience.” Quarterly Journal
of Experimental Psychology. 32 (3): 365–386.
Weiskarantz, Lawrence. 1986. Blindsight: A Case Study and Implications. Oxford:
Clarendon Press.
Weiskrantz, Lawrence. 1988. “Some Contributions of Neuropsychology of
Vision and Memory to the Problem of Consciousness.” In Consciousness in
Contemporary Science, edited by A.J. Marcel and E. Bisiach, 183–199. Oxford:
Weslake, Brad. 2014. “Statistical Mechanical Imperialism.” In Chance and
Temporal Asymmetry, edited by Alastair Wilson, 241–257. New York: Oxford
University Press.
Westfall, Richard S. 1971. Force in Newton’s Physics: The Science of Dynamics in
the Seventeenth Century. New York: American Elsevier.
Weyl, Hermann. 1952. Space-Time-Matter. 4th Edition. Translated by Henry L.
Brose. Mineola, NY: Dover.
Wheeler, J.A. 1998. With Kenneth Ford. Geons, Black holes, and Quantum Foam:
A Life in Physics. New York: W.W. Norton and Company, Inc.
Whewell, William. 1967. The Philosophy of the Inductive Sciences. 3rd Edition.
Vol. 1. London: Frank Cass & Co. LTD. Cited as “vol. 1”
Whewell, William. 1967. The Philosophy of the Inductive Sciences. 3rd Edition.
Vol. 2. London: Frank Cass. & Co. LTD. Cited as “vol. 2”
White, Roger. 2010. “You Just Believe That Because . . . ” Philosophical Perspectives,
Epistemology 24: 573–615.
Whitrow, G.J. 1980. The Natural Philosophy of Time. 2nd Edition. Oxford:
Clarendon Press.
Bibliography 357
Whittaker, E. 1987. A History of the Theories of Aether and Electricity: I. The
Classical Theories (The History of Modern Physics 1800–1950). College Park,
MD: American Institute of Physics.
Wigner, Eugene. 1985. “Events, Laws of Nature, and Invariance Principles.” In
How Far Are We from the Gauge Forces, edited by A. Zuchichi, 699–708. New
York: Plenum.
Williams, D.C. 1953. “On the Elements of Being I.” Review of Metaphysics 7 (1): 3–18.
Williamson, Timothy. 2000. Knowledge and Its Limits. New York: Oxford
University Press.
Williamson, Timothy. 2007. The Philosophy of Philosophy. Malden, MA: Blackwell.
Wilson, Leonard G. 1975. “The Clinical Definition of Scurvy and the Discovery
of Vitamin C.” Journal of the History of Medicine and Allied Sciences 30 (1):
40–60.
Wilson, N.L. 1974. “Facts, Events and Their Identity Conditions.” Philosophical
Studies 25 (5): 303–321.
Woodward, James. 2003. Making Things Happen: A Theory of Causal Explanation.
Woodward, James. 2009. “Agency and Interventionist Theories.” In The Oxford
Handbook of Causation, edited by Helen Beebee, Christopher Hitchcock, and
Peter Menzies, 234–262. New York: Oxford University Press.
Woodward, James. 2015. “Methodology, Ontology, and Interventionism.” Synthese
192 (11): 3577–3599.
Woodward, James. 2016. “Causation and Manipulability.” The Stanford Encyclopedia
of Philosophy. Edited by Edward N. Zalta. December 21. Accessed December 11,
2017. https://plato.stanford.edu/archives/win2016/entries/causation-mani/.
Woodward, James, and Christopher Hitchcock. 2003. “Explanatory Generalizations,
Part I: A Counterfactual Account.” Noûs 37 (1): 1–24.
Wootton, David. 2011. “Galileo: Reflections on Failure.” In Causation and Modern
Philosophy, edited by Keither Allen and Tom Stoneham, 13–30. New York:
Routledge.
Worrall, John. 1989. “Structural Realism: The Best of Both Worlds?” Dialectica
43 (1/2): 99–124.
Worrall, John. 2007. “Miracles and Models: Why Reports of the Death of Structural
Realism May Be Exaggerated.” In Philosophy of Science: Royal Institute of
Philosophy Supplement: 61, edited by Anthony O'Hear. Cambridge University
Press. Volume 61: 125–154.
Worrall, John. 2012. “Miracles and Structural Realism.” In Structural Realism:
Structure, Object, and Causality, edited by Elaine M. Landry and Dean P. Rickles,
77–95. Dordrecht: Springer.
Wudka, J. 2006. Space-Time, Relativity, and Cosmology. Cambridge: Cambridge
University Press.
Yablo, Stephen. 1997. “Wide Causation.” Philosophical Perspectives, Volume 11,
Mind, Causation, and World 11: 251–281.
Yablo, Stephen. 2004. “Advertisement for a Sketch of an Outline of a Prototheory
of Causation.” In Causation and Counterfactuals, edited by John Collins, Ned
Hall, and L.A. Paul, 119–137. Cambridge, MA: MIT Press.
Yang, C.N., and R.L. Mills. 1954. “Conservation of Isotopic Spin and Isotopic
Gauge Invariance.” Physical Review 96: 191–195.
Zee, A. 2013. Einstein Gravity in a Nutshell. Princeton, NJ: Princeton University Press.
358 Bibliography
Zee, A. 2016. Group Theory in a Nutshell for Physicists. Princeton, NJ: Princeton
University Press.
Zimmerman, Dean. 2008. “The Privileged Present: Defending an ‘A-Theory’ of
Time.” In Contemporary Debates in Metaphysics, edited by Theodore Sider, John
Hawthorne, and Dean W. Zimmerman, 211–225. Malden, MA: Blackwell.
Zynda, Lyle. 2016. “Subjectivism.” In The Oxford Handbook of Probability and
Philosophy, edited by Alan Hájek and Christopher Hitchcock, 360–381. New
Index
ɅCDM model 165 – 166 how to defeat argument from physics

with minimal assumptions 257 – 261;
abduction 44 – 51 minimal fundamentalist account
aboutness 9, 59, 86, 171 261 – 263; objections to causal
absolute space 121 interpretation of GTR 273 – 280;
acausal representation 61 – 62 overview 252 – 255; on pushing
Achinstein, Peter 87 causation out of physics and on
Acta Eruditorum 116 how to reintroduce it 257 – 263
ADM-formulation 53, 258 argumentive loophole, interaction for
agency/perspectival analysis of closing 267 – 268
causation (AP-A) 158 – 159, 289 – 290 argumentum ad populum 198
agents as causal relata 231 – 232 argumentum ad verecundiam 52
Aguilar, A.C. 122 Aristotle 40, 89, 102, 117, 182,
Albert, David Z. 62, 153 – 154 199, 234
Alkon, Daniel L. 38 Armstrong, David M. 2, 12 – 13, 104,
Allori, Valia 157, 230 110, 169, 203 – 204, 221, 226, 239,
Alston, William P. 22 241, 262, 284, 287, 297 – 300
Ampère, Andre Marie 68 Armstrong’s door 298 – 301
Analytical Mechanics (Lagrange) 68 Aronson, Jerrold L. 124
Analytical Theory of Heat (Fourier) 68 Audi, Robert 38, 163
Anjum, Rani Lill 13, 110, 169 Aune, Bruce 231
anomalous monism (AM) 4
Anscombe, G.E.M. 101 back-reaction, causal asymmetry and
anti-reductionists 124 – 125 277 – 278
anti-reductive proposal 295 – 298 backward causation 113, 142, 152,
anti-reductive theories 268, 287, 295 274 – 275
AP-A see agency/perspectival analysis Baker, Lynne Rudder 90
of causation Banks, Tom 55, 137, 167
AP-AM see modified AP-A Barbour, Julian 67
a priori: evidence for deterministic Barrow, Isaac 116
interpretation of QM 157; intuition Bayesianism, subjective 44 – 47
38 – 44, 71; justification 38, 43; in Bayesian probability theory 44 – 47
science 39 – 44 Bayes net 208
a priori intuition 38 – 44 Bayes theorem 46
argument from physics and general BBN see big bang nucleosynthesis
relativity 252 – 286; argument Bealer, George 38 – 39
255 – 256; domains of causal Beck, Jacob Scott 31
influence in cosmology 272 – 273; Beebee, Helen 228
gravitational field as cause 263 – 272; Bell, John Stewart 70, 156, 230
360 Index
Bennett, Jonathan 140, 227 – 228, 240 Carroll, Sean M. 221 – 222
Bennett, Karen 111, 271 Carruthers, P. 92
Bergmann, Michael 27, 29, 38, 219 Cartesian vortices 67
Bernoulli, Daniel 68 Cartwright, Nancy 12 – 13, 111, 154,
Bernstein, Sara 239, 241 169, 257
best systems account (BSA) of laws of causal anti-reductionists 124 – 125
nature 270 causal arrow 153, 300
Bianchi identities 267 causal asymmetry, back-reaction and
biconditionals 9 – 10, 30, 33, 124, 292 277 – 278
big bang nucleosynthesis (BBN) 165 – 166 causal claims (CCs) 107, 138, 159, 241,
bijuncts 30 – 31 289 – 290, 302
Bird, Alexander 101 – 102, 297 causal eliminativists 24, 93 – 99
Black, Max 221, 232, 294 causal futures 272 – 273
Blandford, Roger D. 206, 222 causal interpretation of GTR,
Bloom, Paul 200 objections to 273 – 280; back-
Blundell, Stephen J. 237 reaction and causal asymmetry
Bohm-de Broglie approach 157 277 – 278; causal GTR presupposes
Bohmian mechanics 229 substantivalism 278 – 279; causal
Bohr, Niels 66, 69 – 70 talk is gloss 279 – 280; GTR is not
Boltzmann, Ludwig 62, 96 – 98, 129 fundamental 273 – 274; time-reversal
Bolzano, Bernard 14 invariance and closed timelike curves
BonJour, Laurence 171 274 – 277
Bonk, Thomas 164 causal loops 204 – 207
Bonnell, Jerry 15 causal models 67, 208, 236 – 238, 268
Boyle, Robert 37, 61 – 62 causal pasts 272 – 273
Boyle’s ideal gas law 61 – 62 causal phenomena 65, 67, 92 – 93, 112,
Bradley’s regress 104 121, 124, 135 – 137, 187, 254, 257,
Braithwaite, R.B. 199 261, 264, 268
Brand, Myles 224 causal power (CP) 98 – 104, 123, 262,
Brett, J.M. 123 292, 297
Briggs, Rachael 291 causal priority: from physics 130 – 133;
Brown, Harvey R. 264 – 265 from temporal priority 112 – 130
Brush, Stephen G. 129 – 130, 130 causal reductionists 124 – 125
brute asymmetry of causation causal relata 215 – 251; causal
110 – 151; causal priority from eliminativists and 93 – 99; changes
physics 130 – 133; causal priority as 224 – 225; concrete changes
from temporal priority 112 – 130; as 224 – 225; concrete particular
motivation and thesis 110 – 111; objects as 223 – 224; in defense of
objections 133 – 143; overview of 89 – 109; events as, new account
110 – 112; PRE-CD 111 – 113, 124, of 232 – 234; forces as 234 – 238;
133 – 134, 135, 137 – 138, 140, fundamental 134, 231; Lewis and
141, 143 the void 103 – 105; Lowe’s case
BSA of laws of nature see best systems against 99 – 103; multigrade relation
account (BSA) of laws of nature as 216 – 217; negative causation
Burge, Tyler 31, 160 and 238 – 244; obtaining 89 – 93;
Burns, J.J. 243 overview of 215 – 216; physics and
93, 95 – 99; property exemplifications
CAC see Counterfactual Analysis as 228 – 231; proposals, competing
of Causation 217 – 232; sets of space-time regions
Callender, Craig 96 as 225 – 227; substances and agents
Carnap, Rudolf 25 as 231 – 232; theory of, making
Carpenter, Kenneth 243 217 – 223; tropes as 227 – 228
Carroll–Chen model 224, 231 causal representation 62 – 71
Index 361
causal skepticism, hyperrealism and CLC (classical logical consequence)
168 – 171 23 – 24
causal talk in GTR 279 – 280 Cleland, Carol 225
causation 287 – 307; after reductionism Clement, Corey 189
287 – 288; anti-reductive proposal closed-causal-curves (CCCs) 274 – 276
295 – 298; Armstrong’s door and closed timelike curves (CTCs) 274 – 277
298 – 301; counterfactual dependence CMBR see cosmic microwave
account of 137 – 143; getting into background radiation
physics 93, 95 – 99; interventionist Coady, C.A.J. 51, 163
manipulability account and 288 – 292; cognizers: in epistemological isolation
irreflexivity of 198 – 207; mechanistic objection 160, 163, 168 – 175;
account and 292 – 295; transitivity metaphysical theories 1, 3, 5 – 6, 20,
of 207 – 210; well-foundedness of 26 – 28, 35 – 40, 43, 46, 48, 49, 66;
210 – 211 in obtaining causal relations 89 – 92;
“Causation and the Price of in property exemplifications 230;
Transitivity” (Hall) 111 puzzlement of properly functioning
causation hunting 62 – 71 199; in transitivity 208
causation-is-a-multigrade-relation Cohen, I. B. 120
thesis 217 Coleman, S. 132
causation schema 33 Collier, John 221
cause, gravitational field as 263 – 272 Collins, John 55, 298 – 300
Cavendish, Henry 68 completeness, doctrine of 7 – 8
CCs see causal claims Complete Works (Aristotle) 40, 182
CCCs see closed-causal-curves composition, as many-one relation 271
CERN Large Hadron Collider 122 Compton, Arthur H. 130
Cetto, Ana Maria 130 concept interactionc 123
CFOL see classical first-order logic conceptual relativity 21
Chalmers, David J. 5, 157 conceptual view (CV) 161 – 164
changes: causal relata 224 – 225; concrete changes 225
concrete 225 concrete particulars 89, 99, 223 – 224
change theory 225 concrete phase (CP) 225
Chase, Carl Trueblood 69 Connors, Michael H. 92
Chen, Jennifer 123, 179, 221, 224 conservation laws 136, 253 – 254
Chew, Geoffrey 236 conserved-quantity theory of causation
Chisholm, Roderick 27, 215 136 – 137
Choquet–Bruhat, Yvonnne 120 constitution, as relation between
Church, Alonzo 23, 156, 164, 183 – 190 collocated massive bodies 271
Church–Fitch argument from causal contingent events/facts 157, 190, 194
explanation 183 – 190 continuity 157, 186, 221, 232
Church–Fitch knowability result 156 contrastive underdetermination
Churchland Materialism 6 164 – 168
Churchland, Paul 91 – 92, 94, 204 Copenhagen non-relativistic quantum
circularity 142, 158, 295 mechanics 53
Clapham, Christopher 15 correspondence/aboutness relation 9
Clarke, Randolph 240 correspondence rules 53, 78
classical Boltzmannian SM 96 correspondence theory 9, 59
classical first-order logic (CFOL) 23 – 25 cosmic microwave background
classical logical consequence (CLC) radiation (CMBR) 166 – 168
23 – 24 cosmological model, standard 254
classical propositional logic (CPL) cosmology, domains of causal influence
23, 100 in 272 – 273
Class of Physics of the Royal Swedish Counterfactual Analysis of Causation
Academy of Sciences 259 (CAC) 30 – 31
362 Index
counterfactual arguments in Pruss and direct warrant 27
Weaver 190 – 195 DiSalle, Robert 269 – 270
counterfactual dependence account domains of causal influence in
137 – 143 cosmology 272 – 273
CP see causal power Dowe, Phil 34 – 35, 113, 134 – 137, 220,
CPL see classical propositional logic 253 – 254, 256
credence 44 – 47 Drummond, Jack Cecil 243 – 244
Crude Sufficient Condition Account of Dubson, Michael A. 55, 130
Causation (CSC) 202 – 203 Ducasse, C.J. 169
CSC see Crude Sufficient Condition Duhem, P. 164
Account of Causation Dummett, Michael 21, 113
CTCs see closed timelike curves Dyson, Freeman J. 235 – 236
Curiel, Erik 267
CV see conceptual view early preemption 299 – 300
CWP see water’s causal power Earman, John S. 207, 220, 264, 276
Edwards, Ward 46
d’Alembertian operator 115 Eells, Ellery 143, 203, 256
Daly, Helen 4 Eells-Sober Probabilistic Analysis 203
Danks, David 169 EFEs see Einstein’s field equations
Darrigol, Olivier 66, 129 (EFEs)
Davidson, Donald 4, 28 – 29, 182, Ehrenfest, Paul 55
223 – 224 Ehring, Douglas 111, 228
declarative sentences 2, 14, 20, 56 Einstein, Albert 54 – 55, 66, 69,
de Coulomb, Charles Augustin 68, 119 120 – 121, 128 – 130, 218 – 220, 254,
default setting 21, 210, 217 258 – 259, 263 – 267, 269 – 270, 278
definitional dependences 278 Einstein’s field equations (EFEs):
De Gandt, F. 116 applied to Robertson-Walker metric
de la Peña, Luis 130 165 – 166; de Sitter or vacuum
Dennett, Daniel 94 solution to 59; energy-momentum
density 62, 115, 128, 132, 166, 263, tensor applied to 128, 166, 253,
267, 277 259, 263 – 264, 267; generalized 267;
dependence: counterfactual 141; geodesic equations of motion and
existential 302; functional 61, 63; 161, 263 – 265, 267, 268; Lorentzian
negative 82; ontological 97; positive metric in 126 – 127; partial
51; probabilistic 261; relation 11, 61, interpretation 59 – 60; QCD
97, 261 – 262, 301 – 302; transitive 16 and 56; time travel and 204 – 205
Descartes, René 116, 145 electrodynamics 11, 54, 113 – 114,
designative formulations 56 – 59, 64, 119, 121, 127, 129, 131, 204, 222,
218, 220, 236 257 – 258, 268
de Sitter solution to Einstein’s field electrons 2, 59, 129 – 130, 165, 203,
equations (EFEs) 59 204, 235, 242, 249, 294
de Sitter space-time 167, 221, 231, 234 Elga, Adam 138 – 140
d’Espagnat, Bernard 70 eliminativists 93 – 99
determination, non-causal 268 – 269 Ellis, G.F.R. 127, 186, 254, 273,
determinism 102, 139, 157, 187 284 – 285
deterministic causation 2, 30 – 31, 137, emission 55, 150, 166, 204, 238
232, 298, 303 energy 55, 96, 101, 129 – 130, 166,
Dialogo (Galileo) 116 186, 213, 242, 252 – 254, 257, 259,
Dirac, P.A.M. 55, 69, 114, 213 261, 281
directionality 110 – 113, 128, 130, 134, energy bands 242
138, 152, 169, 171, 173 energy-momentum tensor 126, 128,
directives 3 – 6, 19, 22, 24, 25, 30, 137, 166, 176, 205, 253 – 254, 259,
33 – 36, 38, 44, 51 – 52, 163, 233, 248 263 – 264, 267, 277 – 278
Index 363
EP see equivalence principle 121 field equations, Newtonian 70
epiphenomenalism 5 Field, Hartry 160 – 162
epistemic justification 30, 72, 80, first-order logic 22
163 – 164, 172, 221; a priori intuition Fitch, F.B. 156, 164, 183 – 190
and 38 – 44; basic sources of 36 – 38; FLRW metric 254
expert testimony and 52; IBEs and Fodor, Jerry A. 31
47 – 48; non-basic sources of 51 – 52 Folkman, Susan 91
epistemic structural realism (ESR) formalism 53 – 58
11 – 13, 71 Forschungen und Fortschritte
epistemological isolation objection (Einstein) 266
152 – 181; see also hyperrealism Fourier, Joseph 68
epistemology of metaphysics 36 – 71; Friedman equations 59, 166 – 167
a priori intuition 38 – 44; epistemic Friedmann–Robertson–Walker solution
justification 36 – 38; other non-basic 59, 165 – 166, 166
sources 51 – 52; physical theory Frölich, Theodor 243 – 244
structure and, propositional view of fundamental causal relata 134, 231
52 – 71; probability and abduction
44 – 51 Gale, R. 187
equivalence principle (EP) 121 Galileo 39 – 44, 67; Dialogo 116; space-
ESR see epistemic structural realism time 229 – 230; thought experiment
“Essay for the Apostles on ‘Analogies in 39 – 44
Nature’ ” (Maxwell) 69 Galison, Peter 40, 66
ESSI (Events as States of Substances gauge theory, non-Abelian 131 – 132
at Indices) 117, 233 – 234, 238, Gauss’s law of classical
240 – 241, 252, 261, 298, 301 electromagnetism 114
Euler, Leonhard 68, 278 Gearhart, Clayton A. 129
events as causal relata, new account of Gendler, Tamar Szabó 42
232 – 234 general theory of relativity (GTR):
Events as States of Substances at Indices context of 278 – 279; formalism 261,
(ESSI) see ESSI (Events as States of 272; instantaneous causation and
Substances at Indices) 125, 125 – 129; interpretation of 120,
“Experimental Studies Relating to Ship- 126, 127; laws of geodesic motion
Beri-Beri and Scurvy” (Frölich and 269; in Newtonian limit 120 – 121;
Holst) 243 – 244 partial interpretation of 258 – 259;
expert testimony, epistemic justification see also causal interpretation of
and 52 GTR, objections to
geodesic motion in physics 263 – 267
facts, non-causal 150, 156, 159 – 164, geodesic principle 128, 265,
172, 175 267, 270
Fadin, S. 132 Geroch, Robert 121, 267, 272
Fair, David 136, 252 – 253, 256 Ghirardi, G.A. 157, 229 – 230
Fales, Evan 169, 180 Giere, Ronald N. 260
Fara, Michael 241 Gillies, Donald 232
Faraday, Michael 69, 258 Gillmor, C. Stewart 68
Faraday’s Law (FL) 258 Glennan, Stuart 125, 292 – 295
Farhi/Guth/Guven mechanism for baby gluons 122, 131 – 132, 190, 234 – 235,
universes 221 238, 295
Festinger, Leon 200, 212 Glymour, Clark N. 124
Feynman diagrams 235 – 238, 268 God 5, 49, 266
Feynman, Richard P. 204 – 205, Gödel metric 204
235 – 237, 242 Gold model of the universe 168, 180
field equations, Einstein’s see Einstein’s Goldstein, Sheldon 156, 178
field equations (EFEs) Goodman, Nelson 26
364 Index
Gott, J. Richard, III 113 Hitchcock, Christopher 63
Gowers, Timothy 14 Hoefer, Carl 273 – 274
Gravetter, Frederick J. 123 Hofweber, Thomas 22
gravitational field as cause 263 – 272; Holst, Axel 243 – 244
causal interaction for closing Holton, Gerald 130
argumentive loophole 267 – 268; Hooke, Robert 67
geodesic motion in contemporary Horgan, Terence 4 – 5
physics 263 – 267; gravitational Horwich, Paul 44, 113, 138, 140
interactions, non-causal determination Howard-Snyder, Daniel 179
and 268 – 269; interaction as causal Howson, Colin 44
phenomena 268; Mill–Ramsey–Lewis Hubble constant 166
best-systems treatment 269 – 270; Huemer, M. 113 – 115
other treatments 270 – 272 Hulse-Taylor pulsar 138 – 140
gravitational theory, Newtonian Hume, David 6 – 7, 112 – 113, 117 – 118,
115 – 120, 126 – 127 141, 152, 154 – 155, 158
Green’s function 114 Humeanism, Sider’s 6 – 11
Grice, H.P. 185 Humphreys, P. 98
Grommer, J. 264, 266 Hutchins, Robert Maynard 68
grounding: direct coupling and Huygens, Christiaan 67, 120
definitional dependency facts 278; hyper-eliminative materialism
Schaffer’s theory of 50, 271 – 272, (hyper-EM) 95
288, 301 – 304; theories of causal hyperrealism 152 – 181; causal
relata and 223; truth-conditions and skepticism 168 – 171; debunking
16; truthmaking and 61; universal arguments 171 – 175; overview
causal determination and 184 – 185 of 152 – 155; underdetermination
grueness 26 argument 155 – 168
GTR see general theory of relativity hyper-realist-consistent theory
(GTR) 168
Gupta, Rajan 55
IBE see inferences to the best
haecceitistic properties 190 – 191 explanation
Hájek, Alan 232 ideal gas law, Boyle’s 61 – 62
Halligan, Peter 91 – 92 I-FL see interpreted FL
Halvorson, Hans 54 IMA see interventionist manipulability
Hamiltonian mechanics 53, 96 account
Hanks, Patrick 123 IMA*-P see possibility constraining
Hardy, Lucian 274 interpretation of interventions
Harman, Gilbert 46 – 47 IMA*-S see settings intervention
Hartle, James B. 147 approach
Hasker, William 91 Infeld, L. 196 – 197
Hauptsatz of scientific representation inferences to the best explanation (IBE)
56 47 – 48, 51, 93, 282
Havas, Peter 264 inferential causal knowledge 160
Hawking, Stephen 127, 161, 254, 273 instantaneous causation 125 – 129
Healey, Richard 70, 132 – 133, 157, interaction: as causal concept 123;
176, 178 as causal phenomena 121 – 125; as
Heathcote, Adrian 124, 132 – 133, 203 causal phenomena in gravitational
Heisenberg, Werner 53, 66 field 268; as causal term 123 – 124;
Helliwell, T.M. 55, 114 for closing argumentive loophole
Hempel, Carl G. 199 267 – 268; non-causal determination
Herbrand’s theorem 23 and 268 – 269; quark 235
Hertz, Heinrich 129 inter alia loca 8
Hiddleston, Eric 291 intermediate value theorem 15
Hilbert spaces 54, 186 interpreted FL (I-FL) 258
Hirsch, Eli 24 – 25 intervention 125, 289 – 290
Index 365
interventionist manipulability account Lancaster, Tom 237
(IMA) 288 – 292 Landau, Lev 236
intrinsicness thesis (IT) 167 – 168, Lange, Marc 55, 93, 114
297, 302 Laplace, Pierre-Simon 182
intuition, a priori 38 – 44 Laplacian operator 115, 120
Ioffe, B.L. 132 late preemption 298
irreflexive causation from irreflexive lattice gauge 55 – 56
explanation 198 – 201 law of classical electromagnetism 114
irreflexivity 198 – 207; argument for law of inertia 54
201 – 204; causal loops and 204 – 207; laws of motion 113 – 114
from irreflexive explanation 198 – 201; Lebesgue measure 98
time travel and 204 – 207 Lehmkuhl, Dennis 277 – 278
Irzik, G. 111 Leibnizian space-time 120, 207, 221,
Ismael, J.T. 63 232 – 233, 294
IT see intrinsicness thesis Leighton, Robert B. 242
Lemmon, J. 223 – 224
Jackson, John David 55 Lewis, David K. 30, 89, 91, 95, 103,
James, L.R. 123, 163 – 164, 208 103 – 107, 110, 135, 137 – 143,
Jang, S. 267 159, 182, 184 – 185, 204, 215,
Janssen, M. 121 221, 225 – 227, 240, 252, 256, 261,
Jeffrey, Richard 44 269 – 270, 290; Mill–Ramsey–Lewis
Jenkins, C.S. 157 best-systems treatment 269 – 270;
J-facts 271 similarity metrics and 137, 139 – 140,
Jörg, Ton 123 252, 290 – 291; theory of causation
Joyce, James M. 44 – 45 111, 138, 140 – 143, 150 – 151, 204,
226, 291; void and 103
Kaiser, David 236 – 237 Leyton, Michael 123
Kania, Andrew 191 Liddle, A.R. 165 – 166
Kant, Immanuel 32 Lincoln, Abraham 101
Kaplan, David 32 Liouville measure 98
Kent, Adrian 229 – 230 Lipatov, N. 132
Kentridge, R.W. 37 Locke, John 169
Kepler, J. 116 – 117 Loewer, Barry 141, 153
Khan, M.D.N. 130 logic choice, quantifiers and 22 – 24
Killing vector field 254 Lombard, Lawrence 224 – 225
Kim, Jaegwon 98, 117, 134, 228 – 231 Lorentz, H.A. 206, 235, 257, 268;
King, Jeffrey C. 2, 18 – 20, 30 force law of classical electrodynamics
Kirchoff, Gustav 129 113, 268; gauge condition 126;
Kittel, Charles 242 invariance 219 – 220, 278; metric
Klein, Sanford 96 tensor 126 – 128, 259, 263;
Koizumi, Ayako 200 transformations 55, 114, 219 – 220
Koons, Robert 111, 182, 185, 210 Lorentz–Heaviside–Maxwell equations
Korman, Daniel Z. 171 of classical electrodynamics 257 – 258
Kovitz, B. 113 – 115 Lormand, Eric 37
Kripke, Saul 22, 32, 105, 230 Loux, Michael 21
Kugler, Peter 123 Lowe, E.J. 61, 99 – 103, 262
Kuhn, Thomas 129 Lycan, William 48
Kutach, Douglas 34 – 36, 124, 136, 201 Lyth, D.H. 165 – 166
Kvanvig, J.L. 187
Kvart, I. 209 MA see mechanistic account
Mach, Ernst 93, 207
Ladyman, James 19 – 20 Machian space-time 207
Lagrange, Joseph-Louis 53, 68, macrostate 96 – 98
257, 278 Majors, B. 226
Lambourne, Robert J.A. 263 Malament, D.B. 167, 266 – 267
366 Index
many-one relation 271 epistemology of metaphysics 36 – 71;
materialism 6, 92, 94 – 95, 204 metaphysical methodology 2 – 13;
mathematics 6, 9, 12, 14 – 16, 53, 55, metaphysical theorizing 22 – 71;
58, 62, 64, 185, 221, 237, 278; non- propositions and truth-conditions
mathematical explanations 201; 14 – 20
principles 44 metaphysical theory: in metaphysicalC
Mattuck, Richard D. 237 – 238 system building 22 – 71; of omissions
Maudlin, Tim 21, 118, 154, 187 239 – 241; of propositions 20
Maxwellian electromagnetic theory metaphysical theses 19, 39, 49 – 52, 221
129, 222, 268 metaphysical truth-conditions 9 – 10
Maxwellian space-time 207 metaphysical worldview 1 – 88; building
Maxwell, James Clerk 11, 66, 69 – 70, 30 – 36; epistemology of metaphysics
119, 127, 129, 131, 161, 207, and 36 – 71; true 1, 25, 30
222, 257 methodological principles (MPs): MP1
MC see Mellor’s Criterion 3, 11, 71; MP2 5, 13 – 14, 89, 90;
McGrew, T. 171 MP3 6
Meaning, Soames 18 metrics of similarity 291
mechanics, classical 39, 53 – 54, 96, microstates 96 – 98
113 – 115, 222, 257 Millikan, Robert A. 130
mechanistic account (MA) 125, Mill, John Stuart 91, 269 – 270
292 – 295 Mill–Ramsey–Lewis best-systems
mechanistic anti-reductive theory 292 treatment 269 – 270
Mellor, D.H. 93 – 94, 98, 240 – 241 Mills, R.L. 131, 269 – 270
Mellor’s Criterion (MC) 93 minimal fundamentalist account
mental causation 4 – 5, 12 – 13, 261 – 263
89 – 90, 173 Minkowski, Hermann: interpretation
mental states 3 – 6, 11, 13, 21 – 22, 26, of special relativity 217 – 218; metric
30, 37, 89 – 90, 94 – 95, 99, 101, 303 126 – 127; space-time 186, 272 – 273,
Menzies, Peter 150, 154, 156 – 159, 278
162, 174, 256, 262, 298 modal logic, quantified 22, 183
mereological universalism 217 modified AP-A (AP-AM) 159
Merlin’s spell-casting event 272 Molnar, George 102
Mermin, N. David 218 – 220 Moon, Andrew 93
Merricks, Trenton 2, 173 – 174 Moore, Michael S. 70, 169, 239
metametaphysics 2, 27 Moreland, J.P. 104, 226
metaphysical doctrine 49 Morris, Michael S. 127 – 128,
metaphysical inquiry 1, 30, 33, 36, 38, 264 – 265, 275
47, 51 – 52, 61; conducive 24 – 25, Moser, Paul K. 38, 42
27 – 28; context of 44; knowledge- motivation and thesis 110 – 111
seeking 28 Mulaik, S.A. 123
metaphysical methodology 2, 6, 11, multifarious theories of causal
20, 26, 74, 90; epistemic structural explanation 163, 261
realism and metaphysics of science multigrade relation as causal relata
11 – 13; grueness 26; methodological 216 – 217
principles 3 – 6; naturalistic 28; Mumford, Stephen 13, 101 – 103, 169
obtaining causal relations 90; of Mundy, B. 118
Ross, Ladyman, and Spurrett 19 – 20;
self-stultification in, avoiding 2 – 13; Nagel, Ernest 199
Sider’s Humeanism 6 – 11, 28; truth- naturalistic metaphysical methodology
aimed 1 28
metaphysicalC system building 1 – 71, ND see neutron degeneration
90 – 91, 110, 217, 241, 290; analysis negative causation 238 – 244; cases of,
30 – 36; default setting 21 – 22; in science 241 – 244; causal relata as
Index 367
238 – 244; metaphysical theories of Oppy, Graham 187
omissions and 239 – 241 Opticks (Newton) 116, 120
Nellis, Gregory 96 Oxford Dictionary of Physics 124
Nemiroff, Robert 15 Oxford English Dictionary (OED)
neo-Russellianism 17 – 18, 63 – 64, 123 – 124
93 – 95, 99
neuron firings 141, 216, 298 – 301 Page, Don 166 – 167
neutron degeneration (ND) 138 Pais, Abraham 69, 266
Newtonian limit 120 – 121, 126 – 127 P&W see Price, Huw; Weslake, Brad
Newtonian mechanics 54, 58, 121, 152, P&W-O see Price-Weslake objection
238, 257, 298 Parfit, Derek 226
Newton, Isaac 54, 58, 68, 70, 87, partial interpretation of a true or
113 – 120, 144 – 148, 303; conception approximately true F (PIFs) 58 – 60
of absolute space 121; field equations partial interpretations of conjectural
70; gravitational theory 115 – 120, or theoretical formulations (PICs)
126 – 127; law of inertia 54; laws of 59 – 60, 167, 267
motion 113 – 114; Opticks 116, 120; Paul, L.A. 124, 136, 153, 159, 179,
Poisson equation 115, 127; Principia 202, 209, 287, 298 – 301, 304
54, 115 – 116, 120 Pearl, Judea 208, 236, 288
Ney, Alyssa 229 Pearsall, Judy 123
Nicrosini, O. 230 Penrose, Roger 96, 254
non-Abelian gauge theory 131 – 132 Perlmutter, S. 166
non-causal determination 268 – 269 Perry, John 32, 221
non-causal facts 150, 156, 159 – 164, perturbative QCD 55 – 57
172, 175 Phillips, Lawrence D. 46
non-causally interpreted SC 167 philosophy 21, 26, 31, 63, 67
non-grue-like predicates 27 – 28, photoelectric effect 129 – 130
233, 297 physical interactions as causal
non-mathematical explanations 201 phenomena 121 – 125; concept
non-relativistic QM 53, 113, 258 interactionc 123; evidence from
Norton, John D. 40 – 43, 273 consensus among causal reductionists
Novick, L.R. 123 and causal anti-reductionists 124 – 125;
interaction as causal term 123 – 124;
Obama, Barack 33 prima facie evidence 121, 122 – 123,
objections 133 – 143; to causal 268
interpretation of GTR 273 – 280; physical reductive explanation of
conserved quantity theory of causal direction (PRE-CD) 111 – 113,
causation 136 – 137; counterfactual 124, 133 – 134, 135, 137 – 138, 140,
dependence account of causation 141, 143
137 – 143; epistemological isolation physical theory structure 52 – 71;
152 – 181; reductive accounts acausal representation and 61 – 62;
135 – 136; reductive theories of causal representation and causation
causation 133 – 135; time-reversal hunting and 62 – 71; formalism and
invariance 133 53 – 58; full theory interpretation and
obtaining causal relata 89 – 93 60 – 61; partial interpretation and
O’Connor, Timothy 231 – 232 58 – 60; propositional view of 52 – 71
OED see Oxford English Dictionary physics: argument on pushing causation
omissions, metaphysical theories of out of, and on how to reintroduce it
239 – 241 257 – 263; causal priority from 130 – 133;
operators 48, 96, 109, 113, 115, geodesic motion in 263 – 267; getting
120, 126, 156, 186, 188, 258, causation into 93, 95 – 99; how to
263, 293 defeat argument from, with minimal
Oppenheim, Paul 199 assumptions 257 – 261
368 Index
PICs see partial interpretations probability theory, Bayesian 44 – 47
of conjectural or theoretical projection, problem of 26 – 30
formulations property exemplifications as causal
PIFs see partial interpretation of a true relata 228 – 231
or approximately true F property instances 16, 97, 203,
Pincock, Christopher 64 – 65 225 – 228, 271
Planck’s constant 129, 222 propositions: metaphysical theory
Plantinga, Alvin 26 – 29, 36 – 38, of 20; nature of 17, 20; new
180, 183 verificationism 19 – 20, 71; physical
Plantinga/Bergmann theory of theory structure and 52 – 71;
warrant 38 representational properties of
PLO see plural first-order logic 16, 18 – 19, 101; true negative
plural first-order logic (PLO) 22 existential 240; truth-conditions
PNC see principle of naturalistic closure in 14 – 20
Poincaré, Henri 65 – 66 Pruss, Alexander R. 187, 190
point coincidence argument 220 Psillos, Stathis 136
Poisson equation 115, 127 PSR see principle of sufficient reason
Polchinski, Joseph 275 purely platonic events (PPEs)
Polger, T.W. 97 191 – 192, 194
Popper, Karl 199 Putnam, Hilary 21, 32, 71
possibility constraining interpretation Pylyshyn, Zenon W. 31
of interventions (IMA*-P) 289 – 292
PPC see principle of platonic causality QCD see quantum chromodynamics
PPEs see purely platonic events QED see quantum electrodynamics
PRE-CD see physical reductive QFT see quantum field theory
explanation of causal direction QG see quantum gravity
preemption 279 – 280, 287, 298 – 301; QM see quantum mechanics
early 299 – 300; late 298; trumping QPS see quantum physical system
300 – 301 QSM see quantum statistical
Premise 1 210, 214, 256 mechanics
Premise 2 211, 256 quantified modal logic 22, 183,
Premise 3 211, 256 196 – 197
Premise 4 189 – 190, 211, 256 quantifier 7, 9, 14, 22 – 28, 30
Price, Huw 153 – 160, 162, 174, quantum chromodynamics (QCD)
180, 262 40, 55 – 57, 121 – 122, 125,
Price-Weslake objection (P&W-O) 131 – 1 35, 142, 206, 234 – 236,
154 – 156, 159, 161, 164 – 165, 168, 238, 295; causal priority from
171 – 172, 174 – 175 physics 131 – 133; fundamental
Priest, Graham 183 quanta of 132, 235; interpreting
prima facie evidence 121, 122 – 123, 132; lattice gauge 55 – 56;
268 literature 132
principal root 15 quantum electrodynamics (QED)
Principia (Newton) 54, 115 – 116, 120 121, 131, 204 – 206, 213, 235 – 236,
principle K-A 1, 4, 6, 227, 288 238, 268
principle of naturalistic closure (PNC) quantum event 186 – 187
19 – 20 quantum fields 55, 132 – 133, 242
principle of platonic causality (PPC) quantum field theory (QFT) 55, 108,
191, 194 124, 131, 133, 157, 178, 186 – 187,
principle of sufficient reason (PSR) 231, 242, 275
68, 138 quantum gravity (QG) 274
principle P 63 quantum mechanics (QM) 11, 23, 54,
Prinz, Jesse J. 180 69, 113, 147, 156 – 157, 161, 229,
probabilism 44 – 46 259, 273 – 274
probability 44 – 51 quantum physical system (QPS) 157, 229
Index 369
quantum statistical mechanics (QSM) Robertson-Walker metric 165 – 166
96, 222, 245 Rodriguez-Pereyra, Gonzalo 104
quantum theory 70, 157, 187, 229, Rosenberg, Alex 111
238, 258 – 259, 266, 273 – 274 Ross, Don 19 – 20
quark interactions 235 Ross, Ladyman, and Spurrett (RLS)
Quine, W.V.O. 7, 223 – 224 19 – 20
Rovelli, Carlo 122, 128, 279 – 280
Ramsey, Frank P. 91, 269 – 270 Rueger, Alexander 129, 150, 253
Ramsey sentence 12, 95 Ruetsche, Laura 258 – 260
Rasmussen, Joshua 157, 178, 186 Rumfitt, Ian 23
realism 21 – 22, 217, 222, 226, 260; Russell, Bertrand 14, 17, 63, 93,
adopted 11; causal 89, 110, 153; 104, 109
default setting and 210; favored 191;
robust 29; scientific 63, 65, 134, 185, Salerno, Joe 156, 187
217, 256, 260; structural 11 Salmon, Wesley C. 199
Realism Regained (Koon) 210 Sands, Matthew 242
realists 4 – 5, 24, 28, 102 Santos, Laurie R. 200
realization, as many-one relation 271 Savage, C.W. 55
received view 53 – 54 Savitt, Steven 219
Redhead, Michael 63, 186 SC see standard cosmology
reductio ad absurdum 8, 40 scalar quantity 186, 219, 278
reductionism 92, 153 – 155, 159 – 163, Schafferian grounding 50, 271 – 272,
177, 287 – 288; causation after 288, 301 – 304
287 – 288 Schaffer, Jonathan 2, 16, 50, 62 – 63,
reductionists 10, 124 – 125, 133 – 136, 111, 124, 136, 153, 209, 239,
153 – 154, 161, 164 – 165, 167 – 168, 271 – 272, 288, 301 – 304
171, 252, 255 – 256, 264, 273 – 275, Scheines, R. 124
279 – 280, 287, 298 Schlick, Moritz 93, 121
reduction strategies, local 280 Schlieder, S. 186
reductive accounts 135 – 136; of Schrödinger equation 229
causation 135 – 136 Schrödinger, Erwin 53, 70, 156, 229
reductive theories of causation 133 – 143 Schwitzgebel, Eric 92
redundancies, truths without logical 8 science: a priori in 39 – 44; metaphysics
Reeh, H. 186 of 11 – 13; negative causation in
Reeh–Schlieder theorem 186 241 – 244
Reid, T. 169 scientific realism 63, 65, 134, 185, 217,
relational predicate 156, 187 256, 260
relativity, conceptual 21, 71 scientific theory 11, 19, 136, 256, 260
relatum 13, 215 SCP see self-consistency principle
representational properties of Scriven, Michael 212
propositions 16, 18 – 19, 101 Seager, William 5
Rescher, Nicholas 140 Segal, Ariel 129, 213
restricted underdetermination argument self-consistency principle (SCP)
163 – 164 275 – 276
Ricci tensor 206 self-stultification in metaphysical
Richards, D.G. 56 methodology, avoiding 2 – 13
Richter, Maurice N. 200 settings intervention approach
Rickles, Dean 219, 274 (IMA*-S) 289 – 290
Riemann curvature tensor 206 Shafer-Landau, Russ 226
Riess, A.G. 166 Shankar, R. 118, 130
Rimini, A. 156 – 157, 229 – 230 Shapiro, L.A. 97
RLS see Ross, Ladyman, and Spurrett Shoemaker, Sydney 98, 221
Roberts, C.D. 122, 133 Sider, Theodore 6 – 11, 22, 24 – 25, 28,
Robertson, T. 59, 166 71, 241
370 Index
Siegel, Susanna 169 Summers, S.J. 186
similarity-based semantics 291 Sun’s gravitational force 116, 118
similarity metrics 137 – 140, 228, 252, supervenience 96 – 97, 270 – 271
290 – 291 Suppe, Frederick 260
simpliciter 64, 114, 288, 291 Suppes, Patrick 260
Sitter space-time 167, 221, 231, 234 Swampman scenario 28 – 29
Slavnov, A.A. 132 syntactic view 53 – 54
Sloman, Steven 66, 123, 237
SM see statistical mechanics Tabery, James G. 125
S-matrix theory 236 Tamir, M. 264, 267
Smith, John 55 Taylor, John R. 55, 113, 130
Smith, Sheldon R. 114 Taylor, Richard 113
Soames, Scott 14, 18, 32 TDB see truth or true things depend on
Sober, Elliott 203, 256 reality or being
Sosa, Ernest 28 – 30, 38 – 39, 170 temporal priority, causal priority
space-time: de Sitter 167, 221, 231, from 112 – 130; GTR and 120 – 121,
234; Galilean 229 – 230; Leibnizian 125 – 129; Newtonian gravitation
221, 232 – 233, 294; Machian 207; 115 – 120, 126 – 127; Newtonian limit
Maxwellian 207; Minkowskian 186, 120 – 121, 126 – 127; photoelectric
272 – 273, 278; regions as causal effect 129 – 130; physical interactions
relata 225 – 227; regions, causal relata as causal phenomena 121 – 125
as sets of 225 – 227; substantivalism Tennant, Neil 32
103, 223, 278 – 279 testability 50
special relativity (SR) 54 – 55, 113, testimony 42, 51 – 52, 163 – 164, 179
113 – 114, 217 – 218, 218 – 220, 253 theoretical unification 42, 50
special theory of relativity (STR) 55, theory interpretation 60 – 61
69, 126, 272, 277 – 278 theory of causal relata 217 – 223
Spirtes, P. 124 theory of grounding 16, 50, 184, 288,
spontaneous collapse theory 156 – 157, 301 – 302
178, 229 – 230 theory of omissions 209, 239 – 241
spontaneous wave function 230 – 231 theory of propositions (T-P) 20
Spurrett, David 19 – 20 thesis, motivation and 110 – 111
SR see special relativity ‘t Hooft, Gerard 236
standard cosmology (SC) 39, 165 – 168, Thorne, Kip S. 127 – 129, 206, 213,
180, 202 – 203, 254 – 255, 272, 222, 264 – 265, 275
275 – 276 thought experiment 39 – 41, 43 – 44
Stanley, J. 160 “Three Lectures on Hadron Physics”
statistical mechanics (SM) 62, 96 – 99, (Roberts) 122, 133
131, 168, 222 time-reversal invariance 133, 274;
Stefánsson, H. Orri 44 closed timelike curves and 274 – 277;
Steglich-Petersen, A. 194 objections and 133
Stein, Howard 11, 113 time travel 204 – 207
stellar evolution, SC and 166 TMT see truth-maker theory
Sterman, George 55 Tooley, Michael 111, 136, 143
Steup, M. 180 176 – 177, 180
Stich, S. 91 “Topics in the Foundations of GR”
stimulatory signal 298 – 300 (Malament) 167, 267
STR see special theory of relativity T-P see theory of propositions
Strawson, Galen 155, 177 transitivity 207 – 210; counter-examples
Strevens, Michael 180, 185, 199, 261 to (cases) 208 – 210; overview
structural realism 11 207 – 208
subjective Bayesianism 44 – 47 Treatise on Electricity and Magnetism
substances as causal relata 231 – 232 (Maxwell) 69
substantivalism 103, 223, 278 – 279 Treatise on Light (Huygens) 67, 69
Index 371
Trinity device 216 Veltman, Martinus J.G. 236
tropes as causal relata 227 – 228 verificationism 19 – 20, 71
true metaphysical worldview 1, 25, 30 Verificationist Principle (VP)
true negative existential propositions 19 – 20, 273
240 void, Lewis and 103 – 105
true things depend on reality or being von Helmholtz, Herman 69
(TDB) 2 – 3, 50, 61, 104, 272, vortices 67, 69
300 – 301 VP see Verificationist Principle
trumping preemption 300 – 301
truth-aimed metaphysical Wald, Robert M. 59, 126, 218, 254,
methodology 1 263, 266, 272, 275, 279
truth and meaning: propositions Wall, Aron C. 274
and truth-conditions 14 – 20; self- Wallnau, Larry 123
stultification in metaphysical warrant 1, 3, 26 – 28, 30, 38 – 39,
methodology 2 – 13 41, 45 – 46, 48, 71, 90, 163, 172;
truth-bearers 2, 9, 19 – 21, 30, 61, 173 knowledge-conducive 33, 164, 177;
truth-conditions 8 – 9, 14 – 20, 16 – 17, theory of 169
58, 191, 252 warrantK 26 – 28, 164
truth-evaluable 58, 104, 173 Wasserman, Ryan 221
truth-maker theory (TMT) 2 water’s causal power (CWP) 100
truth or true things depend on reality or Weatherson, Brian 45
being (TDB) 2 – 3, 50, 61, 104, 272, Weaver, Christopher G. 23, 183,
300 – 301 186 – 187, 190, 221
Tumulka, Roderich 156, 230 – 231 Weber, T. 156 – 157, 229 – 230
Turvey, Michael 123 Weinberg, Steven 129, 165, 272
Weinert, Friedel 266
UCP see universal causal principle Weiskrantz, Lawrence 37
underdetermination 155 – 168, 175, well-foundedness of causation 210 – 211
177; causal concepts 162 – 163; Weslake, Brad 153 – 165, 168, 171 – 172,
characterization of 155; contrastive 174, 262
164 – 168; formula 155 – 156; Weyl, Hermann 264
restricted argument 163 – 164 Wheeler, John A. 127 – 129, 264 – 265,
universal causal determination 182 – 197; 275, 277
Church–Fitch argument from causal Whewell, William 68 – 69
explanation 183 – 190; counterfactual Williams, D.C. 104
arguments in Pruss and Weaver Williamson, Timothy 241
190 – 195; overview of 182 – 183 Woodward, James 124 – 125, 158, 169,
universal causal principle (UCP) 157, 178, 261 – 262, 288 – 291
159, 182 – 195 Worrall, John 11 – 12
would-counterfactual statement 159
vacuum solution to Einstein’s field Wright, N.T. 59
equations (EFEs) 59
vacuum states 186 – 187 Yablo, Stephen 4
Vaidman, Lev 157 Yang, C.N. 131
Vallinder, Aron 45 Yurtsever, Ulvi 127 – 128, 264 – 265, 275
van Cleve, James 226
van Fraassen, Bas C. 55 – 56, 260 Zafiratos, Chris D. 55, 130
van Inwagen, Peter 21, 24, Zanghì, Nino 156
150, 241 Zee, A. 127, 213
variables 7, 14, 124, 187, 207, 230, Zermelo-Fraenkel set theory 22
236 – 237, 288 – 291 Zynda, Lyle 44

Fundamental Causation Physics, Metaphysics, and The Deep Structure of The World

Uploaded by

Document Information

Original Title

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

Fundamental Causation Physics, Metaphysics, and The Deep Structure of The World

Uploaded by

Copyright:

Available Formats

Fundamental Causation

Fundamental Causation addresses issues in the metaphysics of deterministic

Christopher Gregory Weaver received his PhD in philosophy from Rutgers

The Future of the Philosophy of Time

Properties, Powers and Structures

The Puzzle of Existence

Neo-Aristotelian Perspectives in Metaphysics

Nominalism about Properties

Substance and the Fundamentality of the Familiar

Philosophy of Time and Perceptual Experience

For a full list of titles in this series, please visit www.routledge.com

Christopher Gregory Weaver

List of Figures viii

1 A Metaphysical Prolegomena for the Theory of

2 In Defense of the Causal Relation 89

3 The Brute Asymmetry of Causation 110

4 On the Epistemological Isolation Objection to Causal

5 Universal Causal Determination 182

6 On the Irreflexivity, Transitivity, and Well-Foundedness

8 On the Argument From Physics and General Relativity 252

9 Fundamental Causation 287

3.1 Newtonian Gravitational Force #1 119

Section 1: Avoiding Self-Stultification in Metaphysical

(Methodological Principle #1 (MP1)): Necessarily, for any cognizer C,

An actual mental state defeater for a cognizer’s belief B is a believed reason

(Directive #1 (D1)): One’s metaphysicalC system ought not preclude or

(Directive #2 (D2)): One’s metaphysicalC system ought not preclude or

(Directive #3 (D3)): One’s metaphysicalC system ought not preclude or

If the guiding principle behind D3 is correct, then one’s metaphysicalC sys-

(Directive #4 (D4)): One’s metaphysicalC system ought not preclude or

(Methodological Principle #2 (MP2)): Necessarily, for any cognizer C

(Directive #5 (D5)): One’s metaphysicalC system ought not preclude or

And the methodological principle underneath D5 is,

(Methodological Principle #3 (MP3)): Necessarily, for any cognizer C

Section 1.1: Sider’s New-Fangled Humeanism

[T]he world according to an ideologically bloated theory has a vastly

My reading is at odds with other ways Sider characterizes structure. For

[Interpretive Conjecture]: Substantive notions of the ideology of the

Sider’s new-fangled Humeanism includes a recommendation for achiev-

(1) It is not the case that there are(fundamentally) truths.30

is true only if,

But a non-trivial necessary condition for the truth of (4) is,

That is to say, γ corresponds to fundamental reality or structure. It follows

Section 1.2: Epistemic Structural Realism and Metaphysics of Science

According to ESR (quoting Worrall), “it is a mistake to think that we can

(2) (d): Brandon’s decision to ask a question about Newton’s theory of

(3) If (2), then Brandon’s decision is best categorized as theoretical stuff

ESR precludes knowledge of (at least) instances of mental causation when

(Sol. 1&2) x = ± 2 (Let (Sol. 1) include the plus/positive value.)

(Eq. 2): σ = 2 (where, again, I have in mind the principal root of 2)

(Directive #6 (D6)): One’s metaphysical worldview (and one’s metaphysicalC

Section 2.1: Against the New Verificationism

(Verificationist Principle (VP)): (ai) Theories in metaphysics should have

Beside this Verificationist Principle, RLS add the Principle of Naturalistic

(Principle of Naturalistic Closure (PNC)): One should only take seri-

Realism is incompatible with Hilary Putnam’s doctrine of conceptual rela-

descriptions which are cognitively equivalent . . . in the sense that any

(Directive #7 (D7)): One’s metaphysical worldview (and one’s meta-

Section 4: Metaphysical Theorizing

Section 4.1: Quantifiers and Logic Choice

(Principle #1): Every well-formed formula’s (wff’s) truth-value on some

Continuing to follow (Weaver forthcoming), I note how classical logic has

(Directive #8 (D8)): One’s deductive reasoning in one’s metaphysical

List of Figures viii

1 A Metaphysical Prolegomena for the Theory of

2 In Defense of the Causal Relation 89

3 The Brute Asymmetry of Causation 110

4 On the Epistemological Isolation Objection to Causal

5 Universal Causal Determination 182

6 On the Irreflexivity, Transitivity, and Well-Foundedness

8 On the Argument From Physics and General Relativity 252

9 Fundamental Causation 287

3.1 Newtonian Gravitational Force #1 119

It is a definite plus that one’s potential explanation is an explanation (gener