Jessica M Wilson - Metaphysical Emergence (2021, Oxford University Press)

OUP CORRECTED PROOF – FINAL, 2/2/2021, SPi
Metaphysical Emergence
Metaphysical Emergence
J E S SIC A M . W I L S O N
1
3
Great Clarendon Street, Oxford, OX26DP
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Jessica M. Wilson 2021
The moral rights of the author have been asserted
First Edition published in 2021
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2020943339
ISBN 978–0–19–882374–2
DOI: 10.1093/oso/9780198823742.001.0001
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
For my mother and my husband

List of Figures
2.1. Case 1 of the problem of higher-level causation: S causes S* 42

2.2. Case 2 of the problem of higher-level causation: S causes P* 43
2.3. The Strong emergentist’s response to case 2 52
2.4. The Strong emergentist’s response to case 1 53
2.5. The nonreductive physicalist’s response to case 1 70
2.6. The nonreductive physicalist’s response to case 2, version 1 70
2.7. The nonreductive physicalist’s response to case 2, version 2 71
5.1. Beginning state of the glider 187
5.2. Four-step periodic evolution of the glider 188
7.1. Two-dimensional matrix of intensions for ‘water’ 230
Acknowledgments
This book covers research carried out over the past two decades, and I
correspondingly have many people and institutions to thank. The heart of the
approach to emergence that I endorse here can be traced back to my thesis,
Physicalism, Emergentism, and Fundamental Forces (Cornell, 2002), prior to that
to my first paper, ‘How Superduper does a Physicalist Supervenience Need to
Be?’ (Philosophical Quarterly, 1999), and prior to that to John Heil’s 1996 NEH
seminar in the metaphysics of mind, which Heil graciously allowed me to attend.
It was during this seminar that Michael Watkins came up with the ‘proper subset
of powers’ strategy for addressing Jaegwon Kim’s concerns that real, distinct,
and dependent mental properties would problematically causally overdetermine
effects already produced by their physical dependence base properties. I still
vividly remember Michael arriving at a cocktail party hosted by co-participant
Amie Thomasson and announcing, “I’ve solved the problem of mental causation!”
And so he had, it seems to me, at least from a nonreductive physicalist perspective.
Much of my work on physical realization (i.e., Weak emergence) traces back to
Watkins’s core insight. In the early years developing this approach to realization,
along with a ‘new power’ approach to physically unacceptable (i.e., Strong)
emergence, I had the benefit of a dream team committee—Sydney Shoemaker,
Richard Boyd, and Jason Stanley—and of supportive interactions with many
exceptional philosophers, including Heil, Terry Horgan, and Jaegwon Kim, whose
work has been methodologically and dialectically foundational in my own.
For assistance in the ensuing years in developing this project, I am indebted,
to start, to my home departments at the University of Michigan (2001–5) and
the University of Toronto (2005–present), each of which created environments
in which my research was able to flourish. I would like to thank the National
Science Foundation for a Professional Development Fellowship providing me
with a year off to gain further education in physics, and the Social Science and
Humanities Research Council for fellowship support for the project ‘Metaphysical
Emergence: The Ontological and Causal Autonomy of Special-Science Entities’,
which resulted in a first draft of this book, and the in-progress project ‘How
Metaphysical Dependence Works’, which has seen the book’s completion. I am
also grateful to several departments which hosted me as a visiter for extended
periods, including the Centre for Consciousness, Philosophy Program, Research
School of Social Sciences, Australian National University (as Visiting Fellow), the
Logic, Language, and Cognition Research Group at the University of Barcelona
(as Visiting Scholar), the Arché Centre for Metaphysics, Language, Logic, and
viii acknowledgments
Epistemology at Saint Andrews (as Visiting Fellow), the Department of Philosophy

at the University of Cologne (as Visiting Fellow), the Eidyn Research Centre at
the University of Edinburgh (as Regular Distinguished Visiting Professor), and
the Institute for Philosophical Research at the National Autonomous University
of Mexico (as Visiting Fellow). My experiences and interactions with the many
helpful and intellectually incisive philosophers at my home and other institutions
have been of enormous benefit to the development of my research on emergence.
I have given many dozens of talks at colloquia, conferences, and workshops
on topics related, one way or other, to the present work—too many to list
here. For stimulating feedback and conversation I thank the organizers of and
audiences at these events. I am also indebted to certain individuals, in addition
to those previously mentioned, for valuable assistance in refining aspects of
this work, one way or another. These include Ken Aizawa, Torin Alter, Fatema
Amijee, Louise Antony, Joseph Baltimore, Umut Baysan, George Bealer, Mark
Bedau, Karen Bennett, Sara Bernstein, Ricki Bliss, Stephen Biggs, Alexander
Bird, Amanda Bryant, Claudio Calosi, Alex Carruth, David Chalmers, Jonas
Christensen, Lenny Clapp, Troy Cross, Louis deRosset, Ranpal Dosanjh, Douglas
Ehring, Kit Fine, Steven French, Sophie Gibb, Carl Gillett, Matthew Haug,
Robin Hendry, Eric Hiddleston, Robert Howell, Paul Humphreys, Gordon
Kane, Kathrin Koslicki, David Kovacs, Maurice Lamb, Tom Lancaster, Pat
Lewtas, Cynthia and Graham Macdonald, Cristian Mariani, Jonathan McCoy,
Michael McKenna, Brian McLaughlin, Andrew Melnyk, Sandra Mitchell, Barbara
Montero, Margie Morrison, Adam Murray, Alyssa Ney, Daniel Nolan, Paul
Noordhof, Tim O’Connor, Asya Passinsky, Laurie Paul, Derk Pereboom, Thomas
Polger, Diana Raffman, Gonzalo Rodriguez-Pereyra, Luke Roelofs, Jonathan
Schaffer, Christopher Shields, Ted Sider, Michael Silberstein, Daniel Stoljar,
Michael Strevens, Tuomas Tahko, Elanor Taylor, Mariam Thalos, Alessandro
Torza, Bob van Gulick, Jennifer Wang, Gene Witmer, Stephen Yablo, David Yates,
and Jeffrey Yoshimi. I thank each of these philosophers for their contributions
and for taking the time to engage with me and/or my work, and I apologize to
anyone I have inadvertently left off this list. I would like to extend special thanks
to Catharine Diehl, Kevin Morris, Michele Paolini Paoletti, and Bill Seager, as well
as two anonymous referees for Oxford University Press, for detailed comments
on previous drafts; this book was greatly improved by taking these comments into
consideration. Thanks as well to Peter Momtchiloff, the paradigm of a patient and
encouraging editor, and to others at OUP, including Henry Clarke and Vaishnavi
Ananthasubramanyam, who so helpfully assisted in the production of this book.
Some of the material in this book is based on previously published work.
For permission to reproduce the occasional short passage I’d like to thank Brill
(‘Metaphysical Emergence: Weak and Strong’, in Metaphysics in Contemporary
Physics), Cambridge University Press (‘Free Will and Mental Quausation’, in
Journal of the American Philosophical Association), Oxford University Press (‘Non-
reductive Physicalism and Degrees of Freedom’, in The British Journal for the
acknowledgments ix
Philosophy of Science; ‘Non-Reductive Realization and the Powers-based Subset

Strategy’, in The Monist), and Springer Publishing (‘The A Priority of Abduction’
and ‘Determination, Realization, and Mental Causation’, each in Philosophical
Studies). Full citations can be found in the bibliography.
Finally, I want to thank two members of my family, to whom this book is
dedicated. Thanks to my wonderful mother, Marilyn Wilson, for being a constant
source of love, wisdom, and inspiration. And thanks to my beloved husband, Benj
Hellie, for being the most brilliant interlocuter and partner I could ever hope for,
in philosophy and in life.
1
Key issues and questions
1.1 Metaphysical emergence: dependence with autonomy
1.1.1 Dependence with autonomy
Consider some entities (objects, systems, or other particulars) of the sort charac-
teristic of the special sciences: cells (treated by cellular biology), organs (treated by
biology), trees (treated by botany), birds (treated by zoology), lakes and mountains
(treated by geology), hurricanes (treated by meteorology), and humans (treated by
anthropology and psychology, among other disciplines), to name just a few. Such
macro-entities, the scientists tell us, are ultimately dependent on complex con-
figurations of fundamental physical entities—‘micro-configurations’, for short—
in that, at any given time, a macro-entity inherits its matter from some micro-
configuration at that time, and a macro-entity’s features (its states, properties,
behaviors, or other ways to be) over a given temporal interval are at least in part
a function of features of its underlying micro-configuration(s) over that temporal
interval. Notwithstanding this cotemporal material dependence, however, these
special-science entities also seem, from both theoretical and experiential points
of view (to be discussed in more detail shortly), to possess a certain degree of
ontological and causal autonomy—that is, they appear to be distinct from, and
distinctively efficacious with respect to, the micro-configurations upon which they
depend.
It is the coupling of cotemporal material dependence with ontological and causal
autonomy which is most basically definitive of the notion of emergence, at least
as suggested by the central cases of special-science entities with respect to the
physical micro-configurations which are their constant companions. (Note that
the dependence at issue here need not be instantaneous, but is broad in perhaps
holding over a temporal interval. Note also that the notions of autonomy at issue
here follow common use in indicating that some goings-on are distinctive in a
specific way, leaving open, as in developmental, political, and other contexts, that
autonomous goings-on are not entirely independent of other goings-on.) This
general notion of emergence is also motivated by attention to cases of artifacts,
including tables, paintings, buildings, and so on; for these too are reasonably taken
to be cotemporally materially dependent on, yet distinct from and distinctively
efficacious with respect to, underlying micro-configurations. And insofar as the
Metaphysical Emergence. Jessica M. Wilson, Oxford University Press (2021). © Jessica M. Wilson.
DOI: 10.1093/oso/9780198823742.003.0001
2 metaphysical emergence
dependence, distinctness, and distinctive efficacy seemingly at issue in all these

cases appear to be characteristic of the entities themselves, as opposed to what
we can know or represent about such entities, the notion of emergence initially
motivated by these cases is that of specifically metaphysical emergence.
As we’ll see, there are questions about how to interpret the appearances of
metaphysical emergence, and relatedly, about whether these appearances can be
taken at realistic face value. Such questions will occupy much of this book. First,
though, it’s worth saying a bit more about the prima facie motivations for there
being metaphysical emergence, to substantiate that there is good reason to attend
to the content, viability, and applicability of this interesting notion.
1.1.2 The prima facie motivations
Cotemporal material dependence

Why think that special-scientific and artifactual macro-entities are cotemporally
materially dependent on micro-configurations of fundamental physical entities—
that is, on collections (‘pluralities’) or structural aggregates of entities of the sort
treated by fundamental physics?1 The motivations here are broadly empirical, and
reflect common scientific consensus on two points. The first is that the only matter
or substance is physical matter or substance, such that any and all macro-entities
(objects, systems, or other particulars) inherit their matter from (configurations
of) their ultimately physical constituents. The second is that the features (states,
properties, behaviors, or other ways for a particular to be) of any macro-entity
over a given temporal interval are at least in part a function of the features of the
micro-configuration(s) which materially constitute the macro-entity during that
temporal interval.
The introductory section of practically any scientific textbook is likely to register
one or both of these baseline scientific assumptions, as in these examples, drawn
at random from my bookshelf:
Everything is made of atoms. That is the key hypothesis.

(Feynman 1963, Vol I, 1–9)
Quarks and leptons are the fundamental objects of which all matter is composed;
they interact via the exchange of gauge bosons. (Kane 1993, 1)
In the present state of scientific knowledge, quantum mechanics plays a fun-
damental role in the description and understanding of natural phenomena.
1 Here and throughout I follow physicists and others in being generous in my use of the expression
‘fundamental’ when speaking of physical goings-on of the sort treated by fundamental physics. Nothing
deep for present purposes turns, for example, on whether physical particles such as electrons are
nonfundamental waves in more fundamental physical fields. I say more about the operative notion
of fundamentality as relevant to the present project in §1.4.3.
key issues and questions 3
[…] when we are concerned only with macroscopic physical objects […] it is
necessary, in principle, to begin by studying the behaviour of their various
constituent atoms, ions, electrons, in order to arrive at a complete scientific
description. There are many phenomena which reveal, on a macroscopic scale,
the quantum behaviour of nature. It is in this sense that it can be said that
quantum mechanics is the basis of our present understanding of all natural
phenomena, including those traditionally treated in chemistry, biology, etc.
(Cohen-Tannoudji et al. 1977, 9)
Every day, whether we know it or not, we witness changes in matter that are a
result of the properties of the atoms and molecules composing that matter—ice
melts, iron rusts, gasoline burns, fruit ripens, water evaporates. […] If we want
to understand the substances around us, we must understand the physical and
chemical properties of the atoms and molecules that compose them—this is the
central goal of chemistry. (Tro et al. 2016, 2)
These introductory gestures leave a lot open, but the core idea that macro-entities
are cotemporally materially dependent on configurations of physical entities is
about as uncontroversial as it gets in the sciences. That said, we will see that there is
controversy about exactly how to understand the form(s) of dependence at issue
here; moreover, we will also later discuss views, such as substance dualism and
panpsychism, which reject the appearances of cotemporal material dependence,
on at least some understandings. At this point we are simply canvassing prima
facie reasons to think that there is cotemporal material dependence in the cases at
issue; and there isn’t any real question about there being an enormous amount of
empirical support for such a view.
Ontological and causal autonomy

Next, why think that special scientific and artifactual macro-entities are
autonomous, ontologically and causally, with respect to the micro-configurations
upon which they cotemporally materially depend? Three prima facie motivations
for one or both forms of autonomy have to do with how special-science entities
are classified or characterized, as follows:
• Distinctive special-scientific taxonomy: Special-science entities are classified

as falling under distinctive types. For example, gases, cells, hurricanes, trees,
birds, and humans are classified as falling, respectively, under distinctive
types in, respectively, thermodynamics, biology, meteorology, botany, zool-
ogy, and anthropology (as well as psychology and other disciplines); and
similarly for countless other macro-entities treated by these and other special
sciences. On the face of it, such special-science types are different from
those under which (configurations of) fundamental physical entities fall.
Classification practices thus provide prima facie support for thinking that
special-science entities are ontologically autonomous with respect to—that
is, distinct from—their underlying micro-configurations.
• Distinctive special-science features: Special-science entities are characterized
as having distinctive features, constitutive of the distinctive types under
which they fall. A tree, for example, has roots, a trunk, branches, stems,
leaves; it obtains nutrients from air, sun, soil, and water through leaves and
roots; it reproduces via seeds and may bear fruit; it is deciduous or evergreen;
it is hardy in certain climate zones, and so on. On the face of it, such
features are not appropriately attributed to even complex configurations of
fundamental physical entities; and the same is true for the characteristic fea-
tures of other special-science entities. The mental features of human persons
are especially distinctive, as involving, among other salient characteristics,
qualitative aspects of sensory experience; representational aspects of belief,
desire, intention, and other mental states; appreciation of aesthetic, moral,
and other broadly normative values; and seemingly free agency. By Leibniz’s
Law, of course, entities with different features (at the same time, at least)
are distinct; hence that special-science entities are characterized as having
features different from those had by the micro-configurations upon which
they cotemporally materially depend again supports special-science entities’
being ontologically autonomous with respect to these micro-configurations.
• Distinctive special-science laws: Special-science entities are taken to be gov-
erned by special-science laws (or regularities) describing states, properties,
and behaviors of, including associated causal interactions involving, such
entities—laws that, on the face of it, are different from those governing even
complex configurations of fundamental physical entities. To be sure, there
are deep questions here (which will be a primary focus in what follows) about
how to understand the distinctive efficacy of special-science goings-on, and
relatedly, about whether such efficacy is compatible with the assumed efficacy
of micro-configurations and constituent fundamental physical entities.2 At
present, we can say this much: that special-science laws are seemingly distinc-
tive and seemingly causal provides prima facie support for special-science
entities’ being causally autonomous—that is, distinctively efficacious—with
respect to their underlying micro-configurations.
Two other reasons for thinking that special-science entities are autonomous
reflect, more specifically, that special-science entities, features, and laws often
2 It is sometimes claimed, typically by appeal to the discussion in Russell 1912, that physics has
dispensed with the notion of ‘cause’, but these complaints, old and new, are directed at specific
and sometimes implausible conceptions of causation, leaving open the possibility of broader and
empirically informed conceptions of physical causation. I revisit the issue of physical causality in §1.4.4.
abstract from details relevant to characterizing physical (more generally: depen-

dence base) entities, features, and laws:
• Universal properties and behavior: Many special-science entities, including

thermodynamic complex systems such as liquids and gases, exhibit features
that are functionally independent of various features of their underlying
micro-configurations. For example, complex systems near critical points
exhibit ‘universal’ properties and behavior across widely diverse underlying
micro-configurations. Such cases provide support for thinking that some
special-science entities are causally, hence ontologically, autonomous, in that
their behaviors are sensitive to comparatively abstract causal joints or levels
of grain.
• Elimination of microphysical ‘degrees of freedom’: Degrees of freedom are
independent parameters needed to specify states relevant to an entity’s law-
governed properties and behaviors. In some cases, certain such states of
special-science entities (e.g., the configuration state, encoding an object’s or
system’s spatial location) are specifiable by reference to strictly fewer degrees
of freedom than are needed to specify the corresponding (e.g., configuration)
state of the micro-configurations upon which the special-science entities
cotemporally materially depend. For example, rigid bodies cotemporally
materially depend upon quantum-mechanical micro-configurations, but the
behavior of rigid bodies does not generally rely upon quantum-mechanical
degrees of freedom such as spin direction and magnitude, as is reflected in
the fact that the laws governing rigid bodies do not contain reference to such
degrees of freedom (or their values). That certain law-governed features of
rigid bodies can be specified using strictly fewer degrees of freedom than
are required to specify the associated law-governed features of their underly-
ing quantum-mechanical micro-configurations suggests that the former are
ontologically autonomous with respect to the latter.
Four other prima facie motivations for autonomy apply to artifacts of the sort avail-
able to ordinary experience as well as certain special-scientific entities, and reflect
perceptual, individuative, and semantic considerations, in ways that support the
scientific characterizations of such entities as having features which abstract from
certain details of the micro-configurations upon which they depend:
• Perceptual unity: Though the macro-entities of our acquaintance are, sci-

entists tell us, materially constituted by massively complex and constantly
changing micro-configurations, macro-entities do not perceptually appear
to us as massively complex, constantly changing, configurations of micro-
entities. A tree, for example, does not perceptually appear as a complicated
structure of cells or tissues, much less as a buzzing array of subatomic
particles or other physical fundamenta; rather, a tree perceptually appears as a

comparatively stable and unified entity, both at and over time; and the same
is true for statues and other familiar macro-entities. Moreover, in the case
of human persons there is the evidence of introspection—a kind of internal
perception—of ourselves as fairly unified and persisting ‘selves’.
• Compositional flexibility: The identity of many macro-entities appears to
transcend that of their underlying micro-configurations, in the sense that
the existence of a given special-science entity (a hurricane, a cell, a tree,
a bird, a human) does not depend on the existence of any specific micro-
configuration(s), and relatedly, in that many macro-entities, both in scien-
tific practice and in our ordinary practices of individuation, are taken to
be capable of surviving at least some changes in their underlying micro-
configuration(s)—as when, e.g., the Venus de Milo survived the loss of her
arms.
• Proper names and definite descriptions: Our practices of giving names or
definite descriptions to certain artifacts and special-science entities—the
Mona Lisa, the CN Tower, the Atlantic Ocean, Hurricane Katrina, Benj
Hellie—suggests that we often treat macro-entities as individuals, distinct
from the ever-changing micro-configurations upon which they cotemporally
materially depend.
• Truth and meaning: Many of the sentences that we take to be true appear to be
about macroscopic goings-on, and relatedly, to contain subject terms which
appear to denote macro-entities (as either token individuals or types) and
predicate terms which appear to be used to attribute macro-features to these
macro-entities—as with, e.g., ‘That table is well-made’, ‘Tigers typically have
tails’, and ‘Her novels are spellbinding’.
One final motivation for ontological and causal autonomy is worth mentioning, as
especially dear to our hearts and minds:
• Seemingly free will: The status of our actions as genuinely free remains
up for debate. However, a starting point in this debate is that it at least
introspectively seems as if creatures like ourselves are capable of making free
choices to produce (or intend to produce) certain effects, where this efficacy
appears to be quite different from that associated with the (deterministically
or indeterministically) lawfully governed micro-configurations upon which
we and our mental states cotemporally materially depend.
Summing up: many considerations, drawn from science, perception, language,

our practices of individuation, and introspective experience, provide prima facie
support for thinking that many broadly natural entities are cotemporally materially
dependent on micro-configurations of fundamental physical entities, yet are also
ontologically and causally autonomous with respect to these underlying micro-

configurations. And that is just to say that, on the face of it, there is metaphysical
emergence.
1.2 Two key questions
Given that the prima facie motivations for metaphysical emergence span such a
wide range of phenomena, the interest in exploring and illuminating this notion
is clear. Towards this end, two questions (and related sub-questions) are key.
The first pertains to the nature and varieties of metaphysical emergence. Here
we ask: just what is metaphysical emergence, more precisely? How is it, exactly,
that special scientific and artifactual entities can cotemporally materially depend
on complex configurations of fundamental physical entities, while retaining some
degree of ontological and causal autonomy? And is there more than one way in
which this can be—is there more than one form of metaphysical emergence?
The second pertains to whether there actually is any metaphysical emergence.
To start: are there any insuperable problems with the notion(s) of metaphysical
emergence, such that emergence is, at best, an epistemic or representational phe-
nomenon? And supposing that one or more varieties of metaphysical emergence is
in-principle viable, are there any actual cases of such emergence? If we can make
sense of metaphysical emergence, both in principle and as actually instantiated,
then we would be on track to vindicating and illuminating the appearances of
emergence in science and ordinary experience. But if we cannot make sense of
emergence as viable or as actual, that would still be worth knowing. Either way,
the result promises to have widespread ramifications for our understanding of the
world around us.
There is good reason, then, to pursue answers to the key questions of what,
more specifically, metaphysical emergence is, and whether there actually is any.
Indeed, in past decades there has been an explosion of interest in emergence, in
both philosophical and scientific contexts. Unfortunately, as I’ll now discuss, all
this attention has left the answers to the key questions less clear than ever.
1.2.1 What, more specifically, is metaphysical emergence?
One source of unclarity as regards the question ‘What is metaphysical emergence?’

stems from the availability of multiple interpretations of the notions of dependence
and autonomy at issue, which interpretive options have given rise to a huge
diversity of accounts offered as providing comparatively specific answers to this
question. Before surveying this diversity, however, it’s worth first registering
certain ways in which accounts of metaphysical emergence directed at the target

cases typically agree, since these commonly held assumptions will be largely taken
for granted in what follows.3
The main points of agreement are as follows:
• Accounts of metaphysical emergence typically agree in taking the only sub-

stance to be material or physical substance, and relatedly, in taking emergence
not to involve any new substance, of the sort posited, e.g., by Cartesian
dualists or vitalists.⁴ As Stephan (2002) puts it:
The first feature of contemporary theories of emergence, the thesis of
physical monism, is a thesis about the nature of systems that have emergent
properties (or structures). The thesis says that the bearers of emergent
properties are made up of material parts only. It denies that there are
any supernatural components responsible for a system’s having emergent
properties. Thus, all substance-dualistic positions are rejected […]. (79)
Hence, for example, Mill (1843/1973), the father of British Emergentism,
supposes that emergent entities are entirely materially composed:
All organised [living] bodies are composed of parts similar to those com-
posing inorganic nature, and which have even themselves existed in an inor-
ganic state; but the phenomena of life which result from the juxtaposition of
those parts in a certain manner bear no analogy to any of the effects which
would be produced by the action of the component substances considered
as mere physical agents. (243)
3 As regards the identification of these common components, it is worth appreciating that not every
account or treatment of so-called ‘emergence’ is relevant to investigations into emergence of the sort
motivated by the target cases of special-scientific and artifactual entities and features. For example, the
diachronic conception of emergence at issue in Morgan’s (1923) notion of ‘emergent evolution’, whereby
emergence reflects increasingly complex combinations of lower-level constituents, is not relevant to
making sense of emergence in the target cases. Perhaps relatedly, nor are ‘part/whole’ conceptions on
which it suffices for emergence that a whole has a feature not had by any individual part (see §1.4.2 for
further discussion).
Also not to the point of accommodation of the target cases is the sort of intra-level diachronic
emergence advanced in Humphreys (1997), involving the exhaustive (non-mereological) ‘fusion’ of
multiple physical entities into another physical entity (see also the ‘horizontal’ account of emergence in
Santos 2015 and accounts of ‘transformational emergence’ in Humphreys 2016 and Guay and Sartenaer
2016). More generally, I am inclined to think that such cases of diachronic intra-level ‘emergence’ can
be understood simply as involving causation (see Wilson in progressa); hence I do not follow Paolini
Paoletti (2018b) in taking it to be an advantage of an account of emergence that it be general enough
to accommodate such cases.
⁴ As in Stephan’s remarks to follow, ‘material’ and ‘physical’ are sometimes treated as synonyms
in the contemporary literature, at least so far as characterizing substance is concerned. I discuss the
operative notion of the physical in §1.4.1; for present purposes the supposition of physical monism can
be understood as the view that there is only one kind of substance, of the sort that was traditionally
characterized as ‘matter’ or ‘material’ substance.
Moreover, on contemporary accounts of emergence, it is also typically sup-

posed that what it is for the entities constituting the ‘stuff ’ of natural reality
to be physical is for them to be among the target subject matter of funda-
mental physics.⁵ In what follows, I will also suppose that all substance is
physical substance, and that what this comes to is that the compositionally
basic entities are physical.⁶ That said, accounts of emergence typically admit
of generalization beyond the specific commitment to material or physical
substance monism, in the sense that they could be applied to other forms
of substance monism.
• Accounts of emergence typically agree that core to the notion is the combi-
nation of cotemporal material dependence and (some degree of) autonomy.
These core components are sometimes explicitly flagged (see Bedau 1997),
and are sometimes implicitly encoded in specific accounts of the dependence
and autonomy at issue, as when Kim (2006) says
[T]wo […] necessary components of any concept of emergence that is true
to its historical origins […] are supervenience and irreducibility. (548)
As previously noted, in accounts of metaphysical emergence the intended
notion of cotemporal dependence is broad, in applying either at a time or
over a temporal interval—i.e., is one allowing for temporally extended base
and dependent phenomena. And in these accounts the intended notion of
autonomy is one entailing not just that emergents are distinct (ontologically
autonomous) from their dependence bases, but also that emergents are dis-
tinctively efficacious (causally autonomous) with respect to these bases. To be
sure, even as regards certain of these components there is some dispute; such
variations, however, are either subsumable under the core understandings
or else not to the point of accommodating the target phenomena involving
special-scientific or artifactual entities (or features) and their material bases.⁷
⁵ As Paolini Paoletti (2018b) puts it, “Microphysical entities are here taken to be all the relatively
compositionally basic entities in the universe, as they are studied (or they can be studied) by
physics” (4).
⁶ What if there are no compositionally basic entities? I’ll address this issue in §1.4.1, when I discuss
the operative notion of the physical.
⁷ Again, not every account of so-called ‘emergence’ is directed at the target cases of special-scientific
and artifactual entities and features; the accounts discussed in this note, however, are directed at these
cases.
In re purportedly diachronic emergence: Mill (1843/1973) suggests that certain (“heteropathic”)
effects diachronically emerge from temporally prior causes, but as he observes, one can understand
such emergence as involving the cotemporal emergence of powers of a composite entity to produce
such effects. A similar treatment accommodates the view of O’Connor and Wong (2005), according to
which emergent features are diachronically caused by dependence base features. More generally, since
whether an entity causes a given effect typically requires that certain circumstances be in place, it is
preferable to treat ‘causal’ accounts of emergence in terms of the emergence of powers of a composite;
• Accounts of metaphysical emergence typically agree that the emergence of

entities can be investigated by attention to the emergence of features of
the entities at issue. On this understanding, any emergence there might be
involves an emergent feature. As Bedau (2002) puts it:
[A]n entity with an emergent property is an emergent entity and an emer-
gent phenomenon involves an emergent entity possessing an emergent
property—and they all can be traced back to the notion of an emergent
property. (6)
Here the supposition is that if some entity is emergent, this is because it has
some characteristic feature which is emergent—e.g., the feature (characteris-
tic of a liquid or gas near a critical point) of undergoing a change of phase, or
the feature (of a human or other minded creature) of being in pain—which
can be the direct target of investigation. Relatedly, a focus on features is useful
in that talk of entities can be naturally translated into talk of features—namely,
the property or feature of being an entity of the type in question.⁸
I too will usually (though not always) focus on the emergence of features as
offering a comparatively straightforward and systematic way of investigating
the emergence of associated entities.⁹ Three clarificatory points in re this
focus are in order. First, insofar as emergence involves cotemporal material
dependence, which in the first instance applies to entities, we need to extend
the operative understanding of cotemporal material dependence to apply to
features, as follows:
for otherwise the emergence of the composite would implausibly depend on whether it happened to
occupy circumstances enabling it to enter into the production of the effect in question.
Another supposed motivation for a diachronic notion of emergence is that emergence can be
temporally extended, as with, e.g., Rueger’s (2000) account of diachronic emergence as involving
temporally extended processes, and Mitchell’s (2012) account of emergence as involving ‘dynamic
self-organization’ (see also Kirchhoff 2014). But again, the emergence of such processes is compatible
with these cotemporally depending on a temporally extended base; compare Lepore and Loewer’s
(1989) discussion of how, even if externally individuated content properties do not supervene on
spatiotemporally local (i.e., neurophysiological) properties, “there may be a more global physical
property whose instantiation explains the possession of content properties” (181).
In re purportedly epiphenomenal emergence: Chalmers (1996) and Morris (2014) characterize
an epiphenomenalist form of emergence, on which emergents are ontologically but not causally
autonomous. However, such emergence is beside the point of accommodating the emergence of
special-science entities as entering into distinctive, typically causal, laws, and so (modulo discussion in
Ch. 4, §4.4.1) will be put aside here.
⁸ A third reason to focus on the emergence of features is that, and notwithstanding that the initial
motivations for metaphysical emergence are naturally expressed as involving the emergence of entities,
one might want to leave open the possibility that one feature of an entity might emerge from another
feature of that same entity. In Ch. 4 (§4.2.3) I will revisit the question of whether the emergence of a
feature always brings the emergence of an entity in its wake, and if so, whether this is in tension with
the assumption of physical monism.
⁹ Nominalists who reject the existence of properties or other features are invited to apply their
preferred strategies for converting talk of features into talk of objects or other entities in what follows.
What it is for a feature S to cotemporally materially depend on a feature P

is for the entity bearing S to cotemporally materially depend on the entity
bearing P.1⁰
Second, sometimes the emergence of the type of feature (entity) will be at
issue, where a type of feature may have many instances or tokens; other times
the emergence of a given token of a feature (entity) will be at issue. When
the difference between type and token matters, this will be made clear. Third,
while most accounts of metaphysical emergence take this to involve a one-
one relation between base and emergent features, some accounts characterize
the relation between features as a many-one relation, with a given feature
emerging from a plurality of features or feature instances. I am inclined
to think that many-one approaches to metaphysical emergence can either
be subsumed under a one-one approach or else are not to the point of
characterizing emergence understood as an inter-level relation,11 but I will
track the difference where it matters.
• Accounts of metaphysical emergence typically agree that emergence has cer-
tain correlational or modal implications, sometimes expressed by saying that
emergent features ‘supervene’ on base features with at least ‘nomological’—
i.e., natural law-based—necessity. As I’ll sometimes put it: emergent features
‘minimally nomologically supervene’ on base features.12 Here the idea is
that in every world (actual or hypothetical) with the same or relevantly
similar laws of nature, the occurrence of an emergent feature S requires the
occurrence of some or other base feature P, and in every such world, the
occurrence of any such P will be accompanied by the occurrence of such an S.
For example, Broad (1925) maintains that emergent features of a compound
are “completely determined” by features of its parts when appropriately
configured, in that “whenever you have a whole composed of these […]
1⁰ If one wants to make room for the possibility that a feature of an entity can emerge from that same
entity (as per note 8), one should allow here that an entity can cotemporally materially depend on itself.
11 I’ll revisit this issue in Ch. 3 (§3.4.2), when discussing the ‘many-one’ approach endorsed in
Gillett 2010, 2002a, and 2016. On Gillett’s approach, emergence involves the cotemporal material
dependence of a single (e.g., special-scientific) feature on multiple, comparatively nonrelational (e.g.,
physical) features, where the latter are understood as combining in various (e.g., lower-level causal,
spatiotemporal, or mereological) ways agreed by all parties not to involve emergence, whereas the
one-one approach takes emergence to involve the cotemporal material dependence of a single (e.g.,
special-scientific) feature on a single (e.g., physical) feature of a micro-configuration (plurality or
structural aggregate), understood as already combined in various ways agreed by all parties not to
involve emergence. So understood, the many-one approach is subsumable under the one-one approach;
see related discussion in Bennett 2017, 10. One reason to stick with a one-one approach is that intra-
level causation may also take the form of a many-one relation; but such causation is beside the point
of emergence as an inter-level relation.
12 See Kim 1990 and McLaughlin and Bennett 2018 for discussion of the notion and varieties of
supervenience.
elements in certain proportions and relations you have something with the
[compound’s] characteristic properties” (64).
As we’ll see, some accounts of metaphysical emergence take the correla-
tions to hold with just nomological necessity, whereas others take them to
moreover hold with metaphysical necessity (that is, in every possible world—
not just worlds with the same or relevantly similar laws of nature). Either
way, these accounts all agree that emergent features minimally nomologically
supervene on base features.
• Finally, accounts of metaphysical emergence typically agree that natural real-
ity exhibits a kind of ‘leveled’ or ‘layered’ structure, reflected in the broadly
hierarchical structure of the special sciences. Certain views contrasting with
any form of emergentism—e.g., vitalism or substance dualism—also take
natural reality to have a leveled structure, so the supposition that there
are multiple levels is not equivalent to the supposition that there is emer-
gence: accounts of metaphysical emergence differ from these other views
in aiming to make sense of natural reality’s having a leveled structure in a
way compatible with the appearances of cotemporal material dependence
coupled with autonomy. That qualification registered, emergent entities and
features are often characterized as ‘higher-level’ with respect to the ‘lower-
level’, and ultimately physical, goings-on upon which they depend. Talk of
levels is convenient, especially as making room for there being different yet
concurrently existing systems of entities, features, and laws of the sort that are
associated with the special sciences or different scales of experience or reality,
and in what follows this terminology will frequently make an appearance (as
in, e.g., discussion of Kim’s ‘problem of higher-level causation’).
There remains the independent and important question of how to indi-
viduate levels in a way that does not prejudice the question of whether
there is genuine metaphysical emergence against the ontological or meta-
physical reductionist. The reductionist maintains, contra any metaphysical
emergentist, that all goings-on, including apparently ‘higher-level’ goings-on,
are in fact identical to some or other complex combination of ‘lower-level’,
ultimately physical goings-on. Correspondingly, the operative conception of
levels needs to be broad enough to not immediately rule out of court the
reductionist view that there is only one level, such that, e.g., the level of
physical goings-on that all parties agree (in the present debate) contains the
compositionally basic entities needs to include not just, e.g., physical particles
such as quarks and electrons standing in relatively noncomplex physical
relations, but also massively complex pluralities or structural aggregates of
such particles and relations (where the relations at issue include lower-level
causal, spatiotemporal, mereological, and other ontologically ‘lightweight’
modes of combination), of the sort that might potentially serve as the
candidate physicalist reduction base for macro-entities such as tables, planets,

and persons, and their characteristic features. I’ll return to the question of
how to individuate levels in §1.4.2. Meanwhile, one should keep in mind that
the notion of a ‘level’ in what follows must be suitably expansive if substantive
debate between the reductionist and emergentist is to proceed.13
So there are important core points of agreement among accounts of meta-

physical emergence. Beyond these core points of agreement, however, accounts
of such emergence diverge into a bewildering variety, primarily reflecting that
the core notions of dependence and autonomy have multiple, often incompatible
interpretations. The extent of this diversity has led some to claim that references to
emergence “seem to have no settled meaning” (Byrne 1994, 206); that accounts of
emergence are “not obviously reconcilable with one another” (O’Connor 1994, 91);
that there has been “a historical load of confusion surrounding the metaphysical
aspects of the concept, reflected in the fact that it has been used in a long series
of different ways” (Emmeche et al. 1997, 84); that “those discussing emergence,
even face to face, more often than not talk past each other” (Kim 2006, 548); that
“‘emergent’ and all its semantic kin have come to stand for a hopeless jumble of
different ideas” (Ladyman and Ross 2007, 193); and that “within philosophy and
the sciences the term ‘emergence’ is used in such a bewildering variety of ways
that it seems the word itself is the only thing shared across these various usages”
(Silberstein 2009, 254).
To be sure, some of the confusing diversity to which these philosophers are
adverting reflects that some accounts of emergence take this to be a merely epis-
temic or representational phenomenon. On these other approaches, the seeming
autonomy of emergent phenomena is understood not in metaphysical terms, but
rather in terms of such phenomena’s being, e.g., unpredictable or underivable from
lower-level theories, and where the failures of predictability or derivability are not
taken to have any clear metaphysical consequences for whether there are distinct
and distinctively efficacious higher-level entities.1⁴ But even restricting ourselves
13 It may also be worth registering that there need be no supposition that natural (or artifactual)
reality neatly divides into anything resembling a layer cake. The complete characterization of crea-
tures like us, for example, would presumably draw upon goings-on at multiple (physical, chemical,
biological, psychological, ecological) levels. This complexity provides another reason why a focus on
the emergence of features rather than entities may in some cases be more perspicuous.
1⁴ Discussions of emergence in scientific contexts are commonly cashed in (purely) epistemic
terms. Hence, for example, Anderson (1972) qualifies the metaphysical import of his discussion of
the unpredictability-based emergence of certain features of an ammonia molecule by saying that “we
must all start with reductionism, which I fully accept” (22). See also, e.g., the overview of scientific
accounts of emergence in terms of unpredictability or surprise in Damper 2000. Such purely epistemic
accounts are for the most part irrelevant to the present project.
to accounts of emergence intended to have such metaphysical consequences, there

remains a bewildering variety of options.1⁵
Candidate accounts of the dependence at issue in metaphysical emergence
include mereological (‘part/whole’) determination,1⁶ causation or nomological
connection,1⁷ functional realization,1⁸ constitutive mechanism,1⁹ the deter-
minable/determinate relation,2⁰ inheritance of causal powers,21 and primitive
‘Grounding’.22 Candidate accounts of the ontological and/or causal autonomy at
issue are even more various. Explicitly metaphysical accounts of such autonomy
include nomological but not metaphysical supervenience,23 nonfundamental
novelty (of features, powers, laws, entities),2⁴ fundamental novelty (of features,
powers, forces/interactions, laws, entities),2⁵ nonadditivity/nonlinearity,2⁶
‘downward’ causal efficacy,2⁷ multiple realizability/universality/compositional
plasticity,2⁸ causal proportionality/difference-making/counterfactual considera-
tions,2⁹ elimination in degrees of freedom,3⁰ sometimes associated with symmetry
breaking,31 and the holding of a proper subset relation between token powers,32
sometimes cashed in terms of a proper parthood relation between properties or
behaviors.33 And ‘epistemic criteria’ accounts of ontological and/or causal auton-
omy include in-principle failure of deducibility/predictability/explicability,3⁴
predictability, but only by simulation,3⁵ lack of conceptual or representational
entailment,3⁶ and theoretical/mathematical singularities.3⁷ Given this plethora
1⁵ The lists and citations to follow are representative rather than exhaustive; there are many
hundreds of papers and books on these notions and their variations, as entering into accounts of either
physically unacceptable emergence (a.k.a. ‘strong’ emergence) or physically acceptable emergence
(a.k.a. ‘realization’ or ‘weak’ emergence).
1⁶ See Stephan 2002, Gillett 2002a. 1⁷ See Searle 1992, O’Connor and Wong 2005.
1⁸ See Putnam 1967, Boyd 1980, Poland 1994, Antony and Levine 1997, Melnyk 2003, Yates 2012.
1⁹ See Craver 2001, Haug 2010, Gillett 2016.
2⁰ See Macdonald and Macdonald 1986, Yablo 1992, Ehring 1996, Wilson 2009.
21 See Kim 1992a, Wilson 1999 and 2015b, Shoemaker 2000/2001.
22 See Schaffer 2009, Dasgupta 2014.
23 See van Cleve 1990, Chalmers 1999, Seager 1999/2016, Noordhof 2010.
2⁴ See Humphreys 1996, Wimsatt 1996, Crane 2001, Pereboom 2002, Megill 2013.
2⁵ See the British Emergentists (e.g., Mill 1843/1973, Alexander 1920, Broad 1925), Kim 1992a,
Cunningham 2001, O’Connor 1994, Wilson 2002a and 2015b, Barnes 2012, Paolini Paoletti 2017.
2⁶ See the British Emergentists, Newman 1996, Bedau 1997, Silberstein and McGeever 1999,
Mitchell 2012.
2⁷ See Morgan 1923, Sperry 1986, Klee 1984, Thompson and Varela 2001, Searle 1992, Schroder
1998, Stephan 2002.
2⁸ See Putnam 1967, Fodor 1974, Boyd 1980, Klee 1984, Lepore and Loewer 1989, Wimsatt 1996,
Antony and Levine 1997, Batterman 1998, Aizawa and Gillett 2009, Morrison 2012.
2⁹ See Yablo 1992, Lepore and Loewer 1987 and 1989, Bennett 2003, List and Menzies 2009.
3⁰ See Batterman 1998, Wilson 2010b, Lamb 2015. 31 See Morrison 2012.
32 See Wilson 1999. 33 See Shoemaker 2000/2001, Clapp 2001, Rueger and McGivern 2010.
3⁴ See Broad 1925, Hempel and Oppenheim 1948, Klee 1984, Lepore and Loewer 1989.
3⁵ See Newman 1996, Bedau 1997. 3⁶ See Chalmers 1996, Van Gulick 2001.
3⁷ See Batterman 2002.
of options, it’s no surprise that many discussions of specifically metaphysical

emergence aim primarily to taxonomize its varieties.3⁸
Now, as I have previously observed (Wilson 2015b), though in general a thou-
sand flowers may fruitfully bloom, this much diversity is unhelpful as regards
answering the first key question, concerning the nature and varieties of specif-
ically metaphysical emergence. It would be one thing if different accounts of
metaphysical emergence targeted different phenomena. But different accounts
often target the same phenomena, while disagreeing about whether these are
metaphysically emergent. And when accounts do agree that a given phenomenon
is emergent, there is often no clear basis for the agreement—or, relatedly, for
determining whether the emergence at issue is compatible with physicalism, the
view that all broadly scientific goings-on are, to speak schematically, ‘constituted
by’, ‘grounded in’, or ‘completely metaphysically dependent on’ physical goings-on.
More generally, the extent of variability in both content and application here might
well lead one to suppose that accounts of specifically metaphysical emergence, like
accounts of emergence generally, have (beyond the comparatively abstract core
components) nothing systematic in common.
1.2.2 Is there actually any metaphysical emergence?
The answer to the second key question, of whether there actually is any emergence
of a metaphysical variety, has also remained unclear, in large part owing to still-
live concerns about whether the appearances of such emergence (or of associated
levels of natural reality) are genuine. Among these concerns are that metaphysical
emergence is naturalistically unacceptable; that considerations of parsimony push
against taking the appearances of metaphysical emergence ontologically seriously;
that the notion of metaphysical emergence is either trivially fulfilled or trivially
never fulfilled; and that metaphysically emergent entities or features, were they to
exist, would give rise to problematic causal overdetermination of effects already
produced by dependence base entities or features.
Here the diversity of accounts of emergence again muddies the waters; for
while some accounts have resources to respond to some of these concerns, the
absence of any systematic treatment of the notion of metaphysical emergence
renders it unclear whether the notion can survive all the various attacks. And to the
extent that the in-principle viability of metaphysical emergence remains unclear,
3⁸ See Klee 1984, Van Gulick 2001, Stephan 2002, Gillett 2002b, O’Connor and Wong 2015. See
also the introduction to Gibb et al. 2018 for discussion of the continuing diversity of approaches to
emergence.
the further project of determining whether there actually is any such emergence
cannot even get off the ground.
1.2.3 My aim: to provide clear, compelling, systematic

answers to the key questions
The point and purpose of this book is to provide clear, compelling, and systematic
answers to the two key questions of what, more precisely, metaphysical emergence
is, and whether there actually is any such emergence.
In response to the first key question, I will argue that for the sort of target
cases motivating the notion of metaphysical emergence, there are two and only
two schemas for—schematic characterizations of—metaphysical emergence, one
of which is compatible with physicalism (on the assumption that the dependence
base goings-on are physical), and the other of which is not so compatible (on
that assumption). And I will show that a wide range of existing accounts of
metaphysical emergence plausibly aim to instantiate one or the other schema, such
that much of the apparent diversity of these accounts is superficial.
In response to the second key question, I will first argue that each of these
two forms of metaphysical emergence is viable—coherent, metaphysically sub-
stantive, naturalistically acceptable, such as to avoid problematic causal overde-
termination, and more generally such as to vindicate and illuminate the prima
facie scientific and ordinary experiential motivations for thinking that there is
metaphysical emergence. I will go on to consider, for a variety of interesting
actual phenomena, whether these phenomena are metaphysically emergent in
one or the other of these two ways; I will argue that one form of metaphysical
emergence (the sort compatible with physicalism) is actually quite common, and
the other (the sort incompatible with physicalism) remains, for some cases, an
open and in-principle empirically verifiable possibility, and in one special case—
arguably the most important case, for creatures like ourselves—is plausibly actually
instantiated.
1.3 Outline of the book
The plan for carrying out this project is as follows.

In Chapter 2 (‘Two schemas for metaphysical emergence’), I present what is seen
by many as the most pressing challenge to taking the appearances of metaphysical
emergence as genuine—namely, the problem of higher-level causation, made
salient by Jaegwon Kim in his 1989, 1993a, 1998, and elsewhere. Kim’s general
concern is that any purported effects of higher-level features are already produced
by the lower-level features upon which they minimally nomologically supervene,
such that the metaphysical emergentist is committed to such effects’ being prob-
lematically causally overdetermined—that is, problematically caused twice over.
I argue, following discussions in Wilson 1999, 2001, 2011b, and elsewhere, that
there are two and only two strategies of response to this problem that make sense
of seemingly higher-level entities and features’ being metaphysically emergent—
that is, as being cotemporally materially dependent on yet also ontologically and
causally autonomous with respect to dependence base (in particular, physical)
entities and features. One strategy provides a schematic basis for ‘Weak’ (physically
acceptable) emergence; the other provides a schematic basis for ‘Strong’ (physically
unacceptable) emergence. And for each of these strategies and associated schemas,
I show that a representative range of seemingly diverse accounts of emergence are
plausibly seen as satisfying the conditions in one or the other schema, and thus are
more unified than they appear.
Since the schemas play a large and structuring role in this book, it is worth
prefiguring their content and the associated strategies for avoiding problematic
causal overdetermination. On the view to be advanced and defended in what
follows, a Strongly emergent feature has, on a given occasion, at least one token
power not had by the base feature upon which it cotemporally materially depends
on that occasion; overdetermination is avoided by denying that the base feature
produces the effect (or, more weakly, by denying that the base feature produces
the effect in the same way as the higher-level feature). Strong emergence is of the
anti-mechanistic or anti-physicalist variety associated with British Emergentism,
according to which, at certain levels of compositional complexity, fundamentally
novel features and associated powers or laws come to exist (be instantiated, obtain).
By way of contrast, a Weakly emergent feature has, on a given occasion, a proper
subset of the token powers had by the base feature upon which it cotempo-
rally materially depends on that occasion; problematic overdetermination is thus
avoided insofar as every token power of the higher-level feature is identical to
a token power of its base feature, while the distinctive efficacy of the higher-
level feature is preserved as a result of its having a distinctive power profile.
Weak emergence is the sort associated with nonreductive physicalism, according
to which some higher-level features are, while completely metaphysically depen-
dent on complex configurations of ultimately physical goings-on, nonetheless
distinct from and distinctively efficacious with respect to the latter. For purposes
of appreciating the generality of the schemes, it is worth registering that the notion
of ‘power’ here is metaphysically highly neutral, reflecting commitment just to the
plausible thesis that the causes an entity may potentially bring about are associated
(perhaps only contingently) with how the entity is—that is, with its features. More
generally, no controversial theses pertaining to the nature of powers, properties,

causation, or laws are presupposed.3⁹
The results of Ch. 2 establish that we have prima facie reason to think that
satisfaction of the conditions in the schemas for Weak and Strong emergence is,
as I put it, “core and crucial” to metaphysical emergence of physically acceptable
and physically unacceptable varieties, respectively. I prefer this terminology to the
usual though to my mind overly coarse-grained terms of necessary and sufficient
conditions, since any schematic account needs to be sensibly filled in. But modulo
this caveat, the results of this chapter can also be seen as providing prima facie
reason to think that the conditions in the schemas are, when sensibly filled in, both
necessary and sufficient for metaphysical emergence of both physically acceptable
and physically unacceptable varieties—a bold claim, but one that, as I argue in
ensuing chapters, is surprisingly robust.
In Chapter 3 (‘The viability of Weak emergence’) I consider and respond to
a representative range of objections to the schema for Weak emergence and the
associated ‘proper subset of powers’ approach to realization presented and devel-
oped in Wilson 1999, 2009, 2010b, 2011b, and 2015b, among other venues.⁴⁰ These
objections include that satisfaction of the conditions in the schema is compatible
with anti-realism or reductionism about the purportedly emergent features, is
compatible with the physical unacceptability of the emergent features, and is not
necessary for physically acceptable emergence. These diverse challenges can, I
argue, be answered. As we’ll see, each challenge admits of one or more responses
that are generally available on any sensible implementation of the schema for Weak
emergence. Upon occasion, additional responses draw on features of my preferred
accounts of Weak emergence—one appealing to the determinable/determinate
relation (as per my 1999 and 2009, developing the proposals in Macdonald and
Macdonald 1986 and Yablo 1992), according to which Weakly emergent features
are determinables of lower-level realizers, and another appealing to an account of
Weak emergence as involving an elimination in degrees of freedom (as per my
2010b, developing the proposal in Batterman 1998 and elsewhere), according to
which (roughly speakly) at least one state of a Weakly emergent entity can be
specified using strictly fewer degrees of freedom (independent parameters needed
to specify states relevant to an entity’s law-governed properties and behaviors) than
are needed to specify the corresponding state of the system of entities upon which
it cotemporally materially depends.
3⁹ As I argue in Wilson 2015b, and as I’ll discuss further in §1.4.4, even a contingentist categoricalist
Humean—that is, someone who thinks that causation is a matter of contingent regularities, and that
powers are reducible to categorical, non-dispositional features—can accept powers in the metaphysi-
cally neutral sense at issue in the schemas.
⁴⁰ As will become clear, the approach to Weak emergence/realization that I endorse bears certain
similarities to, but has certain crucial advantages over, the accounts of realization endorsed in
Shoemaker 2000/2001 and 2007 and Clapp 2001.
In Chapter 4 (‘The viability of Strong emergence’), I consider and respond to

a representative range of objections to the schema for Strong emergence and the
associated ‘new power’ approach to physically unacceptable emergence presented
and developed in Wilson 1999, 2002a, and 2015b, among other venues.⁴1 These
objections include that satisfaction of the conditions in the schema renders
Strong emergence naturalistically unacceptable or ‘scientifically irrelevant’, is
compatible with physicalism, is impossible owing to the base feature’s inheriting
any purportedly novel power, and is not necessary for physically unacceptable
emergence. Here again, I argue that the diverse challenges can be answered. And
here again, for each challenge one or more strategies of response are available on
any sensible implementation of the schema for Strong emergence. Upon occasion,
however, additional responses draw on features of my preferred ‘fundamental
interaction-relative’ account of Strong emergence (as per my 2002a), according to
which a Strongly emergent entity (feature) has at least one power that is grounded,
at least in part, in a novel (nonphysical) fundamental interaction.
Having established the in-principle viability of both Weak and Strong concep-
tions of metaphysical emergence, I go on to put this result to work, in considering
whether complex systems, ordinary (inanimate) objects, consciousness (charac-
teristic of persons), and free will (characteristic of agents), are plausibly seen as
actually either Weakly or Strongly emergent.
In Chapter 5 (‘Complex systems’), I first discuss how the historical assumption
that nonlinearity is a marker of fundamental novelty of the sort at issue in Strong
emergence was undermined by the discovery of complex nonlinear systems which
were clearly physically acceptable. I then suggest (drawing on Wilson 2002a) an
alternative empirical criterion of Strong emergence, as located in an apparent
violation of a conservation law and associated new fundamental interaction. By
lights of this criterion, the Strong emergence of complex systems remains an
empirically open but currently unmotivated possibility. Turning next to Weak
emergence, I argue (drawing on Wilson 2010b and 2013b) that while bare appeals
to common features of complex systems such as algorithmic incompressibility
(Bedau 1997 and 2008), dynamic self-organization (Mitchell 2012), and universal-
ity (Batterman 1998) do not themselves provide a decisive basis for taking complex
systems to be Weakly emergent, cases can be made that these or related features
entail satisfaction of the conditions in the schema. Most promisingly, I argue,
complex systems exhibiting universality of the sort Batterman focuses on also
have (as he observes) degrees of freedom (DOF) that are eliminated relative to
the systems of their composing lower-level entities, and so are Weakly emergent
by lights of a DOF-based account; and I offer reason to think that certain other
⁴1 Variations on this approach to Strong emergence can be found in, e.g., Broad 1925, McLaughlin
1992, and O’Connor 1994.
complex systems (Bedau’s gliders in the game of Life; Mitchell’s flocks of birds) may
also be seen as Weakly emergent by these lights. Here also I address the concern,
suggested by discussions in Morrison 2012 and Lamb 2015, that complex systems
involve not fewer but more DOF than their dependence base systems, associated
with ‘order parameters’ that arise near critical points.
In Chapter 6 (‘Ordinary objects’), I consider whether ordinary (inanimate)
objects, of either natural or artifactual varieties, are either Weakly or Strongly
emergent. I offer three motivations for thinking that such objects are ‘at least’
Weakly emergent in having at least one feature satisfying conditions in the asso-
ciated schema: first, I argue that insofar as quantum DOF are eliminated from the
specification of ordinary objects of the sort appropriately treated by classical (or
‘Newtonian’) mechanics, such objects are Weakly emergent by lights of a DOF-
based account; second, I argue that a common conception of artifacts as associated
with sortal properties and distinctive functional roles supports thinking of these
as being at least Weakly emergent by lights of a functional realization account;
third, I argue that ordinary objects typically have metaphysically indeterminate
boundaries, which when coupled with an attractive determinable-based account
of such indeterminacy, indicates that such objects are at least Weakly emergent, by
lights of a determinable-based account of such emergence. While the Strong emer-
gence of ordinary objects remains an open empirical (if not commonly endorsed)
possibility, the best such case involves artifacts. More specifically, I suggest that
artifacts might be Strongly emergent—if the states of consciousness that determine
what powers are possessed by artifacts are themselves Strongly emergent, as is
explored in Ch. 7. I close by noting that the previous results undercut the meta-
ontological view, endorsed in Thomasson 2010 and elsewhere, according to which
investigations into the status of ordinary objects should proceed differently from
investigations into the status of special-science entities.
In Chapter 7 (‘Consciousness’) I turn to considering whether consciousness of
the sort that we and other creatures enjoy is plausibly seen as either Weakly or
Strongly emergent. Existing arguments for the Strong emergence of consciousness
rely, one way or another, on the supposition that certain of the characteris-
tic features of consciousness—notably, its subjective or qualitative aspects—lie
beyond the explanatory reach of any lower-level physical goings-on. Though, as
will have been established by this point, the presence of even an insuperable
explanatory gap is not in itself a sufficient indicator of Strong emergence, the
proponents of explanatory gap arguments take such gaps to be metaphysically
significant, in reflecting not just broadly mathematical barriers to explanation
such as nonlinearity, but rather that the features at issue depart so greatly from
physical features that this divergence provides reasonable grounds for thinking
that no physicalist account of consciousness of either reductive or nonreductive
(i.e., Weakly emergent) varieties could possibly be correct. I consider the two
most promising forms of explanatory gap argument: knowledge arguments of the

sort advanced in Nagel 1974 and Jackson 1982 and 1986, and the conceivability
argument advanced in Chalmers 1996, 2009, and elsewhere, and I argue that
each form of explanatory gap argument is uncompelling, for reasons not much
previously explored. The upshot will be that while it remains an open empirical
possibility that consciousness is Strongly emergent, at present we have no com-
pelling philosophical or empirical motivation for taking this to actually be so. I go
on to argue that, on the supposition that consciousness is not Strongly emergent,
attention to the irreducibly determinable nature of qualitative conscious states
provides good reason to see certain such states as realized in determinable-based
fashion by lower-level physical states, and hence as Weakly emergent.
In Chapter 8 (‘Free will’), I consider whether free will of the sort that we appear
to have and to exercise is either Weakly or Strongly emergent. I start by drawing
on Bernstein and Wilson 2016 to present a framework for connecting existing
positions on free will—most importantly, compatibilism and libertarianism—to
existing positions in the mental (more generally: higher-level) causation debates.
In our paper, Bernstein and I argue that a representative range of compatibilist
accounts are appropriately seen as implementing a ‘proper subset’ strategy rel-
evantly similar to that implemented by nonreductive physicalists; here I extend
this result to establish that the compatibilist strategy also entails satisfaction of the
conditions in the schema for Weak emergence. I then argue that a representative
range of libertarian accounts are appropriately seen as implementing a ‘new power’
strategy entailing satisfaction of the conditions in the schema for Strong emer-
gence. This setup established, I first argue that free will of the compatibilist/Weakly
emergent variety is plausibly widespread, then present a novel argument for the
conclusion that at least some instances of seemingly free choice are properly
taken to be of the libertarian/Strongly emergent variety. At the end of the day—
and contra discussions which have primarily focused on qualitative aspects of
conscious experience (as in Chalmers 1996)—I suggest that libertarian free will
provides the best case for there actually being Strong emergence.
I finish up, in Chapter 9 (‘Closing remarks’) by calling attention to some
phenomena whose status as metaphysically emergent deserves further attention,
and by making some methodological observations that point towards other ways
in which the present project might be profitably extended.
1.4 Operative notions
In the remainder of this chapter, I say a bit more about certain operative notions
playing a regular background role in the investigations to follow. These are, first,
the physical; second, levels and their individuation; third, the fundamental; fourth,
causes and powers; fifth, methodology. Those interested in cutting to the chase of
later chapters are invited to use what follows as a reference section, if and when
needed.
1.4.1 The physical
As above, debates over the status of some goings-on as metaphysically emergent

typically take physical monism for granted, and more generally suppose that
the entities and features upon which emergent goings-on depend are ultimately
physical. But what is it for some goings-on to be physical?
The usual understanding is one according to which the physical goings-on
are those treated by physics. Such a view reflects a transition from an a priori
to an a posteriori characterization of the compositionally basic goings-on. As
Crane and Mellor (1990) tell the story, the compositionally basic entities were
historically characterized as having certain features definitive of matter—being
impenetrable, being conserved, being such as (only) to deterministically interact,
and so on. But contemporary physics has shown that the compositionally basic
entities actually have few, if any, of these characteristics; hence the characterization
of such entities—deemed ‘physical’—is now taken to be determined a posteriori
by physics. (More precisely, in the first instance the usual understanding is one
according to which physics determines the extension of the ‘basic’ or ‘narrowly’
physical entities. It remains to say how the range of basic physical goings-on
should be expanded so as to constitute the physical level—a task I address
in §1.4.2.)
A physics-based approach to the physical, though common, faces the concern,
articulated in ‘Hempel’s Dilemma’ (acknowledging Hempel 1979), that neither
current nor future physics will do so far as characterizing the domain of the
(basic) physical is concerned—at least if this domain (in combination with the
operative expansive resources) is supposed to serve as a substantive dialectical
basis for exploring whether some form of physicalism is correct. For if the physics
at issue is current physics, then (since current physics is to some extent both
inaccurate and incomplete) so will be the associated domain. But if the physics
at issue is future physics, then (since we don’t know what future physics will
end up positing) the domain of the physical will be presently indeterminate.
Indeed, for all we now know, future physics might end up positing basic entities
or features that are intuitively physically unacceptable—most pressingly, in being
fundamentally mental, at odds with the intended conception of the physical as
contrasting with panpsychism or other doctrines that neither physicalists nor their
Strong emergentist rivals would accept (Papineau 1993, Loewer 2001).
Hempel’s dilemma can be avoided, however (see Wilson 2006), in a way con-
sonant with historic and contemporary debate on the status of physicalism and
competing doctrines.⁴2 To start with the concern that future physics might end
up positing fundamentally mental goings-on: this difficulty may be avoided by
noting that Crane and Mellor’s genealogy omits a crucial fact—namely, that phys-
icalists have not handed over all authority to physics to determine, a posteriori,
what is physical. Reflecting the historical roots of physicalism in materialism as
foundationally committed to understanding mentality as nothing over and above
complex material goings on, one feature has remained definitive of the term
‘physical’: namely, that the compositionally basic physical entities and features
are not fundamentally mental—that is, do not individually either possess or
bestow mentality. Hence a physics-based account of the physical should not be
understood as the view that any and all entities treated by physics—current,
future, or ideal—are physical, but must rather incorporate a ‘No Fundamental
Mentality’ constraint along these lines. Strong emergentists would agree.⁴3 Note
that this constraint, properly understood, attaches to the basic or comparatively
noncomplex entities and features which are the targeted subject matter of physics,
and so does not rule out configurations of physical entities or their features
(what I called the ‘physically acceptable’ entities and features in my 2006) from
individually having or bestowing mentality.
The ‘No Fundamental Mentality’ constraint serves, in the main, to address
Hempel-style concerns with a future physics-based account of the physical. The
problem remains, however, that if the entities posited by future physics are of too
different a character from those posited by present physics, the present content
and applicability of such an account will be compromised—an observation which
lies at the heart of efforts (as in Melnyk 1997) to make sense of a present-physics
account of the physical. Though, as I argue in my (2006), present-physics accounts
are not ultimately sustainable, the apt concern about indeterminacy can, I believe,
be accommodated by taking the (basic) physical entities to be those that are
⁴2 See Stoljar 2001, Dowell 2006, and Ney 2008 for discussion of alternative conceptions of the
physical and associated approaches to resolving Hempel’s dilemma. A full discussion of these alter-
natives would take us too far afield; here I’ll just register that these conceptions are for various reasons
(as in, e.g., Ney’s conception of physicalism as an ‘attitude’ as opposed to a metaphysical thesis, or
Stoljar’s conception of the physical as ostensively defined by reference to paradigmatic inanimate
objects and their components, where nothing rules out the components’ being conscious) not well-
suited to characterizing the compositionally basic entities as physical in terms consonant with the
presuppositions of accounts of emergence.
⁴3 Notwithstanding the focus in Hempel’s Dilemma on formulating physicalism, the dilemma
also attaches to the formulation of Strong emergentism, since this view shares with (all versions of)
physicalism commitment to the compositionally basic entities’ being physical. Like physicalists, Strong
emergentists maintain that the compositionally basic entities neither have nor bestow mentality, as
per the intended contrast of Strong emergence with views, such as pan- or proto-psychism, on which
consciousness is had or possessed by individual compositionally basic entities.
approximately accurately treated by present or future (in the limit of inquiry, ideal)
physics.⁴⁴
Positing the (basic) physicality of nonfundamentally mental entities treated
approximately accurately by present or future versions of physics prevents physics’
present failures from immediately falsifying physicalism, while providing contin-
uous content to the account of the physical through the needed revisions. In what
follows this account of the (basic) physical will be operative, though as we’ll see not
much will turn on the specific details of this account. The main take-home point is
that there is at least one physics-based account of the (basic) physical suitable for
our dialectical purposes.
One question remains: what if there are no compositionally basic entities, as per
a ‘gunky’ world, in which everything can be further decomposed (i.e., has further
proper parts)?⁴⁵ Here I am inclined to think that even if the physical entities are
further decomposable, one may treat them as (effectively) compositionally basic
so long as features of the composing goings-on are either inherited or irrelevant to
the influence of goings-on at the physical level or above. In such a case of infinite
decomposability, the ‘No Fundamental Mentality’ would need to be revised,
perhaps along lines of a ‘No Low-Level Mentality’ constraint (as per Montero 2006)
or along lines of an ‘Only Structural-Dynamic Mentality’ constraint (as per Alter
2020). In what follows, I will assume that certain physical entities are (or can be
treated as) compositionally basic, but will keep track of any concerns arising from
this supposition.
1.4.2 The individuation of levels
It is natural to think of metaphysical emergence in the target cases, whether natural

or artifactual, as going hand-in-hand with the suggestion that emergent entities
and features are ‘higher-level’ with respect to the ‘lower-level’ goings-on upon
which they depend. As Rueger and McGivern (2010) observe:
Talk of levels or layers of reality is ubiquitous in science and in philosophy. It

is widely assumed, for instance, that physical, chemical, biological, and mental
phenomena can be ordered in a hierarchy of levels […] . (379)
Still, as Wimsatt (1994) notes, “the notion of a compositional level of organization

is left unanalyzed by virtually all extant analyses of inter-level reduction and
emergence” (203), notwithstanding that “levels and other modes of organization
⁴⁴ See Hellman and Thompson 1975, Papineau 1993, Ravenscroft 1997, Papineau 2001, and Loewer
2001 for variations on this theme.
⁴⁵ See Zimmerman 1995 and Schaffer 2003 for discussion.
cannot be taken for granted, but demand characterization and analysis” (204).
How should we understand talk of levels in what follows, and most importantly,
which entities and features should be taken to exist at a given level? The question
is a delicate one in various dialectical respects.
To start, at the present stage of investigation—antecedent to having considered
anti-realist or reductionist reasons for rejecting realist treatments of metaphysical
emergence and associated multiple levels—talk of there being ‘higher’ and ‘lower’
levels should be taken to reflect the appearances, so as not to rule ‘one-level’
positions out of court. As Rueger and McGivern (2010) note, in the face of certain
pressures, we might well wonder whether levels have “no ontological significance
and are merely different modes of representing an un-layered reality” (380).
Also important for dialectical purposes is that (as previously observed) levels
(or the one level, if anti-realism or reductionism turns out to be correct) be
individuated so as to include any combinations or configurations of entities and
features to which the anti-realist or reductionist may reasonably appeal. Suppose,
for purposes of illustration, that fundamental physics is atomic, such that the
basic physical entities are atoms and the basic physical relations include spatial
relations and pairwise bonding relations between atoms. Then, beyond the atoms
and atomic relations, we should allow as existing, at the atomic level, not just
small numbers of atoms standing in atomic relations, but also (among other
aggregative combinations) large numbers of atoms standing in highly complex
atomic (including spatial) relations, constituting pluralities or aggregates of the
sort that might, if reductionism is correct, be identical with a rock, a plant, or a
person, at least at any given time. Similar resources are needed to make sense of
anti-realist views, such as Heil’s (to be discussed further in Ch. 3), which reject
both reductionism and the independent posit of higher-level features.⁴⁶
If we are suitably generous to the anti-realist or reductionist, substantive debate
can proceed over whether some apparently higher-level entity or feature really is
higher-level, or is rather identical to some lower-level entity or feature (or perhaps,
as on Heil’s account, doesn’t exist at all). Being suitably generous means that
we cannot rest with certain ways of individuating levels. For example, Wimsatt’s
(1994, 226) understanding of “compositional levels of organization […] as consti-
tuted by families of entities usually of comparable size and dynamical properties,
which characteristically interact primarily with one another” rules reductionism
out of court. Indeed, unless the compositionally basic level can contain complex
configurations of (perhaps vastly) different sizes, one will not be able to make
⁴⁶ As Heil (2003a) puts it:

I am inclined to think that ‘this is a statue’ can be, and often is, literally true. What makes
it true is a complex, dynamic, arrangement of particles […] We deploy predicates like ‘is a
statue’ to mark off salient features of the world. These features are grounded in properties
and arrangements of the fundamental constituents. (217)
sense of Wimsatt’s reductionist-friendly claim that “Because any complex material

objects can be described at a number of different levels of organization, identity
relations must hold between descriptions of the same object at different levels”
(227–8).⁴⁷ For related reasons, the occasional suggestion in the literature that it is
indicative of a new ‘level’ (or more specifically, of emergence) that a feature of a
configuration be different, one way or another, from any feature of the individual
entities entering into the configuration (as when the mass of a composite has a
different value from the mass of any of its parts, or when a diamond has a property,
such as hardness, that no carbon atom has), is unsatisfactory for present dialectical
purposes; for a reductionist can happily agree that configurations (pluralities or
structural aggregates) have features not had by the individuals entering into the
configurations.⁴⁸
On the other hand, neither do we want to be so generous to the reductionist (or
other deflationist) as to rule the possibility of either Weak or Strong emergence out
of court.
With these constraints in mind, the question to be answered is: which combi-
nations of entities and associated features should be taken to exist at a given level
L of broadly scientific reality—most crucially, the fundamental physical level of
reality—beyond the entities and features typically taken, by lights of the associated
science S, to be characteristic of L? In short: how should levels be individuated?
In what follows, I’ll discuss two different approaches to answering this question,
neither of which is perfect, but either of which suffices to get discussion off the
ground.
The ‘lightweight combination’ approach

One common approach to the individuation of levels proceeds by allowing that
various ontologically ‘lightweight’ combinations of the characteristic entities and
features treated by a given science S and placed at a level L are also appropriately
placed at L. For example, Hellman and Thompson (1975) first individuate the
compositionally basic level in terms of the entities and features taken to be
characteristic of fundamental physics, as including any
⁴⁷ That said, Wimsatt does aim to make sense of such identity claims by reference to a number of
modes of ‘aggregation’, in a way that is friendly to the ‘lightweight combination’ approach to be discussed
shortly.
⁴⁸ Discussions of emergence in scientific contexts, when not appealing merely to epistemic notions,
frequently appeal to these sorts of conceptions of emergence, tracking novelty of features of a whole
with respect to features of the parts (as in, e.g., the characterization in Blitz and Richmond 1994
of “emergence in a system [as involving] a property not possessed by any of its parts”, 1)—though
these discussions typically then walk this back by introducing aggregative considerations which more
accurately capture the import of the oft-cited slogan that ‘the whole is more than the sum of the parts’.
… satisfying any predicate in a list of basic positive physical predicates of [fun-

damental physical theory]. Such a list might include, e.g., ‘is a neutrino’, ‘is an
electromagnetic field’, ‘is a four-dimensional manifold’, and ‘are related by a force
obeying the equations (Einstein’s, say) listed’, etc. (554)
(See also Melnyk 1997 and others.) Hellman and Thompson then expand beyond
the entities and features picked out by the basic physical predicates to include
at the basic physical level (to simplify somewhat) any and all mereological sums
(i.e., ‘fusions’ of parts into wholes) of spatiotemporally located instances of basic
physical entities and features.⁴⁹
Several other modes of combination are usually allowed, applying to entities,
features, or both, which are supposed by all parties not to result in any interesting
form of emergence. These are typically operations which are uncontroversially
aggregative using resources of the science at issue (e.g., spatial relations, or itera-
tions of pairwise relations, as in the atomic example just given), or are broadly log-
ical, as involving certain Boolean, classical mereological, or set-theoretic combi-
nations of entities or features. Some mathematical modes of combination are also
considered ontologically lightweight in this context; in particular, it is common
to suppose that features corresponding to linear combinations of L-level features
should also be placed at L. Hence at a level L associated with a given science S,
the L-level entities and features (including properties, states, and relations) would
typically be taken to include:
• any characteristic entities or features treated by S

• any entity (or feature) consisting in a set or plurality of L-level entities (or
features), understood as (merely) jointly existing
• any entity (or feature) consisting in a disjunction of any L-level entities (or
features)
• any entity (or feature) consisting in a conjunction of L-level entities (or
features)
• any entity (or feature) consisting in a mereological fusion of L-level entities
(or features)
• any feature consisting in a linear (scalar or vector) combination of L-level
features⁵⁰
⁴⁹ Classical mereology is a theory of parts and wholes on which composition (‘fusion’) is ontolog-
ically lightweight, in that any objects qua parts automatically form a whole. There is controversy over
whether material composition (as when, e.g., some atoms compose a molecule, and so on) can or should
be understood in terms of mereological composition; see Simons 1987, van Inwagen 1990, McDaniel
2001, Paul 2002, Koslicki 2008, Bennett 2015, and Varzi 2019 for discussion.
⁵⁰ Wimsatt (1994) more generally suggests that “the conditions required for a system property to
be an aggregate of the properties of the parts of the system—conditions on the ‘composition function’
• any relational entity consisting in any number of L-level entities standing in

L-level relations
• any relational feature of the form “being composed of some L-level entities
with L-level features standing in L-level relations.”⁵1
Note that the specification here allows for iterative closure under the operations;
for example, an entity consisting in a disjunction of conjunctions of L-level
relational entities would also be L-level. Though in principle there might be further
ontologically lightweight operations, it is common to suppose, on this approach,
that the closure of L-level entities under these operations is more or less exhaustive
of the (individual and aggregative) entities at L. Hence it is that debate over the
status of a given seemingly higher-level entity or feature frequently proceeds by
considering whether the entity or feature can be reduced to one or other of these
lightweight combinations of characteristic lower-level entities or features. For
example, a nonreductive physicalist might argue that the multiple realizability of a
given mental feature type indicates that it is not reducible to any type of lower-level
physical feature, while a reductive physicalist might respond by suggesting that the
multiple realizability of the mental type can be accommodated, compatible with
reduction, by taking the mental type to be identical to a disjunction of physical
types.
For most purposes in what follows, no harm comes from taking levels to be
individuated along lines of the ontologically lightweight approach, as again is
commonly done; I will flag any exceptions to this rule.
relating system and parts’ properties” are “associativity, commutativity, inter-substitutivity, linearity,
and invariance under decomposition and reaggregation” (237).
⁵1 Features of this form are sometimes called “micro-based” or “micro-structural” features; see Kim
1998 and Shoemaker 2007, following Armstrong 1978. Kim (1998) offers on the reductionist’s behalf
the feature of being a water molecule as a case in point: “it is the property of having two hydrogen atoms
and one oxygen in such and such bonding relationship” (84). More generally, Kim (1998) characterizes
a micro-based property as follows:
P is a micro-based property just in case P is the property of being completely decomposable
into nonoverlapping proper parts, a1 , a2 , . . . ,an , such that P1 (a1 ), P2 (a2 ), . . . ,Pn (an ), and
R(a1 , a2 , , . . . , an ). (84)
Shoemaker’s related (2007) characterization is as follows:
[Micro-structural] properties […] can be specified entirely in terms of the micro-manifest
powers of the constituent micro-entities together with how these micro-entities are related
i.e., in terms of what could be known about them prior to their entering into emergence
engendering combinations. Such a property will be the property of being composed of
particles with such and such micro-manifest causal powers and related in such and such
a way. […] If emergentism is false, manifest causal powers are the only ones the micro-
entities have, and physical micro-structural properties are the only ones macro-objects have,
and the other properties of macro-objects are realized in their physical micro-structural
properties. (55)
The law-consequence approach

An alternative approach to the individuation of levels, which I here put on the
table, expands upon the appeal to scientific laws, understood as applying to
entities at a distinctive level of natural reality. Laws so understood are, in the
first instance, metaphysical; they are, or encode, the ‘rules’ governing the entities
and features at issue, though, consonant with a suitably fallibilist realism about
theories, discussion of laws typically focuses on claims made in or by scientific
theories as appropriate stand-ins for claims about the associated laws.
On the law-consequence approach, the suggestion is that the laws governing
entities characteristic of a given level can also do the work of expanding the domain
of entities and features at that level in such a way that an anti-realist or reductionist
view is not ruled out of court. The laws of fundamental physics, for example, are
capable of taking as input or initial conditions various complex configurations
(pluralities or structural aggregates) of characteristic physical entities and features;
hence the laws/theories themselves have resources to expand beyond the explicit
focus of the laws/theory treating some L-level entities and features, to admit
at that level any entities and features whose (potential) existence is deemed a
metaphysical consequence—not to be confused with either mere necessitation or
representational entailment—of the L-level laws.
A law-consequence approach to the individuation of levels has certain advan-
tages over a lightweight combination approach. For example, a law-consequence
approach need not antecedently specify whether nonlinear entities or features
are or are not to be placed at a given level L; whether this is so will follow
from the laws governing the characteristic entities at that level. Relatedly and
more generally, on this approach entities and features which are causal conse-
quences just of the L-level laws may also be placed at L. Another advantage of a
law-consequence approach is that it need not be committed to any and every
ontologically lightweight combination of entities or features at L’s also being placed
at L—which exclusion might be desirable if, say, some complex configurations of
atoms could not exist, for some law-based reason.
That said, one might be concerned that, while a law-consequence approach to
levels clearly makes room for the possibility of Strong (physically unacceptable)
emergence (since fundamental higher-level powers or other features of reality
will not be metaphysical consequences of lower-level laws), it rules out Weak
(physically unacceptable) emergence as associated with nonreductive physicalism.
After all, nonreductive physicalists grant that the higher-level entities and features
that they take to be genuine, as well as the higher-level laws governing these entities
and features, are in some sense metaphysical consequences of the fundamental
physical laws, even if these higher-level goings-on are (as nonreductive physicalists
suppose) different from any lower-level goings-on (and moreover, as some nonre-
ductive physicalists think, are in some cases epistemically beyond our ken).
However, a law-consequence approach, properly understood, does not rule out

(the possibility of) Weak emergence. For if there are any metaphysical conse-
quences of the sort that the Weak emergentist supposes there to be, these will not
be appropriately placed at L. In Ch. 3, I’ll develop this point in terms of degrees
of freedom. The basic idea is intuitive, however. To start, laws require certain
kinds of information in order to operate. For example, in order for the quantum
laws to operate on a configuration of fundamental physical particles, the input
into the laws has to contain information about the individual quantum spins of
those particles. Suppose also that it is a metaphysical consequence of the quantum
configuration’s existing and evolving as per the quantum laws that a given planet
exists and evolves as per the laws of astronomy, where (as a Weak emergentist
might well maintain) the characterization of planets and their behaviors is taken
to abstract away from any details concerning quantum spin. Were the planet, so
characterized, to be given as input into the quantum laws, this input would not
contain the information about quantum spin required for the laws to operate.
Correspondingly, and more generally, planets and their features would not be
appropriately placed at the fundamental physical level, since the physical laws
would not be able to operate on such entities or features.
The upshot is that a law-consequence approach to the individuation of levels
does not entail that any and all metaphysical consequences of the laws at a
given level L should be placed at that level. Rather, only those consequences
which preserve the degrees of freedom required to allow the laws at level L to
operate should be placed at that level. Nor does a law-consequence approach to
the individuation of levels build in the truth of emergentism: whether there are
actually any goings-on whose characterization eliminates lower-level degrees of
freedom is a matter for empirical and philosophical determination, to be further
considered down the line.
1.4.3 The fundamental
In what follows, we will have occasion to consider some goings-on—entities,

features, interactions—as fundamental. For example, both physicalists and their
Strong emergentist rivals suppose that the physical goings-on are fundamental,
at least to the extent that certain physical entities are compositionally basic; where
they disagree is over whether there are any other fundamental goings-on. But what
is it for some goings-on to be fundamental?
There are three main approaches to this question (see Tahko 2018 for dis-
cussion), divided into primitivist and non-primitivist varieties. Non-primitivist
approaches aim to metaphysically analyze what it is for some goings-on to be fun-
damental in terms of dependence, one way or another. On independence-based
conceptions, what makes it the case that some goings-on at a world w are funda-
mental at w is that those goings-on are (individually) metaphysically independent
at w.⁵2 On dependence-based conceptions, what makes it the case that some
goings-on at a world w are fundamental at w is that those goings-on are part of
a complete minimal basis for everything that exists at w, such that everything that
exists at w is either fundamental or completely metaphysically dependent on what
is fundamental.⁵3 By way of contrast, on primitivist accounts, what makes it the
case that some goings-on at a world w are fundamental at w is not metaphysically
analyzable in other terms, positive or negative.⁵⁴
My own preference is for a primitivist account. I follow Fine (2001) in thinking
that one should “reject the idea that the absolute notion of fundamental reality is
in need of a relational underpinning” (25); as I previously put it, the fundamental
goings-on “play a role analogous to axioms in a theory—they are basic, they are ‘all
God had to do, or create’ ” (Wilson 2014a). (Nota bene that it is not the fundamenta
themselves, but what makes it the case that some goings-on are fundamental, that
is taken to be primitive.) Such an account has (so I argue in Wilson in progressb)
numerous advantages over either variety of non-primitivist account. For the most
part, however, this dispute will not matter for what follows.
The main exception to this rule concerns the usual understanding of an
independence-based conception, according to which a given entity x is funda-
mental just in case it does not depend on any other entity (see, e.g., Schaffer
2009, 373 and Bennett 2017, 138). Such a conception problematically rules out
numerous accounts of fundamental structure, including
• Cases involving self-dependent phenomena (as per a self-sufficient God)

• Cases involving fundamenta which metaphysically depend on some but not
all other fundamenta (say, quarks, which in coming only in pairs or triplets
appear to be existentially dependent on other quark types and/or tokens)⁵⁵
• Cases involving fundamenta each of which metaphysically depends on all
other fundamenta (as per Hua-yen Buddhism or ontic structural realism)
and most importantly for present purposes,
• Cases involving fundamenta that are partly but not completely metaphys-
ically dependent on other fundamenta (as per typical understandings of
Strongly emergent phenomena).
⁵2 See, e.g., Schaffer 2009, Audi 2012a, Dasgupta 2014, Raven 2016, and Bennett 2017.
⁵3 See, e.g., Sider 2011, Jenkins 2013, Paul 2002, and Tahko 2014.
⁵⁴ See, e.g., Fine 2001 and Wilson 2014a and in progressb.
⁵⁵ See Tahko 2018 and Wilson 2020 for discussion.
Whether some other version of an independence-based conception of fundamen-

tality might make room for such Strong emergence remains to be seen. Here it is
worth noting, however, that proponents of independence-based conceptions do
not distinguish between the claim that individual fundamenta do not depend on
any other goings-on, and the claim that the fundamental goings-on collectively
do not depend on any other goings-on. The former claim seems to me to be
unmotivated, given the many seeming counterexamples, whereas the latter claim
seems true and capable of accommodating the typical understanding of Strongly
emergent entities or features. Correspondingly, appeals to fundamentality in what
follows, as applying either to physical or to Strongly emergent phenomena, may
be understood as broadly neutral between a primitivist, a dependence-based
(‘complete minimal basis’), or a collective-independence-based conception of
fundamentality.
1.4.4 Causes and powers
The discussions to come often advert to causal relations and associated powers
to produce effects, as had by both physical and macro-level (special-scientific or
artifactual) goings-on. There are vast literatures on causation and powers, as well as
on how these notions enter, metaphysically and modally, into the characterizations
of entities and features. Fortunately, for purposes of the present project it is possible
to remain almost entirely neutral as regards these more specific details.
To start, the operative notion of ‘power’ in what follows is metaphysically highly
neutral, following the presuppositions operative in my (2015b):
[T]alk of powers is simply shorthand for talk of what causal contributions pos-
session of a given feature makes (or can make, relative to the same laws of nature)
to an entity’s bringing about an effect, when in certain circumstances. That
features are associated with actual or potential causal contributions (‘powers’)
reflects the uncontroversial fact that what entities do (can do, relative to the
same laws of nature) depends on how they are (what features they have). So, for
example, a magnet attracts nearby pins in virtue of being magnetic, not massy; a
magnet falls to the ground when dropped in virtue of being massy, not magnetic.
Moreover, a feature may contribute to diverse effects, given diverse circumstances
of its occurrence (which circumstances may be internal or external to the entity
possessing the feature). Anyone accepting that what effects a particular causes
(can cause, relative to the same laws of nature) is in part a function of what
features it has—effectively, all participants to the present debate—is in position
to accept powers, in this shorthand, metaphysically neutral and nomologically
motivated sense. (354)
Note that whether internal or external circumstances are at issue, the sense in
which a given feature (potentially) contributes to causing a given effect in certain
circumstances is minimally one according to which those circumstances alone are
not up to the task of (potentially) causing that effect.
Relatedly, the operative notion of causation in what follows is also metaphys-
ically highly neutral. Again, there are many specific accounts of this notion, but
it may serve as an initial proof of concept (to be filled in a bit in Ch. 2, after
presenting the two schemas for emergence) that even a contingentist categoricalist
Humean—someone who thinks that causation is a matter of regularities, such
that features have their powers contingently, and that all features are ultimately
categorical—can accept powers and the associated notion of causation in the
neutral sense(s) here: for such a Humean, to say that an (ultimately categorical)
feature has a certain power would be to say that, were a token of the feature to occur
in certain circumstances, a certain (contingent) regularity would be instanced.
Contemporary Humeans implement more sophisticated variations on this theme;
but the point remains that no ‘heavyweight’ notion of powers or causation need
be presupposed in what follows.⁵⁶
More generally, no controversial theses pertaining to the nature of powers, cau-
sation, properties, or laws are here presupposed. Four points of non-commitment
are worth highlighting. First, nothing in what follows requires accepting that
it is essential to features that they have the powers they actually have—that
is, that they make the causal contributions they actually do, when in certain
circumstances. Perhaps features are essentially (or necessarily) associated with
certain powers; perhaps they aren’t. (I do assume, as is uniformly done by both
‘contingentists’ and ‘necessitarians’, that features have the same powers relative to
the same laws of nature.) Second, nothing in what follows requires accepting that
features are exhaustively individuated by powers. Perhaps they are, perhaps they
aren’t; perhaps features are also or ultimately individuated by quiddities or other
noncausal aspects of features. Third, nothing in what follows requires accepting
that powers are or are not reducible to categorical (non-dispositional) features,
or that attributions of powers are or are not reducible to certain conditionals or
counterfactuals, etc. Perhaps powers, or talk of them, are reducible to other entities
or terms; perhaps they aren’t. Fourth, nothing in what follows requires endorsing a
specific understanding of features—as, e.g., universals (of some variety), or rather
tropes (of some variety). And as per usual, nominalists rejecting properties or
⁵⁶ Carruth (2018) argues that some heavyweight conceptions of powers might not be suited for cer-
tain emergentist purposes, since the conceptions open the door to certain reductionist strategies. A full
engagement with Carruth’s discussion would take us too far afield; here I’ll just register that certain of
the reductionist strategies he advances fail in the face of methodological considerations to be discussed
shortly, and others are along lines of those operative the ‘collapse objection’ to be treated in Ch. 4 (§4.2).
other features are invited to implement their preferred metaphysical translation

strategy.
One last issue pertaining to the operative understanding of causation and
powers deserves mention here. Both Kim’s problem of higher-level causation and
the schemas for metaphysical emergence motivated by this problem presuppose
that there is physical causation and associated powers—but some (notably, Russell,
1912 and Field, 2003) maintain that such a supposition is problematic. As Field
characterizes Russell’s line of thought, “not only does the word ‘cause’ not appear in
the advanced sciences, but the laws that these sciences state are incompatible with
causation as we normally understand it” (435). To be sure, that there is no physical
causation remains an outside view, as is reflected in the standard understanding
of physicalism as committed to Physical Causal Closure (to be discussed in more
detail in the next chapter), according to which any physical effect has a sufficient
purely physical cause; so I could perhaps be forgiven for simply bracketing the
concern. But it may be worth briefly registering why I take the assumption of
physical causes and powers to be justified.
To start, that physical equations are expressed in broadly mathematical rather
than explicitly causal terms seems to me to be best understood as reflecting the
usual scientific focus on the prediction of quantities, coupled with a kind of expres-
sive underdetermination associated with logical and mathematical expressions, as
opposed to encoding that physical goings-on do not enter into causal relations.
In ‘F = ma’, to take a toy example, is the import of the ‘=’ sign causal or not?
What matters for this and related questions is whether the usual descriptions
and interpretations of physical theories make reference to causal notions; and on
the face of it (as in the quotation by Kane near the start of this chapter) they
frequently do so, in adverting to interactions, transfers of energy, behaviors, and
other notions naturally seen as having causal connotations (see Earman 1976 for
similar observations).
A more serious concern would be if the laws of physics were “incompatible
with causation as we normally understand it”. But here too doubts can be raised.
To start, Russell’s argument for this conclusion presupposes a certain implausible
conception of causation, according to which true causal claims take the form of
universal generalizations. Here the fact that the notions of causation and powers
operative in what follows are metaphysically highly neutral is relevant, since for
all Russell establishes there may be alternative conceptions of physical causation
compatible with our best science and one’s preferred metaphysical commitments.
Field discusses another potential source of incompatibility, according to which
(unlike physical laws) causation is ‘directed’, such that causes typically precede and
determine their effects, rather than vice versa; but as he notes, there are available
strategies for locating directionality in the physical laws.
Field ends up taking the main concern about physical causation to be that “since
there is always a possibility of interventions from afar”, a specification of the cause
of an effect up to the task of guaranteeing the associated effect will have to include
“each part of the past light cone of an event among the causes of the event” (439).
But this expression of the concern reverts to Russell’s implausible conception of a
cause as that which guarantees the associated effect (such that causal claims take
the form of universal generalizations). In fact, that causal relations are typically
ceteris paribus—i.e., interfering factors aside—is par for the course so far as causal
relations in the special sciences are concerned, and from what I can tell there is no
barrier to understanding physical causation as similarly admitting of exceptions.
I conclude that it is the supposition that physical causation would have to be
exceptionless, rather than the notion of physical causation itself, that is “a relic
of a bygone age”.
1.4.5 Methodology
Notwithstanding the Cartesian caricature of metaphysicians as aiming to discover

the nature of reality from the comfort of their armchairs, contemporary meta-
physicians more typically endorse a broadly abductive methodology, also known
as ‘inference to the best explanation’ (see, e.g., Harman 1965 and Douven 2017),
whereby candidate metaphysical accounts of a given phenomenon are assessed by
attention to how well they do, overall, at satisfying various theoretical desiderata.
As Sider (2009) describes ‘mainstream metaphysics’:
Competing positions are treated as tentative hypotheses about the world, and
are assessed by a loose battery of criteria for theory choice. Match with ordinary
usage and belief sometimes plays a role in this assessment […] Theoretical
insight, considerations of simplicity, integration with other domains (for instance
science, logic, and philosophy of language), and so on, play important roles.
(358)
Notwithstanding this broadly common metaphysical methodology, there is con-

siderable variation in exactly which theoretical desiderata are operative (i.e., how
to fill in Sider’s reference to ‘and so on’) as well as in how these desiderata, which
may push in different directions, should be weighted.
Though some see the lack of current consensus as regards fine-grained meta-
physical methodology as reason to be somewhat skeptical about metaphysical
progress, my own view is that (as discussed in Wilson 2011a, 2016b, and 2016c)
there is at present no warrant for such pessismism. The lack of current fixed
standards is unsurprising, given the wide purview of metaphysical investigations
as potentially open to any and all data relevant to a given phenomenon, and
given our present distance, metaphysically and epistemologically, from the end
of inquiry. There is moreover reason to think that we are slowly but surely making
methodological advances. Indeed, the present common acceptance of inference
to the best explanation as operative in philosophy as well as in the sciences itself
reflects an advance in methodological understanding, associated with the demise
of verificationism in particular. In any case, in circumstances where there is some
but not complete consensus regarding methodological standards, progress can
nonetheless be made, so long as one is suitably explicit about which theoretical
desiderata are primarily guiding one’s investigations. If one is so explicit, the
hope is that even those with somewhat different methodological sensibilities may
find one’s investigations useful, in mapping the domain of associated options as
regards understanding the phenomenon at issue. Here, then, I put on the table
two methodological desiderata which will guide the investigation to follow.
The first is that we should aim to realistically accommodate the appearances
of metaphysical emergence, in the absence of specific reason(s) to think that
this cannot be done. To be sure, we have yet to consider the range of objections
that might be and have been raised against such emergence. Even so, in what
follows I take accommodation of these appearances on their own terms—that
is, accommodation of the natural (i.e., straightforward, default) understanding
of the relevant scientific and ordinary experiential data, as suggesting that there
are phenomena coupling cotemporal material dependence with ontological and
causal autonomy—to be a heavily weighted virtue of an account of this data. As I
see it, we should give up such natural understandings of the structure of reality as
admitting metaphysically emergent entities and features only if forced to do so—
e.g., by causal overdetermination considerations of the sort to be considered in Ch.
2, or by arguments showing that reasonable conditions on emergence cannot be
satisfied. In particular, I do not take a supposed desire for ontological parsimony
for its own sake, of the sort that might lead one to attempt to ‘explain away’
the prima facie motivations for metaphysical emergence even in the absence of
specific objections to this notion, as having a similar weight. (That’s not to say
that parsimony considerations don’t carry any weight, of course.) The intended
desideratum can then be expressed as follows:
Criterion of Appropriate Accommodation: An adequate account of metaphysical

emergence should make natural (straightforward, default) and realistic sense of
the appearances of metaphysical emergence, in the absence of specific reasons to
think that this cannot be done.
Reflecting the Criterion of Appropriate Accommodation, though I do take it to

be a central part of the burden in what follows to show that various purported
problems with metaphysical emergence can be addressed, I do not take it to
be part of the burden to show that no deflationary (anti-realist or reductionist)
account of the appearances of metaphysical emergence (or of the considerations
motivating these appearances, including, e.g., there being seemingly distinctive
special-science taxonomies and laws) is viable. My ultimate goal, in other words,
is not to knock the anti-realist or reductionist off their horse, but rather to show the
metaphysical emergentist who aims to accommodate the appearances at realistic
face value how to stay on their own horse. Again, I hope that those with different
methodological sensibilities will nonetheless find the ensuing discussion useful,
at least as revealing the extent to which the antecedent weighting of parsimony
considerations, as opposed to any specific problem with the notion of meta-
physical emergence itself, may be playing a role in deflationary accounts of such
emergence.
A second criterion reflects my supposition that (again, modulo reasons to
think that no realistic account of the appearances can be given) an adequate
account of metaphysical emergence must not only appropriately accommodate
these appearances, but moreover do so in a way that provides an illuminating—
that is, explanatorily relevant—basis for understanding just how the characteristic
features of such emergence might be in place:
Criterion of Illuminating Accommodation: An adequate account of metaphysical

emergence should provide an illuminating basis for accommodating the appear-
ances of metaphysical emergence in natural (straightforward, default) fashion.
Relatedly, an adequate account of metaphysical emergence must provide a clear

basis for the in-principle resolution of disputes over whether some phenomenon
is metaphysically emergent, in a way going beyond appeal to brute intuitions or
irrelevant distinctions. For example (as I noted in discussing a version of this
criterion in Wilson 2019), even supposing that an oracle existed who could infal-
libly report whether a given phenomenon is metaphysically emergent, an account
pitched in terms of such oracular pronouncements would fail to illuminate what
it is to be metaphysically emergent, and so would fail to satisfy the criterion. More
plausibly, an account according to which metaphysically emergent phenomena are
cotemporally materially dependent on but ontologically and causally autonomous
with respect to lower-level physical phenomena, but which did not provide any
explanatory insight into how, exactly, such dependent goings-on might be so

autonomous, would lead directly to stalemate between emergentists and their
rivals (e.g., with regard to the status as emergent of certain mental states), and
so would fail to satisfy the Criterion of Illuminating Accommodation.
Our operative notions on the table, let us begin.
2
Two schemas for metaphysical emergence
In this chapter, I start by presenting what is seen by many as the most pressing
challenge to taking the appearances of emergent structure as genuine—namely,
the problem of higher-level causation, made salient by Jaegwon Kim in his
1989, 1993a, 1998, and elsewhere, according to which metaphysically emergent
higher-level features would problematically causally overdetermine effects already
brought about by the base features upon which they cotemporally materially
depend.1 I argue, following discussions in Wilson 1999, 2011b, 2015b, and
elsewhere, that there are two and only two strategies of response to this problem
that appropriately accommodate there being metaphysically emergent entities and
features—that is, entities and features that are cotemporally materially dependent
on, yet also ontologically and causally autonomous with respect to, dependence
base entities and features. One of these strategies provides a schematic basis
for ‘Weak’ metaphysical emergence, of the sort that would be compatible with
physicalism, on the assumption that the compositionally basic entities are physical;
the other provides a schematic basis for ‘Strong’ metaphysical emergence, of the
sort that would be incompatible with physicalism, on that assumption.2 Along
the way, I show that representative ranges of seemingly diverse accounts of
realization (i.e., Weak emergence) and of Strong emergence are plausibly seen
as providing an illuminating basis for satisfying the conditions in the relevant
schema, such that accounts of metaphysical emergence are more unified than they
appear.
I conclude that we have prima facie reason to think that satisfaction of the
conditions in the schemas for Weak and Strong emergence is, as I put it, ‘core
and crucial’ to metaphysical emergence of physically acceptable and physically
unacceptable varieties, respectively. Of course, any schematic account needs to be
filled in and moreover filled in sensibly, if it is to be really adequate. But modulo
this caveat, the results of this chapter can also be seen as providing prima facie
1 See Malcolm 1968 for an early variation on the theme, focused on an apparent conflict between
mechanical and intentional (psychological) explanations. Recall that ‘emergent’ and ‘higher-level’ are
not exact synonyms, given that vitalist and substance pluralist views of the sort intended to contrast
with any form of emergentism also countenance higher-level entities and features.
2 As previously noted, although the assumption that the base-level entities and features are
physically acceptable is typically operative in what follows, the schemas generalize to characterize
emergence of two different varieties, whatever the precise ontological status of the base-level goings-on.
DOI: 10.1093/oso/9780198823742.003.0002
reason to think that the conditions in the schemas for Weak and Strong emergence
are, when sensibly filled in, both necessary and sufficient for (appropriate and
illuminating accommodation of) metaphysical emergence of physically acceptable
and physically unacceptable varieties, respectively—a bold claim, but one that, as
we will see in ensuing chapters, is surprisingly robust.
2.1 The problem of higher-level causation
I start with three clarificatory remarks. First, following Kim and standard practice,
I assume that entities (objects, systems, or other particulars) are efficacious in
virtue of having efficacious features (states, properties, behaviors, or other ways
for entities to be). For example, the effects that a billiard ball causes (can cause)
are a matter of what features it has—its mass, shape, velocity, and so on. Cor-
respondingly, in what follows talk of entities’ causing effects is suppressed in
favor of talk of their features’ causing effects. The assumption that the efficacy
of entities lies in their having efficacious features is conveniently consonant with
the usual assumption that the emergence of entities is ultimately a matter of the
emergence of certain features. Second, again following Kim and standard practice,
I assume that causation is in the first instance a relation between spatiotemporally
located goings-on; hence reference to ‘features’ in what follows is to be understood,
unless otherwise qualified, as reference to spatiotemporally located tokens (e.g.,
particular states or property instances, or associated particular events) of a given
(state, property, event) type.3 Third, to fix ideas I set up the problem as directed
at special-science entities and features, but nothing deep hangs on this focus;
the problem arises more generally for higher-level entities and features of either
natural or artifactual varieties.
2.1.1 Kim’s overdetermination argument
Six premises lead to the problem of higher-level causation.⁴ Four of these concern
special-science features, and are motivated by considerations similar to those
giving rise to there seeming to be metaphysical emergence. These are:
3 That said, I will sometimes gloss the type/token distinction—e.g., when discussing the necessita-
tion of one feature by another, later in this section.
⁴ What follows reflects my preferred way of presenting the problem and slate of candidate resolu-
tions, as set out, e.g., in Wilson 2009, 2011b, 2015b. As I’ll discuss, Kim’s own presentations aim to
motivate reductive over nonreductive versions of physicalism.
two schemas for metaphysical emergence 41
1. Dependence. Special-science features cotemporally materially depend on

lower-level physical features (henceforth, ‘base features’) in such a way
that, at a minimum, the occurrence of a given special-science feature on a
given occasion minimally nomologically supervenes on base features on that
occasion.⁵
2. Reality. Both special-science features and their base features are real.
3. Efficacy. Special-science features are causally efficacious.
4. Distinctness. Special-science features are distinct from their base features.
The remaining two premises concern causation. The fifth is a standard physicalist
commitment, sometimes called ‘the causal closure of the physical’:
5. Physical Causal Closure. Every lower-level physical effect has a sufficient

purely lower-level physical cause.⁶
The sixth reflects the common supposition that there is no systematic causal
overdetermination (henceforth, just: ‘overdetermination’) of effects by distinct
individually sufficient causes, with the exception of cases involving causes forming
part of a single diachronic causal chain, and ‘double-rock-throw’-type cases, where
a given effect (e.g., a window’s breaking) is, on a given occasion, the result of two
distinct causes (e.g., two rock-throwings), each of which is individually sufficient
for an effect of the type at issue. Note that in neither type of ‘exception’ case does
one of the competing causes stand in a relation of cotemporal material dependence
to the other.
6. Non-overdetermination. With the exception of cases of the double-rock-

throw variety, effects are not causally overdetermined by distinct individ-
ually sufficient cotemporal causes.⁷
⁵ Recall: by ‘minimal nomological supervenience’ is meant that in worlds with relevantly similar
laws of nature, any given token of the supervenient (e.g., special-scientific) type requires, for its
occurrence, a token of some base (e.g., physical) type; and in such worlds, if any token of that base
type occurs, then a token of the supervening type will occur.
⁶ In being formulated in terms of lower-level physical causes and effects, this characterization of
Physical Causal Closure is similar to those in Baker 1993, 79 (“Every instantiation of a micro-physical
property that has a cause at t has a complete micro-physical cause at t” and Sturgeon 1998, 124 (“Every
quantum event has a fully disclosive, purely quantum history”). See Montero 2003 and Garcia 2014 for
discussion of more generous characterizations of the closure principle in terms of any (even higher-
level) physically acceptable features, which characterizations are not useful for purposes of setting up
Kim’s problematic.
⁷ This characterization of Non-overdetermination (also sometimes called ‘the causal exclusion
principle’) is along lines of that in Yates 2012, 3 (“No single event can have more than one sufficient
cause occurring at any given time—unless this is a genuine case of causal over-determination”).
On to the problem. There are two cases to consider, reflecting two sorts of effect
that might be at issue. In the first case, a special-science feature S is assumed to
cause another special-science feature S*; in the second case, S is assumed to cause
a base feature P*. In Kim’s classic presentation, S is taken to be a mental state (e.g.,
a state of being thirsty); P is taken to be a base state upon which mental state S
depends; and mental state S is taken to cause either another mental state S* (e.g., a
desire to quench one’s thirst) or a base state P* (e.g., a physical reaching for a glass
of water). More generally, however, the considerations to follow raise a concern
about how any real and distinct higher-level feature might be unproblematically
efficacious.
First (case 1), suppose that S causes special-science feature S* on a given
occasion (compatible with Efficacy). S* is cotemporally materially dependent on
some base feature P* (Dependence), such that P* necessitates S*, with at least
nomological necessity. Moreover, P* has a sufficient purely lower-level physical
cause (Physical Causal Closure)—plausibly, and without loss of generality, P. If P
causes P*, and P* (at least nomologically) necessitates S*, then it is plausible that P
causes S*, by causing P*. So, it appears, both P and S cause S*, and given that P and
S are both real and distinct (Reality, Distinctness), S* is causally overdetermined;
moreover (given Dependence), this overdetermination is not of the double-rock-
throw variety (contra Non-overdetermination). Diagrammatically, the case is as in
Figure 2.1, with bold lines representing causation:
S S*
P P*
Figure 2.1 Case 1 of the problem of higher-level causation: S causes S*
Second (case 2), suppose that S causes some base feature P* on a given occa-
sion (compatible with Efficacy). P* has a sufficient purely lower-level physical
cause (Physical Causal Closure)—plausibly, and without loss of generality, P. So,
it appears, both P and S cause P*, and given that P and S are both real and
distinct (by Reality and Distinctness), P* is causally overdetermined; moreover,
(given Dependence) this overdetermination is not of the double-rock-throw variety
(contra Non-overdetermination). Diagrammatically, the case is as in Figure 2.2:
P P*
Figure 2.2 Case 2 of the problem of higher-level causation: S causes P*
So goes Kim’s argument that real, distinct, dependent, and efficacious higher-level
features induce problematic overdetermination, whether their purported effects
are higher-level or lower-level.⁸
Kim sees his argument as motivating rejection of the premise that special-
science features are distinct from their base features—that is, he sees it as motivat-
ing reductionism (more specifically: reductive physicalism). For present purposes,
however (and following Wilson 2011b and elsewhere), it is useful to more generally
note that rejection of each of the premises of the argument is associated with one
or more fairly comprehensive positions in the metaphysics of science. Rejection
of one or other of the first four premises gives rise to the following strategies of
response, and associated positions:⁹
1. Substance dualism. Deny Dependence: avoid overdetermination by denying

that S and S* cotemporally materially depend on base features P and P*,
respectively.1⁰ If higher-level features S and S* do not so depend on lower-
level features, there is no motivation for positing a base feature P as a
dependence base for M, hence no motivation for positing a competing causal
chain from P to M* (case 1) or from P to P* (case 2).
2. Eliminativism. Deny Reality: avoid overdetermination by denying that S and
S* are real.11
⁸ Kim also argues that cases of the first type (involving ‘same-level’ causation) involve cases of the
second type (involving ‘downward’ causation), on grounds that if a given higher-level feature (e.g., S*)
is realized by a lower-level feature (e.g., P*), any cause of the higher-level feature must also be a cause of
the lower-level feature, as per what he calls ‘The Causal Realization Principle’; see, e.g., Kim 1993a, 439.
I will later offer my own reasons for thinking that ‘same-level’ causation might also involve downward
causation, at least in the case of Strong emergence.
⁹ In discussing these strategies and associated positions, I flag certain proponents of the positions,
for the reader’s reference; some such proponents explicitly appeal to the problem of higher-level
overdetermination as motivating the position, but not all do so.
1⁰ See, e.g., Descartes 1641–7/1984 and Chalmers 1996. Pan- or proto-psychists are also appropri-
ately categorized as rejecting Dependence, on the assumption (operative here; and reflecting dialectical
and historical considerations discussed in Wilson 2006) that (basic) physical goings-on do not
individually have or bestow mentality. See, e.g., Seager’s (2012) description of a form of panpsychism
requiring “that the mental not be ontologically dependent on non-mental features of the world” (19).
11 See, e.g., Paul Churchland 1981 and 1984, and Patricia Churchland 1986, for eliminativism about
certain mental higher-level features, and Merricks 2003 for eliminativism about certain non-mental
higher-level features.
3. Epiphenomenalism. Deny Efficacy: avoid overdetermination by denying that

S is efficacious.12
4. Reductive physicalism. Deny Distinctness: avoid overdetermination by deny-
ing that S is distinct from P.13
Each of these strategies avoids overdetermination, but not in a way that makes
sense of higher-level features as metaphysically emergent. In the case at hand,
for S to be emergent, it must cotemporally materially depend on base feature P
while being both ontologically and causally autonomous with respect to P—that
is, while being distinct from and distinctively efficacious with respect to P. But
the substance dualist strategy denies that S cotemporally materially depends on
P, on grounds that S is instantiated in a nonphysical substance.1⁴ The eliminativist
and reductive physicalist strategies each deny that S is ontologically autonomous—
that is, distinct—from P: the eliminativist denies this on grounds that S doesn’t
exist, and the reductive physicalist denies this on grounds that S is identical with
P. Finally, the epiphenomenalist and reductive physicalist strategies each involve
denying that S is causally autonomous—that is, distinctively efficacious—with
respect to P: the epiphenomenalist denies this on grounds that S isn’t efficacious at
all, and the reductive physicalist denies this on grounds that S and P are efficacious
in just the same way, since they are identical.
2.1.2 The two ‘emergentist’ strategies for responding to the problem
The remaining strategies for responding to the problem of higher-level causation,

and associated positions, do better by way of accommodating emergence. These
are as follows:
5. Strong emergentism. Deny Physical Causal Closure: avoid overdetermination

by denying that every lower-level physical effect has a sufficient purely lower-
level physical cause.
6. Weak emergentism/Nonreductive physicalism. Deny Non-overdetermination:
allow that effects caused by S are overdetermined by P, but maintain that
the overdetermination here is of an unproblematic non-double-rock-throw
variety.
12 See, e.g., Hodgeson 1962 and Huxley 1874; see Robinson 2012 for contemporary literature.
13 See, e.g., Smart 1958, Lewis 1966, and Kim 1993a.
1⁴ Similarly, the panpsychist will deny that S cotemporally materially depends on physical P, on the
usual assumption that the (basic) physical goings-on do not themselves have or bestow mentality.
As I argue in the next two sections, these two strategies and associated positions
are perspicuously seen as motivated by two conditions on the powers of a given
special-science feature, where satisfaction of one or other condition provides a
prima facie plausible and principled (i.e., appropriate and illuminating) basis
for taking the feature to be emergent, in ways that standard proponents of
the strategy/position would endorse. In each of these sections, treating Strong
emergence and Weak emergence, respectively, I start by motivating the associated
condition on powers by attention to standard versions of the position; I then
show how satisfaction of the condition dovetails with the associated strategy for
responding to the problem of higher-level causation; I then provide prima facie
reasons for thinking that satisfaction of the condition provides an appropriate
and illuminating basis for taking special-science features to be both cotemporally
materially dependent and ontologically and causally autonomous; finally, I use the
condition to formulate the associated schema for metaphysical emergence.
Before getting started, three points of clarification are worth noting. First, as
prefigured in Ch. 1 (§1.4.4), talk of ‘powers’ in what follows is simply shorthand for
talk of what causal contributions possession of a given feature makes (or can make,
relative to the same laws of nature) to an entity’s bringing about an effect, when
in certain circumstances (where the circumstances alone are not up to the task of
bringing about that effect). Anyone who accepts that the effects an entity causes (or
can cause, relative to the same laws of nature) are in part a function of what features
the entity has—effectively, all participants to the present debate—is in position
to accept ‘powers’, in the shorthand, metaphysically neutral and nomologically
motivated sense at issue here. Besides commitment to the platitude that what
entities can do (cause), relative to the same laws of nature, depends on how they are
(what features they have), only one metaphysical condition is required in order to
make sense of the powers-based conditions to follow; namely, that one’s account
of (actual or potential) causal contributions (powers) has resources sufficient to
ground the identity (or non-identity) of a token causal contribution associated
with a token of a higher-level feature, with a token causal contribution associated
with a token of a lower-level feature. Here again, effectively all participants to the
debate can make sense of such identity (non-identity) claims as applied to token
(actual or potential) causal contributions (token powers).1⁵ Second, the qualifier
1⁵ For example, suppose a contingentist categoricalist Humean wants to take a physicalist approach
to the problem of higher-level causation, and so aims (as I will expand on §2.3) to identify every
token power of a token higher-level feature with a token power of its lower-level base feature. As
previously discussed, such a Humean understands powers in terms of actual or potential instances
of a (contingent) regularity. Where the aim is to avoid overdetermination, the Humean may suppose,
to start, that the (relevant instances of the) regularities overlap, both with respect to the (single) effect,
and with respect to the (single) circumstances in which the two token features occur. If the Humean
aims to be a reductive physicalist, they may suppose that such overlap motivates identifying the token
features at issue, and hence the associated powers. If the Humean aims to be a nonreductive physicalist,
‘prima facie’ in the previous paragraph reflects that a full defense of the claim
that the two conditions on powers serve as the basis for two viable conceptions
of metaphysical emergence requires detailed treatment of the sort I’ll conduct
in Chapters 3 and 4. Third, though it is an interesting question how to more
specifically understand the forms of dependence at issue in the different schemas
for emergence, entering into these details now would take us too far afield. Hence
in formulating the schemas, the condition on dependence is expressed simply as
requiring cotemporal material dependence, as different forms of emergence agree.
Here again, more detailed discussion will be found in later chapters.
2.2 Strong emergentism and the New Power Condition
Strong emergentists maintain that some special-science features are real, cotem-
porally materially dependent on, distinct from, and distinctively efficacious with
respect to their base features. So far, Weak emergentists agree—but the conception
of higher-level efficacy at issue in Strong emergentism is, as we have seen, one
which denies Physical Causal Closure, and is correspondingly incompatible with
physicalism. And while different accounts of Strong emergentism emphasize
different aspects of this distinctive efficacy as located in fundamentally novel
features, laws, effects, forces, interactions, and the like, core and common to these
accounts is that Strongly emergent features have fundamentally novel powers—
powers to produce effects entailing the violation, in particular, of Physical Causal
Closure.
This is true, to start, on British Emergentism, as endorsed most systematically
by Mill (1843/1973), Alexander (1920), Lewes (1875), and Broad (1925).1⁶ Again,
notwithstanding some variations in emphasis, the core claim that Strong emer-
gence involves a fundamentally novel power is a recognizable running theme in
these accounts.1⁷ Hence in his classic survey, McLaughlin (1992) describes British
they can reject this identification of features, on difference-making or other grounds of the sort to
be discussed §2.3. Such a Humean will suppose that attention to broader patterns of regularities can
provide a basis for identifying token powers of token features, even when the token features are not
themselves identical. Whether reductive or nonreductive, the contingentist categoricalist Humean can
make sense of the claim that some, all, or none of the token powers of token features are identical. As I
observed in my (2015b, 35), this case is like the case of New York: if we can make it (out) here, we can
make it (out) anywhere.
1⁶ Evidently Lewes was the first to use the term ‘emergent’ to characterize certain higher-level entities
and features.
1⁷ In the case of Alexander, there is some controversy about the content (and consistency) of his
view, and relatedly, about whether his intended account is supposed to be a version of Strong or of Weak
emergence. Here I follow McLaughlin (1992) in grouping Alexander with other Strong emergentists on
the basis of certain natural interpretations of some of his statements. Nothing deep hinges on exactly
who is categorized as a Strong or Weak emergentist; what is crucial is that accounts of emergence that
Emergentism as “the doctrine that there are fundamental powers to influence

motion associated with types of structures of particles that compose certain
chemical, biological, and psychological kinds” (52), where the powers at issue are
typically taken to be “powers to generate fundamental forces not generated by any
pairs of elementary particles” (71).
For example, Mill, the father of British Emergentism, characterizes emergence
(in Book III, Ch. 6 of his 1843/1973, ‘On the Composition of Causes’) in terms
of ‘heteropathic’ effects of joint causes, which effects cannot, unlike ‘homopathic’
effects, be understood as the mere aggregative sum of the effects of the causes when
operating independently—a conception reflecting the supposition, natural at the
time, that a violation of broadly additive composition laws of the sort associated,
e.g., with the vector addition of forces, sufficed to indicate a fundamentally novel
causal influence and associated law. Mill claims that
To whatever degree we might imagine our knowledge of the properties of the

several ingredients of a living body to be extended and perfected, it is certain
that no mere summing up of the separate actions of those elements will ever
amount to the action of the living body itself. (Book III, Ch. 6, §1)
On Mill’s account, then, emergents are distinctively efficacious in being able

to produce novel effects—effects going beyond those that base-level goings-on
in additive or aggregative combination are able to produce. And though Mill
motivates his view by attention to a distinction between diachronically produced
effects, it is (as discussed in Ch. 1, note 7) straightforward and desirable to translate
Mill’s talk of diachronically produced effects into talk of cotemporally had powers:
to say that an effect of a feature of a composite entity is heteropathic is just to
say that the feature has a power not had by features of its components when
in linear combination. Mill himself moves seamlessly from talk of heteropathic
effects to talk of new properties of and laws governing entities capable of causing
such effects:
[W]here the principle of Composition of Causes [ . . . ] fails [ . . . ] the concur-

rence of causes is such as to determine a change in the properties of the body
generally, and render it subject to new laws, more or less dissimilar to those
to which it conformed in its previous state. (1843/1973, 435)
As McLaughlin (1992) notes, “Mill holds that collocations of agents can possess
fundamental force-giving properties” (65).
are plausibly seen as compatible with physicalism, or as not so compatible, encode, one way or another,
certain identifiable conditions on powers of emergent and base features.
Both Mill’s reference to “new laws” and his supposition that such cases contrast
with “the extensive and important class of phenomena commonly called mechani-
cal” indicate that Mill takes heteropathic effects to be associated with novel powers,
enabling the feature (or its possessing “body”) to override the usual composition
laws in the production of certain effects. Moreover, as Mill’s emphasis on the
contrast with effects obeying the ‘Composition of Causes’ attests, the novel powers
at issue in this variety of emergence are fundamentally novel, rather than merely
reflecting a nonfundamental summation or other aggregation of existing lower-
level powers.1⁸
For a second example: in his (1920), Alexander (at least on one common
reading; see McLaughlin 1992 for discussion) similarly takes fundamentally novel
powers to be core to emergence. Consider the following passage:
Physical and chemical processes of a certain complexity have the quality of

life. [ . . . ] The higher quality emerges from the lower level of existence and
has its roots therein, but it emerges therefrom, and it does not belong to that
level, but constitutes its possessor a new order of existent with its special laws
of behaviour. The existence of emergent qualities thus described is something
to be noted, as some would say, under the compulsion of brute empirical
fact, or, as I should prefer to say in less harsh terms, to be accepted with the
“natural piety” of the investigator. It admits no explanation. (46–7)
Here Alexander takes an emergent feature to be indicative of a “new order of

existent” with “special laws of behaviour”, which novel behaviors are associated
with novel powers, as per the usual reciprocal relation between powers and causes.
Moreover, Alexander’s famous claim that emergent features must be accepted with
“natural piety” is strongly suggestive of—or in any case compatible with—a view
on which the novel powers at issue are fundamentally novel.
Broad (1925) also focuses on laws, in a way that is yet more explicit in the
supposition that Strong emergence involves fundamental novelty of powers. In
particular, Broad maintains that emergent features arise in accord with funda-
1⁸ Indeed (and echoing remarks in the Ch. 1, §1.4.2 discussion of the individuation of levels), any
powers of a configuration that are novel merely in reflecting a nonfundamental summation or other
aggregation of powers of individual components of the configuration are irrelevant to metaphysical
emergence of either Strong or Weak varieties, for all parties to the present debate, including reductive
physicalists, will happily agree that configurations (pluralities or structural aggregates) can and indeed
typically do have such nonfundamentally novel powers. Correspondingly, metaphysical emergence,
whatever else it may be, is not a matter of such nonfundamental (i.e., merely aggregative) novelty of
powers. At issue in Strong emergence are powers that are fundamentally novel, in being over and above
any merely aggregative novelty associated with the coming-to-be of base-level micro-configurations
and features, and (as we will shortly see) in Weak emergence no novelty of powers, fundamental or
nonfundamental, is at issue.
mental ‘configurational’ or ‘trans-ordinal’ laws, connecting lower-level structures

with higher-level entities and features, which laws are just as metaphysically
and scientifically fundamental as the ‘intra-ordinal’ laws governing lower-level
physical phenomena:
[T]he law connecting the properties of silver-chloride with those of silver and
of chlorine and with the structure of the compound is, so far as we know, an
unique and ultimate law. (64–5)
Broad’s reference here to emergent laws’ being “unique and ultimate” again indi-
cates that the novelty of emergent laws is intended to be of the fundamental rather
than nonfundamental variety; and given that in Broad’s discussion of candidate
cases of emergence the laws at issue are again laws of behavior and associated
effects and powers, it follows that core to his account is that Strongly emergent
features have fundamentally novel powers.1⁹
Contemporary accounts of Strong emergence also typically agree in taking
emergent features to have or bestow fundamentally novel powers, not had (or had
only in derivative fashion) by base features or associated micro-configurations.2⁰
For example, Silberstein and McGeever (1999) speak of ‘fundamental’ irreducibil-
ity, and understand emergent features as having causal capacities irreducible to
(that is, not identical to) any capacities of features of the parts when in relation:
Ontologically emergent features are features of systems or wholes that possess

causal capacities not reducible to any of the intrinsic causal capacities of the
parts nor to any of the (reducible) relations between the parts. (186)
Van Gulick (2001) similarly characterizes what he calls “radical” emergence, high-
lighting that the novel powers at issue would undercut the physicalist assumption
of Physical Causal Closure:
1⁹ Broad put his official account of emergence in terms of ‘in-principle’ failure of deducibility,
but this reflected his assuming that such failures were indicative of fundamental novelty—that is, of
Strong emergence. As McLaughlin (1992) notes, “the Emergentists do not maintain that something is
an emergent because it is unpredictable. Rather, they maintain that something can be unpredictable
because it is an emergent” (73). I’ll return to this issue in Ch. 4.
2⁰ An interesting variation on this theme is found in Paolini Paoletti 2017, Ch. 7, and 2018a. Paolini
Paoletti suggests that it is in the first instance powers which are Strongly emergent; by way of filling in
his schematic (2018b) account of emergence as involving partial but not complete dependence, he takes
the dependence at issue to reflect that the having of the power by a higher-level entity is determined by
lower-level powers, whereas the autonomy at issue reflects its being the case that no lower-level goings-
on can manifest the power. My own view is that there is insufficient daylight between the bestowal and
the manifestation of a given power to quite capture the notion of Strong emergence, but in any case
Paolini Paoletti’s suggestion is a recognizable cousin of the more standard accounts I will discuss here.
If [ . . . ] system-level powers were not determined by the laws governing the

powers of their parts, then that would seem to imply the existence of powers
that could override or violate the laws governing the powers of the parts [ . . . ]
It is in this respect that radically emergent powers would pose such a direct
challenge to physicalism, since they would threaten the view of the physical
world as a closed causal system. (18–19)
O’Connor and Wong (2005) characterize emergent features as “fundamentally

new”, not just in being (perhaps epiphenomenally) different, but more specifically
in having fundamentally novel causal capacities:
[A]s a fundamentally new kind of feature, [an emergent feature] will confer
causal capacities on the object that go beyond the summation of capacities
directly conferred by the object’s microstructure. (665)
Hendry (2010) offers an account of Strong emergence continuous with that

associated with the British Emergentist tradition, as involving ‘configurational
Hamiltonians’, and similarly maintains that “the emergentist sees [special-science
properties] as distinct and nonreducible just because the causal powers they
confer are not exhausted by those conferred by their physical bases” (211), contra
views committed to “the completeness of physics”. More recently, Hendry (2017)
maintains that “the conferment of additional causal powers [is] a mark of the
distinct reality that is characteristic of something that is strongly emergent” (147).
Finally, in Wilson 1999, I maintain that “the [Strong] emergentist response to
the possibility of [mental] causal overdetermination is to deny that the physical
is causally closed [ . . . ] and [assert] that mental properties have causal powers
not possessed by any physicalistically acceptable base properties” (41). In Wilson
2002a, I go on to argue that, reflecting that scientific practice suggests that powers
are plausibly grounded, one way or another, in fundamental forces or interactions
(as when the power of a magnet to attract a pin is grounded in the electromagnetic
interaction), naturalistic good sense can be made of the Strong emergentist posit
of fundamentally novel powers, as reflecting novel fundamental interactions that
come into play only at certain levels of compositional complexity, such that Strong
emergentism “is committed to there being at least one other fundamental force
beyond those fundamental forces currently posited” (74).
2.2.1 The New Power Condition
Summing up, Strong emergentists suppose that at least some cotemporally

materially dependent features have powers (“causal capacities”, etc.) not had by
(“not directly conferred by”) their (lower-level, physical) base features, as per the
following New Power Condition:
New Power Condition: Token feature S has, on a given occasion, at least one
token power not identical with any token power of the token feature P upon
which S cotemporally materially depends, on that occasion.
In having a power that its base feature does not have, S has a novel token power.
Moreover, and notwithstanding that (as just substantiated), the novel powers at
issue in Strong emergence are fundamentally novel (as opposed to being non-
fundamentally novel, reflecting merely the difference between unaggregated and
aggregated lower-level entities and features), there is no need to explicitly reg-
ister this qualification in the New Power Condition. For the condition as stated
blocks this sort of nonfundamental novelty, since the condition characterizes the
emergence of S in terms of a relation between S’s powers and the powers of a
single lower-level feature P (had by a lower-level configuration) upon which S
cotemporally materially depends, rather than as a relation between S’s powers and
the powers of multiple lower-level features (had, e.g., by individual components
of the entities entering into the configuration having P). In other words: if S
has a power not identical to any power of P, where P is a feature of a con-
figuration (plurality or structural aggregate) of lower-level (ultimately physical)
entities, there is no chance that the novelty of S’s power might merely reflect
the difference between aggregated and nonaggregated lower-level goings-on; for
P’s powers will include any powers associated with such aggregative phenomena.
This clarification made, here and throughout, the New Power Condition should
be read as involving a fundamentally rather than a merely nonfundamentally
novel power.
Moreover, Strong emergentists typically assume that among the (fundamen-
tally) novel powers of a Strongly emergent feature are powers to influence lower-
level physical goings-on, in a way entailing the rejection of Physical Causal Closure
and ergo the falsity of physicalism. Hence the Strong emergentist position is
naturally associated with a strategy of response to the problem of higher-level
causation that proceeds by rejecting Closure. Moreover, as we’ll shortly see, only
against the backdrop of the rejection of Closure does satisfaction of the New Power
Condition provide a basis for avoiding overdetermination of higher-level as well
as lower-level effects.
The Strong emergentist response to the problem of higher-level causation

Let’s now return to the problem of higher-level causation, starting with the second,
simpler case, to see how satisfaction of the New Power Condition enters into
implementation of the Strong emergentist strategy for responding to this problem.
In the case where special-science feature S causes a base feature P* (case 2), the
Strong emergentist strategy involves, to start, the supposition that S satisfies the
New Power Condition specifically in having a fundamentally novel power to bring
about P*. For example, S might be a Strongly emergent state of being thirsty,
which depends on base feature P, and which in the circumstances causes a physical
reaching for a nearby glass of water P*. On this assumption, P* does not, contrary
to the assumption of Physical Causal Closure, have a sufficient purely lower-level
physical cause: as per the New Power Condition, P has no token power identical
with S’s token power to cause P*; hence either P is not at all a cause of P* (does not
have any power to cause P*), or else, if P can be understood to cause P* (that is, to
have a power to cause P*), P has this power only in a derivative way, in virtue
of P’s being a dependence base for S, which non-derivatively has the power at
issue.21 Either way, P fails to be a sufficient purely lower-level physical cause of P*;
and without loss of generality, it moreover follows that P* has no sufficient purely
lower-level physical cause, contra Physical Causal Closure, and overdetermination
is avoided, as in Figure 2.3:
P P*
Figure 2.3 The Strong emergentist’s response to case 2
Next, suppose (as per case 1), that S causes another special-science feature S*—
say, a desire to drink some water. Here the Strong emergentist supposition is
that S satisfies the New Power Condition specifically in having a fundamentally
novel power to bring about S*—that is, a power that P doesn’t have (either
at all, or non-derivatively). Interestingly, even though the novel power at issue
here is not directed at the production of a lower-level physical effect, it remains
that satisfaction of the New Power Condition in this case requires the falsity of
Physical Causal Closure. Why so? Because, if Closure held in this case, P would
have a non-derivative power to cause S*—by being a sufficient purely lower-level
physical cause of P*, which in turn nomologically necessitates S*. But in that case
overdetermination would not be avoided, and moreover the claim that S has a
fundamentally novel power to cause S* would be undermined. Hence the Strong
emergentist must deny Physical Causal Closure, even when the novel power had by
21 S’s causing of P* might be entirely independent of P, or it might be that S and P jointly cause P*;
either route to the production of P* is compatible with the denial of Physical Causal Closure. I’ll revisit
these options down the line.
Strongly emergent S is for the causing of a special-science feature S*. In this case,
the strategy is as in Figure 2.4:
S S*
P P*
Figure 2.4 The Strong emergentist’s response to case 1
A remaining question about the Strong emergentist treatment of this case

concerns what is responsible for S*’s having the base feature P* that it does. Given
that P is not (on this view) itself up to the task of causing P*, there are two
possibilities here: first is that S* carries with it its own dependence base P*, such
that S, in causing S*, also causes P* (such that cases of type 1, in which S causes S*,
turn out also to be cases of type 2, in which S causes P*); another is that S and P
jointly cause P* (with S either independently causing S* or else causing S* jointly
with P). Either way, the Strong emergentist response to case 1, like the response
to case 2, involves so-called ‘downward causation’. I will explore these options in
more detail in Ch. 4.
2.2.2 The schema for Strong emergence
Prima facie, satisfaction of the New Power Condition by a special-science feature S

which cotemporally materially depends on a base feature P provides an appropriate
and illuminating basis for avoiding overdetermination while guaranteeing that S is
both ontologically and causally autonomous with respect to P. First, since S has a
token power (at a time or over a temporal interval) that P doesn’t have (at that time
or over that interval), S is distinct from P (by Leibniz’s law); hence S is ontologically
autonomous with respect to P. Second, in having a novel token power, S can cause
an effect that P can’t cause, or that P can’t cause in the same (non-derivative) way as
S; hence S is causally autonomous—that is, distinctively efficacious—with respect
to P. The New Power Condition at the heart of the Strong emergentist’s strategy for
resolving the problem of higher-level causation thus provides the basis for our first
schema for metaphysical emergence:
Strong Emergence: What it is for token feature S to be Strongly metaphysically

emergent from token feature P on a given occasion is for it to be the case, on
that occasion, (i) that S cotemporally materially depends on P, and (ii) that S
has at least one token power not identical with any token power of P.
Here the locution ‘what it is for’ is intended to flag that Strong Emergence provides
a schematic metaphysical basis for a given case of such emergence, encoding what
is core and crucial to that notion. Of course, any schematic account needs to be
sensibly filled in. Modulo this caveat, the conditions in the schema are, I claim, nec-
essary and sufficient for (appropriate and illuminating accommodation of) meta-
physical emergence of the physically unacceptable variety.22 Some clarifications:
• The schema is formulated so as to apply to features of artifacts and the like as

well as to features of special-science entities.
• The first condition minimally specifies cotemporal material dependence, typ-
ically and comparatively neutrally understood as involving both (physical)
substance monism and the minimal nomological supervenience of emergent
feature types on base feature types.
• The second condition (effectively, the New Power Condition) captures the
comparatively strong sense in which an emergent feature may be causally,
hence ontologically, autonomous with respect to the base feature upon which
it cotemporally materially depends, in virtue of having a fundamentally novel
power relative to the powers of its base feature.
Note that while the initial presentation of the New Power Condition built
in the cotemporal material dependence of the feature having the new power
on its base feature, the schema for Strong Emergence treats the new power
and the dependence conditions separately. Henceforth, references to the New
Power Condition should be understood as references just to condition (ii) in
that schema.
• As I have been emphasizing throughout, the base feature at issue is a feature
of a given (ultimately physical micro-) configuration (plurality or structural
aggregate), and the novelty of the power at issue in condition (ii) should be
understood accordingly. In particular, the novelty of this power is not to be
understood as relative to the powers of features of individual entities entering
into the dependence base configuration.
• Again, S’s novel power is not just nonfundamentally novel (as reflecting a
merely aggregative difference between powers of a configuration and powers
of components of the configuration) but is moreover fundamentally novel. I
do not include the qualifier ‘fundamentally’ in the schema because including
it would misleadingly suggest that the nonfundamentality of S’s novel power
was among the available options, which (as previously discussed) it isn’t,
under the operative understanding of P as having any nonfundamental
powers attaching to mere aggregation of lower-level components.
22 This is not to presuppose, of course, that all implementations of Strong Emergence are on a par or
equally well suited to accommodate a given macro-phenomenon. As we’ll see in Ch. 4, my preferred
implementation has certain advantages over others.
• The schema is relativized to occasions (times or temporal intervals), but it is

worth noting that it would be reasonable to suppose that it suffices for the
Strong emergence of S, simpliciter, that the condition is ever satisfied, and to
suppose that it suffices for the Strong emergence of the feature type (of which
S is a token), simpliciter, that any token feature S on any occasion satisfies (or
would satisfy) the condition. These complications won’t play a role in what
follows.
2.3 Nonreductive physicalism and the Proper Subset

of Powers Condition
Nonreductive physicalists, like Strong emergentists, maintain that (some) special-

science features are real, cotemporally materially dependent, distinct, and
distinctively efficacious with respect to their base features. They too are non-
reductionists, maintaining that certain configurations (pluralities or structural
aggregates) of entities whose constituents are ultimately physical have features
from which other features emerge, constituting emergent entities and a novel
level of natural reality. But as physicalists, their response to the problem of
higher-level causation cannot entail the rejection of Physical Causal Closure,
which thesis is core to the physicalist view that the physical goings-on are an
existential and causal basis for all other broadly scientific phenomena. Rather,
nonreductive physicalists reject Non-overdetermination, maintaining that distinct
special-science and base features can each be a sufficient cause of a single effect,
in virtue of standing in a relation that, while not identity, is intimate enough both
to avoid overdetermination of the problematic (since implausible, for the cases at
issue) double-rock-throw variety and to retain compatibility with Physical Causal
Closure, hence with physicalism.23
In presenting their view, nonreductive physicalists (or those aiming to char-
acterize this position) typically endorse some or other ‘realization’ relation as
holding between tokens or types of the features at issue, which relation, either in
itself or in combination with some consonant condition, is supposed to illustrate or
illuminate how problematic overdetermination is avoided, compatible with both
physicalism and nonreduction, as in these representative cases:
• Functional realization (Putnam 1967, Fodor 1974, Papineau 1993, Antony

and Levine 1997, Melnyk 2003, Witmer 2003, Polger 2007, Yates 2012):
realized features are functionally implemented by lower-level physically
acceptable features. As applied to mental properties:
23 How nonreductive physicalists aim to establish that higher-level features are not just efficacious,
but distinctively so, will be addressed shortly.
Let “functionalism” be the view that, first, each mental property M

is identical with the second-order property of having some property
or other which plays the causal role associated with M, and second,
each actual instance of a mental property is realized by some physical
property. If anything counts as a received theory of the mental in
contemporary philosophy of mind, this is it. (Witmer 2003, 198)
[T]he fact that psychological properties are realized in some physical
property or other seemed to ensure that there was nothing occult or
dualistic about psychological causation: the causal interactions among
the inputs, outputs, and internal states of a device that realized a higher-
order functional system would be effected by the causal mechanisms
characteristic of the first-order realizers.
(Antony and Levine 1997, 86)
• Constitutive mechanism (Cummins 1975, Craver 2001, Haug 2010): realized
features of a system are implemented by lower-level micro-structural features
specifying causal mechanisms involving parts of the system:
The components of mechanisms, in contrast to those of mere aggregates,
have an active organization; they act and interact with one another in
such a way that the ψ-ing of [realized feature] S is more than just the
sum of [individual component] φ properties. (Craver 2001, 59)
[T]he realizer P must itself provide a causal mechanism for the instanti-
ation of M—a constitutive mechanism, in which the instantiation of M
in s is explained by the components of s being propertied and related in
certain ways. (Haug 2010, 321)
• Mereological realization (Shoemaker 2000/2001, Clapp 2001, Rueger and
McGivern 2010): realized features are proper parts of lower-level physical
features:
[T]he instantiation of a realizer property entails, and might be naturally
said to include as a part, the instantiation of the [ . . . ] property realized.
(Shoemaker 2000/2001, 28)
[M]ultiply realized mental properties, though real and causally effica-
cious, are better thought of as parts of their physical realizers. [ . . . ]
Just as there is no causal and/or explanatory competition between a
whole and its parts, so there is no causal and/or explanatory competition
between instances of mental properties and instances of their physical
realizers. (Clapp 2001, 133; italics in text)
• The determinable/determinate relation (Macdonald and Macdonald 1986,

Yablo 1992, Wilson 1999 and 2009): realized features are determinables of
lower-level physical features. Again as applied to mental phenomena:
Traditionally, the paradigm of one-way necessitation was the relation
of determinate to determinable [ . . . ]. What if mental phenomena are
determinables of physical phenomena in something like the traditional
sense [ . . . ]? (Yablo 1992, 250)
[W]e know that [determinables and determinates] are not causal rivals.
This kind of position is of course familiar from other contexts. Take
for example the claim that a space completely filled by one object can
contain no other. Then are even the object’s parts crowded out? No. In
this competition wholes and parts are not on opposing teams [ . . . ].
(Yablo 1992, 183)
• ‘Superdupervenience’ or explanatory necessary connection (Lepore and
Loewer 1987 and 1989, Horgan 1989 and 1993). As Horgan (1993) describes
the view:
[T]he sort of inter-level relation needed by the materialist who is also
a realist about a given mode of discourse (e.g., mental discourse) is not
bare supervenience, but rather what I hereby dub superdupervenience:
viz., ontological supervenience that is robustly explainable in a materi-
alistically explainable way. (566)
Here the unproblematic nature of the overdetermination at issue is reflected
in the association of the schematically characterized realization relation
with an additional, broadly counterfactual condition encoding the causal
relevance—or as Horgan evocatively puts it, ‘quausal’ relevance, whereby a
feature is efficacious ‘qua’ the type of feature it is—of the realized feature:
What seems required for quausal relevance [of a property F vis-à-vis
an effect having property G], then, is a wider pattern of counterfac-
tual dependence of the occurrence and non-occurrence of events with
property G upon the occurrence or non-occurrence of events with
property F. (Horgan 1989, 59)
[S]ubjunctive conditionals [ . . . ] lie at the heart of the view that [mental]
content properties are causally potent.
(Lepore and Loewer 1989, 188)
Now, underlying the seeming diversity in these and many other accounts of
nonreductive physicalism hides a deeper unity of strategy, which again can be put
in terms of a certain condition on powers (see Wilson 1999 and 2011b). Motivating
this condition—the Proper Subset of Powers Condition—and seeing how it provides
a unifying basis for a wide range of nonreductive physicalist responses to the

problem of higher-level causation takes a bit more doing, however, than in the
corresponding case of the New Power Condition as operative in accounts of Strong
emergence. I’ll start by putting the Proper Subset of Powers Condition on the table,
and arguing that each of the aforementioned accounts of realization aims to ensure
its satisfaction; I’ll then argue that satisfaction of this condition provides a basis
not just for a cotemporally materially dependent feature S to be distinct, but also
for S to be distinctively efficacious; I’ll then show how this condition provides an
illuminating basis for understanding the nonreductive physicalist’s response to the
problem of higher-level causation; finally, I’ll use the condition as the basis for the
schema for Weak emergence.
2.3.1 The Proper Subset of Powers Condition
Again, physicalists, unlike Strong emergentists, cannot allow that special-scientific

features have powers different from the physical base features upon which they
cotemporally materially depend, on a given occasion; for if some special-scientific
feature S were to have a power its physical base feature didn’t have, then this
would violate the physicalist supposition that the physical goings-on provide a
complete ontological and causal basis for all other broadly scientific goings-on.
Correspondingly, physicalists are committed to every special-science feature S’s
satisfying the following condition:
Token Identity of Powers Condition: Every token power of token feature S, on

a given occasion, is identical with a token power of the token feature P on
which S cotemporally materially depends, on that occasion.
Note that reductive physicalists can and do accept the Token Identity of Powers
Condition, so long as it is allowed that a feature can depend on itself in the relevant
sense. For example, if a given mental type is (as the reductive physicalist believes)
identical to a given physical type, then tokens of these types, along with their
associated token powers, will also be identical, and the Token Identity of Powers
Condition will be satisfied.
As I’ll now argue by attention to the aforementioned accounts of realization,
nonreductive physicalists moreover typically accept another, stronger condition,
according to which at least some higher-level features S stand in a relation to their
lower-level base features P satisfying the following condition:
Proper Subset of Powers Condition: Token feature S has, on a given occasion, a

non-empty proper subset of the token powers of the token feature P on which
S cotemporally materially depends, on that occasion.2⁴
Functional realization
First, consider functionalist accounts of realization, on which realized types are
higher-order types associated with causal roles that, on a given occasion, are played
by tokens of realizer types. A causal role is just a collection of powers. Hence if
S is a token feature of functional type, then, on any given occasion, every token
power of S will be numerically identical with a power of the base feature P that
plays S’s causal role on that occasion. Witmer (2003) concurs, saying that on a
functionalist approach, “any effect—not just physical ones—to which a mental
property may seem causally related is one which can be explained by reference
to the causal powers of the realizer property” (200).2⁵ Functionalist accounts of
realization thus satisfy the Token Identity of Powers Condition. Do such accounts
also satisfy the Proper Subset of Powers Condition? It is sometimes suggested that
instances of functionally realized features inherit all of the token powers of their
realizing feature instances:
A functional reduction of pain has the following causal and ontological

implications: Each occurrence of pain has the causal powers of its neural
realizer [ . . . ] In general, if M occurs by being realized by N on a given
occasion, the M-instance has the causal powers of the N-instance.
(Kim 2006, 554)
In cases of multiple realizability, however, a functionally realized feature arguably

has only a proper subset of the powers of its realizing feature(s), at both the type
and token levels.
To see this, first recall the hardware/software analogy motivating functionalism,
initially highlighted by Putnam (1967): the realizing systems are similar with
respect to powers needed to implement a given software program, but different
with respect to powers associated with their distinctive varieties of hardware. More
generally, when a type of functionally characterized feature is multiply realizable,
its realizing types will each have all the powers associated with the functional
role, and more besides (where the further powers reflect differences between the
2⁴ The requirement that the proper subset of powers be non-empty reflects the rejection of epiphe-
nomenal features as metaphysically emergent, in the relevant sense.
2⁵ See also Yates 2016: “According to functionalism, mental properties are defined by roles that
are filled by physical properties, so a novel causal role seems out of the question” (4); “Functional
properties are physically realized, and their realizers do all the causal work associated with their
defining roles” (11).
multiple realizers). Hence the powers of the realized type will be a proper subset
of those of each of its realizing types.
Moreover, this proper subset relation between powers of the types will arguably
hold between token powers of instantiations of the types, as per the Proper Subset
of Powers Condition. It may make sense for a token feature to have fewer powers
than its feature type, reflecting restrictions associated with circumstances in which
the feature occurs or is instantiated.2⁶ But as I’ve previously discussed (see Wilson
2011b and 2015b), it makes no sense for a token feature, whether functionally
realized or not, to have more powers than its type; for if a token feature purportedly
of a certain type had token powers not associated with the purported type, that
would be reason for taking the token not to be of the type. So functionally realized
features arguably satisfy the Proper Subset of Powers Condition.2⁷
Constitutive mechanism realization

Second, consider constitutive mechanism accounts of realization. Such accounts
trace back at least to the discussions in Cummins 1975 and 1983, which in
turn draw on explanatory strategies in the sciences aimed at “understanding the
capacities of systems by analyzing them into the capacities of their components”
(Craver 2001, 54), as in the following classic case of a mechanistic explanation:
[T]he circulatory system’s (S’s) capacity to deliver goods to body tissues (W)
is explained by decomposing it into its parts (e.g., hearts (X1 ), arteries (X2 ),
kidneys (X3 ), and valves (X4 )) and capacities (e.g., to pump (φ1 ), to convey
(φ2 ), to filter (φ3 ), and to regulate the direction of blood flow (φ4 )) and linking
those parts together in the programmed φ-ing of the circulatory system.
(Craver 2001, 55–6)
2⁶ Clarke (1999, 307) suggests something in this ballpark, albeit as applied to what he calls ‘exem-
plifications’ of properties in events as opposed to token instantiations of properties.
2⁷ The preceding point serves as my primary response to Morris’s (2013) claim that someone
maintaining that instances of higher-level features inherit all of the token powers of their realizers,
as per what he calls ‘Full Inheritance’, can accommodate the presumption that tokens of a multiply
realized property have the same powers regardless of how they are realized. As Morris puts it:
[T]here is a sense in which the defender of Full Inheritance can accept the causal unity of
multiply realized properties—namely, in the sense that all instances of a property M have
a certain set of causal powers. A functionalist about pain who endorses Full Inheritance
may, for example, allow that all instances of pain have the power to bring about avoidance
behavior. It is just that the defender of Full Inheritance will then contend that insofar as pain
is multiply realized, there may be many other powers that are possessed by some instances
of pain but not others. (207)
But as noted in the text, that some token feature has powers not associated with a given type is good
reason to think that the feature is not of that type. Coupled with the usual supposition that functional
types are associated with comparatively abstract roles, this observation provides independent good
reason to endorse Subset rather than Full inheritance as characteristic of functionally realized features.
Such explanatory strategies have inspired an approach to realization of higher-

level features of a system explicitly encoding reference to the parts of the system
and the specific ways (associated, e.g., with individual ‘role functions’) in which
the parts and their features enter into implementing the higher-level capacity
or function.2⁸ On this approach, a mechanistically realized feature is realized
by a single lower-level feature—sometimes called a ‘micro-structural’ feature—
schematically along lines of the feature of having such-and-such spatiotemporally
related parts engaging in such-and-such causal interactions (exercising such-and-
such capacities, engaging in such-and-such activities). Such an approach is sug-
gested by Craver’s (2001) observation that precisely what is distinctive about
mechanistic approaches is that the parts are engaging in organized as opposed
to merely aggregative behavior: “The components of mechanisms, in contrast to
those of mere aggregates, have an active organization; they act and interact with
one another in such a way that the ψ-ing of [realized feature] S is more than
just the sum of [individual component] φ properties” (59). Kim describes such
a mechanistic account of realization in such terms, as follows:
When P is said to ‘realize’ M in system s, P must specify a microstructural

property of s that provides a causal mechanism for the implementation of
M in s [ . . . ]. (Kim 1993, 197)
As Haug (2010) more recently puts it (expanding on the earlier quotation):
[T]he realizer P must itself provide a causal mechanism for the instantiation
of M—a constitutive mechanism, in which the instantiation of M in s is
explained by the components of s being propertied and related in certain
ways [ . . . ] in a constitutive mechanism, the properties [ . . . ] involved in the
mechanism belong to proper parts or sub-systems of s. [ . . . ] a constitutive
mechanism explains how it is that a given special science property is instan-
tiated [ . . . ] For example, the microstructure of acids (including the crucial
activity of proton donation) provides a constitutive mechanism for being an
acid and is part of an integrative [i.e., nomological] mechanism that can be
used to explain why acids behave the way they do (e.g., dissolve metal, form
a salt when combined with a base, turn litmus paper red). (321)
2⁸ In fact, mechanistic explanations have inspired two different accounts of realization: a one-one
approach, to be next treated, and a many-one approach, of the sort developed by Gillett in his 2002b,
2002a, 2016, and elsewhere. I will treat Gillett’s approach in Ch. 3 (§3.4.2).
A constitutive mechanism account of realization along these lines arguably

ensures satisfaction of the Proper Subset Condition on Powers. One way to see this
is to note that such an approach is a clear variation on the theme of functional real-
ization, with the main difference reflecting that mechanistic approaches, in being
concerned to identify just how a given higher-level functional role is implemented
by lower-level goings-on, explicitly require that the realizing feature encode the
micro-structural and causal mechanistic means by which the role associated with
the higher-level feature is implemented. Notwithstanding this emphasis, it remains
that on this view, realized features are associated with causal roles that, on a given
occasion, are played by tokens of realizing features, and that a causal role is just a
collection of powers. Hence if S is a token feature of a mechanistically realized type,
then, on any given occasion, every token power of S will be numerically identical
with a token power of the micro-structural base feature P that mechanistically
implements S’s causal role on that occasion, as per the Proper Subset Condition on
Powers. And indeed, Haug includes satisfaction of a subset-of-powers condition
in his considered mechanistic account:
I propose the following positive account of the way in which physical prop-
erties realize mental (and other high-level) properties. A physical property
X realizes property Y if and only if (a) the set of causal powers contributed
by Y is a subset of the causal powers contributed by X [ . . . ] and (b) the
components of X provide a constitutive mechanism for the instantiation of
Y, which (c) fits into a family of realizers that, in turn, provides an integrative
mechanism for Y’s causal or nomological relations to other (usually high-
level) properties. (323)
As I read Haug, he does not mean here to suggest that satisfaction of conditions (b)
or (c) does not entail or ensure satisfaction of the Proper Subset of Powers Condition
at issue in (a); rather (as his discussion of alternative accounts of realization
satisfying the latter condition indicates), he simply aims to make explicit the
mechanistic means by which the Proper Subset of Powers Condition is met.
Mereological realization
Third, consider mereological (parthood-based) accounts of realization, according
to which realized features are proper parts of their realizing features at the level of
both types and tokens (Shoemaker 2000/2001 and 2007, and Clapp 2001). Proper
parthood appears to satisfy the nonreductive physicalist’s desiderata: proper parts
are distinct from and yet in a sense completely metaphysically dependent on
(grounded in, constituted by, etc.) a whole that is antecedently given, and the parts
may be efficacious without inducing overdetermination, as when both I and my
eye cause a wink, or both a plane and its wheels are causes of a runway’s being
touched (as in Paul’s 2002 example). In the present case, the suggestion is that a
lower-level physical feature is something like a more fundamental whole of which
a higher-level feature is a less fundamental proper part.2⁹
Both Shoemaker and Clapp suppose that mereological accounts satisfy the
Proper Subset of Powers Condition; indeed, both see satisfaction of the condition
as core to their accounts of realization, with the appeal to mereology serving to
illustrate their preferred strategy for satisfying the condition.
To start, Shoemaker (2000/2001) presents an account of realization based on a
type-level version of the Proper Subset of Powers Condition:
Property X realizes property Y just in case the conditional powers bestowed

by Y are a subset of the conditional powers bestowed by X (and X is not a
conjunctive property having Y as a conjunct). (78)
He then claims that multiply realized feature types satisfy this condition:
Where the realized property is multiply realizable, the conditional powers

bestowed by it will be a proper subset of the sets bestowed by each of the
realizer properties. (78–9)
Shoemaker supports this claim by appeal to considerations similar to those can-

vassed for functional realization, with the main difference being that (following the
causal account of properties advanced in his 1980) he takes all broadly scientific
properties to be essentially characterized by distinctive sets of powers. When such
a feature is multiply realized, he plausibly maintains, its realizing types will share all
the powers of the realized type, but will differ from each other in respect of further
powers. Shoemaker goes on to argue that this proper subset relation between
powers of the types will plausibly hold between token powers of the instantiated
types. In general, Shoemaker (2000/2001) notes, it is reasonable to suppose that if
realized and realizer types are not identical, in virtue (at least in part) of bestowing
different sets of conditional powers, then neither will be their instantiations:
[I]t seems doubtful that we should identify the mental property instance
with the instance of the physical property that realizes it—or that we should
identify the instance of red and the instance of scarlet. If we think of the
instantiation of a property as the conferring on something of the conditional
2⁹ It is important to this strategy that the whole be antecedently given as (more) fundamental.
Were the parts to be antecedently given as (more) fundamental, it might rather be the whole that was
appropriately seen as completely metaphysically dependent on the parts. See Wilson 2014a for further
discussion.
powers associated with that property, then when properties confer different
sets of conditional powers, the instantiation of one of them is not identical
with the instantiation of the other. (28)
These remarks suggest that in cases of multiple mereological realization, and given
that one endorses a causal account of properties, the Proper Subset of Powers
Condition is satisfied.3⁰
Alternatively, one can backwards-engineer the need to satisfy the Proper Subset
of Powers Condition from a mereological approach, independently of whether
one endorses a causal account of properties. In particular: if features have non-
causal aspects, then the supposition that a realized state is a proper part of a
realizing state need not indicate that the realized state has any powers at all (if
the overlap concerns only the noncausal aspect), much less that it is distinctively
efficacious. Hence if proper parthood is to provide a basis for higher-level efficacy,
the overlap must be specifically in respect of powers, as per the Proper Subset of
Powers Condition.31
Determinable-based realization
Fourth, consider accounts of realization in terms of the determinable/determinate
relation, the relation of increased specificity paradigmatically holding between
(determinable) colors and their (more determinate) shades (see Macdonald and
Macdonald 1986 and 1995, Yablo 1992, Ehring 1996, Wilson 2009). Yablo (1992)
expected the suggestion that, e.g., mental features stand to their physical realiza-
tions in the relation that colors bear to their shades to be met with some incredulity.
One way to make his conjecture more plausible is to put the point in terms of the
causal powers of the properties involved (see Wilson 1999 and 2009). Consider a
patch that is red, and more specifically scarlet. Sophie the pigeon, trained to peck
at any red patch, is presented with the patch, and she pecks. The patch’s being red
caused Sophie to peck—after all, she was trained to peck at red patches. But the
patch’s being scarlet also caused Sophie to peck—after all, to be scarlet just is to
3⁰ Similarly for the mereological approach presented in Rueger and McGivern 2010, on which
higher-level behaviors are proper parts of lower-level behaviors: “[W]e can understand the claim
about parthood between behaviours in terms of the subset relation between their associated causal
powers—so that the causal powers of the macro behaviour form a subset of those of the micro
behaviour [. . .]” (395).
31 Importantly, and notwithstanding that Shoemaker and Clapp each take a mereological account
of realization to naturally flow from a causal theory of properties, such an understanding of properties
is not required in order for an account of realization to be seen as ensuring satisfaction of the Proper
Subset of Powers Condition. I’ll return to this issue in Ch. 3, when considering the objection (pressed
by Melnyk 2006, among others) that satisfaction of the conditions in the schema for Weak emergence
will guarantee the physical acceptability of the associated emergent feature only if a causal theory of
properties is assumed.
be red, in a specific way. Nonetheless, Sophie’s pecking was not problematically

overdetermined. Plausibly, this is because each token power of the determinable
red instance is numerically identical to a token power of its determining scarlet
instance. Similarly, the proponent of a determinable-based account of realization
maintains, for the case of S and P, in which case such an account satisfies the Token
Identity of Powers Condition.
Here too it has been suggested that instances of determinable features inherit
all of the token powers of their realizing feature instances, on grounds that
determinable and determinate instances are identical (Macdonald and Macdonald
1995, Ehring 1996). But again, there is a case to be made that the powers of a
determinable feature are a proper subset of those of its determinate features, at
both the type and the token levels.
To start, note that given Sophie’s training, she would have pecked even had the
patch been a different shade of red (burgundy, say); but not so for Sophie’s cousin
Alice, trained to peck only at scarlet patches. This suggests that the determinable
feature type red has fewer powers than its determinate types (scarlet, crimson, etc.).
More generally, since broadly scientific determinables are associated (as per usual:
at least in worlds with the same laws of nature) with distinctive sets of powers, and
may typically be determined by multiple more specific determinates, the powers
of determinable feature types will typically be a proper subset of those of their
determinate feature types. Moreover, this relation will plausibly hold between
token powers of determinable and determinate instances. Again, while it might
make sense for an instance of a given feature type to have fewer powers than
are associated with its type (reflecting restrictions on available circumstances of
instantiation, or the like), it does not make sense for an instance of a given feature
type to have more powers than are associated with the type. In particular, were a
token feature purportedly of a determinable type to have more token powers than
are associated with the type, that would be reason to think that the token was not,
after all, of that type. Hence a determinable/determinate account of realization
arguably satisfies the Proper Subset of Powers Condition.
Ontologically explanatory realization

Finally, consider the aforementioned schematic accounts of realization as involv-
ing an ontologically explanatory necessary connection; here I focus on that offered
in Lepore and Loewer 1989, which draws on Lepore and Loewer 1987 and Horgan
1989. Lepore and Loewer sketch a strategy whereby a feature that is so realized
might be distinctively efficacious with respect to its base feature in virtue of certain
subjunctive conditionals being true:32
32 In the quotations to follow, I’ve corrected Lepore and Loewer’s spelling of ‘quausation’ for
consistency with Horgan’s discussion.
When explaining why one thinks that possession of a [higher-level] property

makes a causal difference one is likely to make reference to certain subjunc-
tive conditionals. [ . . . ] We suggest that [such] conditionals lie at the heart
of the view that [higher-level] properties are causally potent. To apply this
idea to characterizing causal potency we will define a relation between event
property pairs which, after Terry Horgan (1989), we call “quausation” (as in
c qua F causes e qua G). As a tentative account of the quausation relation we
propose the following:
(c, F) is quausally related to (e, G) iff c and e occur and are respectively F
and G and there is some time before the occurrence of c at which these
two conditionals obtain: (1) if c were to occur and be F then that would
cause an event e to be G; (2) if c were to occur but not be an F then it
would not cause an event which is G. (188–9)
Crucially, the proposal eschews the suggestion that distinctive efficacy requires
new powers, on grounds that this route to such efficacy would be at odds with
physicalism. Rather, the distinctive efficacy at issue reflects the sense in which
higher-level features make a difference to whether the effect occurs:
As far as we can see an event’s causal powers are completely determined by

its basic causal properties. [Higher-level] properties are not needed for that.
However, if (c, F) is quausally related to (e, G) then there is a perfectly good
sense in which c’s having F makes a difference to c’s causal powers. This is
the sense captured by the counterfactuals 1 and 2. To say that instantiating F
makes a difference to what c causes is to say, in part, that had c not had F then
it would not have caused a G. Of course, this counterfactual is true in virtue
of certain physical facts obtaining. But that doesn’t make these conditionals
any less true or any less explanatory. (190)
For reasons similar to those associated with a determinable-based approach to

realization, a counterfactual or difference-making approach to distinctive higher-
level efficacy is arguably one entailing satisfaction of the Proper Subset of Powers
Condition.
Summing up: a representative range of seemingly diverse accounts of realiza-
tion are unified in each arguably ensuring satisfaction of the Proper Subset of
Powers Condition as a means of avoiding problematic overdetermination. Other
accounts of realization arguably ensure satisfaction of this condition, as well—
for example (as I argue in Wilson 1999), Pettit’s (1995) account as appealing to
a ‘dot-shape’ analogy, according to which the shape is distinct from, but nothing
over and above, a collection of dots, and (as I argue in Wilson 2010b and will
discuss in detail in Ch. 5) my account in terms of an elimination in degrees of

freedom.
Distinctive power profiles as a basis for ontological and causal autonomy

In having only a proper subset of the token powers of the token physical feature
P upon which it depends, on a given occasion, a token feature S satisfying the
Proper Subset of Powers Condition will clearly be distinct from P, by Leibniz’s Law.
Might S also be causally autonomous—distinctively efficacious—with respect to P,
as required if S is to be genuinely metaphysically emergent?
Yes, supposing that a feature may be distinctively efficacious in virtue of having
a distinctive set or collection of causal powers—i.e., a distinctive power profile. The
underlying promise of nonreductive physicalism as making sense of the distinctive
efficacy of (at least some) special-science (more generally, higher-level) entities and
features lies in the claim that S’s causal autonomy does not require that S have a
distinctive power: it is enough that S have a distinctive set (collection, plurality) of
powers.
One case for taking the having of a distinctive power profile to be sufficient for
causal autonomy adverts to the sort of difference-making or other ‘proportionality’
considerations operative in cases where S (or S’s type) is multiply realizable
(following discussion in, e.g., Lepore and Loewer 1987 and 1989, Horgan 1989,
and Yablo 1992).33 Again, suppose that S is a state of feeling thirsty, which causes
a physical reaching for a glass of water (effect E). Now suppose that S (or another
token of S’s type, etc.) were realized by P′ rather than P, in circumstances relevantly
similar to those in which S caused E. Would E (or an event of E’s type) have still
occurred? Intuitively, yes, since the only powers that matter for the production
of E are the powers associated with S: powers differing between P and P′ (e.g., to
produce different readings on a neuron detector) don’t make a difference to, and
hence are in this sense irrelevant for, E’s production. That S’s distinctive power
profile contains just those powers relevant or ‘proportional’ to E’s production
provides a principled reason for taking S’s efficacy vis-à-vis E to be distinctively
different from P’s, notwithstanding that (as per S’s satisfaction of the Token Identity
33 Yablo suggests that a candidate determinable cause (e.g., the patch’s being red) might be more
proportional to a given effect (e.g., Sophie’s pecking), on a given occasion, than the associated candidate
determinate cause (e.g., the patch’s being scarlet), in that the determinable has an ‘essence’ tracking
both sufficiency and difference-making considerations (e.g., if the patch had been crimson rather
than scarlet, Sophie would still have pecked). Here I focus on difference-making considerations, since
(against the background assumption of Physical Causal Closure and associated satisfaction of the Token
Identity of Powers condition, in particular) it is difference-making rather than sufficiency which is
distinctive of higher-level efficacy, on a ‘proper subset of powers’ approach to realization.
and Proper Subset conditions on token powers), S doesn’t cause anything that P (or
other realizers of S’s type, on other occasions) doesn’t (or don’t) also cause.3⁴
Another case for causal autonomy reflects that distinctive power profiles are
typically associated with distinctive systems of laws—for example, the special-
science laws treating entities of S’s type. Plausibly, systems of laws track causal
joints in nature; hence when S is of a special-science type, its distinctive power
profile is similarly plausibly understood as tracking such a distinctive causal joint.
Moreover, given the holding of the Proper Subset of Powers Condition, and conso-
nant with a general understanding of special-science phenomena as abstracting
away from various lower-level details, the causal joints at issue here concern
comparatively abstract goings-on. This is, I think, what Antony and Levine (1997)
have in mind when they say that for the causal autonomy of functional properties,
“What we really need is a ‘realization-indifferent’ regularity: a contingent regularity
that essentially involves the [realized] property, and that applies to any instance
of the property, no matter the form of realization” (92). Moreover, as Antony
(2003) notes, even in the absence of multiple realizers, one can still make sense
of the presence of properties and laws that are, as she puts it “at a higher level of
abstraction” (8, emphasis in text) relative to lower-level properties and laws:
[M]ultiple realizability is something of a red herring. What matters, funda-

mentally, is not whether there could be minds embodied in things other
than brains, but rather whether there is a level of reality beyond the level at
which brains are normally studied—whether psychological kinds are “really
there” [in addition to] the already recognized kinds in chemistry, biology
and the other established sciences. If this is what is at stake, then it would not
matter if brains turned out to be the only kinds of things that realize minds
in any nomologically possible worlds. [ . . . ] The functional descriptions, and
the generalizations given in terms of the psychological categories defined
at the functional level would still, in this case, be autonomous from the
descriptions, generalizations and categories that turned up at the level of the
realizers. (8)
This line of thought seems right to me. Taking it on board, one might also say that
while multiple realizability is a good indicator of when a comparatively abstract
ontological and causal joint is in place, that there is such a comparatively abstract
joint does not hinge on multiple realizability. Also worth noting is that causal
joints may overlap—both in respect of a given token power and in respect of an
3⁴ Note that nothing in this line of thought requires that one accept a ‘difference-making’ or
counterfactual account of causation, or relatedly, that one reject P as being a cause of E. The suggestion
is simply that attention to difference-making considerations provides a principled basis for S’s being
distinctively efficacious with respect to P. I’ll return to this issue in Ch. 3.
associated effect E. If the joints as a whole are different, this provides a principled
reason for taking a realized feature S to be distinctively efficacious vis-à-vis E, in
that S produces E as part of a different system of laws (different causal joint) than S’s
realizer, P. Relatedly, insofar as causal relations involve exercises of a given power,
this appears to be what Macdonald and Macdonald (1995) have in mind when they
say that the causal autonomy of realized mental features reflects that “any instance
of a cause-effect relation can be an instance of more than one pattern” (71).
The key suggestion operative in the nonreductive physicalist’s approach is that
there are two ways for a higher-level feature to be distinctively efficacious with
respect to the lower-level feature(s) upon which it depends. One way, emphasized
by Kim and others, is for the higher-level feature to be associated with a new
power to produce the effect; here the distinctive efficacy (characteristic of accounts
of Strong emergence) is located in the having of a distinctive (fundamentally
novel) power. Another way—that at issue in the powers-based subset strategy,
and characteristic of accounts of nonreductive physicalism/Weak emergence—
is for the higher-level feature to be associated with a distinctive subset of powers
that are relevantly proportional to the effect, in the ways indicated by difference-
making considerations and comparatively abstract special-science laws or (more
generally) causal joints. In short: distinctive efficacy may reflect either the having
of a distinctive power or the having of a distinctive power profile, and it is the latter
variety of efficacy that is at issue in accounts of physical realization.
The nonreductive physicalist response to the problem of higher-level

causation
Let’s now see in more detail how satisfaction of the Proper Subset of Powers Condi-
tion enters into the nonreductive physicalist’s response to the problem of higher-
level causation, and in particular into their rejection of Non-overdetermination
(according to which, with the exception of double-rock-throw cases, effects are not
causally overdetermined by distinct individually sufficient cotemporal causes).
In case 1, special-science feature S again depends on base feature P, and S causes
another special-science feature S*, which depends on base feature P*; as previously,
we might suppose that S is a state of feeling thirsty, and S* is a desire to reach for a
nearby glass of water. Here the nonreductive physicalist’s strategy involves, to start,
the supposition that S satisfies the Proper Subset of Powers Condition in such a way
as to have the power, on a given occasion, to bring about S*. As per the condition,
this token power is identical to one had by P; hence when S causes S*, so too does
P. S and P are each sufficient causes of S*; they are distinct, by Leibniz’s Law, since
S, in satisfying the Proper Subset of Powers Condition, has fewer token powers than
P; and since S and P are cotemporal, they are not parts of a diachronic causal
chain. Consequently, S* is overdetermined, contra Non-overdetermination. Yet,
the nonreductive physicalist maintains, the overdetermination here is not of the
double-rock-throw variety that is supposed to be problematic (since implausible).
In a double-rock-throw case, distinct token powers and associated causal chains

converge on a single effect (a window breaking). In cases where the higher-level
cause satisfies the Proper Subset of Powers Condition, however, two distinct token
features standing in an intimate dependence relation are each associated with the
same token power and hence the same causing. The overdetermination here is
benign—indeed, is no more problematic than in cases where, e.g., both a plane
and its wheels are causes of a runway’s being touched.3⁵ Hence it is that Non-
overdetermination, at least in full generality, must be rejected, leaving the way clear
for higher-level efficacy. Here it’s worth representing the features at issue as having
overlapping sets of powers, with each power represented as a dot, as in Figure 2.5:
S*
S
P P*
Figure 2.5 The nonreductive physicalist’s response to case 1
In case 2, special-science feature S rather causes a base feature P*. In the first
instance, the treatment of this case, shown in Figure 2.6, is a variation on the same
theme:
P P*
Figure 2.6 The nonreductive physicalist’s response to case 2, version 1
3⁵ Note that, contra a guiding supposition of Morris 2011, the motivations for thinking that
satisfaction of the Proper Subset of Powers Condition blocks problematic overdetermination do not
hinge on this sort of illustrative analogy to cases of benign part/whole overdetermination. Rather, that
the overdetermination is benign follows just from the fact that, if the condition is satisfied, only one
token power is manifested on the occasion in question, in which case the overdetermination here is
nothing like that at issue in double-rock-throw cases.
That said, there is a subtlety here that must be addressed, associated with the
possibility that no other lower-level feature besides P can, in the relevant circum-
stances, cause P*. In this case, the supposed distinctive autonomy of S cannot rely
on difference-making considerations, according to which if S had been realized by
some lower-level physical property other than P, S would still have caused P*. Nor
can S’s distinctive efficacy vis-à-vis P* be a matter of S and P* mutually occupying
a distinctively abstract system of laws or causal joint, since P* is, by assumption, a
lower-level physical feature.
In such a case, we can still accommodate the seeming efficacy of S vis-à-
vis physical goings-on by taking these appearances to concern, not lower-level
goings-on (at, e.g., the quantum level), but rather what we might call ‘physically
acceptable’ goings-on P′ at some level lower than that at which S is properly
located, but higher than that at which S’s ultimate lower-level physical realizer P is
located. For example, the nonreductive physicalist can accommodate mental state
S’s being distinctively efficacious vis-à-vis some sort of physical behavior P′ —say, a
reaching for a glass—which is also realized by P*, but for which difference-making
considerations vis-à-vis S would be present. In this case, a more accurate picture
would be as shown in Figure 2.7:
S P'
P P*
Figure 2.7 The nonreductive physicalist’s response to case 2, version 2
Indeed, irrespective of whether P is the only lower-level physical feature capable

of causing P*, this sort of model might more generally make better sense of
the seeming efficacy of special-science features vis-à-vis physically acceptable
goings-on.
It is also worth noting that the strategy here is compatible with physicalism. The
main concern about S’s physical acceptability, given its reality, cotemporal material
dependence on, and distinctness from its dependence base feature P, turns on the
possibility that S might be Strongly emergent from P—that is, that S might have,
as per the New Power Condition, a (fundamentally novel) power not had by P, of
the sort undermining Physical Causal Closure. But S’s satisfaction of the Proper
Subset of Powers Condition is incompatible with S’s satisfaction of the New Power
Condition, ruling out S’s being Strongly emergent.3⁶
2.3.2 The schema for Weak emergence
Prima facie, satisfaction of the Proper Subset of Powers Condition by a special-

science feature S which cotemporally materially depends on a base feature P
provides an appropriate and illuminating basis for avoiding overdetermination
while guaranteeing that S is both ontologically and causally autonomous with
respect to P. Satisfaction of this condition guarantees that S is ontologically
autonomous (distinct) from P: since S has a proper subset of the token powers
of P, S is distinct from P, by Leibniz’s Law. It moreover makes room for S to
be causally autonomous (distinctively efficacious) with respect to P, in virtue of
S’s having a distinctive power profile, tracking difference-making/proportionality
considerations or a comparatively abstract system of laws or (more generally, and
compatible with the distinctive efficacy of artifacts) causal joint.
We have thus arrived at our second schema for metaphysical emergence:
Weak Emergence: What it is for token feature S to be Weakly metaphysically

has a non-empty proper subset of the token powers had by P.
Here again, the locution ‘what it is for’ is intended to flag that Weak Emergence
provides a schematic metaphysical basis for a given case of such emergence,
encoding what is core and crucial to that notion. Of course, any schematic
account needs to be sensibly filled in. Modulo this caveat, the conditions in the
schema are, I claim, necessary and sufficient for (appropriate and illuminating
accommodation of) metaphysical emergence of the physically acceptable variety.3⁷
Some clarifications:
3⁶ As discussed in Wilson 1999, satisfaction of the Proper Subset of Powers Condition also appears
to block other live routes to physical unacceptability, associated with S’s being non-natural (see
Moore 1903) or supernatural (as per Malebranchean occasionalism). As it happens, Moore used the
term ‘non-natural’ as indicative of epistemological irreducibility (more specifically: indefinability),
which is arguably compatible with physicalism (see Wilson 2002a). But supposing that S’s being
epistemologically irreducible is deemed naturalistically (hence physically) problematic, this must be
because such irreducibility indicates that S’s existence involves something metaphysically new relative
to (‘over and above’) P; on the nonreductive physicalist’s operative assumption that S is efficacious, the
problematic addition in question would presumably either be or entail S’s having of a non-natural or
supernatural causal power, not had by P. But if the Proper Subset of Powers Condition is satisfied, the
having of such a power is ruled out.
3⁷ Here again, this is not to presuppose that all implementations of Weak Emergence are on a par, or
equally well suited to accommodate a given macro-phenomenon. For example, Bealer (1997) argues
that functional realization cannot properly accommodate self-consciousness, and Ehring (1996) and
others argue that certain features of the determinable/determinate relation render it unsuited as an
account of the realization of mental states (though see my responses in Ch. 7, §7.2.3). And as we’ll see
in Ch. 3, certain implementations of Weak Emergence have distinctive advantages.
• Again, the first condition minimally specifies cotemporal material depen-

dence, understood as involving both (physical) substance monism and the
minimal nomological supervenience of emergent feature types on base fea-
ture types.
• The second condition (effectively, the Proper Subset of Powers Condition)
captures the comparatively weak sense in which an emergent feature may be
both ontologically and causally autonomous with respect to its base feature,
in virtue of having a proper subset of the token powers of the base feature.
Note that while the initial presentation of the Proper Subset of Powers Con-
dition built in the cotemporal material dependence of the feature having the
proper subset of powers on its base feature, the schema for Weak Emergence
treats the proper subset of powers and dependence conditions separately.
Henceforth, references to the Proper Subset of Powers Condition should be
understood as references just to condition (ii) in that schema.
• Again, the base feature at issue is a feature of a given (ultimately physical
micro-) configuration (plurality or structural aggregate), and the overlap
in powers at issue in condition (ii) should be understood accordingly. In
particular, this overlap in powers is not to be (implausibly) understood as
involving overlap of powers of a macro-entity with powers of features of
individual entities entering into the configuration upon which the macro-
entity cotemporally materially depends.
• Again, the schema is relativized to occasions (times or temporal intervals),
but it is reasonable to suppose that (given that S’s type is not Strongly emer-
gent) it suffices for the Weak emergence of S, simpliciter, that the condition
is ever satisfied, and to suppose that it suffices for the Weak emergence of the
feature type (of which S is a token), simpliciter, that any token feature S on
any occasion satisfies (or would satisfy) the condition. These complications
won’t play a role in what follows.
2.4 The schemas for Strong and Weak emergence as core

and crucial to metaphysical emergence
Let’s sum up the results so far. Attention to the problem of higher-level causation
points towards two strategies of response to this problem, associated with Strong
emergentism and with nonreductive physicalism, respectively, each of which pro-
vides an appropriate and illuminating basis for accommodating the metaphysical
emergence of higher-level entities and features, understood as coupling cotem-
poral material dependence with ontological and causal autonomy (that is, with
distinctness and distinctive efficacy). Each response is associated with a specific
condition on the token powers of a higher-level feature with respect to the token
powers of the lower-level feature upon which it depends on a given occasion, which
condition, along with a minimally specified condition on cotemporal material

dependence, is encoded in the associated schema for Strong or Weak emergence:

has at least one token power not identical with any token power of P.
has a non-empty proper subset of the token powers had by P.
Indeed, attention to these responses makes clear the limited ways in which a
cotemporally materially dependent higher-level feature can be causally, hence
ontologically, autonomous with respect to its base feature, as the operative concep-
tion of metaphysical emergence requires. First, the feature may have more powers
than its base feature; second, the feature may have fewer powers than its base
feature. In terms of effects: the higher-level feature may be distinctively efficacious
in potentially contributing to causing more effects than its base feature, or it may
be distinctively efficacious in potentially contributing to fewer effects than its base
feature. Since complete coincidence of token powers doesn’t make room for causal
autonomy (distinctive efficacy), these routes to metaphysical emergence exhaust
the available options.
I conclude that satisfaction of the conditions in either schema is, as I put it, ‘core
and crucial’ to metaphysical emergence of the sort relevant to realistically vindi-
cating the seeming appearances of emergence as pertaining to special-scientific
and artifactual entities and features. Modulo the supposition that the schemas are
sensibly filled in, the results of this chapter can be seen as providing prima facie
reason to think that the conditions in the schemas are both necessary and sufficient
for (appropriate and illuminating accommodation of) metaphysical emergence of
both physically acceptable and physically unacceptable varieties—a bold claim, but
one that, as I argue in ensuing chapters, is surprisingly robust.
We thus have a provisional schematic answer—or rather, two answers—to the
first key question, ‘What is metaphysical emergence’? The answers are schematic,
since as we have already seen and will see further in the chapters to follow, there are
several ways in which either schema might be (or might aim to be) implemented.
The answers are also at this point provisional, since a number of objections have
been raised to the viability of physically acceptable or physically unacceptable
emergence, either in general or as specifically directed at the schemas for Weak or
Strong emergence or the conditions on powers therein. In the next two chapters,
I’ll consider these objections, and show that they can be answered.
3
The viability of Weak emergence
In Ch. 2, I provided prima facie reasons for thinking that satisfaction of the
conditions in the schema for Weak emergence is core and crucial to a feature’s
being metaphysically emergent from—cotemporally materially dependent on, yet
ontologically and causally autonomous with respect to—a lower-level feature,
in a way compatible with physicalism, given that the base feature is physically
acceptable. Again, the schema is as follows:

that occasion, (i) that S cotemporally materially depends on P, and (ii) that S has
a non-empty proper subset of the token powers had by P.
Henceforth, by ‘Weak emergence’ I will intend to refer to the view (or the
phenomenon, understood in light of the view) that conformity to the conditions in
Weak Emergence is core and crucial—and when sensibly filled-in, both necessary
and sufficient—for physically acceptable emergence (i.e., realization).
In this chapter, I consider and respond to a representative range of objec-
tions that have been or could be made to the viability of Weak emergence, so
understood. These objections fall into four main categories, according to which
satisfaction of the conditions in Weak Emergence is compatible with anti-realism
about higher-level features (§3.1), is compatible with reductionism about higher-
level features (§3.2), is compatible with the emergent feature’s being physically
unacceptable (§3.3), or is not necessary for metaphysical emergence of a physically
acceptable variety (§3.4). The primary focus of many of the objections is on
condition (ii) in the schema—i.e., the Proper Subset of Powers Condition. As we’ll
see, each of these objections admits of at least one response that could be endorsed
by any proponent of Weak emergence, whatever their preferred implementation
of the schema. Upon occasion, however, a response to a given objection is available
that relies on a specific implementation of the schema. In particular, certain attrac-
tive responses appeal to either a determinable-based account of Weak emergence
or a degree of freedom-based (DOF-based) account of such emergence, of the sorts
I have previously endorsed.
DOI: 10.1093/oso/9780198823742.003.0003
3.1 Objection: compatibility with anti-realism
Certain philosophers, who seem to me to be properly deemed anti-realists, main-

tain that seeming satisfaction of the Proper Subset of Powers Condition reflects
not an independently existing higher-level feature, but rather a (mere) mind-
dependent abstraction from, or pragmatically motivated way of conceiving of,
lower-level physical goings-on.
For example, Ney (2010) argues, after discussing how Shoemaker (2000/2001,
2003, and 2007) takes satisfaction of the Proper Subset of Powers Condition to
serve as the basis for realization, that an alternative way of understanding seeming
satisfaction of this condition is available:
[One] can […] think of things in the following way (nothing has been said
to rule out this way of thinking): on this view, really the mental event (and
realized tokens more generally) are just abstractions from concrete microphysical
situations. They are abstractions in the sense that they are what we attend to when
we focus only on a proper subset of a microphysical state’s causal powers. (442)
Ney sees this reading as indirectly supported by Shoemaker’s (2007) claim that
realization relations form a hierarchy, where “those higher in such a hierarchy will
be realized by those further down” (23), and where properties whose instantiations
lie at the bottom of the hierarchy are, unlike other properties, “self-constituted”. As
she says:
Plausibly, [one] can tolerate the existence of entities that are not reducible, i.e.,
identical to, physical entities, so long as these things are mere abstractions.
Since realized events can very naturally be seen on Shoemaker’s account as mere
abstractions from their realizer events, since they are not self-constituted, I don’t
see why [one] cannot endorse this approach. (443)
On such an interpretation, “The only entities with genuine, (mind-)independent

existence here are the microphysical states of affairs” (444). Though there might be
a sense in which, on Ney’s interpretation, mental states understood as ‘mere’ mind-
dependent abstractions exist, in the sense relevant to our investigations here this
interpretation is properly deemed ‘anti-realist’.1
1 Ney herself presents such abstractionism as compatible with reductionism, but in my view this
muddies the terminological waters, insofar as standardly, and as I am understanding it here (and as Ney
herself suggests in the previous quotation), ontological reductionism entails that the reduced goings-
on are identical to the reducing goings-on. Similar remarks apply to Heil’s view (to be next discussed),
which he sometimes presents as being compatible with a form of realism about higher-level features.
the viability of weak emergence 77
Heil (2003b) also argues for a kind of anti-realist conceptualism about higher-
level features. As Heil sees it, embrace of nonreductionism reflects an uncritical
tendency for philosophers and others to suppose that “we can ‘read off ’ features
of reality from our ways of speaking about it” (207), in a way that involves the
following ‘Picture Theory’, whereby representation always corresponds to reality:
The Picture Theory: When a predicate applies truly to an object, it does so in virtue
of designating a property possessed by that object and by every object to which
the predicate truly applies (or would apply). (210)
By way of illustration of what is problematic about this principle, Heil argues

that notwithstanding that the predicate ‘is red’ truly applies to some objects—e.g.,
tomatoes, stoplights, apples—it would be a mistake to uncritically suppose that
this is in virtue of these objects’ possessing the “very same property”. Rather, he
suggests, one may take the predicate and associated concept to here be serving the
broadly pragmatic function of enabling us to categorize objects that are inexactly
similar:2
The concept expressed by the predicate ‘is red’ […] seems tailor-made for picking
out a range of objects that are, in a particular way, less-than-perfectly-similar to
one another. The concept applies to objects by virtue of properties possessed by
those objects, presumably an extremely complex and diverse class of physical
properties. There is, I gather, no prospect of defining or analyzing redness in
terms of these physical properties. This is due, in some measure, to the fact that
the properties in question are salient—to us—partly owing to the nature of our
perceptual system. Were we built differently, were we made of different materials,
the diverse collection of properties that satisfy our concept of redness could well
fail to stand out. In that case we should have no use for the concept. (215)
The same holds more generally, Heil continues, for special-science predicates such
as ‘is a tree’, ‘is a planet’, or ‘is in pain’. Rather than taking diverse applications of a
given such predicate as indicating the existence of a common property—whether
reducible or irreducible, no matter—we should see such applications as reflecting a
merely conceptual means of tracking inexact similarity among lower-level physical
properties.
2 Berkeley (1710) argues similarly against the assumption that general terms denote abstract ideas:
“[I]t is thought that every name has, or ought to have, only one precise and settled signification, which
inclines men to think there are certain abstract, determinate ideas, which constitute the true and only
immediate signification of each general name […] whereas, in truth, there is no such thing as one
precise and definite signification annexed to any general name, they all signifying indifferently a great
number of particular ideas” (§18).
Heil here endorses (what is reasonably described as) a form of anti-realism

about higher-level features, at least in cases of seeming multiple realizability. Such
cases are frequently appealed to in support of the holding of the Proper Subset of
Powers Condition (with the powers associated with the higher-level feature being
those in the intersection of the sets associated with its multiple realizers); hence
Heil’s deflationary conceptualism constitutes an objection to the viability of Weak
emergence, to the effect that insofar as the condition on powers appears to be
satisfied, this has at best pragmatic implications.
My response to the anti-realist begins by first granting that, as Ney puts it,
“nothing has been said to rule out” an abstractionist or pragmatist line on seeming
satisfaction of the Proper Subset of Powers Condition; but second, denying that the
viability of Weak emergence hinges on accomplishing such a ‘ruling out’. As per the
operative methodological criterion of appropriate accommodation (Ch. 1, §1.4.5),
I take the natural (straightforward, default) accommodation of the appearances of
metaphysical emergence to be a heavily weighted desideratum—more weighty, in
particular, than any antecedent parsimony-based motivation, according to which
one should strive to minimize the number of ontological posits invoked in a
metaphysical account of some phenomenon, independent of considerations of the
usual understanding of the phenomenon, and in the absence of any identified
problems with such posits. Correspondingly, to the extent that the appearances
support the posit of metaphysically emergent features, the burden is on the anti-
realist to provide reasons for not taking the appearances at face value.3 As I’ll
now argue, however, neither Ney nor Heil provides any such good reason—in
particular, as telling against a Weak emergentist treatment of the appearances.⁴
To start, one might wonder whether Heil’s ‘debunking’ explanation of moti-
vations for nonreduction, according to which these motivations really reflect an
unreflective inclination to take any true predication as tracking a genuine property,
would count as undermining the operative presumption of the criterion of appro-
priate accommodation. By way of brushclearing, it is worth noting that the focus
on predicates is not really to the point, since predicates or concepts do not after
3 This is even more true in the present case, where the considerations offering prima facie support
for there being metaphysical emergence are very extensive, drawn from the special sciences, perception,
linguistics, and ordinary practices of individuation, among others. I doubt that all these phenomena
would admit of anti-realist treatment, but even if some such treatments could be cooked up, they would
be to some extent revisionary. For example, on Heil’s approach, special-science laws which appear to
make no reference to lower-level entities and features would turn out to really be about such entities
and features (and their inexact similarities, about which more shortly).
⁴ Morris (2018) offers a developed ‘argument by cases’ for the conclusion that physicalism must take
the form of a ‘one-level’ anti-realist (or reductionist; as in Ney and Heil’s discussions, the distinction
here is somewhat fluid) view similar to that endorsed by Ney and Heil, which proceeds by raising
concerns for a range of accounts of realization, including what Morris calls ‘subset realization’. In this
section I’ll mention two salient concerns raised by Morris (one drawn from his 2013); what I take to
be his most pressing concerns, however, are best treated in §3.2.3 and §3.3.1.
all appear from nowhere, to be interpreted (or misinterpreted, as the case may be)
as referring to genuine features, but are rather typically introduced into scientific
and ordinary discourse in response to a felt need to refer to features that we have
some or other independent reason to think exist. Heil’s concern, more properly
understood, is that any seeming motivations for positing higher-level or more
specifically metaphysically emergent entities or features are better interpreted
simply as tracking inexact similarities between lower-level entities or features.
This ‘debunking’ account of the seeming motivations for metaphysical emer-
gence is unsuccessful, however. For even granting, e.g., that seeming attributions
of higher-level features were in some sense tracking inexact similarities among
lower-level features, this would not show that the higher-level features did not
exist, unless it was antecedently clear that the inexact similarities at issue were
not themselves higher-level. But this is not antecedently clear; on the contrary,
someone aiming to take the appearances at realistic face value might happily grant
that higher-level features are in some sense tracking inexact similarities between
lower-level features, but maintain that such inexact similarities are themselves
metaphysically emergent features of reality. Upon closer examination, then, nei-
ther considerations pertaining to the Picture Theory nor considerations pertaining
to the connection between higher-level feature attributions and inexact lower-level
similarities provide independent reason to reject the prima facie appearances of
higher-level reality, beyond the background methodological supposition (which
we are here rejecting) according to which ontological parsimony is weighted more
heavily than a natural accommodation of the appearances.
Heil’s anti-realism is also motivated by Kim-style concerns that higher-
level properties are (on pain of problematic causal overdetermination) causally
excluded by their lower-level realizers:
I need not remind you of difficulties a levels conception breeds. Consider just the
problem of the ‘causal relevance’ of higher-level properties. Suppose that mental
properties are higher-level properties realized in the nervous systems of sentient
creatures. How could such properties affect the behavior of creatures possessing
them? The potential causal contribution of any higher-level property would seem
to be preempted by its lower-level realizing property. […] This worry about the
causal relevance of mental properties extends smoothly to higher-level properties
generally. If you like to think of the special sciences as occupied with higher-
level properties and events, then you will need some accounting of how these
properties and events could make a causal difference in our world.
(Heil 2003b, 213)
In Heil’s remarks, we again see operative the assumption, following Kim, that the
only way for a feature to be distinctively efficacious is by its having a distinctive
power. But as previously, the Weak emergentist has provided an “accounting of

how [higher-level] properties and events could make a causal difference in our
world”, in spite of not having any distinctive powers—namely, by having distinctive
power profiles, reflecting either causal difference-making considerations and/or
comparatively abstract, metaphysically real, systems of laws or causal joints. What
is needed at this point, then, is anti-realist reason to think that the strategy of
locating distinctive efficacy in the having of distinctive power profiles is somehow
problematic.
Ney offers such a reason, but in focusing solely on Shoemaker’s discussion, her
objection misses the mark. After presenting her mental abstractionist alternative,
she says, on Shoemaker’s behalf:
But maybe this is too quick. At times Shoemaker has tended to suggest that in
cases of mental causation, it is only the mental event [S] that is efficacious with
respect to the effect. In Physical Realization [Shoemaker 2007], he makes the
slightly weaker but similar suggestion that in cases of mental causation, only the
mental event is directly efficacious, with the physical or microphysical realizer
events being efficacious only in virtue of containing the causal powers of the
mental event as parts (53). My speculation is that the reason Shoemaker makes
claims like these is in order to emphasize that he is really providing a nonreductive
account of mental causation […] The goal is to do this by securing some kind
of distinctive causal efficacy for the mental, claiming that in cases of mental
causation, it is only mental events (not underlying physical or microphysical
events) that are efficacious, or that in cases of mental causation, it is only mental
events that are directly efficacious.⁵ (443)
Ney goes on to say that, insofar as every power of a mental feature on a given
occasion is supposed to be identical with a power of its physical realizer, and
insofar as satisfaction of this token identity condition is crucial to preventing
overdetermination, “it doesn’t make sense to say that the realizers are only indi-
rectly efficacious vis-à-vis the efficacy of the mental event. There is only one causal
relation here […] there is no reason to try to secure any distinctive causal efficacy
for mental events. Assuming we are all physicalists, the challenge is to secure
nonredundant causal efficacy for mental events, not causal distinctiveness” (444).
Hence, she concludes, Shoemaker’s strategy for securing the distinctive efficacy of
higher-level features, via satisfaction of the Proper Subset of Powers Condition, fails.
⁵ Morris (2011) similarly suggests “perhaps we could take Shoemaker to be saying that […] when
a property M is subset realized by a property P on some occasion, and M and P seem to overdetermine
an effect, in fact it is the M-instance that is the ‘real’ cause of that effect” (370).
It is true that Shoemaker has several times suggested that a higher-level feature
satisfying the Proper Subset of Powers Condition might be appropriately deemed
‘the’ cause of a given effect. For example, in his (2000/2001), Shoemaker says
[W]here only the causal features of a property P that play a role in producing
an effect are ones that belong to a property M, of which P is a […] realizer
property, there seems a good sense in which considerations of proportionality
[including sufficiency and difference-making] favor the instantiation of M over
the instantiation of P as the cause of the effect. (436)
In his (2003), he confirms:
One advantage of this approach is that it provides a basis for saying that in some
cases it is the instantiation of a mental property, rather than the instantiation
of one of its realizer properties, that caused a certain effect, or contributed to
causing it. (3)
But here, in my view, Shoemaker’s discussion goes awry. I agree with Ney that
the supposition that a Weakly emergent feature might be appropriately deemed
‘the’ cause (or a ‘more direct’ cause) of a given effect is problematic: she is right
that it doesn’t make sense to deny that the lower-level realizer is a cause (and
moreover a direct cause) of the effect, given that the supposition that the Proper
Subset of Powers Condition is satisfied entails that the realizer has the power to
produce (or contribute to producing) the effect, and this power is manifested (or
the associated regularity instanced, etc.) on the occasion in question. Moreover, as
I note in Wilson 1999, 2011b, and elsewhere, the supposition that a higher-level
feature causes something that its base feature does not cause is, for reasons that
Kim’s causal overdetermination concern brings out, in tension with physicalism
and the associated supposition of Physical Causal Closure, whether the purported
effect is a higher-level special-science feature or a lower-level physically acceptable
feature.
There is, however, no need to follow Shoemaker in suggesting that realized
features can cause effects that their realizing features don’t cause, since what the
Weak emergentist requires is not that higher-level features be uniquely effica-
cious but just that they be distinctively efficacious. For similar reasons, the Weak
emergentist can and should reject Yablo’s (1992) claim that difference-making
and other proportionality considerations can support identifying a determinable
feature (e.g., red) rather than an associated determinate feature (e.g., scarlet) as the
cause of a given effect (e.g., the pecking of a pigeon at a scarlet patch), as well as List
and Menzies’ (2009) account of higher-level autonomy as involving ‘downwards
exclusion’:
If it is correct that realization-insensitivity is a general requirement in higher-

level causal claims, then it follows that the conditions for downwards exclusion
are generally satisfied. But this in turn entails that higher-level causal relations
such as that between [realized] M and [effect] B obtain even though there is no
underlying causal relation between the neural realizer N and B. In such cases, we
have good reason to believe in the causal autonomy of higher-level properties.⁶
(499)
Once it is seen that distinctive efficacy can be achieved either via the exercise of
a distinctive power or via a distinctive power profile (and associated difference-
making and/or abstract causal joint/law considerations), one can allow that both
realizing and realized features are causes—distinctive causes—of a given effect: one
can have one’s physicalism and one’s distinctive higher-level efficacy, too.⁷
Having gotten clear about the strategy operative in Weak emergence for achiev-
ing distinctive efficacy of higher-level features, one can also see that Ney’s dis-
cussion fails to accurately take this strategy into account. This distinctive efficacy
is not (even for Shoemaker, notwithstanding certain of his remarks’ suggesting
otherwise) a matter of a higher-level feature’s being able to cause things that its
realizer(s) cannot cause; rather, it is a matter of its profile’s being such as to track
relevant difference-making considerations, or comparatively abstract systems of
laws or causal joints. Relatedly, Ney is wrong to claim that for physicalists “the
challenge is to secure nonredundant causal efficacy for mental events, not causal
distinctiveness”, for ensuring the former without ensuring the latter will still fail
to accommodate the prima facie appearances of emergence, contra our guiding
methodology, and relatedly, will lead to the need to drastically revise our under-
⁶ It is worth noting that even if one endorses a response to Kim-style overdetermination concerns
appealing not just to difference-making considerations (as per, e.g., Horgan 1989 and Lepore and
Loewer 1987 and 1989; see also Bennett 2003) but moreover to a difference-making or counterfactual
account of causation (as List and Menzies do), one might not be committed to gaining the distinctive
efficacy of a higher-level feature vis-à-vis a given effect at the cost of denying that its physical
dependence base feature is also a cause of the effect (contra Physical Causal Closure); for one may
rather suppose that higher-level and lower-level features are each causes, relative to different contrast
classes (see, e.g., Hitchcock 1996.)
⁷ My response to Ney applies, mutatis mutandis, to the ‘symmetry’ concern raised in Morris (2013),
according to which attention to causal implications of cases of the determinable/determinate relation
fails to indicate that higher-level features, on a determinable-based account of realization, will satisfy
the Proper Subset Condition on Powers. Morris correctly observes that if (as both Yablo 1992 and
Shoemaker 2000/2001 suggest, by appeal to proportionality considerations), e.g., the patch’s being red
is the best candidate for being the cause of Sophie’s pecking and the patch’s being scarlet is the best
candidate for being the cause of Alice’s pecking, then we have symmetry here, with each feature having
at least one power that the other doesn’t have, contra the supposed satisfaction of the Proper Subset of
Powers Condition. Like Ney, Morris assumes that in the cases at hand, the assignment of a power to
an instance of a feature depends on taking the feature to be the cause of the effect in question: red but
not scarlet causes Sophie to peck, while scarlet but not red causes Alice to peck. Morris is correct that
this reading of the cases is problematic, but as per the discussion in the main text this reading can and
should be rejected.
standing of the truth and/or content of a large range of scientific and ordinary
beliefs.
Unlike Ney, Morris (2018) registers the strategy of appealing to difference-
making considerations as a means of gaining the distinctive efficacy of subset-
realized features in a way compatible with nonreductive physicalism/Weak emer-
gentism, but maintains that more work needs to be done to show that such a
strategy “can provide a nonreductive response to the exclusion problem” (154),
as involving mental features, in particular:
[C]an the powers of a mental occurrence really be a subset of the powers of a

physical occurrence, if the former is a difference-making cause of certain events
while the latter is not? Even if an affirmative answer is possible, talk of mental
and physical occurrences only involving a single “causing” becomes strained if
mental causes are difference-making causes while physical causes are not—how
could a mental occurrence’s bringing about some event involve the very same
“causing” as some physical occurrence’s bringing about that event, if the one
is more proportional or difference-making than the other with respect to that
event? (154, note 32)
In response to Morris’s initial question, I say ‘yes’: satisfaction of the Proper Subset
of Powers Condition is not only compatible with cases of higher-level difference-
making, but in the motivating cases—that is, cases where a mental or other higher-
level feature is properly seen as tracking certain difference-making considerations
in the production of a given effect that both it and its lower-level realizer have the
power to cause (and manifest, on the occasion in question)—provides the basis
for explaining why the higher-level feature is distinctively efficacious in this way.
It is because the higher-level feature has fewer powers than its realizer (on the
occasion in question, and so on), such that changes in powers not shared between
realizer and realized don’t matter—don’t make a difference—to the production of
the effect, that the higher-level feature tracks difference-making considerations.
Nor does it seem to me that when a higher-level feature is distinctively efficacious
in this way, there is any difficulty in understanding how both the higher- and
lower-level features could cause the very same effect. The having of a power is
one thing—one way for a feature to be efficacious; the having of a power profile
is another thing—another way for a feature to be efficacious. Correspondingly,
two properties can exercise the same power on a given occasion, but only one
have a power profile tracking difference-making considerations (or comparatively
abstract systems of laws or causal joints), as cases such as that of the differently
trained pecking pigeons illustrate. If there is more work to be done here, it lies
in identifying some specific problem for this means of accommodating the causal
autonomy at issue in physically acceptable emergence.
So far, then, we have not been given good reason to resist taking the prima
facie appearances of higher-level reality seriously, as treated along lines of Weak
emergence, in particular.⁸
3.2 Objection: compatibility with reductionism
In this section, I consider objections targeting Weak emergence on grounds

that even granting that satisfaction of the conditions in Weak Emergence by
features S and P is realistically understood, such satisfaction is compatible with
S’s being ontologically reducible to—that is, identical with—some other lower-
level physically acceptable feature P′ . The general line of thought is as follows.
Let the level of base property P be the fundamental physical level—the only level
the reductionist thinks exists. To be sure, if S is Weakly emergent from P on a
given occasion, then S has, on that occasion, a proper subset of the token powers
of P. Still, the reductionist maintains, given that S satisfies the Proper Subset of
Powers Condition, surely S must be reducible to—that is, identical to—some other
physically acceptable feature P′ . As Yates (2012) puts the concern:
[I]f P has as a constituent a physical property P′ whose powers are the same
proper subset of P’s that M inherits, then there’s nothing to prevent identification
of M with P′ . (6)
Whether both P and P′ should be considered ‘realizers’ of S is a topic for further

debate.
In what follows I present and address three different strategies for achieving
such a reduction.
3.2.1 Reduction to a conjunct of a lower-level conjunction?
On the first reductive strategy, satisfaction of the Proper Subset of Powers Condi-
tion is compatible with seemingly higher-level feature S’s being identical with a
⁸ Here I put aside another motivation for anti-realism about (some) higher-level features, associated
with eliminativist physicalism, according to which if we cannot explain a seemingly existent higher-
level feature in physical terms—if an insuperable explanatory gap exists between the seeming feature
(or associated laws and theories) and lower-level physical features (laws, theories)—then we should
reject the seeming existence of the higher-level feature as genuine (see, e.g., Feyerabend 1963, Paul
Churchland 1981, and Patricia Churchland 1986). Insofar as the primary eliminativist target has
been qualitative mental features, this objection is not generally directed against the viability of Weak
emergence, since many special-science features, of the sort that would be candidates for satisfying the
schema for Weak emergence, are uncontroversially explainable in lower-level physical terms (and as
I’ll argue in §3.2.3, such explainability, including deducibility, is compatible with Weak emergence). I
revisit the question of the metaphysical import of explanatory gaps in Ch. 7 (§7.1).
conjunct P′ of a lower-level conjunctive feature P (of form P′ (= S) ∧ P′′ ), such

that S, though different from P, would nonetheless be identical to a lower-level
physical feature. Indeed, one might reasonably suppose that conjunct features
typically satisfy the Proper Subset of Powers Condition with respect to associated
conjunctions, insofar as conjunctive features, being more specific, can contribute
to producing more effects (in the same circumstances) than their conjunct fea-
tures can individually do, for reasons similar to those motivating taking (more
specific) determinates to have more powers than their associated (less specific)
determinables. So, for example, the conjunctive feature being massy and being
charged has powers to produce effects that being massy alone does not have (e.g.,
attract pins) and that being charged alone does not have (e.g., attract other massy
objects). But if P is a conjunctive lower-level physical feature having P′ (= S) and
P′′ as conjuncts, then plausibly, S will also be a lower-level physical feature, and so
reducible rather than emergent.⁹
That conjuncts of conjunctions may satisfy the Proper Subset of Powers Con-
dition is especially pressing for Shoemaker, who characterizes realization in terms
just of satisfaction of the Proper Subset of Powers Condition. It is also a difficulty for
accounts of Weak emergence involving functional realization, since if the higher-
level functional role is associated with a lower-level conjunct feature, one might
naturally suppose that any conjunction containing that conjunct will count as
implementing or realizing the role.
Shoemaker (2000/2001) is sensitive to this concern, and in response stipulates
that for a feature to count as realized, it must not be a conjunct of a conjunctive
realizer:
Property P realizes property S just in case the conditional powers bestowed by S

are a subset of the conditional powers bestowed by P (and P is not a conjunctive
property having S as a conjunct).1⁰ (78)
The need to stipulatively rule out conjunct/conjunction cases as genuine cases of

Weak emergence—i.e., to rule out conjunctive realization—is perhaps a bullet one
could bite. There are, however, two non-stipulative strategies for ruling out such
cases.
The first strategy, due to Umut Baysan, is compatible with any implementation
of Weak Emergence. Here one rules out cases of conjunctive realization as satisfying
the conditions in the schema by saying more about the intended notion of
cotemporal material dependence operative in condition (i). However exactly such
dependence is cashed out, in a context where physicalism is at issue the operative
⁹ Note that the concern here remains even if, as Gibb (2013) compellingly argues, it is not always
the case that conjunctive features have more token powers than each of their conjunct features.
1⁰ Here I have replaced Shoemaker’s variables with S and P for consistency with my discussion.
understanding should be one according to which cotemporally materially depen-

dent features are less fundamental than their base features. But if a conjunctive
feature exists at a given level, it is plausible that the conjuncts of that feature
are as fundamental as or more fundamental than the conjunctive feature. As
such, a conjunct feature S would not be appropriately taken to satisfy the relevant
condition on dependence in the schema for Weak emergence, notwithstanding
that a conjunctive feature (at least nomologically) necessitates its conjuncts; hence
such features pose no threat to the sufficiency of this schema, properly interpreted.
As Baysan (2014) puts it:
[A] property P realizes a property Q if and only if the causal powers of Q are
a proper subset of the causal powers of P, and P is more fundamental than Q.
Thanks to the requirement that a realized property is less fundamental than its
realizers, two things that the original version of the subset view cannot explain
are guaranteed: first, fundamental properties are not realized; second, arbitrary
conjunctions of properties do not realize their conjuncts. (2)
I like the idea of filling in the operative dependence relation at issue, since as it
stands, the notion of cotemporal material dependence at issue in condition (i) of
Weak Emergence is fairly weak. As above, Shoemaker characterizes realization just
in terms of the Proper Subset of Powers Condition—plausibly because he supposes
that this condition also ensures satisfaction of the relevant dependence condition
(see also Clapp 2001). But this supposition is dispensable: nothing prevents a
proponent of Weak emergence from offering an independent characterization of
the cotemporal material dependence at issue, with the Proper Subset of Powers
Condition serving mainly to ensure the autonomy that is characteristic of emer-
gence. It would, after all, be unsurprising that two conditions would be needed
to characterize metaphysical emergence of whatever variety: cotemporal material
dependence is one thing, ontological and causal autonomy another.
The second strategy aimed at blocking conjunctive realization is one appealing
to a specific implementation of Weak Emergence on which conjunct/conjunction
cases are non-stipulatively excluded. A determinable-based account of Weak
emergence, on which base features are taken to be determinates of higher-level
determinables, would work for this purpose, for it is definitive of the deter-
minable/determinate relation that it is not properly metaphysically characterized
in terms of anything like the conjunct/conjunction (or relatedly, genus/species)
relations (see Wilson 2017 for discussion). In mereological terms: determinates,
unlike classical wholes, do not satisfy ‘weak supplementation’, according to which a
whole having one proper part must also have at least one other proper part disjoint
from the first (see Simons 1987 for discussion). In Karen Bennett’s memorable
terms (pers. comm.): determinates are not determinables with frosting on top. As
such, on a determinable-based implementation of Weak Emergence, the nature of
determinate features alone non-stipulatively rules out that conjuncts of lower-level

conjunctions would count as Weakly emergent.
An implementation of Weak Emergence in terms of an elimination in degrees
of freedom (DOF), as proposed in Wilson 2010b and as I will describe in Ch. 5,
might also non-stipulatively rule out conjunctive realization, if no features satisfy-
ing the condition on degrees of freedom at issue in this account also stand in the
conjunct/conjunction relation. Since which degrees of freedom are associated with
which features is an empirical matter, the exclusion here would be contingent as
opposed to following (as in the case of the determinable/determinate relation) just
from features of the specific relation taken to satisfy the Proper Subset of Powers
Condition itself; but even so the exclusion would not be stipulative.
Summing up: that conjuncts can satisfy the Proper Subset of Powers Condition
with respect to associated conjunctions poses a prima facie problem for taking
satisfaction of the conditions in Weak Emergence to be sufficient for metaphysical
emergence, since the relation between conjuncts and conjunctions is plausibly one
according to which conjuncts are at least as fundamental as their associated (here,
presumed physical) conjunctions. The concern can be addressed, however, either
by stipulating the exclusion of such cases (a less than optimal but non-fatal adden-
dum), by denying that conjunct features properly satisfy the cotemporal material
dependence condition (a reasonable addendum in cases where the dependence at
issue is supposed to conform to physicalism), or by endorsing an implementation
of Weak Emergence (along determinable-based and perhaps also DOF-based lines)
that non-stipulatively rules out conjunct/conjunction cases as instances of such
emergence.
3.2.2 Reduction to a disjunction of lower-level disjuncts?
The next reductive strategy takes as its starting point an account on which
purportedly higher-level features are identical with disjunctions of base features.
The strategy is commonly motivated by a seeming objection to reductionism,
according to which the multiple realizability of a given type rules out the type’s
being identical with any one of its realizing types. As Antony (2003) expresses the
objection:
[M]ultiple realizability has ontological consequences. Clearly, a property P cannot

be identical with a property Q if there can be instances of P that are not instances
of Q. But to say that a property S is multiply realizable [by types P1 and P2 ] is
to imply that […] there can be instances of S that are not instances of P1 , and
instances of S that are not instances of P2 , and S cannot be identical with either
P1 or P2 . (3)
A popular reductionist response implements the ‘disjunctive strategy’ or ‘disjunc-

tive move’ (see Fodor 1974, Jaworski 2002, and Dosanjh 2014 and 2019), according
to which multiply realizable types may be identified with a lower-level disjunctive
type, where each disjunct is a type of lower-level realizer of S’s type.11 As Heil
(1992) describes the strategy (without endorsing it):
Multiple realizability […] need not deter a determined identity theorist [that is,
reductionist]. It is open for such a theorist, for instance, to argue that the relevant
[…] characteristic is, in fact, disjunctive in character. That is, it might be that, in
you, mental feature S is realized in neural structure N, whereas in an octopus, S
is realized in a different sort of neural structure N′ . Would this undermine type
identity? It would not, unless we assume that S [could] not be identical with the
disjunctive characteristic ⟨N ∨ N′ ⟩. (64)
Now, a number of objections have been raised against the disjunctive strategy.12
Even putting these concerns aside, there is as yet no clear problem for Weak emer-
gence here; for on the usual understanding of what it is for a disjunctive type to
be instanced or tokened, the disjunctive strategy is incompatible with satisfaction
of the Proper Subset of Powers Condition. On the usual understanding, what it is
for a disjunctive feature type D to be tokened on a given occasion is for one of the
disjunct types to be tokened on that occasion. In the case at hand, the disjunct types
(P1 , P2 , …, Pi ) correspond to the physically acceptable realizers of S’s type. It fol-
lows that if S’s type is identified with disjunctive type D, then any token of S’s type
would, on a given occasion, be identical with a token of one of its realizer types,
on that occasion. But then the Proper Subset of Powers Condition would fail to be
met: S’s token powers would be, on any given occasion, the same as, rather than a
proper subset of, the token powers of the base feature realizing S on that occasion.
But by assumption, S satisfies the condition. It follows that on the assumption that
S satisfies this condition, as Weak emergence requires, S’s type cannot be identified
with a type consisting of a disjunction of S’s lower-level realizers.
11 See Polger and Shapiro 2016 for arguments that empirical reasons to think that special-scientific
types (in particular, mental state types) are multiply realizable are overblown, and that many such cases
can be accommodated in type-theoretic terms, one way or another. Discussion of this argumentation
is beyond the scope of this investigation; for present purposes I will just note that some of Polger and
Shapiro’s cases are subject to challenge (see Levin 2016) and that at any rate they are clear that some
special-science or artifactual types are multiply realizable by their lights, and that moreover there might
well be cases of mental multiple realizability. Correspondingly, I do not treat them here as general
proponents of a reductionist approach to higher-level goings-on.
12 These include that disjunctive features don’t exist (as per Armstrong 1978, 19–23), that disjunctive
features are not available for purposes of physicalist reduction of scientific properties, because they
are too heterogenous to form a natural kind (as per Putnam 1967, Fodor 1974, and Kim 1992b), that
disjunctions are ‘open-ended’ (having an indefinite or infinite number of disjuncts), rendering them
unsuitable to enter into type-identities (as per Lepore and Loewer 1987 and Pereboom and Kornblith
1991), and that among the metaphysically possible realizers of a multiply realizable feature type such
as S will be some physically unacceptable types (as discussed, though not endorsed, in Dosanjh 2014).
Granting that the Proper Subset of Powers Condition is met, a disjunctivist

gambit remains. Drawing in part on discussion in Clapp 2001 and Antony 2003,
Dosanjh (2014) argues that it is not generally the case that what it is for a
disjunctive property to be instanced is for one of its constituent disjunct properties
to be instanced. To be sure, Dosanjh allows, this is the case when the disjuncts of
disjunctive properties are gerrymandered or otherwise dissimilar (as with, e.g.,
the properties being red or round and being sour or prime). But sometimes, he
maintains, the instantiation of the disjunction is not the same as the instantiation
of one or other disjunct. In particular, when the disjuncts are relevantly similar,
in sharing powers associated with the purportedly higher-level feature, and when
the disjunction contains all and only such disjuncts (that is, when the disjunctive
property has disjuncts that “exhaustively overlap”), then, Dosanjh maintains, the
powers of the disjunction should rather be seen as a proper subset of the powers of
whichever disjunct is instanced. And this proper subset relation between powers
will, as per usual, be inherited by the tokens of the disjunction and the disjunct
types. So, Dosanjh concludes, there is no clear barrier to identifying the type of
a realized feature S with the disjunction of its realizer types, since tokens of both
types will satisfy the Proper Subset of Powers Condition.
My response is to deny that this gambit is available to the reductionist, for two
reasons.
First, whether or not the disjuncts of a disjunct are relevantly similar, there
is a case to be made that disjunctive features do not have a proper subset of the
powers of their disjuncts. Consider the powers of P ∨ Q to produce effects when in
circumstances C (restrict attention to these powers, for simplicity). For the cases of
multiple realization at issue, we can without loss of generality assume that different
realizers P and Q cannot be co-instantiated, in which case there will be two ways
for P∨Q to be instanced, and so (at least) two powers which may be exercised in C:
1. If in C ∧ P ∧ ¬Q, then E1 ; and

2. If in C ∧ Q ∧ ¬P, then E2 .
(There may well be others, but that won’t matter for making the point.) What
powers will P have, in C? It will have at least one of the powers of P ∨ Q, in C—
namely,
If in C ∧ P ∧ ¬Q, then E1 .
(Here the conjunct P is redundant, but no matter.) However, P will not have
another of the powers of P ∨ Q, in C—namely,
If in C ∧ Q ∧ ¬P, then E2 .
P would have such a power only if it could be both instanced and not instanced in
C at the same time (powers being relativized to times or temporal intervals), which
it can’t. So in this case (and more generally), P ∨ Q does not have a proper subset
of the powers of P, or any other of its disjuncts.
Second, the disjunctive strategy is most naturally implemented against the
assumption that the reductionist can help themselves to any ‘ontologically
lightweight’ combinations of characteristically physical entities and features,
including—all parties agree—any combinations resulting from Boolean opera-
tions (with the possible exception of negations). Hence it is that, were it possible
to identify a feature S with a disjunction of lower-level physical features, that would
suffice for S’s being ontologically reducible to—that is, identical with—some
lower-level physical feature. As Dosanjh (2014) himself notes, “If there are any
ontologically innocent combinations of properties, boolean logical combinations
are among them. For an important example: when we talk about a disjunction
of properties, we are committed to little beyond the properties that serve as the
disjuncts” (17). But if the disjunctive property D with which S is identified is
not such that its instantiation consists just in the instantiation of its disjuncts,
then resources for constructing D have gone beyond mere Boolean combination,
and it is no longer clear that D is appropriately taken to be a lower-level physical
property. On the contrary: to the extent that D, like S, is supposed to satisfy the
Proper Subset of Powers Condition with respect to S’s lower-level realizer, one might
rather suppose that D (hence S) is a higher-level feature, contra reductionism.13
3.2.3 Reduction to a metaphysical consequence of lower-level laws?
The third strategy for reductively treating a feature S satisfying the conditions in
Weak Emergence adverts to the metaphysical consequences of the laws governing
goings-on at the level of S’s base feature—that is, to the fundamental physical laws
governing entities and features at the one level that the reductionist thinks exists.1⁴
Here the reductionist starts by observing that the Weak emergentist is commit-
ted to taking higher-level features and laws to be metaphysical consequences
13 Dosanjh recognizes the concern here, and in response, attempts to undermine the reasons for
thinking that higher-level features are irreducible to base features. Most pressingly, he argues that
attention to distinctive systems of laws and associated comparatively abstract causal joints need not
be seen as supporting higher-level causal autonomy, on grounds that one may “deny that such laws
are distinct enough to ground causal autonomy”, insofar as “every law statement about a property that
satisfies [the Proper Subset of Powers Condition] is entailed by law statements about physical properties”
(73). This objection is of a piece with, and admits of the same response to, the objection to be next
discussed.
1⁴ Recall that the qualifier ‘metaphysical’ when applied to consequences of laws is intended to
sidestep concerns about the lower-level laws containing vocabulary local to the special sciences: the
notions of ‘consequence’ or ‘deducibility’ here are metaphysical, not representational.
of, and perhaps even to be theoretically deducible or predictable from, lower-level

physical features and laws—after all, it is this which ensures that S is physically
acceptable. But in that case, the reductionist continues, what prevents S, even with
its reduced set of powers, from itself being a lower-level physical feature?
Variations on the theme of this concern are common. Consider these remarks
by Klee (1984), following Nagel (1961), directed against an account of purportedly
physically acceptable emergence on which emergent entities and laws simply
involve new relational structures:
[I]n what sense are these new regularities emergent? To be sure, they may be
regularities and structures of a type not found on lower-levels of organization,
but it has seemed to some (Nagel 1961, 367–74) that this fact by itself would not
justify the label of ‘emergent’ if they had been predictable on the basis of a thor-
ough understanding of those lower-levels of organization. If the new relational
structure which grounds the new regularities could have been predicted on such
a basis, then the new regularities could have been predicted and the force of any
emergence claim, at least partially, compromised. (46)
Indeed, it might seem practically definitional that (in-principle) theoretical

deducibility entails ontological reducibility. As Owens (1989) puts it:
Reductionism is sometimes expressed as the thesis that the laws of the non-
physical sciences can be deduced from those of the physical sciences together
with certain bridging generalizations […]. (63)
To be sure, some higher-level goings-on—in particular, certain complex sys-

tems of the sort we will discuss in Ch. 5—might be thought not to be even in-
principle deducible from lower-level goings-on, at least given a wide but empir-
ically informed purview of the available resources; but the deeper concern here
remains even so. For so long as the nonreductive physicalist maintains that the
purported higher-level goings-on are metaphysical consequences of the lower-
level goings-on, as they are committed to doing, the concern remains that meta-
physically, if not representationally, the higher-level goings-on must be reducible
to—that is, identical to—some or other lower-level goings-on, after all. This line
of thought seems operative in Kim’s (2010) discussion of Shoemaker’s subset-of-
powers approach to physical realization:
[Shoemaker’s view] strikes me as a very strong form of physicalism, something

like type physicalism […] On an intuitive level, my picture of Shoemaker’s
scheme is something like this: The fundamental ontological items of this world
are physical causal powers […] These are packaged, or bundled, into properties
[…] In this picture, where do mental properties find a place? If all properties are
bundles of physical causal powers […] what makes some properties mental and
others physical? (110)
Note that Kim is not asking here what makes it the case that some powers are
associated with, say, mental properties. Which collections of powers are associated
with what features is presumably a matter for scientists (or other relevant practi-
tioners) to decide. Rather, the concern is simply one registering a kind of generic
suspicion that if some feature is associated with only physical powers, then that
alone suffices to render the feature identical to some physical feature. Hence Kim
goes on:
[I]f physical properties are constituted by purely physical causal powers and
mental properties are physically realized in Shoemaker’s sense, there seems no
good reason not to consider these supposedly mental properties to be physical
properties. (111–12)
Morris (2018) endorses this line of thought, and aims to develop it by filling in
the operative understanding of physical properties and powers (which Kim leaves
open) as satisfying the ‘No Fundamental Mentality’ (NFM) constraint (as per
Wilson 2006; see Ch. 1, §1.4.1):
Suppose that physical properties and instances of them have all and only powers
that are NFM-physical. If all properties that are instantiated are either physical
or subset-realized by physical properties and only physical properties, it follows
that there are all and only NFM-physical powers. It follows, moreover, that any
property or property instance associated with a subset of the powers of a physical
property and its instances also has all and only NFM-physical powers. And if
properties are the properties that they are in virtue of their powers, it is difficult
to see why a property that has all and only NFM-physical powers should not be
regarded as NFM-physical. (135)
My response here consists in saying more about how an appeal to laws should
enter into the individuation of levels. It is correct that nonreductive physicalists
will allow, qua physicalists, that all special-science entities and features are meta-
physical consequences of lower-level physical goings-on, and that (consonant with
Physical Causal Closure) the only powers are physical powers (satisfying the NFM
constraint, in particular). However, they can reasonably deny that it thereby fol-
lows that special-science entities and features are themselves lower-level physical
entities or features. In Wilson 2010b, I develop this idea by attention to the role that
degrees of freedom (DOF) play in individuating scientific entities and features.
To start, note that specification of the law-governed properties and behaviors of

lower-level physical entities and features requires all the information needed for
the lower-level physical laws to operate, including information about quantum-
mechanical DOF and associated values, such as spin and quark color charge. By
way of contrast—the nonreductionist can reasonably maintain—specification of
the law-governed properties and behaviors of at least some special-science entities
and features does not require information about, e.g., quantum-mechanical DOF
such as spin or color charge: such DOF are eliminated from these specifications,
as unneeded to characterize higher-level goings-on. For example, and as I will
develop in more detail in Ch. 5 (§5.2.4) and Ch. 6 (§6.1.1), there are empirical
cases to be made that various quantum DOF are eliminated from the DOF needed
to specify entities and features of the sort treated by thermodynamics and classical
mechanics.
Such eliminations in quantum DOF explain, in part, why special-science enti-
ties and features are insensitive to certain micro-level details (and relatedly, make
room for such entities and features’ being multiply realizable): higher-level macro-
entities are typically insensitive to spin-theoretic details, among other quantum
features. More importantly for present purposes: that the specification of special-
science entities and features does not include quantum-level information means
that even though these entities and features are metaphysical consequences of
physical laws, it is not appropriate to place them at the physical level: the quantum
laws wouldn’t know what to do with them! Related considerations, which I expand
on in Ch. 5, §5.2.4, provide a basis for seeing how it could be that an entity or
feature satisfying the Proper Subset of Powers Condition, where each token power is
a power of its lower-level physical realizer, could nonetheless not itself be a lower-
level physical entity or feature.
As such, one must distinguish two sorts of metaphysical consequences of the
lower-level physical laws. First are the entities and features which are consequences
in the broadest sense—which, if physicalism is correct, will include any and all
special-science (more generally: macro-) entities and features. Second are those
metaphysical consequences which retain all the DOF and associated information
(pertaining to quantum spin, color charge, etc.) needed for the lower-level physical
laws to operate. If there is Weak emergence, then as an empirical matter of fact
some entities and features which are metaphysical consequences of the physical
laws in the first sense will not be metaphysical consequences in the second sense.
Such entities and features will have specifications that fail to include all the DOF
and associated information needed for the lower-level physical laws to operate,
and so these entities and features will not be appropriately seen as identical to any
lower-level physical goings-on—as the nonreductionist maintains.
This line of thought—that laws require certain kinds of information in order
to operate, and that among the entities and features that are consequences of
level-L laws, only those preserving the information needed for the level-L-laws
to operate are appropriately placed at L—is explicitly encoded in a DOF-based
implementation of Weak Emergence (see Wilson 2010b), and it is some advantage
of a DOF-based account that it clearly has the resources to address the concern
that theoretical deducibility suffices for or entails ontological reducibility. That
said, there does not appear to be any barrier to proponents of other accounts of
Weak emergence appealing to the aforementioned general scientific facts about
laws and associated DOF by way of responding to the threat of law-consequence
reducibility.
3.3 Objection: compatibility with physical unacceptability
Another general line of objection to Weak emergence—again, understood as

conforming to Weak Emergence—is that a feature S satisfying the conditions in the
schema might nonetheless be, one way or another, ‘over and above’ its dependence
base feature P, in a way rendering S physically unacceptable. In what follows I
present and address several strategies for establishing this result.
3.3.1 Quiddities
Melnyk (2006, 141–3) supposes that for a feature S that cotemporally materially
depends on some base feature P to be physically acceptable, S must satisfy not only
the aforementioned ‘necessitation’ condition, according to which a realized feature
is minimally nomologically necessitated by its realizing (i.e., base) feature, but
also the ‘constitution’ and ‘truthmaking’ conditions, according to which a realized
feature is constituted by base features, and truths about realized features are made
true by truths about realizing features. Properly restricted to worlds with physical
laws of nature similar to those actually governing P, Melnyk’s suppositions are
plausible: it is commonly supposed that S’s physical acceptability requires that S be
necessitated by P in worlds containing physical laws;1⁵ the constitution condition
is a clear variant on the theme of complete metaphysical dependence operative in
physicalism; and the truthmaking condition might be thought to follow from the
1⁵ Not everyone discussing necessitation or entailment of features by base features is careful to make
explicit that the necessitation or entailment at issue is operative against the backdrop assumption that
relevant instantiations of the base features hold fixed the operation of the physical laws. Without this
qualification, failures of necessitation or entailment wouldn’t clearly count against Weak emergence, at
least given that (as some believe), features can enter into different laws and thus have different powers;
but whether a feature P is associated with a feature S in worlds with different laws of nature is neither
here nor there for purposes of motivating S’s being realized by P in worlds with laws relevantly similar
to ours.
general physicalist commitment to the lower-level physical goings-on providing a

basis, hence a truthmaking basis, for all else.
Does satisfaction of the conditions in Weak Emergence guarantee satisfaction
of these conditions on physical acceptability? Melnyk is willing to grant that the
answer is ‘yes’, if one endorses a ‘causal’ account of properties along lines of that
endorsed by Shoemaker (1980 and 1998), on which features are essentially and
exhaustively constituted by the powers they have or bestow. The resulting version
of a powers-based account of realization, Melnyk observes, “has the important
virtue of meeting the necessitation, constitution, and truthmaking conditions”
(144), with “[t]he key move [being] to identify property-instances with something
like clusters of causal power-tokens of particular types” (140). Morris (2018)
agrees, moreover suggesting that a causal account is required for purposes of
guaranteeing the physical acceptability of features satisfying the Proper Subset of
Powers Condition:
[H]ow could the subset relation between powers entail that higher-level proper-
ties are nothing over and above physical realizers if there is more to higher-level
properties than powers […]? (142)
Granting for the moment the supposition that implementing the schema for
Weak emergence requires endorsement of a causal account of properties (features),
the problem remains, in Melnyk’s view, that for some properties, “it’s implausible
to identify their instances with clusters of causal power-tokens” (146).1⁶ The most
pressing case among those Melnyk considers has to do with properties whose
individuation involves some sort of primitive identity or ‘quiddity’: the property
equivalent of a ‘haecceity’, or primitive identity, serving to individuate objects or
other particulars.1⁷ By way of illustration, Melnyk refers to Hawthorne’s (2001)
case of properties which are intuitively distinct but which play the same nomic
role (perhaps positive and negative charge are actual such properties). If special-
science feature S has such a noncausal quiddity, then, it seems, S’s satisfaction of the
Proper Subset of Powers Condition won’t guarantee that S’s quiddity is constituted
by or otherwise ‘nothing over and above’ P (or P’s quiddity, as the case may be),
or that truths about S are made true by truths about P, or even that instances of
P’s type (physically) necessitate instances of S’s type. As such, Melnyk suggests,
1⁶ A similar concern applies to variations on the theme of Shoemaker’s account, associated with
views on which properties and laws are essentially intertwined (as per Swoyer 1982 and Bird 2001 and
2007), and which would also (at least provisionally) entail that satisfaction of the conditions in Weak
Emergence would ensure satisfaction of the constitution, truthmaking, and necessitation conditions.
1⁷ Effectively, a haecceity makes room for the identity and individuation of an entity to float free
from any of the entity’s features, and a quiddity makes room for the identity and individuation of a
feature to float free from any of the feature’s powers.
the possibility that features have noncausal quiddities poses a dilemma for the
Weak emergentist: either properties may have noncausal quiddities, in which case
satisfaction of the Proper Subset of Powers Condition doesn’t ensure satisfaction
of the constitution, truthmaking, and necessitation conditions as required for S’s
physical acceptability; or quiddities are rejected and a causal account of prop-
erties maintained, which ensures S’s physical acceptability, but only at the price
of endorsing a controversial and (given the seeming possibility of Hawthorne-
style cases) not obviously satisfactory account of properties (or features, more
generally).
The Weak emergentist can sidestep this dilemma, as follows (see Wilson 2006
and 2011b). To start, they can observe that the individuation of scientific features
is neutral on the presence or absence of quiddities: in scientific contexts, the
occurrence of scientific features, and any truths about such features, does not
depend on or otherwise track whether such features have quiddities, much less
track how the noncausal quiddities of seemingly distinct features are related. This
is true, in particular, for properties such as positive and negative charge, which at
some level of abstraction play the same causal role. That physics posits more than
one charge property reflects global considerations pertaining to the structure of
the laws as requiring that there be two or more distinct properties playing what
is in some sense the ‘same’ role; the posit of primitive quiddities plays no role in
this story. As such, the Weak emergentist can reasonably maintain that whether
S and/or P have quiddities, shared or not, is irrelevant to whether S satisfies
the constitution, truthmaking, or necessitation conditions, and more generally,
is irrelevant to whether S is physically acceptable; and similarly for artifactual
features satisfying the conditions in Weak Emergence.
Morris (2018) registers dissatisfaction with this line of thought, saying that “I
am not sure that science is exclusively concerned with powers. It is certainly not
the case that all sciences characterize all entities and properties in all and only
causal terms” (145); here Morris cites Polger (2007), who notes that science may
be concerned with, e.g., formal computational properties. My own view is that
even formal computational properties are associated with powers. In any case, my
suggestion is not that the individuation of scientific properties proceeds only by
way of powers; rather, it is that the individuation of such features does not proceed
by way of noncausal quiddities.1⁸ Morris also makes the interesting point that
even if scientific features are individuated in a way that is neutral on whether a
causal account of properties is correct—or more specifically, for present purposes,
on whether scientific features have quiddities, it remains that “physicalism is a
comprehensive thesis about the nature and structure of the world […] the question
1⁸ In context, Morris also appears to suppose, as is commonly done, that phenomenal aspects of
features are not causal; I respond to this sort of concern in §3.3.2.
of whether subset realization suffices for physical acceptability without a causal

view of properties immediately reappears as the question of whether the ‘scientific’
account of the world is the whole world” (145). I take Morris’s point, and in
response clarify that in my view, quiddities do not enter into the individuation
of any features, scientific or otherwise, at a world.
Importantly, this response does not require endorsing a causal theory of prop-
erties; indeed, this response might be endorsed even by those taking properties
to have noncausal quiddities. For the primary reason for endorsing property
(more generally, feature) quiddities concerns not a supposed need to distinguish
properties with shared roles at a world, but rather a supposed need to identify
or individuate properties at or across worlds with different laws of nature, on the
assumption that the powers of properties may vary in worlds with different laws
(see, e.g., Lewis 1983a and Schaffer 2004). Since we do not have any access to
noncausal quiddities, if properties are individuated within or at a world even partly
by their quiddities, then all bets are off so far as actual property individuation
is concerned (see Shoemaker 1980 for a similar epistemological point). As such,
those taking properties to have quiddities should agree that the identification and
individuation of broadly scientific features within or across worlds with the same
(or relevantly similar) laws of nature proceeds by reference to their powers in a
way that is neutral on the presence or absence of quiddities.
The upshot is that the possibility of noncausal quiddities poses no threat to
the claim that satisfaction of the conditions in Weak Emergence is sufficient for
physically acceptable emergence.
3.3.2 Phenomenal aspects
A second concern about the physical acceptability of features satisfying the con-
ditions of Weak Emergence adverts to the phenomenal or qualitative aspects of
certain mental features. As Walter (2010) summarizes the concern, “phenomenal
properties just cannot be characterized in terms of their causal role, and thus they
cannot be individuated in terms of the causal powers of their bearers to which
they contribute” (220). Indeed, it is common to claim that phenomenal aspects of
mental features cannot be characterized in terms of causal roles or associated pow-
ers, on grounds that such aspects are associated with a phenomenal ‘character’ or
qualitative ‘feel’ (sometimes described as there being “something that it is like” to
have the feature in question) not exhausted by the powers of the associated feature.
Call a feature having phenomenal or qualitative aspects a ‘qualitative feature’.
As Jonas Christensen notes (pers. comm.), qualitative features needn’t be
epiphenomenal—an important qualification for present purposes, since on the
operative conception of emergent entities and features, these are efficacious
(indeed, distinctively efficacious) as well as distinct.1⁹ Still, Christensen suggests,

attention to qualitative feel provides reason to think that the phenomenality of a
feature is not exhausted by the powers of the feature. The intense feel of pain might
indeed cause one to cry out and/or behave in certain pain-attending or pain-
mitigating ways; nonetheless, the qualitative feel of being in pain is reasonably
taken to be an extra feature of reality, in addition to any such powers. Supposing so,
then even if a qualitative feature S satisfied the Proper Subset of Powers Condition,
S could, in virtue of being associated with a distinctive phenomenal aspect, be
over and above its base feature P, and hence physically unacceptable.
Moreover, Christensen observes, the means of resisting taking the possibility
of quiddities to pose a problem for Weak emergence do not carry over to the
case of phenomenal aspects, for two reasons. First, unlike the case of quiddities,
scientific theory and practice do appear to be concerned with phenomenal aspects
of mental features—for example, pain and how to allay it are part of the subject
matter of pharmacology. Second, unlike the case of quiddities, it is commonly
thought that if mental features have rich phenomenal aspects that are not fully
explained by or (more weakly) somehow metaphysically accommodated in terms
ultimately involving lower-level physical powers, then such mental features would
be physically unacceptable. So the previous response to the possibility of quiddities
doesn’t carry over to the case of phenomenal aspects.
I see at least two strategies of response to Christensen’s concern. On the first,
one maintains that any phenomenal aspects of features there may be are reducible
to non-phenomenal representational aspects. Here one might follow reductive
representationalists (e.g., Harman 1990, Dretske 1995, Tye 1995, Byrne 2001, Hill
2009; see Chalmers 2004 for discussion) in thinking that phenomenal aspects are
reducible to non-phenomenal features of reality, which non-phenomenal features
are, in turn, standardly taken to be amenable to treatment in terms of powers.
On the second strategy of response, which I prefer, one rather maintains that
phenomenal aspects of mental features are fully incorporated into the powers of
these features (compatible with a view on which powers are contingently associ-
ated with features, relative to a given set of laws). The suggestion, in other words,
is that phenomenal aspects are ‘causally loaded’. After all, as Christensen notes,
qualitative mental features are plausibly taken to enter into causal relations—in
virtue, at least in part, of their phenomenal aspects. In my view, it is reasonable to
believe that, as our immediate introspective access to the phenomenal aspects of
mental features suggests, any (discernible) differences in phenomenality would
result in causal differences—including, if one wants a systematic hook to hang
this point on, differences in what sort of qualitative experience the bearer of the
1⁹ Moreover, as I’ll discuss later in the section, good reasons for thinking that phenomenal or
qualitative features might be entirely epiphenomenal are in short supply.
mental feature will have and (upon reflecting) take themselves to be having. Call
the thesis that phenomenal aspects of a feature are fully incorporated into powers
of that feature, perhaps in a way that varies with different laws at different worlds,
‘Phenomenal Incorporation’. Given Phenomenal Incorporation, it is reasonable to
believe that the powers of a qualitative feature S fully incorporate its phenomenal
aspects, in such a way that, were each power of S to be (on an occasion, etc.)
token-identical with a power of its lower-level physical feature P, that would
indeed suffice for S’s being physically acceptable.2⁰ Similar views have recently
been proposed and defended by Langsam (2011) and Mørch (2018) and (2020).21
Are there reasons to reject Phenomenal Incorporation? One purported reason
might advert to arguments (see Chalmers 1996, 2009, and 2003) for the (suitably
ideal) conceivability of zombies, according to which there could be a world
physically and functionally (including causally) the same as the actual world, and
in particular containing creatures physically and functionally like us, but with a
complete absence of phenomenal character. Supposing that zombies are (suitably
ideally) conceivable, and supposing also that (suitably ideal) conceivability is a
guide to metaphysical possibility, then one might reasonably think that higher-
level qualitative mental features are, even if nomologically connected to lower-
level physical features and powers, nonetheless wholly distinct from such features
and powers in a way at odds with Phenomenal Incorporation. A related scenario
(see Block and Fodor 1972 and Shoemaker 1975) involves the (suitably ideal)
conceivability of creatures who are physically and functionally identical to us,
yet spectrally inverted (with one seeing green where the other sees red, and
so on); here again one might see such a possibility, if genuine, as indicating that
2⁰ Here it is worth recalling that the physicalist is under no obligation to deny that there are
phenomenal aspects of natural reality—they merely maintain that any such aspects are metaphysically
dependent on suitably complex lower-level physical goings-on. Note also that the supposition that, as
per Phenomenal Incorporation, the powers of a Weakly emergent qualitative feature will be identical
to powers of its base feature poses no threat to the ‘No Fundamental Mentality’ (NFM) constraint in
the operative conception of the physical (see Ch. 1, §1.4.1), since the dependence base feature will
be a feature of a micro-configuration, and as previously disussed, the NFM constraint only rules out
mentality as had or bestowed by individual or ‘basic’ physical entities or features. (Similarly, mutatis
mutandis, for variations on the theme of the NFM constraint aimed at making sense of physicalism in
infinitely decomposable worlds.) All (non-eliminativist) physicalists will grant that suitably complex
physical configurations have features that are identical with or which realize qualitative mental features.
21 Langsam (2011) claims that conscious states have powers that “flow in an intelligible way from the
relevant intrinsic features of consciousness” (4), and Mørch (2020) claims that “phenomenal properties
(i.e. properties which characterize what it is like to be in conscious states) have non-Humean causal
powers […] in virtue of how they feel, i.e. in virtue of their phenomenal character” (2). Each cites by way
of illustration the power of a state of pain to cause the desire to avoid it and the power of phenomenal
properties to cause beliefs. I am friendly to these views, though as in the main text, I am inclined to
think that one can maintain that the powers of phenomenal mental states incorporate reference to
their phenomenal aspects while remaining neutral as regards the metaphysics and modal status of the
associated powers.
qualitative mental features float free of lower-level physical features and powers,
contra Phenomenal Incorporation.
There is a huge contemporary literature on whether the zombie and/or inverted
spectrum scenarios are genuinely (ideally) conceivable, and on whether (suitably
ideal) conceivability is a guide to metaphysical possibility.22 In Ch. 7 (§7.1.2), I
will revisit Chalmers’s conceivability argument, in particular, in more detail. For
now, I aim to lay out two ways of maintaining Phenomenal Incorporation, and the
associated viability of Weak emergence, in the face of these sorts of scenarios.
First, and perhaps most importantly, even if the scenarios are taken to be ideally
conceivable and metaphysically possible, they do not undercut Phenomenal Incor-
poration per se; rather, they undercut Phenomenal Incorporation understood
as coupled with physicalism. This much is compatible with qualitative features’
satisfying Phenomenal Incorporation, but being Strongly emergent (among other
options). In that case, however, the scenarios pose no difficulty for the in-principle
viability of Weak emergence, for what they establish is not the falsity of Phenom-
enal Incorporation but rather (at best) the failure of qualitative features to satisfy
the Proper Subset of Powers Condition. Again: the zombie and inverted spectra
scenarios do not show that qualitative aspects float free of powers; they show, at
best, that such aspects float free of physical powers.
Second, though it is not the present order of business to defend the live
possibility of a Weak emergentist account of qualitative features (a possibility to
which we will return in Ch. 7), it is worth observing that a Weak emergentist
(nonreductive physicalist) has a fairly straightforward reason to deny that such
scenarios are genuinely possible. As Perry (2001) correctly notes as regards the
zombie case, physicalists (whether of reductive or nonreductive stripe) taking
qualitative features to be efficacious will maintain that the absence of such features
at a world would entail the absence of the corresponding powers or associated
effects at that world, resulting, contra Chalmers’s assumption, in a physical or
functional causal difference (see Wilson 2002b for discussion). Similarly, physical-
ists accepting Phenomenal Incorporation may reasonably deny that either zombie
or inverted spectrum cases are genuinely possible. More generally, independently
of whether physicalism is true, it is implausible that inverted spectral differences
would fail to make a physical or functional causal difference—one has only
to look at a spectrally inverted image of food, for example, to appreciate how
comparatively unappetizing the represented contents seem. Given the availability
of such physicalist responses, it appears that the zombie and spectral inversion
22 Re zombies, see the discussion and references in Kirk 2015; re inverted spectrum cases, see the
discussion and references in Byrne 2020; for reasons to think that such scenarios import problematic
propositional presuppositions, see Hellie 2019b; re whether conceivability is the best way to implement
‘epistemic two-dimensionalism’ (by way of regaining, post-Kripke, some a priori access to modal
truths), see the discussion and references in Biggs and Wilson 2020.
scenarios presuppose rather than establish that qualitative aspects float free of
physical goings-on (and so are physically unacceptable). Relatedly, only if one
antecedently rejects Phenomenal Incorporation, either in general or as compat-
ible with physicalism, will one be inclined to think that there might be worlds
physically and functionally identical to ours in which our doppelgangers have
mental features entirely lacking in or spectrally inverted with respect to their
phenomenal character. As such, the scenarios pose no clear threat to Phenomenal
Incorporation, either in general or as compatible with physicalism, and so pose
no threat to the claim that satisfaction of the conditions in Weak Emergence is
sufficient for physical acceptability.23
How else might one argue against Phenomenal Incorporation in such a way as
to call into question the sufficiency of Weak Emergence for physically acceptable
emergence? What seems to be required is that it be genuinely possible that
qualitative features be epiphenomenal—capable of being instantiated without
any powers whatsoever, physical or otherwise. But since, as above, the zombie
and inverted spectra scenarios are compatible with the truth of Strong emer-
gence, these scenarios do not establish the genuine possibility of epiphenomenal
qualitative features; and given that the having of a qualitative feature crucially
involves the having of a qualitative experience—a seemingly causal affair—one
might reasonably surmise that no other scenario is going to establish this, either.
I conclude that, modulo the presentation of some new and better reasons to reject
Phenomenal Incorporation, phenomenal aspects of qualitative features pose no
problem for the physical acceptability of features satisfying the conditions in Weak
Emergence.
3.3.3 Lack of entailment
Several cases have been offered as showing that satisfaction of the Proper Subset
Condition on Powers by features S and P is compatible with P’s failing to entail
or metaphysically necessitate S, and so (it is moreover maintained) with S’s being
physically unacceptable:
• Melnyk (2006) suggests that if S’s possession requires having a causal history
of some sort (e.g., being a member of the species Homo sapiens, or being a
mother), or requires standing in noncausal (e.g., spatiotemporal) relations
(e.g., being to the right of a rock), and on the assumption that the base feature
P is spatiotemporally local, then even if S satisfies the conditions in Weak
23 As prefigured, in Ch. 7 (§7.1.2), I will consider and address a more sophisticated line of thought,
due to Chalmers, for taking the conceivability of zombies to have anti-physicalist import.
Emergence, S might fail to be metaphysically necessitated by P, and hence not

be guaranteed to be physically acceptable. A similar concern might attach to
intentional mental features having ‘broad’ content: if S is a state of believing
that water is wet, and the extension of ‘water’ encodes facts about the chemical
composition of the predominantly watery stuff in the environment, and P is
a given neurophysiological state, then even granting that S’s token powers are
a proper subset of P’s token powers, P might not entail S.
• McLaughlin’s (2007) case pertains to a view of properties (more generally,
features) on which they are individuated not just by what effects they may
contribute to causing (powers, properly speaking), but also by how they may
be caused (as per what Shoemaker calls ‘backwards-facing powers’). Insofar
as the Proper Subset of Powers Condition makes no reference to backwards-
facing powers, satisfaction of this condition by S with respect to P won’t
ensure that S is, as McLaughlin puts it, ‘entailed by’ (i.e., metaphysically neces-
sitated by) the occurrence of P. What is additionally needed to ensure that P
entails S, McLaughlin suggests, is that the backwards-facing powers of P entail
the backwards-facing powers of S (as the Proper Subset of Powers Condition
does for the forward-facing powers of S with respect to P); but there doesn’t
seem to be any clear way of ensuring this. In particular, Shoemaker’s (2007)
revision of his account of realization to incorporate reference to backwards-
facing powers doesn’t ensure that P even nomologically necessitates S, for this
revision requires that the token backwards-facing powers of an instance of a
higher-level realized feature S be a proper superset of those of the feature P
that realizes it on a given occasion. In that case, as McLaughlin observes, the
occurrence of P will have fewer backwards-facing powers than S, in which
case P will not entail S (at least, not without further conditions).
• Gibb (2013) offers a case involving circuits in which the powers of a con-
junctive feature S are a proper subset of the powers of one of its conjuncts P,
but where (since conjuncts typically do not entail associated conjunctions),
satisfaction of the Proper Subset of Powers Condition is compatible with S’s
being physically unacceptable—if, in particular, other of S’s conjuncts are
physically unacceptable.
A first, general response to these cases proceeds by noting that each case
is directed against an account of realization (e.g., that endorsed in Shoemaker
2000/2001) appealing just to the satisfaction of the Proper Subset of Powers Condi-
tion. The schema for Weak emergence also explicitly involves a cotemporal mate-
rial dependence condition requiring, among other things, that Weakly emergent
features be minimally nomologically supervenient on base features. But in each of
the cases, the base feature P at issue fails (not just to metaphysically necessitate,
but also) to even nomologically necessitate the higher-level feature S. (This is
clear for Melnyk’s and McLaughlin’s cases; the reader will have to take my word
for it for Gibb’s case, which is a bit too complicated to present in detail here.)
Correspondingly, the cases don’t in fact satisfy the conditions in the schema for
Weak emergence, and hence do not bear on the physical acceptability of features
satisfying those conditions.
More specific responses are available for the individual cases:
• As regards Melnyk’s cases, a Weak emergentist can maintain that what such
cases show is that if the instantiation of a higher-level feature S requires
that certain spatiotemporal or causal-historical features or facts be in place,
then establishing that S is Weakly emergent will require identifying a lower-
level feature P capable of encoding such facts in such a way as to satisfy
the dependence condition. For example, features such as being to the right
of a rock will require realizing features covering the relevant spatiotemporal
extent; and if S is a historically sensitive species feature—say, being human—
then if P is to minimally nomologically necessitate S, then P will presumably
have to be a spatiotemporally non-local, extrinsic, and presumably massively
complex lower-level physical feature.2⁴ Such spatiotemporal extensions of the
base entities and features are familiar from the literature on supervenience
as a potential realization relation (to be discussed down the line), where, it
is suggested, accommodating extrinsically constituted higher-level features
requires taking the physical supervenience base feature to be regional or even
global (see, e.g., Horgan 1982, Kim 1984, 1989, and Paull and Sider 1992).2⁵
• As regards McLaughlin’s case, the Weak emergentist can reject as clearly
incorrect a view of properties on which they are typically individuated (in
part) by reference to the ways in which they can be caused. For example,
scientific properties do not seem to be individuated in this way: it isn’t
any part of what it is to be an H2 O molecule (even given the actual laws)
that such molecules might arise either as the result of natural or artificial
processes. If higher-level features are not typically individuated by the ways
in which they may be caused, then the fact that higher-level features can
typically be caused in more ways than lower-level features poses no threat to
either the necessitation of or the physical acceptability of higher-level features
satisfying the Proper Subset of Powers Condition. Perhaps for such reasons,
Shoemaker has since given up individuating properties in terms referencing
2⁴ Perhaps there are other strategies for encoding such historical facts, but this will do for purposes
of illustration.
2⁵ That non-locality of certain higher-level features might be seen as indicating just that the
dependence base feature must be appropriately spatiotemporally extended is also relevant to assessing
certain purported cases of Strong emergence (e.g., the case of quantum entanglement); see Ch. 4, note
3 for discussion.
‘backwards-facing powers’, and reverted to characterizing realization in terms

referencing only powers of the usual, forward-facing variety. To be sure,
it is plausible that some higher-level properties are individuated in part by
reference to how they were in fact caused (as in the case of species kinds);
but in that case the Weak emergentist can offer a response similar to that just
given to Melnyk, according to which such cases show only that lower-level
feature P must be spatiotemporally broad enough to ensure satisfaction of
the dependence condition.
• As regards Gibb’s case, suppose that this case is tweaked to ensure satisfaction
of the dependence condition, such that P qua conjunct feature (at least)
nomologically necessitates S qua conjunctive feature. The Weak emergentist
might then maintain that given the satisfaction of the Proper Subset of
Powers Condition, there’s no clear reason to think that S would be physically
unacceptable.
3.3.4 Fundamentally mental powers
Another case aimed at showing that satisfaction of the Proper Subset of Powers
Condition is compatible with physical unacceptability is due to Baltimore (2013),
as follows:
[A]ccording to Wilson, when the set of causal powers associated with a mental
property is a proper subset of the set of causal powers associated with its physical
base, the mental property, although distinct from its physical base, will still
be physicalistically acceptable […] There is reason to question, however, the
safety of such a retreat. Consider […] a micro-object at the fundamental level
of the micro-macro hierarchy [;] channel the spirit of panpsychism and endow
the fundamental micro-object with mentality. However, instead of it having a
mental property that is identical with one of its physical properties, suppose
that it has a mental property that non-reductively supervenes on one of its
physical properties. Suppose further that […] each causal power associated with
the mental property is identical with a causal power associated with its physical
base. But the physical base here is a fundamental property and, so, mental
causal powers are thereby associated with a fundamental property, which seems
physicalistically unacceptable. (20)
This is an interesting case, and I agree with Baltimore that it is not ruled out
just by satisfaction of the Proper Subset of Powers Condition (or more generally,
the conditions in Weak Emergence). It is ruled out, however, by the operative
physics-based conception of the physical, which as discussed in Ch. 1 (§1.4.1)
incorporates a ‘No Fundamental Mentality’ (NFM) (or related) constraint, as

encoding the historical philosophical motivations for characterizing the physical
in such a way as to avoid Hempel’s Dilemma. Baltimore considers this response,
and replies that the NFM constraint should not be imposed, since it “builds too
much metaphysics into the notion of the physical” (15)—ruling out, in particular,
that the physical entities might turn out to be as panpsychists suppose. But given
the need to preserve the historical contrast between physicalism and panpsychism,
and to secure the other advantages of a physics-based NFM account of the physical
(see Wilson 2006), I take the import of Baltimore’s case to mainly be that of
reminding us that nonreductive physicalist implementations of Weak Emergence
will need to appeal to an account of the physical goings-on suited for physicalist
purposes.
3.3.5 Physically unacceptable constraints
Physically acceptable emergence, according to Weak Emergence, is ultimately a

matter of a higher-level feature S having, on a given occasion, only a proper
subset of the token powers of the lower-level feature P upon which S cotemporally
materially depends, on that occasion. But presumably not every proper subset of
powers had by such a P corresponds to a distinct higher-level feature. So, one
might ask, what distinguishes subsets of powers that are associated with a given
higher-level feature from those that aren’t? Now, in general it’s not the burden of
the metaphysician aiming to make sense of emergence to say which higher-level
features exist—that’s an empirical matter, whether natural or artifactual features
are at issue. Still, the Weak emergentist does need to say enough to block answers
to the above question having as a consequence that the existence or instantiation
of the higher-level feature would be physically unacceptable. As Melnyk (2006)
puts it:
[N]ot just any old subcluster of a given cluster of causal power-tokens constitutes
a genuine property-instance […]. Hence, some further condition must be met by
those subclusters that do (see Shoemaker, 2001, pp. 85–6). And it is presumably
a task for metaphysics to say what this further condition is […] the meeting of
the further condition must be a purely physical affair […]. (146)
By way of motivation, Melnyk provides a “frivolous” example, in which a sub-

cluster of a given cluster of token powers constitutes a genuine property instance
only if it is divinely classified as natural, and to observe that such a scenario would
call into question the physical acceptability of the higher-level feature at issue—
even were such a feature to satisfy the conditions in Weak Emergence.
Before responding, I want to clarify just what the concern here is supposed to
be. At least as I read Melnyk, the concern isn’t that any property instantiation
that might be caused by a divine being would thereby be physically unacceptable.
Suppose that a divine being said, “Let there be the big bang and the laws of physics”,
and in so saying brought our universe into being, to evolve henceforth according to
those initial conditions and laws. This sort of scenario would not obviously falsify
physicalism—some physicalists are theists—and relatedly, it would not obviously
render all natural features (including physical ones!) physically unacceptable. The
concern is rather along the following lines: whatever makes it the case that some
proper subsets of token powers of a given lower-level physical feature correspond
to (instantiated) higher-level features, while other subsets do not do so, had better
itself be physically acceptable if the higher-level features are to be physically
acceptable; yet satisfaction of the conditions in Weak Emergence is silent on why
a given higher-level feature S has the distinctive power profile it has, and so is
compatible (one might think) with the instantiation of a higher-level feature’s
being, somehow or other, the outcome of a physically unacceptable process.
Melnyk’s concern can be more finely pitched by observing that in many cases
the existence of higher-level entities and features goes hand in hand with the
presence of certain constraints. As the qualifier ‘special’ suggests, the special
sciences concern broadly natural goings-on which are restricted to certain ‘spe-
cial’ circumstances—corresponding, e.g., to energies and associated temperatures
conducive for the formation of atoms or stable molecules, or where conditions are
favorable for life, or where creatures with mentality exist, and so on. Effectively,
such restricted circumstances encode the presence of constraints which enter
into individuating the associated special-scientific entities, features, and laws as
necessarily subject to those constraints. Now, entities and features that can exist
only in some circumstances—i.e., only when associated constraints are in place—
will be able to do less than, hence will have fewer associated powers than, entities
and features that can exist under the restricted circumstances as well as other
circumstances; as Lamb (2015) puts it, “constraints are limits on possible states
and behaviours” (3). For example, a system of atoms (or associated feature of
being such-and-such system of atoms) will have more powers than the molecule
(or associated feature of being such-and-such a molecule) for which it may serve
as a dependence base entity (feature), since the system of atoms may exist in, and
contribute to the production of effects in, circumstances in which the molecule
cannot exist as well as circumstances in which the molecule exists. And one might
reasonably suggest that many special-science features which are good candidates
for physically acceptable emergence are ones similarly reflecting the holding
of certain constraints. That constraints may be operative in the existence and
individuation of higher-level entities or features satisfying the Proper Subset of
Powers Condition provides a way for Melnyk’s concern to be pitched—namely as
the concern that these constraints might arise as a result of physically unacceptable
processes. Though this concern is a live one, the Weak emergentist has two ways
to respond.
First, the Weak emergentist can maintain that, as per the historical dispute
between physicalists (materialists) and their Strong emergentist rivals, what dis-
tinguishes ‘over and above’ (i.e., physically unacceptable) features is just that they
have powers their base features don’t have. So even if some physically unacceptable
process were to be operative in making it the case that, e.g., temperatures were
within the range needed for stable molecules to exist, this wouldn’t in itself show
that molecules were physically unacceptable. This response seems to be generally
available to proponents of any implementation of Weak emergence.
Second, the Weak emergentist can build into the operative implementation of
Weak emergence that any constraints underlying or responsible for the holding of
the Proper Subset of Powers Condition must occur as a result of physically accept-
able processes (or, more generally, as a result of law-governed processes at the level
of P). In my (2010b) DOF-based implementation of Weak Emergence (which I will
present in more detail in Ch. 5), I include just such a condition.2⁶ If there is a well-
motivated revision of Weak Emergence and the associated conception of physically
acceptable emergence, I am inclined to see it as involving the inclusion of such a
condition on any operative constraints.
3.4 Objection: non-necessity
I turn now to considering certain accounts of physically acceptable emergence

which can be seen as challenging the claim that satisfaction of the condi-
tions in Weak Emergence is necessary for physically acceptable emergence—
notwithstanding that, as argued in Ch. 2, a wide range of accounts of realization
clearly do aim to satisfy these conditions. There are three main alternative
2⁶ See condition 3, below. Since in this implementation I aimed to characterize a Weakly emergent
entity (as opposed to just a Weakly emergent feature), the condition more generally requires that any
constraints associated with any feature of the entity satisfy the condition:
Weak Ontological Emergence (DOF): An entity E is weakly emergent from some entities ei
if
1. E is composed by the ei , as a result of imposing some constraint(s) on the ei .
2. For some characteristic state S of E: at least one of the DOF required to characterize a
realizing system of E (consisting of the ei standing in the ei -level relations relevant to
composing E) as being in E is eliminated from the DOF required to characterize E as
being in S.
3. For every characteristic state S of E: Every reduction, restriction, or elimination in the
DOF needed to characterize E as being in S is associated with ei -level constraints.
4. The law-governed properties and behavior of E are completely determined by the law-
governed properties and behavior of the ei , when the ei stand in the ei -level relations
relevant to their composing E. (Wilson 2010b, 292)
approaches on offer, in terms of token identity, constitution, and primitive

Grounding, respectively.2⁷ In what follows, I provide reasons for thinking that
each of these alternative approaches to physically acceptable emergence is
unsatisfactory.
3.4.1 Token identity
On a token identity approach, a realized feature S is token but not type-identical to

the base feature P upon which it depends, on a given occasion (see, e.g., Davidson
1970, Macdonald and Macdonald 1995, Ehring 2003, and Robb 1997). Such an
approach entails that every token power of S, on an occasion, is identical to a token
power of base feature P, on that occasion, and hence avoids token-level causal
overdetermination between S and P. It doesn’t gain S’s (token-level) ontological
autonomy (distinctness), but one might think that this isn’t as important as gaining
S’s reality and efficacy. Moreover, the token-identity theorist maintains, the causal
autonomy characteristic of emergence can be accommodated by taking S to be
distinctively efficacious in virtue of falling under a distinctive type.
One concern with this approach is that appeal to non-identical types as the
ground of the causal autonomy of S with respect to P reintroduces a threat
of higher-level causal overdetermination, and associated threat of exclusion. As
Ehring (2003) puts the concern: “Since mental types are not identical to physical
types (because of multiple realizability) even if mental tokens are identical to
physical tokens, there are no causes of physical effects that are efficacious in virtue
of mental property types” (364). To gain S’s causal autonomy, the proponent of a
token identity account must provide an account of the relation between S and P’s
associated types, and show that the associated means of gaining autonomy does
not reintroduce problematic overdetermination.
Ehring (2003) aims to provide an account of the relation between types on
which problematic overdetermination is avoided. To start, he takes S and P to be
tropes—particularized properties, such as this redness, or that complex configura-
tion of charges—and their associated types to be collections of resembling tropes.
He then argues that S and P’s types are related as part to whole. Here the order
of the part/whole relation is reversed from Shoemaker’s (2000/2001) and Clapp’s
(2001) understanding: for Shoemaker and Clapp, a realized type is part of each
2⁷ A fourth line of objection to the necessity claim appeals to supervenience, with the general idea
being that it suffices for physically acceptable emergence that a higher-level feature asymmetrically
metaphysically supervenes on lower-level physical features. However, since appeals to supervenience
in this context are primarily aimed at distinguishing physically unacceptable emergence from any sort
of physically acceptable relation, I postpone discussion of supervenience as a basis for metaphysical
emergence until the next chapter.
of its realizing types; for Ehring, a realized type is a whole, having as parts the
subclasses (of resembling tropes) of its realizing types.
What is interesting for present purposes is that Ehring takes appeal to a type-
level ‘proper subset of powers’ strategy to be required by the token identity
theorist to establish the requisite causal autonomy without inducing problematic
overdetermination. He first motivates the view for the determinable type red and
its associated determinate shade types:
It seems clear that the class of red tropes as a whole, the type “red”, has certain
causal powers. […] [W]e are still left with the question of how the causal powers
of this class as a whole are related to the causal powers of the subclasses of each
determinate shade of red. I believe the answer is that the causal powers of the type
“red” are those exactly similar causal powers shared by each of these subclasses.
[…] For any causal power of a shade of red not matched by an exactly similar
causal power belonging to each of the other shades of red, “red” lacks any such
power. (374)
Here the powers of the type red are a proper subset of those of each of its
constituent determinate types. Ehring takes similar considerations to indicate that
the powers of mental types are a proper subset of the powers of their realizing
physical types. Macdonald and Macdonald (1995) also plausibly implement a
subset-of-powers-based strategy at the level of types, for they take mental state
types to be relevantly analogous to determinables, and as previously, a case can be
made that determinable types have a proper subset of the powers of their realizing
determinate types.
Such hybrid approaches, combining token identity of features with a proper
subset relation between powers of associated types, are problematic, however, for
a reason that I have already observed. On the hybrid view, S’s type does not have
powers that differ between its realizer types; but a token of S’s type can, when
identical with a token of P’s type, have such powers. Hence token feature S can
have powers that S’s type doesn’t have. But it arguably makes no sense for a token
feature to have more powers than its type, at least if types are supposed to track
similarities among associated tokens. If a token feature has more powers than a
given type, then that is itself compelling reason to think that the token feature is
not of that type, or so it seems to me.
Avoiding this difficulty requires that the proper subset relation between powers
hold at the level of tokens as well as types—that is, that the Proper Subset of Powers
Condition be imposed. More precisely, it requires imposing the Proper Subset of
Powers Condition if the account is to be a version of nonreductive physicalism
(physically acceptable emergence). Alternatively, the proponent of a token identity
account could endorse reductionism at both token and type levels. At the end of the
day, token identity accounts of realization either do not establish the ontological
and causal autonomy of higher-level features, and so are not really accounts
of physically acceptable emergence; or else impose the Proper Subset of Powers
Condition, and so are not really token identity accounts.
3.4.2 Many-one constitutive mechanism
In Ch. 2, I discussed constitutive mechanism accounts of realization of a higher-

level feature of a system by a lower-level micro-structural feature, where the
latter feature encodes information about the system’s parts and provides a causal
mechanism for the implementation of the causal role and associated powers
associated with the higher-level feature. Such accounts, I argued, plausibly ensure
satisfaction of the Proper Subset of Powers Condition, and more generally are
properly seen as implementing the schema for Weak emergence. There I also
noted, however, that a second variety of constitutive mechanism approach to
realization, notably endorsed by Gillett in his 2002b, 2002a, 2016, and elsewhere,
rather takes realization to be a many-one relation between the many token features
of individual parts of a system and a token feature of the system as a whole.
One might think that Gillett’s version of mechanistic realization would not
satisfy the Proper Subset of Powers Condition, on pain of saying, implausibly, that
every token power of a mechanistically realized feature is identical to a token
power of some or other individual part of the system having the feature. And
indeed, Gillett does charge those endorsing what he calls the ‘Subset View’ with
this implausible consequence. In fact, however (and echoing remarks made in Ch.
1, note 11), there is a case to be made that a many-one mechanistic approach to
realization is committed to satisfaction of the Proper Subset of Powers Condition. As
Craver (2001) emphasizes, a given constitutive mechanistic explanation does not
proceed simply by citing the features of individuals which compose or otherwise
serve as a cotemporal material base for a given macro-entity. Indeed, Craver goes
further and says that “In fact, the φ properties of mechanisms are not really mere
properties [of individuals] at all; they are the activities of and among the entities
in the mechanism” (59). More specifically, he observes:
An analytic account for a mechanism is not just a list of entities and activities;
it is a description of a mechanism. And that description involves, in addition to
a list of entities and activities, a description of how they are organized together
actively, spatially, and temporally in S’s φ-ing. Specifying the mechanistic role of
some component X, accordingly, involves describing how X is organized with the
other entities in S such that it contributes to S’s φ-ing. (61)
In metaphysical terms, Craver’s point tells against an understanding of mech-

anistic realization according to which a single higher-level feature of a system
is somehow realized by—not a list, but a plurality—of co-instantiated features
of individual parts of the system. Nor would Gillett deny that in order for
mechanistic realization to occur, the parts and associated features need to be
appropriately spatiotemporally and causally integrated. But if a many-one account
of mechanistic realization is feasible only if the many parts and their many
features are appropriately integrated, there must needs be some metaphysical
locus of this integration; and the natural thing to say here is that this locus is
something like a micro-structural feature of the sort at issue in the previously dis-
cussed mechanistic accounts, which again plausibly conforms to the conditions in
Weak Emergence.
That said, Gillett expresses three concerns with an account of realization or
Weak emergence involving satisfaction of the Proper Subset of Powers Condi-
tion: first, that such an account is committed to realizing and realized features
being instantiated in the same entity (as per what he calls a ‘flat’ rather than a
‘dimensioned’ approach); second, that such an account is unable to accommodate
qualitative differences between features of macro-entities and their composing
micro-entities (as when, e.g., a diamond, but not a carbon atom, is hard); third (and
related to the previous two concerns), that such an account fails to accommodate
the idea that individual entities and features ‘realize’ configurations and associated
micro-structural features.
I address each concern, in turn. First, the Proper Subset of Powers Condition is
neutral on whether realized and realizing features are had by the same entities.
To be sure, some (including myself, in my 1999) take the condition to apply in
cases of what I there called ‘same-subject necessitation’. But as I now maintain (see
Wilson 2011b and 2015b), this assumption is dispensable. Second, as for qualita-
tive differences between features (and associated powers) of, e.g., carbon atoms
and diamonds: these are understood and accommodated, in the first instance,
as reflecting differences between features of individuals (e.g., carbon atoms) and
features of micro-configurations (e.g., complex arrays of carbon atoms). Against
the backdrop assumption of physicalism, qualitative differences are accommo-
dated not via an inter-level realization relation, but rather via the various intra-
level relations that are operative in bringing individuals at a given level together
into configurations at that level. Third, one may deny that the latter sorts of
intra-level many-one spatiotemporal and causal relations holding between, e.g.,
(many) lower-level physical individuals and their features and (one) lower-level
micro-configuration and its features should be subsumed under the rubric of
‘realization’—most importantly, because this sort of intra-level relation does not
provide a basis for making sense of higher-level reality.
The upshot is that accounts of realization ensuring the Proper Subset of Powers
Condition do not have the problematic implications Gillett attributes to them—a
result he should embrace, since he is arguably also committed to satisfaction of
this condition, if constitutive mechanisms are to be properly integrated.
3.4.3 Constitution
On Pereboom’s (2002) account of “robust” nonreductive physicalism—i.e., of

physically acceptable emergence—a higher-level feature S is neither type- nor
token-identical with the lower-level physical feature P upon which it depends;
and contra both the Token Identity of Powers Condition and the Proper Subset
of Powers Condition, S’s token powers are “irreducible to” powers of P: “robust
nonreductive physicalism affirms various token-diversity claims for mental causal
powers” (500). Such a view will clearly make sense of S’s ontological and causal
autonomy. But how are S’s token-irreducible powers supposed to avoid prob-
lematic causal overdetermination (as discussed in Ch. 2, §2.1.1) while retaining
compatibility with physicalism? According to Pereboom, this is because S’s powers
are “constituted” by P’s powers, in a way piggybacking on the notion of token
feature constitution:
Token Power Constitution: The causal powers of a token of kind F are constituted
of the causal powers of a token of kind G just in case the token of kind F has the
causal powers it does in virtue of its being constituted of a token of kind G.
(504)2⁸
The notion of constitution of one token feature by another is broadly primitive,

but is (as per Pereboom 2011) to be grasped as relevantly analogous to the ‘made
up of ’ relation holding between one particular and another (e.g., a statue and
a lump of clay). The account of feature constitution, coupled with Token Power
Constitution, is intended to motivate taking the powers of a realized feature S to
be, while irreducible to, still nothing over above the powers of P. As Pereboom
puts it:
[Though S’s token powers are irreducible to P’s] there would be a sense in which
the token causal powers of S would be “nothing over and above” the token
causal powers of P […] S’s causal powers would nevertheless be “absorbed” or
“swallowed up” by P’s causal powers. But there are importantly distinct modes
of this sort of absorption: identity and constitution without identity. […] token
mental causal powers are wholly constituted by token microphysical causal
powers. (503–4)
2⁸ See also Pereboom and Kornblith 1991, 131.

(I have changed Pereboom’s notation for continuity with my discussion.)

But such appeals to token feature and power constitution do not establish that
rejection of the Token Identity of Powers Condition, hence Proper Subset of Powers
Condition, is compatible with physicalism. ‘Constitution’ is a term of art, applied
mainly (as Pereboom notes) to objects. Where token features are at issue, and
where conformity to physicalism is presumed, ‘constitution’ is usually just another
name for ‘realization’. But as previously argued, standard accounts of realization
presuppose satisfaction of the Token Identity of Powers Condition. The expression
‘in virtue of ’ entering into the account of token power constitution is also a term
of art, compatible with many underlying relations, including identity (satisfying
the Token Identity of Powers Condition) and the determinable/determinate relation
(satisfying the Proper Subset of Powers Condition).2⁹
Pereboom offers further considerations in support of irreducible mental powers’
being compatible with physicalism and with the avoidance of overdetermination,
but these also fail to establish his case. In re compatibility with physicalism, he says
that “correlated with the possibility of this sort of constitutional explanation is the
fact that the existence and nature of token higher-level causal powers would be
predictable in principle from their microphysical constituents together with the
laws governing them” (504). But if the powers of S, at either the type or token
level, are not identical with the powers of P, what guarantees that the powers of S
would be so predictable? Perhaps such predictability could be guaranteed if every
power of S was type-identical (though token-distinct) with a power of P, but this
understanding appears to give rise to pervasive causal overdetermination.3⁰
Pereboom resists this conclusion, saying that “no competition arises in the case
of mere constitution”:
For if the token of a higher-level causal power is currently wholly constituted by

a complex of microphysical causal powers, there are two sets of causal powers at
work which are constituted from precisely the same stuff […] and in this sense
we might say that they coincide constitutionally. (505)
To the extent that I understand why constitutional coincidence blocks overdeter-

mination, however, this is because (the relevant sort of) coincidence would entail
identity of the token powers had by features of the ‘stuff ’. Pereboom acknowledges
“that they now coincide in this way might tempt one to suppose that these
causal powers are token-identical, but […] there is a good argument that they
are not” (505).
2⁹ One might be tempted to interpret this ‘in virtue of ’ talk in terms of a primitive notion or relation
of ‘Grounding’ or ‘Ground’. This would be a mistake, for reasons I’ll discuss in the next section.
3⁰ See Audi 2012b, 14 for discussion of a similar concern.
Here Pereboom is referring to a line of thought presented in his (2002). There he

first suggests that a token higher-level mental state S might be multiply realizable,
and that if so S would not be identical with its token base state P:
Suppose that S is realized by a complex neural state [P]. It is possible for S to be

realized differently only in that a few neural pathways are used that are token
distinct from those actually engaged. […] [I]t is evident that this alternative
neural realization is itself realized by a microphysical state P′ that is token distinct
from P. It is therefore possible for S to be realized by a microphysical state not
identical with P, and thus S is not identical with P. (503)
He then continues:
[T]his reflection would also undermine a token-identity claim for mental causal
powers—should they exist—and their underlying microphysical causal powers.
For if the token microphysical realization of S had been different, its token
microphysical causal powers would also have been different. We therefore have
good reason to suppose that any token mental causal powers of S would not be
identical with the token microphysical causal powers of its realization. (503)
As I discuss in Wilson 2015b, the Weak emergentist can respond to this line
of thought in either of two ways. First, they can deny that tokens of higher-level
features are ever multiply realizable. What is (fairly) uncontroversially true is that
types of higher-level features are multiply realizable, in the sense that different
tokens of the realized type can be realized by tokens of different lower-level types.
Pereboom’s case, above, for taking token higher-level feature S to be multiply
realizable isn’t compelling, and may be rejected. Compare: could that very instance
of red, currently realized by an instance of scarlet, have been realized by an instance
of burgundy? One might reasonably judge not, and continue to reasonably judge
not even supposing the alternative shade to be only “slightly” different from the
original. One may similarly reasonably deny that S (as opposed to another token of
S’s type) could be realized by a base feature other than P, whether this alternative
feature is very different from P, or only different with respect to “a few neural
pathways” or other lower-level physical details.
Second, the Weak emergentist can grant that a token feature S may be multiply
realizable, but maintain that in such a case, S’s token powers are relativized
to realizers (or occasions). Pereboom presupposes that S has its token powers
essentially; but why think this? As he later observes, “stable tokens […] often retain
their identity over certain changes in their constitutions and configurations” (529).
One might similarly maintain that token feature S can retain its identity across
changes in its realizers and associated token causal powers. The Proper Subset
of Powers Condition requires only that S’s token powers on a given occasion be
numerically identical with the powers of the lower-level physical feature realizing
it on that occasion; hence the Proper Subset of Powers Condition can be satisfied
even if S is token multiply realizable.
3.4.4 Grounding
It has recently been suggested that complete metaphysical dependence can or

should be understood in terms of a primitive relation or notion—what I’ve called
(‘big-G’) ‘Grounding’, to distinguish it from historically common schematic uses
of the terms ‘grounding’ (or ‘ground’)—which is supposed to be operative in any
context where some goings-on or facts hold ‘in virtue of ’ some others (see Fine
2001, Schaffer 2009, Rosen 2010, Audi 2012a, Raven 2015; see Bliss and Trogdon
2016 for an overview and other references). Though proponents disagree about
which formal features are characteristic of Grounding, it is commonly supposed
that Grounding is (at a minimum) asymmetric and irreflexive, in which case
Grounded goings-on are ontologically autonomous (distinct) from Grounding
goings-on.
One of the main suggested applications of Grounding (see, e.g., Schaffer 2009,
Rosen 2010, and Dasgupta 2014) is as providing a basis for formulating non-
reductive physicalism, as follows:
Physicalism (Grounding): All broadly scientific goings-on are Grounded in lower-

level physical goings-on.
The suggestion is supported by appeal to the following form of argument:31
1. Physicalism is the thesis, schematically speaking, that all broadly scientific

goings-on are completely metaphysically dependent on (‘nothing over and
above’, etc.) lower-level physical goings-on.
2. The operative notion of complete metaphysical dependence cannot be suc-
cessfully characterized in semantic/representational, epistemic, or purely
modal (i.e., supervenience-based) terms.
3. No other non-primitive approach to characterizing complete metaphysical
dependence is available.
∴ The operative notion of complete metaphysical dependence in physicalism
should be characterized in terms of Grounding.
31 See Schaffer (2009, 364), Rosen (2010, 111–12), and Dasgupta (2014, 557).
Coupled with the usual supposition that Grounding is asymmetric and irreflexive,
such an argument might be seen as offering an account of the realization relation
making no reference to the Proper Subset of Powers Condition, and so as suggesting
that satisfaction of the conditions in Weak Emergence is not necessary for emer-
gence of a physically acceptable variety.
As I’ve argued in a series of papers (Wilson 2014a, 2016d, 2018/2016), there
are many ways in which the lines of thought leading to the posit of Grounding
and its application to physicalism go wrong. A full discussion of my concerns here
would take us too far afield; in what follows I simply sketch two concerns which
are relevant to the question at hand, and direct those interested in further details
to my other work on the topic.
First, as I note in Wilson 2014a, arguments like that above for positing Ground-
ing are unsound, since premise (3) is false. As will already be clear from discussion
in this and previous chapters, during the past several decades philosophers work-
ing on physicalism have identified and explored many specific relations of the sort
that might enter into formulating physicalism, either in general or as applied to a
specific case, which relations are explicitly assumed to go beyond representational,
epistemic, or merely modal notions. These small-‘g’ grounding relations (as I call
them in my 2014a)—including type and token identity, functional realization,
constitutive mechanism, the part/whole relation, the determinable/determinate
relation, the proper-subset-of-powers relation, and so on—serve, against the
backdrop assumption that the physical goings-on are the sole fundamenta, to
characterize diverse forms of physical dependence (some reductive, some nonre-
ductive) in an explanatory and illuminating way. Given the availability of these
properly metaphysical alternatives, there is no route from the failure of epistemic,
representational, and/or merely modal (i.e., supervenience-based) conceptions of
metaphysical dependence to a primitive Grounding-based understanding of this
notion (much less to one tracking just metaphysical dependence of a nonreductive
variety). Other motivations for positing primitive Grounding have been advanced,
but as I have argued, these are also uncompelling.32
32 One stated motivation appeals to unity considerations (see Schaffer 2016b and 2016a, Berker
2017), with the basic idea being that Grounding is needed in order to formally and otherwise unify
the small-‘g’ relations. But as I and others have argued (see Wilson 2014a, 2016d; Koslicki 2012,
2016), the small-‘g’ relations are not formally or otherwise unified. Even if they were unified, more
would be required to establish that such unity motivates a generic worldly posit, as opposed to merely
motivating a schematic or generic concept. To highlight just one case in point, determinable features
formally and otherwise unify determinate features, but it is nonetheless commonly supposed either
that determinables don’t exist or that they are reducible to a disjunctive or other combination of
determinates (see Wilson 2017 for discussion). My own view is that determinables do exist and may
be as fundamental as their determinates (see Wilson 2012), but my point here is that this requires
argument: unity considerations alone aren’t enough to motivate a metaphysical posit, much less a
primitive one.
Second, even granting that Grounding exists, the lack of non-stipulative content
associated with this primitive posit renders it incapable in itself of shedding any
illuminating light on the notion of metaphysical dependence it is introduced as
explicating.33 Most importantly, appeal to Grounding in itself provides no basis for
a physicalist response to the problem of higher-level causation, according to which
the effects of a higher-level entity or feature are causally overdetermined, since
already brought about by base-level entities or features. As we’ve seen, considerable
nonreductive physicalist action concerning this problem has involved identifying
one or other specific relation (e.g., functional realization, constitutive mechanism,
the part/whole relation, the determinable/determinate relation, and so on) whose
holding would illuminate how, exactly, some higher-level entities or features
might be both physically acceptable and distinctively efficacious in a way avoiding
problematic (double-rock-throw-type) overdetermination. By way of contrast, the
holding of a primitive Grounding relation says nothing—at least, nothing non-
stipulative—about whether and how causal overdetermination might be avoided
as between higher- and lower-level entities and features, or even about whether a
Grounded entity has any powers at all; hence in itself this relation fails to provide
a basis for responding to the main barrier to making sense of physically acceptable
emergence. Correspondingly, a Grounding-based approach to Weak emergence
does not satisfy the criterion of illuminating accommodation—that is, does not
provide a clear and illuminating basis for accommodating the appearances of
metaphysical emergence in natural (straightforward, default) fashion.
All this said, nothing prevents a proponent of Grounding from stipulating or
otherwise imposing a further condition as doing that work. One option would be
to stipulate that in general or in the case at hand a Grounded feature has powers
which are moreover a proper subset of those of its Grounding feature, on any given
Another stated motivation appeals to a purported need for Grounding to fix the direction of priority
of instances of the specific relations, since (as is illustrated by the dispute between Monists and atomists
over whether the Cosmos qua whole is prior to or posterior to its proper parts) instances of these
relations do not come with a built-in direction of priority. But as discussed in Wilson 2014a, 2016d,
2018/2016, and in progressb, what more is needed is not a primitive pointer (i.e., Grounding) but
rather a specification of what is or serves as fundamental. For example, in the case above: given
that the Cosmos is the sole fundamental entity, proper parts of the Cosmos are nonfundamental,
whereas given that the atomic parts are the sole fundamenta, mereological fusions of the parts
(including the Cosmos) are nonfundamental. For another example: given that the basic physical
goings-on are the sole fundamenta, then lower-level configurations of the basic physical goings-on
are nonfundamental, as are higher-level goings-on standing in various specific relations (functional
realization, constitutive mechanism, the part/whole relation, the determinable/determinate relation,
and so on) to such lower-level configurations of physical goings-on. Priority between nonfundamenta,
where it exists, is then a function of various considerations, including considerations registering the
nature of the nonfundamenta as determined by how they small-‘g’ stand to the fundamenta, and
considerations registering how the nonfundamenta are related to each other.
33 See Bryant 2020 for a recent broadly pessimistic assessment along these lines of Grounding’s
usefulness for formulating physicalism.
occasion. Another option would be not so different from the one that Horgan
(1989) and Lepore and Loewer (1987 and 1989) implement—namely, to require
that, in addition to the holding of the (schematically characterized) realization
relation at issue, certain difference-making conditions be in place, as reflected
in the truth of certain subjunctive conditionals. Either way, however, it would
turn out that Grounding qua realization would conform to the schema for Weak
emergence—in which case that relation, even granting that it exists, poses no threat
to taking satisfaction of the conditions in Weak Emergence to be necessary for
physically acceptable emergence.
3.5 Concluding remarks
In Ch. 2, I considered and responded to what is perhaps the most pressing concern
with the viability of physically acceptable emergence—namely, that associated
with Kim’s problem of higher-level causation—with the key components of the
response being the identification of the Proper Subset of Powers Condition lying
at the heart of the schema for Weak emergence, and the related appreciation of
a second way (different from that associated with the having of a new power)
in which a higher-level feature might be distinctively efficacious—namely, by
having a distinctive power profile, tracking difference-making considerations or
a comparatively abstract system of laws or causal joint.
In this chapter, I have considered a wide range of objections to the viability of
Weak emergence, according to which satisfaction of the conditions in the schema
is compatible with anti-realism, reductionism, or physical unacceptability, or is
not necessary for physically acceptable emergence; and I have argued that each
objection admits of one or more responses available to proponents of any of the
diverse implementations of the schema.3⁴ In two cases, additional responses are
available which rely on features specific to either a determinable-based or a DOF-
based implementation of Weak Emergence. In particular, a determinable-based
account (and perhaps also a DOF-based account) provides a non-stipulative basis
for ruling out the Weak emergence of conjuncts from lower-level conjunctions,
and a DOF-based account explicitly includes a condition specifying that the
holding of any constraints entering into or responsible for the holding of the Proper
Subset of Powers Condition must be a matter only of physically acceptable (more
3⁴ Additional objections and responses to the viability of Weak emergence were more conveniently
treated in Ch. 2. See my §2.3.1 response to Kim’s (2006) claim that realized features inherit all, not just
some, of the token powers, my note 27 response to Morris’s (2013) attempted motivation of Kim’s claim
(as per what Morris calls ‘Full Inheritance’), and my note 35 response to Morris’s (2011) concern that
the motivations for taking satisfaction of the Proper Subset of Powers Condition to block problematic
overdetermination hinge on a problematic analogy to cases of part/whole overdetermination.
generally, base-level) processes. Again, my sense is that if any of the considered

objections requires tweaking the conditions in Weak Emergence, it is that (due
to Melnyk) pertaining to the need to require the physical acceptability of any
operative constraints associated with the higher-level feature and its distinctive
power profile. The addition of such a condition would be straightforward, and
hardly fatal to the overall approach; but again, this is a choice point, and how
exactly a Weak emergentist chooses to respond to this or other objections may
depend on further of their commitments. In any case, these results collectively
indicate that Weak Emergence is not just a viable and indeed attractively robust
means of accommodating physically acceptable emergence, but that this schema
moreover captures what is core and crucial to (and when suitably filled in, what is
necessary and sufficient for) such emergence.
4
The viability of Strong emergence
I now turn to a project similar to that of the last chapter, only as directed at
the schema for Strong emergence. Recall that in Ch. 2, I provided prima facie
reasons for thinking that satisfaction of the conditions in the schema for Strong
emergence is core and crucial to a feature’s being metaphysically emergent from—
cotemporally materially dependent on, yet ontologically and causally autonomous
with respect to—a lower-level feature, in a way incompatible with physicalism,
given that the base feature is physically acceptable. Again, the schema is as
follows:

emergent from token feature P on a given occasion is for it to be the case, on that
occasion, (i) that S cotemporally materially depends on P, and (ii) that S has at
least one token power not identical with any token power of P.1
Henceforth, by ‘Strong emergence’, I will intend to refer to the view (or the
phenomenon, understood in light of the view) that conformity to the conditions in
Strong Emergence is core and crucial—and when sensibly filled-in, both necessary
and sufficient—for emergence of the sort incompatible with physicalism.
In this chapter, I consider and respond to a range of objections that have been
or could be made to the viability of Strong emergence, so understood. These
objections fall into four main categories, according to which satisfaction of the
conditions in Strong Emergence is incompatible with scientific theory or practice
(§4.1); is impossible, since any purportedly novel powers of Strongly emergent
features are inherited by (or “collapse” into) base features (§4.2); is compatible
with physical acceptability (§4.3); or is not necessary for emergence of a physically
unacceptable variety (§4.4). As I’ll argue, each of the objections admits of one or
1 Recall that the novelty of powers at issue in Strong Emergence is fundamental novelty, but that
the qualification that the novelty is fundamental rather than nonfundamental (stemming merely
from various less complex goings-on entering into certain aggregative relations) is not needed, given
that P, as a feature of a lower-level configuration (plurality or structural aggregate), has any such
nonfundamental (merely aggregative) powers. Recall also that for a feature’s being Strongly emergent
it plausibly suffices that the conditions are satisfied on at least one occasion by at least one instance of
the feature in worlds with laws relevantly similar to the actual laws; for continuity with the schema for
Weak emergence I stick with the schema expressed in terms of occasions.
DOI: 10.1093/oso/9780198823742.003.0004
the viability of strong emergence 121
more responses that could be endorsed by any proponent of Strong emergence,

however the schema is implemented. And as in the case of Weak emergence,
upon occasion an additional response to an objection is available which relies on a
specific implementation of the schema. In particular, certain attractive responses
appeal to a fundamental interaction-relative account of Strong emergence, of the
sort I have previously endorsed.
4.1 Objection: incompatibility with scientific theory or practice
Strong emergentists maintain that some higher-level features have new powers,
associated with higher-level laws, forces, or interactions that are as metaphysically
and scientifically fundamental as the fundamental physical powers, laws, forces
and interactions. In terms of forces, for example: a Strongly emergent feature might
be supposed to have a new power to produce a fundamental ‘configurational’
force—that is, a force present only upon the occurrence of certain lower-level
configurations (pluralities or structural aggregates). When such forces are present,
the operative forces are a combination of physical and Strongly emergent forces.
The powers at issue might include powers to contribute to causing some special-
scientific effect, but in any case will, for reasons brought out in Ch. 2 (§2.2.1),
include powers affecting the motion of lower-level physical entities, conforming
to a common understanding of Strong emergence as involving the nomological
possibility of ‘downward causation’. Several concerns with Strong emergence, so
understood, have to do with its being in tension with the content or practice of
scientific theories; I address these in turn.
One such concern is that Strong emergence is incompatible with contemporary
physics. In response, McLaughlin (1992) convincingly argues that Strongly
emergent configurational forces or interactions (of the sort posited by the British
Emergentists, in particular) are compatible with the laws and conservation
principles of physics (see also Horgan 1993 and Papineau 2001). For example,
Newton’s second law of motion, F = ma, is neutral as regards which component
forces enter into the net force F. As Kane (1993) notes:
F = ma is used to compute the motion of an object, given any force F on the

object. And specific classical forces have been discovered, such as gravity with
G mM
F = N 2 […] . Hamilton’s or Lagrange’s equations are equivalent to F = ma in
r
a different formulation. In quantum theory there is an analogous structure. The
Schrödinger equation […] is like F = ma. It holds for any Hamiltonian. Specific
forces lead to specific Hamiltonians. (2–3)
Accordingly, this law is compatible with the operative forces’ including a fun-
damental configurational force. Similarly, McLaughlin (1992) observes, for the
𝜕ψ
contemporary descendant of Newton’s law, Schrödinger’s equation Hψ = ih ,
𝜕t
into which is inserted the Hamiltonian H specifying the energies of the state (forces
and energies being inter–translatable2):
It is not that British Emergentism is logically incompatible with nonrelativistic

quantum mechanics. It is not. Schrödinger’s equation could be the fundamental
equation governing motion in a world with energies that are specific to types of
structures. (54)
Nor are fundamental configurational forces (more generally: interactions) or

energies incompatible with conservation laws, such as the relativistic principle of
conservation of mass-energy. As McLaughlin notes:
[C]onfigurational forces need not involve any violation of this principle. […]
Configurational forces could involve various compensating shifts in mass and
energy that maintained conformance to the principle of mass-energy. (74)
A seemingly different defense of the scientific compatibility of Strong emer-

gence is offered by Cucu and Pitts (2019), who suggest that were any form of
interactionist dualism to be true (as regards minds, in particular), this would
violate energy conservation—but that such a violation would not be problematic,
but rather just what one would empirically expect. As they put it:
The view defended here is that the best response to the energy conservation
objection—the response that reflects an understanding of the relevant theoretical
physics—is what has been called the ‘conditionality response’ (Pitts 2020) that
energy is conserved when and where minds do not act on bodies, but is not
conserved when and where minds act on bodies. (100)
Cucu and Pitts’s understanding of the relevant conservation law is one, however,
which builds in that the conserved energy is physical (is ‘physically conserved’);
hence their view is not really in tension with the previous response, according
to which mental or other Strongly emergent energies or forces would not violate
conservation laws, neutrally understood. In any case, their interesting survey of
scientists—including Descartes, Newton, and Euler—who appear to have counte-
nanced the possibility of nonphysical energies or forces provides further support
for thinking that Strong emergence is not incompatible with physical science.
A second concern is that even if there is nothing in-principle problematic about
adding fundamental configurational forces/interactions or energies to the mix of
2 See Wilson 2007 for discussion and some derivations.

physical forces/interactions or energies entering into the operative equations of

motion, such posits are incompatible with scientific practice, or relatedly, with a
‘naturalist’ outlook, according to which metaphysical investigations and accounts
should be consonant with such practice. As I discuss in Wilson 2002a, however,
scientific theorizing itself provides a blueprint for how conservation laws might
enter into the warranted posit of Strongly emergent configurational interactions
and associated powers, laws, and features. In the 1930’s, the law of conservation of
mass-energy appeared to be violated in nuclear β-decay interactions. Rather than
accept the apparent violation as genuine, physicists posited a new fundamental
interaction—the weak nuclear interaction—as carrying away the missing energy
(see Greiner 1996 for discussion). Nuclei are composite entities; hence evidently
scientists have no problem with positing fundamental configurational forces/in-
teractions. As it happens, the nuclear interactions are now understood as ulti-
mately due to interactions between sub-nuclear entities, but the point remains: the
posit of fundamental configurational forces/interactions and associated powers,
which come into play only at certain levels of comparatively complex organization,
is compatible with scientific practice, and hence with a naturalistic approach to
metaphysical theorizing.
A third concern is that, even granting that there is nothing in-principle prob-
lematic about there being fundamental configurational novelty (of powers, forces/
interactions, laws), at present science provides no empirical support for such
posits. This is the considered judgement of McLaughlin (1992), who maintains
that there is “not a scintilla of evidence” in favor of there being Strongly emergent
features; Ladyman and Ross (2007) register a similar opinion. Strictly speaking,
such a lack of evidence would not constitute an objection to the viability of Strong
emergence per se, as opposed to its actual applicability. Even so, it’s worth noting
that the claim that we presently don’t have any reason to think that there is
any Strong emergence is overstated. That certain historical candidates for Strong
emergence (e.g., the sorts of chemical interactions that Broad discussed) now
admit of lower-level physical explanations doesn’t show that all such candidates
have been so explained. Indeed, as we will see in later chapters, there continues to
be considerable debate over whether phenomena such as consciousness and free
will are Strongly emergent, in ways in keeping with a naturalistic methodology.3
3 Two other recently discussed cases are worth mentioning here. First, Silberstein and McGeever
(1999, 187–9) offer quantum entanglement as a case of emergence (see also Humphreys, 1997, 216);
second, Hendry (2010) argues for “the ontological emergence of molecular structure with respect to
quantum mechanical systems of nuclei and electrons interacting via Coulomb forces” (220; see also
Hendry 2017). As discussed in Ch. 2 (§2.2), the operative accounts of emergence in these discussions
are reasonably supposed to be of the Strong variety. That said, I postpone detailed discussion of these
purported cases of such emergence for a future occasion (see the to-do list at the end of Ch. 9), in part
because quantum entanglement is both poorly understood and not among the target cases of inter-level
emergence, and in part because certain difficulties attach to the stated motivations for each purported
case of Strong emergence. The considerations Hendry offers in in his (2010) appear to presuppose that
A final concern about the compatibility of Strong emergence and scientific

theory and practice has to do with the usual understanding of such emergence as
involving commitment to the nomological possibility of downward causation.
Wouldn’t the efficacy of Strongly emergent properties vis-à-vis physical or
physically acceptable effects violate Physical Causal Closure, according to which
every lower-level physical effect has a sufficient purely lower-level physical cause?
And isn’t Physical Causal Closure widely accepted? Yes, and yes. However, Physical
Causal Closure is not a principle of contemporary physics (though no doubt
many physicists, and scientists more generally, believe it). The acceptance of this
principle is rather a constraint on physicalist theorizing (which motivates, in
particular, reductive and nonreductive physicalist approaches to the problem of
higher-level causation); hence that Strong emergentists deny it is not in itself a
strike against their view.
4.2 Objection: collapse
The second line of objection aims to show that the conditions in Strong Emergence
are never jointly satisfied, as per what Taylor (2015) evocatively calls the ‘col-
lapse’ objection, versions of which have been raised and/or addressed by several
philosophers (see, e.g, van Cleve 1990, Kim 1999, O’Connor 1994, Wilson 2002a,
Francescotti 2007, Howell 2009, Taylor 2015, and Carruth 2018). The general
concern is that Strong emergence makes no sense, since any purportedly Strongly
emergent features or associated powers ‘collapse’, one way or another, into the
lower-level base features upon which they depend, undermining the supposed
ontological and causal autonomy of the emergent features.⁴
it suffices for molecular structure to be Strongly emergent that the structure (or associated features) is
not determined solely by the locally interacting parts. In particular, he says “the explanation of why
molecules exhibit the lower symmetries they do would appear to be holistic, explaining the molecule’s
broken symmetry on the basis of its being a subsystem of a supersystem (molecule plus environment).
This supersystem has the power to break the symmetry of the states of its subsystems without acquiring
that power from its subsystems in any obvious way. That looks like downwards causation” (215–16).
However, as I observed in Ch. 3 (§3.3.3) when considering the bearing of spatiotemporally non-local
higher-level features on satisfaction of the conditions in Weak Emergence, such non-locality might
be seen as indicating just that the dependence base feature must be appropriately spatiotemporally
extended. (See also note 8 of this chapter for another observation in re Hendry’s view.) The main
consideration offered by Silberstein and McGeever in their (1999) is that in the case of entanglement
“the parts exhibit a holistic (emergent) correlation property possessed by the system but not locally
carried by the separate parts” (187), but as discussed in Ch. 1 (§1.4.2), the emergence at issue here
concerns the possibility of features that are novel with respect to features of underlying configurations,
not just with respect to features of individuals entering into the configurations. Again, there is more to
say here but given the present target this will have to do.
⁴ A related concern is that such collapse, combined with the supposed physical unacceptability of
Strongly emergent features and powers, threatens the physical acceptability of the base features (see
Howell 2009). Since Howell offers this objection in support of a supervenience-based approach to
physically unacceptable emergence, I postpone discussion of his concern until §4.4.2.
As discussed in Baysan and Wilson 2017, there are two main versions of the
objection, to be discussed in the following sections. After presenting these versions
and noting, along the way, certain difficulties with previous responses, I offer more
promising responses, drawing on Wilson 2002a and Baysan and Wilson 2017.
4.2.1 Collapse via power possession
The first route to the collapse objection is one according to which an intuitive
way of assigning powers to features entails that any purportedly new power of
a Strongly emergent feature S will be inherited by its base feature P. The concern
here underlies what Kim (2006) calls the “critical” question of emergence:
M, as an emergent, must itself have an emergence base property, say P. Now we

face a critical question: if an emergent, M, emerges from basal condition P, why
cannot P displace M as a cause of any putative effect of M? […] If causation is
understood as nomological (law-based) sufficiency, P, as M’s emergence base, is
nomologically sufficient for it, and M, as P*’s cause, is nomologically sufficient
for P*. It follows that P is nomologically sufficient for P* and hence qualifies as
its cause. (558)
Given that, as we are assuming, a feature’s powers are a matter of what effects the
having of that feature can contribute to causing, when in certain circumstances,
the threat to the viability of Strong emergence is clear. For the Strong emergen-
tist standardly supposes that the cotemporal material dependence of a Strongly
emergent feature S involves, at a minimum, the base feature’s being nomologically
sufficient for S; moreover, nomological sufficiency (in the circumstances, as per
usual) is transitive. Consider, then, any power of S to contribute to causing an
effect E in circumstances K. If causation is a matter of nomological sufficiency in
appropriate circumstances, if P is nomologically sufficient for S in K, and if S is
nomologically sufficient for E in K, then P will also be nomologically sufficient for
E in K, and so also have the power to contribute to causing E in K, ruling S’s Strong
emergence out of court. The upshot is that any supposedly novel powers of S will
‘collapse’ into those of P.
Insofar as the line of thought here depends on certain assumptions about causa-
tion, one might wonder whether the Strong emergentist can respond by rejecting
a view on which nomological sufficiency in the circumstances is sufficient for
causation, or by denying that causation is transitive, or by denying that causation
can be (as with P and S) cotemporal. Such responses are unsatisfactory, however.
To start (following Hall 2004), accommodation of many intuitive cases of causation
requires a notion of causation as ‘production’, involving nomological sufficiency in
the circumstances; moreover, other accounts of causation seem likely to introduce

similar or other difficulties.⁵
Most importantly, even if it is possible to block taking P to cause S in cases
of Strong emergence, there would remain a case to be made that P inherits any
powers of a feature S that at least nomologically cotemporally depends on P. Here
O’Connor’s (1994) presentation of the following “strong objection” to a powers-
based account, which he credits to Carl Ginet, is apropos:
If an emergent property is a necessary consequence of certain base-level prop-

erties (as is implied by the supervenience [i.e., dependence] condition), then
its instantiation is one of the potentialities of that set of properties. But then
are not the further potentialities of this emergent property also a subset of
the total set of potentialities of the base properties, in virtue of the necessary
connection between the base properties and it? These further potentialities are
simply potentialities of the base properties at one remove. And now one is led to
wonder why we might ever think to postulate an emergent property at all, since
it provides no explanatory gain over an account which excises the mediating link
by taking the “further” potentialities as directly tied to the base properties. This
objection implies, in effect, that the features of supervenience and novel causal
influence are incompatible. (98)
The deeper collapse objection raised here, as well as in Kim’s (1998) framing of the
objection, does not hinge on the supposition that Strong emergence can be seen as
a causal relation, but rather just on the supposition that P cotemporally necessitates
S, with at least nomological necessity. Such necessitation alone suggests that
anything that S can do in circumstances K is also something that P can do in
circumstances K, in which case there is no way for S to have a novel power, and so
no way for it to be Strongly emergent.
O’Connor (1994) offers a response to the deeper collapse concern, but it is less
than satisfactory. He suggests that if P is taken to inherit S’s powers, then the
⁵ Most saliently, if causation is counterfactual dependence (the other main category of causation
that Hall identifies), such that a power is associated with a feature only if the associated effect is coun-
terfactually dependent on the feature, two difficulties ensue. First, if the counterfactual dependence
concerns the token instances of S, P, and E, then it might be reasonably thought that if S’s power to
cause E reflects S’s being necessary in the token circumstances for E, then P also has the power to cause
E, in being necessary in the token circumstances for S. So collapse remains. (Note also that the previous
line of thought is not affected by taking counterfactual dependence to be relative to a given contrast
class.) If the necessity rather attaches to the types at issue, then a different problem arises—namely, that
any higher-level feature with multiple dependence bases will be deemed Strongly emergent, including
multiply realized features that are intuitively physically acceptable. It shouldn’t be that easy to falsify
physicalism! And while it would be less costly to deny that causation must be transitive or to require
that it be diachronic, these denials are both overly committing and ad hoc.
lower-level physical laws would have a “very odd complexity, involving tacked-
on disjuncts to cover the special cases” (98). Effectively, O’Connor’s suggestion is
that collapse would entail that lower-level entities (e.g., atoms) would interact with
each other in a uniform way until entering into a complex aggregate, at which point
they would (collectively) start doing “quirky” things, and that such discontinuous
behaviors are better explained by positing Strongly emergent features. O’Connor’s
response presupposes that complex behavior, if it is to be physically acceptable,
must be smoothly aggregative. That’s incorrect, however, for physicalists (reduc-
tive or nonreductive) are happy to allow that quirky behavior can come about
simply as a result of complexity (as with, e.g., chaotic nonlinear systems).
4.2.2 Collapse via lower-level dispositions
A second version of the collapse objection focuses on the question of when a

feature is appropriately placed at the presumed lower level of physical goings-on.
An early version of this objection is registered by van Cleve (1990), who after
arguing that physically unacceptable emergence represents the best option for
making sense of dependent but irreducible higher-level mental features, says of
Broad’s ‘in-principle failure of deducibility’ account:
There is one more point about Broad’s account that needs to be discussed. It could
be objected to what has so far been said that there is simply no room for the
concept of an emergent property, since for any property P of any whole w, there
will always be properties of the parts from which P may be deduced. For example,
is it not true of sodium that it comes with chlorine to form a whole having such-
and-such properties, including its odor and anything else one might have claimed
to be emergent? And from such properties of the parts, may not all properties of
the whole be deduced? The answer, of course, is yes; but it is also clear that if
properties of this sort are admitted in the “supervenience base,” the doctrine of
anti-emergence […] becomes completely trivial. (223–4)
Taylor (2015) develops this line of thought, observing that Broad took sodium
chloride to be Strongly emergent (that is, to have fundamentally novel powers,
etc.), on grounds that from complete knowledge of the properties of sodium and
chlorine in isolation, or in compounds different from that associated with sodium
chloride, one could not deduce that salt will dissolve in water. But, Taylor argues, it
seems that dispositional properties are among the features that can be had by the
components “in isolation”, in which case the characteristic features and associated
powers of sodium chloride will be deducible, after all:
This case of emergence ‘collapses’ when […] dispositional properties are included
among the micro-level properties. […] For example, one of the characteristic
properties of sodium chloride is its solubility in water. Accordingly, sodium has
the following dispositional property: to generate a compound that is soluble in
water when combined with chlorine into sodium chloride. In Broad’s terms, this
property is a property of sodium ‘in isolation’. […] The emergent features of
the whole R(A, B, C) can obviously be deduced from complete knowledge of the
features of the parts A, B, and C and the knowledge that they are arranged as a
whole R(A, B, C), so long as the features of the parts include these dispositional
properties. (736)
Taylor sees a general problem here for accounts of Strong emergence:
[C]ases of emergence presuppose a distinction between micro-level and macro-

level properties. For any purported case of emergence, there are properties that
prima facie belong to the micro level, but if they are included in the micro
level then the purported emergent fails to meet a necessary condition for emer-
gent autonomy. I call these problematic properties collapse-inducing properties
because when they are included in the micro level, the purported emergent
effectively ‘collapses’, and yet it seems arbitrary to exclude them. […] This is the
problem of collapsing emergence (or, for short, the collapse problem).
(732–3, emphases in the original)
Again, the problem such a ‘dispositional move’ poses for the viability of
Strong emergence is clear. Both van Cleve and Taylor focus on Broad’s ‘failure
of deducibility criterion’, but as above, the intended import of this criterion is
to track the fundamental novelty of a Strongly emergent feature, as reflected in
such a feature’s having powers not had by the lower-level physical features upon
which it depends (or, for that matter, by any other lower-level physical feature).
And notwithstanding that the dispositional features of the ‘isolated’ lower-level
entities at issue in van Cleve’s and Taylor’s discussions are, to use O’Connor’s
(1994) terminology, at various ‘removes’ from either P or S (understood, as per
usual, as features of lower-level configurations or of macro-entities, respectively),
nonetheless such dispositions call into question the intended fundamental novelty
of a Strongly emergent feature. Here again O’Connor’s (1994) discussion is useful
in highlighting the deeper concern at issue, which he credits to Sydney Shoemaker
(pers. comm.), according to which one can always insist that purportedly Strongly
emergent features are in fact “further (hitherto undetected) micro-properties”
which are manifested only in certain complex circumstances.
O’Connor (1994) and van Cleve (1990) offer responses to this version of the
collapse objection, but these responses are again less than satisfactory.
O’Connor maintains that it would be “implausible” and ad hoc to posit micro-

properties that make their presence known only in highly complex systems: “the
only motivation one could have for postulating [such a] micro-property is a very
strong methodological principle to the effect that one is to avoid emergentist
hypotheses at all costs” (98). But it seems clear that in general, dispositions “make
their presence known” only when certain conditions are in place, and sometimes
such conditions might well involve highly complex states of affairs (here again
cases of complex nonlinear but presumably physically acceptable behavior are
relevant); so the mere fact that micro-dispositions would manifest in complex
circumstances is not enough to show that the collapse-inducing suggestion is
either implausible or ad hoc.
Van Cleve suggests that restricting the base features to those that are manifested
in non-emergence-engendering combinations might do the trick:
Clearly, some sort of anti-triviality stipulation is required. Perhaps the required

work can be done by Broad’s phrase “taken separately and in other combinations,”
for one could plausibly refuse to regard the property “forming a whole with such-
and-such features when combined with chlorine” as a property of sodium taken
separately. (223)
But it is not clear that Broad’s qualification provides a basis for plausibly refusing
to regard the property “forming a whole with such-and-such features when com-
bined with chlorine” as a property of sodium “taken separately”; for it is commonly
assumed (see, e.g., Martin 1996) that dispositions can be had by individuals even
when the dispositions aren’t being manifested: a vase can be fragile, for example,
even if it is never broken. Taylor considers another response—namely, to require
that lower-level features be non-dispositional. As she correctly notes, however,
this would be overly restrictive, since many uncontroversially lower-level physical
features—e.g., having a mass of 5 g—are to some extent dispositional.
4.2.3 Three responses to the collapse objection(s)
I now present three more satisfactory responses to the collapse objection(s); again,
see Wilson 2002a and Baysan and Wilson 2017 for further discussion.
Direct vs. indirect powers

Perhaps the simplest line of response is one distinguishing between direct and
indirect having of powers. Here the Strong emergentist grants that while in cases
of Strong emergence there is a loose sense in which P or other lower-level physical
features inherit S’s purportedly new power (due either to P’s being nomologically
sufficient for S, or to P or some other lower-level features’ being disposed to give

rise to S), in a stricter sense S’s novel power is not had or manifested by lower-level
features in the same direct or immediate way as they are had or manifested by S.
Notwithstanding that P (at least nomologically) necessitates S, P has the power
at issue only in that P is a nomologically sufficient precondition, in the circum-
stances, for S, which is the more direct locus of the power. Similarly for lower-level
dispositions of isolates which are even further removed from S than is P: granting
that there are such dispositions and that these in some sense refer to S and its
novel power, the notion of disposition here is again simply that of a precondition
or precursor of a feature (namely, S) that more directly has the power in question.⁶
Such a strategy seems intuitively well-motivated, given the Strong emergentist
understanding of lower-level physical goings-on as being precisely such nomolog-
ically sufficient preconditions for Strongly emergent features. There are, moreover,
two ways to substantiate the intuition and associated strategy.
First, the Strong emergentist can appeal to an analogy to temporally extended
causal chains: even if each link in the chain is, in the circumstances, nomologically
sufficient for the next link, one can nonetheless distinguish more and less direct
causes of the end result; and the mere fact that, say, a person lights (or could
light) a fuse leading to the explosion of some fireworks doesn’t entail that the
explosion isn’t a novel phenomenon, or that there is any but an indirect sense
in which that person has the power to produce such an explosion. Similarly, the
Strong emergentist can maintain that in cases of Strong emergence the base feature
P is metaphysically, if not temporally, antecedent to S in the chain of feature
instantiations potentially leading to the effects associated with S’s novel power.
Second, the Strong emergentist can appeal to an analogy to sets and subsets to
make the notion of the cotemporal yet indirect having of a power more precise,
as reflecting different circumstances associated with S and with P with respect to
the having of the power at issue. As is uncontroversial, powers are individuated, in
part, by the circumstances in which they manifest and contribute to the production
of a given effect; but just as we can distinguish between a set and its subsets
at a time, there seems to be no in-principle reason why we cannot distinguish
between different sets of circumstances associated with a single temporal interval
(instantaneous or extended). In particular, the Strong emergentist can say that P
has the power to contribute—nomologically, if not causally—to the production of
S, in circumstances K which do not include the presence of S. In virtue of having
⁶ Note the contrast here with Shoemaker’s suggestion as discussed in Ch. 3 (§3.1) that in cases of
physical realization (Weak emergence), certain of the powers of a realized feature may be had in a more
‘direct’ way than as had by the realizing feature. As previously discussed, given physicalist acceptance
of the theses of Physical Causal Closure and, relatedly, the Token Identity of Powers Condition, there isn’t
any room for such a distinction between ways of having powers to get a grip. The Strong emergentist
rejects both theses, however, and so can accommodate this distinction.
this power, P indirectly has the power to contribute to causing anything that S can
cause. By way of contrast, S has at least one power to contribute to the production
of a given effect E—namely, its novel power—directly, which power is manifest in
circumstances K′ which, whatever else they might be or contain, do not include
the absence of S.
Perhaps the main concern with the direct/indirect having strategy is that there
may be some indeterminacy in what counts as direct (as opposed to indirect)
having or manifestation of a power, just as there might be indeterminacy as regards
which link in a causal chain is most temporally proximal to a given effect. Here
the Strong emergentist has two responses. First, they can maintain that, as per
usual, the presence of indeterminacy or borderline cases needn’t undermine the
usefulness of a given distinction. Second and relatedly, they can avail themselves of
one or other of two strategies for accommodating indeterminacy in properly meta-
physical (as opposed to merely semantic or epistemic) terms: first, a metaphysical
supervaluationist approach along lines of Akiba (2004), Barnes (2010), Barnes
and Williams (2011), and others; second, a determinable-based approach along
lines of Wilson (2013a and 2016a), and recently applied by Bokulich (2014), Wolff
(2015), and (Calosi and Wilson, 2019 and forthcoming) to the case of quantum
metaphysical indeterminacy. In Ch. 6 (§6.1.3) I’ll discuss these accounts in more
detail.
Powers relativized to fundamental interactions

A second response to the collapse problem appeals to an independent way of
sorting powers, based in the notion of a fundamental interaction, which makes
room for higher-level features to have powers that are in some sense new, as Strong
emergence requires (see Wilson 2002a). It is a scientific truism that powers are
metaphysically dependent on one or more fundamental interactions. The power
of being able to bond with an electron, in circumstances where one is in the
vicinity of a free electron, is grounded in the electromagnetic (or electroweak)
interaction, as opposed to the strong nuclear or gravitational interactions. The
power of being able to fall when dropped, in circumstances where one is poised
above Earth’s surface, is grounded in the gravitational force (or its geometric
correlate), as opposed to the other fundamental interactions in operation. The
power of being able to bond with other atomic nuclei in a stable configuration
is grounded in the strong nuclear interaction, as opposed to the electromagnetic,
weak, or gravitational interactions. The power of being able to sit on a chair
without falling through it is grounded (at least) in the gravitational and the
electromagnetic interactions. And so on. In providing a metaphysical basis for the
powers bestowed by properties, fundamental interactions systematically explain
vast ranges of natural phenomena. As Auyang (1999) puts it, in discussing the
currently accepted fundamental interactions:
There are four fundamental interactions. Gravity holds our feet on earth and
the earth in orbit; it is responsible for the large-scale properties of the universe
[…] Electromagnetism binds electrons and nuclei into atoms and atoms into
molecules; it is responsible for all physical and chemical properties of solids, liq-
uids, and gasses. The strong interaction binds quarks into nucleons and nucleons
into atomic nuclei. The weak interaction is responsible for the decay of certain
nuclei. (46)
Similarly, Greiner (1996) says:
All the known interactions that occur in nature can be reduced to four inter-
actions between material particles. Listed in order of decreasing strength, these
are: the strong (nuclear) interaction, electromagnetism, the weak (nuclear) inter-
action, and gravity. (1)
The metaphysics of fundamental interactions, treating the nature of such inter-

actions and how they serve as a basis for powers, is an underdeveloped area of
research, and a full exploration of these interesting issues would take us too far
afield. Here I will limit myself to saying just enough about these issues to motivate
the interaction-based strategy of response to the collapse objection.
To start, the notion of an interaction is a contemporary generalization of the
notion of a force: whereas forces are pushes or pulls (or component contribu-
tions thereof), now commonly seen as ultimately involving particle exchanges,
interactions may involve not just forces but other sorts of interactions, such as
particle creations and annihilations. As in the case of (what used to be called)
fundamental forces (e.g., gravity, electromagnetism), talk of a fundamental inter-
action is shorthand for talk of token interactions of a given type, that do or may
occur in certain circumstances, and which are at least partly constituted by the
presence of features (e.g., charge) lawfully associated with the interaction. As
above, certain interactions are deemed fundamental, in the sense of providing a
metaphysical basis for all other interactions and associated phenomena. Which
criteria are operative in deeming a given form of interaction fundamental is
again a large question; for present purposes what is most crucial is that there
is an operational test for the introduction of a novel fundamental interaction—
namely, that the posit of the interaction is needed to ‘balance the books’ as regards
various quantities (e.g., energy) which are taken to be conserved. Hence it was, as
discussed previously, that the weak nuclear interaction was introduced in response
to seeming violations of conservation laws associated with radioactive decay.
Though the operative test for positing a new fundamental interaction does not
hinge, it seems, on any particular metaphysical account of such interactions or
how these provide a basis for powers, it may nonetheless be worth registering
certain options on this score. To start, it is common to take a given fundamental
interaction to either be or be associated with a specific collection of fields. If fields

are understood as objects (or some other kind of entity), then it might be natural to
see them as having features and associated powers of their own, in which case one
part of the answer to the question ‘how are powers grounded in fundamental inter-
actions?’ would be that certain powers—ones plausibly deemed fundamental—are
grounded (schematically speaking) in fundamental interactions in virtue of being
associated with features of fundamental fields. The question would remain of how
exactly the powers of ordinary entities (objects, systems, and other particulars) are
grounded (again, schematically speaking) in fundamental interactions; and here
one answer might be that these nonfundamental powers are second-order powers
of fields: powers of fields to contribute to producing powers of ordinary entities or
their ultimate non-field substantial components (e.g., protons and electrons). A
somewhat more lightweight metaphysical variation on this theme would interpret
fields not as objects (entities), but as collections of comparatively fundamental
features and associated powers at spacetime points or regions; here again one
might take powers of ordinary entities or their constituents to be second-order
powers of powers of spacetime points or regions. And of course, as per usual,
there here remain the usual options for understanding talk of powers in more
or less heavyweight terms. Independent of further metaphysical details, however,
given that the claims that distinct fundamental interactions exist and serve as
a foundational basis for the spectrum of diverse powers of ordinary objects are
claims in unassailably good scientific standing, a Strong emergentist is within their
rights to speak of a feature’s having (or not having) a power, relative to a given set
of fundamental interactions.
Of course, physicalists of whatever stripe think that fundamental physical inter-
actions are the only fundamental interactions there are, while (as McLaughlin 1992
emphasizes) the Strong emergentist thinks that, in addition, there are one or more
nonphysical ‘configurational’ fundamental interactions. Strong emergentists can
thus grant that, taking both physical and nonphysical fundamental interactions
into account, a base feature P has every power an associated Strongly emergent
feature S has; but also coherently maintain that such an S will have powers that
are ‘new’ relative to those powers of P grounded only in fundamental physical
interactions. Such a conception clarifies the sense of (fundamental) novelty at
issue in the New Power Condition in Strong Emergence, making explicit that
this novelty—hence Strong emergence itself—is interaction-relative, along the
following schematic lines:
Interaction-relative Strong Emergence: Token feature S is Strongly emergent from

token feature P relative to the fundamental interactions in a set {F} just in case (i) S
cotemporally materially depends on P, and (ii) S has at least one token power that
is not identical with any token power of P that is grounded only in fundamental
interactions in {F}.
Condition (i) again minimally specifies cotemporal material dependence, under-

stood as involving both (physical) substance monism and the minimal nomo-
logical supervenience of higher-level on lower-level features, while condition (ii)
refines the New Power Condition in Strong Emergence, making explicit that the
sense of ‘new’ at issue adverts in part to a fundamental interaction that is new
relative to some specified set. Again, the use of ‘grounded in’ at issue here and else-
where is intended as schematic for some or other specific metaphysical relation, to
be further determined (see Wilson 2014a). For purposes of characterizing Strong
emergence in a way that appropriately contrasts with physicalism, one specifies
that the interactions in {F} are the fundamental physical interactions.
An interaction-relative implementation of Strong Emergence is clearly in the
spirit of the original British Emergentist suggestion that Strong emergence
involves what “we may call ‘configurational forces’: fundamental forces that
can be exerted only by certain types of configurations of particles” (McLaughlin
1992, 52). And it makes room for there to be Strong emergence in the face of
the collapse objection(s). To start: even if, taking all fundamental interactions
into account, features of the composing system in some sense inherit all the
powers of any features they cotemporally nomologically necessitate, it remains
that higher-level features may be associated with powers that are new, in not being
grounded only in the set of physical fundamental interactions. Properly relativized,
the novel powers of Strongly emergent features do not collapse. Relatedly,
relativizing powers to fundamental interactions provides a principled basis for
distinguishing dispositions expressing mere preconditions for the occurrence of
Strongly emergent features from those that are more directly implicated in the
having of the novel powers at issue.
One might be concerned that, as with conceptions of emergent features as
“surprising”, or with Taylor’s alternative (2015) conception in terms of what
is (perhaps only currently and contingently) scientifically unexplained, a rela-
tivized conception of physically unacceptable emergence will fail to track anything
metaphysically interesting or ‘joint-carving’. Interaction-relative Strong Emergence
doesn’t have this problem, however: new fundamental interactions are interesting
and joint-carving, if any natural phenomena are. Another concern might be that
the conception requires realism about fundamental forces or interactions. It is
unclear, however, just what is supposed to be problematic about such notions.⁷
As previously, fundamental interactions are plausibly understood as second-
order dispositions of fundamental fields (namely, dispositions to give rise to
further dispositions of non-field entities), and dispositions are not just familiar but
⁷ See Wilson 2007 for a defense of the reality and irreducibility of Newtonian forces; as least some
aspects of this defense would carry over to defense of the reality and irreducibility of fundamental
forces and/or interactions.
moreover admit of more or less metaphysically lightweight interpretations. In any

case, participants in the debates over physicalism or Strong emergence typically
are happy to take a (fallibilist) realist stance towards the posits of science, including
any fundamental interactions there might be; so even bracketing or leaving open
further metaphysical details, the Strong emergentist is within their rights to appeal
to such interactions in framing their account.
Strongly emergent objects

The third available response to the collapse problem is motivated by the thesis that
features have their powers derivatively on the powers of their bearers, as suggested
by Baysan 2016 (see also Baysan and Wilson 2017). Drawing on this idea, the
Strong emergentist might maintain that the novelty of powers at issue in Strong
emergence can be understood as always involving the coming-into-existence of a
new object (more generally, entity; I stick with ‘object’ for continuity with Baysan’s
discussion), suited to be the bearer of S and its novel power, and which object
is different from the object(s) bearing base feature P. Call this ‘the new object
strategy’. Indeed, though it is common to assume that what powers an entity has are
a matter of what features (and associated powers) it has, one might rather maintain
that the association of powers with features is derivative on the association of
powers with objects. As Baysan (2016) puts it:
What do we mean when we attribute powers to properties? […] Being knife-

shaped has the power to cut bread—conditionally on being instantiated with
certain other properties, of course. When we attribute this power to the property
of being knife-shaped, do we really mean that the property itself has this power?
Unless we want to identify properties with bundles of powers, I don’t think that
we have any good reason to give an affirmative answer to this question. Properties
don’t cut bread. Their bearers might. To generalize, properties don’t (literally or
fundamentally) have powers; their bearers do. (386)
Accordingly, the Strong emergentist can maintain that the novel powers asso-
ciated with a Strongly emergent feature are in the first instance powers of a
novel object. Such a view provides the basis for a principled response to the
collapse objection(s): in cases of the Strong emergence of a feature S, S’s novel
power presupposes the coming-into-existence of a new object, different from the
bearer(s) of P. Since, on this approach, powers are derivative on the objects having
the powers (and associated features), P would inherit S’s power only if P were
born by the same object as S; but again, the Strong emergentist can reasonably
maintain that this is incompatible with S’s having the novel power at issue, on
grounds that new powers require new objects. Effectively, the new object strategy
turns the collapse objection on its head: given that Strong emergence requires a
novel power, and given that powers of features are derivative on powers of their
bearers, Strong emergence requires a novel object to have the power, which is then
associated with emergent feature S. Indeed, the Strong emergentist can implement
the strategy even if they don’t agree with Baysan that the powers of features are
always derivative on the powers of objects. They can simply maintain that new
objects are required to be the bearers of any fundamentally novel powers or features
there might be.
One might wonder whether the implied commitment, on the new object
strategy, to the existence of distinct but spatiotemporally coincident objects is
problematic. Here I think that the Strong emergentist is within their rights to shrug
their shoulders. To start, the prima facie appearances of emergence encourage such
a pluralist commitment: special-science entities appear to be spatiotemporally
coincident with lower-level physical configurations, individual persons appear
to be spatiotemporally coincident with their bodies, and so on. To be sure,
there remains controversy over how best to treat seemingly coincident objects,
a debate which is often characterized as over the nature of material constitution
(see Wasserman 2010 for an overview); but antecedent to the identification of
some specific difficulty with Strongly emergent coincident objects, the Strong
emergentist who endorses the new object strategy can maintain that their view
constitutes one among the available pluralist options for accommodating material
constitution.
Another concern with the new object strategy is that it might be seen as avoiding
the collapse objection only by giving rise to an ‘explosion’ objection—namely, by
committing the Strong emergentist to a form of substance dualism, contra the
traditional supposition that emergence of whatever variety is supposed to conform
to substance (more specifically: physical) monism.
Here the Strong emergentist has two main lines of response. First, they may
grant that a Strongly emergent object counts as a new (type of) substance, and
moreover one that is in some sense nonphysical, but maintain that notwith-
standing the traditional characterization of emergence as a form of substance
monism, what is most important is that viable forms of such emergence suitably
contrast with views on which the additional substances or associated subjects
of Strongly emergent features are immaterial or otherwise very different from
physical substances. A commitment to Strongly emergent objects or associated
‘substances’ doesn’t entail anything of this sort. Indeed, the assumed cotemporal
material dependence of Strongly emergent entities and features on lower-level
physical entities and features is typically offered as a basis for contrasting this view
with serious (e.g., Cartesian) forms of substance dualism.
Second, the Strong emergentist can deny that from the mere positing of a new
object (entity) a new substance is thereby posited. To start, the claim that all and
only objects are substances is controversial, and may be rejected. For example,
Lowe (1998, 181) argues that some entities (e.g. surfaces, holes, heaps, events) are
objects, in being countable and in having determinate identity conditions, but are
not substances, since incapable of independent existence. More generally, on a
conception of substance as capable of independent existence, a Strongly emergent
object does not count as a substance, since such an object requires the existence of
whatever entity or entities are the proper bearers of the base feature. Nomologically
as well as conceptually, an emergent entity requires the existence of something else
from which it ‘emerges’. So there is reason to deny that the new object strategy of
response to the collapse objection(s) leads to substance dualism, much less to a
problematic form of such dualism.
Finally, it is worth noting that although the new object strategy supposes that
fundamentally novel features and powers bring new objects in their wake, it doesn’t
thereby follow that having Strongly emergent features or novel powers is required
for a new object to emerge. Indeed, Weak emergentists also commonly assume (as
per Bedau’s guiding assumption that an emergent entity is one having an emergent
feature) that Weak emergence involves the coming-into-existence of new entities,
associated with a distinctive subset of the powers of the base entities/features.
4.3 Objection: compatibility with physical acceptability
I turn now to the concern that satisfaction by a feature S of the conditions in

Strong Emergence is compatible with S’s being physically realized, hence physically
acceptable.
Yates (2016) offers an argument to this effect. By way of illustrative motivation
he argues that the molecular geometry of a water molecule—feature G—is a mul-
tiply realized feature which bestows powers not bestowed by its realizers. Key to
Yates’s proposal is the suggestion that some higher-level features are ‘qualitatively’,
as opposed to functionally, realized. Functionally realized features have causal
specifications, and correspondingly inherit their powers from their realizers; hence
they are at most Weakly emergent. By way of contrast, Yates suggests, qualitatively
realized features have noncausal—e.g., mathematical—specifications, which in
turn makes room for such features to be causally fundamental, and so Strongly
emergent:
Functionally realized properties […] are at most weakly emergent, but qual-
itatively realized properties don’t share this principled limitation. The reason
for this is that the bearer of a qualitatively realized property can have certain
causal powers: (i) in virtue of meeting the defining specification, but (ii) not
in virtue of basic physical properties and relations in virtue of which it meets
that specification. It follows that qualitative realization is consistent with strong
emergence […]. (820)
In the case of an H2 O molecule’s having molecular geometry G, the suggestion is

that the specification of G is mathematical and spatial rather than causal; that G
bestows certain powers upon its bearer—in particular, those, including hydrogen
bonding in water, associated with the molecule’s dipole moment, which is itself
explained in terms of G; and that (notwithstanding that G is physically realized by
lower-level physical goings-on) these powers are not had/bestowed by G’s realizers.
Granting that the specification of G is mathematical and spatial, and that the
possession of G is associated with various powers, why think that these powers are
not had by the base feature F that ‘qualitatively’ realizes G on a given occasion?
Yates’s line of thought seems to be that if such power inheritance were in place,
references to G could be eliminated in broadly deductive explanations of the dipole
moment and associated powers:
If we attempt to deduce the dipole moment of H2 O from basic physics without

appealing to molecular geometry, we draw a blank, for no sooner have we started
the deduction than we find ourselves deducing that H2 O has G as an essential
intermediary step. I conclude that molecular geometry plays a unique role in
determining the dipole moment of H2 O, and in virtue of this bestows a condi-
tional power φ that it doesn’t inherit from its basic physical realizers. (832)
I respond as follows. First, Yates’s methodology is suspect, since it is unclear

what metaphysical import should be assigned to considerations about what is
required for a given explanation or deduction. Relatedly, nothing in physicalism
or in the physicalist supposition that higher-level features inherit their powers
from physical base features requires that elements of higher-level explanations,
deductive or otherwise, be ‘dischargeable’ in terms referring only to lower-level
physical goings-on. On the contrary, physicalists of all stripes are typically happy
to allow that there’s no prospect of any such explanatory reductions; relatedly, their
reasons for maintaining that higher-level features are at best Weakly emergent
(as per their acceptance of Physical Causal Closure and the associated strategy for
addressing the problem of higher-level causation) do not hinge on the availability
of such reductions.
Second, it is especially unclear what import should be assigned to the need to
appeal to molecular structure G in explanations of the existence and powers of
the dipole moment of H2 O. If G couldn’t be deduced from lower-level physical
goings-on, but was nonetheless required in order to metaphysically account for
the explananda, that at least would provide some reason for thinking that G has
powers not had by its dependence base features, along standard (anti-physicalist)
Strong emergentist lines; but Yates maintains that G can be deduced from lower-
level physical goings-on, as an “intermediary step”. But then why think that the
need to appeal to G indicates that G has new powers, as opposed to thinking that
this need simply reflects that the explanation of the existence and powers of the
dipole moment has to proceed in steps, compatible with the physicalist assumption
that any powers of deducible features such as G are inherited?⁸
Third (and related to the previous point), Yates observes that his account faces
a difficulty:
At this point, however, a problem arises. I appeal to ‘in virtue of ’ relations twice
here: once in the realization relation between the basic physical properties of H2 O
and G, and again between G and conditional powers such as φ, which reflect G’s
putatively unique role in determining how H2 O molecules are disposed. But if ‘in
virtue of ’ is transitive, then H2 O has φ in virtue of its basic physical properties,
precluding G’s causal fundamentality. (834)
Yates develops this concern as tacitly appealing to a univocal notion of Grounding

of the sort I discussed in Ch. 3 (§3.4.4); he then responds that if we rather
distinguish two small-‘g’ grounding relations, then even if there is a kind of abstract
Grounding chain here, it is not one entailing that G’s powers are inherited:
The basic physical properties that realize H2 O’s molecular geometry are those in
virtue of which it meets a mathematical specification; and in virtue of meeting
that specification, H2 O has φ. These are two quite distinct grounding relations,
but this crucial fact is obscured by focusing on their shared abstract properties.
Following Wilson (2014a), I distinguish this abstract grounding relation from
specific grounding relations such as realization by labelling the former ‘Ground-
ing’ (with a capital ‘g’), and shall argue that Grounding is too coarse-grained to
capture the metaphysical structure of my emergentist proposal, which requires
two distinct little-g grounding relations. Basic physics mediately grounds φ
via one grounding relation—qualitative realization—between basic physics and
molecular geometry, and a distinct grounding relation—causal power bestowal—
between G and φ. That these relations are both instances of Grounding does not
entail that φ is really bestowed by the H2 O molecule’s basic physical properties.
. (835)
⁸ A related observation attaches to Hendry’s (2017) claim that the Strong emergence of molecular
structure is motivated by attention to the fact that distinguishing isomers (e.g., ethanol and dimethyl
ether) which share the same Schrödinger equation requires “putting in by hand the parameters that
specify an important difference between the two cases: the nuclear positions” (153–4). Representational
underdetermination aside, presumably the nuclear positions are completely determined by lower-level
physical goings-on; but in that case, why think that the need to appeal to such structural details indicates
that molecular structure has new powers?
My concern with Yates’s response here is that it does not block the natural thought
that the relation of qualitative realization is (like functional and other forms of
realization) also a relation of causal power bestowal—or, alternatively, that the
relation of causal power bestowal is also instantiated as holding between G and
its physical realizers. For example, suppose some entity has the property of being
circular, which specification is given by the usual mathematical formula. Granting
that satisfaction of the formula doesn’t in general entail anything about the having
of powers—perhaps there are circular abstracta—in any case whenever concretely
realized this geometric feature will be associated with powers. Round cookie
cutters have the power to cut round cookies, and so on. But if so, then the concern
remains that attention to the grounding chain in Yates’s case suggests that G does
not have any new powers, after all.
I conclude that the considerations that Yates raises do not provide a compelling
basis for thinking that satisfaction of the conditions in Strong Emergence is com-
patible with physical acceptability.
4.4 Objection: non-necessity
I turn next to considering and responding to objections to the claim that satisfac-
tion of the conditions in Strong Emergence is necessary for physically unacceptable
emergence. There are four main alternative approaches on offer, in terms of
epiphenomenalism, supervenience, primitivism, and epistemic criteria, respec-
tively. In what follows, I provide reasons for thinking that each of these alternative
approaches to physically unacceptable emergence is unsatisfactory.
4.4.1 Epiphenomenalism
Some attempts to characterize a form of physically unacceptable emergence have

aimed to do so in ways that suggest that such emergent features might be epiphe-
nomenal, and so fail not only to have a new power (as per the New Power
Condition), but to have any powers at all. Hence Morris (2014) says that “we
could attempt to characterize varieties of noncausal nonphysicalist novelty that
are compatible with supervenience” (353), and Chalmers (1996) endorses a form
of what he calls ‘naturalistic dualism’ which is naturally seen as involving “a
limited form of epiphenomenalism” (158–9). Indeed, Chalmers goes on to argue
that the sort of ‘interactionist dualism’ endorsed by the British Emergentists, and
which the schema for Strong emergence aims to characterize, doesn’t represent any
real advantage over epiphenomenalist naturalistic dualism, as regards conscious
experience:
[O]n a close analysis, [interactionist dualism] leaves consciousness as superfluous

as before. To see this, note that nothing in the story about emergent causation
requires us to invoke phenomenal properties anywhere. The entire causal story
can be told in terms of links between configurations of physical properties.
There will still be a possible world that is physically identical but that lacks
consciousness entirely. It follows that at best phenomenal properties correlate
with causally efficacious configurations. (1996, note 41, 378–9)
The best response to this line of thought, and the associated suggestion that
physically unacceptable emergent features might be epiphenomenal, rehearses
the response I previously gave in Ch. 3 (§3.3.2) to the concern that satisfaction
of the conditions in Weak Emergence is compatible with there being noncausal
phenomenal aspects of higher-level features. There I argued that it is reasonable
to maintain that any phenomenal or qualitative aspects of features would be
fully encoded in powers of those features, as per the thesis of Phenomenal
Incorporation. There I also argued that the Phenomenal Incorporation thesis is
neutral between a physicalist (Weak emergentist) and a nonphysicalist (Strong
emergentist) conception of qualitative features. As such, neither the Weak nor
Strong emergentist need agree that “nothing in the story about emergent causation
requires us to invoke phenomenal properties anywhere” such that “The entire
causal story can be told in terms of links between configurations of physical
properties”. On the contrary, either variety of emergentist can reasonably maintain
that phenomenal properties are crucially referenced as part of the causal story,
and similarly for consciousness more generally (a point to which we will return
in Ch. 7).
4.4.2 Supervenience
A common baseline assumption of accounts of emergence, whether of Strong or

Weak varieties, is that emergent features minimally nomologically supervene on
base features, such that, at least in worlds with the same laws of nature as actually
hold, the occurrence of an emergent feature requires the occurrence of some base
feature, and (again, in such worlds) any such base feature necessitates the emergent
feature. A number of more specific conceptions of supervenience are on offer,
aimed at precisifying glosses such as that ‘there can be no change in supervenient
features without a change in base features’ or (equivalently) that ‘duplicating the
base features duplicates the supervenient features’ (see McLaughlin and Bennett
2018 for discussion). For present purposes what is most relevant is, first, that
supervenience is an abstract modal correlational notion or relation, typically
understood as holding between lower-level and higher-level features, and second,
that some have thought that the difference between physically acceptable and
physically unacceptable higher-level features can be cashed simply in terms of a
difference in the strength of the correlations holding between higher-level and base
features. More specifically, the suggestion is that physically acceptable (Weakly)
emergent features supervene with metaphysical necessity on lower-level physical
features, such that the correlations are guaranteed to hold in any worlds where the
physical features (and, for reasons discussed in Ch. 3, note 15, associated physical
laws) are in place, whereas physically unacceptable (Strongly) emergent features
supervene with only nomological necessity on lower-level physical features, such
that the correlations are guaranteed to hold only in worlds where, in addition to
the physical laws, any emergent laws there might be are also in place. For example,
van Cleve (1990) characterizes Strong emergence as follows:
If P is a property of w, then P is emergent iff P supervenes with nomological neces-

sity, but not with logical necessity, on the properties of the parts of w. (222)
(Here by ‘logical necessity’, Van Cleve has in mind metaphysical necessity, as

characterized above.) Chalmers (2006a) endorses a similar conception:
[W]e can say that Strong emergence requires that high-level truths are not
conceptually or metaphysically necessitated by low-level truths. (244, note 1)
The proposed distinction in modal strength here is sometimes taken to serve as a

basis for formulating physicalism (understood as per usual as incompatible with
Strong emergence) as the thesis that all broadly scientific goings-on supervene with
metaphysical necessity on lower-level physical goings-on. Hence Lewtas (2013)
says
The supervenience physicalist [ . . . ] claims that a suitable physical base metaphys-

ically necessitates [higher-level] properties. But a [Strong] emergent supervenes
on its base and yet stands over and above it. Emergent supervenience therefore
amounts to lawful correlation and nothing more. (532)
In terms of a global supervenience thesis, the suggestion is that physicalism is true

just in case any world duplicating the physical goings-on (including physical laws)
duplicates the rest of natural reality:
Physicalism is true of our world iff any world that is a physical duplicate of our
world either is a duplicate of our world simpliciter or contains a duplicate of our
world as a proper part.⁹ (Howell 2009, 85)
⁹ The latter condition is intended to accommodate the intuition that worlds that are just like our
world except for the addition of ghosts or immaterial souls and the like would not falsify physicalism
as holding in this world.
Such supervenience-based characterizations form the basis for an objection to the

necessity of Strong Emergence, according to which a distinction between strength
of modal correlations, rather than a distinction between powers, is what is key to
making sense of physically unacceptable emergence.
The supposition that a difference in strength of modal correlation suffices to
distinguish physically unacceptable from physically acceptable dependent features
is problematic, however. To start, as has frequently been observed, Strongly
emergent features might supervene with metaphysical necessity on physical base
features (see Horgan 1993, Kim 1998, Levine 2001, Melnyk 2003, Tye 1995, and
Wilson 2005). In my (2005), I offer several scenarios illustrating how this might
be. One somewhat fanciful but still metaphysically coherent scenario is a version of
Malebranchean occasionalism, in which a consistent Malebranchean God brings
about certain mental features upon the occasion of certain lower-level physical
features, in every world where the latter exist. Another less fanciful scenario
involves a view on which features are essentially individuated by (all) the laws
of nature into which they directly or indirectly enter (along lines proposed in,
e.g, Shoemaker 1980, Swoyer 1982, and Bird 2001, 2002, and 2007). On such
a view, any Strongly emergent features there might be would be metaphysically
necessitated by base features, since the latter could be instantiated only in worlds
containing the emergent as well as the physical laws. Yet another scenario with this
consequence is one in which Strongly emergent features involve a nonphysical fun-
damental interaction, and where it is supposed that the fundamental interactions
are unified, in coming as a package deal.
Such scenarios indicate that neither the distinction between metaphysical and
nomological necessity, nor the distinction between supervening and not superven-
ing on physical goings-on in worlds where the actual physical laws are operative,
can serve to distinguish cotemporally materially dependent features which are
physically unacceptable from those that are physically acceptable.1⁰
There are two main lines of resistance to this claim; I address each in turn.
The first proceeds by rejecting the seeming counterexamples on grounds that
these violate what is sometimes called ‘Hume’s Dictum’, according to which there
are no metaphysically necessary connections between wholly distinct existences.
As Stoljar (2001) describes the strategy:
[One] suggestion points out that the problem is only genuine if the cases that
generate it are coherent—and are they? One reason against supposing so is that
both seem to violate Hume’s dictum that there are no necessary connections
between distinct existences. According to [Strong] emergentism, for example,
mental and physical properties are metaphysically distinct, and yet are necessarily
connected.
1⁰ Note that the concerns here are properly metaphysical. See Kovacs 2019 for critical assessment
of broadly epistemological concerns about supervenience according to which it is not an ‘explanatory’
relation.
Adherence to something like Hume’s Dictum is present, for example, in Noord-

hof ’s (2010) endorsement of a supervenience-based characterization of the dis-
tinction between physically acceptable and physically unacceptable supervenient
features:
The intuitive thought is that the reason why emergent dualists [should] reject
appeal to metaphysical necessity is that they suppose that some of the target
properties determined by narrowly physical property causes are wholly distinct
from them, whereas non-reductive physicalists are committed to thinking that
they are not. (71)
A similar thought is operative in Lewtas’s (2013) response to certain of the seeming

counterexamples, according to which in cases of Strong Emergence, “the correla-
tion connects two entities where [ . . . ] each is over and above the other” (538).
But appeal to Hume’s Dictum is ultimately unsatisfactory by way of response.
To start, as Stoljar notes, Hume’s Dictum is controversial:
Hume’s dictum is itself a matter of controversy, so it is unclear if the cases can be

dismissed in this way (see Jackson 1994, Stalnaker 1996, Stoljar 2010, and Wilson
2005, 2010c).
Indeed, post-Humean reasons for believing Hume’s Dictum are in short supply. To
be sure, if one is a strict empiricist like Hume, who takes the content of our ideas
and beliefs to ultimately be a matter of fairly superficial sense experiences, one
might well be inclined to endorse Hume’s Dictum: any such ideas might, at least
in principle, either go together or come apart. But contemporary philosophers are
not strict empiricists, and as MacBride (1999) notes,
[I]t is a curious fact that the proponents of the contemporary Humean pro-
gramme […] having abandoned the empiricist theory of thought that under-
writes Hume’s rejection of necessary connections provide precious little by way
of motivation for the view. (127)
Moreover, in a series of papers (see Wilson 2010a, 2010c, 2014b, and 2015a) I have
considered several potential reasons for accepting Hume’s Dictum, and argued that
none withstands scrutiny.
Even if Hume’s Dictum is accepted, endorsement of this thesis won’t sidestep
many of the seeming counterexamples, for Strongly emergent features need not be
‘wholly distinct’ from lower-level physical goings-on (see Wilson 2002a and Stoljar
2007). For example, Strongly emergent features might share some (though not all)
powers with lower-level physical base features, or be governed by physical as well
as nonphysical interactions. As such, there is no principled route to maintaining

a supervenience-based approach to realization that proceeds by way of Hume’s
Dictum.
A more principled response to defending a supervenience-based approach is
suggested by Howell (2009), who argues that the seeming counterexamples to this
approach can be rejected, on grounds that if a Strongly emergent feature S were to
be metaphysically necessitated by a lower-level base feature P, then we would have
good reason to take P to itself be physically unacceptable:
The basic argument is that if emergence laws are [metaphysically] necessary,11

and the emergent properties are “new” enough to count as non-physical, then
the supervenience base will be polluted and will no longer be purely physical. If
this is the case, then [supervenience physicalism] will judge an emergence dualist
world to be non-physical, because duplicating the purely physical properties will
not duplicate the world simpliciter. (93)
The general suggestion here is that if Strongly emergent feature S arises from lower-
level feature P with metaphysical necessity, then one should take P to be individ-
uated, in part, by the disposition to give rise to S; and if P is so individuated—if
part of what it is to be an instance of P is to be disposed to give rise to instances
of S—then one should not regard P as a physical property. Howell illustrates his
intended point as follows:
If it turns out that part of what makes electrons what they are is that they give rise
to ‘unpredictable’ qualitative experiences when in a certain setting, then it seems
that electrons are somewhat magical and are at least partly constituted by non-
physical dispositions. […] In such a world, a sort of quasi-panpsychism is true:
at least some of the basic stuff in our world is not conscious, but it is infused with
mentality in that it is individuated by the brute tendency to produce it. (93–4)
In that case, Howell suggests, the possibility of metaphysically necessitated

Strongly emergent features poses no threat to characterizing the distinction
between Weak and Strong emergence in terms of stronger or weaker modal
correlations, since any such features would ‘pollute’ the supervenience base
features in such a way that the latter would no longer be properly considered
(lower-level) physical, contra the standard assumption of physicalists and Strong
emergentists alike (93).
11 Note that the antecedent here isn’t one expressing that the ‘emergence laws’ hold in every possible
world, but rather that the emergents are connected to their bases in every possible world where those
bases exist.
The Strong emergentist has two available responses. The first is due to Morris
(2014), according to which “reflection on the base pollution maneuver reveals
a new worry about supervenience physicalism” (356). Morris starts by noting
that the proponent of a supervenience-based approach to realization (Weak
emergence) will, like the Strong emergentist, distinguish the lower-level physical
goings-on in the supervenience base from the higher-level features and entities
that supervene on this base. In that case, one might naturally wonder whether
such higher-level features would also end up ‘polluting’ the base:
The base pollution defense supposes […] that the physicality of a property may be
called into question by that property necessarily giving rise to another property
that is not itself physical. But in this case, it may seem that supervenient properties
will generally end up “polluting the base”. Why, in other words, does the problem
of “base pollution” specifically concern [Strong] emergentism? Why doesn’t it
threaten the very idea of properties distinct from [lower-level] physical properties
supervening on [lower-level] physical properties? (356)
Howell (2009) is aware of this concern, and addresses it by saying that if the
supervenient properties are not “substantively new”, then there is no difficulty
with maintaining that the base properties are (lower-level) physical (93, note
18). But as Morris (2014) notes, the need to provide an account of when some
supervenient features are or are not substantially new introduces a new challenge
for supervenience physicalism:
Granting that the [Strong] emergentist’s base cannot be regarded as physical,

some account is needed of the conditions under which a supervenient property is
not “substantially new” with respect to subvenient, putatively physical properties.
Further, without such an account, a supervenience definition of physicalism
will provide no guidance for distinguishing between a physical supervenience
base and the kind of polluted base associated with [Strong] emergentism, and at
least in this way would appear incomplete. The challenge for a supervenience
physicalist is to account for the difference between the polluted base and a
physical supervenience base without, in effect, rendering talk of supervenience
superfluous [if] the only or best way to mark the requisite distinction appeals to
the very resources at work in alternative formulations of physicalism or, likewise,
alternative accounts of what it is to be a physicalist about some [higher-level]
feature of reality. (357)
Morris then goes on to consider two candidate accounts of when a supervenient

property is not ‘new enough’ to result in ‘base pollution’—one appealing to a
second-order functionalist account, and one appealing explicitly to satisfaction
of the Proper Subset of Powers Condition—and to observe that while each

account plausibly wards off base pollution, each at the same time delivers a
non-supervenience-based account of realization—that is, of physically acceptable
emergence—rendering the appeal to supervenience in characterizing such
realization otiose. (See also Morris 2018, 45–7.) A similar thought seems
appropriately directed at Lewtas’s (2013) suggestion that, even if Strong as
well as Weak emergents metaphysically supervene on their bases, the latter
are distinguished in that “the correlation [is] underlain [ . . . ] by a more basic
constitutive relation” (539).
I agree with Morris’s assessment, and moreover add that (as previously argued)
insofar as functionally realized features also satisfy the Proper Subset of Powers
Condition, the more general moral of his assessment is that supervenience-based
accounts of realization ultimately do not provide an approach to such realization
alternative to one implementing the conditions in Weak Emergence.
A second response to Howell’s argument is available—namely, to deny his claim
that base features will be rendered physically unacceptable if they are disposed to
bring about metaphysically necessitated Strongly emergent features or powers. To
start, one need not interpret Howell’s illustrative case of electrons giving rise to
‘unpredictable’ qualitative experiences as showing that if base features or entities
metaphysically necessitate higher-level features, then ‘part of what makes the [base
features] what they are’ is that they give rise to these higher-level features; nor
need one see his case as showing that if base features have dispositions to give rise
to physically unacceptable features, the base features will thereby be physically
unacceptable. Consider the scenario I previously offered, in which a consistent
Malebranchean God brings about mental features on the occasion of certain
physical features in every world where the latter exist (that is, with metaphysical
necessity). In such a scenario it need not be any part of ‘what it is to be’ the
occasioning physical features that God takes them to be such occasions. Relatedly,
one might reasonably maintain that if lower-level physical features do have ‘dispo-
sitions’ to bring about Strongly emergent features, the sense of ‘disposition’ here
is a weak one, reflecting just that the lower-level features are, for whatever reason,
metaphysically sufficient preconditions for the Strongly emergent features, which
sense implies nothing about the natures of the occasioning features that would
impugn their physical acceptability.
That said, on certain proposed counterexamples to a supervenience-based
approach to metaphysical emergence, the sense in which a lower-level feature
might be disposed to produce a Strongly emergent feature would plausibly inform
some part of its nature—for example, if (as previously mentioned; see Wilson
2005 for discussion) features are individuated by all of the laws into which they
enter. It remains, even here, that the sense of ‘disposition’ at issue is weak, again
reflecting that the lower-level features are metaphysically sufficient preconditions
for the Strongly emergent features. Still, for such cases it seems that more is needed
to block the threat of ‘pollution’. And what more is needed seems to be some
principled way of characterizing the lower-level features as physical, in spite of
their having ‘nature-involving’ dispositions to produce Strongly emergent features.
One strategy for doing this appeals to the operative characterization of physical
entities and features, according to which (as discussed in Ch. 1, §1.4.1) these are
the entities and features treated (approximately accurately) by present (and in
the limit of inquiry, ideal) physics, and which are not fundamentally mental (see
Wilson 2006). Here it is crucial to appreciate two qualifications of the intended
characterization of the physical goings-on as satisfying the ‘No Fundamental
Mentality’ (NFM) constraint (or related constraints as discussed in Ch. 1, §1.4.1):
first, that the restriction on fundamental mentality is intended to apply to entities
and features in comparatively noncomplex combination (what I earlier called
‘basic’ physical entities and features), reflecting the crucial contrast with panpsy-
chism, according to which individual fundamenta might have or bestow mentality;
and second, that the restriction on fundamental mentality pertains not to any
unmanifested dispositions there may be, but rather only to manifested features of
entities that, again, are comparatively noncomplex. So long as individual lower-
level entities and features do not have or bestow mentality, then that they or
associated configurations might have dispositions to give rise to Strongly emergent
mentality poses no threat to their status as physical or physically acceptable.
Interestingly, while Howell’s criterion of physicality imposes a restriction on
mentality, his case for taking the restriction to be violated depends on ignoring
these qualifications. Howell’s argument relies on the following necessary condition
on something’s being physical:
ND: Something is physical only if it does not ineliminably involve mental

features.
It is the supposed violation of ND by lower-level dispositional features that is

supposed to establish that metaphysically necessitated Strongly emergent phe-
nomenal features would undermine the physical acceptability of the dependence
base features. But—and notwithstanding that Howell claims that my (2006) char-
acterization of the physical involves ND as a necessary condition—ND is too
broad, in failing to specify that the necessary condition here applies only to
comparatively noncomplex physical entities and features and moreover only to
features manifested under such comparatively noncomplex circumstances. Since
the dispositional features under discussion are not so manifest, they pose no threat
to the physical acceptability of lower-level features.
A second strategy for characterizing the lower-level features as physical, even
given that they have ‘nature-involving’ dispositions to produce Strongly emer-
gent features, adverts to fundamental interactions. Again, even if, taking all
fundamental interactions into account, lower-level physically acceptable features

are essentially disposed to bring about certain Strongly emergent features, it
remains that the occurrence of these lower-level features is just a matter of physical
fundamental interactions, reflecting the Strong emergentist understanding of
lower-level physical features as something like cotemporally sufficient (nomo-
logically or metaphysically, no matter) preconditions for any Strongly emergent
features there might be. In short: the Strong emergentist can maintain that the
occurrence or instantiation of lower-level physical features ultimately relies just
on the operation of fundamental physical interactions, whereas the occurrence
or instantiation of a Strongly emergent feature requires the operation of an
additional, non-physical fundamental interaction. As in the case of a fundamental
interaction-based response to the collapse objection(s), fundamental interactions
provide a basis for distinguishing lower-level physical from Strongly emergent
goings-on, even when these are deeply dispositionally connected.
4.4.3 Primitivism
I turn next to Barnes’s (2012) ‘meta-ontological’ account of emergence, according

to which emergent goings-on are those which are both fundamental and depen-
dent, and where the notions of fundamentality and dependence are each taken
to be primitive. Barnes does not distinguish between Strong and Weak forms of
metaphysical emergence in her discussion, but in appealing to fundamentality her
account plausibly aims to characterize emergence of the Strong, ‘over and above’
variety.
The general problem with this account is that it is too abstract to satisfy the
criteria of appropriate and illuminating accommodation.12 As we have seen in
both historical and contemporary contexts, an understanding of Strong emergence
as combining dependence with fundamentality is the traditional starting point
of discussions of such emergence, with the bulk of effort devoted to filling in
the operative notions of dependence and fundamentality so as to illuminate how
these characteristics could be jointly instantiated in ways appropriately accom-
modating the target cases, while blocking concerns that Strong emergence is
naturalistically unacceptable, gives rise to problematic overdetermination, is inco-
herent (since subject to collapse), and so on. An account of Strong emergence
on which the operative notions of dependence and fundamentality are primitive,
and where there is no further (or adequate) explication of these notions, goes
12 See Wilson 2019 for further discussion of concerns with ‘abstractionist’ accounts of metaphysical
emergence, which also include the Grounding-based approaches discussed in Ch. 3 (§3.4.4).
no distance towards illuminating such emergence or addressing the associated

concerns.13
In particular, though I am friendly to a primitivist account of fundamentality
(see Ch. 1, §1.4.3, and Wilson 2014 and in progress)—where better to put prim-
itivity than in what makes it the case that some goings-on are fundamental?—in
the absence of further explication about the locus of fundamentality in cases of
Strong emergence (e.g., as involving powers that are associated with fundamental
interactions), which Barnes does not provide, her account does not ensure that
Strong emergents are distinctively efficacious (if, say, the dependent fundamental
entity is a noncausal quiddity or other epiphenomenon). A concomitant difficulty
is that Barnes’s account fails to provide any means of engaging with or addressing
either Kim’s problem of higher-level causation or the collapse objection.
Another difficulty is that a primitivist account of dependence does not ensure
that goings-on that are emergent by lights of Barnes’s account are incompatible
with goings-on of the sort posited by non-emergentist accounts, such as substance
dualism. Barnes does offer an “intuitive gloss” on the notion of dependence, as
follows:
Ontological Dependence (OD): An entity x is dependent iff for all possible worlds
w and times t at which a duplicate of x exists, that duplicate is accompanied by
other concrete, contingent entities in w at t. (880)
But this conception of dependence is too weak to ensure material dependence—

again, consider a Malebranchean occasionalist scenario involving a God who
consistently (at every possible world) brings about mental states when and only
when certain (contingent) physical states are on the scene.1⁴
OD also renders Barnes’s account subject to a range of counterexamples. Sup-
pose that (as either a Strong or a Weak emergentist might believe) a fundamental
physical state P necessitates some mental state M, distinct from P, in any world
where P exists. Then P turns out to be dependent by lights of OD, since for all
possible worlds w and times t at which a duplicate of P exists, that duplicate
will be accompanied by a distinct contingent entity (M, or a duplicate thereof).
Moreover, since P is fundamental, it follows on Barnes’s account that P is Strongly
emergent. That’s the wrong result. For similar reasons, Barnes’s account interpreted
as involving OD will deem contingent fundamental determinates, which are
13 Pearson (2018) and Paolini Paoletti (2018b, 14–15) register similar difficulties with Barnes’s
account.
1⁴ As Pearson (2018, 396) points out, the conception is also too strong, in the sense that it rules
out that Strong emergence might hold with only nomological necessity, as many proponents of such
emergence have maintained.
accompanied by their associated determinables in any world where they exist,

Strongly emergent. Yet another counterexample is due to Pearson (2018):
[S]uppose that it were necessary that every time we have charge we have spin and
vice versa, and that these were concrete, contingent, Fundamental properties.
That wouldn’t mean that charge would be an emergent property. Rather, neither
charge nor spin appear to be good candidates for emergent properties. (397)
As a supplement to OD, Barnes offers a characterization of ‘dependence on’:
x is dependent on the ys iff x is dependent because it is part of its intrinsic nature

that it bears relation R to things intrinsically like the ys. (881, note 16)
This is no help, however, since relation R is again left uncharacterized, and since the
fundamental xs at issue might well have (and in the case of determinates, definitely
do have) as part of their intrinsic nature that they are intimately related to things
intrinsically like the ys.
The moral of these multiple difficulties is that there is no shortcut to an adequate
characterization of physically unacceptable emergence via primitivist specification
of the core notions of fundamentality and dependence. These notions must be
filled in, and moreover in a way not leading to immediate counterexample.
4.4.4 Epistemic criteria
A final objection to the necessity claim adverts to satisfaction of one or other

epistemic criterion as sufficient for Strong emergence. Indeed, accounts of Strong
emergence as involving one or other epistemic failure have been historically
common. For example, Broad (1925) characterized emergence as involving in-
principle failure of deducibility:
Put in abstract terms the emergent theory asserts that there are certain wholes,
composed (say) of constituents A, B, and C in a relation R to each other; that all
wholes composed of constituents of the same kind as A, B, and C in relations of
the same kind as R have certain characteristic properties; that A, B, and C are
capable of occurring in other kinds of complex where the relation is not of the
same kind as R; and that the characteristic properties of the whole R(A, B, C)
cannot, even in theory, be deduced from the most complete knowledge of the
properties of A, B, and C in isolation or in other wholes which are not of the
form R(A, B, C). (61)
And recall Chalmers’s (2006a) suggestion that
[W]e can say that Strong emergence requires that high-level truths are not
conceptually or metaphysically necessitated by low-level truths. (note 1)
Even if metaphysical necessitation or supervenience, understood as a purely cor-

relational notion, is too weak to characterize physically unacceptable emergence,
Chalmers here suggests that a failure of conceptual entailment will do the trick.
Finally, consider Horgan’s (1993) epistemic account of ‘superdupervenience’—that
is, supervenience of the sort guaranteeing that supervening properties are nothing
over and above their physically acceptable base properties:
Horgan’s Constraint: Any genuinely physicalist metaphysics should countenance

ontological inter-level supervenience connections only if they are robustly
explainable in a physicalistically acceptable way. (563)
Conversely, one might suggest, it suffices for physically unacceptable emergence

that a dependent higher-level feature fails to be “robustly explainable in a physi-
calistically acceptable way”.
Though not uncommon, there are two good reasons to reject ‘epistemic failure’
characterizations of physically unacceptable emergence.
First, these epistemic characterizations are intended by their proponents to
track a metaphysical distinction relevant to properly metaphysical emergence. For
example, as previously discussed, Broad’s epistemic characterization was intended
to characterize metaphysical emergence as involving fundamentally novel (‘trans-
physical’) laws and associated fundamentally novel powers. Similarly, the case
studies that Chalmers and Horgan offer as illustrating the sort of conceptual
entailment or robust explanation that is supposed to be lacking in cases of Strong
emergence involve something like functional or causal realization, which are
reasonably seen as satisfying the (metaphysical) conditions (in particular, the
Proper Subset of Powers Condition) in Weak Emergence.1⁵ Conversely, it would
1⁵ Hence Chalmers (2006a) says of heat:

The concept of heat that we had a priori—before the phenomenon was explained—was
roughly that of ‘the thing that plays this causal role in the actual world.’ Once we discover
[a posteriori] how that causal role is played, we have an explanation of the phenomenon.
(45)
And Horgan (1993) says of water:
Explaining why liquidity supervenes on certain microphysical properties is essentially a
matter of explaining why any quantity of stuff with these microphysical properties will
exhibit these macro-features [tendency to flow, to assume shape of vessel that contains it,
etc.] […] this suffices to explain the supervenience of liquidity because those macro-features
seem that, on these accounts, the intended sense in which a dependent feature
might fail to be conceptually entailed or robustly explainable would, as on Broad’s
account, reflect the feature’s having a new power, as per the schema for Strong
emergence. As such, the Strong emergentist might maintain that even if Strongly
emergent features were always and distinctively (as compared to Weakly emergent
features, in particular) accompanied by certain epistemic failures, it would still be
advisable to characterize physically unacceptable emergence in terms of what is
metaphysically at issue—that is, in the terms encoded in Strong Emergence.
Second, in any, case epistemic failures are not distinctive of physically unaccept-
able emergence. As I will discuss in more detail in Ch. 5: while it made sense in
Broad’s day to suppose that in-principle failures of deducibility or predictability
from lower-level physical goings-on (including laws) were indicative of funda-
mentally novel laws and associated powers, it has since become clear that many
uncontroversially physically acceptable dependent goings-on are not deducible,
even ‘in principle’, from lower-level physical goings-on, for reasons having to do
not with fundamental novelty but rather with, e.g., sensitivity to initial conditions
(à la the ‘butterfly effect’) or mathematical limitations of the sort discussed in
Boyd 1980, rendering predictions about such goings-on impossible, even given
the resources of the entire universe. As such, the proponent of an epistemic
characterization of physically unacceptable emergence will need to provide some
means of distinguishing unexplained physically unacceptable from unexplained
physically acceptable higher-level features.
Here the Strong emergentist can register a variation on the complaint, made
by Melnyk (1999) and Morris (2014) against supervenience-based accounts of
realization/Weak emergence, that insofar as ensuring the physical acceptabil-
ity of supervenient entities or features ultimately requires appeal to some or
other specific metaphysical relation(s), the appeal to supervenience is rendered
‘superfluous’ (Morris), ‘an idle wheel’ (Melnyk).1⁶ In particular, and assuming
satisfaction of the condition on cotemporal material dependence, it is unclear what
distinction might be appealed to by way of differentiating physically unacceptable
from physically acceptable unexplained phenomena besides one encoding the
powers-based conditions in the schemas for Strong and Weak emergence. In that
case, however, one can and should dispense with the appeal to lack of explanation
(deducibility, conceptual entailment) and rather stick with the powers-based
conditions in the schemas, which encode what is really at issue in the contrast
between Weak and Strong emergence.
are definitive of liquidity [and because] it seems explanatorily kosher to assume a “connect-
ing principle” linking the macro-features to liquidity, precisely because those features are
definitive; the connecting principle expresses a fact about what liquidity is. (579)
1⁶ See also Hill 2009, 66 and Morris 2018, 48–9.

In this chapter, I have considered a wide range of objections to the viability of

Strong emergence, according to which satisfaction of the conditions in Strong
Emergence is incompatible with scientific theory or practice, is impossible (as per
variations on the ‘collapse’ objection), is compatible with physical acceptability,
or is not necessary for physically unacceptable emergence; and I have argued that
each objection admits of one or more responses available to proponents of any
of the diverse implementations of the schema. In two cases, additional responses
are available which rely on features specific to a fundamental interaction-relative
account of such emergence. In particular, such an account provides the basis for
distinctive responses to the collapse objection and the associated ‘base pollution’
objection. My sense is that if any of the considered objections requires tweaking
the conditions in Strong Emergence (so as to explicitly incorporate reference to fun-
damental interactions), it is these. If so, the needed relativization to fundamental
interactions would be straightforward, and hardly fatal to the overall approach;
but again, this is a choice point, and how exactly a Strong emergentist chooses to
respond to the various objections may depend on further of their commitments. In
any case, these results collectively indicate that Strong Emergence is not just a viable
and indeed attractively robust means of accommodating physically unacceptable
emergence, but that this schema moreover captures what is core and crucial to
(and when suitably filled in, what is necessary and sufficient for) such emergence.
5
Complex systems
At this point, I have answered the first key question driving this book: What,
more precisely, is metaphysical emergence of the sort seemingly motivated by
the special sciences as well as ordinary experience, coupling cotemporal material
dependence with ontological and causal autonomy (distinctness and distinctive
efficacy)? Underlying the seeming diversity associated with the many accounts of
such emergence are two and only two schemas, expressing conditions which are
core and crucial to (and when sensibly filled-in, necessary and sufficient for) meta-
physical emergence of physically acceptable and physically unacceptable varieties:

emergent from token feature P on a given occasion is for it to be the case, on that
occasion, (i) that S cotemporally materially depends on P, and (ii) that S has at
least one token power not identical with any token power of P.
I have moreover argued that each form of emergence is viable: coherent, metaphys-
ically substantive, naturalistically acceptable, such as to avoid problematic causal
overdetermination, and more generally such as to accommodate, in principle, the
appearances of seeming metaphysical emergence in an appropriate and illuminat-
ing way.
Since there are no in-principle barriers to there being metaphysical emergence
of either variety, we are now in position to address (or finish addressing) the sec-
ond key question driving this book: Is there actually any metaphysical emergence?
In this and the following chapters, I will consider this question as applied to four
classes of phenomena that have been frequently taken to involve metaphysical
emergence: complex systems, ordinary objects, consciousness, and free will.
I begin with complex systems, as perhaps the phenomena that have been most
often offered as emergent, by scientists as well as philosophers. Such systems take
many forms, both natural (as in cases of turbulent water flows, phase transitions,
and weather patterns) and artificial (as in Conway’s ‘Game of Life’, to be discussed
in §5.2.1). While there is no agreed-upon definition of what it is to be a complex
DOI: 10.1093/oso/9780198823742.003.0005
system (see Ladyman et al. 2013), the commonly highlighted characteristics of

such systems are as follows:
• Nonlinearity: Complex systems are nonlinear, in that certain of their features

(including associated powers and behaviors) cannot be seen as linear or other
broadly additive combinations of features of the system’s composing entities,
and relatedly, in that mathematical expressions describing the evolution of
such systems contain nonlinear terms.
• Extreme sensitivity to initial conditions: For ‘chaotic’ complex systems, small
differences in initial conditions can result in huge differences in trajectory or
behavior.
• Unpredictability: The precise behaviors of complex systems are unpredictable
and relatedly, often surprising and seemingly novel.
• Algorithmic incompressibility: The dynamical equations of complex systems
do not admit of analytic or ‘closed’ solutions.
• Universal behavior (‘Universality’): Compositionally different complex sys-
tems may exhibit highly similar behavior.
• Self-organization: Complex systems exhibit coherent patterns arising as a
result of interactions among the parts, in a way suggesting that they are ‘self ’-
organizing.
Such characteristics have been offered as supporting taking certain complex

systems to be either Strongly or Weakly emergent; here I present and assess these
cases.
In the first section, I consider whether any complex systems might be Strongly
emergent. I start with a compressed historical discussion of why the British Emer-
gentists took nonlinearity and in-principle failures of predictability to suffice for
fundamental novelty—a view that, while reasonable at the time, was undermined
by the discovery and creation of complex systems clearly not involving any funda-
mentally novel powers/interactions/laws. This discussion is useful for appreciating
how nonlinearity moved from being a criterion of Strong emergence to being a
criterion of Weak emergence, and for seeing how a recognizable descendant of
nonlinearity as a criterion of Strong emergence is present in the aforementioned
motivation for new fundamental interactions, reflecting seeming violations of
conservation laws. By lights of the latter criterion, I observe, there is little moti-
vation for taking non-mental complex systems to be Strongly emergent—though
the case is less clear for certain mental phenomena, a topic to which I return in
later chapters. (Those primarily interested in whether complex systems are Weakly
emergent can skip this section without loss of continuity.)
In the second section, I consider three existing cases for the Weak emergence of
complex systems, appealing to one or other characteristic of such systems, as found
in Bedau 1997 and 2008, Mitchell 2012, and Batterman 2000 and 2002. As we’ll see,
the cases made in these discussions fall short of establishing that complex systems
complex systems 157
are Weakly emergent, in failing to rule out certain reductionist strategies for
accommodating the characteristics at issue. That said, the prospects for developing
these cases in a way that reveals an associated satisfaction of the conditions in
Weak Emergence are good. In particular, after expanding a bit on my degree-of-
freedom (DOF)-based account of Weak emergence, and arguing that attention
to order parameters does not undercut its general strategy, I argue that each of
the complex systems at issue in Bedau’s, Mitchell’s, and Batterman’s discussions is
reasonably seen as Weakly emergent, by lights of a DOF-based account.1
5.1 Are complex systems Strongly emergent?
5.1.1 Nonlinearity and unpredictability in the British

Emergentist tradition
Perhaps the first suggestion that nonlinearity is indicative of metaphysical emer-

gence is found in Mill 1843/1973, Ch. X, ‘On the Composition of Causes’. Recall
that Mill distinguishes between two kinds of effects: first, ‘homopathic’ effects,
which conform to “the principle of composition of causes” in being (in some
sense) mere sums of the effects of the component causes when acting in relative
isolation—as when the weight of two massy objects on a scale is the scalar sum
of their individual weights, or when the joint operation of two forces conforms
to vector addition in bringing an object to the place it would have occupied had
the forces operated sequentially; second, ‘heteropathic’ effects, which violate the
composition principle in not being mere sums in any clear sense. Mill takes this
distinction to be crucial, in that (he supposes) the advent of heteropathic—i.e.,
nonadditive or more generally, nonlinear—effects is indicative of the operation of
new laws:
This difference between the case in which the joint effect of causes is the sum of
their separate effects, and the case in which it is heterogeneous to them; between
laws which work together without alteration, and laws which, when called upon
to work together, cease and give place to others; is one of the fundamental
distinctions in nature. (408–9)
By way of illustration, Mill offers chemical compounds and living bodies as entities
capable of producing heteropathic effects. Recall also that, given the reciprocal
connection between powers and effects, it is straightforward to translate Mill’s talk
of effects into talk of powers: to say that an effect of a feature of a composite entity is
heteropathic, relative to effects of features of its composing parts acting separately,
is just to say that the feature of the composite has a power not had by associated
1 In this chapter I draw on and extend Wilson 2010b and 2013b.

lower-level base features when in linear combination. Finally, recall that Mill’s
appeal to the nonlinearity of effects is aimed at identifying a criterion of a higher-
level feature’s having a fundamentally novel power, enabling it (or its possessing
‘body’) to override the usual composition laws in the production of certain effects.
Other British Emergentists followed Mill in characterizing Strong emergence
in terms of nonlinearity, including Alexander (1920), who also characterized
emergent features as having powers to produce heteropathic effects; Morgan
(1923), who contrasted emergent features with resultant “additive and subtractive”
features; and Broad (1925), who offered scalar and vector addition as paradigms
of the compositional principles whose violation was characteristic of emergence.
These appeals to nonlinearity, like Mill’s, are best seen as attempts to provide
substantive metaphysical criteria of the operation of new fundamental powers,
forces/interactions, or laws that come into play only under certain complex
conditions, as when Broad (1925) says, “[T]he law connecting the properties of
silver-chloride with those of silver and of chlorine and with the structure of the
compound is, so far as we know, an unique and ultimate law” (64–5). Correspond-
ingly, McLaughlin (1992) accurately characterizes British Emergentism as
[…] the doctrine that there are fundamental powers to influence motion associ-
ated with types of structures of particles […] In a framework of forces, the view
implies that there are what we may call “configurational forces”: fundamental
forces that can be exerted only by certain types of configurations of particles […].
(52)
Before continuing, it is worth noting that the operative notion of nonlinearity

in the British Emergentist tradition was not one according to which it sufficed for
nonlinearity (hence, in this tradition, for fundamental novelty) that a feature of a
composite entity fail to be a linear combination of intrinsic features (powers, etc.)
of its composing entities—i.e., a linear combination of features of its composing
entities when ‘in relative isolation’. Such a contrast would render the view imme-
diately implausible. Consider, for example, the shape of a molecule composed of
some atoms. The molecule’s shape (and associated powers, etc.) is presumably not
the product of any new fundamental interactions; but on the other hand, this shape
is clearly not a function (linear or otherwise) just of the shapes of the atoms when
in relative isolation—also involved are the spatiotemporal and bonding relations
between the atoms associated with these being at some distance from each other.
In light of such cases, the most careful of the British Emergentists—namely,
Broad—included pairwise and other relatively noncomplex (e.g., spatiotemporal
and lower-level causal) relations between the lower-level entities (or states of affairs
consisting of lower-level entities standing in relatively noncomplex lower-level
relations) as among the physically acceptable ‘summands’ apt to be combined in
broadly additive fashion, and against which a given claim of emergent nonlinearity
complex systems 159
was to be assessed. Hence it was that Broad characterized “pure mechanism” as

involving broadly additive deducibility of all higher-level features from features of
lower-level entities “either individually or in pairwise combination”, and couched
his official formulation of emergence in terms of failures of in-principle deducibil-
ity of higher-level features from features of lower-level entities both “in isolation”
and “in other wholes”. The notion of linearity in these construals clearly adverts
to lower-level relational as well as intrinsic features of lower-level entities. Such an
appropriately broad understanding of linearity can accommodate, e.g., the shape
of a molecule, as a (vector) additive function of shapes associated with atoms in
pairwise or other relatively noncomplex combination.
Closely related to these appeals to nonlinearity as indicative of fundamental
novelty is the British Emergentist supposition that unpredictability, understood to
be in some sense ‘in principle’, is indicative of such novelty. As just noted, Broad’s
official characterization of emergence was couched in epistemic terms involving
in-principle unpredictability:
The emergent theory asserts that there are certain wholes, composed (say) of
constituents A, B, and C in a relation R to each other and that the characteristic
properties of the whole R(A, B, C) cannot, even in theory, be deduced from the
most complete knowledge of the properties of A, B, and C in isolation or in other
wholes which are not of the form R(A, B, C). (1925, 64)
As remarked in Ch. 4, the discussion preceding the formulation makes clear

that Broad’s appeal to failure of deducibility aims to characterize a metaphysical
notion of emergent autonomy, and moreover one tracking a form of emergence
incompatible with complete lower-level determination. A similar metaphysical
import attaches to Alexander’s (1920) claim that insofar as emergents cannot be
predicted or explained, they must be accepted with ‘natural piety’.
5.1.2 The fall of nonlinearity and unpredictability

as guides to fundamental novelty
Given an appropriately sophisticated understanding of the notion of linearity at

issue, it was quite reasonable for the British Emergentists to take nonlinearity
to be a mark of Strong emergence. To start, various paradigm cases of non-
emergent features of composite entities are simple scalar sums of features of
their composing entities, as when (bracketing relativistic considerations) the mass
of a composite entity is the sum of the masses of its composing entities. More
generally and more importantly, at the time it was common to suppose that
causal relations ultimately involve the exertion of various fundamental forces,
including gravity and electromagnetism, operating either singly or in tandem. As

Mill’s discussion makes especially clear, the combination of fundamental forces
was taken to proceed in accord with linear composition laws. Again recalling the
correspondence between powers and effects, linearity thus looked to provide a
general handle on when the features of composite entities did not involve or invoke
any new fundamental powers, forces, or laws. Conversely, failures of features or
behaviors of composite systems to be subject to linear analysis would have been
reasonably interpreted as indicating that some additional fundamental force—a
force not operative at lower, less complex levels of natural reality—was now on
the scene.
Though reasonable at the time, the British Emergentist supposition that non-
linearity is sufficient indication of the fundamental novelty at issue in Strong
emergence is no longer plausible. For one thing, the background supposition of
causation as involving additive combinations of fundamental push-pull forces is
now seen as largely heuristic, or in any case not generalizable; it is fundamental
interactions, involving particle exchanges, or yet more abstract accounts of the
existence and evolution of natural phenomena, that provide the ultimate story as
regards the ‘go’ of events; and to the extent that Newtonian forces can be seen
as real (as they arguably can be; see Wilson 2007) they are now assumed to be
nonfundamental (as constituted, somehow or other, by fundamental interactions).
For another, in the course of the 20th century, investigations into a wide range of
complex systems revealed not just that genuine nonlinearity was rampant, but that
in many of these cases the nonlinearity was generated in ways clearly not involving
any fundamental novelty. As I noted in my (2013b):
Even as early as the late 1880s, there were difficulties in seeing chaotic complex
systems of the sort associated with turbulence in fluids and gasses, and with
phase transitions, as linear. Attempts were made to explain away failures of linear
prediction in these cases as due to noise or imprecision in measurement; but
[…] the anomalies and epicycles associated with the supposition of linearity
eventually gave way to an understanding of complex systems as being genuinely
nonlinear. This is not to say, of course, that failures in prediction were thereby
(always) overcome; rather, such failures were given an alternative explanation as
reflecting, most saliently, the typical highly sensitive dependence of the associ-
ated nonlinear functions on initial conditions (a.k.a. ‘the butterfly effect’). (206)
As I further noted, that predictive anomalies in complex systems were not due
just to noise or imprecision was confirmed by attention to natural and artificial
nonlinear systems for which the relevant initial conditions could be specified with
complete accuracy:
complex systems 161
Population growth, for example, is straightforwardly modeled by the nonlinear

logistic map:
xn + 1 = axn − ax2n
Here a is a parameter representing birth and death rates, and is different for dif-
ferent systems. The behavior of a given system is heavily dependent on a. For most
values of a, the system evolves to a fixed point; as a approaches 4, the system’s
behavior becomes periodic, and subject to increasingly rapid bifurcation; for
a = 4, the system’s behavior becomes chaotic, with very small differences in initial
conditions xi , associated with distant decimal places, eventually leading to wildly
different trajectories. The discovery of natural nonlinear systems encouraged
attention to nonlinear systems in general, and with the advent of computers in
the latter half of the 20th century, much attention focused on artificial complex
systems such as cellular automata, where, as in Conway’s ‘Game of Life’, the
stipulated dynamics are nonlinear. (207)
The recognition of genuinely nonlinear systems clearly not involving any addi-
tional fundamental forces/interactions or associated novel powers fatally under-
mined the British Emergentist supposition that nonlinearity is indicative of Strong
emergence.
Similar remarks apply to the British Emergentist supposition that unpredictabil-
ity is so indicative. It was reasonable enough at the time to assume that an
insuperable failure of predictability would be an epistemic marker of fundamental
novelty: after all, if even a Laplacian demon couldn’t deduce certain higher-level
phenomena from lower-level goings-on, then what else, besides something fun-
damentally novel, could explain the occurrence of the higher-level phenomena?
But again, increased awareness and appreciation of the distinctive characteristics
of complex systems undermined the supposition. In particular, given the extreme
sensitivity to initial conditions associated with chaotic complex systems, it turns
out to be in-principle impossible to deduce the features of such systems, at least if
it counts as a failure of ‘in-principle’ deducibility that these features could not be
deduced even given the resources of the entire universe.
5.1.3 Might some nonlinear phenomena be Strongly emergent?
In light of the previous considerations, it is often taken for granted that complex
systems are at best Weakly emergent. Hence Bedau (1997) says:
An innocent form of emergence—what I call “weak emergence”—is now a

commonplace in a thriving interdisciplinary nexus of scientific activity […]
that includes connectionist modeling, nonlinear dynamics (popularly known as
“chaos” theory), and artificial life. (375)
As I’ll discuss in the next section, there is indeed a good case to be made that many
complex systems are (at best) Weakly emergent. Still, it is worth noting that stated
reasons for the claim that complex systems are never Strongly emergent aren’t
compelling.
Newman (1996) cites the fact that such systems are ‘strictly deterministic’
in support of the claim; but nothing prevents Strongly emergent features from
entering, both as regards their emergence and their subsequent evolution, into a
deterministic nomological net. Nor does the fact that features of complex systems
are ‘derivable’ from nonlinear equations and initial (or boundary) conditions
establish physical acceptability, since—as the British Emergentist tradition makes
explicit—unlike linear combinations, that nonlinear combinations of physically
acceptable features are themselves physically acceptable is not obvious. Bedau
(1997) claims that features of complex systems are physically acceptable in being
“structural”—that is, in being features of a configuration consisting in some
lower-level entities standing in lower-level relations; but given that the features
of the configurations at issue do not consist solely in additive combinations of
features of their parts, that such features are merely structural (as with, e.g.,
the shape of a molecule), in a sense that would entail physical acceptability, is
again not obvious. One might aim to support the general claim via an argument
by analogy, maintaining that insofar as various surprising features of complex
systems (period doubling, extreme sensitivity to initial conditions) can be modeled
in comparatively simple and artificial systems for which it is uncontroversial that
no new fundamental novelty is at issue, there is no reason to suppose that more
complex natural complex systems involve fundamental novelty, either. But this
argument by analogy fails, for precisely what is at issue is whether, in the more
complex natural cases, the nonlinear behaviors at issue have a physically acceptable
source.2
Indeed, there is in-principle room for maintaining that Strong emergence is at
issue in at least some cases of complex systems. Consider, for example, cases where
the nonlinear phenomena involve feedback between the micro-entities constitut-
ing the base, associated with strange attractors and other dynamic phenomena.
2 The thought here is not so different from that undercutting the supposition that Strong emergence
has been generally discredited by scientific advances; from the fact that some previous candidates for
Strong emergence are no longer such candidates, it doesn’t follow that all relevant phenomena will
admit of treatment in physically acceptable terms.
complex systems 163
As Silberstein and McGeever (1999) note, the nonlinearity at issue in complex

systems might be taken to involve a kind of system-level holism:
What is the causal story behind the dynamics of strange attractors, or behind
dynamical autonomy? The answer, it seems to us, must be the nonlinearity found
in chaotic systems. […] But why is non-linearity so central? […] Non-linear
relations may be an example of what Teller calls ‘relational holism’ […].3 (197)
Silberstein and McGeever go on to suggest that relational holism of this sort might
reflect emergent features’ possessing fundamentally new powers (“irreducible
causal capacities”), in line with Strong emergence.
It has also been suggested (or interpreted as being suggested) that the singular-
ities standardly associated with thermodynamic phase transitions are indicative
of Strong emergence. Hence Menon and Callender (2013) interpret Batterman’s
claim that “thermodynamics is correct to characterize phase transitions as real
physical discontinuities and it is correct to represent them mathematically as
singularities” (Batterman 2005, 234) as signaling Batterman’s commitment to
phase transitions’ being emergent along British Emergentist lines. Whether Teller
and Batterman would agree that relational holism or thermodynamic singularities
should be understood as involving new fundamental powers/interactions/laws is
disputable (as I’ll shortly discuss in re Batterman’s view). Still, for present purposes
the crucial point is that one could coherently take Strong emergence to underlie
some features associated with some complex natural nonlinear systems.
Let’s sum up the results so far.
First, whether or not all complex systems are physically acceptable, in any
case there’s no doubt that some are. The general moral to be drawn from the
identification of straightforwardly mechanistic and artificial complex systems is
that, contra Mill, Broad, and the other British Emergentists, neither nonlinearity
nor (even ‘in-principle’) unpredictability are sufficient indication of fundamental
powers/forces/interactions/laws.
Second, though many complex systems clearly do not involve Strong emer-
gence, stated reasons for the general claim that all complex systems are at best
Weakly emergent are uncompelling; as it stands, the general claim is something
of an article of faith. It would be nice if, given that nonlinearity and in-principle
unpredictability can no longer be seen as criterial of Strong emergence, there
were an alternative criterion which could distinguish physically acceptable from
physically unacceptable cases of complex systems (assuming any of the latter exist).
3 In re relational holism, see also the provisional definition of emergence offered by Thompson and
Varela (2001).
I turn now to identifying such a criterion, as it enters into my preferred account of

Strong emergence.
5.1.4 Nonlinearity’s descendant
As previously noted, in Wilson 2002a I offer an account of Strong emergence along

British Emergentist lines, which fills in the sense in which the power of a Strongly
emergent feature is fundamentally novel in terms which explicitly (as opposed
to implicitly, as on Broad’s characterization in terms of ‘in-principle failure of
deducibility’) involve the coming into play of a new fundamental force/interaction.
As above, at the heart of the Strong emergentist position is that some composite
systems have features associated with new powers, grounded in new fundamental
forces or interactions, suited to enter into producing effects that lower-level
configurations can’t (directly) enter into producing. McLaughlin (1992) claims
that those suspicious of forces (and presumably also of interactions) can dispense
with this aspect of the view, retaining only the appeal to new powers. But as we saw
in the previous chapter, there is some motivation for thinking that an appeal to
fundamental forces or interactions provides perhaps the best response to (among
other objections) the ‘collapse’ objection, according to which the supposedly novel
powers associated with Strongly emergent entities or features are always inherited
by the base-level entities or features. Recall that on this strategy of response, the
collapse objection is avoided by understanding Strong emergence in interaction-
relative terms:
Interaction-relative Strong Emergence: Token feature S is Strongly emergent from

token feature P relative to the fundamental interactions in a set {F} just in case (i) S
cotemporally materially depends on P, and (ii) S has at least one token power that
is not identical with any token power of P that is grounded only in fundamental
interactions in {F}.
Again, Interaction-relative Strong Emergence makes room for there to be Strong

emergence: even if, taking all fundamental interactions into account, any power
of a cotemporally materially dependent feature is inherited by its base feature,
it remains that higher-level features may be associated with powers that are
relevantly ‘new’, in not being grounded only in the set of physical fundamental
interactions.
Interaction-relative Strong Emergence also has three features relevant to under-
standing the bearing of the nonlinearity characteristic of complex systems on
Strong emergence.
complex systems 165
First, the account nicely accommodates the supposition that features of a com-
posite entity that can be analyzed as broadly additive combinations of physically
acceptable features of the composing entities will not be Strongly emergent.
For whatever the more specific account of how exactly powers are grounded
(schematically speaking) in fundamental interactions, in any case it is clear—to go
back to Mill’s original discussion—that every power of a feature that is a broadly
additive combination of physically acceptable features will be grounded only in
fundamental physical interactions, and so fail to be Strongly emergent.
Second, the account suggests an alternative criterion for Strong emergence,
which not only survives the advent of physically acceptable nonlinear systems,
but moreover provides the means of distinguishing, at least in principle, between
cases of nonlinearity that do and cases that don’t involve Strong emergence. Here
I have in mind the aforementioned criterion of a new fundamental interaction
(operative, for example, in the posit of the weak nuclear interaction), adverting
to apparent violations in conservation laws. A similar strategy makes in-principle
empirical room for testing whether the unusual features associated with complex
natural nonlinear systems are or are not due to configurational fundamental
interactions, by comparing the values of relevant conserved quantities predicted
by fundamental physical theory as attaching to features of composite entities (or
associated configurations), with the actually observed values of these quantities.
Again: if there’s less (or more) energy coming out than going in, for example, we
might well be inclined to conclude, following accepted scientific procedure and as
per the Strong emergentist thesis, that a new configurational force/interaction has
come into play.
Third, in the appeal to apparent violations of conservation laws as a sufficient
criterion of Strong emergence we have, it seems to me, a recognizable descendant
of the British Emergentist appeal to apparent violations of linearity as such a
criterion. For an apparent violation of a conservation law serves, as an apparent
violation of a linear composition law was reasonably but incorrectly taken to do,
to flag that the whole is more than the mere sum of its parts, such that some
fundamentally novel powers (forces, etc.) and laws must be posited, if the sum—of
forces, of conserved quantities—is to come out right.
All this said, I take it that there is not much motivation for thinking that any
complex natural nonlinear systems involve new fundamental interactions, with
the possible exception of those systems associated with qualitative consciousness
and free choice (to which we will return in Chs. 7 and 8). Still, it is useful to observe
that there is a criterion for Strong emergence upon which (and unlike bare appeals
to nonlinearity, relational holism, unpredictability, or representational mismatch)
all parties are likely to agree, and which could, in principle, be tested for; and that
by lights of this criterion the Strong emergence of at least some complex systems
has not been empirically discredited, since the relevant experiments simply have
not been done.
5.2 Are complex systems Weakly emergent?
I turn now to considering whether any complex systems are reasonably taken to
be Weakly emergent. My focus will be on three prominent accounts where it is
claimed that one or more of the aforementioned characteristics supports seeing
complex systems of a certain variety as emergent: first, Bedau’s (1997 and 2008)
appeal to algorithmic or explanatory incompressibility, applied to properties such
as being a glider in Conway’s Game of Life; second, Mitchell’s (2012) appeal to
self-organization, applied to group behaviors such as the flocking of birds; and
third, Batterman’s (2002 and 2000) appeal to asymptotic universality, as applied
to thermodynamic systems undergoing phase transitions.⁴ As prefigured at the
beginning of this chapter, while the discussions at issue in these accounts do not,
as they stand, establish even the Weak emergence of the complex systems at issue,
the cases made can be supplemented so as to provide a sound basis for taking
complex systems of the target variety to be Weakly emergent.
5.2.1 Bedau’s appeal to algorithmic incompressibility
Bedau’s focus is on a feature of nonlinear systems shared by both chaotic and

nonchaotic nonlinear systems; namely, that such systems typically fail to admit
of analytic or closed solutions. The absence of analytic or otherwise ‘compressible’
means of predicting the evolution of such systems means that the only way to find
out what this behavior will be is by going through the motions: set up the system,
let it roll, and see what happens. Such algorithmic incompressibility serves as the
basis for Bedau’s (1997) account of physically acceptable emergence, applicable in
contexts where a composed system S has “microstates” (encoding intrinsic states
of its parts) whose time evolution is governed by a microdynamic D, and where S’s
“macrostates” are structural properties constituted wholly out of its microstates:
⁴ I do not discuss Newman’s (1996) treatment of the feature being in the basin of a strange attractor,
since his treatment (according to which such a feature is identical to some lower-level feature, albeit in
such a way that it is ‘epistemically impossible’ for us to discover which one) is explicitly ontologically
reductive. Newman’s discussion is representative of approaches on which the emergence of complex
systems is purely epistemological (see also Popper and Eccles 1977, Klee 1984, and Rueger 2000).
However illuminating such accounts may be as regards why we find the features and behavior of such
systems interesting, novel, or unpredictable, they do not provide any basis for taking complex systems
to be even Weakly metaphysically emergent. The accounts I will consider in the text aim, or can be seen
as aiming, to provide such a basis.
complex systems 167
Macrostate P of S with microdynamic D is weakly emergent iff P can be derived

from D and S’s external conditions but only by simulation. (378)
Derivation of a system’s macrostate “by simulation” involves iterating the sys-

tem’s microdynamic, taking initial and any relevant external conditions as input.
The broadly equivalent conception in Bedau’s (2002) takes physically acceptable
emergence to involve ‘explanatory incompressibility’, where there is no shortcut
explanation of certain macrostates of a composite system. In being derivable by
simulation from a microphysical dynamic, such macrostates are plausibly physi-
cally acceptable, such that emergence understood as involving incompressibility
“is consistent with reasonable forms of materialism” (Bedau 1997, 376).
By way of illustration, Bedau focuses on Conway’s Game of Life, an example of
a nonchaotic nonlinear map. The game consists in a set of simple rules, applied
simultaneously and repeatedly to every cell, each of which may be ‘alive’ or ‘dead’,
in a two-dimensional lattice. At each step in time, cells are updated as per the
following rules: (1) any live cell with fewer than two live neighbors dies; (2) any
live cell with two or three live neighbors stays alive; (3) any live cell with more
than three live neighbors dies; (4) any dead cell with exactly three live neighbors
becomes alive. Here there is no problem of sensitivity to initial conditions, since
these conditions consist just in the discrete ‘seeding’ of the lattice. Still, Bedau
argues that the property of being a glider in the Game of Life is emergent, in
being algorithmically incompressible. That this property does not involve any
Strong emergence is clear, since for cellular automata the long-term behavior of
the system is completely metaphysically determined by (‘derived from’) the lower-
level rules applying to cells in the grid. But that a given system will evolve in such a
way as to generate a glider can typically not be predicted from knowledge of initial
conditions (seeding) and these rules.
Before continuing, it is worth registering that, notwithstanding that Bedau
describes the failure of predictability at issue as being diachronic (involving an
inability to antecedently predict the later evolution of a given system), he aims
to motivate a view of the relevant complex systems as emerging from “underlying
processes” (1997, 395), and (as I’ll shortly discuss) as manifesting macro-level pat-
terns. Correspondingly, we can think of Bedau’s account as characterizing a form
of emergence in which certain temporally extended macrostates (or collections
thereof) cotemporally depend on underlying microstates (or collections thereof),
in ways that due to algorithmic incompressibility could not be antecedently
predicted.
Now, Bedau’s appeal to algorithmic incompressibility, like Newman’s appeal to
epistemically inaccessible identity (see note 4 of this chapter), might be thought
to support complex systems’ being merely epistemically emergent, an impression
seemingly confirmed when Bedau says that “weakly emergent phenomena are
autonomous in the sense that they can be derived only in a certain non-trivial way”
(2002, 6). Indeed, Bedau explicitly claims that emergent features of composite sys-
tems are both ontologically and causally reducible to features of their composing
systems:
[W]eakly emergent phenomena are ontologically dependent on and reducible to

micro phenomena. (2002, 6)
[T]he macro is ontologically and causally reducible to the micro in principle.
(2008, 445)
Notwithstanding these reductive claims, Bedau maintains that algorithmic incom-

pressibility provides a basis for genuinely metaphysical emergence. He offers two
reasons for thinking this, but (as I discuss in Wilson 2015b), neither establishes
the point.
The first is that the incompressibility of an algorithm or explanation is an
objective metaphysical (if broadly formal) fact:
The modal terms in this definition are metaphysical, not epistemological. For
P to be weakly emergent, what matters is that there is a derivation of P from
D and S’s external conditions and any such derivation is a simulation. […]
Underivability without simulation is a purely formal notion concerning the
existence and nonexistence of certain kinds of derivations of macrostates from
a system’s underlying dynamic. (1997, 379)
But such facts about explanatory incompressibility, though objective and hence
in some broad sense ‘metaphysical’, are not suited to ground, in particular, the
ontological or causal autonomy of emergent entities. What is needed for such
autonomy is not just some or other metaphysical distinction between macro- and
micro- goings-on, but moreover one which plausibly serves as a basis for rendering
the higher-level features at issue ontologically and causally autonomous—that is,
distinct and distinctively efficacious—with respect to the lower-level features upon
which they cotemporally materially depend.
The second reason Bedau gives is more promising—namely, that the algorithmi-
cally incompressible features of complex systems typically enter into macro-level
patterns and laws. As Bedau says:
[T]here is a clear sense in which the behaviors of weak emergent phenomena are
autonomous with respect to the underlying processes. The sciences of complexity
are discovering simple, general macro-level patterns and laws involving weak
complex systems 169
emergent phenomena. […] In general, we can formulate and investigate the basic
principles of weak emergent phenomena only by empirically observing them
at the macro-level. In this sense, then, weakly emergent phenomena have an
autonomous life at the macro-level. (1997, 395)
As such, Bedau maintains, “weak emergence is not just in the mind; it is real and
objective in nature” (2008, 444). Attention to macro-level patterns is a move in
the right direction towards gaining emergent autonomy; but, two points. First, I
don’t see how Bedau can maintain that some emergent goings-on are not “merely
epistemological” (2008, 451) and rather reflect an “autonomous and irreducible
macro-level ontology” (2002, 38) , while also maintaining that these same goings-
on are ontologically and causally reducible to micro-level goings-on. Either the
ontological/causal autonomy or the ontological/causal reducibility has to go.
Second, for purposes of blocking the potential reducibility of higher-level to
lower-level features, it isn’t enough merely to point to the fact that the higher-level
feature enters into macro-level patterns (and associated laws) which are in some
sense more general than the patterns (and associated laws) into which the lower-
level features enter. For the reductionist has various strategies for accommodating
such general patterns (by, e.g., taking the feature whose having corresponds to
instantiating a given pattern to be identical with a disjunction of specific lower-
level ways of implementing the pattern). Hence while something seems right about
attending to the comparative generality of the patterns associated with features like
being a glider, more needs to be said if reductionist accommodation of this sort of
pattern is to be blocked. Bedau does not say more along these lines, however, and
the upshot is that this strategy, like the previous one, does not in itself establish
that any complex systems are even Weakly metaphysically emergent.
That said, as prefigured I will later argue that there are good reasons to think
that a DOF-based account can accommodate the Weak emergence of the sort of
features that Bedau discusses. Moreover, it seems likely that the sorts of arguments
that I gave in Ch. 2 (§2.3) supporting the ontological and causal autonomy of
realized features on various accounts of realization (including functionalist, mech-
anist, mereological, determinable-based, and ‘ontological explanation’ accounts)
could be brought to bear here; for these accounts also typically aim to realistically
accommodate macro-patterns and general laws, and (as I argued) are reasonably
seen as blocking reductionism via satisfaction of the conditions in Weak Emer-
gence. Bedau’s account of physically acceptable emergence might well be profitably
understood as combining one or more of these accounts of realization with the
further stipulation, unique to certain nonlinear complex systems, that the behavior
of these systems is algorithmically or explanatorily incompressible.
5.2.2 Mitchell’s appeal to self-organization
I next turn to Mitchell’s (2012) account of the emergence of chaotic complex sys-
tems exhibiting what is sometimes called ‘dynamic self-organization’. In general,
Mitchell follows Kauffman (1993, 1995), Wimsatt (1994, 1996, 2007), Bedau (1997,
2008), Thompson and Varela (2001), Thompson (2007), Camazine et al. (2018),
and others in taking emergence to involve “certain types of non-aggregative
compositional structures” (179). Mitchell moreover maintains that in the case
of certain chaotic complex systems, the non-aggregativity is dynamic, in arising
in a process-like fashion from interactions of the constituents, and in involving
feedback loops of the sort characteristic of self-organized systems:
Self-organized systems are ones in which feedback interactions among simple

behaviors of individual components of a system produce what appears to be an
organized group-level effect. (183)
The flocking of birds is a case in point:
Simple additive relations and simple linear equations […] will fail to make sense
out of much of the complexity that we find in nature even though patterns
and structures emerge from the simple interactions of the constituents. The vee
pattern that emerges in a flock of geese or the more complex patterns of flocking
starlings are not predictable by an aggregation of behaviors of individuals in
solo flight, but only from the non-aggregative interaction or self-organizing that
derives from the local rules of motion plus feedback among the individuals
in group flight (see Couzin 2007; see Rosen 2007 for photos of the starling
patterns). Ontologically, there are just physical birds; there is no new substance,
no director at a higher level choreographing the artistic patterns the flocks make.
Nevertheless, this type of behavior is emergent. (179)
Here again, the characterization of the purported emergence in dynamic and

diachronic terms can be reconceived as involving the emergence of certain macro-
level patterns (e.g., a ‘vee’ pattern) from the spatiotemporally distributed lower-
level configurations of individuals upon which the patterns cotemporally materi-
ally depend.
Mitchell is sensitive to the threat of ontological and causal reduction faced
by certain accounts of physically acceptable emergence, but maintains that fea-
tures arising from dynamic self-organization are not subject to such reduction,
on grounds, first, that “interaction among the parts generates properties which
none of the individual components possess” and second, that “these higher-order
properties in turn can have causal efficacy, i.e., novelty” (179). I address each of
these motivations for blocking reduction, in turn.
complex systems 171
First, as previously discussed (e.g., in Ch. 1, §1.4.2), a conception of emergence

according to which it suffices for the operative notion of ‘non-aggregativity’ that a
composed whole have “properties which none of the individual components pos-
sess” is too weak to serve as an appropriate dialectical basis for this notion. Again,
the reductionist will happily allow that configurations (pluralities or structural
aggregates) often and indeed typically have properties that are ‘novel’ in the weak
sense of not being had, either by way of type or token, by individual components.
Hence it is that if there is to be an interesting distinction between reductionists and
emergentists of whatever stripe, the reductionist must be granted resources suited
unto making sense of (massively complex, spatiotemporally extended) lower-level
configurations and associated features.
Moreover, these additional resources need not be restricted to linear combina-
tions, contra what Mitchell seems to suggest in observing that “simple additive
relations and simple linear equations” do not suffice to accommodate flocking
behavior. For—not least because fundamental physics is itself nonlinear—the
reductionist is within their rights to take any lower-level interactions (linear
or nonlinear, as the lawful case may be) between components into account as
input into their theorizing—at least when those interactions are manifest when
in relatively non-complex combinations.⁵ But in that case, it is unclear why the
reductionist cannot agree with Mitchell that there is “scientific emergence” as
involving “concrete accounts of how and why rules of interaction among compo-
nents produce difficult-to-predict emergent behaviors”. The upshot is that neither
the fact that certain properties of flocks of birds are not had by individual birds,
nor the fact that these group-level properties are nonlinear functions of properties
of individuals, blocks the ontological reduction of the group-level features at issue.
Nor does Mitchell’s discussion of the causal novelty associated with group-level
behaviors block causal reduction. Mitchell notes a number of ways in which group-
level dynamics is associated with interesting causal interactions. These include
that self-organization results in stable properties (pertaining, e.g., to the density
of flocks being concentrated at the interior) which can be “the target of natural
selection [and hence] exhibit causal saliency (see Carere et al. 2009)” (180); that
behaviors giving rise to group-level properties can also, if heritable, explain certain
facts about natural selection; and that (as per experiments run by Page and Mitchell
on ‘computer bees’, discussed in their 1991 and 1998) groups can come to manifest
division of labor, even when the individuals are assigned very ‘minimal’ behaviors.
Mitchell (2012) suggests that these forms of causal interaction are distinctive in
incorporating historical and contextual features:
⁵ This is one place where a law-based conception of the individuation of levels appears to be superior
to an ‘ontologically lightweight’ conception. At a minimum, the latter approach needs to be somehow
modified to allow for some nonlinear combinations of features at a level to be placed at that level.
It is still the case that the properties and behavior of the component parts cause
the ensuing behavior of the system, but there is a shift of emphasis to the features
of the history and context that the system experiences to understand why one
outcome occurred rather than another. (182)
But here again, the reductionist will presumably maintain that, while these
forms of causal interaction are massively complex, there is nothing—at least for
all Mitchell’s discussion of the cases establishes—to prevent their being given a
reductionist treatment. In particular: nothing requires that causal interactions on
a reductionist account be spatiotemporally local or somehow free of history or
context; on the contrary (and as a causal variation on the theme of my Ch. 3,
§3.3.3 response to Melnyk), the reductionist can allow that spatiotemporally ‘wide’
circumstances crucially enter into what causes what. It is moreover worth noting
that the reductionist will point to Mitchell’s reference to ‘a shift in emphasis’
from local to broader features as suggestive of, and in any case as consonant
with, a merely pragmatic reading of the higher-level efficacy at issue. So far, then,
Mitchell’s stated considerations do not establish that self-organizing behaviors are
anything more than highly complex lower-level processes.
As prefigured, I will later argue that a DOF-based account can accommodate
the Weak emergence of the sort of features that Mitchell discusses. Moreover,
it is again worth noting that the considerations that Mitchell raises, pertaining
to group-level effects and stable patterns of behavior, are consonant with the
considerations highlighted in the various accounts of realization discussed in Ch.
2 (§2.3), and it is correspondingly likely that Mitchell could avail herself of certain
associated strategies for achieving ontological and causal autonomy. In particular,
it is plausible that, whatever exactly ‘self-organization’ comes to, in any given case
this will be associated with a functional role that can be implemented by diverse
(collections of) configurations involving lower-level interactions between lower-
level components. In the case of flocking behavior, for example, the functional role
will advert to the instantiation of a vee pattern, which pattern and associated role
can be implemented by diverse configurations of locally interacting birds. Hence
it is natural to see the features to which Mitchell is calling attention as a variation
on the theme of functional realization, on which the functional role at issue has
the distinctive characteristics of self-organizing systems.
5.2.3 Batterman’s appeal to asymptotic singularities
Batterman has written a great deal concerning the status as emergent or reducible
of special-science entities (see his 2002, 2005, and elsewhere), though there has
remained unclarity as regards whether he takes emergence to be a metaphysical
complex systems 173
rather than merely epistemic or representational phenomenon, and if so, of what

strength. As I’ll shortly discuss, there is reason to think that Batterman does not
have even a Weak account of metaphysical emergence in mind, much less the
Strong account some attribute to him. Still, Batterman’s work on emergence in
asymptotic regions of the sort associated with phase transitions is relevant to the
present discussion, both because chaotic nonlinear systems are associated with
such transitions, and because, whatever Batterman’s view of the matter, at least
one of the features of systems in asymptotic regions that he mentions can, as
I’ll later discuss, support taking complex systems to be Weakly metaphysically
emergent.
Let’s start with the basics of Batterman’s account of asymptotic emergence. An
asymptote in mathematics is a limiting value of a function that is approached
indefinitely closely, but never reached. For example, as x → 0 the function 1/x
goes to infinity; in this (though not in every) case the asymptote is associated with
a discontinuity. Interestingly, many ‘near-neighbor’ scientific theories involve
asymptotes: special relativity asymptotically approaches Newtonian mechanics in
the limit as v/c → 0, wave optics asymptotically approaches ray optics as 1/λ → 0,
quantum mechanics asymptotically approaches Newtonian mechanics as Planck’s
constant approaches 0, and statistical mechanics asymptotically approaches
thermodynamics in the ‘thermodynamic limit’, where particle number N and
volume V → infinity. Now, in the latter three cases, the asymptotic limits at
issue are associated with discontinuities in the regions near the asymptote. In
such cases of ‘singular’ asymptotic limits, Batterman suggests, we have reason
to take various objects or features associated with the asymptotic region (or
associated theory) to be emergent. In particular, Batterman suggests that various
features of systems undergoing phase transitions, including those associated with
certain critical exponents, are emergent features of such systems. Why so, and in
what sense?
Batterman’s most explicitly stated reasons for thinking that phase transitions
involve emergence cite certain explanatory factors. One explanatory concern
reflects a kind of theoretical mismatch between the near-neighbor theories at
issue, insofar as the discontinuities associated with taking the thermodynamic
limit, and which are commonly supposed to be needed to accommodate the
associated asymptotic phenomena, find no representational mirror in the analytic
functions of statistical mechanics. Even if there were no problem with deriving
specific instances of asymptotic features from the micro-theory, however, a second
explanatory factor would remain—namely, that the characteristic universality
of asymptotic phenomena cannot be properly explained by reference just to
lower-level ‘causal-mechanical’ explanations. As above, the behavior of systems
undergoing phase transitions is characterized by critical exponents. As Batterman
(1998) says,
What is truly remarkable about these numbers is their universality […] the
critical behavior of systems whose components and interactions are radically
different is virtually identical. Hence, such behavior must be largely independent
of the details of the microstructures of the various systems. This is known in the
literature as the “universality of critical phenomena”. Surely one would like to
account for this universality. (198)
Lower-level causal-mechanical explanations, even if available, cannot account for

universality, for, as Hooker (2004) puts it, these “will be infinitely various in detail
and this will block any reconstruction of what is universal about them” (442).
By way of contrast, Batterman argues, various methods for modeling asymp-
totic phenomena—most notably, the Renormalization Group (RG) method—do
provide an explanation of the universal features of systems undergoing phase
transitions. The RG method takes a system’s governing laws (e.g., its Hamiltonian)
and iteratively transforms these into laws having a similar form but (reflecting
moves to increasingly ‘larger’ scales) fewer parameters—i.e., DOF. In the limit of
applications of the method, the resulting Hamiltonian describes the behavior of a
single ‘block’, corresponding to the macroscopic system:
[T]he transformation effects a reduction in the number of […] degrees of

freedom within the correlation length. […] In effect, the renormalization group
transformation eliminates those degrees of freedom (those microscopic details)
which are inessential or irrelevant for characterizing the system’s dominant
behavior at criticality. (200)
That the method is suitably applied to systems undergoing phase transitions

reflects that near critical points, such systems cease to have any characteristic
length scale, and are ‘self-similar’ in that the laws governing the systems take the
same form at all length scales.⁶ Application of the RG method to such systems
enables calculation of the critical exponents associated with phase transitions, and
so provides an explanation of the universal behavior of systems near critical points:
[I]f the initial Hamiltonian describes a system at criticality, then each renor-
malized Hamiltonian must also be at criticality. The sequence of Hamiltonians
thus generated defines a trajectory in the abstract space that, in the limit as the
⁶ See also Lamb 2015: “In terms of phase spaces, RG methods […] systematically [reduce] the
dimensionality of infinitely or nearly infinitely dimensional systems with the result that certain
behaviors of distinct complex systems represented by distinct high-dimensional phase spaces may be
represented by an identical low-dimensional phase space. After renormalization, only the dimensions
required to account for the behavior of interest are included in this new, ‘renormalized’, phase space”
(90–1).
complex systems 175
number of transformations goes to infinity, ends at a fixed point. The behavior

of trajectories in the neighborhood of the fixed point can be determined by an
analysis of the stability properties of the fixed point. This analysis also allows for
the calculation of the critical exponents characterizing the critical behavior of
the system. It turns out that different physical Hamiltonians can flow to the same
fixed point. Thus, their critical behaviors are characterized by the same critical
exponents. This is the essence of the explanation for the universality of critical
behavior: Hamiltonians describing different physical systems fall into the basin of
attraction of the same renormalization group fixed point. […] This stability under
perturbation demonstrates that certain facts about the microconstituents of the
systems are individually largely irrelevant for the systems’ behaviors at criticality.
(201)
Batterman’s account of asymptotic emergence cites three features potentially

relevant to the emergence of many complex systems:
1. elimination of micro-level degrees of freedom (DOF)

2. universality of certain features or behavior
3. stability of certain behavior under perturbation.
As Hooker (2004) notes, these features are generally characteristic of chaotic

nonlinear systems:
In every case of so-called ‘critical phenomena’ […] the asymptotic domain shows
a universally self-similar spectrum of fluctuations. […] This is indicative of
chaos and occurs when behaviours are super-complexly, but still systematically,
interrelated. (440)
Indeed, the core similarities between critical phenomena in statistical mechanics

and chaotic nonlinear phenomena, including period doubling and intermittency
routes to chaos of the sort displayed by the logistic map, have led to an active
area of investigation in which “the logistic map is [understood as] a prototypical
system […] for the assessment of the validity and understanding of the reasons
for applicability of the nonextensive generalization of […] Boltzmann-Gibbs
statistical mechanics” (Mayoral and Robledo 1970, 339). If one or more of the
features entering into Batterman’s account can be seen as supporting metaphysical
emergence, this result would apply to a wide range of complex systems.
Now, as it happens (and as I discuss further in Wilson 2013b) there is good
reason to believe that Batterman does not intend to be seen as offering either an
account of or a case of metaphysical emergence. As above, he is rather focused on
the question of what is required if the critical behaviors of the systems in question
are to be explained. Hence Morrison (2012) correctly reads Batterman as offering

an ‘explanatory’ account of emergence:
I characterize Batterman’s account [of emergence] as explanatory insofar as the

main argument centers on how asymptotic methods (via the RG) allow us to
explain features of universal phenomena that are not explainable using either
intertheoretic reduction or traditional causal mechanical accounts […]. (143)
That said, nothing prevents us from considering whether any of the aforemen-
tioned features—elimination in micro-level DOF, universality, and/or stability
under perturbation—might serve as sufficient indicators of metaphysical emer-
gence. With an eye to sticking somewhat closely to Batterman’s work, we might
look especially to universality and stability under perturbation, since it is these
features which he has most consistently highlighted as motivating the ‘emergence’
not just of systems near critical points, but also of special-science entities to which
the RG approach and its associated strategy for eliminating DOF do not apply.
Universality and stability under perturbation are really two sides of the same
coin; as Batterman (2000) says, “most broadly construed, universality concerns
similarities in the behavior of diverse systems” (120). The suggestion that phys-
ically acceptable emergence might be a matter of universality or of stability
under micro-level perturbations is common enough; indeed, we saw something
similar in Bedau’s (1997) conjecture that the fact that certain features of nonlinear
automata (e.g., gliders) enter into ‘macro-level patterns’ might support such
features’ being ontologically autonomous. It is unsurprising, then, that the same
concerns with Bedau’s conjecture also attach to attempts to locate metaphysical
emergence in universal or stable features of composite entities—namely, that
reductionists have various strategies for accommodating such features. What is
needed, if these features are to be seen as tracking the ontological and causal
autonomy of composite entities, is a better response to the usual reductivist
strategies; but Batterman does not provide such a response—unsurprisingly, given
that his primary concern is with whether an appropriately explanatory account
of universality is available, not with whether such features are ontologically or
causally autonomous.
That said, it is likely that attention to the powers associated with features
manifesting universality or stability under perturbation would provide a basis for
establishing such autonomy. In particular, these characteristics are clear variations
on the theme of multiple realizability; and as I argued in Ch. 2, there are cases to
be made, on various accounts of realization, that multiply realizable features have
only a proper subset of the token powers of their realizers, at both the level of types
and the level of tokens, and so (in combination with the cotemporal dependence
complex systems 177
condition, which is clearly satisfied in cases of complex systems undergoing phase

transitions) satisfy the conditions in Weak Emergence.
In the remainder of this chapter, I want to focus on the third feature Batterman
discusses—namely, that of a system’s having degrees of freedom (DOF) that are
eliminated with respect to the system of base entities upon which it cotemporally
materially depends. In the next section, I’ll present a DOF-based account of Weak
emergence, as developed in Wilson 2010b, and argue that it is applicable not only to
complex systems undergoing phase transitions, but also to other complex systems,
including Bedau’s gliders and Mitchell’s flocking birds.
5.2.4 A DOF-based account of Weak emergence
Call states upon which the law-governed properties and behavior of an entity E
(object, system, or other particular) functionally depends the ‘characteristic states’
of E. A DOF is then, roughly, a parameter in a minimal set needed to describe an
entity as being in a characteristic state. Given an entity and characteristic state, the
associated DOF are relativized to choice of coordinates, reflecting that different
sets of parameters may be used to describe an entity as being in the state. More
precisely, the operative notion of DOF is as follows:
Degrees of Freedom (DOF): For an entity E, characteristic state S, and set of

coordinates C, the associated DOF are parameters in a minimal set, expressed
in coordinates C, needed to characterize E as being in S.
I’ll sometimes speak for short of ‘characterizing an entity’, with state and coordi-
nates assumed.
Some common characteristic states, and DOF needed to characterize certain
entities as being in those states, are as follows:
• The configuration state: tracks position. Specifying this state for a free point
particle requires 3 parameters (e.g., x, y, and z; or r, ρ, and θ); hence a free
point particle has 3 configuration DOF, and a system of N free point particles
has 3N configuration DOF.
• The kinematic state: tracks velocities (or momenta). Specifying this state
for a free point particle requires 6 parameters: one for each configuration
coordinate, and one for the velocity along that coordinate; hence a free point
particle has 6 kinematic DOF, and a system of N free point particles has 6N
kinematic DOF.
• The dynamic state: tracks energies determining the motion. Specifying
this state typically requires at least one dynamic DOF per configuration
coordinate, tracking the kinetic energy associated with each position

coordinate; other dynamic DOF may track internal/external contributions
to the potential energy.
Why attend to DOF in hopes of illuminating Weak emergence? To start, as

above, different entities, treated by different sciences, may be functionally depen-
dent on the same characteristic state (e.g., the configuration state). Moreover, as
above, the DOF needed to characterize intuitively different entities as being in
these states typically vary. Following these observations, I take the main cash value
of attention to DOF to lie in the fact that DOF track the details of an entity’s
functional dependence on its characteristic states, in a more fine-grained way than
the mere fact that the system is in the state does. Driving my account is the idea
that the fine-grained details concerning functional dependence that are encoded
in the DOF needed to characterize broadly scientific entities serve as a plausible
ontological basis for the individuation of such entities.
I start by observing an important tripartite distinction (again, see Wilson
2010b) relevant to such functional dependence, reflecting that the DOF needed
to characterize an entity may be reduced, restricted, or eliminated in certain
circumstances (typically associated with the imposition of certain constraints or
more generally, the presence of certain energetic interactions), compared to those
needed to characterize a (possibly distinct) entity, when such circumstances are
not in place. To prefigure: eliminations in DOF, in particular, enter into the
DOF-based account of Weak emergence. Let’s get acquainted with these different
relations and note an example of each.
First, constraints may reduce the number of DOF needed to characterize an
entity as being in a given state. So, for example, a point particle constrained to
move in a plane has 2 configuration DOF, rather than the 3 configuration DOF
needed to characterize a free point particle. In cases where a particular DOF is
given a fixed value, the laws governing an entity so constrained are still functionally
dependent on the (now constant) value of the DOF; hence such constraints do not
eliminate the DOF, but rather reduce it to a constant value. By way of example:
rigid bodies treated in classical mechanics have such a reduced set of DOF relative
to the unconstrained system of their composing entities.
Second, constraints may restrict the DOF needed to characterize an entity. A
point particle may be constrained, not to the plane, but to some region including
and above the plane. Characterizing such a particle still requires 3 configuration
DOF, but the values of one of these DOF will be restricted to only certain of the
values needed to characterize the unconstrained particle. Cases of restriction in
DOF are more like cases of reduction than cases of elimination of DOF, in that,
again, the entity’s governing laws remain functionally dependent on specific values
complex systems 179
of the DOF. By way of example: molecules, whose bonds are like springs, have
a restricted set of DOF relative to the unconstrained system of their composing
entities.
Third, constraints may eliminate the DOF needed to characterize an entity. For
example, N free point particles, having 3N configuration DOF, might come to
compose an entity whose properties and behavior can be characterized using fewer
configuration DOF, not because certain of the DOF needed to characterize the
unconstrained system are given a fixed value or restricted to a range of values,
but because the properties and behavior of the composed entity are functionally
independent of these DOF. A case in point is that of a spherical conductor of
the sort treated in electrostatics, which has DOF that are eliminated relative to
the system of its composing entities; for while the E-field due to free particles
depends on all charged particles, the E-field due to a spherical conductor depends
only on the charges of particles on its surface. Certain quantum DOF are also
eliminated in the classical (macroscopic) limit. For example, entities of the sort
treated by classical mechanics are ultimately composed of quantum entities, but
the characteristic states of classical-mechanical entities do not functionally depend
on the spins of their quantum components.
Two features of such special-science case studies are commonly assumed, and
I will follow the presentation in my (2010b) in taking them for granted in what
follows as holding of the special-science entities under discussion. First is that the
holding of the constraints relevant to reducing, restricting, or eliminating DOF
occurs as a matter entirely of physical or physically acceptable processes. Such
processes suffice to explain why sufficiently proximate atoms form certain atomic
bonds, why atoms or molecules engage in the energetic interactions associated
with SM ensembles, and so on. More generally, for each of the aforementioned
special-science entities E, the constraints on the ei which enter into their com-
posing E are explicable using resources of the theory treating the ei (or resources
of some more fundamental theory, treating the constituents of the ei ). Call such
constraints ‘ei -level constraints’. A second important feature of the case studies is
that it is again plausible that all of the properties and behavior of the special-science
entities E at issue are completely determined by the properties and behavior of their
composing ei , when these stand in the relations relevant to their composing E.
In my (2010b), I drew on the case studies and associated features as motivating
the following (here, renamed) DOF-based account of Weak emergence:
DOF-based Weak Emergence: An entity E is Weakly emergent from some entities

ei if
1. E is composed by the ei , as a result of imposing some constraint(s) on the ei .

2. For some characteristic state S of E: at least one of the DOF required to

characterize the system of unconstrained ei as being in S is eliminated from
the DOF required to characterize E as being in S.
3. For every characteristic state S of E: every reduction, restriction, or elimination
in the DOF needed to characterize E as being in S is associated with ei -level
constraints.
4. The law-governed features of E are completely determined by the law-
governed features of the ei , when the ei stand in the relations relevant to
their composing E.
The conditions should be understood as relativized to occasions, as per usual. Here

the first and fourth conditions encode the supposition of cotemporal material
dependence of an emergent entity E and its features; on the usual supposition
that the ei are physically acceptable, the first condition more specifically encodes
physical monism and the fourth condition more specifically encodes Physical
Causal Closure. The second condition, against the backdrop assumption of the
fourth condition, encodes the satisfaction of the Proper Subset of Powers Condition
for at least one feature (characteristic state) of E, for reasons that I will present
shortly. Finally, the third condition encodes the assumption that any constraints
associated with features of E (which constraints, as discussed in Ch. 3, are typically
operative in cases where a feature satisfies the Proper Subset of Powers Condition)
are a result of only ei -level (i.e., physically acceptable) processes.
One last point before we proceed: in this account I cut out the feature ‘middle-
man’ and focus directly on the emergence of an entity E. If we want to express a
DOF-based account of physically acceptable emergence in terms of features, this is
perhaps most straightforwardly done by replacing reference to E with reference to
the token feature of being an entity of type E, and similarly for the configuration of
ei upon which E cotemporally materially depends. Here I’ll stick with the entity-
based formulation, for continuity with discussion in my previous article.
The ontological irreducibility of DOF-eliminated entities

Systems that are emergent by lights of the above account are physically acceptable,
given that the unconstrained system of composing entities is physically acceptable.
In my (2010b), I argue for this in some detail, but here I’ll just observe that this
result is to be expected, given that both the constraints relevant to composing E,
as well as all of E’s law-governed features (hence associated powers), are physi-
cally acceptable (in addition to, as above, satisfying the cotemporal dependence
condition). More crucial is to establish that entities satisfying the conditions in
DOF-based Weak Emergence are ontologically and causally autonomous.
I start by expanding on the general DOF-based strategy for accommodating the
ontological and causal autonomy of Weakly emergent goings-on. Let us suppose
complex systems 181
that some entities ei are treated by the lower-level physical laws Lp , and that an
entity E satisfies the conditions in DOF-based Weak Emergence with respect to
the ei , such that (among other things) the DOF needed to characterize E as being
in some characteristic state are strictly fewer than those needed to characterize
the system of lower-level physical ei as being in that state. Suppose also, to
fix ideas and as is typically the case for macro-entities in the classical limit (a
point to which I will return in Ch. 6, §6.1.1), that among the DOF eliminated
in characterizing E are those pertaining to quantum spin. In this case, E will
not be appropriately identified with any lower-level physical goings-on, whatever
they might be. Why not? Because (as previously noted in Ch. 3, §3.2.3) the
goings-on appropriately placed at Lp will be those whose specification includes
all the DOF needed for the physical laws to operate, including quantum spin.
Since E’s characterization, by assumption, fails to include (DOF containing) any
information about quantum spin, the lower-level physical laws will not be able to
take an entity such as E as input. E, and more generally any goings-on whose DOF
are eliminated vis-à-vis their lower-level physical realizers, cannot be identified
with any lower-level goings-on, since their specifications fail to include all the
DOF needed for the lower-level physical laws to operate. This much shows that an
entity E’s satisfaction of the conditions in DOF-based Weak emergence suffices to
ensure that E is ontologically autonomous—that is, distinct—from any lower-level
goings-on.
Relatedly, attention to the relation between the DOF characterizing E and
the DOF characterizing its realizing system(s) also provides a basis for E’s
satisfaction of the Proper Subset of Powers Condition, and so for E’s being causally
autonomous—that is, distinctively efficacious—with respect to E’s realizers.
Indeed, there are two results to this result.
The first route proceeds by attention to the bearing on E’s powers of E’s having
strictly fewer DOF than the realizing system of composing ei . To start, as per the
fourth condition in DOF-based Weak Emergence, the law-governed features of E
on a given occasion are completely determined by the law-governed features of
the system of ei on that occasion. On the intended understanding of ‘complete
determination’, it follows that every token power of E, on a given occasion, is
identical to some token power of the system of ei , such that E satisfies the Token
Identity of Powers Condition. The reverse is not the case, however: for the system
of ei , but not E, has token powers associated with quantum spin. It follows that
the system of ei has more token powers than E, and the Proper Subset of Powers
Condition in the schema for Weak emergence is satisfied. Correspondingly, and as
previously discussed, considerations of difference-making and/or comparatively
abstract systems of laws provide a sound basis for taking E to be distinctively
efficacious with respect to the system of ei .
The second route proceeds by attention to the bearing on E’s powers of its being
treated by a special science that is ‘special’, in particular, in treating entities such as
E only under certain conditions which are in some or other sense prerequisites for
the eliminations in DOF at issue. Given that (as we are assuming, for purposes of
illustration) E is a composed entity whose satisfaction of the conditions in DOF-
based Weak Emergence involves elimination of DOF associated with quantum spin,
the relevant conditions would be those pertaining to the existence of composed
entities in the classical limit, including that temperatures and energies be low
enough to allow for composed entities to exist. This provides a second basis for
satisfaction of the Proper Subset of Powers Condition. As I put it in my (2015b):
[W]hat powers an entity has are plausibly a matter of what it can do; and the
sciences are plausibly in the business of expressing what the entities they treat
can do. It follows that, plausibly, what powers an entity has are expressed by
the laws in the science treating it. The powers of E are thus those expressed
by the laws in the theory treating (constrained) entity E, while the powers of
[the system of ei ] are those expressed by the laws in the more fundamental
theory treating the (relatively unconstrained) lower-level constituents of [the
system of ei ]—that is, the constituents of [the system of ei ] as existing both inside
and outside the constraints associated with E. Consequently, the laws of the
theory treating E express what happens when certain lower-level entities stand
in relations associated with certain lower-level constraints, and the laws treating
[the system of ei ] express what happens when certain lower-level entities stand
both in these relations and in other relations not associated with the constraints.
Hence [the system of ei ] has more powers than E, and the proper subset relation
between powers in Weak emergence is thus in place. (387)
Order parameters and DOF

Before going on to apply a DOF-based account of Weak emergence to the case
of complex systems, I want to address the potential import of a related feature of
certain complex systems having eliminated DOF—namely, the association of such
systems with an order parameter. Both Morrison (2012) and Lamb (2015) agree
that elimination of DOF is characteristic of Weak emergence in complex systems,
but they also each discuss order parameters as equally and perhaps more crucial.
For example, Morrison says:
The basis of the idea of universality is that the fixed points are a property of
transformations that are not particularly sensitive to the original Hamiltonian.
What the fixed points do is determine the kinds of cooperative behavior that are
possible, with each type defining a universality class. The important issue here is
not just the elimination of irrelevant degrees of freedom; rather it is the existence
complex systems 183
or emergence of cooperative behavior and the nature of the order parameter

(associated with symmetry breaking) that characterizes the different kinds of
systems. (160)
Lamb (2015) suggests something similar, albeit pitched in terms of explanation:
A reductive explanation is one in which all of the constraints are characterized in

terms of small-scale DOF. […] [R]epurposing some of Wilson’s terminology, for
an explanation to be reductive it must be the case that the DOF needed to explain
why E exhibits S includes only DOF that characterize the smallest scale DOF of
ei […]. When a system’s components have minimal or linear coupling it is very
likely that an adequate explanation of the system’s behaviors and features will be
reductive. A non‐reductive explanation is one for which an order parameter is
identified in the explanans. In this case, the order parameter identifies large‐scale
features or behaviors that constrain the phenomena of interest. In terms of DOF,
non-reductive explanation is an explanation for which at least one of the DOF
used to explain E as being in [C] is not a DOF required to characterize the
smallest-scale [ei ] in the set [ei ] exhibiting σ. (71)
The order parameter identifies a large-scale constraint on relatively smaller fea-
tures of the system, such that the behaviors of the smaller features are dependent
on system-wide organizational features. (74)
One might correspondingly wonder if the real indicator of metaphysical emer-

gence in the cases at hand is reflected in the introduction of an order parameter
rather than an elimination in lower-level DOF. And one might moreover wonder
whether, as Lamb seems to suggest, such a parameter might be seen as introducing
a new DOF, associated with a new power to constrain lower-level goings-on,
contra the assumption, in DOF-based Weak Emergence, that operative constraints
on the lower-level entities are themselves lower-level. Supposing so, it might be
thought that attention to the DOF associated with complex systems near critical
points suggests that such systems are Strongly, not Weakly emergent. As Lamb
(2015) notes, “If the order parameter characterization identifies a novel physical
feature or power, not characterizable in terms of a smaller scale, then that is our
best evidence for strong emergence” (117).
Lamb goes on, however, to register that he suspects that “strong emergence of
this form would not be unscientific or a violation of physicalism” (117). Similarly,
Morrison (2012) is clear that, for the cases of symmetry breaking (as in the
transition to a ferromagnetic or superconducting state) that are her primary
focus, the underlying micro-physics is completely determinative of the higher-
level phenomena:
Symmetry breaking is reflected in the behavior of an order parameter that

describes both the nature and magnitude of a broken symmetry. In the ferro-
magnetic state, the order parameter is represented by the vector describing the
orientation and size of the material’s magnetization and the resulting field. In the
superconducting case, the order parameter is the amplitude φ of the macroscopic
ground state wave function of the Cooper pairs. The electromagnetic properties
in a superconductor are dominated by Cooper pairs, whereas electrons in a metal
normally behave as free particles that are repelled from other electrons because
of a negative charge. Because Cooper pairs appear only at T (their presence
indicates that the system has undergone a phase transition), they give rise to
the order parameter, which involves the Cooper pairs forming a single wave
function. In general, the order parameter can be thought of as an extra variable
required to specify the macroscopic state of a system after the occurrence of a
phase transition. In nonsuperconducting metals, gauge invariance ensures that
φ = 0. (152)
Three observations are relevant at this point. First, notwithstanding talk of phase
transitions’ (or symmetry breakings’) ‘introducing’ or ‘giving rise to’ order param-
eters, both Lamb and Morrison register that the parameter is present at the micro-
level. Indeed, standard discussions of order parameters take these to be present at
the micro-level, so to speak: what is new upon the occurrence of a phase transition
is not a given order parameter itself, but rather the value of the parameter, such
that (in a simple toy case) the parameter for a metal prior to magnetization has
the value zero, and the same parameter after magnetization has the value one.
More realistically, that a given order parameter is present prior to a given phase
transition is required in order for it to be able to register the range of values
associated with the coming-to-be of the ordered state, as when more and more
electrons become ‘Cooper pairs’ on their way to composing a superconductor. But
in that case, the parameter, or associated DOF, isn’t in fact ‘new’.
Second, it is worth noting that on one common understanding of order param-
eters, they are not properly seen as DOF, understood as parameters needed in
order to specify the law-governed properties and behavior of some characteristic
state of a given entity (object or system). Rather, order parameters are seen as
mere ‘phenomenological descriptions’ of the extent to which an entity manifests a
certain variety of order. As Kleman and Lavrentovich (2003) put it:
The concept of an order parameter has appeared with the attempt to describe the
order-disorder transition of alloys, specifically, to define a degree of disorder […]
First elaborated by Gorsky and Bragg and Williams to describe order-disorder
transitions in alloys, it has been developed in its modern form by Landau for the
purpose of a phenomenological description of phase transitions. (76)
complex systems 185
This much suggests that, pace Morrison, it is the eliminations in DOF rather than
the ‘introduced’ order parameter that are most important in understanding the
sense in which complex systems undergoing phase transitions are emergent.
Third, it is worth recalling that metaphysical emergence involves cotemporal
material dependence of emergents on base goings-on. As such, in consider-
ing whether phase transitions involve genuine metaphysical emergence, it isn’t
obviously to the point to consider whether a given feature associated with a
phase transition (say, superconductivity) is metaphysically emergent from some
states of affairs prior to the phase transition. The concern here, as in cases of
purported diachronic intra-level emergence (as in Humphrey’s cases of fusion, or
‘transformative’ emergence, of which particle annihilation is the primary case in
point) is that what is being offered as a case of emergence is better seen as simply
an interesting case of diachronic causation (see Ch. 1, note 7).
For the foregoing reasons, it is reasonable to assume that attention to order
parameters as involved in complex systems’ undergoing phase transitions poses
no threat either to the physical acceptability of such systems or to the potential
applicability to such systems of an account of emergence in terms of eliminations
in DOF.
5.2.5 DOF-based Weak Emergence and complex systems
Let’s now return to the question of whether any complex systems are plausibly seen
as Weakly emergent. As I’ll now argue, there are cases to be made that many such
systems, including those having the characteristics discussed by Bedau, Mitchell,
and Batterman, are Weakly emergent by lights of a DOF-based account.
I start with Batterman’s account of complex systems undergoing phase transi-
tions, having the features of universality and stability under perturbation. Given
Batterman’s observation (discussed in §5.2.3) that the Renormalization Group
(RG) method applies to such complex systems, and given that eliminations in DOF
(along with certain other suppositions which are here in place, concerning ei -level
constraints and ei -level determination) are sufficient for Weak emergence, we can
be comparatively brief.
As previously noted, the RG method applies to systems undergoing phase
transitions, which are relevantly similar to and indeed can be understood as
chaotic nonlinear systems; and that the RG method applies to a given system
provides as good an indication as we are likely to get that the system has DOF that
are not just reduced or restricted, but eliminated as compared to the unconstrained
system of its composing entities. We can thus argue as follows:
1. Systems that can be modeled by the RG method have eliminated DOF (Bat-
terman 1998 and elsewhere).
2. Chaotic nonlinear systems are modeled by the RG method.
3. Therefore, chaotic nonlinear systems have eliminated DOF.
4. Systems with eliminated DOF (and which satisfy other conditions satisfied by
chaotic nonlinear systems) are Weakly emergent (Wilson 2010b).
∴ Chaotic nonlinear systems are Weakly emergent.
As confirmation of the fact that many chaotic systems have eliminated DOF, it
is worth noting that one of the puzzles that Batterman raises for thermodynamic
systems carries over to chaotic complex systems, and is answered in just the same
way. The puzzle he raises concerns how thermodynamic systems can be a viable
object of study. Such systems—e.g., an isolated gas E—are composed of massively
large numbers of particles or molecules ei . Since (boundary restrictions aside) the
composite entity E in this case is effectively unstructured, shouldn’t it have the
same DOF as the system of unconstrained ei ? Supposing so, however, the success
of statistical mechanics (SM) is mysterious, since obviously we are not in position
to track such massive numbers (on the order of 1026 ) of DOF. As Batterman (1998)
puts it:
One wants to know why the method of equilibrium SM—the Gibbs’ phase
averaging method—is so broadly applicable; why, that is, do systems governed
by completely different forces and composed of completely different types of
molecules succumb to the same method for the calculation of their equilibrium
properties? (185)
The answer reflects that while the ei are not bonded, they are interacting via
exchanges of energy, and such interactions may not only restrict or reduce, but
eliminate DOF, as indicated by the RG method’s being appropriately applied to
such systems. This, then, is the answer to the puzzle: such systems are tractable
since the modes of interaction of their composing entities result in their having
DOF that are massively eliminated compared to the unconstrained system of
composing entities. Again:
[T]he renormalization group transformation eliminates those degrees of freedom

(those microscopic details) which are inessential or irrelevant for characterizing
the system’s dominant behavior at criticality. (Batterman 1998, 200)
A similar puzzle applies to chaotic complex systems. Recall that chaotic complex
systems are characterized by their extreme sensitivity to initial conditions. If
nonlinear systems are so sensitive and their resulting trajectories so “chaotic”, how
complex systems 187
is it that they can be, as they are, a viable object of scientific study? The answer,
I take it, is effectively the same: the composing entities, though not bonded, are
energetically interacting, in ways that, as application of the RG method reveals,
massively eliminate DOF needed to characterize the composite system. Here we
have a solution to the puzzle, and more to the present point, a decisive, empirically
supported case for taking the important class of chaotic nonlinear systems to be
Weakly metaphysically emergent.
Though applicability of the RG method to a given complex system provides a
quick route to taking the system to be Weakly emergent, such applicability is not
necessary in order to establish that a given complex system satisfies the conditions
of Weak emergence (DOF). Indeed, there are cases to be made that the complex
systems targeted in Bedau’s and Mitchell’s discussions also satisfy the conditions
of this account, and so are appropriately taken to be Weakly emergent.
First, recall Bedau’s suggestion that certain phenomena, such as gliders in Con-
way’s Game of Life, manifest a sort of autonomy characteristic of ‘innocent’—that
is, physically acceptable—emergence, as involving behaviors that are “autonomous
with respect to the underlying processes [involving] simple, general macro-level
patterns and laws” (1997, 395). In this artificial case, we do not have laws of nature
on hand to look to as a guide to the DOF of the entities at issue. Nonetheless, the
rules of the Game of Life will do for these purposes.
In particular, we can consider what DOF are required to specify the location
of the system of ‘live’ cells composing a ‘macro-entity’ in the Game of Life as
compared to the DOF required to specify the location of the macro-entity itself.
To focus on a simple case, consider a glider (the composed entity E) composed of
five ‘live’ cells, which starts at the origin in the configuration shown in Figure 5.1:
Figure 5.1 Beginning state of the glider
The glider ‘moves’ in a four-step sequence, at the end of which it returns to its
original configuration, one grid diagonally to the SE, so to speak, as shown in
Figure 5.2:
Now, let us suppose that the initial seeding of the grid contains only the initial
state of a glider, so that we can ignore any potential collisions of the glider with
1 2 3 4 5
Figure 5.2 Four-step periodic evolution of the glider
other ‘entities’. Given that the Game of Life takes place on a (potentially infinite)
two-dimensional grid, the location of each of the cells composing the glider E will
require two DOF, corresponding to values (relative to some arbitrarily chosen
origin point) along the x and y axes; hence determining the position of each of
the five ‘live’ cells composing a glider, in ways that can then be input into the
‘lower-level’ rules of the Game of Life, will require ten DOF. By way of contrast,
the position of a glider at n time-steps from t0 requires only two DOF, specifying
the x, y coordinates of the relevant diagonal cell, which at a time-step n/4 will be
associated by the ‘glider laws’ with the relevant stage of the glider. As such, a glider
E in the Game of Life has DOF that are eliminated relative to the system of its
composing ei .
Moreover, a glider E clearly satisfies the other conditions at issue in DOF-based
Weak Emergence. It is composed by the ei (the cells of the grid), as a result of
imposing some constraint(s) on the ei , where the constraints here simply reflect
how the rules of the Game of Life introduce a stable configuration. There is a
characteristic state S—namely, position—the specification of which for E requires
strictly fewer DOF than those required to characterize the system of ei . Every
reduction, restriction, or elimination in the DOF needed to characterize E as
being in a characteristic state is associated with ei -level constraints, since any such
constraints are a matter only of the operation of the rules of the Game of Life.
And more generally, the law-governed features of E are completely governed by
the law-governed features of the ei .
Correspondingly, the considerations previously brought to bear showing that
an entity E satisfying the conditions of DOF-based Weak Emergence with respect
to its system of composing ei can here be brought to bear in support of Bedau’s
claim that gliders in the Game of Life are Weakly emergent.
Next, recall Mitchell’s suggestion that certain complex phenomena, such as
flocking birds, manifest dynamic self-organization. Here the composing entities
are in the first instance the individual birds making up the flock. Specifying
the position of each bird requires at least three DOF (actually, it would require
many more, but for present purposes this won’t matter); hence specifying the
position of 100 birds would require 300 position DOF. When the birds compose a
flock, however, they mutually constrain each others’ positions, moving (to idealize
complex systems 189
somewhat) as an ensemble. Such constraints, as noted, can introduce restrictions,

reductions, or eliminations in the DOF needed to specify the position of the birds.
Now, since a flock isn’t a rigid body, and since knowing the position of the flock
at a time requires knowing its shape at that time, the DOF needed to specify the
position are all those needed to specify this shape. For this it suffices to know the
position of every bird on the external boundary of the flock; the positions of birds
in the interior of the flock are irrelevant.⁷ Since some birds in a flock will be in the
interior, it follows that the DOF needed to specify, at a time, the position of a flock,
will be strictly fewer than—will be eliminated as compared to—the DOF needed
to specify, at a time, the positions of all the individual birds constituting the flock.
That said, the case of flocking birds, unlike the case of the other complex systems
we have considered, introduces a new consideration as entering into the system at
issue—namely, consciousness, at least of a perceptual variety. Presumably it is some
part of birds’ flocking the way they do that they perceive and are otherwise aware
of other birds in the flock and what they are doing, and adjust their own behavior
accordingly. Of course individual birds are also perceptually conscious, so this fact
doesn’t in itself suggest that flocks fail to satisfy the conditions in DOF-based Weak
Emergence. But if the entities composing the flock are further decomposed into
lower-level physical entities, then whether flocks are Weakly emergent from lower-
level physical entities will hinge on the status as either Strongly or Weakly emergent
of such conscious mental features—a topic to which we will return in Ch. 7.⁸
Complex systems have frequently been offered as cases of actual metaphysical

emergence, but as I’ve argued in this chapter, previous accounts of such systems
as either Strongly or Weakly emergent do not, as they stand, establish this result.
This is commonly acknowledged as regards British Emergentist claims that
nonlinearity and unpredictability are sufficient marks of fundamental novelty of
the sort associated with Strong emergence—though as I’ll discuss in Chs. 7 and 8,
the empirical jury is still out over whether complex systems associated with
mentality manifest apparent violations in conservation laws, which criterion is,
⁷ How exactly to determine this boundary is a good question. Perhaps some sort of density measure
would come into play. In any case, the point in the main text doesn’t hinge on this epistemological issue.
⁸ This point highlights another way (beyond the relativization to fundamental interactions asso-
ciated with my preferred account of Strong emergence) in which attributions of emergence may be
relative. For all that has been established (or ruled out) thus far, it might be that flocks are Weakly
emergent from birds, but birds (qua conscious entities) are Strongly emergent from lower-level physical
phenomena.
I’ve here suggested, a recognizable descendant of nonlinearity as a marker of

fundamental novelty.
By way of contrast, it has frequently been claimed that certain characteris-
tic features of complex systems, including, in addition to unpredictability and
nonlinearity, the manifesting of macro-patterns, self-organization, universality,
and stability under perturbation, themselves serve as a sufficient basis for taking
the complex systems at issue to be Weakly emergent. As I’ve argued here, however,
such claims require supplementation, to the extent that there are available reduc-
tionist strategies for accommodating these features that need to be blocked if the
claim of emergence is to be defended.
It may be—indeed, it is likely—that such supplementation can be achieved
by attention to the sorts of considerations canvassed in the Ch. 2 discussion
of functionalist and other accounts of realization. In any case, there is another
feature of complex systems—namely, having DOF some of which are eliminated
as compared to the DOF needed to characterize the associated realizing system—
which provides a principled basis for blocking the threat of ontological and causal
reduction, in a way compatible with physicalism, given the physical acceptability
of the realizing system. This feature is most clearly present in cases of complex
systems undergoing phase transitions, to which the RG method (which effectively
works by eliminating DOF) applies. But as I’ve also argued, there are cases to be
made that certain macro-patterns in cellular automata (e.g., gliders) and certain
biological complex systems (e.g., flocks of birds) also have eliminated DOF, and
more generally satisfy the conditions in DOF-based Weak Emergence.
The upshot is that we now have confirmation of the common but previously
unsubstantiated belief that at least some complex systems are emergent in a way
compatible with physicalism—that is, are Weakly emergent.
6
Ordinary objects
In this chapter, I turn to the question of whether ordinary objects are either
Strongly or Weakly metaphysically emergent. By ‘ordinary’ objects I have in mind
objects which are uncontroversially inanimate (as Thomasson, 2007 puts it) or
nonliving (as Merricks, 2003 puts it), and of the sort with which creatures like
us are or may be perceptually acquainted.1 Such objects might be either natural
(rocks, feathers, mountains, planets) or artifactual (tables, baseballs, statues).2
While there are several competing metaphysical accounts of the nature of objects,3
the discussion to follow will be broadly neutral on which metaphysical account of
objects is correct, so long as this account does not rule out of court the possibility
that ordinary objects are metaphysically emergent.
Here the order of consideration is reversed from that of the previous chapter.
In the first section, I consider whether any ordinary objects are either Weakly
emergent or (as I will sometimes put it) are ‘at least’ Weakly emergent, in having
at least one feature satisfying the conditions in the schema for Weak emergence.⁴
First, I argue that ordinary objects of the sort appropriately treated by classical (or
‘Newtonian’) mechanics are Weakly emergent by lights of a DOF-based account,
for reasons anticipated in the previous chapter; second, I argue that a common con-
ception of artifacts as associated with sortal properties and distinctive functional
roles supports thinking of these as being at least Weakly emergent by lights of a
functional realization account; third, I argue that ordinary objects typically have
metaphysically indeterminate boundaries, which when coupled with an attractive
determinable-based account of such indeterminacy, indicates that such ordinary
1 In its focus on inanimate ordinary objects, the purview of the present chapter is more restricted
than that at issue in Korman 2020, according to which “ordinary objects are objects belonging to kinds
that we are naturally inclined to regard as having instances on the basis of our perceptual experiences:
dog, tree, table, and so forth”.
2 Some (e.g., Grandy, 2007) suggest that there is no deep difference between artifactual and natural
objects; here I’ll assume that there is a distinction, though as we’ll see, much of my discussion of the
one type of object applies, mutatis mutandis, to the other.
3 Among the many approaches are those on which objects are bundles of properties (either tropes,
as per Campbell 1990, or universals, as per Paul 2002), combinations of substrate and property (as per
Locke 1690 and Simons 1994), and hylomorphic compounds (as per Fine 2003 and Koslicki 2008).
⁴ Recall (as prefigured in Ch. 1, §1.2.3) that an object’s being ‘at least’ Weakly emergent is sufficient
for its not being reducible to lower-level goings-on, but leaves open the possibility that some other of
its features satisfies the conditions in the schema for Strong emergence, in which case the object would
be deemed Strongly, not Weakly, emergent.
DOI: 10.1093/oso/9780198823742.003.0006
objects are at least Weakly emergent, by lights of a determinable-based account of

such emergence.
In the second section, I consider whether any ordinary objects are Strongly
emergent. Here I argue that the best case for this stems from artifactual ordinary
objects whose functional or other characterization reflects or encodes certain
social practices involving normative or aesthetic goings-on. The ultimate status
of such objects as Strongly or rather just Weakly emergent hinges, like the status of
certain complex systems involving mentality, on the status as Weakly or Strongly
emergent of the associated mental features of persons, of the sort to be discussed
in the next chapters.
I close by observing that the results of this chapter serve not just to establish
the status of ordinary objects as at least Weakly emergent, but also to undercut the
motivations for Thomasson’s meta-ontological view, as discussed in her (2010)
and elsewhere, according to which investigations into the ontological status of
artifactual ordinary objects should proceed differently from investigations into the
ontological status of special-science entities.
6.1 Are ordinary objects Weakly emergent?
6.1.1 Classical objects
What I will call ‘classical’ objects are ordinary objects of the sort whose static and
dynamic behaviors are appropriately treated by classical or Newtonian mechanics,
understood as comprising, roughly, Newton’s three laws of motion and the grav-
itational and electromagnetic force laws. Classical objects include special-science
entities such as rocks, feathers, planets, and other comparatively structurally stable
inanimate objects. Artifacts may also count as classical ordinary objects; but to
avoid introducing complications associated with mentality (to be discussed later
in this chapter) I will restrict my focus here to special-science objects.
Classical mechanics as a special science

To start, it’s worth saying a few words in support of taking classical mechanics to be
a special science. It is sometimes suggested that classical mechanics is at best a false
but useful approximation to more fundamental theories. Hence Cohen-Tannoudji
et al. (1977) say
Classical laws cease to be valid for material bodies […] on an atomic or subatomic
scale (quantum domain). However, it is important to note that classical physics
[…] can be seen as an approximation of the new theories, an approximation
which is valid for most phenomena on an everyday scale. (9)
ordinary objects 193
Though natural, this line of thought can and should be resisted, at least if it is taken
to suggest that classical mechanics is irrelevant to understanding the metaphysical
structure of natural reality.
To start, it is characteristic of special sciences that they have restricted applica-
tion: the laws of thermodynamics, chemistry, cell biology, geology, botany, neuro-
biology, psychology, and so on, traversing the ladder of theories of constitutional
and biological complexity, do not hold in every context, but rather require for
their appropriate application that certain constraints or boundary conditions be
in place (as requisite for, e.g., the occurrence of thermodynamic systems, complex
inorganic and organic structures, specific genetic or environmental factors, and
so on). Hence while the restricted application of classical mechanics, along with
its presumably derivative status with respect to fundamental physics, spells its
failure as a fundamental theory, it remains a candidate for being a nonfundamental
theory, with something important to tell us about the nature of nonfundamental
reality. Indeed, like other special sciences, classical mechanics is reasonably seen
as tracking an important and distinctive level of ontological grain, and associated
laws of nature, salient in circumstances wherein lower-level details are irrelevant
to the dynamics of (relatively) large-scale goings-on. This is just how Feynman
(1965) characterizes classical mechanics:
Newton’s laws are the “tail end” of the atomic laws, extrapolated to a very large
size. The actual laws of motion of particles on a fine scale are very peculiar, but if
we take large numbers of them and compound them, they approximate, but only
approximate, Newton’s laws. Newton’s laws then permit us to go on to a higher
and higher scale, and it still seems to be the same law. In fact, it becomes more
and more accurate as the scale gets larger and larger. […] As we apply quantum
mechanics to larger and larger things, the laws about the behavior of many atoms
together do not reproduce themselves, but produce new laws, which are Newton’s
laws, which then continue to reproduce themselves from, say, micro-microgram
size, which still is billions and billions of atoms, on up to the size of the earth, and
above. (Vol. I, 19–2)
Hence it is that classical mechanics, qua special science, provides the basis for most
scientific and engineering-based investigations into and treatments of ordinary
objects.
Nor does the transition from force-based to energy-based mechanics—in par-
ticular, as was advisable for purposes of characterizing quantum phenomena—
pose any in-principle difficulty for taking classical mechanics to be a special sci-
ence in good standing. In particular, as I argue in detail in Wilson 2007, forces and
energies are interdefinable, and even if fundamental quantum processes involve
(e.g.) particle exchanges rather than Newtonian forces (which is not entirely clear,
insofar as analogical descriptions of such exchanges tend to involve forces), it

remains that Newtonian forces can be seen as nonfundamental goings-on which,
along with ordinary objects, are properly seen as part of the subject matter of
classical mechanics.⁵
To be sure, there remain further questions about the ontological status of
Newtonian forces, which have periodically been the target of anti-realist concerns.
In Wilson 2007, I address and respond to these concerns; to enter into these
details at this point would take us too far afield, however, so I invite the interested
reader to attend to that other discussion. Here I will follow Feynman and many
other scientists in seeing classical mechanics as a special science in good standing,
associated with distinctive laws of nature. Seeing classical mechanics in this light
provides our first route to the Weak emergence of some actual ordinary objects—
namely, classical ordinary objects—by lights of a DOF-based account.
The classical limit

Classical objects are associated with the so-called ‘classical limit’, in which (as in
Feynman’s remarks above) certain quantum features cease to be relevant to the
properties and behaviors of macro-objects, in ways that are plausibly interpreted
as involving an elimination (and not just a reduction or restriction) of quantum
DOF.
For example, consider DOF associated with quantum spin. Classical objects
are ultimately composed of quantum entities, whatever exactly these entities may
be; but the characteristic states of classical objects do not functionally depend on
the spins of the quantum components of these entities. Hence notwithstanding
that the values of quantum parameters may in some cases lead to macroscopic
differences—for example, readings on a measurement apparatus, and the like, as
in the case of Schrödinger’s cat—it remains the case that DOF such as quantum
spin are eliminated (and not just reduced or restricted) from those needed to
characterize entities of the sort appropriately treated by classical mechanics. Hence
for reasons discussed in Ch. 5 (§5.2.4), classical objects satisfy the Proper Subset
of Powers Condition, and more generally are appropriately seen as satisfying the
conditions in a DOF-based account of Weak emergence.
Another potential and very general source of elimination in quantum DOF
in the classical limit reflects that the probabilistic values of quantum mechanical
observables average out to their mean values in this limit, as per the main strategy
for understanding how classical mechanics ‘emerges’ from quantum mechanics
⁵ At least this is so for non-gravitational Newtonian forces. The prospects for seeing gravitational
forces as real, nonfundamental goings-on are less promising, at least given a General Relativistic
understanding of gravitational influence in terms of inertial motion.
(see, e.g., Messiah 1970, 215). Forster and Kryukov (2003) provide a useful
explanation by analogy of how this occurs:
It may be surprising that deterministic laws can be deduced from a probabilistic

theory such as quantum mechanics. Here, curve-fitting examples provide a useful
analogy. Suppose that one is interested in predicting the value of some variable
y, which is a deterministic function of x, represented by some curve in the x − y
plane. The problem is that the observed values of y fluctuate randomly above and
below the curve according to a Gaussian (bell-shaped) distribution. Then for any
fixed value of x, the value of y on the curve is well estimated by the mean value
of the observed y values, and in the large sample limit, the curve emerges out of
the noise by plotting the mean values of y as a function of x.
To apply the analogy, consider x and y to be position and momentum, respec-
tively, and the deterministic relation between them to be Newton’s laws of
motion. Then it may be surprising to learn that Newton’s laws of motion emerge
from quantum mechanics as relations between the mean values of quantum
mechanical position and quantum mechanical momentum. These deterministic
relations are known as Ehrenfest’s equations. In contrast to curve fitting, the
Heisenberg uncertainty relations tell us that the quantum mechanical variances
of position and momentum are not controllable and reducible without limit.
Nevertheless, it is possible for both variances to become negligibly small relative
to the background noise. This is the standard textbook account of how Newton’s
laws of motion emerge from quantum mechanics in the macroscopic limit.
(1040)
It is natural to read references to probabilistic variances becoming ‘negligibly

small’ as indicating that the DOF encoding such variances are eliminated in the
classical limit, and so as providing support for classical objects’ being Weakly
emergent—at least, on the assumption that the probabilistic aspects of quantum
theory correspond to objective features of quantum goings-on. This is true for
many interpretations of quantum mechanics, including versions of a Copenhagen
(or ‘orthodox’) interpretation on which quantum probabilities are grounded in
objective properties of particles to be, e.g., in a certain location if measured, and
versions of a spontaneous collapse theory, à la Ghirardi et al. 1986, on which
quantum probabilities are DOF of the wave function, but where these DOF are
understood as tracking properties of the associated quantum entities.⁶ It may
also be possible to see probabilistic quantum DOF as eliminated even if the
⁶ Hence Frigg (2009) notes, “A crucial assumption of the theory is that hits occur at the level of the
micro-constituents of a system [e.g.,] at the level of the atoms that make up the marble” (267). See also
Monton’s (2004) ‘object’ interpretation of a spontaneous collapse theory.
probabilities are ‘epistemic’, if what this comes to is something like statistically

random behavior, and the properties and/or behavior of (some) quantum entities
is functionally dependent on such randomness.
To sum up: various ordinary objects of the sort appropriately treated by clas-
sical mechanics have DOF required to specify the objects as being in a given
characteristic state relevant to the object’s law-governed properties (including
behavior) that are eliminated as compared to the systems of lower-level entities
upon which they cotemporally materially depend, and hence are at least Weakly
emergent by lights of a DOF-based account.⁷ Now, as above, for an object to be
Weakly emergent it is required not just that at least one of its features satisfy the
conditions in the schema for Weak emergence, but that the remaining features
and behaviors of the object are such as to either also satisfy the conditions in
the schema, or else be appropriately identified with (i.e., reducible to) lower-level
goings-on. On the assumption that a given ordinary object is in fact appropriately
treated by classical mechanics, this further constraint is plausibly in place, since
otherwise the object would have new powers to affect lower-level physical goings-
on in ways not entirely encoded (just) in classical mechanics. Hence we may say,
conditionally, that any object that is appropriately treated by classical mechanics
is Weakly emergent, simpliciter. Insofar as many special-science objects, including
rocks, planets, and the like, are reasonably supposed to be appropriately treated by
classical mechanics, all such entities are reasonably supposed to be actually Weakly
emergent, by lights of a DOF-based account.
6.1.2 Sortal features and functional realization
Artifacts are commonly taken to be associated with functional roles that to some
significant extent serve to characterize what it is to be an object of the type in
question. Hence Mitchell (2012) says:
A chair is describable as an artifact designed to function as something suitable

for humans to sit on. It is also the case that chairs are always made out of
something material: wood, iron, plastic, etc. So a particular chair could also be
described in terms of its material components. Chairs in general are described by
their capacity to function as something humans use for sitting. Chair-function is
realized by an individual entity’s physical components and structure. (174)
⁷ As discussed in Ch. 5 (§5.2.4), the eliminations in DOF at issue pertain to the system of composing
entities ei even when the latter are constrained to stand in the relations relevant to their composing
a given higher-level entity, since the lower-level laws applying to the ei , whether constrained or
unconstrained, require the full complement of lower-level DOF.
And Searle (2007) says:
Many of the most common concepts that we use in dealing with the world, for
example, concepts like ‘cars’ and ‘bathtubs’ and ‘tables’ and ‘chairs’ and ‘houses’,
involve the assignment of function. (8)
(While Searle speaks here of concepts, he also clearly takes the notion of function
as applying to the referents of the concepts.)
The view that artifacts are functionally characterized no doubt reflects that
artifacts are typically created with some purpose in mind. Here I want to develop,
in addition, another route to seeing artifacts as functionally characterized, that
proceeds by attention to what are sometimes called ‘sortal features’ (or ‘primary
kinds’, as in Baker 1993) of such objects. Lowe (2007) observes that “The idea that
objects are sortally individuated has a long and distinguished pedigree” (514); see
also, e.g., Wiggins 2001, 150–1 and Lowe 1989, 11–13. Candidate sortal features
for ordinary objects of the varieties at issue here would be features expressing
membership in the category at issue, such as ‘being a table’ or ‘being a statue’.
Why is attention to sortal features useful for purposes of assessing whether
a given object is metaphysically emergent? To start, note that while there are
different understandings of sortal features (see Grandy 2016 for discussion of
several definitions and applications), the common thread is that a sortal feature
encodes various conditions characterizing the object, at both type and token
levels. At the level of types, an object’s sortal feature specifies its ‘individuation
conditions’, encoding what distinguishes objects of the type in question from other
objects or types of entity. At the level of tokens, an object’s sortal feature specifies
its ‘identity conditions’, encoding what it takes for a particular object of the type to
be identical to another particular object of the type, either at a time or over time.
Identity conditions applying at a time are useful for purposes of counting how
many objects of the type associated with the sortal are in a given collection (how
many chairs are in the room right now?). Identity conditions applying over time—
sometimes called ‘re-identity’ or ‘persistence’ conditions—are useful for purposes
of determining whether an object of a given type has persisted through time, and
more importantly, change (is this the same chair as was in the room yesterday?).
It is the persistence conditions encoded in a given sortal feature that I want to
highlight as relevant for present purposes. To start, on a standard understanding
of the persistence conditions of artifacts, artifacts are compositionally flexible, in
that objects such as tables and statues can survive some (and sometimes quite
considerable) changes to the lower-level configurations (pluralities or structured
aggregates) upon which they depend. Relatedly, it is frequently supposed that
artifacts have compositionally flexible material origins, such that, e.g., a given table
(‘Woody’) might have originated from a block of wood somewhat different from
that from which it actually originated (see, e.g., Kripke 1972/80 and Salmon 1989);
such ‘modal’ compositional flexibility might also be encoded in the sortal features
of artifacts, as reflecting the conditions under which such objects⁸ are identical
across worlds, as well as times.
Now, given that the sortal features of artifacts typically encode that such objects
are compositionally flexible, both temporally and modally, the question naturally
arises: why so? What explains why artifacts are typically compositionally flexible?
And here a plausible answer, supporting the sort of characterizations of artifacts
mentioned above, is that artifacts are functionally characterized, in that what it
is to be an object of the sortal type at issue is to be an object that plays or is
capable of playing a specific functional role. For if artifacts are associated with
functional roles, and these functional roles are sufficiently abstract that they can
be implemented by multiple lower-level configurations, then the compositional
flexibility of artifacts would follow as a matter of course—effectively, as an objec-
tual variation on the theme of multiple realizability as usually applied to features.
So both the intuitive descriptions of artifacts, and a natural metaphysical basis
for the compositionally flexible persistence conditions of artifacts, support taking
artifacts to be functionally characterized (again, such that what it is to be an object
of the sortal type at issue is to be an object that plays a specified functional role);
and I will henceforth assume that this account of artifacts is correct.
Now, the fact that a given artifact is properly seen as functionally characterized
does not in itself guarantee that the artifact is Weakly as opposed to Strongly
emergent, for reasons having to do with the bearing of mentality on the intentional
specification and social implementation of the functions at issue. I’ll expand on
this further issue down the line. For now, it remains that such a characterization
pushes towards thinking that the object is at least Weakly emergent.
To see this, first note that while numbers or other abstracta may be functionally
characterized in ways that cannot (or at least cannot obviously) be cashed in causal
terms, the functional roles at issue in the case of artifacts are more specifically
causal roles. Hence Rosenberg (1994) says:
Causal descriptions are often called ‘functional’ role descriptions in the phi-
losophy of science, and I will use this terminology hereafter, understanding
‘functional’ to mean simply “role in a network of causes and effects”. Thus, the
functional description of an [axe] identifies it by its functional role, the causes
and effects into which it typically enters. An [axe] is a tool […] for cutting wood
(that is, its effect when applied with sufficient momentum will be to sever wood
into smaller pieces). (23–4)
⁸ Or their counterparts, if an object cannot exist at more than one world, as per Lewis 1986.
Given that the functional roles associated with artifacts (e.g., tables) are more
specifically causal roles, it follows that what it is to be an artifact of a certain sortal
type is to be an object characterized by a certain causal role, and in particular, to
be an object with certain powers. These powers are encoded in the sortal feature
of the artifact which, as Lowe put it, individuates the artifact. Hence at this point
we can return to a focus on the features of a given object as the potential locus of
the metaphysical emergence of the object itself.
Now, independent of the possibility that some of the powers associated with
the sortal feature of a given artifact are fundamentally novel in a way indicative
of Strong emergence, it seems likely that many of the powers of a sortal feature
will be token-identical with the powers of whatever lower-level configuration
(plurality or structural aggregate) serves as the dependence base for the artifact on
a given occasion. Indeed, it is commonly assumed that all of the powers of artifacts
are ultimately powers of their dependence base configurations. Hence Rosenberg
(1994) continues:
Now consider the class of objects that meet the purely functional characterization
of an [axe]. They all have to be material objects and have some structural
composition or other. But nothing in the functional definition requires that their
handles be made of wood or that their heads be made of steel. […] An [axe] is
thus “nothing but” the material out of which it is composed, even though two
different axes may share no fact of structural composition in common (above
the atomic level). (24)
Granting that an axe is “nothing but” its material base on a given occasion, every
power of the axe (or associated sortal feature) must be identical with a power of that
base (or associated sortal feature) on that occasion. Merricks (2003) is even more
explicit, saying “a baseball and its constituent atoms cannot do any more than those
atoms all by themselves” (59), and later, “everything that a baseball causes is caused
by its parts at some level of composition” (62). Even some of those who explicitly
take into account the role of human intentions in the creation of or social practices
involving artifacts seem to assume that an artifact’s lower-level dependence base is
“capable of fulfilling the functions” of artifacts, as in these remarks of Thomasson
(2007):
[I]f you think that the rules and practices that make up the game of baseball,
and intentions of those who rearrange atoms into appropriate spherical shapes,
are necessary for there to be baseballs, consider that for atoms to be arranged
baseballwise requires atoms tightly bonded in a spherical shape of such-and-such
diameter, such that they are jointly capable of fulfilling the functions of baseballs,
[i.e., being such that they are] bonded by people with intentions to make baseballs
that meet major league regulations, are usable and to-be-used in such games, and
so on. (17)
Supposing that the powers of (the sortal feature) of a given artifact do not go
beyond those had by its dependence base configuration, as the physicalist assumes,
we are a short step from seeing the artifact as Weakly emergent. For as above,
and as per the compositional flexibility of artifacts, these powers are associated
with a comparatively abstract functional or causal role, which can be implemented
by diverse base-level goings-on. Consider a baseball, for example. It cannot be
among the powers of a given baseball (or its sortal feature) to produce a certain
precise weight reading on a scale, since given its compositional flexibility, the
baseball might exist in the same extrinsic circumstances, and yet, if its dependence
base configuration is different, not produce the same reading; by way of contrast,
the dependence base configuration is not so compositionally flexible, and so it
(or its sortal feature) does have that power. It follows that the baseball (or its
sortal feature) has, on a given occasion, fewer powers than its dependence base
configuration (its sortal feature) on that occasion, and so (given the empirically
supported satisfaction of the dependence condition) satisfies the conditions in
Weak Emergence. And the same is true, more generally, for any compositionally
flexible artifact. Since artifacts are typically (always?) compositionally flexible,
artifacts are typically (always?) at least Weakly emergent.
It may be that the previous considerations also support, mutatis mutandis,
taking natural ordinary objects to be at least Weakly emergent. After all, special-
science objects, ranging from cells to mountains, tectonic plates, and planets, are
also commonly functionally characterized, and fall under sortal types encoding
their compositional flexibility (cf. Lowe’s 2007, 519, remarks that “sortal per-
sistence conditions [are] conditions determining what changes an object of any
given sort may or may not undergo, as a matter of natural law, while remaining
an object of that same sort”). Indeed, according to Rosenberg (1994), “[t]he
complexity of nature above the level of the molecule is the result of selection for
function and its blindness to structure” (55). It is also worth observing that in
the case of natural objects, the functional specification occurs as a result of natural
processes or laws as opposed to intentions or social practices, sidestepping the pri-
mary basis for thinking that functionally characterized objects might be Strongly
emergent.
Be all this as it may, since the primary examples of functionally specified special-
science objects are drawn from the biological sciences, and we are here focused
on nonliving ordinary objects, I won’t here develop this line of thought, but will
rather move on to a different reason for thinking that both artifacts and natural
ordinary objects are at least Weakly emergent, which has not yet been discussed
in the literature.
6.1.3 Metaphysically indeterminate boundaries
Ordinary objects, of either natural or artifactual varieties, typically appear to

have indeterminate—that is, imprecise—boundaries. As Tye (1990) observes,
“common sense has it that the world contains countries, mountains, deserts, and
islands […] and these items certainly do not appear to be perfectly precise” (396).
And as has been frequently noted, even the seeming distinctness of spatial bound-
aries of ordinary objects such as rocks, tables, and statues, dissolves upon closer
examination into an array of lower-level constituents (e.g., atoms) of decreasing
but non-zero concentration in the region of a given macro-object boundary—
a phenomenon which again suggests that such boundaries are in some sense
indeterminate.
How should such seeming boundary indeterminacy be understood? There
are three main strategies. The first takes indeterminacy to have its source in
how we represent the world (this is ‘semantic’ indeterminacy); the second takes
indeterminacy to reflect the limits of our knowledge of the world (this is ‘epistemic’
indeterminacy); the third takes indeterminacy to have its source, somehow or
other, in the world itself (this is ‘metaphysical’ indeterminacy). While some sorts
of indeterminacy may be best treated in (merely) semantic or epistemic terms,
object boundary indeterminacy is most naturally treated in metaphysical terms. It
is unclear how indeterminacy in ordinary object boundaries might be a semantic
matter, reflecting that we haven’t gotten around to drawing certain precise lines:
we are not inclined to draw such lines—most importantly, because attribution
of any particular precise boundary to an ordinary object (e.g., a mountain, or a
table) would be inappropriately arbitrary. Such arbitrariness renders it similarly
implausible to take seeming boundary indeterminacy to reflect our inability to
discern which perfectly precise boundary is in fact possessed by the object at issue,
even granting (what might be questioned⁹) that the boundaries of the lower-level
configurations upon which ordinary objects depend are perfectly precise.
Though boundary indeterminacy is most naturally treated in metaphysical
terms, many have found metaphysical indeterminacy problematic, as incoher-
ent (Evans 1983), “not properly intelligible” (Dummett 1975), or underspecified
(Lewis 1986). Lately, however, two specific accounts of metaphysical indetermi-
nacy have been advanced, each of which is claimed to be coherent and intelligible.
In what follows, I’ll first argue that one of these—the account initially presented
⁹ Various considerations call into question the common assumption that fundamental physical
entities, either individuals or micro-configurations, have perfectly precise boundaries. Perhaps the
most salient consideration concerns the quantum value indeterminacy associated with position and
momentum as incompatible observables, which pushes against taking lower-level entities or configu-
rations to have precise locations, on pain of an infinite variance in momentum. See Calosi and Wilson
2018 for discussion.
and defended in Wilson 2013a—is better than the other, so far as treating the
indeterminacy of ordinary object boundaries is concerned. I’ll then show that
the specific preferred treatment entails that ordinary objects are at least Weakly
emergent, by lights of a determinable-based account of Weak emergence.
‘Meta-level’ and ‘object-level’ accounts of metaphysical indeterminacy

Here I very briefly present the two main approaches to metaphysical indetermi-
nacy, highlighting certain of their differences along the way; see Wilson 2013a
and 2016a and Calosi and Wilson 2019 and forthcoming for more detailed
comparisons.
At a general level, the two approaches differ structurally as regards where
worldly indeterminacy is supposed to be located. On the metaphysical superval-
uationist approach endorsed in Akiba 2004, Barnes 2006, Barnes and Williams
2011, and elsewhere, what it is for a given state of affairs—e.g., an object’s having
a property—to be metaphysically indeterminate is for it to be metaphysically
indeterminate which precise state of affairs, out of some range of candidate precise
states of affairs, is actually the case. Here the operating assumption is that all
states of affairs are perfectly precise, and worldly indeterminacy consists in the
world’s being unsettled about which precise state of affairs is in fact the case;
indeterminacy is thus structurally located at the ‘meta-level’, so to speak, as inde-
terminacy among precise options. On the contrasting approach to metaphysical
indeterminacy presented and endorsed in Wilson 2013a (see also Bokulich 2014,
Wilson 2016a, and Calosi and Wilson 2019 and forthcoming), what it is for a
given state of affairs to be metaphysically indeterminate is for the constitutive
object (more generally, entity) of the state of affairs to have an irreducibly
indeterminate property (more generally, feature), where this is understood, in
turn, in terms of the constitutive object’s having a determinable property, but no
unique determinate of that property. Such a failure of unique determination can
happen in two ways: first, if many candidate determinates of the determinable are
instantiated, in such a way that no one determinate is non-arbitrarily taken to be
the ‘unique’ determinate of the determinable instance (this is ‘glutty’ indetermi-
nacy); second, if no determinate of the determinable is instantiated (this is ‘gappy’
indeterminacy). Here the operating assumption is that some states of affairs are
irreducibly imprecise, and worldly indeterminacy consists in a certain pattern of
determinables and determinates (namely, an object’s having a determinable but
no unique determinate of that determinable); hence indeterminacy is structurally
located at the ‘object-level’, so to speak, in states of affairs themselves.
In a bit more detail: on a meta-level supervaluationist approach, what it is
for the world to be ‘unsettled’ between various precise options is a primitive
matter; proponents aim, however, to render primitive metaphysical indeterminacy
intelligible by offering a logic and semantics of such indeterminacy along lines
of the logic and semantics offered on a supervaluationist approach to semantic

indeterminacy. On a semantic supervaluationist approach, indeterminacy reflects
our language’s failing to determine a precise extension for certain expressions (e.g.,
‘bald’), rendering the truth of certain claims (e.g., ‘Bob is bald’) indeterminate;
this failure is then modeled as entailing that precisifications of our language, each
compatible with existing (determinate) use, disagree about the extension of the
relevant term(s), such that, e.g., according to one linguistic precisification ‘Bob is
bald’ is true, whereas according to another linguistic precisification ‘Bob is bald’
is false. On the metaphysical version of a supervaluationist approach, it is the
world rather than language that is unsettled about whether a given state of affairs
obtains, rendering the truth of certain claims (e.g., ‘Mount Everest has B’, for some
precise boundary B) indeterminate; this failure is then modeled as entailing that
‘precisificationally possible worlds’, each compatible with existing (determinate)
facts, disagree on whether the state of affairs obtains, such that, e.g., according
to one precisificationally possible world ‘Mount Everest has B’ is true, whereas
according to another precisificationally possible world, ‘Mount Everest has B’ is
false. Note that on both semantic and metaphysical supervaluationist accounts,
meta-level indeterminacy gives rise to sentential or propositional indeterminacy,
and so to the need to introduce an indeterminacy operator into the semantics.
As for an object-level determinable-based approach to metaphysical indetermi-
nacy, the account I endorse (Wilson 2013a) is more specifically as follows:
Determinable-based MI: What it is for a state of affairs to be metaphysically

indeterminate in a given respect R at a time t is for the state of affairs to
constitutively involve an object (more generally, entity) O such that (i) O has a
determinable feature P at t, and (ii) for some level L of determination of P, O does
not have a unique level-L determinate of P at t. (366)
One might wonder: how can it be that a determinable property is not uniquely
determined? It has standardly been assumed (e.g., by Funkhouser 2006, among
many others) that it is a requisite feature of determinables that, for every level
of specification L at which they may be determined, a determinable instance at a
time has one and only one determinate at that time. However, as I argue in Wilson
2013a and 2016a (see also Wilson 2017 for historical discussion), the standard
supposition that determinables must be uniquely determined reflects an overly
quick generalization from certain paradigm cases of determinable and determinate
instances, and should be rejected as generally characterizing determinables and
determinates.
Indeed, there is a case to be made that even instances of color, the feature most
often highlighted as illustrative of a determinable that can be further determined,
are not always uniquely determined. In particular, the case of an iridescent feather,
the color of which is relative to perspective (the feather is red from one perspective,
blue from another), is reasonably interpreted as one where the feather is colored
(has the determinable color) but is not any unique (one and only one) color (does
not have a unique determinate of color), primarily because any attribution of a
specific shade (e.g., red rather than blue) as its purported unique color would be
inappropriately arbitrary. At best, one can say that the feather is red relative to
one perspective (more generally, circumstance), and is blue relative to another.
This and other cases of what I call ‘multiple relativized determination’ provide
good reason for thinking that the conditions in Determinable-based MI can be
satisfied—in the case of a feather, in ‘glutty’ fashion (again, adverting to there being
too many candidate determinate instances on the scene).1⁰
There is, of course, much more to be said about each approach, and I will say a
bit more down the line; but this sketch will be enough to proceed.
Application to the case of ordinary object boundaries

I now want to apply the two accounts to a candidate case of indeterminacy in
an ordinary object boundary, and argue that a determinable-based account better
treats the case.
On a supervaluationist account, such indeterminacy is treated as follows:
Supervaluationist MI (ordinary object boundaries): What it is for an ordinary

object O to have an indeterminate boundary at a time t is for it to be primitively
metaphysically indeterminate at t which precise boundary O has (i.e., for it to be
primitively metaphysically indeterminate at t which precisificationally possible
world is actual).
Here, what it is for Mount Everest to have an indeterminate boundary at a given

time is for the world to be unsettled at that time about which precise boundary
Mount Everest has—or, in terms of precisificationally possible worlds, for it to be
primitively indeterminate which such possible world is actual. Notwithstanding
that it is unsettled at t which precise boundary Mount Everest has, it is nonetheless
settled (since ‘super-true’: true at every precisificationally possible world) that
Mount Everest has a precise boundary, and it is moreover settled (for the same
reason) that for every precise boundary B, either Mount Everest has B or Mount
Everest does not have B. And similarly for the indeterminate boundaries of other
natural and artifactual ordinary objects.
On a determinable-based account, such indeterminacy is treated as follows:
1⁰ Open future cases (see Wilson 2016a) and some interpretations of certain quantum phenomena
(see Bokulich 2014, Wolff 2015, and Calosi and Wilson 2019) provide motivation for taking the
conditions in Determinable-based MI to be satisfiable in ‘gappy’ fashion, whereby no determinates are
available to provide the unique determinate of the determinable.
Determinable-based MI (ordinary object boundaries): What it is for an ordinary

object O to have an indeterminate boundary at a time t is for O (i) to have a
determinable boundary property P but (ii) for some level L of determination of
P, not to have a unique level-L determinate of P at t.
For example, on a determinable-based account, what it is for Mount Everest to

have an indeterminate boundary at a given time is for Mount Everest to have a
determinable boundary property at that time, but not to have a unique determinate
of that boundary property at that time. What it is for a table (call it ‘Woody’) to
have an indeterminate boundary at a time is for Woody to have a determinable
boundary property at that time, but no unique determinate of that property at
that time. And similarly for other natural and artifactual ordinary objects.
Why think that mountains, tables, and other ordinary objects can have a
determinable boundary property at a time, but no unique determinate boundary
property at that time? And given that such ordinary objects satisfy the conditions
in Determinable-based MI, is the indeterminacy at issue glutty or gappy? On
the view I endorse, the indeterminacy of ordinary object boundaries (at a time;
henceforth I’ll leave this qualification implicit) is plausibly seen as reflecting that
there are multiple candidate determinate boundary properties, associated with
different and typically overlapping micro-configurations (pluralities or structural
aggregates) in the vicinity of the ordinary object’s boundary, such that it would
be inappropriately arbitrary to pick out just one of these determinate boundary
properties as being the unique determinate such boundary had by the ordinary
object. This sort of case, involving multiple relativized determination, involves
glutty, not gappy, indeterminacy.
Consider Mount Everest. This mountain has the determinable property of
having a boundary around here (gesturing at the general vicinity of the mountain).
Now, does it make sense to take Mount Everest to have some unique maximal
determinate of this determinable boundary property? It’s hard to see how. If Mount
Everest were to have a unique maximally determinate boundary property, it would
presumably be inherited, either by way of identity or realization, from the (pre-
sumed) maximally determinate boundary property of some micro-configuration
in its vicinity. But at any given time, there are multiple micro-configurations in
the vicinity of Mount Everest. These overlap as regards the clear interior regions
of Mount Everest, but differ in their spatial extent: some have more dirt, some
less, some are bigger, some smaller. Which micro-configuration is ‘the’ micro-
configuration whose (presumed maximally determinate) boundary is possessed
by Mount Everest?
There is not, it seems, any good answer to this question except one according to
which the question is ill-formed. For none of the candidate micro-configurations
has any more metaphysical claim to be considered ‘the’ dependence base of Mount
Everest than any other. Correspondingly, taking Mount Everest to have, as its
purported unique maximally determinate boundary, the boundary of any one of
these micro-configurations would be inappropriately arbitrary.11
Summing up: Mount Everest has a determinable boundary property, but (owing
to the presence of too many candidate maximal determinates of this deter-
minable) does not have a unique maximal determinate of this property; hence
the state of affairs of Mount Everest’s having a boundary satisfies the conditions
in Determinable-based MI.12 And similarly for the indeterminate boundaries of
other natural and artifactual objects.
Now, in previous work I’ve highlighted several advantages of a determinable-
based approach to metaphysical indeterminacy over a metaphysical supervalu-
ationist approach, including that while a supervaluationist account introduces
sentential or propositional indeterminacy, and so requires some revisions to both
classical logic and semantics (see Williamson 1994, Ch. 6), a determinable-based
account does not introduce sentential or propositional indeterminacy, and so
requires no such revisions.13 Here I want to highlight two important advantages
specifically applying to the case at hand.
The first advantage is that a supervaluationist treatment of object boundary
indeterminacy is, unlike a determinable-based approach, not clearly intelligi-
ble. Recall that on a supervaluationist approach, metaphysical indeterminacy is
11 This observation is along lines of what Unger (1980) evocatively calls ‘the problem of the many’,
according to which attempts to identify any given macro-object (in his preferred case: Tibbles the cat)
with some lower-level configuration runs into the problem that there are many equally good candidates,
at a given time, for being the lower-level dependence base entity, such that the identification of the
macro-object with any one of these candidates would be inappropriately arbitrary. The problem of
the many can also be seen as an interesting variation on cases of multiple realizability, where a given
feature that is in fact realized by a given micro-feature might possibly be alternatively realized by a
different micro-feature. In the case of Mount Everest, the multiplicity of realization of the determinable
boundary property is actual, not merely possible; and as in Unger’s problem, any attempt to pick out
one of the micro-configurations in the vicinity of Mount Everest and attribute its maximally precise
boundary to Mount Everest as the unique determinate of the mountain’s determinable boundary
property would be inappropriately arbitrary.
12 Again, the satisfaction of the conditions in Determinable-based MI in the cases at hand is
indicative of glutty rather than gappy indeterminacy. Like the color of an iridescent feather, the
determinable boundary of Mount Everest admits, at a time, of multiple determinations, each associated
with one of the micro-configurations in the vicinity. As with the case of the feather, some may see
it as appropriate to attribute the precise boundaries to the mountain, albeit in relativized fashion;
others may maintain that even relativized attributions of precise boundaries are inappropriate unto
characterizing ordinary objects. This is a choice point; but either way, in cases of such multiple
relativized determination, it is not an option to attribute a unique, or indeed any, precise boundary
to the mountain (more generally, ordinary object) in unrelativized fashion.
13 Consider our test case, involving Mount Everest’s indeterminate boundary. On this treatment, it
will be true that Mount Everest has a boundary, and false that Mount Everest has a perfectly precise
boundary. It will be false, for every unrelativized attribution of such a precise boundary, that Mount
Everest has that boundary. It may or may not be true, depending on the choice point mentioned in
note 12 and on the specific facts at hand, that Mount Everest has such a precise boundary in relativized
fashion. And so on. Similarly, all the usual theorems of classical logic and semantics hold.
primitive, but is supposed to be rendered intelligible by analogy to semantic super-

valuationism. But at least for the case at hand, a metaphysical supervaluationist
account does not naturally inherit the intelligibility of its semantic precursor; for
though it is clear enough what it would be for our language to be unsettled between
various ways of linguistically precisifying a vague term such as ‘bald’, it is less clear
what it would be for the world itself to be unsettled, at a given time t, between
various precise ways it might be at t—especially given that (as we have seen) the
logic of precisificational possibilities entails that it is settled at t that the world is one
of those precisificationally precise ways. As applied to the case of Mount Everest:
how could it be that the world is settled, at t, that Mount Everest has a precise
boundary, but is unsettled, at t, about which boundary that is? Conversely, if the
world is unsettled at t about which precise boundary Mount Everest has, how could
it be settled at t that it has a precise boundary?1⁴ By way of contrast, a determinable-
based account of object boundary indeterminacy is clearly intelligible; for on this
view indeterminacy is not primitive, but is rather reducible to a certain pattern
of features of the sort with which we are already experientially and theoretically
familiar, and which pattern is clearly intelligibly modeled (by, e.g., such mundane
cases as an iridescent feather).1⁵
The second advantage is that a supervaluationist account incorrectly presup-
poses that it would be appropriate, in principle, to assign a perfectly precise
boundary to an ordinary object such as Mount Everest, in unrelativized fash-
ion. Relatedly, on a metaphysical supervaluationist account, a claim such as
‘Mount Everest has a unique, maximally determinate boundary’ comes out true
(since ‘super-true’: true in every precisificationally possible world). A metaphysical
supervaluationist approach thus fails, like semantic and epistemic approaches to
boundary indeterminacy, to accommodate intuitions according to which it is part
of what it is to be Mount Everest, or pretty much any other ordinary object, to
not have a precise boundary. By way of contrast, a determinable-based account
accommodates this intuitive central feature of ordinary objects.
I take it that, even setting aside the general advantages of a determinable-based
account, these considerations provide good reason to prefer a determinable-based
treatment of ordinary object boundary indeterminacy.
1⁴ Here it is worth comparing a better case for the supervaluationist—the case of the open future. I
can see how the world might be settled at t that, at some future time t′ , either a sea battle will be going on
or a sea battle won’t be going on, but unsettled at t about which of these options will be the case at t′ . The
difficulty attaching to the case of ordinary object boundary indeterminacy lies in understanding how
to apply this pattern of worldly settledness and unsettledness to a case of present-tense indeterminacy.
1⁵ In reductively analyzing metaphysical indeterminacy using off-the-shelf resources, a
determinable-based account is also more parsimonious than a metaphysical supervaluationist
account; but since (as per Ch. 1, §1.4.5) I don’t place much weight on parsimony considerations, I
won’t press this advantage.
The connection to Weak emergence

The previous result provides a basis for taking the vast store of ordinary objects,
which typically have indeterminate boundaries, to be at least Weakly emergent,
that proceeds by looking more closely at the analogy between multiple determi-
nation and multiple realization, and considering its implications. As above, on a
determinable-based treatment, the metaphysical indeterminacy in ordinary object
boundaries is of the glutty variety: the determinable feature is associated with mul-
tiple candidate determinates of the determinable at a given time, each associated
with a micro-configuration in the vicinity, such that the determinate boundary
properties can be had by the ordinary object, at best, in relativized fashion. This
phenomenon of multiple relativized determination echoes the phenomenon of
multiple realizability, with the main difference being that in cases of multiple
determination the multiplicity is actual, rather than merely possible.1⁶ Now, as
previously discussed, while there are reductive strategies for accommodating cases
of multiple realizability, attention to the relations between the powers associated
with a multiply realized feature and the powers associated with any one of its
realizing features suggests that reductive strategies fail, since multiply realized
types have fewer powers than each of their realizing types (reflecting effects that
the latter can cause that depend on differences between the realizing types), and
this proper subset relation between the powers of the realized and realizing types
will be inherited by instances of the types, satisfying the Proper Subset of Powers
Condition and blocking reduction at both the type and token levels.
Here I want to observe that in cases where a given determinable admits of
multiple (relativized) determination, the determinable type will also be reasonably
taken to have fewer powers than each of the more determinate types, again
reflecting effects that the latter types can cause that depend on differences between
determinate types, at a given level of specification. For already-rehearsed reasons,
this proper subset relation as holding between powers of determinable and deter-
minate types will also hold between tokens of those types, hence satisfying the
Proper Subset of Powers Condition. That the multiple determinates are all actually
instantiated (albeit in relativized fashion) in the case of the determinable boundary
properties of ordinary objects, as opposed to being merely possibly instantiated
(in non-relativized fashion) in typical cases of multiple realization, doesn’t affect
the relation between powers of the higher-level feature and lower-level feature at
issue. Hence given that the determinable boundary properties of ordinary objects
satisfy the conditions in Determinable-based MI in glutty fashion (as involving
multiply relativized determination), and coupled with associated satisfaction of
the cotemporal material dependence condition, it will follow that the determinable
1⁶ Even this difference is superficial, in that it has been observed that certain features might be
actually multiply realized at a time.
boundary property of an ordinary object satisfies the conditions in a determinable-

based implementation of Weak Emergence, such that the ordinary object at issue
will be at least Weakly emergent.
Adding further support to this result is that the connection between realization
and the determinable/determinate relation has in past decades been recognized
and appealed to (by Macdonald and Macdonald 1986, Yablo 1992, Wilson 1999
and 2009, Shoemaker 2000/2001, and others) as providing the basis for an account
of realization (in particular, one allowing that determinable and associated deter-
minate instances may be possessed by different objects).1⁷ Considerations of how
best to treat the metaphysical indeterminacy of ordinary object boundaries thus
provide independent support for thinking that ordinary objects are at least Weakly
emergent, by lights of a determinable-based account of such emergence.
6.2 Are ordinary objects Strongly emergent?
The considerations of the previous section suggest that many ordinary objects,
of both natural and artifactual varieties, are at least Weakly emergent. Might any
ordinary objects be, moreover, Strongly emergent? While the Strong emergence
of some natural ordinary objects remains an empirical possibility, the best case
for Strong emergence here attaches to the case of artifacts. In the remainder of
this section I’ll briefly lay out this case. I will not (in this chapter) assess this case,
since as we will see, whether it goes through ultimately hinges on the status as
Strongly emergent, or not, of conscious mentality—a topic to be discussed in the
next chapter.
6.2.1 Two routes to the Strong emergence of artifacts
Recall that for an object to be Strongly emergent, it must have at least one feature
associated with a fundamentally novel power—a power not token-identical to any
power of its base feature on a given occasion. Yet as above, philosophers commonly
assume that every power of (every feature of) an artifact, on a given occasion, is
uncontroversially a power of (some feature of) the physically acceptable micro-
configuration upon which it depends, on that occasion.
1⁷ Most objections to determinable-based accounts of realization have focused on whether such

realization would be suited for accommodating mental-physical realization. I return to this issue in
Ch. 7 (§7.2.2).
One might think that this is a reasonable assumption. After all, while it might
remain at least an open question whether mental features and their bearers (e.g.,
persons) have fundamentally novel powers, associated (as on the British Emer-
gentist view, as well as my own interaction-relative version of Strong emergence)
with nonphysical fundamental interactions, it’s considerably less plausible that any
such fundamental interactions enter into determining the existence or behaviors
of baseballs, tables, and the like.
This is too quick, however. For artifacts are associated with functional roles in
which mentality is deeply implicated, in ways that, were the associated mental
features to be Strongly emergent, might open the door to artifacts’ having fun-
damentally novel powers.
Mentality potentially enters into the functional roles of artifacts in two ways:
first, mentality enters into the association of a given functional role with a given
artifact; second, mentality is, one way or another, typically to some extent constitu-
tive of the role in question. Hence to be a baseball is to be an object that enters into a
highly complex set of human practices, ranging from the playing of the game itself
and all that mentally entails, to the emotional and even ethical appreciation that is a
concomitant of the game, its players, and its values. To be a statue is to be an object
that enters into a highly complex set of human practices, ranging from the creation
and exhibition of said object, to the perceptual and aesthetic, not to mention
economic, interactions that are some part of these practices. And so on. Given
that mental features are typically implicated in the characterization of artifacts as
having specific functional roles as well as being to some extent constitutive of the
roles themselves, we can ask the question: supposing that the mental features in
question turned out to be Strongly emergent, would it reasonably follow that the
associated artifacts were also Strongly emergent?
In general, it seems to me, the bare involvement of a given mental state in either
of these aspects might not be enough to render the associated artifact Strongly
emergent. For example, were someone with a Strongly emergent mental state to
stipulate that a given rock is to be deemed a ‘Faller’—an artifact whose role is to
fall when dropped in accord with the laws of nature, their doing so would probably
not be seen as sufficient unto bestowing a fundamentally novel power on the Faller.
Still, for other functional roles, and contingent on the status as Strongly emergent
of normative, intentional, and/or perceptual features, it seems plausible to think
that an intentional act of stipulation of a given functional role could reasonably
be seen as bestowing upon the artifact in question one or more fundamentally
novel powers, as an extension of the idea that the presence of a novel fundamental
interaction can serve to ground novel powers. For example, contingent on the
status of the associated mental act, perhaps Andy Warhol’s stipulation that a given
Brillo box was to be considered an artwork suffices to bestow certain Strongly
emergent powers on the box—to produce certain aesthetic responses, to fetch a

high price at auction, and so on.
The latter considerations indicate that we can say this much: insofar as various
forms of mentality typically enter into both constructing and constituting the
functional roles associated with artifacts, and insofar as artifacts are plausibly
taken to be to some important extent characterized by the contributions of such
mental features, then any features or associated powers of artifacts which are
constituted or caused to exist by Strongly emergent mental features would also
be appropriately considered Strongly emergent.
It is worth connecting this result to Merricks’s conclusions in his (2003). There
Merricks argues that while ordinary objects, such as baseballs, do not exist, persons
do exist—precisely because, he supposes, persons (in particular, certain of their
mental features) have fundamentally novel powers relative to the powers had
by their dependence base goings-on. In other words, for Merricks, persons are
Strongly emergent. The present discussion shows that there is a potential tension
in Merricks’s view, at least for the artifacts that are the focus of his discussion;
for supposing that the existence and features of artifacts crucially depend on the
existence and mental features of persons, the Strong emergence of persons and
associated mental features might well be inherited by any artifacts that they create.
And in that case, Merricks would be committed to the existence of baseballs and
other ordinary objects, after all.
Let’s sum up the results of this chapter. I’ve argued that there are three different
cases to be made for the claim that some ordinary objects are Weakly emergent
or at least Weakly emergent. First, classical objects—special-science objects of the
sort appropriately treated by classical mechanics—are arguably Weakly emergent,
by lights of a DOF-based account, thanks to the elimination of quantum DOF in
the classical limit. Second, artifacts are arguably at least Weakly emergent, by lights
of a functional realization account, thanks to the support for such a treatment that
can be extracted from the compositionally flexible persistence conditions typically
encoded in the sortal features of artifacts. Third, both natural and artifactual
ordinary objects are arguably at least Weakly emergent, by lights of a determinable-
based account, thanks to the metaphysical indeterminacy of the boundaries of
ordinary objects, which indeterminacy is best treated in determinable-based
terms.
Moreover, the possibility remains, especially for artifacts, that these are ulti-
mately not just Weakly but Strongly emergent, owing to the role mentality plays
both in the specification and the constitution of the functional roles which are
typically associated with artifacts. As such, the status of any artifacts as Strongly
emergent arguably ultimately depends on the status of conscious mentality, to be
explored in the next chapter. These considerations more generally suggest that
those who are committed to the Strong emergence of persons and certain of their
conscious mental features might well find themselves with considerably more
entities in their ontology than they are on record as accepting.
Before moving on, I pause to note one other implication of these results. In her
(2010), Thomasson discusses a number of metaphysical concerns directed at the
supposition that ordinary objects exist, and argues that these concerns arise from
inappropriately applying broadly ‘scientistic’ methodology in service of answering
philosophical questions:
Although the arguments against the existence of ordinary objects do not rely on
any particular piece of scientific knowledge, many of them do rely on a certain
scientistic approach to metaphysics: the view that metaphysics is of a piece with
(and indeed part of the same total theoretical enterprise as) natural science. […]
Lying behind many of the arguments against ordinary objects is the assumption
that metaphysics is engaged in explanatory theory construction following the
same principles as those governing the natural sciences—or (perhaps better)
engaged in one and the same enterprise of constructing the best ‘total theory’.
(596)
Thomasson takes such an assimilation of scientific methodology to the metaphys-

ical enterprise to be problematic, as giving rise, for example, to concerns about
causal overdetermination:
Straightforward causal redundancy arguments rely on accepting the Eleatic

Principle: that we should only accept the existence of those entities that ‘make
a causal difference’. This may be a reasonable principle for deciding whether or
not we should accept the existence of neutrons as well as protons and electrons—
theoretical particles posited to explain experimental data. But whether or not it
carries over to the question of whether we should ‘posit’ tables ‘as well as’ the
particles that make them up is less clear. (596)
Thomasson suggests that this sort of concern is effectively a spandrel of a mistaken

assimilation of scientific to metaphysical methodology. Correspondingly, she
maintains, the correct response to this and other concerns about ordinary objects
is to reject this methodological assimilation, and the associated assumption that
questions pertaining to the existence and nature of ordinary objects should be
treated in ways consonant with those questions as pertaining to special-science

entities.
How to respond? To start, one might think that investigations in the meta-
physics of science are not really ‘scientistic’, at least in that the methodology of
metaphysicians of science is not empirical per se. One might also observe that it
would be surprising if the metaphysical treatment of special-science entities were
required to be different from the metaphysical treatment of ordinary inanimate
objects, since after all, many ordinary objects are also (e.g., classical mechanical)
special-science entities.
In any case, the results of this chapter show that there is no in-principle reason
to think that investigations into the existential (more generally, metaphysical)
status of ordinary objects should proceed differently from investigations into
the existential (metaphysical) status of special-scientific entities. Thomasson’s
suggestion to the contrary seems to have been primarily motivated by thinking,
first, that the usually stated concerns with ordinary objects (e.g., Kim-style causal
overdetermination concerns) arise from trying to give scientific and ordinary
objects (including artifacts) a unified treatment, and second, taking the concerns
as attaching to scientific goings-on not to admit of any good answers. But as I
have argued, there are good responses to the concerns at issue, whether natural or
artifactual ordinary objects are at issue. Nothing stands in the way of a systematic
treatment of natural and artifactual ordinary objects as at least Weakly emergent,
and contingent upon future empirical results and the import of mentality to be
next considered, perhaps even Strongly emergent.
7
Consciousness
I turn now to considering whether consciousness or conscious experience of the

sort that we and other creatures enjoy is either Weakly or Strongly emergent.
There are, to be sure, many forms or species of consciousness, including perceptual
awareness of the external world, conscious awareness of internal states (e.g.,
pain), and self-consciousness—consciousness of ourselves as conscious beings. As
Chalmers (1996) notes,
Conscious experiences range from vivid color sensations to experiences of the

faintest background aromas; from hard-edged pains to the elusive experience of
thoughts on the tip of one’s tongue; from mundane sounds and smells to the
encompassing grandeur of musical experience; from the triviality of a nagging
itch to the weight of a deep existential angst; from the specificity of the taste
of peppermint to the generality of one’s experience of selfhood. All these have
a distinct experienced quality. All are prominent parts of the inner life of the
mind. (4)
Notwithstanding this diversity, little in this chapter hinges on differences between

these forms of consciousness; so unless some specific variety is under discussion,
I will speak generically of consciousness or conscious awareness (or associated
mental features), which may have as its seeming object the external world, one’s
internal states, or (as a special case of the latter) consciousness itself.1
Interestingly, and in some contrast to the cases of complex systems and ordinary
objects considered in previous chapters, the cases proferred for the emergence
of consciousness are most frequently aimed at showing that consciousness is
Strongly rather than merely Weakly emergent. Hence Chalmers (2006a) says,
“there is exactly one clear case of a strongly emergent phenomenon, and that is
the phenomenon of consciousness” (3). In turn, it is often suggested that the main
1 One respect in which the discussion to follow is not neutral is in supposing that conscious mental
states are an apt subject of metaphysical attention, as potentially (in the theoretical sense) reducible to,
Weakly emergent from, or Strongly emergent from lower-level physical states. See Hellie 2019a, 2019b,
and forthcoming for development of a competing view on which this supposition is incorrect, and
consciousness is rather given an expressivist treatment. Hellie’s work challenges certain foundational
suppositions of the debates addressed in this chapter, and a full engagement with these challenges must
await a different occasion; what follows should accordingly be seen as conditional on there good being
responses to these challenges.
DOI: 10.1093/oso/9780198823742.003.0007
consciousness 215
motivation for taking consciousness to be Strongly emergent reflects a commonly

acknowledged failure of consciousness to be predictable from or explainable in
terms of lower-level physical phenomena. Hence Broad (1925) says:
[An archangel] would know exactly what the microscopic structure of ammonia
must be; but he would be totally unable to predict that a substance with this
structure must smell as ammonia does when it gets into the human nose. The
utmost that he could predict on this subject would be that certain changes would
take place in the mucous membrane, the olfactory nerves and so on. But he could
not possibly know that these changes would be accompanied by the appearance
of a smell in general or of the peculiar smell of ammonia in particular, unless
someone told him so or he had smelled it for himself. (71)
More recently, Bedau (2010) says:
Our inability to have found any plausible micro-level explanation for conscious
mental states might reflect just our ignorance, but another possibility is that these
phenomena really are strongly emergent. (60, note 3)
Indeed, existing arguments offered as supporting the Strong emergence of con-

sciousness nearly all rely, one way or another, upon the supposition that conscious-
ness, or certain of its characteristic features or aspects,2 lies beyond the explanatory
reach of any lower-level physical goings-on.
As the reader may suspect, the presence of even an insuperable or ‘in-principle’
explanatory gap can’t be the whole supporting story, however, since—to take
one previously discussed example—many complex nonlinear phenomena are
clearly physically acceptable, in spite of having features that are, in some relevant
sense of ‘insuperability’ (as, e.g., beyond the access of any empirically possible
determination), insuperably unpredictable or otherwise unexplainable in lower-
level physical terms. What’s moreover at issue in discussions of consciousness is
that the explanatory gaps are taken to be metaphysically significant, in reflecting
not just broadly mathematical barriers to explanation, such as nonlinearity and
associated sensitivity to initial conditions, but rather that certain characteristic
features of consciousness—in particular, subjective or qualitative aspects of con-
scious experience—depart so greatly from lower-level physical features that this
divergence provides reasonable grounds for thinking that no physicalist account
2 Talk of characteristic features (or aspects) of consciousness or conscious states should be under-
stood as broadly neutral on exactly how such features enter into such states. In particular, such talk
isn’t intended to suggest that characteristic features of conscious states are second-order features of
features.
of consciousness of either reductive or nonreductive (Weakly emergent) varieties

could possibly be correct.3
In being so motivated, explanatory gap arguments in favour of the Strong
emergence of consciousness deserve independent consideration. As I will argue,
however, these arguments are ultimately uncompelling for reasons not much
previously explored, notwithstanding the considerable critical attention that has
been given to these arguments. More generally, I will argue that while it remains
an open empirical possibility that consciousness is Strongly emergent, at present
we have no compelling philosophical or empirical motivation for taking this to
actually be so.⁴
On the assumption that consciousness isn’t (doesn’t turn out to be) Strongly
emergent, might it be Weakly emergent rather than ontologically reducible? Here
I again present an underexplored reason to endorse a positive answer to this
question. The argument proceeds via the claim that qualitative conscious states—
e.g., states of conscious awareness of colors or pains—are typically determinable
rather than (maximally) determinate, in a way that defensibly renders them
suitable (again, assuming that they are not Strongly emergent) for being realized
in determinable-based fashion, and hence Weakly emergent.
The overall conclusion is that there are reasons to think that consciousness is at
least Weakly emergent from lower-level physical goings-on.
3 An early expression of this line of thought can be found in Ewing’s (1951) reasons for rejecting
the identity theory:
Nineteenth-century materialists were […] inclined to identify thinking, and mental events
generally, with processes in the central nervous system or brain. In order to refute such views
I shall suggest your trying an experiment. Heat a piece of iron red-hot, then put your hand
on it, and note carefully what you feel. You will have no difficulty in observing that it is quite
different from anything which a physiologist could observe, whether he considered your
outward behaviour or your brain processes. The throb of pain experienced will not be in the
least like […] anything described in textbooks of physiology as happening in the nervous
system or brain. I do not say that it does not happen in the brain, but it is quite distinct
from anything that other people could observe if they looked into your brain. […] We know
by experience what feeling pain is like and we know by experience what the physiological
reactions to it are, and the two are totally unlike. […] the difference is as plainly marked and
as much an empirical matter as that between a sight and a sound. The physiological and the
mental characteristics may conceivably belong to the same substance […] but at least they
are different in qualities, indeed as different in kind as any two sets of qualities. (101–2)
⁴ This result applies to the qualitative or subjective conscious features that are the usual focus of
discussions of the ontological status of consciousness or conscious experience. As we’ll see in the next
chapter, I think there is a good existing case for taking libertarian free will to be Strongly emergent;
hence if exercises of such free will involve consciousness, then the result here should be accordingly
understood as restricted.
consciousness 217
7.1 Is consciousness Strongly emergent?
In this section, I consider two different forms of argument that have been or might
be offered in support of consciousness’s being Strongly emergent: first, ‘knowledge
arguments’ as presented by Nagel (1974) and Jackson 1982 and 1986; second, the
‘conceivability’ or ‘zombie’ argument as presented by Chalmers 2009. The qualifier
‘might be’ reflects that while proponents of these arguments take the conclusions
to undermine the physical acceptability of conscious mental states, they do not
thereby go on to conclude that consciousness is Strongly emergent.⁵ For present
purposes, what is important is that the considerations these authors raise can and
have been offered in support of taking consciousness to be physically unacceptable,
and hence (along with other premises) for taking consciousness to be Strongly
emergent, in the sense operative in the associated schema for emergence.
7.1.1 The knowledge arguments
The suggestion that even complete knowledge of physical goings-on might not
suffice for knowledge of certain aspects of conscious experience can be found
throughout history. Echoes of the basic idea can be found in Leibniz’s (1714) ‘Mill
argument’:
[W]e must confess that perception, and what depends upon it, is inexplicable
in terms of mechanical reasons, that is through shapes, size, and motions. If we
imagine a machine whose structure makes it think, sense, and have perceptions,
we could conceive it enlarged, keeping the same proportions, so that we could
enter into it, as one enters a mill. Assuming that, when inspecting its interior,
we will find only parts that push one another, and we will never find anything to
explain a perception. (§17; GP: VI, 609/AG, 215)
Leibniz concluded that perceptual consciousness should be understood as located

in a ‘simple substance’ rather than ‘the composite’. As above, the British Emer-
gentist Broad also offered a version of a knowledge argument (Broad 1925, 71),
involving a mathematical archangel who knows all the mechanistic facts about
chemical goings-on, in support of taking qualitative aspects of conscious
⁵ Indeed, Nagel (1979) elsewhere registers his supposition that there is no emergence of a physically
unacceptable variety (see van Cleve 1990 for discussion). Both Jackson and Chalmers appear to endorse
the possibility of an anti-physicalist form of emergence (albeit of an epiphenomenalist variety) as a
further concomitant of their arguments, though Chalmers moreover allows that the anti-physicalist
upshot might rather be a form of panpsychism.
experience to be Strongly emergent. More recently, both Nagel (1974) and Jackson
(1982 and 1986) have offered arguments aiming to show that one could have
complete physical knowledge of some entity or subject matter, but nonetheless
fail to know certain facts pertaining to conscious states associated with the
entity or subject matter; from this they conclude that physicalism, at least as
commonly understood, is false. Such an anti-physicalist conclusion, coupled
with the assumption that states of conscious awareness cotemporally materially
depend on lower-level physical states, would provide positive motivation for
consciousness’s being Strongly emergent.
Nagel’s discussion proceeds by attention to the question of what we ought to be
able to understand about the conscious experience of creatures relatively different
from ourselves.⁶ He first connects the notion of consciousness to what he calls “the
subjective character of experience”:
Conscious experience is a widespread phenomenon. It occurs at many levels of

animal life […] [N]o matter how the form may vary, the fact that an organism
has conscious experience at all means, basically, that there is something it is like
to be that organism. There may be further implications about the form of the
experience; there may even (though I doubt it) be implications about the behavior
of the organism. But fundamentally an organism has conscious mental states if
and only if there is something that it is like to be that organism—something it is
like for the organism. We may call this the subjective character of experience.⁷
(436)
Nagel goes on to more specifically consider the case of a bat—a creature plausibly
having conscious experience, but of a very different sort from any to which we
are privy—and to raise to salience difficulties in our being able to comprehend
what it is like to be that sort of creature. As he observes, there is no clear way
to extrapolate from our own perceptual experience to that of a creature who
navigates the world using echolocation. And nor, crucially, would any amount of
knowledge of the physiology of bats, whether pitched at a lower or at a higher level
of physical goings-on, enable us to gain knowledge of this comparatively alien form
of conscious experience. Nagel concludes that such knowledge is beyond our ken,
⁶ See Feigl 1959 for a historical antecedent of attention to this issue, involving a Martian in
possession of all the physical facts about humans, who would nonetheless “be lacking completely in
the sort of imagery and empathy which depends on familiarity (direct acquaintance) with the kinds of
qualia to be imaged or empathized” (431).
⁷ For anticipations of such ‘what it’s like’ talk, see also Farrell 1950, 185–8 and Feigl 1967, 139–40.
For recent discussion of the locution, see Hellie 2007.
consciousness 219
and goes on to suggest that the latter gap spells trouble for any view on which
consciousness (and its subjective character) are either ontologically reducible to
or realized in (that is, Weakly emergent from) lower-level physical goings-on:
While an account of the physical basis of mind must explain many things, this
appears to be the most difficult. […] if physicalism is to be defended, the phe-
nomenological features must themselves be given a physical account. But when
we examine their subjective character it seems that such a result is impossible.
The reason is that every subjective phenomenon is essentially connected with a
single point of view, and it seems inevitable that an objective, physical theory will
abandon that point of view. (437)
Nagel sees here a “divergence between […] two kinds of conception: subjective
and objective” (438). An account of consciousness as physically acceptable must
include a treatment of the subjective character of consciousness, but on the
assumption that physical theory is solely concerned with the objective or third-
person point of view, any such treatment will be, he maintains, “impossible”.
Jackson’s (1986) knowledge argument (for an early variation on the theme, see
Jackson 1982) also aims to establish that there is an insuperable, metaphysically
significant gap in relevantly comprehensive knowledge of certain physical goings-
on and knowledge of what it is like to have certain conscious experiences. Jackson’s
focus differs from Nagel’s in highlighting the gap as attaching to the qualitative
aspects of conscious experiences, as per the following thought experiment:
Mary is confined to a black-and-white room, is educated through black-and-

white books and through lectures relayed on black-and-white television. In this
way she learns everything there is to know about the physical nature of the
world. She knows all the physical facts about us and our environment, in a wide
sense of ‘physical’ which includes everything in completed physics, chemistry,
and neurophysiology, and all there is to know about the causal and relational facts
consequent upon all this, including of course functional roles. If physicalism is
true, she knows all there is to know. For to suppose otherwise is to suppose that
there is more to know than every physical fact, and that is just what physicalism
denies. […] It seems, however, that Mary does not know all there is to know.
For when she is let out of the black-and-white room or given a color television,
she will learn what it is like to see something red, say. This is rightly described
as learning—she will not say “ho, hum.” Hence, physicalism is false. This is the
knowledge argument against physicalism in one of its manifestations. (291)
As Jackson qualifies, at issue here is not that Mary comes to learn something about
her own experiences, but rather that Mary comes to learn something about the
experiences of others:
The trouble for physicalism is that, after Mary sees her first ripe tomato, she
will realize how impoverished her conception of the mental life of others has
been all along. She will realize that there was, all the time she was carrying out
her laborious investigations into the neurophysiologies of others and into the
functional roles of their internal states, something about these people she was
quite unaware of. (292, emphases in text)
How might a physicalist best respond to the knowledge arguments? In what

follows, I’ll focus on Jackson’s version; the application to Nagel’s argument and
other variations on the theme will be clear. I’ll moreover focus on the version of
Jackson’s argument presented in Nida-Rümelin 2015, which helpfully articulates
the key premises and conclusions, as follows:
P1: Mary has complete physical knowledge about human color vision before
her release.
C1: Therefore, Mary knows all the physical facts about human color vision
before her release. (P1)
P2: There is some (kind of) knowledge concerning facts about human color
vision that Mary does not have before her release.
C2: Therefore, there are some facts about human color vision that Mary does
not know before her release. (P2)
C3: There are nonphysical facts about human color vision. (C1, C2)
The knowledge arguments have generated an enormous amount of literature,

and a full treatment of all the variations on, ramifications of, and responses to
these arguments is beyond the scope of this chapter (see Nida-Rümelin 2015 for
extensive discussion and references). Here I’ll present my preferred strategy, which
is apparently not much on the books—namely, to deny P1. Along the way I’ll
positively contrast this approach with certain other more popular strategies.
The ‘Incomplete Physical Knowledge’ strategy

According to P1 of the knowledge argument, Mary has complete physical knowl-
edge about human color vision before her release. The strategy of denying P1
is not much considered; hence Nida-Rümelin passes it over, saying, “it seems
hard to deny that it is in-principle possible to have complete physical knowledge
about human colour vision (or about an appropriately chosen part thereof). If
so, premise P1 should be accepted as an appropriate description of a legitimate
consciousness 221
thought experiment”. (Note that Nida-Rümelin’s remarks here presuppose not just
that it is in-principle possible to have complete physical knowledge about human
color vision, but more strongly that Mary could have such knowledge prior to her
release from the black-and-white room.) A good case can be made for denying
P1, however, as per what I hereby dub the ‘Incomplete Physical Knowledge’
strategy.
To start, note that nearly all participants in the debate over the import of
the knowledge arguments take for granted that physical knowledge does not
include knowledge of subjective or qualitative aspects of consciousness. Hence
Nagel links physical knowledge to physical theory, and takes the latter to concern
only “objective” phenomena, such that it would be “impossible” for physics to
accommodate a subjective point of view, and Jackson motivates P1 on grounds
that “it is plausible that lectures over black-and-white television might in principle
tell Mary everything in the physicalist’s story. You do not need color television
to learn physics or functionalist psychology” (295). Opponents of the knowledge
argument similarly take for granted that complete physical knowledge fails to be
knowledge of any subjective or qualitative aspects of reality there might be. This is
true on the ‘Ability Hypothesis’, according to which what Mary gains upon leaving
her room is a new ability rather than a new piece of knowledge (see, e.g., Lewis
1990, Nemirow 2006). It is true on what Nida-Rümelin calls the ‘Complete Physical
Knowledge without Knowledge of all the Physical Facts’ strategy, which proceeds
by accepting P1 but denying the inference from P1 to C1 (according to which
Mary knows all the physical facts about human color vision before her release),
on grounds that since physical knowledge is not of subjective or qualitative facts,
and since some physical facts are subjective or qualitative, Mary’s having complete
physical knowledge does not entail her knowing all the physical facts (see, e.g.,
Harman 1990, Flanagan 1992, and Alter 1998). It is true on the ‘Acquaintance
Hypothesis’, which again grants that Mary has complete physical knowledge, and
moreover knows all the physical facts, prior to leaving her room, but maintains
that what she gains upon leaving the room is ‘knowledge by acquaintance’, which,
in not being ‘informational’, doesn’t threaten physicalism. And it is true on the
popular ‘two ways’ or what Nida-Rümelin calls the ‘New Knowledge; Old Fact’
approach, according to which, while Mary does gain new knowledge upon leaving
her room, this is simply knowledge about a different, qualitatively informed way
of thinking about a fact she already knew, such that the supposition in P1 that
prior to leaving her room Mary had complete physical knowledge, sufficient unto
knowing all the physical facts, is not undermined.
But need the physicalist agree that physical knowledge is ‘objective’ in the sense,
in particular, of failing to be of any subjective or qualitative aspects of reality there
may be? No, for two reasons.
First, such a view is in tension with physicalism. The ontological physicalist of

whatever stripe maintains that the lower-level physical goings-on, either individu-
ally or in appropriately complex combination, provide a basis for all of natural real-
ity. Given that natural reality includes consciousness and its associated subjective
and qualitative aspects, and given that the physicalist denies that ‘basic’ physical
goings-on either have or bestow mentality, the physicalist should correspondingly
maintain that some appropriately complex physical goings-on—namely, those that
are either identical with (on a reductive physicalist view) or which realize (on
a nonreductive physicalist view) these aspects of consciousness—constitutively
involve the instantiation of subjective and/or qualitative features which (again,
given the supposed truth of physicalism) are themselves physical or physically
acceptable features. The questions of whether these features should be seen as had
by the highly complex physical micro-configurations that accompany the instan-
tiation of the features, and of whether, if so, the associated micro-configurations
should be taken to have a first-person perspective, represent choice points: what
answers are given to these questions will depend on further details about the form
of physicalism at issue, as well as what account is given of the creatures having
such conscious states. However the physicalist answers these questions, they are
in any case committed to complex lower-level physical goings-on’s involving the
instantiation of lower-level physical features which, in being either identical to
or such as to realize subjective and qualitative aspects of natural reality, are
themselves properly deemed subjective and qualitative.
Now, complete physical knowledge is, as a matter of definition, complete
knowledge of any and all actual and/or potential physical goings-on, including
knowledge of the potential instantiation of any subjective and qualitative physical
features which suffice, either by way of identity (on a reductive view) or realization
(on a nonreductive view), for any subjective and qualitative aspects of conscious-
ness there might be.⁸ How is one to gain knowledge of subjective and qualitative
physical features? The physicalist can and should maintain, plausibly enough and
without any clear threat to their position, that such knowledge can only be gained
by acquaintance—that is, by being in position to experience these subjective and
qualitative features either directly (on a reductive physicalist view) or indirectly
(on a nonreductive physicalist view). Consequently, the physicalist can and should
deny (in particular) that Mary has complete physical knowledge about human
color vision before her release—that is, they can and should deny P1.
⁸ Nothing in this claim turns on whether knowledge has to be broadly propositional—i.e., of states
of affairs as opposed to properties or other features—since knowledge of (the potential instantiation)
of properties can be translated, if so desired, into propositional terms (as, e.g., knowledge that a certain
property would be instantiated under such-and-such circumstances).
consciousness 223
I will shortly offer a diagnosis of why, in spite of its being in clear tension
with the truth of physicalism, P1 has been taken for granted even by opponents
of the knowledge argument. First, though, it’s worth noting that the Incomplete
Physical Knowledge strategy improves on certain alternative strategies of response
to the knowledge argument, while arguably preserving the insight common to
many of these responses—namely, that the fact that certain kinds of experience
are prerequisite to grasping certain features of reality is no barrier to the phys-
ical acceptability of these features. Consider, for example, Conee’s (1994) case
for taking the Acquaintance hypothesis to block the physical unacceptability of
phenomenal qualities:
The physical facts may include every fact about qualia. Still, the physical story
does ‘leave out the qualia’, in the sense that knowledge of the physical facts does
not imply knowledge of the qualia. Gaining knowledge of phenomenal qualities,
though, is no more than a matter of making their acquaintance by attentive
experience. It requires only entering into a new cognitive relation to the qualities,
not learning any new information. It gives us no reason to doubt that everything
is physical. (148)
As Nida-Rümelin points out, a proponent of the knowledge argument is likely to

respond that if the having of an experience with the qualitative physical property
(Q) in question is required for knowledge of it by acquaintance, such acquaintance
is moreover “a necessary condition for being able to know (in the relevant sense)
that an experience has Q”. What has not been previously appreciated is that the
physicalist can and arguably should simply grant that acquaintance is a necessary
condition for knowing certain physical facts—namely, those providing a subjective
or qualitative basis for any subjective or qualitative aspects of consciousness there
may be. And similarly for the usual physicalist responses to the Ability hypothesis,
the ‘two-ways’ strategy, and the ‘Complete Physical Knowledge without Knowl-
edge of all the Physical Facts’ strategy. Rather than stand opposed to the intuitively
compelling take on Mary’s reaction to seeing a ripe tomato—‘so this is what it is like
to see red!’—according to which she does come to know a new fact after leaving her
room, the physicalist can rather simply agree that Mary gains new knowledge, and
then ‘modus tollens’ the anti-physicalist conclusion as rather showing that prior to
leaving her room, Mary didn’t have complete physical knowledge (of human color
vision, in particular), after all.
Why has this strategy of response been previously neglected? I diagnose this,
and the associated acceptance of P1 by opponents and proponents of the knowl-
edge argument alike, as reflecting a mistaken characterization of the physical
goings-on—one which is (a) overly representational, (b) overly restricted, and (c)
problematically qualitatively etiolated.
To start, recall that the motivation for P1 reflects the supposition that physics
does not contain reference to subjective or qualitative phenomena. Hence it was
that Nagel claimed that physical ‘theory’ concerns only ‘objective’ goings-on, and
that Jackson claimed that qualitative experience is not needed to learn physics.
The first problem with this motivation for P1 is that it characterizes the physical
basis in overly representational terms. Yes, physicalists look to physics as a guide to
the compositionally basic goings-on, but representation (not to mention theory) is
one thing, reality another. Recall the operative conception of the physical goings-
on, according to which these are the goings-on that are treated, approximately
accurately, by present or (in the limit of inquiry, ideal) physics (see Ch. 1, §1.4.1;
Wilson 2006). On this and other physics-based conceptions of the physical, physics
gives us a working handle on the compositionally basic goings-on, but even in the
limit of inquiry, there is no supposition that physical theory will be up to the task
of offering a complete description of these goings-on.
The second problem with this motivation is that (and even bracketing the
previous point) it presupposes an overly restricted characterization of the physical
goings-on. As discussed in Ch. 1 (§1.4.2) and elsewhere, the goings-on explicitly
referenced in physics provide the starting point, not the end point, of the lower-
level physical base. To be sure, the immediate targets of physical theory—e.g.,
subatomic particles or their field-theoretic or wave-functional correlates—do not
individually instantiate or otherwise encode subjective or qualitative aspects of
consciousness of the sort we enjoy. But that doesn’t show that sufficiently complex
physical goings-on do not do so, any more than the fact that the goings-on
explicitly referred to in physical theory do not individually instantiate or otherwise
encode thermodynamic processes and properties shows that sufficiently complex
physical goings-on do not do so.
A third problem might be thought to reflect a somewhat more substantive
motivation for P1—namely, that (so the story usually goes) insofar as physics
is concerned only with the trajectories and other behaviors of the goings-on
that are the immediate target of physical theory, and insofar as such behaviors
and associated laws are completely described in structural, non-qualitative terms,
no amount of complex or even causal combination could ever eventuate in the
instantiation of physical properties up to the task, so to speak, of either being or
serving as a physical basis for such qualitative aspects of consciousness. Structure,
one might say, can only beget more structure.
Now, this is a more substantive motivation for P1, but drawing upon the previ-
ous responses, the physicalist can reasonably maintain that it mischaracterizes the
physical goings-on, even at low orders of complexity, as completely qualitatively
etiolated.
consciousness 225
Such an abstract, completely structuralist conception of the compositionally

basic entities is not forced by physical theory. On the contrary: it is common
for physicists to speak of particles or other lower-level physical phenomena as
‘feeling’ forces, where the reference here might be thought to be both qualitative
and sensitive to perspective—the force coming from this direction rather than that,
and applied to the entity in its location in ways that then enter into its moving in
such-and-such a way. This degree of qualitativity, in which a noncomplex physical
entity is capable of registering a local perspective and associated sensitivity to the
environment, in a way that is not entirely qualitatively etiolated, clearly need not be
seen as involving any measure of consciousness or mentality, even of a ‘protopsy-
chic’ variety. On the contrary, it reflects a naturalistic point of view according to
which physical goings-on are real and substantial, and when they move, they do
so not because, e.g., they must play a predetermined part in the harmony of the
spheres, but rather for some salient reason or reasons—for instance, the influence
of one or more felt forces or interactions. This understanding of noncomplex phys-
ical goings-on suggests that we should distinguish qualitativity and consciousness:
sometimes these go together, as in creatures like us; but qualitativity is a weaker
notion, that may be present, e.g., in the response of a particle to the forces acting
on it, or more generally, in a particle’s engaging in interactions in ways reflecting
sensitivity to the relevant environmental circumstances.
To be sure, from recognition of this amount of qualitativity in the physical
goings-on it’s still a considerable leap to get to the instantiation of subjective or
qualitative aspects of consciousness. But given that some explanatory gaps (e.g.,
in the case of complex nonlinear systems) do not have anti-physicalist import, the
burden is now on the anti-physicalist to establish that the seeming explanatory gap
between physical goings-on and subjective and qualitative aspects of conscious
experience must be taken to have such import. The strategy of the knowledge
arguments crucially proceeds by way of P1 and the associated characterization
of complete physical knowledge, as failing to be knowledge of any subjective or
qualitative aspects of consciousness there may be. But the physicalist can and
should reject this characterization, both as effectively begging the question against
their view (since for the physicalist, complete physical knowledge must include
knowledge of those physical features that are identical with or serve as a subjective
or qualitative basis for such aspects of consciousness), but also on grounds that
there’s no independent reason to accept such a characterization.
I conclude that the knowledge arguments do not provide compelling reason to
think that consciousness and its associated subjective and qualitative aspects are
actually physically unacceptable, much less actually Strongly emergent.
7.1.2 The conceivability argument
I turn now to the conceivability argument advanced and developed by Chalmers

(1996, 1999, 2009, and elsewhere), according to which the conceivability of
zombies—creatures which are functional and physical duplicates of creatures
like us, but which are lacking in any conscious mentality—is taken, in
combination with certain other commitments, to establish the Strong emergence
of consciousness.
Like the knowledge arguments, the conceivability argument relies on the pres-
ence of an explanatory gap, as needed to make room for the conceiving in question.
As Levine (1983) describes the connection:
If there is nothing we can determine about C-fiber firing that explains why having
one’s C-fibers fire has the qualitative character that it does—or, to put it another
way, if what it’s particularly like to have one’s C-fibers fire is not explained, or
made intelligible, by understanding the physical or functional properties of C-
fiber firings—it immediately becomes imaginable that there be C-fiber firings
without the feeling of pain, and vice versa. We don’t have the corresponding
intuition in the case of heat and the motion of molecules—once we get clear
about the right way to characterize what we imagine—because whatever there
is to explain about heat is explained by its being the motion of molecules. (359)
Chalmers’s argument goes beyond the knowledge arguments, however. To start,

his argument relies on the conceivability of zombies, as opposed to the mere
absence of explanation or associated failures of knowledge. That said, an appeal to
conceivability alone isn’t much of an advance on the previous arguments, since just
as physicalists need not deny that there are explanatory gaps between conscious-
ness and lower-level physical goings-on, neither need physicalists deny that zom-
bies are conceivable, as long as they are not forced to take such a broadly epistemic
fact to have metaphysical import (in particular, of the anti-physicalist variety).
The primary advance of Chalmers’s argument rather lies in his situating the
conceivability of zombies in an independently motivated framework—‘epistemic
two-dimensionalism’, or E2D, to be explicated in more detail shortly—according to
which certain facts about meaning, which are taken to be a priori accessible, can be
used to identify or establish certain facts about modality, expressing or encoding
what is genuinely metaphysically possible (necessary, contingent, impossible). It is
commonly assumed that the mode of a priori access to meanings that enters into
the E2D strategy proceeds by way of conceiving. Consequently, commitment
to the E2D strategy for gaining access to modal truth, and to this strategy’s
consciousness 227
being implemented by means of a conceiving-based epistemology of meanings,

provides an independent basis for taking the conceivability of zombies to have
anti-physicalist metaphysical import, as a case in point of a systematic connection
between conceivability and metaphysical (as opposed to merely epistemic)
possibility.
Reflecting the crucial role that E2D plays in his argument, Chalmers has come
to call his argument, more accurately (and again, for reasons that will be made clear
shortly), ‘the two-dimensionalist argument against materialism’. The argument
proceeds essentially as follows:
1. It is conceivable that there is a world which is physically exactly like our

world, but in which there is no consciousness.
2. If the world described in (1) is conceivable, then it is metaphysically possible.
(E2D)
3. If the world described in (1) is metaphysically possible, then physicalism is
false.
4. Therefore, physicalism is false.
5. In particular, consciousness is physically unacceptable (and moreover might
be Strongly emergent).
In aiming to provide independent motivation for taking explanatory gaps to have

anti-physicalist metaphysical import, Chalmers’s argument is different from, and
to my mind improves on, the knowledge arguments. Nonetheless, as I’ll presently
argue, Chalmers’s line of thought is also ultimately unsuccessful, and hence fails to
motivate taking consciousness to be physically unacceptable, much less Strongly
emergent.
The focus of my critical attention is on the second premise.⁹ I’ll first present
Chalmers’s two-dimensionalist argument in more detail, highlighting the attrac-
tiveness of the E2D strategy that lies at the argument’s core. Next, drawing on Biggs
⁹ Rejecting the first premise seems likely to lead directly to stalemate (a problem discussed in Biggs
and Wilson 2019 and 2020), not just because what counts as ‘conceiving’ is unclear, but because (even
holding fixed a given understanding of the conceivability at issue) people disagree about what they
conceive. (Similar remarks apply to variations on the theme of conceiving as involving, e.g., rational
intuition as per Bealer 2002.) Rejecting the third premise is ultimately indecisive. One might deny
that the metaphysical possibility of a zombie world suffices to establish the falsity of physicalism,
on grounds that the premise presupposes that any relation between conscious states and lower-level
physical states compatible with physicalism must be one according to which the latter metaphysically
necessitate the former, but (the objector maintains) this presupposition is incorrect. Recall, e.g., that it
would suffice for the physical acceptability of conscious states that the latter be functionally realized by
physical states, in such a way that the powers of the former are, on any given occasion, a proper subset
of those of the latter; but if powers are only contingently associated with features, then the physical
states that realize conscious states in worlds with our laws of nature might not do so in worlds with
different laws of nature—compatible with, in particular, the existence of zombie worlds. This response
to Chalmers’s two-dimensional argument is ultimately indecisive, however, since he might maintain
that the argument applies even when the holding of the actual physical laws is built into the specification
of worlds in which the lower-level physical base goings-on are present.
and Wilson (2019 and 2020), I’ll suggest that there is an alternative, and superior,
way in which the E2D strategy might be implemented—namely, by appeal to
an abduction-based rather a conceiving-based epistemology of the meanings
entering into this strategy. I’ll then argue that it is far from clear that the genuine
possibility of zombies, or the associated Strong emergence of consciousness, is
output from E2D, when this framework is implemented using abduction rather
than conceiving. One might wonder, as against this line of thought, whether
abduction would be suited for purposes of implementing E2D, given that (as
above) the access to the meanings which are in turn supposed to provide a basis
for access to modal truths is supposed to proceed in a priori fashion. Here again,
I draw on joint work with Biggs (Biggs and Wilson 2017), where we argue that,
contra common assumption, abduction is an a priori mode of inference—as a
priori as conceiving, in particular. The upshot at the end of the day will be that,
like the knowledge arguments, Chalmers’s two-dimensional argument fails to
establish that consciousness is actually physically unacceptable, much less Strongly
emergent. That’s the overview; now for the details.
Epistemic two-dimensionalism
What is missing from the knowledge arguments is independent good reason to
take the explanatory gaps that they highlight to have anti-physicalist metaphysical
import. Chalmers’s intended route to such good reason takes as its starting point
Kripke’s (1972/80) undermining of the traditional supposition that modal truths
are (always) a priori accessible. Traditionally, claims about what is possible or
necessary were taken to be knowable a priori, reflecting that such modal claims
were presumed not to hinge on actual facts. Kripke argues, however, that many
necessary truths about individuals and natural kinds can be known only a poste-
riori. For example, one could not have determined a priori that water is identical
to H2 O; that water is H2 O is an empirical matter. Nonetheless, Kripke plausibly
asserts, given that water is H2 O, then water is necessarily H2 O; so, that water is
necessarily H2 O is knowable only a posteriori. For another example, one could not
have determined a priori that this lectern is made of wood; that a certain lectern
is made of wood is an empirical matter. Nonetheless, Kripke plausibly maintains,
given that a certain lectern is made of wood, it follows that it is not possible for
that very lectern to be made of some completely different kind of material—again,
a modal claim knowable only a posteriori. More generally, Kripke’s results might
be seen as undercutting the prospects of our having much, if any, a priori modal
knowledge about broadly scientific goings-on.
These consequences would be undesirable, not just for purposes of assessing
the truth of physicalism, but more generally because much theory and practice
(philosophical, legal, semantic, scientific, etc.) presupposes that we can know
modal truths independently of, or at least prior to the end of, empirical inquiry.
We might thus hope that, notwithstanding Kripke’s results, the traditional link
consciousness 229
between necessity and a priority could be restored, to at least some considerable

extent.
This is the promise of epistemic two-dimensional semantics (E2D), advocated
by Chalmers (1996, 2006b, 2009), Chalmers and Jackson (2001), and others.1⁰
E2D can be heuristically seen as refining and generalizing Frege’s (1892/2010)
suggestion that there are two kinds of meaning: sense and reference. On the
Fregean picture, sense represents an aspect of meaning that is a priori accessible
to a competent speaker, whereas reference represents an aspect of meaning that
may fail to be so accessible. For example, Frege maintained that a competent
speaker could know a priori the senses of ‘Hesperus’ and ‘Phosphorus’ as ‘the
first star seen in the evening’ and ‘the first star seen in the morning’, respectively,
while not knowing the empirical fact that these expressions have the same referent
(namely, the planet Venus). Moreover, on a standard reading of Frege’s view, our
a priori access to sense provides a basis for a priori knowledge more generally, in
that whether a claim is knowable a priori depends entirely on the nature of the
relations among the senses of its constituent expressions and the nature of our
access to them. For example, on Frege’s view ‘bachelors are male’ is knowable a
priori because the senses of ‘bachelor’ and ‘male’ appropriately overlap, and one
can grasp the senses of ‘bachelor’ and ‘male’ in a way revealing this overlap.
E2D similarly aims to characterize two distinct aspects of meaning, each repre-
sented as ‘intensions’: functions from possible worlds to extensions. More care-
fully, and following Chalmers, intensions are functions from either scenarios
(centered worlds, incorporating a subject’s perspective) or scenario-world pairs
to extensions. The ‘primary’ intension of an expression E takes each scenario s to
the extension of E in s on the supposition that s is actual. The ‘secondary’ intension
takes each scenario-world pair < s, w > to the extension of E in w given that s is
actual. The E2D strategy supposes that many natural kind expressions of the sort
entering into a posteriori necessities are associated with primary and secondary
intensions, so understood. (Note that the exposition to follow is abbreviated, owing
to considerations of space. The interested reader is directed to Chalmers 2006b and
2009 and Biggs and Wilson 2019 and 2020 for more detailed expositions.)
There are several useful ways of characterizing these intensions. On one,
primary intensions track extensions in contexts of utterance, and secondary
intensions track extensions in contexts of evaluation given a context of utterance.
On another, primary intensions track extensions in worlds ‘considered as actual’,
and secondary intensions track extensions in worlds ‘considered as counterfactual’,
holding fixed which world is actual. In any case, the promise of E2D ultimately
1⁰ See, e.g., Peacocke 1993, Chalmers 1996, Boghossian 1996, and Gertler 2002.
lies in the suggestion that primary and secondary intensions are connected in
such a way as to provide a basis for a priori knowledge of a wide range of modal
truths. To understand the suggested connection, it is useful to think in terms of
a two-dimensional matrix, with worlds ‘considered as actual’ along the vertical,
and worlds ‘considered as counterfactual’ along the horizontal (cf. Chalmers
2006b, following Stalnaker 1999). For example, the two-dimensional matrix for the
intensions associated with the term ‘water’ would look something like Figure 7.1:
‘water’ H2 O-world XYZ-world …
H2 O-scenario H2 O H2 O …
XYZ-scenario XYZ XYZ …
… … … …
Figure 7.1 Two-dimensional matrix of intensions for ‘water’
The entries in the matrix encode that our access to appropriate intensions, on the
E2D strategy, provides a basis for our knowing a priori that if water (or the watery
stuff) is actually H2 O, then water is necessarily H2 O; and if water (or the watery
stuff) is actually XYZ, then water is necessarily XYZ; and so on. Correspondingly,
while it is (as per Kripke’s results) a posteriori that water is necessarily H2 O, the a
posteriori contribution to this and related necessity claims concerning natural kind
terms is limited to discharging the antecedent of a conditional known a priori—
compatible with our having a priori access to a great deal of modal knowledge.
Granting that we have independent reason to accept the E2D framework and
associated strategy for reforging the link between a priority and modality, the
question remains which epistemology of intensions (meanings) should be taken to
be operative in implementing this strategy. As prefigured at the start of this section,
it has commonly been taken for granted that the epistemology of modality must
rely on conceiving. Hence Gertler (2006) says:
Conceivability is the only guide to necessity; our concepts, and the intuitions
about possibility that derive from them, provide our only grip on modal claims.
[…] Since a claim of impossibility cannot be established by considering the actual
world alone (though of course it can be refuted in this way), [one] must consider
whether certain non-actual scenarios are possible. And the only way to determine
this is to use the method of conceivability. (205)
Given a conceiving-based epistemology of intensions (CEI), we are close to

the anti-physicalist conclusion of Chalmers’s conceivability argument. CEI implies
that ‘zombie’ has a positive extension at some world if one can conceive of a zombie
consciousness 231
world; E2D coupled with CEI implies that, if one can conceive of a zombie world,
then zombies are metaphysically possible. But as per the first premise of Chalmers’s
argument (which I am granting), one can conceive of a zombie world. So zombies
are metaphysically possible, which given the third premise of Chalmers’s argument
(which premise I am also granting) implies that consciousness is not physical, and
moreover might be Strongly emergent.
This is an interesting line of thought, but as I’ll now argue, we have both negative
and positive reason to reject it.
Conceiving-based vs. abduction-based epistemologies of intensions

Let’s start by asking: why think that the only epistemic access to the intensions of
our concepts or terms proceeds via conceiving, as per CEI? Why not, in particular,
allow that such access might proceed via inference to the best explanation, or
abduction, as per an abduction-based epistemology of intensions (AEI)? Why
not allow that various theoretical desiderata or abductive principles—plausibility,
compatibility with other beliefs, unifying power, ability to resolve certain prob-
lems, explanatory fruitfulness, ontological and ideological parsimony, and the
like—can come into play in identifying the extensions of certain concepts or terms
in non-actual scenarios (which extensions enter into constituting the associated
intensions)? After all, on the face of it such considerations might well be relevant.
Plausibly, AEI has not been taken seriously as a basis for implementing the
E2D strategy on grounds that, if the strategy is to do the job of reforging an a
priori accessible link between meaning and modality, the operative epistemology
of intensions must be a priori (involve an a priori mode of inference); and while
conceiving is a priori, abduction is not. As Biggs and I argue in our (2017 and
2019), however, abduction is a priori. In our papers, we offer many positive and
defensive motivations in support of taking abduction to be a priori. Here I offer
sketches of two argumentative lines of thought for this position, and call attention
to some historical and contemporary precursors of our view.
First, those taking abduction to be a posteriori typically do so on grounds that
whether abduction has epistemic value, sufficient unto conferring justification on
conclusions output from that mode of inference, depends on contingent facts. As
Beebe (2009) notes:
[P]ractically everyone who works on abductive inference believes that such

inferences are justified empirically and that the theoretical virtues are broadly
empirical and contingent marks of truth. (625)
What motivates this understanding of the epistemic value of abduction? To

fix ideas, consider a specific abductive principle—say, Parsimony, a version of
ontological parsimony according to which, ceteris paribus, one should choose
the theory involving the fewest fundamental type-level ontological commitments.

Why think that the epistemic value of Parsimony (hence abduction, as partly
constituted by Parsimony), depends on experience? Something like the following
sort of case seems to be at issue:
Sam gains a bunch of evidence about the world and identifies two theories, T and
T*, that would appropriately accommodate the evidence. At t1 , she picks T as the
most parsimonious theory, but later, at t2 , gains new evidence supporting T*. So
choosing T on the basis of Parsimony led her astray.
As Biggs and I argue, however, such a case (or series of cases, or proportion of
such cases) at a world does not show that abduction lacks epistemic value in that
world, but rather that the supposition that other things were equal was, as it turned
out, mistaken.11 As we put it in our (2017):
Consider how the objector would describe cases in which, they claim, Parsimony
is taken to initially support a theory that further evidence later disconfirms.
According to the objector, at t1 Parsimony supports theory T over theory T*
given explananda E, and at t2 we discover additional explananda E′ , which T*
explains better than T does. […] A more accurate description of such cases is as
follows: at t1 we mistakenly believe that T and T* are otherwise equal (because
we are unaware of E′ ), but at t2 we discover our mistake. Accordingly, rather than
the discovery of E′ disconfirming a theory initially supported by Parsimony, that
discovery reveals that we initially applied Parsimony too hastily. And similarly,
mutatis mutandis, for any cases that might be thought to figure in empirically
[disconfirming] Parsimony. (745)
Generalizing, we suggest that the usual supposition that abduction is a posteriori

reflects a failure to appreciate that the ceteris paribus clauses associated with
abductive principles operate to shield them from empirical disconfirmation.
By way of support for this line of thought, note that it is commonly recognized
that failure to possess all the relevant data can lead to false belief via modes of
inference whose epistemic value is taken to be a priori. As B. Russell (2017) recently
observed:
You might initially be a priori justified in believing that no matter how happiness
has been produced it is intrinsically good […], or that it is always wrong to punish
an innocent person […], and later think of counterexamples to such claims (e.g.,
happiness had through the suffering of others or punishing an innocent person
to prevent some evil men from punishing him and many other innocent people).
11 As I’ll discuss shortly, this is not to say that the initial application of Parsimony was inapt.
consciousness 233
For another example: were one to initially apply modus ponens to some premises
later revealed to be false, that would show not that modus ponens is a posteriori,
but rather that one was working with faulty (incomplete or inadequate) data.
Similarly, we maintain, for the case of Sam’s abductive inference: that new evidence
might undermine the conclusion of a given abductive inference goes nowhere
towards establishing that abduction is a posteriori.
To be sure, there are further choice points associated with cases in which one
applies a given mode of inference to faulty (incomplete or inaccurate) data. As
above, B. Russell (2017) supposes that in such cases the conclusions, even if false,
may be a priori justified (such that, notwithstanding that Sam was in fact mistaken
in thinking that all else was equal between T and T*, she was nonetheless justified
in taking T to be the best explanation at t1 ). By way of contrast, Chalmers builds
into his (2002) account of conceivability as an a priori guide to possibility that
the conceivers are suitably ideal and in possession of all relevant information, and
that the products of such conceivings are indefeasible, which suggests that the
conclusions of non-ideal, not-fully-informed conceivings might not be a priori
justified by Chalmers’s lights (such that Sam was not a priori justified in taking T
to be the best explanation at t1 ). Another choice point might reflect the distinction
between internalist and externalist understandings of the justification in these
cases. For present purposes, what is important is that cases such as that involving
Sam do not in themselves provide a basis for taking abduction to be a posteriori.
On the contrary, such cases are reasonably interpreted as indicating, as above,
that insofar as the ceteris paribus clauses associated with abductive principles
operate to shield them from empirical disconfirmation, the epistemic value of
these principles (and of abduction more generally) is a priori, with failures of
abductive inference to lead to true conclusions being due not to any purported
absence of epistemic value of abductive principles or abduction, but rather to
failure of the abductor to be in possession of all the relevant data.12
A second line of thought in favor of taking abduction to be a priori proceeds
by attention to the roles experience may play in inferential reasoning. Consider a
claim p. There are four ways in which experience might play a role in the course of
a particular belief in p coming to be justified. Experience might play a role in
1. acquiring the concepts required to entertain p,

2. acquiring the evidence required to justify belief in p,
3. justifying belief in the epistemic value of the mode of inference used to justify
belief in p, or
4. acquiring or learning to deploy the mode of inference used to justify belief
in p.
12 There is more to say by way of filling in and defending this line of thought; I direct the interested
reader to Biggs and Wilson 2017 and 2019.
Can a belief for which experience plays an ineliminable role along one or more
of (1)–(4) be justified (‘entirely’) a priori? As Biggs and I discuss in our (2017), it
depends on which role is at issue. Re (1): it is commonly maintained that belief
in p can be justified a priori even if experience is needed to acquire the concepts
required to entertain p. For example, belief in ‘sisters are siblings’ can be justified
a priori even if we need experience to acquire the concepts or meanings of ‘sister’
and ‘sibling’. Re (2): it is commonly maintained that belief in p cannot be justified
a priori if experience is (at all) needed to acquire the evidence supporting p. This is
the sense in which any reliance on experience or empirical facts suffices to render
the associated justification a posteriori. For example, belief in ‘water is H2 O’ is
commonly taken to be justified a posteriori, on grounds that justifying this belief
requires, among other things, acquiring empirical evidence to the effect that water
and H2 O are spatiotemporally coextensive. Re (3): it is commonly maintained that
belief in p cannot be justified a priori if experience is needed to justify belief in the
epistemic value of the mode of inference required to justify belief in p. Re (4):
although the role of experience in an agent’s acquiring or learning how to deploy
a given mode of inference is not much discussed, it seems reasonable to maintain
that this role is relevantly similar to that of (1): in each case, experience contributes
to belief (or supposition) formation, not to justification per se. For example,
students often need encouragement to think in an appropriately imaginative way
about what is possible, but that such imaginative ‘training up’ is required in order
to engage in suitably competent conceiving is not taken to undermine the status of
conceiving as an a priori mode of inference. Hence it seems reasonable to maintain
that belief in p can be justified a priori even if experience is required in order to
acquire or learn how to deploy the inferential mode at issue. Correspondingly,
among the roles that experience may play in inferential reasoning, only (2) and
(3) are such that the playing of this role in the justification of a given belief would
render that justification a posteriori.
Now, suppose that belief in a given claim p is justified via abduction. Following
the considerations just canvassed, whether that belief is justified a priori turns on
whether experience enters into its justification via either role (2) or role (3)—that
is, on whether experience provides evidence for p without which that belief in p
would not be justified, or whether experience plays (played) a role in justifying
belief in the claim that abduction has epistemic value.
Following the discussion above, experience does not enter into abductively
justifying the belief in p via role (3): the epistemic value of abduction does not
rely on experience. Does abductive justification require that experience enter via
role (2)? And in cases where experience does enter via role (2) in the abductive
justification of a given belief, should the contribution of abduction to this jus-
tification be considered a posteriori? No, and no. To start, notwithstanding that
abduction can and often does operate on empirical evidence (e.g., ‘water and H2 O
consciousness 235
are spatiotemporally coincident’) to produce a claim (e.g., ‘water is H2 O’) that

is a posteriori justified, the underlying inferential transition can be encoded in
conditionals—e.g., ‘if water and H2 O are spatiotemporally coincident, then water
is H2 O’—which are justified by means of a hypothetical form of abduction. Such
a hypothetical abductive inference would be akin to suppositional reasoning in
conditional proof, allowing one to identify what would be the best explanation
of the state of affairs expressed in the antecedent, were this state of affairs to
obtain (were the antecedent to be true). And since such conditional beliefs may be
justified without the antecedent claim’s being either believed or true, abduction in
such cases operates independently of any claim that is justified through experience.
Accordingly (and given that the epistemic value of abduction does not rely on
experience, as we have argued), such abductively justified conditional beliefs
are ‘entirely’ a priori. Hawthorne (2002, 252) makes a similar point, suggest-
ing that abduction can deliver a priori justification for belief in a conditional
whose antecedent describes an ‘experiential life history’ and whose consequent
is whichever theory best explains some aspect of that life history; see also Cohen
2010, 152–3 and Wedgwood 2013.
It follows that abduction is an a priori mode of inference, even when it operates
on (non-hypothetical) empirical evidence (registering, e.g., the actual spatiotem-
poral coincidence of water and H2 O). To be sure, when abduction operates
on empirical evidence, it does not produce beliefs with (completely) a priori
justification, any more than (e.g.) modus ponens does. But the contribution of
empirical evidence in such cases of abductive inference is simply to discharge the
antecedent of a conditional whose justification is entirely a priori. The same is
true of conceiving; and indeed, it is precisely the appreciation that we can have
a priori access to relevant conditionals that serves as the basis of the E2D strategy
for regaining, post-Kripke, significant a priori access to modal claims.
Finally, it is worth noting that the view that abduction is a priori has clear
historical precursors in the work of Kant and Carnap. In brief (see Biggs and
Wilson 2019 for further discussion) Kant took synthetic a priori truths (including
most philosophical claims) to be justified via an ampliative mode of inference—
plausibly, abduction; Carnap took knowledge of conceptual content to proceed
via explicitly abductive ‘explication’. More recently, and in addition to Hawthorne
(2002), Cohen (2010), and Wedgwood (2013), several philosophers, including
Bonjour (1998), Swinburne (2001), and Peacocke (2003), have offered reasons
(different from ours; see the comparative discussion in Biggs and Wilson 2017)
aiming to establish the a priority of certain ampliative modes of inference or
abductive principles.
Summing up: although it has commonly been taken for granted that abduction
is a posteriori, there are cases to be made that abduction is a priori; and this view
is not as unusual as it might first appear.
I next want to highlight the primary independent reason for thinking that, given
that abduction is a priori, we should look to abduction rather than conceiving as
a guide to the intensions at issue in the E2D strategy.13
This reason reflects what might be called ‘access-based’ objections to E2D,
according to which the common presence of indeterminacy or inconsistency
in our concepts indicates that we commonly fail to have a priori access to the
intensions of natural kind expressions (among others), to such an extent that the
prospects of an E2D-led reforging of the link between meaning and modality
are fatally undermined. As Biggs and I argue in our (2020), these access-based
objections are not to E2D per se, but rather to E2D when implemented using CEI;
for while conceiving alone does not have the resources to resolve indeterminacy
or inconsistency in our concepts, abduction does have such resources. More
generally, Biggs and I argue that when E2D is combined with AEI rather than CEI,
the full range of access-based objections can be satisfactorily answered. In what
follows, I discuss as an illustrative case in point one salient access-based objection.
Afterwards I return to Chalmers’s conceivability argument, and the question of
whether an implementation of E2D provides reason to think that zombies are
genuinely metaphysically possible.
According to the objection at issue, widespread conceptual indeterminacy—
better: underdetermination1⁴—undermines the prospects for E2D. The line of
thought here is that implementing the E2D strategy requires that extensions of
expressions (e.g., natural kind terms) have a priori-accessible extensions in every
possible scenario; but in cases of conceptual underdetermination the associated
expressions lack fixed antecedent extensions in some scenarios. Moreover, the
concern runs, such underdetermination is widespread. That there is widespread
conceptual underdetermination is suggested by M. Wilson’s (1982 and 2006b)
arguments to the effect that applications of natural kind predicates can and fre-
quently do depend on arbitrary factors. For example, in one of Wilson’s toy cases,
whether members of an isolated tribe take airplanes to be in the extension of their
predicate ‘bird’ (or linguistic correlate) might depend on whether the first airplane
they encounter is overhead (in which case, airplanes are judged to be birds) or
is rather on the ground (in which case, airplanes are judged not to be birds). If
such historical accident might partly determine whether ‘bird’ applies to airplanes,
then, Wilson reasonably assumes, the full range of the predicate’s intension was
not antecedently determined. And, Wilson suggests, such underdetermination is
13 For a detailed and more general comparison of abductive vs. conceiving-based epistemologies of
modality, see Biggs and Wilson 2019.
1⁴ Though the objection and responses have been pitched as involving conceptual ‘indeterminacy’,
‘underdetermination’ is a better term for the phenomenon at issue, since ‘indeterminacy’ has connota-
tions associated more specifically with vagueness or metaphysical indeterminacy, of the sort discussed
in the previous chapter.
consciousness 237
characteristic of natural kind predicates more generally, undermining hopes that

E2D can regain much modal knowledge about such kinds.
Chalmers grants that there may be underdetermination of the sort at issue in
Wilson-style cases. In his (1996), he says “There may of course be borderline cases
in which it’s indeterminate whether a concept would refer to a certain object if a
given world turned out to be actual” (364), and more recently, Chalmers grants that
it is plausible that “later extensions [of the kind of expressions at issue in Wilson’s
cases] depend on idiosyncratic developments, and verdicts about such cases are
not determinately prefigured in a user’s original use of an expression” (Chalmers
2012, 231). Chalmers maintains, however, that
[B]orderline cases [are] no problem: we can allow indeterminacies in a primary

intension, as we sometimes allow indeterminacies in reference in our own
world. (1996, 364)
But as Biggs and I note, while Chalmers is right that E2D can tolerate some concep-
tual underdetermination, it cannot tolerate widespread underdetermination, on
pain of undermining E2D’s raison d’être of providing an appropriately expansive
basis for our a priori knowledge of a wide range of modal truths, including
conditional such knowledge of necessary a posteriori truths (such that, e.g., we
can know a priori that if the actual world is one where water is H2 O, then it is
necessary that water is H2 O). As such, widespread conceptual underdetermination
would undermine E2D—at least if such underdetermination is insuperable.
Might Chalmers maintain either that conceptual underdetermination is not
really widespread, or that in any case such underdetermination is superable using
the resources of conceiving? Neither strategy is promising.
First, conceptual underdetermination is widespread, for both natural kind and
other predicates (more generally, expressions). Besides the Wilson-style cases,
Biggs and I note two further sources of conceptual underdetermination. The first
concerns cases of vague predicates of the sort associated with Sorites sequences
(e.g., ‘red’, ‘bald’, ‘rich’, ‘cell’, ‘part’, etc.). As Raffman (1994) compellingly argues,
the application of such predicates can be and often is typically determined in
a given case by arbitrary contextual and/or psychological factors. For example,
the breaking point at which one stops (or starts) applying the predicate ‘red’
in a given color-spectrum sequence may non-systematically depend on where
along the spectrum one starts (as per the phenomenon of hysteresis) as well
as on arbitrary psychological factors (fatigue, boredom, etc.). For Wilson-style
reasons, the dependence of extensions on arbitrary historical and other factors
supports thinking that the associated expressions do not come with antecedently
fixed extensions—i.e., these extensions are to some extent conceptually under-
determined. But those discussing the phenomenon of vagueness often register
its being likely that most expressions in ordinary language (including scientific
language) are vague—in which case Raffman’s results show that most expressions
are conceptually underdetermined.
A second source of widespread conceptual underdetermination reflects the
historical scientific record, which supplies many cases of underdetermination
involving natural kind terms. For example, ‘acid’ initially was taken to refer to only
oxygenated substances, but was later applied to HCL, for theoretical reasons now
largely discarded; dispute persists over whether Newtonian uses of ‘mass’ apply in
relativistic contexts; the decision to classify whales as mammals was a controversial
affair; the classification of Pluto as a planet, or not, remains a topic of dispute.
It seems clear, then, that conceptual underdetermination is widespread. Hence
such underdetermination will be ‘no problem’ for E2D only if the operative
epistemology of intensions (meanings) is capable of resolving the underdetermi-
nation. Now, Chalmers (1996) does consider whether conceivers might be able
to eliminate underdetermination by foreseeing relevant accidents. In Wilson’s toy
case, for example, Chalmers says that one “might try to classify these two different
scenarios [airplane first seen in the sky or on the ground, respectively] as different
ways for the actual world to turn out, and therefore retain a fixed, detailed primary
intension” (364). On this strategy, the fully determinate primary intension of ‘bird’
includes planes in its extension if the tribe members first see a plane overhead,
but not if they first see it on the ground. Either way, according to the suggestion,
the underdetermination in the tribe’s expression ‘bird’ is resolved. And Chalmers
might attempt to implement a similar strategy as a general means of overcoming
conceptual underdetermination involving vague or scientific predicates or other
expressions.
Such a strategy has traction against the underdetermination objection only
if conceivers can foresee how intensions are sensitive to accidental or arbitrary
factors. But as Biggs and I see it, a deeper lesson of Wilson’s case, as well as of
Raffman’s discussion, is that the influence of such factors cannot be foreseen, at
least not by conceiving alone. Determinism and such aside, those using conceiving
alone might apply ‘bird’ differently even relative to the same ‘accident’. After all,
there are any number of respects of dissimilarity between airplanes and birds,
even when the former are in flight, and a minor difference in attention to these
features might result in a different decision about whether ‘bird’ applies to a flying
airplane. And as Raffman points out, the factors entering into the application of a
vague predicate are not just arbitrary but are unsystematically so, as is reflected
in different applications of vague expressions even against relevantly the same
background conditions. We can register, post hoc, extensions resulting from
whatever decision was in fact made, but in cases of arbitrary determination there
is no way to antecedently identify these extensions and corresponding intensions
consciousness 239
through conceiving alone—there is simply no fixed extension to ‘conceive’, even

taking relevant circumstances into account.
Similarly, there is no case to be made that even an idealized conceiver could
determine the extension, in every scenario, of natural kind expressions pertaining
to ‘mass’, ‘planet’, and so on. As a matter of historical fact, the considered extensions
of these terms have been heavily informed by abductive considerations, taking
into account plausibility, compatibility with other beliefs, unifying power, ability
to resolve certain problems, explanatory fruitfulness, ontological and ideological
parsimony, and so on. But as Chalmers and others implementing E2D using CEI
emphasize, conceiving does not involve these sorts of abductive or ampliative
resources.1⁵ Correspondingly, conceiving alone is simply not up to the task of
overcoming conceptual underdetermination in these cases.
By way of contrast, AEI is up to the task of overcoming conceptual underdeter-
mination, in the cases of natural kind expressions and more generally. For when
considering whether to apply an expression in a given scenario, abductors can
consider not only historical accident and psychological variability, but also any
non-demonstrative rational grounds that might push towards one extension rather
than another. Hence abduction, unlike conceiving, can be productive (it is amplia-
tive, after all), allowing those who are identifying intensions to consider how the
concept should be applied, given the usual abductive principles and associated the-
oretical desiderata, even when the application conditions are antecedently to some
extent underdetermined. Since abduction can, in a rational way, go beyond what
expressions antecedently encode, AEI has the potential to overcome conceptual
underdetermination, extending applications of natural kind expressions to new
scenarios or situations, on ultimately rational grounds. For example, an idealized
version of a competent user of ‘bird’ who has never before seen an airplane would
plausibly be in position to consider and compare theories of the intensions of ‘bird’
(as ever including or as always excluding airplanes, however first encountered),
by attention to which theory would be, among other desiderata, most plausible,
consonant with our other beliefs, unifying (e.g., of experience of flying entities),
fruitful, ontologically and ideologically parsimonious, and so on. Similarly for the
cases of ‘mass’, ‘Pluto’, ‘water’, and so on. Indeed, in these and many related cases,
the antecedent underdetermination has been resolved (when it has been resolved)
not through accident, but through abductive reasoning. Here it is worth noting,
by the way, that there is no barrier to an idealized abductor engaging in abductive
1⁵ For example, Chalmers and Jackson (2001) are explicit (when arguing, in particular, against Block
and Stalnaker’s 1999 claim that the justification for the conditional claims output by E2D might rely on
broadly abductive considerations) that in their view abductive considerations do not play any ‘essential
justificatory’ role in the cases at hand; see especially 342–50.
deliberations about how best to extend a given expression or concept, for this is
more or less what language-users do as a matter of course.
Similar results attach to other access-based objections (see Biggs and Wilson
2020 for discussion). Even focusing just on underdetermination, however, the
upshot is clear: E2D when implemented using a conceiving-based epistemology
of intensions faces difficulties to which an abduction-based epistemology of
intensions, in having ampliative resources of the sort that conceiving fails to have,
is in position to respond.
Revisiting the possibility of zombies

This result in hand, I now return to the status of Chalmers’s two-dimensionalist
argument against materialism, and in particular, to the question of whether the
independently desirable E2D framework supports taking zombies to be metaphys-
ically possible. Effectively, the question is: are there worlds that are physical and
functional duplicates of our own (including, let us grant, worlds where the physical
laws are the same as our own) in which the intension of ‘zombie’ has a non-null
extension—that is, where there are zombies?
We can start by observing that on AEI, an attempt to answer this question will
proceed by attention to any and all relevant considerations, which may include
but are not restricted to any modal intuitions we might have. The usual theoretical
desiderata will be relevant, as will be considerations such as the desirability of
properly accommodating the causal efficacy of consciousness, and (perhaps relat-
edly) the desirability of conforming, for purposes of systematicity and plausibility,
to what our best sciences give us reason to believe. Accordingly, different ‘theories’
of the intensions at issue will rate better or worse, depending on how well they do
at accommodating or satisfying these diverse considerations.
A theory on which the extension of ‘zombie’ has a non-null extension at some
world scores positively in accommodating the seeming conceivability of zombies.
On the other hand, insofar as such a non-null extension implies that consciousness
is Strongly emergent or otherwise physically unacceptable, the associated theory
of the intension of ‘zombie’ is less ontologically parsimonious (however exactly
such parsimony is construed) than one on which the extension of ‘zombie’ is null
at the relevant worlds. The former theory is also less systematic than the latter,
so far as accommodating the possibility that conscious states can have physical
effects is concerned, for it is not perfectly clear how nonphysical goings-on can
interact with physical goings-on (as was originally pressed by Princess Elizabeth
of Bohemia against Descartes; see Kim 2015 for recent discussion); and though my
own view is that this can be made sense of via the notion of a fundamental mental
interaction, in any case such an account requires further ontology and ideology.
One might sidestep this concern by coupling a theory on which ‘zombie’ has a
non-null extension with the supposition that conscious states are epiphenomenal,
consciousness 241
as Chalmers sometimes seems to suggest. But (and as is reflected in the dearth of

‘live’ epiphenomenalist accounts of mentality) such a denial is highly implausible,
given our experience, and moreover unsystematic, given that no other empirical
goings-on are taken to be epiphenomenal; hence this theory too suffers by way of
comparison with one taking ‘zombie’ to have a null primary intension.
The ultimate abductive comparison of theories of the intension(s) of ‘zombie’
will involve not just these, but many other considerations. Even without entering
further into this issue, however, it is clear that an abductive rather than (merely)
conceiving-based approach to the question at hand is very far from indicating that
‘zombie’ has a non-null extension at some world, as is required on the E2D strategy
if zombies are to be possible. If anything, the all-things-considered weight seems
likely to be on the other side. To be sure, a positive empirical result, indicating
the presence of a novel fundamental interaction at the level of mentality, might
trump all these other considerations; but at present, we are lacking such evidence.
Hence even granting the independent desirability of the E2D framework as well
as the conceivability of zombies, nothing yet follows about whether zombies are
metaphysically possible.
Chalmers’s conceivability argument, like the knowledge argument(s), thus fails
to establish that consciousness and its associated subjective and qualitative aspects
are actually physically unacceptable, much less actually Strongly emergent.
7.2 Is consciousness Weakly emergent?
I turn now to considering whether any conscious mental features are plausibly seen
as Weakly emergent from the lower-level physical states upon which they depend.
As prefigured, my strategy for answering this question will proceed by moti-
vating the claim that qualitative conscious features—e.g., states of conscious
awareness of colors or pains—are reasonably taken to be typically determinable
rather than (maximally) determinate, such that (on the assumption that conscious
states are not actually Strongly emergent) these determinable conscious states are
plausibly seen as actually realized by determinate lower-level physical states, in
accord with a determinable-based account of Weak emergence. As we’ll see, this
application of a determinable-based account faces certain challenges, which I will
argue (following Wilson 2009) can be addressed.
7.2.1 Determinable perceptions
Among the most salient conscious mental states are those which we might call
‘qualitative’, in constitutively involving experience of features such as colors, tastes,
textures, pains, and the like. There are (at least) two motivations for taking
the qualitative aspects of conscious experiences to be determinable rather than
maximally determinate, rendering the states themselves appropriately seen as
determinable rather than maximally determinate.
The first motivation appeals to Sorites phenomena, as indicating that we fail to
perceive fully determinate instances of many properties, including colors, tones,
and textures. A Sorites sequence consists in a sequence of pairwise indiscriminable
instances of such features. An application of the notorious Sorites paradox consists
in first registering the plausibility of a principle according to which, if one starts
with an instance that is (judged to be), e.g., clearly red, and given that the next
patch is indiscriminable from the first, the next patch must also be (judged to be)
red. One then iteratively implements this reasoning, starting with the first (clearly
red) patch, such that the second, indiscriminable patch must also be (judged to be)
red, and similarly for the second and third patches, and so on, in what is sometimes
called a ‘forced Sorites march’—until one arrives, contradictorily, at a patch that is
clearly not red (e.g., a patch that is clearly orange), but which must be (judged to be)
red. As Fales (1990) notes, that the qualitative aspects of perceptual experiences are
to some extent determinable rather than maximally determinate “is a conclusion
which seems forced upon us by the fact that each member of a series of colors,
etc., may be perceptually indistinguishable from its immediate neighbors but
easily distinguishable from more distant members of the series” (172).
The second motivation reflects that our perception of macro-entities and their
features typically fails to register micro-determinate details. As I observed in Ch.
1 (§1.1.2), the macro-entities of our acquaintance do not perceptually appear to us
as massively complex, constantly changing, configurations of micro-phenomena.
The same is true of the qualitative features of such entities. For example, we do
not experience the shapes of macro-entities in fully microscopically determinate
detail; rather we experience these shapes as to some extent determinable.
Two further points about these motivations for taking perceptions to be typ-
ically determinable are worth noting. First, the motivations here do not depend
on the claim that the objects or features perceived are objectively indeterminate.
My own view, as per the previous chapter, is that some features of ordinary
objects (including but likely not limited to their boundaries) are to some extent
metaphysically indeterminate, which when combined with a determinable-based
view of metaphysical indeterminacy entails that the features (hence the objects) are
in fact indeterminate in the relevant respect(s). But even if the objects and features
perceived are themselves completely determinate in all (or the relevant) respects, it
would remain that qualitative conscious states involving perceptual experience of
such objects would still be less than maximally determinate. As I note in my (2012),
in discussing how perception provides motivation for thinking that determinables
exist:
consciousness 243
Perhaps the (instances of) properties perceived are really maximally determinate,
and only perceptual features or modes of presentation are determinable; but
features of perceptual experience are also aspects of reality, so the larger point
remains. (5)
It seems reasonable to suppose, then, that at least some qualitative conscious states
are less than maximally determinate, owing to these states’ having qualitative
aspects that are less than maximally determinate.
Second, the phenomena at issue suggest that there is no hope here of providing a
deflationary—anti-realist or reductionist—treatment of the determinable qualita-
tive conscious states at issue. That we have qualitative conscious states (qualitative
experience) isn’t, in my view, up for grabs—it’s as epistemically foundational
in our experience as any aspect of reality, or more so. Hence any argument
for eliminating such states would be less compelling than the sheer fact of our
experience.1⁶ Nor is it plausible to suppose that these determinable states might
be given some or other reductive treatment in terms of maximal determinates.
There are a number of routes towards the rejection of any such reduction, including
that in Ch. 2, §2.3.1, according to which determinables and associated determi-
nates satisfy the Proper Subset of Powers Condition (see Wilson 2012 for other
arguments for irreducibility). For present purposes it suffices to observe that
any reductive treatment of a determinable state which proceeds by identifying a
determinable type with a disjunction (or some other ‘lightweight’ combination) of
determinate types entails that every token of a determinable type is identical with
a token of a determinate type—and indeed, a token of a maximally determinate
type. But the phenomena associated with qualitative conscious states undercut the
hope of any such reduction, since as previously, our token qualitative conscious
experiences are not maximally determinate. It is thus reasonable to assume that
at least some conscious experiences, understood as having constitutive qualitative
aspects, must be characterized in irreducibly determinable terms.
Now, as previously, one account of realization of the sort satisfying the con-
ditions in the schema for Weak emergence is a determinable-based account,
according to which it suffices for the realization of a feature that the feature
be a determinable of lower-level physical determinates. So, if the determinable
qualitative conscious states at issue can be seen as having lower-level physical
determinates, we will be in position to conclude that such conscious features are
Weakly emergent.
1⁶ This is a common, and to my mind compelling, reason to reject the sort of eliminativism
about qualitative mental states endorsed by the Churchlands (in, e.g., Patricia Churchland 1986 and
Paul Churchland 1981).
7.2.2 The objections from mental multiple realizability

and mental superdeterminates
Ehring (1996), Funkhouser (2006), and Walter (2006) argue, however, that it does
not make sense to see lower-level physical states as determinates of determinable
conscious mental states.1⁷ The common line of argument in these discussions is
schematically as follows:
1. The determinable/determinate relation has feature F.

2. The relation between conscious states and lower-level physical states does
not have F.
3. Therefore, the relation between conscious states and lower-level physical
states is not the determinable/determinate relation.
4. Conscious states are not realized, in determinable-based fashion, in lower-
level physical states.
There are two instantiations of the above argument schema relevant to the
question of whether qualitative conscious states might have lower-level physical
determinates.
The first, associated with ‘the argument from mental multiple realizability’,
appeals to the feature of the determinable/determinate relation according to which
determinates of a determinable differ ‘in respect of ’ their determinable. This
feature reflects that the distinctive form of specification whereby a determinate
is more specific than a given determinable is supposed to contrast with the
genus/species relation and, more generally, the conjunct/conjunction relation
(see Wilson 2017 for historical discussion). The latter specification relations are
compatible with the increase in specificity’s involving the conjunctive addition of
some independent property, as when a genus is conjoined with some differentium
to form a species, or when a state of affairs is made more specific by being
conjoined with another state of affairs. The specification at issue in paradigm
cases of the determinable/determinate relation is not properly understood in
such conjunctive terms. For example, scarlet is not appropriately analyzed as a
conjunctive combination of red and some other property; rather, determinates are
in some sense more intimately specific ‘in respect of ’ their determinables. One
way in which this more intimate variety of specification gets elucidated is in terms
of the determinable’s having certain ‘determination dimensions’, along which the
1⁷ Worley (1997) and Funkhouser (2006) also argue that certain mental states are not appropriately
seen as determinables of lower-level physical determinates; but since these discussions focus on beliefs,
they are not directly relevant to the present question of whether qualitative conscious states can stand
in a determinable/determinate relation to lower-level physical states. See Wilson 2017 for discussion
of Worley’s and Funkhouser’s concerns.
consciousness 245
determinable can be rendered more determinate (as per Funkhouser 2006). For
example, it is common to suppose that the determination dimensions of color are
hue, saturation, and brightness, with different determinates of color differing in
respect of color by differing in respect of how they are specified along one or more
of these determination dimensions.
Now, according to the argument from mental multiple realizability, that con-
scious states are typically multiply realizable rules out that lower-level physical
states differ in respect of any such conscious state, hence rules out taking conscious
states to be determinables of lower-level physical determinates:
[T]he physical realizers of the mental will not differ mentally at all, as they should
if they are determinates of the requisite mental states. (Ehring 1996, 474)
Mental properties are said to be multiply realizable precisely because distinct
physical realizers can be exactly the same with respect to the mental property
they realize. (Walter 2006, 219)
In particular, such multiple realizability is supposed to be characteristic of quali-

tative conscious states. Note that the problem being raised here is not that we can’t
make sense of determinates of a given qualitative conscious state as being exactly
the same with respect to the mental property they realize—there’s no problem
with this (perhaps, e.g., the multiple determinates are exactly similar in sharing the
determinable or its powers, or in their essences’s containing the same determinable
essence, as Yablo suggests in his 1992). Rather, the concern is that there is no clear
way to see qualitative conscious states as having determination dimensions that
can be further specified by their multiple lower-level physical realizers. Hence, for
example, if the determination dimensions of perceived colors are hue, saturation,
and brightness, there is no clear way of understanding how perceived values of
these features could be further specified by lower-level physical realizers: it’s not
as if the lower-level physical realizers are capable of realizing more precise color
perception! And similarly for other qualitative conscious states, such as pain.
The second instantiation of the above argument schema, associated with ‘the
argument from mental superdeterminates’, relies on the feature of the deter-
minable/determinate relation according to which some determinables admit of
maximal specification. In such cases, there is a maximal determinate or ‘superde-
terminate’ of the determinable. According to the argument from mental superde-
terminates, some qualitative conscious states are superdeterminate:
[S]uppose that M is a fully determinate type of mental state. For example, make M
a precise state of searing pain such that there is no room for further specification
of this mental state qua pain state. […] Suppose that M and “being in pain” have a
physical [realizer], P. Suppose that P is a determinate of M and “being in pain”. [If

M is realized by P in determinable-based fashion] M cannot be a fully determinate
pain state. This is so because there are further determinates of that pain state […]
in the form of P. But in fact, M is a fully determinate pain state by hypothesis.
Thus the physical [realizers] of mental properties are not determinates of that
which they realize since if that were true, M would not be a fully determinate
pain property. (Ehring 1996, 473)
Here the concern is clear—namely, that taking qualitative conscious features to

be realized in determinable-based fashion by lower-level physical features falsely
implies that certain qualitative mental superdeterminates can be further deter-
mined.
7.2.3 Responses to the objections
These concerns are important, but as I will now argue, they can be addressed, given
a proper understanding of the determinable/determinate relation.
A powers-based account of determination

Determinables (of whatever ontological category) are less specific than their
determinates. In Wilson 1999, I argued that this increase in specificity reflects a
proper subset relation between the sets of powers of the types of features involved,
as follows:
Powers-based Determination (first pass): Feature P is a determinate of feature Q iff

the set of powers associated with Q is a proper subset of the set associated with P.
Here the idea is that a determinate is more specific than its determinable in being
associated with a more specific set of powers. Hence it is, for instance that in virtue
of being scarlet, a patch can do more (say get Alice the picky pigeon to peck at it)
than the patch can do simply in virtue of being red.
This first-pass proposal has the virtue of ensuring that there is a contrast
between the determinable/determinate relation and the disjunction/disjunct rela-
tion (another specification relation with which determination is traditionally
taken to contrast), for as previously discussed in Ch. 2 (§3.2.2), disjunctions are
associated with more powers than associated disjuncts. As it stands, however, the
first-pass proposal does not ensure the contrast between the determinable/deter-
minate relation and the genus/species and conjunct/conjunction relations. Again,
it is crucial to the determinable/determinate relation that the determinate cannot
be understood or analyzed as a conjunction of the determinable and some other
property.
consciousness 247
We may preserve this intended contrast on a powers-based approach (improv-

ing on the first-pass proposal) by stipulating that the powers in the complement of
the sets associated with a determinable and any of its determinates, respectively,
are not associated with a set that is associated with any property, as per:
Powers-based Determination (second pass): feature P is a determinate of feature

Q iff Q is associated with a proper subset of the powers associated with P, and the
set of powers had by P but not by Q is not associated with any property.
The second pass account characterizes a specification relation that appropriately

contrasts with both the disjunction/disjunct and conjunct/conjunction relations.
Addressing the objection from multiple mental realizability

Recall that the deeper concern at issue in the objection from multiple mental
realizability reflects, first, a conception of the intimate (‘in respect of ’) form of
specification associated with the determinable/determinate relation as involving
increased specification along one or more determination dimensions of the deter-
minable; and second, the supposition that diverse physical realizers of a quali-
tative conscious state cannot differ from each other along mental determination
dimensions of the conscious states at issue, contra the ‘in respect of ’ feature of the
determinable/determinate relation.
My response proceeds by providing empirical reason to think that what deter-
mination dimensions are associated with a given determinable conscious state
(perceived color, pain) is science-relative, in such a way that we can make sense
of such states as having purely ‘psychological’ determination dimensions relative
to certain sciences (e.g., ‘normal’ color science or some branch of psychology),
and as having physical determination dimensions relative to other sciences (e.g.,
metameric color science or pharmacology). Such relativization makes in-principle
room for multiple physical realizers of a qualitative conscious state to have explic-
itly physical as well as psychological determination dimensions. I then use Powers-
based Determination to fill in how this might be, from a metaphysical point of view.
Take perceived color, for example. Hue, saturation, and brightness suffice
to characterize perceived colors in normal light conditions, as experienced by
normally sighted creatures more or less like us. Interestingly, however, things
that appear to be the same color under normal light conditions may appear
to be different colors under different light conditions. The explanation for this
phenomenon—‘metamerism’—has to do with broadly physical features of the
objects and light at issue.1⁸ Most notably, what color we perceive an object to be
will be a function of the spectral power distribution (SPD) of the light hitting
the retina, specifying the power of the light at each wavelength in the visible
1⁸ See, e.g., Wandell 1993.

spectrum; this SPD is itself a function of the SPD of the light incident on the
object’s surface, and the surface reflectance properties of the object. Different SPDs
of light hitting the retina may give rise to the same ‘tristimulus values’ (effectively:
hue, saturation, and brightness); hence it is that samples that appear the same in
normal light conditions may appear different in other conditions, and that samples
that appear different in normal light conditions may appear the same in other light
conditions. Metamers are color appearance properties that are individuated, in
part, by the relevant broadly physical features—let’s assume these are the retinal
SPDs—needed to accommodate the phenomenon of metamerism, such that color
appearance properties not distinguished by hue, saturation, and brightness are
distinguished by the relevant broadly physical features.
As I argue in detail in Wilson 2009, metamers are reasonably taken to be
colors—just colors seen (no pun intended) through a lens of a finer and partly
physical grain. Here I’ll just mention two considerations in favour of this claim.
First, metamers are part of the broader field of color science, and they are charac-
terized as colors in that science. In particular, color science is not concerned only
with colors as individuated by hue, saturation, and brightness.1⁹ On the contrary,
considerable color research is aimed at understanding colors as individuated by
broadly physical features such as retinal SPDs, as relevant to digital photogra-
phy, screen displays, car interiors, etc.2⁰ Second, the role that retinal SPDs play
in this research appears to be compatible with taking colors to be themselves
characterized by the relevant broadly physical features. After all, as above, colors
are understood as perceptual properties; and retinal SPDs are clearly part of the
process of color perception—in particular, retinal SPDs are input into the color-
sensitive cones, which then output the tristimulus values. Whether or not the
input/output function here is causal or rather ‘filter-like’, in any case there seems
to be no in-principle barrier to characterizing (specific kinds of) colors in terms
of the broader process of visual perception—especially since the broader process
and associated features are required to fully characterize color appearances (in
particular, in light of metamerism).
Similar remarks apply to other qualitative conscious states, such as pain. For
example, Funkhouser (2006) suggests that states of pain have mental determi-
nation dimensions along lines of feel and intensity; but it seems reasonable to
suppose that pains that are exactly similar with respect to these psychological
determination dimensions might be furthermore physically specified, as sciences
such as pharmacology take for granted.
I draw two morals from these sorts of case studies. First is that qualitative
conscious states, such as states of perceiving colors or experiencing pains, may have
1⁹ See, e.g., Wyszecki and Styles 1982. 2⁰ See, e.g., Judd and Wyszecki 1975.
consciousness 249
physical as well as psychological determination dimensions. Second is that deter-

mination dimensions may be science-relative: different sciences may treat a given
determinable feature as having different determination dimensions, effectively
characterizing the determinable feature in more or less fulsome ways. Relative
to normal appearance color science, color has determination dimensions of hue,
saturation, and brightness; relative to metameric color science, color has further
determination dimensions. Note that there’s nothing mysterious about taking
determination dimensions to be science-relative, from a physicalist perspective
of the sort we are presently considering. That different sciences may treat the
same determinable as having different determination dimensions simply reflects
that different sciences and their associated laws may treat certain phenomena at
different levels of metaphysical grain.
The observation that contemporary science suggests that conscious mental
states may have physical as well as psychological determination dimensions pro-
vides a wedge for responding to the objection from mental multiple realizability.
Still, the question remains: can we make metaphysical sense of this, compatible
with qualitative conscious states’ being determinables of lower-level physical
determinates?
Here’s where Powers-based Determination comes in.21 The basic idea is that the
phenomenon of a given determinable’s being associated with increasingly fine-
grained determinable dimensions, relative to a given science, can be straightfor-
wardly understood in terms of nonconjunctive specification of a determinable’s
powers. Relative to purely psychological determination dimensions, reflecting
sensitivity to the set of powers associated with a multiply realized determinable,
determinates of the determinable may be exactly alike. Relative to a finer-grained
set of determination dimensions, reflecting sensitivity to powers additionally
possessed by the physical realizers of the determinable, determinates of the deter-
minable will not be exactly alike. On this approach, what it is for determinates
to determine a determinable ‘in respect of ’ a determinable in the sense relevant
to ensuring the distinctive form of specification at issue in the determinable/
determinate relation is cashed in terms of the determinates’ sharing the powers of
the determinable (associated with the psychological determination dimensions),
but differing with respect to powers going beyond these (associated with the more
fine-grained determination dimensions encoding physical as well as psychological
distinctions), where the additional powers associated with any given realizer
do not themselves form a property. All this is compatible with the multiple
realizers’ being exactly similar with respect to the determinable’s psychological
21 As I discuss in my (2009), there are likely other approaches to the determinable/determinate

relation (as per, e.g., the proposal set out in Funkhouser 2006) that can make sense of the science-
relativity of determination dimensions.
determination dimensions in virtue of sharing, at both the type and token levels,
the powers had by the determinable, while differing in respect of other powers—
and correspondingly, with qualitative conscious mental states being Weakly emer-
gent, as per a determinable-based account of realization.
Accommodating mental superdetermination

Recall that the concern at issue in the objection from mental superdeterminates
is that taking qualitative conscious features to be realized in determinable-based
fashion by lower-level physical features is incompatible with the intuitive possi-
bility of there being qualitative mental superdeterminates, since implying, falsely,
that these could be further determined.
My response again starts with the observation that different sciences may
treat a single determinable as having different determination dimensions, such
that mental features may be superdeterminate relative to a purely psychological
science, while being further determined relative to a lower-level science. What is
superdeterminate relative to one science may not be superdeterminate relative to
another.
This much provides a wedge for responding to the objection from mental
superdeterminates. Still, the question remains: can we make metaphysical sense of
this, compatible with qualitative conscious features being determinables of lower-
level physical determinates?
Again, Powers-based Determination provides a comprehensible metaphysical
basis for accommodating the phenomenon at issue. What counts as a superdeter-
minate depends on which determination dimensions are at issue. Relative to one
set of determination dimensions, reflecting sensitivity to powers associated with
the determinable set, a given qualitative conscious state might be characterized as
a superdeterminate. Relative to a finer-grained set of determination dimensions
(reflecting sensitivity to powers in relevant supersets of the determinable set) that
same feature might not be appropriately characterized as a superdeterminate.
All this is compatible both with certain qualitative conscious features
being superdeterminate with respect to certain psychological determination
dimensions, but being further determinable with respect to further, partly
physical determination dimensions—and so with qualitative conscious mental
features being Weakly emergent, as per a determinable-based account of
realization.
Let’s sum up the results of this chapter. Consciousness—more specifically, the

sort of qualitative or ‘what it’s like’ mental features characteristic of conscious
consciousness 251
experience—has been frequently offered up as the best candidate for an actual

Strongly emergent phenomenon. But, I’ve argued, neither the knowledge argu-
ment(s) nor the conceivability argument establishes that qualitative conscious
features are actually physically unacceptable, and so they provide no basis for
moreover taking such features to be actually Strongly emergent. While the Strong
emergence of qualitative consciousness remains an open empirical possibility,
the presently best case for the metaphysical emergence of such consciousness
is one according to which, due to the irreducibly determinable nature of qual-
itative features of conscious experience, such features are Weakly emergent, by
lights of a determinable-based account. As I’ve argued, such an understanding of
qualitative conscious features can be defended against various objections, given
a proper appreciation of the science-relativity of determination dimensions and
a proper understanding of the determinable/determinate relation, as per Powers-
based Determination. I conclude that consciousness—more specifically, qualitative
conscious experience—is at least Weakly emergent.
8
Free will
Free will (or free agency), if such there be, involves the ability to mentally choose
an outcome (an intention to φ, or a φ-ing), where the outcome is ‘free’ in being,
in some substantive sense, up to the agent of the choice. Free will has often been
taken to be core to what it is to be a person, of either human or other variety,
in part (though not exclusively) because such agency seems to be prerequisite for
persons’ being autonomous in the way seemingly crucially relevant to achieving
certain moral, aesthetic, and other goods. In this chapter, I consider reasons for
thinking that free will, as it actually exists, might be either Weakly or Strongly
emergent.
I start by drawing on Bernstein and Wilson 2016 in order to set up a useful
framework for investigating into whether free will is metaphysically emergent.
Recall, to start, that the schemas for Weak and Strong emergence were initially
motivated as associated with two specific responses to the problem of higher-
level causation; namely, nonreductive physicalism and Strong emergentism. A
common focus of this problem concerns the status of mental features; but in the
usual case the mental features at issue are qualitative or intentional features, for
which free choice is supposed not to be at issue. More generally, debates over the
status of free will have tended to proceed in relative independence from debates
over the status of mental features whose governance by natural law is taken for
granted. As Bernstein and I argue, however, the problematics underlying the free
will and the mental causation debates are appropriately seen as special cases of a
more general problem, concerning whether and how mental features of a given
type may be efficacious, qua the types of feature they are (qualitative, intentional,
freely deliberative), given their apparent causal irrelevance—i.e., apparent failure
of distinctive efficacy—for effects of the type in question. (The literature I will be
discussing typically focuses more specifically on mental events of a given type, so
I will present the parallel in these terms.)
That the free will and mental causation debates can be seen as special cases of a
more general problem serves to suggest certain parallels between positions in the
respective debates, which parallels are useful for purposes of assessing whether
free will is either Weakly or Strongly emergent. In particular, as Bernstein and I
argue, a representative range of compatibilist (or ‘soft determinist’) accounts of
DOI: 10.1093/oso/9780198823742.003.0008
free will 253
free will1 implement a strategy that is structurally similar to the ‘proper subset’
strategy that, I have argued, is core and crucial to nonreductive physicalist accounts
of realization, and more generally, to Weak emergence. Here I present and extend
this result, arguing that the compatibilist’s proper subset strategy is reasonably
taken to indicate that compatibilist free will, were it to exist, would be Weakly
emergent. I then argue that a representative range of libertarian treatments
of free will are appropriately seen as committed to such agency’s involving a
fundamentally novel power, such that libertarian free will, were it to exist, would be
Strongly emergent. Against this backdrop, I go on, in the final section, to consider
the prospects of there being metaphysically emergent free will, suggesting that
free will of the compatibilist, Weakly emergent variety is common, and providing
a new argument for there moreover actually being free will of the libertarian,
Strongly emergent variety.
8.1 The generalized problem of mental quausation
In our (2016), Bernstein and I argue that (certain understandings of) the problems
of free will and of mental causation can be seen as special cases of a more
general problem, concerning whether and how mental events of a given type may
be efficacious, qua the types of feature they are—qualitative, intentional, freely
deliberative, and so on—given their apparent causal irrelevance for effects of the
type in question.
In making this connection, we generalize what Horgan (1989) calls “the prob-
lem of mental quausation”.2 As Horgan presents it, this problem is a refinement
of the problem of mental causation (a special case of the problem of higher-
level causation presented in Ch. 2, §2.1.1). The problem of mental causation is
sometimes pitched as the problem of how a (real, distinct, cotemporally materially
dependent) qualitative or intentional mental event M might be efficacious at all,
given that any effect E it might be seen as causing is, by Physical Causal Closure,
already caused by the physical event P upon which M depends. But as Horgan
notes, there is a quick route to gaining M’s efficacy—namely, via identifying M
with P, as per the usual reductive physicalist response—that leaves open what is
arguably the deeper question underlying the original problem:
1 In general, compatibilists maintain, as Lewis (1981) puts it, that “soft determinism may be true”
(113). Beyond this commitment, individual compatibilists may aim just to carry out the ‘negative’
project of undermining arguments for incompatibility, or in addition to provide a positive conception
of the compatibility at issue (see McKenna and Coates 2008 for discussion). Here by ‘compatibilism’ I
mean ‘positive compatibilism’.
2 Though Horgan’s (and our) focus is on mental goings-on, the more general problem of ‘quausation’
can be seen as applying to any seemingly higher-level goings-on.
Even if individual mental events and states are causally efficacious, are they
efficacious qua mental? I.e., do the mental types (properties) tokened by mental
events and states have the kind of relevance to individual causal transactions
which allows these properties to figure in genuine causal explanations? (47)
What is needed to adequately vindicate the efficacy of the mental, Horgan suggests,
is that mental events be shown to be distinctively efficacious: efficacious qua
mental—and more specifically, ‘qua F ’, where “F is schematic for a specific type of
mental event” (50). As such, it is mental ‘quausation’, not mental causation per se,
that is most deeply challenged by the possible truth of Physical Causal Closure. To
wit: how can a mental event M be efficacious vis-à-vis an effect E, in virtue of being
qualitative or intentional, given that E is causally determined by physical events
and associated laws in ways that seem to preclude M’s being causally relevant in
either of these (qualitative or intentional) respects?
Though Horgan’s discussion targets mental events for which freedom is not at
issue, the deeper concern about whether and how mental events can be seen as
causally relevant—that is, efficacious in virtue of, or qua, the distinctive mental
types they are—lies also at the core of what is sometimes called ‘the consequence
argument’, ‘the problem of free will and determinism’, or just ‘the problem of free
will’. This problem highlights a seeming tension between an intuitive conception
of free will as involving the ability to freely choose an outcome φ, and the broadly
scientific thesis of Determinism, according to which every event is a consequence
of the laws of nature and the state of the world at any time.3 Determinism admits
of different interpretations;⁴ what is at issue here is a reading involving causal
determination of present or future events by prior states (broadly construed), as
follows:⁵
3 One might wonder whether Determinism can be ruled out of court, on grounds that quantum
mechanics, which is likely true, is indeterministic. This would be premature, however, both because
there are deterministic interpretations of quantum mechanics (see, e.g., Bohm 1952), and because the
present incompatibility of quantum mechanics and general relativity suggests that we are not quite
yet in position to infer to the likely truth of quantum mechanics. (See O’Connor 2005, 4 for similar
observations.) Of course, even if Determinism is actually false, one might be interested in understanding
what bearing the truth of Determinism would have on free will. More importantly, the problem of free
will doesn’t crucially depend on the assumption of Determinism, since as I’ll discuss in what follows,
concerns about the causal relevance of free will also arise if the outcomes of seemingly free choices
are the product of indeterministic (e.g., quantum) laws. The focus on Determinism is traditional,
and convenient for setting up the parallel with the problem of mental quausation, but is ultimately
dispensable.
⁴ Steward (2015) distinguishes between interpretations based on entailment between propositions
at one time by propositions at another time, and interpretations based on causation between events.
⁵ Our focus here thus excludes compatibilists whose rejection of causation or causal production
between events—Leibniz, and perhaps also Hume—would entail their rejection of a causal reading of
Determinism.
free will 255
Causal Determinism: With the exception of any first events there might be, every
event is a causal consequence of the laws of nature and the state of the world at
any prior time.
The possible truth of Causal Determinism leads to a question: If every event E (e.g.,
an intention to φ, or a φ-ing) purportedly caused by a mental event of free choosing
M is, by Causal Determinism, a causal consequence of the laws of nature and prior
states, what causal role is left for M to play, vis-à-vis E? The answer, according to (a
causal reading of) the consequence argument, is that no role is left. As van Inwagen
(1983) puts it:
If determinism is true, then our acts are the consequences of the laws of nature
and events in the remote past. But it is not up to us what went on before we
were born; and neither is it up to us what the laws of nature are. Therefore, the
consequences of these things (including our present acts) are not up to us.⁶ (16)
The question and prima facie negative answer constitute what van Inwagen calls
“the problem of free will and determinism”, and what Bernstein and I call, for short,
‘the problem of free will’.
In the problem of free will, it is not the mere efficacy of deliberative mental
events (as, perhaps, the most proximate causes of intentions or actions in a causally
determined chain of events) that is at issue; indeed, that events of choosing are
efficacious is typically taken for granted in the free will debates. Rather, what
is at issue is whether events of choosing can be efficacious qua free—again, in
eventuating in outcomes that are, in some substantive sense, up to the agent—
under circumstances where Causal Determinism is presumed to be true. How
could an event of mental choosing M be efficacious vis-à-vis an event E, in virtue
of being free, given that E was causally determined by laws of nature and events
antecedent to M in ways that seem to preclude M’s being causally relevant in this
respect?
This parallel suggests that the problem posed by Physical Causal Closure for
qualitative and intentional mental events, and the problem posed by Causal
Determinism for events of seemingly free choosing, are each instances of a suitably
general problem of mental quausation:
How can a mental event M of a given type be efficacious vis-à-vis an event E, in

virtue of being the type of mental event it is, given that there is reason to think
that events of M’s type are causally irrelevant to the production of events of E’s
type? (Bernstein and Wilson 2016, 314)
⁶ Van Inwagen’s remarks are directed at an entailment-based reading of Determinism, but they
apply, mutatis mutandis, to a causal reading (simply insert ‘causal’ before ‘consequences’).
Note that in drawing this parallel, Bernstein and I are not suggesting that there
is any interesting parallel between the problem of free will and the unrefined
problem of mental (more generally, higher-level) causation; nor are we suggesting
that the problem of free will is an instance of the problem of mental quausation
as applied to qualitative or intentional mental events; after all, there is no clear
tension between Causal Determinism and mental quausation involving mental
causes for which freedom is not at issue. Our point is simply that both the problem
of free will and the refined problem of mental causation (again, typically directed
at qualitative and intentional mental events for which free will is not at issue)
are specific cases of a suitably general problem of mental quausation, whereby
the causal relevance (distinctive efficacy) of a mental event of a distinctive type—
whether intentional, qualitative, or freely deliberative—is called into question by
the holding of certain theses (Physical Causal Closure, Causal Determinism) which
are to some extent open possibilities for being true.⁷
The traditional responses to the problem of free will may be categorized by
reference to the following free will conditional:
If all events are subsumed by deterministic natural laws, then free mental
quausation—the causation of events (e.g., intentions to φ, φ-ings) by mental
choosings qua free—does not exist.
Hard determinists take the conditional to be true, on grounds that both antecedent
and consequent are true; libertarians take the conditional to be (vacuously) true,
on grounds that the antecedent is false; soft determinists or ‘positive’ compatibilists
(henceforth, just ‘compatibilists’) reject the conditional, on (the usual) grounds
that the antecedent is true but the consequent is false. Similarly, the traditional
responses to the refined problem of mental causation can be categorized by
reference to the following mental causation conditional:
If all physical events are subsumed by physical laws, then qualitative and inten-
tional mental quausation—the causation of physical events (e.g., bodily move-
ments) by mental events qua qualitative or intentional—does not exist.
Eliminativist physicalists and epiphenomenalists take the conditional to be true,

on grounds that both the antecedent and consequent are true; Strong emergentists
and substance dualists take the conditional to be (vacuously) true, on grounds that
the antecedent is false; reductive physicalists and nonreductive physicalists/Weak
⁷ Again, though it is convenient to present the parallel in terms of the problem for free will posed
by the assumption of Determinism, concerns can also be raised (as I’ll discuss down the line) for the
causal relevance of free will on the assumption that free choices are the product of indeterministic laws.
free will 257
emergentists reject the conditional, on grounds that the antecedent is true but the
consequent is false.
Since the problem of free will and the refined problem of mental causation may
each be seen as special cases of a more general problem, we might expect there
to be parallels between the primary responses to each problem, corresponding
to parallels in the stances taken towards the components of the corresponding
conditional. As I will argue in the following two sections, this is indeed the case
for compatibilism and libertarianism, respectively.⁸
8.2 Compatibilism and Weak emergence
According to compatibilism, the free will conditional is false, since the antecedent
(Causal Determinism) is true (or might be true), but the consequent (denying
the existence of free will) is false. Here two positions—reductive and nonreduc-
tive physicalism—take the same stance as regards the refined mental causation
conditional. As previously discussed, however, reductive versions of physicalism,
which identify the mental events at issue with physical events, face immediate
difficulties in making sense of how mental events can be efficacious qua the
qualitative or intentional types they are, with the usual strategies (as per the
deflationary accounts discussed in Ch. 3, §3.1, §3.2) being to offer pragmatic or
purely epistemic accounts of the desired ‘higher-level’ efficacy. Given this, and
since the deeper concern at issue in the problem of free will is with whether there is
a genuinely metaphysical basis for free mental quausation, we can cut to the chase
of considering whether there is an interesting parallel between compatibilism and
nonreductive physicalism/Weak emergentism.
8.2.1 The compatibilist’s proper subset strategy
At this point, the reader is well familiar with the characteristic strategy of the
nonreductive physicalist’s response to the problem of higher-level quausation, as
encoded in the schematic conditions in Weak Emergence:

⁸ There is also a parallel between hard determinism and eliminativism; see Bernstein and Wilson
2016 for discussion.
As setup for motivating the structural parallel with compatibilism, it is worth

recalling how satisfaction of these conditions—most crucially, satisfaction of the
Proper Subset of Powers Condition—operates to secure, not just the reality and
efficacy of the higher-level goings-on, but also their causal relevance/distinctive
efficacy. Again, the core idea is that there are two ways for a higher-level feature S
to be distinctively efficacious with respect to its base feature P. One way is for S to
be associated with a new power—a power that P doesn’t have, or doesn’t have in
the same (direct) way as S; this is the form of distinctive efficacy at issue in Strong
emergence. A second way reflects that S is associated with a distinctive power profile
consisting of a proper subset of the powers associated with P; this is the form of
distinctive efficacy at issue in Weak emergence.
Recall also that there are (at least) two strategies which might be used to
motivate taking the having of a distinctive power profile conforming to Weak
emergence to provide a basis for S’s being efficacious qua the type of higher-
level feature it is: first, by observing that such a power profile provides a basis
for difference-making or ‘proportionality’ considerations (where the effect would
still have occurred and been caused by S even had S been differently realized);
second, by observing that such a power profile may be associated with systems
of laws or causal joints that are comparatively abstract, in being insensitive to
certain details needed for the lower-level physical laws governing P to operate.
Applied to the qualitative or intentional mental events associated with the problem
of mental quausation: even if every token power of a given such mental event M,
on a given occasion, is identical with a token power of its physical realizer P on that
occasion, M can be distinctively efficacious—causally efficacious qua qualitative or
intentional—in virtue of M’s distinctive power profile’s tracking either difference-
making considerations, or the comparatively abstract system(s) of laws or causal
joints associated with (non-agential) psychology.
Interestingly, as Bernstein and I argue, a representative range of seemingly
diverse compatibilist accounts implement a structurally similar ‘proper subset’
strategy for responding to the problem of free will. To prefigure: in the case of
compatibilist free will, the operative strategy involves characterizing events of
seemingly free choosing as associated with only a proper subset of the causal
determinants of the outcome (effect) at issue, in such a way as to provide a
principled basis for the claim that the choosing is efficacious qua free. Hawthorne
and Pettit’s (1996) taxonomy of compatibilist strategies serves as a useful basis of
operations for establishing this result. They start by noting:
All compatibilists agree that every choice has antecedents and […] that this
fact puts freedom of choice in doubt. How can a choice be made freely if it
is the product of independent antecedents? The response they make is that
some possible antecedents are better than others from the point of view of free
free will 259
choice and that a choice is free to the extent that its antecedents, or at least its
relevant antecedents, satisfy the inherently vague condition of leaving it up to the
agent. (191)
In schematic form:
X chooses freely to φ if and only if the relevant antecedents of the choice leave
the φ-ing up to X. (191)
As above, for present purposes a choice to φ (the outcome of an event of choosing)

may be either an intention to φ, or a φ-ing. Hawthorne and Pettit identify three
main compatibilist accounts of what it is for a choice to be “up to an agent”,
associated with the notions of freedom as underdetermination, ownership, and
responsibility, respectively. The accounts vary to some extent as regards which
antecedents are supposed to be relevant to establishing whether the choice was
up to the agent, in the intended sense. As Bernstein and I argue, however, each of
these accounts plausibly imposes satisfaction of a certain ‘proper subset’ condition,
as key to the associated strategy of response to the problem of free will.
I start by motivating the condition. First, reflecting endorsement of (the open
possibility of) Causal Determinism, the compatibilist accepts the following condi-
tion on the causal antecedents of any outcome of a free mental deliberation:
Causal Antecedents Condition: The total causal antecedents of an event of free

choosing M completely determine the outcome of M (e.g., a choice to φ).
As a first pass, the compatibilist strategy requires that a free mental choosing M
satisfy the following proper subset condition:
Subset of Causal Antecedents Condition (first pass): The relevant causal

antecedents {C} of a free mental choosing M are (i) a non-empty proper subset
of the total causal antecedents of M, which (ii) satisfy the condition of leaving
the outcome of M up to the agent.
Moreover, as will shortly become clear, if the compatibilist’s strategy of identifying

the relevant causal antecedents of M is to make sense of the idea that these
antecedents leave the choice up to the agent, then the relevant antecedents must
more specifically satisfy the following (final pass) proper subset condition:
Proper Subset of Causal Antecedents Condition: A free mental choosing M result-

ing in a choice to φ satisfies the following: (i) M has relevant causal antecedents
{C} which are a non-empty proper subset of the total causal antecedents of M, and
(ii) it is possible that a choosing M′ of the same type as M occur, having relevant
antecedents {C′ } of the same type as {C}, but where the total antecedents of M′
are such as to completely determine the outcome of M′ ⁹ as either a choice not to
φ or as the absence of a choice to φ. (Bernstein and Wilson 2016, 325)
After sketching Bernstein’s and my arguments that the following three com-
patibilist accounts aim to satisfy this condition, I will go beyond the previous
discussion, to make the case that the compatibilist’s Proper Subset of Causal
Antecedents Condition is not just structurally similar to, but can moreover be
plausibly understood as entailing, the Proper Subset of Powers Condition operative
in the schema for Weak emergence. This will set up for the later discussion
of whether compatibilist free will (hence persons characterized as having such
agency) is Weakly emergent.
Freedom as underdetermination
On underdetermination accounts, a choice to φ is the result of a free choosing M
iff M could have resulted in a choice not to φ. How could this be, given that the
choice to φ was determined, as per the Causal Antecedents Condition (encoding
Determinism)? The underdetermination approach proceeds by identifying a subset
{C} of the causal antecedents of the choice to φ, relative to which it was left open
whether or not M would result in a choice to φ. As Hawthorne and Pettit (1996)
put it:
Taken as a whole, the antecedents of any choice will necessitate that choice under
a deterministic picture and compatibilists of this stripe must take the relevant
antecedents to be a subset of the totality. But which subset? (193)
The relevant subset of antecedent events will include the choosing event, along
with events tracking the standing beliefs and desires of the agent at the time of
choosing, and events tracking whether the choosing took place under conditions
of physical restraint, threat, etc. (cf. Ayer 1954). From broader perspectives, the
relevant antecedents might also include events tracking cultural influences, past
trauma, or other psychological, social, psychiatric, neurological, etc., conditions
holding of the agent. In general, Hawthorne and Pettit note,
The line will be that an agent is free to the extent that the antecedents that can or
have to be countenanced in that perspective leave the choice underdetermined.
[…] To be free, if you like, is to be free relative to that stance. (193–4)
⁹ As per the Causal Antecedents Condition; the possibility of indeterministic scenarios is put aside
here.
free will 261
Here the relevant antecedent events must be a proper subset of the causal
antecedents that, as per the Causal Antecedents Condition, completely determine
the choice, since only if the subset is proper is there any hope that the subset of
antecedents will leave that choice underdetermined. Moreover, the assumption
that the subset of antecedents leaves the choice underdetermined plausibly entails
(indeed, has as its content) that it is possible that a choosing M′ of the same
type as M occur, having relevant antecedents C′ of the same type as M’s relevant
antecedents C, but where the total antecedents of M′ are such as to completely
determine a different outcome, as either a choice not to φ or the absence of a
choice to φ.
This last entailment just is the Subset of Causal Antecedents Condition. As such,
a freedom as underdetermination form of compatibilism explicitly implements a
proper subset strategy, characterizing a mental choosing M as associated with a
proper subset of its causal antecedents, then using this association to accommodate
M’s being causally relevant/distinctively efficacious qua free vis-à-vis the ensuing
choice.
Freedom as ownership
A second compatibilist approach takes freedom to be a matter of ownership:
The ownership line takes a choice to be up to an agent to the extent that it is not
due to anyone or anything other than the agent themselves; it is a choice that the
agent owns, a choice with which the agent identifies, and not something forced
upon them. Suppose that the relevant antecedents in the adjudication of free will
are taken to be […] beliefs and desires. [Then] an agent φs freely just in case
their beliefs and desires combine to lead—or at least lead in ‘the right way’ (see
Davidson 1963) to their φ-ing. (Hawthorne and Pettit 1996, 194)
Underdetermination by the relevant antecedents is not explicitly required here,

since an agent could own or identify with completely determined intentions. But,
Hawthorne and Pettit argue, if the ‘ownership’ line is to be viable, it will have to
ensure underdetermination by these antecedents (hence, Bernstein and I maintain,
satisfaction of the Subset of Causal Antecedents Condition).
Why so? To start, note that if, for example, I am brainwashed with beliefs and
desires leading to my choice to φ, this intention cannot be seen as the effect of
free mental choosing. A well-known response (see Frankfurt 1971) requires that
choices result from desires that the agent X desires, at the second order, to have and
be moved by. The brainwashing problem will re-arise, however, unless “the action
issues from desires that the agent has some measure of second-order control over”
(Hawthorne and Pettit 1996, 195). As O’Connor (2005) puts it:
We can […] imagine external manipulation consistent with Frankfurt’s account

of freedom but inconsistent with freedom itself. […] one might discreetly induce
a second-order desire in me to be moved by a first-order desire—a higher-order
desire with which I am satisfied—and then let me deliberate as normal. Clearly,
this desire should be deemed “external” to me, and the action that flows from it
unfree. (12)
These considerations indicate that one needs to ensure, somehow, that the for-
mation of second-order desires is up to the agent; and the natural compati-
bilist approach will be to restrict attention to a proper subset of the antecedents
determining the desire—e.g., those relevant to whether the agent’s choosing was
constrained by other persons, or by other psychological, social, psychiatric, neu-
rological, etc., conditions. In other words, to accommodate the needed control
of second-order desires on a ‘freedom as ownership’ picture, a proper subset of
the antecedents of the choice to φ must be specified, relative to which it was
underdetermined that the agent had the second-order desires they had; hence
underdetermined that the agent would identify with the first-order beliefs and
desires leading to the agent’s choice to φ; hence underdetermined that the agent’s
choosing would result in a choice to φ. Such underdetermination in turn entails
(indeed, has as its content) satisfaction of the Subset of Causal Antecedents Condi-
tion.
More generally, here again a proper subset strategy is implemented, whereby a
mental choosing M is associated with a proper subset of its causal antecedents, in
service of making room, as per the Subset of Causal Antecedents Condition, for M to
be causally relevant/distinctively efficacious qua free vis-à-vis the ensuing choice.
Freedom as responsibility
Lastly, on the ‘freedom as responsibility’ approach, a choice to φ is the result of a
free choosing M iff the agent of the choosing could be held responsible for the out-
come of M. The criteria for an agent’s being responsible might advert to prevailing
systems of law and morality, or (as per Strawson 1962) to the participant or reactive
attitudes characteristic of human interactions. As Hawthorne and Pettit point out,
this approach again requires that the choice at issue be underdetermined:
To hold an agent responsible in certain choices is to think that it is not inevitable

either that they get things right or that they get them wrong—either that they do
well or that they do ill—and so it is to believe that there is a sense in which they
could have done otherwise […]. (197)
The relevant antecedents in this case would then include those relevant to deter-
mining whether the agent was deliberating under conditions where they would, by
free will 263
the lights of the prevailing system of law, morality or interaction, be held respon-
sible for the outcomes of their choosings. Once more, these might cite events
or states tracking whether various physical, psychological, neurophysiological,
etc., conditions or constraints were in place antecedent to or concurrent with the
choosing.
Here again, the account identifies a relevant proper subset of M’s causal
antecedents, then requires that, relative to these antecedents, the outcome of
M could have been different, as a way of making sense of M’s being causally
relevant/distinctively efficacious qua free vis-à-vis the ensuing choice. In other
words, a freedom as responsibility account implements a proper subset strategy,
encoded in satisfaction of the Subset of Causal Antecedents Condition.
8.2.2 Deepening the parallel: a powers-based interpretation

of the compatibilist’s proper subset condition
Nonreductive physicalist and compatibilist positions thus each respond to the

respective problems of mental quausation by characterizing the mental events at
issue as associated with a proper subset of the ‘causal determinants’ of their associ-
ated effects. As Bernstein and I present the structural similarity, the determinants
are not the same: in the one case, these are powers; in the other, these are causal
antecedents. This much establishes a structural similarity in strategies: in each case,
associating the mental event M with (only) the relevant proper subset of causal
determinants provides a basis for showing that M is causally relevant/distinctively
efficacious vis-à-vis the production of the effect E in question, qua the type of
mental event M is.
As I will now argue, however, the parallel is even deeper: the compatibilist strat-
egy can be understood as entailing the holding of a proper subset relation between
token powers associated with two complex, cotemporal events, corresponding to,
first, the mental choosing M in combination with the relevant antecedents of M
(call this complex event C′ ), and second, the mental choosing M in combination
with the total antecedents of M (call this complex event C).
To start, note that there is no in-principle problem with associating a set of token
powers with either C′ or C. After all, as we have seen, the events, entities, processes,
or other goings-on at issue in debates over the status of higher-level entities are
often complex and typically spatially and temporally extended; and there is no in-
principle problem with assigning powers to such goings-on (reflecting, e.g., the
operative laws of nature). As such, we are in position to see that the compatibilist’s
proper subset strategy entails that the set of relevant causal antecedents C′ has
a proper subset of the powers of the entire set of causal antecedents C. For this
strategy plausibly entails that, on the one hand, every token power of C′ is identical
to a token power of C, but that C′ has fewer powers than C; and this is because,
while C has the power to result in a choice to φ, C′ does not have this power—for
C′ could occur in circumstances in which M (or an event of M’s type) would have
resulted either in a choice not to φ or in the absence of a choice to φ.
On this reconstruction of the compatibilist strategy, it would not be the seem-
ingly free choosing itself that would be appropriately deemed Weakly emergent,
but rather the seemingly free choosing in combination with the relevant proper
subset of antecedents. Such a result makes sense, on a compatibilist picture. Unlike
the case of a Weakly emergent qualitative feature, for example, the status of
the event of choosing as genuinely free requires that the choosing be associated
with causal antecedents that in the relevant sense leave the outcome up to the
agent; hence on a compatibilist view, the freedom of a given act of choosing can
be seen as constituted by the occurrence of a complex temporally and causally
extended event, consisting in the choosing in combination with the relevant causal
antecedents.
The previous considerations indicate that free will on a compatibilist account
can be seen as satisfying each of the conditions in Weak Emergence, and so as
Weakly emergent.
This result provides a new basis for addressing a frequently voiced concern
about the compatibilist’s strategy—namely, that identification of a given subset of
causal antecedents won’t make sense of how a mental choosing M could be free,
since the mere presence of a subset of antecedents doesn’t establish that M ‘selects’
or ‘determines’ the outcomes of the choosing.
To start, the compatibilist, like the nonreductive physicalist, will grant that
M’s distinctive efficacy vis-à-vis the effect at issue doesn’t proceed by way of
M’s having a distinctive power: just as Physical Causal Closure blocks taking a
qualitative or intentional mental event M to have a new power (that would be
Strong emergence, not physicalism), so too does Causal Determinism block taking
a mental event of choosing M to have a new power (that would be libertarianism,
not compatibilism). Even so, just as the nonreductive physicalist has alternative
ways of motivating the distinctive efficacy of qualitative and intentional mental
events—either as tracking difference-making considerations (if the physical real-
izer had been slightly different, I would still have been thirsty), or as tracking a
comparatively abstract psychological system of laws or causal joint—so too may
the compatibilist maintain that even in the absence of new powers to ‘select’ or
‘determine’ outcomes, M may be distinctively efficacious vis-à-vis those outcomes,
either in tracking difference-making considerations (if the causal antecedents of
my choice had been slightly different, I would still have chosen as I did) or as track-
ing a distinctive broadly psychological level of causal grain. The distinctive form
of causal relevance/distinctive efficacy identified by nonreductive physicalists—
namely, that encoded in the schema for Weak emergence—appears to be, mutatis
free will 265
mutandis, just what the compatibilist needs. Of course, compatibilism faces other
challenges (a point to which I will later return), but in any case it is worth noting
that the parallel to nonreductive physicalism is useful in clarifying just what the
compatibilist strategy for achieving the autonomy of free will is supposed to be.
8.3 Libertarianism and Strong emergence
I next turn to libertarianism (a.k.a. ‘incompatibilism’),1⁰ according to which both

the antecedent (Causal Determinism) and the consequent (denying the existence
of free will) of the free will conditional are false. I will argue that there is a parallel
between libertarianism, on the one hand, and Strong emergentism as standardly
directed at qualitative and non-intentional mental states, on the other.
8.3.1 The libertarian’s new power strategy
An initial point of similarity between libertarianism and Strong emergentism is

that each view rejects the broadly empirical thesis that causes trouble for the
supposition that the higher-level goings-on are causally relevant in the intended
sense. As Clarke and Capes (2017) put it:
To have free will is to have what it takes to act freely. When an agent acts freely—
when she exercises free will—it is up to her whether she does one thing or another
on that occasion. A plurality of alternatives is open to her, and she determines
which she pursues. When she does, she is an ultimate source or origin of her
action. So runs a familiar conception of free will.
Incompatibilists hold that we act freely in this sense only if determinism is false.
(1)
Similarly, Strong emergentists maintain that an appropriate understanding of the

efficacy of certain goings-on requires the falsity of Physical Causal Closure.
The deeper point of similarity, however, lies in the positive accounts given of
the existence and causal relevance/distinctive efficacy of the higher-level goings-
on at issue. Recall that the conditions in Strong Emergence, which takes its original
inspiration from the Strong emergentist strategy for responding to the problem
1⁰ Here again we can distinguish between incompatibilist views which simply deny the compatibility
of free will and Determinism, and those which moreover aim to give some positive account of the nature
of incompatibilist free will; at issue in this discussion are what we might call ‘positive incompatibilist’
views.
of higher-level causation, require that a Strongly emergent feature have a fun-

damentally novel power not had (or only indirectly had, in virtue of being a
sufficient precondition of the Strongly emergent feature) by the lower-level phys-
ical feature upon which it cotemporally materially depends, which novel power
in turn provides a principled metaphysical basis for both the ontological and
the causal autonomy of the Strongly emergent feature. As I will shortly argue, a
representative range of libertarian accounts are reasonably seen as committed to
free will’s satisfying these conditions, and in particular to taking free will to be
associated with a fundamentally novel power—namely, the power to freely choose
to φ, where (as previously) a choice to φ (an outcome of an event of choosing) may
be either an intention to φ, or a φ-ing.
In making this case, I’ll help myself to a commonly acknowledged tripartite
taxonomy of libertarian accounts, as falling under noncausal, event-causal, and
agent-causal varieties. As Clarke and Capes (2017) note:
The incompatibilist theories that have been offered fall into three main groups,
depending on which type of indeterminism (uncaused events, nondeterministi-
cally caused events, agent- [or substance-] caused events) they require. Further
variations among accounts concern where in the processes leading to decisions
or other actions they require indeterminism and what other conditions besides
indeterminism they require. (1, brackets in text)
(See also O’Connor 2005.) To fix ideas, I will primarily (though not exclusively)
focus on representative instances of the accounts, as proposed by Ginet, Kane,
and O’Connor, respectively. Each of these accounts can be seen as offering a
different positive account of what makes a given act of choosing free in the relevant
(strong, incompatibilist) sense, against the common backdrop assumption of the
rejection of Determinism. The case for libertarian free will’s being understood as
Strongly emergent is most straightforward for the two causalist accounts, so I’ll
start with those; I’ll then make a case that even so-called ‘noncausalist’ accounts
are plausibly committed to free will’s involving a (fundamentally) novel power, as
Strong emergence requires.
Before getting started, I register three observations that will sometimes enter
into what follows.
First, though it is more or less common ground that any variety of free will
worthy of the name (whether compatibilist or incompatibilist) has to make sense
of a choice to φ being in some sense ‘up to the agent’, the issue of such ‘agential
control’ is often interwoven with what sort of free will might make room for
moral responsibility, character, reasons, values, self-identification, conscious and
unconscious intentions, and so on—‘moral notions’, for short. This is even more
free will 267
the case in presentations or defenses of libertarian accounts, perhaps reflecting a

driving motivation of such accounts to do better by way of accommodating moral
notions than compatibilist or hard determinist accounts. My own view is that this
mixing of the metaphysics of free will with the question of how such agency bears
on moral notions is ill-advised. A more systematic approach, it seems to me, is to
start by getting clear about what options the libertarian has for accommodating
simple cases of seemingly free choosing, such as a case in which one considers
whether to throw a piece of chalk in the air and then seemingly freely determines
to do so (or not), leaving for later treatment the question of how the available
options comport with the deeply complex further issues of moral responsibility,
self-definition, values, and so on.
Second, and again by way of partial explanation, perhaps, of the common
focus on moral notions in discussions of libertarianism, an oft-stated concern
with libertarian accounts is that, insofar as they reject Causal Determinism, they
must be committed to thinking that the outcomes of free choices are some-
how indeterministically caused, via quantum or other ‘chance-y’ processes, in a
way that would undermine agential control and hence the associated basis for
accommodating moral notions. To be sure, some libertarians (e.g., Kane; see
the following subsection) embrace a kind of analogy between the indeterminacy
of free choice and quantum indeterminacy. It is important to realize, however,
that the libertarian is under no obligation to suppose that the only alternative
account to one on which choices are nomologically deterministic is one according
to which such choices are nomologically indeterministic. On the contrary, the
libertarian may endorse the view that agents are nomologically transcendently
free—that is, free, in at least some of their choices, from either deterministic or
indeterministic laws.
Third, some discussions of libertarianism take for granted that ‘event’ causation
and ‘agent’ causation are distinct varieties of causation; relatedly, discussions of
agent causation frequently import the assumption that such causation would be
causation by a ‘substance’, which in turn (some complain) is problematic or in any
case unusual. My own view is that there is no need to introduce a new variety
of causation, much less a new substance, in order to accommodate the sort of
nomologically transcendent freedom that agents possess (or should be thought
to possess), on a libertarian view.
Insofar as I believe that certain methodological presuppositions of the accounts
to follow are problematic, my discussion will at times depart from the letter
of certain libertarian views, and advance theses that to my mind do better at
metaphysically accommodating the spirit of libertarianism.
Event-causal accounts
Clarke and Capes (2017) describe event-causal libertarian accounts as follows:
Compatibilist accounts of free will are typically event-causal views, invoking

event-causal accounts of action. The simplest event-causal incompatibilist theory
takes the requirements of a good compatibilist account and adds that certain
agent-involving events that cause the action must nondeterministically cause it.
When these conditions are satisfied, it is held, the agent exercises in performing
her action a certain variety of active control (which is said to consist in the action’s
being caused, in an appropriate way, by those agent-involving events), the action
is performed for a reason, and there remains, until she acts, a chance of the agent’s
not performing that action.
Standard varieties of such accounts (a.k.a. ‘centered accounts’) locate the indeter-
ministic causation at issue in the immediate causal antecedents of the choice (as
opposed to some prior indeterministic process leading to certain beliefs becoming
salient or certain preferences being formed, which beliefs or preferences enter into
the process of deliberation).
Does the indeterministic causation here involve a fundamentally novel power?
Arguably, event-causal libertarians should think so, if they aim to decisively answer
the so-called ‘objection from luck’, to which such accounts are commonly seen as
vulnerable. As Clarke and Capes (2017) put the concern:
If a decision is nondeterministically caused, and if there remains until it occurs

a chance that the agent will instead (at that moment) make a different decision,
then there is a possible world that is exactly the same as the actual world up until
the time of the decision, but in which the agent makes the alternative decision
then. There is, then, nothing about the agent prior to the decision—indeed, there
is nothing about the world prior to that time—that accounts for the difference
between her making one decision and her making the other. This difference, then,
is just a matter of luck. And if the difference between the agent’s making one
decision and her instead making another is just a matter of luck, she cannot be
responsible for the decision that she makes.
Consider, for example, Kane’s centered causal account.11 Bracketing certain

nuances, Kane maintains that a choice to φ is one that involves a “self-forming
willing”—an indeterministically caused choice or other action for which the
agent is “ultimately responsible” (1996: 35).12 Of course, not all free choices
need be “self-forming”—here we see the unhelpful mixture of the more basic
metaphysical question of in what libertarian free will consists with the much
more complex question of how our free choices enter into self-constitution and
11 See also Nozick 1995, Ekstrom 2001, and Franklin 2018.

12 The primary nuance consists in allowing that a choice might be free even if causally determined,
so long as it at least partly resulted from a self-forming willing.
free will 269
moral responsibility. In any case, as Clarke and Capes (2017) point out, it remains
unclear how appeal to indeterministically efficacious “willings” is supposed to
answer the objection from luck—at least if the indeterminacy of will or effort
is, as Kane unwisely suggests, analogous to the sort of indeterministic processes
associated with quantum phenomena. Such an appeal not only raises the specter
of the objection from luck, but also problematically suggests that free will is, like
quantum goings-on, still caught in the net of nomological causation. That’s not
what a so-called ‘libertarian’ should say, or so it seems to me.
At any rate, there doesn’t seem to be any reason why an event-causal libertarian
such as Kane couldn’t rather maintain that the indeterminacy of agent-involving
events of free choosing reflects not an analogy with (much less a basis in) inde-
terministic quantum goings-on, but rather that agential control involves a funda-
mentally novel power to choose in a way that is not a matter of either deterministic
or indeterministic nomological processes. As the simple cases suggest, the novel
power here is not, at least in the first instance, essentially tied to morally loaded
notions involving character, morality, or the like. Rather, the power is simply the
power to choose, to make a choice to intend or to act, in a way that transcends
any nomological goings-on, and for whatever reason the agent happens to find
compelling in the moment,13 or indeed, for no reason at all (as in the case of an
illustrative throwing of a piece of chalk). To be sure, as embodied, persons are
subject to various physical and psychological limitations: to be capable of choosing
to intend or to act, in a way that is neither deterministically nor indeterministically
nomologically determined, does not mean that all bets, or all laws of nature, are
off. Still, as persons we are capable of choosing to φ in a way that is plausibly
fundamentally novel, insofar as any non-agents with which we are familiar do
not have such a power to freely choose. Or so an event-causal libertarian can and
should say.
I interpret Merricks (2003) as endorsing such a view. Merricks argues, as regards
conscious persons, that “we should say that some of what those objects cause,
in virtue of having those properties, lack microphysical causes” (110). Merricks’s
claim here is compatible with an event-causal account, and he is clear that he
sees exercises of free will as involving a power that lower-level physical goings-
on do not have—that is, as involving a fundamentally novel power, of the sort that
satisfaction of the schema for Strong emergence requires:
Sometimes my deciding to do such and such is what causes the atoms of my arm to
move as they do. Presumably my so deciding won’t ever be the only cause of their
moving. There will also be a cause in terms of microphysics or microbiology, in
13 Here an underappreciated resource might advert to the phenomenon of attention.

terms of nerve impulses and the like. But at some point in tracing back the causal
origin of my arm’s moving (if it is intended), we will reach a cause that is not
microphysical, that just is the agent’s deciding to do something. (110)
Merricks distinguishes his view from that of the British Emergentists, on grounds
that they “seem to explain being emergent in epistemic terms” (111, note 13).
However, as previously discussed, that the British Emergentists sometimes char-
acterized emergence in epistemic terms reflected their (incorrect) assumption that
certain epistemic failures would track fundamental novelty; properly understood,
Merricks and the British Emergentists are on the same metaphysical page. More
generally, in Merricks’s view, we have a form of event-causal libertarianism that
plausibly satisfies the conditions in Strong Emergence.
Agent-causal accounts
Clarke and Capes (2017) describe agent-causal libertarian accounts as follows:
On what are called agent-causal views, causation by an agent is held not to consist
in causation by events (such as the agent’s recognizing certain reasons). An agent,
it is said, is a persisting substance; causation by an agent is causation by such a
substance. Since a substance is not the kind of thing that can itself be an effect
(though various events involving it can be), on these accounts an agent is in a
strict and literal sense an originator of her free decisions, an uncaused cause of
them. This combination of indeterminism and origination is thought to capture
best the idea that, when we act freely, a plurality of alternatives is open to us and
we determine, ourselves, which of these we pursue, and to secure the kind of
freedom needed for moral responsibility. (23)
While in some cases talk of an object or other particular entity’s causing an

effect can be seen as shorthand for talk of the object’s having some feature that
produced the effect (as when, e.g., ‘the rock broke the window’ is more specifically
understood as ‘the rock’s having momentum M at the point of contact with the
window caused the window’s shattering’), the agent-causal libertarian denies that,
for the outcomes of free choices, such a reduction (e.g., to the agent’s recognizing
certain reasons) is available.
Concerns about agent causation often stem from concern about causes being
substances rather than events. One might reasonably deny, however, that making
sense of agent-causation requires thinking of agents as ‘substances’ or positing an
associated new form of causation, as opposed to thinking of agents as objects of a
distinctive variety, no more problematic than the distinctive kinds of objects (cells,
planets, plants, and so on) posited by the various special sciences, which are also
supposed to be capable of causing various effects. No doubt such special-science
free will 271
objects cause these effects in virtue of certain of their properties, as per usual. But
that needn’t be a problem for the agent-libertarian: if the properties relevant to the
exercise of free choice are along lines of ‘having nomologically transcendent free
will’, then there is no threat to the supposition that agents are capable of causing the
outcomes of their acts of free choosing. Again, if there is a salient causal difference
between agents and other kinds of objects, it has to do with whether or not a given
instance of causation is law-governed (deterministically or indeterministically) or
not; and this difference can be registered without taking agents to be substances or
to exercise a new form of causation.
In any case, on agent-causal accounts there is a clear sense in which free will
involves the exercise of a fundamentally novel power. O’Connor (2009a) is explicit
on this score, saying that
[A]n adequate account of freedom requires, in my judgment, a notion of a dis-

tinctive variety of causal power, one which tradition dubs “agent-causal power”.
(191)
The familiar considerations [against locating free will in indeterminacy] lead
certain philosophers to conclude that the kind of control necessary for freedom
of action involves an ontologically primitive capacity of the agent directly to
determine which of several alternative courses of action is realized. (193)
O’Connor is also explicit that the posit of such a power can be understood in terms
of emergence rather than substance dualism:
While the tidiness of substance dualism has its appeal, it is in fact optional for the
metaphysician who believes that human beings have ontologically fundamental
powers (whether of freedom or consciousness or intentionality). For we may
suppose that such powers are […] ontologically emergent powers, ones that are
at once causally dependent on microphysically-based structural states and yet
ontologically primitive, and so apt to confer ontologically primitive causal power.
(191)
What is the power at issue? As desired, the fundamentally novel power is one which
allows one to choose to act, or to intend to act, in one way rather than another:
One important feature of agent-causal power is that it is not directed to any

particular effects. Instead, it confers upon an agent a power to cause a certain type
of event within the agent: the coming to be of a state of intention to carry out some
act, thereby resolving a state of uncertainty about which action to undertake.
(195)
Agent causation is precisely the power to directly determine which of several

causal possibilities is realized on a given occasion. […] [T]he view posits a
fundamental, irreducible power of agents to form intentions. (211–12)
The posit of such a novel fundamental power is supposed to provide a basis for
responding to the luck objection, in virtue of positing “a kind of single-case form
of control by means of which the agent can determine what happens in each case”
(213). For present purposes, what is crucial is that O’Connor’s version of an agent-
causal form of libertarianism clearly satisfies the conditions in Strong Emergence.
Noncausal accounts
Clarke and Capes (2017) describe noncausal libertarian accounts as follows:
Some incompatibilist accounts require neither that a free action be caused by

anything nor that it have any internal causal structure. Some views of this type
require that a free action be uncaused; others allow that it may be caused as long
as it is not deterministically caused. Since any such account imposes no positive
causal requirement on free action, we may call views of this type “noncausal”.
As O’Connor (2005) puts it, on a noncausal account, free will is taken to be “entirely
noncausal in character and is instead a consequence of intrinsic, noncausal features
of the choice itself ”. There are several different conceptions of the intrinsic feature
at issue. On Ginet’s version (see, e.g., Ginet 1990 and 2002), this feature is an “actish
phenomenal quality,” which he describes (1990, 13) as its seeming to the agent as
if they are directly producing, making happen, or determining the event that has
this quality. On McCann’s (1998) version, the intrinsic feature at issue is ‘intrinsic
intentionality’. As Clarke (2003) describes McCann’s view:
[I]n making a decision, McCann maintains (1998, 163), one intends to decide—
indeed, one intends to decide exactly as one does (e.g., when one decides to A,
one intends to decide to A). One’s so intending, though intrinsic to the decision,
is not a matter of the content of the intention that is formed in deciding; nor is it a
matter of one’s having any further intention in addition to that formed in making
the decision. Rather, McCann holds, it is a matter of a decision’s being, by its very
nature, an act that the agent means to be performing. (18)
Finally, on Stump’s (1999) version, the intrinsic feature is one along Thomistic
lines, as involving an act of will or intellect:
free will 273
What is essential to moral responsibility on Aquinas’s view is that a person be the

ultimate source of what she does, that her intellect and will be the ultimate causes
of her acts. By ‘ultimate cause’ here, I mean that there is nothing which is prior to
that person’s acts of intellect and will and which causally determines her intellect
and will to be in the states in which they are. If we can trace a causal chain of any
sort backward from an agent’s act, then the causal chain must originate only in
acts of her will and intellect. That is, for any act which the agent does, if there is
any causal chain at all of which the act is the effect, then the causal chain must
have a first or ultimate cause, and that ultimate cause cannot be anything other
than an act of the agent’s own will or intellect. (414)
I will now argue that, on plausible construals of each of these variations on the
noncausal theme, the sense in which free will is ‘noncausal’ pertains only to the
purported failure of free choices to themselves be effects of prior (deterministic
or indeterministic) causes. This much is compatible, however, with free choosings
on such an account being ‘causal’ in the sense of themselves having effects, and
associated powers. To prefigure: this result, coupled with the supposition that the
powers themselves are fundamentally novel relative to the powers of dependence
base goings-on, supports taking free will on a noncausal account to conform to the
conditions in Strong Emergence.
Consider, to start, Ginet’s (1990) account, according to which free will involves
“an actish phenomenal quality”. As Clarke (2003) observes, one might complain
that such a conception is compatible with an agent’s not really having any “active
control”:
Whatever the correct characterization of this phenomenal quality, the mere feel
of a mental event—the way it seems to the individual undergoing it—although it
may be a (more or less reliable) sign of active control, cannot itself constitute the
agent’s exercise of such control (cf. O’Connor 2000, 25–26). To hold that it does
is to render the exercise of active control wholly subjective (nothing more than
the way things seem), and this is to greatly diminish the significance of active
control. (20)
Ginet is within his rights, however, to insist that just because the quality at
issue is phenomenal, it does not follow that this renders the associated act of
choosing “wholly subjective” or “nothing more than the way things seem”. For
notwithstanding that actish phenomenal qualities are noncausal in not being
caused, this is compatible with such qualities having powers to produce effects—
most saliently, the outcomes of acts of choosing.
Next, consider McCann’s (1998) view, according to which noncausal free will
involves intrinsic intentionality, and where, more specifically, an act of libertarian
choosing is one which is characterized as such by the presence of an intention to

decide, as “a matter of a decision’s being, by its very nature, an act that the agent
means to be performing” (Clarke 2003, 18). Here again, and granting that both the
act of choosing and the outcome of the choice are noncausal in not being appro-
priately seen as effects of prior causes, it remains that an act of choosing is (also)
causal precisely in that it involves the exercise of a power to choose, or decide.
Finally, consider Stump’s (1999) Thomistic view, according to which libertarian
free will is noncausal in the sense that, when a person chooses, “there is nothing
which is prior to that person’s acts of intellect and will and which causally
determines her intellect and will to be in the states in which they are” (414).
Here a given act of choosing may, after all, be caused—so long as those causes
are other acts of intellect or will. Most importantly for present purposes, however,
this Thomistic view is compatible with taking such acts of intellect or will to be
causal in the sense of themselves having powers to cause or contribute to causing
certain effects—most saliently, in a case of free choosing, the outcome of the act
of choosing.
On a representative range of noncausal accounts of libertarian free will, then,
the characterization of acts of choosing as ‘noncausal’ reflects that such acts are
not the effects of (non-will-involving) causes prior to the acts. It remains that such
acts are causal at least in having powers to cause, at a minimum, an outcome
of the choosing, and moreover to contribute to causing any effects associated
with the outcome of the choice (e.g., the reaching for a glass of water). This
distinctive causal asymmetry—free choices are uncaused, but capable of causing—
is reasonably interpreted as suggesting that on noncausal libertarian accounts,
acts of free choosing involve fundamentally novel powers, as Strong emergence
requires.
8.4 Is free will either Weakly or Strongly emergent?
I turn now to considering whether there is actually any free will of either Weak or
Strong emergent varieties.
8.4.1 Is there compatibilist (Weakly emergent) free will?
As discussed in §8.2, the compatibilist is plausibly seen as implementing a proper

subset strategy, on which the freedom of an act of choosing is ultimately a matter
of the act’s being associated with a proper subset of the complete and actual
causal antecedents of the choosing; and as I argued above, this strategy can in
turn be plausibly interpreted as involving certain complex events’ more specifically
free will 275
satisfying the conditions in Weak Emergence. Given all this, what are the prospects
for there actually being free will of the Weakly emergent variety?
The prospects are good. Though free choices are not taken to be part of a higher-
level system of laws on either compatibilist or libertarian accounts, a compatibilist
account is one manifesting the usual Weak emergentist characterization of special-
science goings-on as comparatively insensitive to lower-level physical details, in
the sense that an agent’s reasons for action in a given case float free of many
such details (and in particular, are sensitive only to facts about ‘relevant’ causal
antecedents). Since our deliberations and associated acts of choice clearly are
insensitive to many microphysical details, then given that free will is understood
along compatibilist (Weak emergentist) lines, there is good reason to think that
such free will actually exists, and moreover is abundant. Indeed, even libertarians
grant that there is actually free will—or as they would prefer to call it, ‘active
control’—of a compatibilist variety. Hence Clarke (2003) says:
We make decisions and act even if determinism is true; we are thus unlike
puppets. And unlike agents who are not persons, we can still act on the basis
of our appreciation of practical reasons, including moral reasons. We are also
unlike prisoners, in that we can generally go where we want to go. Further, most
of us most of the time act quite free from coercive threats and compulsive desires.
We are never subject to the direct manipulation of our brains by malevolent neu-
roscientists. All of this is good, and these goods do not require indeterminism.
There is, then, a valuable variety of active control that we can have and exercise
even if determinism is true. However, this compatibilist variety of active control
falls short of free will. (8)
8.4.2 Is there libertarian (Strongly emergent) free will?
Notwithstanding that there is presumably plenty of what the compatibilist counts

as free will, I am inclined to agree with those, like Clarke, who think that
compatibilist accounts are ultimately unsuccessful in accommodating the core
phenomenal and intentional aspects of free will, according to which a freely
choosing agent feels, from the inside, to be causally determinative of the outcome
in a way transcending any nomological goings-on, whether deterministic or
indeterministic. Even those who aim to reject the appearances as genuine admit
as much. Hence Caruso (2012), a determinist, says:
A major part of the folk psychology of free will is the belief that our conscious
intentions cause action. As Patrick Haggard and Benjamin Libet write, “Most of
us navigate through our daily lives with the belief that we have conscious free will:
that is, we have conscious intentions to perform specific acts, and those intentions
can drive our bodily actions, thus producing a desired change in the external
world” (2001, 47). This commonsense intuition plays a major role in our sense of
free will and is essential to the up-to-me-ness that we associate with free will. It is
also well supported by phenomenology. In normal cases of voluntary behavior,
we experience a conscious intention before the onset of action and naturally take
the former to be the cause of the latter. When I switch on my reading lamp, for
example, I feel as though it is I, my conscious self, that controls the movements
of my arms and hands through the conscious formation of goals, intentions, and
decisions. (189)
Two points about this sort of pretheoretical ‘take’ on our capacity for free will
are worth noting. First, note that the phenomenal and intentional motivations for
thinking that our choices are at least sometimes nomologically transcendently free
don’t turn on the choice at issue being directed at anything morally or otherwise
substantive.1⁴ I can experience myself as seemingly transcendently free even if all
that is at issue is whether or not to throw a piece of chalk into the air.1⁵ As such, and
even granting that inquiring into exactly how free will (assuming it exists) inter-
sects with notions such as reasons, character, and moral responsibility, for present
purposes we can restrict our focus to what seems to me to be the prior question of
whether we are ever in position to act in ways that transcend the nomological net.
Second, note that this much is already enough to block the too-quick strategy of
appealing to indeterministic quantum or other laws as a basis for such freedom.
This strategy is often rejected on grounds that an appeal to indeterministic laws
would render the outcomes of choosings subject to random processes rather than
reasons; but for present purposes it is enough to reject this strategy to note that
indeterministic laws are still laws, solidly within the net of nomological goings-on,
in which case appeal to such laws cannot provide a realistic metaphysical basis for
the seeming experience of nomologically transcendent free will.
Now, if the appearance of nomologically transcendent free will is to be taken as
genuine, free will so understood is not properly accommodated via the usual Weak
emergentist approach to higher-level special scientific goings-on, for even if it is
correct (as the compatibilist maintains) that our acts of seemingly free choosing
are insensitive to certain (antecedent) micro-level details, it is not this insensitivity
that constitutes nomologically transcendent free choice, or the experience of
such. Indeed, once it is appreciated that compatibilist accounts of free will are
1⁴ This observation is registered in views on which freedom is not itself sufficient for moral
responsibility; see Clarke 1992.
1⁵ Hence I would disagree with Clarke’s (2003)’s claim that “For an agent to act with free will, she
must be able to regard some considerations as reasons for action” (16), at least if the ‘reason’ has to
contain more content than the choice in itself, as in ‘I hereby choose to throw the piece of chalk into
the air’.
free will 277
implementing a variation on the usual Weak emergentist theme, such that any
token power associated with (a complex event having as a part) an act of choosing
ends up being identical to a token power associated with a lower-level physical
event—to be sure, a highly complex, temporally extended physical event, but no
matter—the usual concerns with a compatibilist approach are thrown into high
relief. It may be that agents on a compatibilist approach are distinctively efficacious,
in ways that complex systems, ordinary objects, and other special-science goings-
on are distinctively efficacious, in having distinctive power profiles that track
difference-making considerations (if I had been differently molecularly config-
ured, I would still have chosen to go to Pasternak for brunch) or comparatively
abstract causal joints. But this form of distinctive efficacy is not of the transcendent
variety that is core and crucial to our phenomenal and intentional experience of
agency.
What seems to be required for realistic accommodation of this experience is
that associated acts of free choosing involve a fundamentally novel power which,
as libertarians standardly maintain, transcends any nomological net. And as
previously discussed, a fundamentally novel power of event-causal, agent-causal,
or ‘noncausal’ variety will do for purposes of such accommodation.
Is there any libertarian (Strongly emergent) free will, so transcendently under-
stood? To start, as with other purported Strongly emergent phenomena, there is a
case to be made that the phenomenon of libertarian free will admits of empirical
confirmation (or disconfirmation), at least in principle, by attention to whether
the phenomenon involves any apparent violations of conservation laws. After all,
even if libertarian free will involves a fundamentally novel power, and even if
such a novel power transcends not just the nomological net of lower-level physical
goings-on, but indeed any nomological net, so long as such a power is at least partly
determinative of a given effect, it will presumably involve some transfer of energy
or other conserved quantity; and if such a transfer involves some fundamentally
novel interaction, then this would provide an in-principle means of empirically
verifying the hypothesis that free will is Strongly emergent.
Independent of this route to empirical verification, however, the fact that we
have direct introspective access to the phenomenon of seemingly nomologically
transcendent free will provides the basis for a new argument for there actually
being libertarian—Strongly emergent—free will, as follows:
1. We experience ourselves as seeming to freely choose, in ways transcending

any nomological (deterministic or indeterministic) goings-on.
2. In the absence of good reasons to think that our experience of nomologically
transcendent free will cannot be taken at face value, we are entitled to take
this experience at realistic face value.
3. There are no good reasons to think that our experience of transcendent free
will cannot be taken at realistic face value.
∴ We are entitled to take our experience of nomologically transcendent free

will at realistic face value.
The argument is valid, and premise 1 is clearly true: again, even non-libertarians
agree that we seem to freely choose, in ways transcending any nomological net,
on at least some occasions. Premise (2) also seems reasonable: if we have clear
experience of some seeming phenomenon, we need good reason not to take that
experience at realistic face value. In what follows, then, I’ll focus on defending
premise (3) against certain salient empirical results purporting to show that the
phenomenon of free will is in some sense an illusion.
The empirical case against libertarianism

Recent results in empirical neuroscience have frequently been taken to show that
seeming experiences of free will as nomologically transcendent cannot be taken at
face value. Hence Caruso (2012) claims:
Although phenomenology supports this commonsense belief [in nomologically

transcendent free will], empirical evidence in neuroscience now seriously ques-
tions it. In fact, a growing number of theorists now conclude that conscious will—
in the sense of consciously initiated action—is incompatible with the evidence
of neuroscience […] Much of the contemporary case for this conclusion is
derived from the experimental work of Benjamin Libet and his colleagues. […]
My thesis will be that the empirical results from neuroscience do in fact reveal
that conscious will is an illusion—at least in the cases we can currently study
empirically. (189)
In what follows I focus on the ‘Libet cases’ which pose the most serious challenge
to taking our seeming experience of transcendent free will at face value.1⁶ These
studies aim to compare the self-reported time of occurrence of a certain conscious
choice to produce a certain physical behavior with the time of occurrence of
certain unconscious brain states associated with the production of the behavior.
The concern raised for nomologically transcendent free will is that the evidence
gathered in these studies has been interpreted as suggesting that the unconscious
initiation of the physical behavior occurs prior to the time of the supposed
choosing, such that the supposition that our free choices are determinative of our
1⁶ See O’Connor (2009b) for discussion of a number of other empirical results that some have seen as
problematizing taking the seeming nomological transcendence of free choice as genuine; these include,
e.g., cases where subjects of induced behaviors confabulate instances of their agency in a post-hoc way,
cases where the outcomes of supposedly free choices (about which index finger to move) are influenced
by external stimulation of the subject’s brain, cases of subjects who report a feeling of agency concerning
distal outcomes subsequent to being instructed to have certain negative thoughts, and cases where
subjects engage in seemingly voluntary action unaccompanied by any feeling of agency (as in ‘alien
hand syndrome’). O’Connor compellingly argues that none of these results poses a serious challenge to
the libertarian supposition; I direct the interested reader to his discussion.
free will 279
physical actions is illusory. O’Connor (2009b) describes the original setup and
interpretation of results as presented in Libet 1999:
Libet devised a study in which people are asked to wiggle their finger within a
short interval of time (thirty seconds or so). The experimenter instructs them to
do so whenever they wish—though spontaneously, not by deciding the moment
in advance. Throughout, they are to watch a special clock with a very fast-moving
dial (a beam of light) and note its location at the precise moment at which they felt
the “urge” or “wish” to move the finger. During the experiment, a device measures
electrical activity on the agent’s scalp. Libet discovered that a steady increase in
this activity (dubbed the “readiness potential,” or RP) consistently preceded the
time the agents cited as when they experienced the will to move. By averaging
results over hundreds of experiments, Libet determined that the RP preceded the
“experience of will” by an average of some 400 milliseconds, a significant interval
in the context of neural activity. Libet and others concluded from this result that
“conscious will” is not the initiator of voluntary acting but instead a consequence
of an unconscious physical process that also (and according to some hypotheses,
independently) triggers the action. (176)
Do experimental studies of this sort really establish that nomologically tran-

scendent free will is an illusion? This question continues to be hotly debated, but so
far as I can tell, the answer is clearly ‘no’. Others—notably, O’Connor (2005b) and
Mele (2009) have offered very detailed discussions of the relevant studies, noting
in particular that these admit of plausible alternative interpretations which do
not have anti-libertarian import. Here I’ll briefly canvass two of these interpretive
options, and offer a new one of my own, which reflects one of the characteristic
features of metaphysically emergent phenomena.
To start, as O’Connor notes, the setup is one where the agent has already decided
to perform a specific action, and while the focus of the experiment is supposedly
on when the agent’s choice to perform that action occurs, both the antecedent
decision and the restriction on timing render the setup sufficiently nonstandard
that one might reasonably deny taking the results to suggest anything general
about the status of nomologically transcendent agency. More to the interpretive
point, O’Connor observes that the instructions to the subject to introspect and
wait for an unplanned “urge” to occur would likely encourage a passive posture in
which “having decided that one will move, one looks for the urge to do so in order
to act upon it”, which in turn motivates an alternative interpretation of the results
according to which “such a preformed intention to act upon the right internal ‘cue’
initiates an unconscious process that promotes the occurrence (or perhaps evolu-
tion) of a conscious state of desire or intention that is not actively formed” (182).
Such an interpretation is consonant with another study by Libet, which confirms
that in such an experimental context, the subject has ‘veto’ power over the
antecedent ‘urge’ in question. As Libet (2002) himself observes:
[T]he conscious function still had enough time to affect the outcome of the
process; that is, it could allow the volitional initiative to go to completion, it
could provide a necessary trigger for the completion, or it could block or veto the
process and prevent the act’s appearance. There is no doubt that a veto function
can occur. (292)
One might thus interpret the antecedent brain activity in the original setup as
simply setting in train a deliberative process having as its ‘default’ object the
performance of a certain physical action, where the act of free choosing consists
either in a decision to allow the process to continue to completion or in a decision
to block the process. A second, more explicitly causal, interpretation is found in
Mele’s (2009) suggestion that the ability of subjects to exercise veto power renders
it “much more likely that [the prior brain activity] is a potential cause of a proximal
intention or decision rather than a proximal intention or decision itself ” (51). Each
of these alternative interpretations is compatible with nomologically transcendent
free will.
Finally, I suggest that a third interpretive option should be put on the table—
namely, one according to which the intention to choose and the brain activity
are cotemporally initiated, but where it takes just a bit of time for this fact
to consciously register as a complete thought in the agent’s mind. Thinking,
and presumably also choosing, takes time. As such, a very small lag—less than
half a second—between initiation and conscious recognition would be a natural
concomitant of our mental decision-making processes, again compatible with
nomologically transcendent free will. More generally, Libet’s assumption that “In
the traditional view of conscious will and free will, one would expect conscious
will to appear before, or at the onset, of the RP [Readiness Potential], and thus
command the brain to perform the intended act” (1999, 49) appears to reflect an
unmotivated understanding of transcendent free will as instantaneous and indeed
as in some sense floating free of underlying physical processes.
Here, it seems to me, attention to the characteristic features of metaphysical
emergence can prove useful. For first, the libertarian qua Strong emergentist will
assume that mental features are cotemporally materially dependent on physical
features; but second, as I have previously emphasized, the cotemporal dependence
base may be temporally extended. As such, one might reasonably maintain, as an
available interpretation of Libet’s results, that the occurrence of a given Readiness
Potential would be not prior to the event of choosing, either as a kind of ‘default’
unconscious deliberative process (as O’Connor suggests) or as a prior cause of
the event of choosing (as Mele suggests), but rather as a part of the temporally
extended dependence base for an event of free choosing.
There are thus at least three ways to resist taking Libet cases to undercut our
seeming experience of nomologically transcendent free will. I conclude that in the
absence of any more compelling route to undercutting this experience—which
experience, again, is plausibly understood as involving satisfaction of the New
Power Condition—and given the empirically supported supposition that acts of
free will 281
free choosing (like other mental events) are cotemporally materially dependent on
lower-level goings-on, we are entitled to take our experience at realistic libertarian
face value as an actual case of Strong emergence.
Let’s sum up. As Bernstein and I argue in our 2016, there is an important and
theoretically powerful connection between the problem of mental causation and
the problem of free will—namely, each can be seen as specific cases of a general
‘problem of mental quausation’—the problem of how a mental event (or other
higher-level feature) can be efficacious qua the type of event it is, given the open
possibility of certain theses which threaten to undermine, one way or another, the
supposed causal relevance/distinctive efficacy of the event.
This deep connection suggests certain parallels between positions in the cor-
responding debates. Here I have argued that compatibilists, like non-reductive
physicalists, implement a strategy entailing satisfaction of the Proper Subset of
Powers Condition, and that libertarians, like Strong emergentists generally, imple-
ment a strategy entailing satisfaction of the New Power Condition. Coupled with
the assumption of cotemporal material dependence, it follows that compatibilist
free will (if such exists) is Weakly emergent, whereas libertarian free will (if such
exists) is Strongly emergent.
Finally, I have considered whether free will is actually either Weakly or Strongly
emergent. Free will on a compatibilist view, I have argued, is easy to come by,
as another case-in-point of the usual Weak emergentist understanding of higher-
level features as comparatively abstract or insensitive to lower-level physical
details. This result is a double-edged sword, however, for unlike cases of higher-
level phenomena which seem amenable to Weak emergentist treatment, including
complex systems, classical ordinary objects, and qualitative perceptual mental
states, such insensitivity to lower-level detail seems beside the point of generally
accommodating free will, which in at least some manifestations appears to have
more to do with an agent’s ability to nomologically transcend, as opposed to
abstract from, any lower-level physical goings-on. As I have argued, however, our
introspective experience of seemingly nomologically transcendent free will itself
(in combination with the usual assumption of cotemporal material dependence)
provides good reason to think that we have free will of a libertarian, Strongly
emergent variety; and the neuroscientific reasons for rejecting this experience
as illusory are uncompelling. Contrary to common assumption, it is libertarian
free will, not subjective or qualitative experience, that provides the best case of a
Strongly emergent phenomenon.
I conclude that there is actually free will of both Weak and Strong varieties—
a nice result, given the importance of free will as a basis for personal and moral
autonomy, and one which, to my mind, provides a fitting closing indication of the
importance of metaphysical emergence to our understanding not just of the world,
but of ourselves.
9
Metaphysical emergence: next steps
This book began with two key questions, inspired by the target cases of special-
scientific and artifactual entities and features, as seemingly cotemporally mate-
rially dependent on, yet ontologically and causally autonomous with respect to,
lower-level micro-configurations and features. What is metaphysical emergence,
more precisely? And is there any such emergence, in principle and moreover in
actual fact? We now have answers to both questions in hand.
First, what metaphysical emergence is, more precisely, is encoded in the
schemas for Weak and Strong emergence, tracking varieties of emergence that are
physically acceptable and physically unacceptable, respectively, on the assumption
that the dependence base goings-on are ultimately physical. The conditions in the
schemas—Weak Emergence and Strong Emergence, respectively—capture what is
core and crucial (and when suitably filled in, necessary and sufficient) for each
form of emergence, in appropriate and illuminating fashion. Moreover, given that
metaphysically emergent entities and features must be causally autonomous, the
two schemas exhaust the available options. For they reflect that (as attention to the
problem of higher-level causation confirms) there are two and only two routes to
distinctive higher-level efficacy: either the higher-level entity or feature has more
powers, or it has fewer powers, than its dependence base, on a given occasion.
Second, there is metaphysical emergence, both in principle and as a matter of
actual fact. Notwithstanding the many attempts to problematize the forms of emer-
gence at issue in the schemas, these attempts ultimately fail. Each form of emer-
gence is viable: coherent, metaphysically substantive, naturalistically acceptable,
such as to avoid problematic causal overdetermination, and more generally such
as to accommodate and illuminate the prima facie appearances of metaphysical
emergence, as per our guiding methodology. Moreover, each form of emergence
is arguably actually instantiated, with certain complex systems, ordinary objects,
qualitative mental states, and compatibilist free will being Weakly emergent, and
libertarian (nomologically transcendent) free will being Strongly emergent.
This is not the end of the journey, of course. To start, there are many candidates
for metaphysically emergent phenomena that I have not considered in any detail
here, including quantum entanglement (see Humphreys 1997, Silberstein and
McGeever 1999), the fractional quantum Hall state (see Morrison 2006, Lancaster
and Pexton 2015), molecular structure (see Hendry 2010 and 2017), polymers (see
McLeish et al. 2019), biological systems (see Mossio et al. 2013, McLeish 2017,
DOI: 10.1093/oso/9780198823742.003.0009
metaphysical emergence: next steps 283
Santos forthcoming), brain dynamics (see Thompson and Varela 2001 and Thomp-
son 2007), collective individuals (see Bouchard and Huneman 2013), ecological
systems (see Bergandi and Blandin 1998, Levin 2005, and Smith 2006), and
beyond. And the status of many phenomena as Strongly emergent, including some
of those we have previously considered, remains an open empirical question,
contingent on as-yet-unconducted experiments establishing that (on the model
of the discovery of the weak nuclear interaction) one or more fundamental
interactions come into play only under certain comparatively complex circum-
stances. It is also worth noting that, notwithstanding that the present investigation
has been directed at identifying what it would be for there to be metaphysical
emergence of the sort motivated by the target cases, as combining cotemporal
material dependence with ontological and causal autonomy, there may well be
other cases of purported emergence—e.g., spacetime (see Huggett and Wüthrich
2013 and Lam and Wüthrich 2018 for discussion) or numbers (perhaps along lines
discussed in Maddy 1997) which (depending on whether the cases really are ones
of metaphysical emergence) require an alternative to or extended understanding
of the schemas for metaphysical emergence presented here.1 However these and
other philosophical and empirical investigations into metaphysical emergence
play out, my hope is that the present project will provide a useful base of operations
going forward.
I want to close with some methodological observations, that point towards some
other ways in which the present project might be profitably extended. At several
junctures along the way, we saw that one or other revisionary (e.g., reductionist
or anti-realist) account of the nature of seemingly higher-level entities or features
reflected the supposition that some such approach is required for a naturalistically
acceptable resolution of the problem of higher-level causation. A related suppo-
sition was tacitly operative in Thomasson’s (2007) claim (discussed at the end of
Ch. 6) that the best response to causal overdetermination-based concerns about
ordinary objects is to maintain that investigations into the nature and existence
of such objects should proceed in some way different from investigations into
the nature and existence of special-scientific goings-on, since (she assumes) the
embrace of a unified methodology would likely entail acceptance of some or other
revisionary account of ordinary objects. Supposing so, she suggests, better to
1 There is a great deal of interest to say about such cases, which must await a future occasion. For
the case of spacetime, I will here just note that insofar as the purported dependence base is understood
as itself not concretely located (if location presupposes spacetime) but rather in some sense abstract,
an alternative understanding of the relation at issue might be as involving instantiation rather than
(anything answering to) metaphysical emergence. Alternatively, if the base is somehow concretely
located in such a way as to functionally implement certain features of spacetime (along lines suggested,
e.g., by Knox 2013 and 2018), then it might be that references to ‘cotemporal’ dependence in the
schemas for emergence could be extended to refer to whatever in the dependence base plays the
associated role.
take for granted that ordinary objects exist and have the natures that we usually
attribute to them, even at the price of turning metaphysical investigation into a
branch of linguistic or conceptual anthropology.
A moral of the results argued for here, however, is that we are not forced
to embrace either a revisionary metaphysics or a revisionary methodology of
metaphysics—at least for all the problem of higher-level causation shows. For the
concerns associated with the problem of higher-level causation can be answered,
whether as attaching to special-science entities or to ordinary objects (and their
features), and correspondingly there is no clear pressure from this direction for
treating ordinary objects differently from (other) special-science entities.
That said, additional concerns have been raised about the posit of macro-
entities and features, understood as cotemporally materially dependent on, yet
autonomous with respect to, lower-level entities and features (see Paul 2010 for an
overview). One of these is the ‘problem of the many’ (see Unger 1980), according
to which the intimate dependence of, e.g., a cat—say, Tibbles—on its constitut-
ing matter, coupled with the seeming fact that there are many candidate ‘cat-
constituters’ in the vicinity of Tibbles, poses a difficulty for the usual assumption
that there is just one cat on the mat. Another is the puzzle of the statue and the clay
(see Gibbard 1975), associated with the ‘grounding problem’ (see deRosset 2011):
given the intimate relation between a statue and its constituting matter, how is it
that these can be associated with different persistence conditions, such that, e.g.,
the clay, but not the statue, can survive squashing, and the statue, but not the clay,
can survive the loss of a significant part?
These problems have also been taken by some to support revisionary meta-
physical or methodological approaches to the phenomena at hand. A natural
next question, then, is whether the powers-based schemas for Weak and Strong
emergence, or some variation on a powers-based theme, might provide the basis
for non-revisionary resolutions of problems besides that of higher-level causation.
For example, perhaps the key to realistic accommodation of the differing persis-
tence conditions of the statue and the clay lies in appreciating that, on any given
occasion of constitution, the token powers of the statue and the token powers of
the clay overlap—perhaps not by way of one collection being a proper subset of the
other, but rather by each sharing some but not all powers with the other. Here—
to venture even further along the speculative branch—there may be a reciprocal
relation between the powers that are in some sense characteristic of an object
and the changes through which it may persist. For example, it might be that the
ability of the clay, but not the statue, to survive being squashed reflects that while
the clay (characteristically, as a matter of its nature) has powers associated with
the determinable feature being shaped, it does not have all the powers associated
with being statue-shaped (or at least does not have those powers in the same
direct way as the statue it constitutes); similarly, it may be that the ability of
metaphysical emergence: next steps 285
the statue, but not the clay, to survive the loss of a significant part, reflects that
while the statue (characteristically, as a matter of its nature) has powers associated
with being statue-shaped, it does not have powers associated with certain specific
determinates of this feature—those powers are had, at best, by certain temporal
parts of the statue.
All this is speculative but to my mind tantalizing food for future thought. Time
will tell, but it may well be that attention to the broadly mereological relationships
between sets of powers associated with entities or features can provide the key to
realistic and illuminating accommodation of other aspects of higher-level reality,
beyond metaphysical emergence.
Bibliography
Aizawa, Kenneth and Carl Gillett, 2009. “The (Multiple) Realization of Psychological and
Other Properties in the Sciences”. Mind and Language, 24:181–208.
Akiba, Ken, 2004. “Vagueness in the World”. Noûs, 38:407–29.
Alexander, Samuel, 1920. Space, Time, and Deity. London: Macmillan.
Alter, Torin, 1998. “A Limited Defense of the Knowledge Argument”. Philosophical Studies,
90:35–56.
Alter, Torin, 2020. “Physicalism Without Fundamentality.” Erkenntnis. https://doi.org/
10.1007/s10670-020-00285-6.
Anderson, P. W., 1972. “More is Different”. Science, 177:393–6.
Antony, Louise M., 2003. “Who’s Afraid of Disjunctive Properties?” Philosophical Issues,
13:1–21.
Antony, Louise M. and Joseph M. Levine, 1997. “Reduction with Autonomy”. Philosophical
Perspectives, 11:83–105.
Armstrong, David M., 1978. Universals and Scientific Realism, Vol II: A Theory of Universals.
Cambridge: Cambridge University Press.
Audi, Paul, 2012a. “Grounding: Toward a Theory of the In-Virtue-of Relation”. Journal of
Philosophy, 109:685–711.
Audi, Paul, 2012b. “Properties, Powers, and the Subset Account of Realization”. Philosophy
and Phenomenological Research, 84:654–74.
Auyang, Sunny, 1999. How is Quantum Field Theory Possible? Oxford: Oxford University
Press.
Ayer, Alfred J., 1954. “Freedom and Necessity”. In his Philosophical Essays, 271–84. London:
Macmillan. Reprinted in Watson 1982, 15–23.
Baker, Lynne Rudder, 1993. “Metaphysics and Mental Causation”. In John Heil and Alfred
Mele, editors, Mental Causation, 75–96. Oxford: Clarendon Press.
Baltimore, Joseph A., 2013. “Careful, Physicalists: Mind–Body Supervenience Can Be Too
Superduper”. Theoria, 79:8–21.
Barnes, Elizabeth, 2006. Conceptual Room for Ontic Vagueness. Ph.D. thesis, University of
St. Andrews.
Barnes, Elizabeth, 2010. “Ontic Vagueness: A Guide for the Perplexed”. Noûs, 44:601–27.
Barnes, Elizabeth, 2012. “Emergence and Fundamentality”. Mind, 121:873–901.
Barnes, Elizabeth and J. R. G. Williams, 2011. “A Theory of Metaphysical Indeterminacy”. In
Karen Bennett and Dean W. Zimmerman, editors, Oxford Studies in Metaphysics volume
6, 103–48. Oxford: Oxford University Press.
Batterman, Robert W., 1998. “Why Equilibrium Statistical Mechanics Works: Universality
and the Renormalization Group”. Philosophy of Science, 65:183–208.
Batterman, Robert W., 2005. “Critical Phenomena and Breaking Drops: Infinite Idealiza-
tions in Physics”. Studies in History and Philosophy of Science Part B, 36:225–44.
Batterman, Robert W., 2000. “Multiple Realizability and Universality”. British Journal for the
Philosophy of Science, 51:115–45.
288 bibliography
Batterman, Robert W., 2002. The Devil in the Details: Asymptotic Reasoning in Explanation,
Reduction, and Emergence. Oxford: Oxford University Press.
Baysan, Umut, 2014. Realization and Causal Powers. Ph.D. thesis, University of Glasgow.
Baysan, Umut, 2016. “An Argument for Power Inheritance”. Philosophical Quarterly,
383–90.
Baysan, Umut and Jessica M. Wilson, 2017. “Must Strong Emergence Collapse?” Philosoph-
ica, 91:49–104.
Bealer, George, 1997. “Self-Consciousness”. Philosophical Review, 106:69–117.
Bealer, George, 2002. “Modal Epistemology and the Rationalist Renaissance”. In Tamar
Gendler and John Hawthorne, editors, Conceivability and Possibility, 71–126. New York:
Oxford University Press.
Bedau, Mark A., 1997. “Weak Emergence”. Philosophical Perspectives 11: Mind, Causation
and World, 11:375–99.
Bedau, Mark A., 2002. “Downward Causation and the Autonomy of Weak Emergence”.
Principia, 6:5–50.
Bedau, Mark A., 2008. “Is Weak Emergence Just in the Mind?” Minds and Machines, 18:
443–59.
Bedau, Mark A., 2010. “Weak Emergence and Context-Sensitive Reduction”. In Antonella
Corradini and Timothy O’Connor, editors, Emergence in Science and Philosophy, 6–46.
New York: Routledge.
Beebe, James R., 2009. “The Abductivist Reply to Skepticism”. Philosophy and Phenomeno-
logical Research, 79:605–36.
Bennett, Karen, 2003. “Why the Exclusion Problem Seems Intractable and How, Just Maybe,
to Tract It”. Noûs, 37:471–97.
Bennett, Karen, 2015. “ ‘Perfectly Understood, Unproblematic, and Certain’: Lewis on
Mereology”. In Barry Loewer and Jonathan Schaffer, editors, A Companion to David
Lewis, 250–61. Oxford: Wiley Blackwell.
Bennett, Karen, 2017. Making Things Up. Oxford: Oxford University Press.
Bergandi, Donato and Patrick Blandin, 1998. “Holism vs. Reductionism: Do Ecosystem
Ecology and Landscape Ecology Clarify the Debate?” Acta Biotheoretica, 46:185–206.
Berkeley, George, 1710. A Treatise Concerning the Principles of Human Knowledge. In A. A.
Luce and T. E. Jessop, editors, The Works of George Berkeley, volume 2. London: Thomas
Nelson and Sons.
Berker, Selim, 2017. “The Unity of Grounding”. Mind, 127:729–77.
Bernstein, Sara and Jessica M. Wilson, 2016. “Free Will and Mental Quausation”. Journal of
the American Philosophical Association, 2:310–31.
Biggs, Stephen and Jessica M. Wilson, 2017. “The A Priority of Abduction”. Philosophical
Studies, 174:735–58.
Biggs, Stephen and Jessica M. Wilson, 2019. “Abduction versus Conceiving in the Episte-
mology of Modality”. In New Directions in the Epistemology of Modality, Special Issue of
Synthese, Antonella Mallozzi, editor. Published online at https://doi.org/10.1007/s11229-
019-02358-8.
Biggs, Stephen and Jessica M. Wilson, 2020. “Abductive Two-Dimensionalism: A New
Route to the A Priori Identification of Necessary Truths”. Synthese 197:59–93. First
published online in 2017 at https://doi.org/10.1007/s11229-017-1444-6.
Bird, Alexander, 2001. “Necessarily, Salt Dissolves in Water”. Analysis, 61:267–74.
Bird, Alexander, 2002. “On Whether Some Laws are Necessary”. Analysis, 62:257–70.
bibliography 289
Bird, Alexander, 2007. Nature’s Metaphysics: Laws and Properties. Oxford: Oxford University
Press.
Bliss, Ricki and Kelly Trogdon, 2016. “Metaphysical Grounding”. The Stanford Encyclopedia
of Philosophy (Winter 2016 edition), Edward N. Zalta, editor. https://plato.stanford.edu/
archives/win2016/entries/grounding/. Accessed August 31, 2020.
Blitz, David, 1994. Emergent Evolution: Qualitative Novelty and the Levels of Reality.
Dordrecht: Springer.
Block, Ned and Jerry A. Fodor, 1972. “What Psychological States Are Not”. Philosophical
Review, 81:159–81.
Block, Ned and Robert C. Stalnaker, 1999. “Conceptual Analysis, Dualism, and the
Explanatory Gap”. Philosophical Review, 108:1–46.
Boghossian, Paul, 1996. “Analyticity Reconsidered”. Noûs, 30:360–91.
Bohm, David, 1952. “A Suggested Interpretation of Quantum Theory in Terms of ‘Hidden
Variables’, Parts I and II”. The Physical Review, 85:166–79, 180–93.
Bokulich, Alisa, 2014. “Metaphysical Indeterminacy, Properties, and Quantum Theory”. Res
Philosophica, 91:449–75.
Bonjour, Lawrence, 1998. In Defence of Pure Reason. Cambridge: Cambridge University
Press.
Bouchard, Frederic and Philippe Huneman, editors, 2013. From Groups to Individuals:
Evolution and Emerging Individuality. Cambridge: MIT Press.
Boyd, Richard, 1980. “Materialism without Reduction: What Physicalism does Not Entail”.
In Ned Block, editor, Readings in the Philosophy of Psychology, volume 1, 67–106.
Cambridge, MA: Harvard University Press.
Broad, C. D., 1925. Mind and Its Place in Nature. (From the 1923 Tanner Lectures at
Cambridge.) Cambridge: Kegan Paul.
Bryant, Amanda, 2020. “Physicalism”. In Michael J. Raven, editor, The Routledge Handbook
of Metaphysical Grounding, 484–500. New York: Routledge.
Byrne, Alex, 1994. The Emergent Mind. Ph.D. thesis, Princeton University,
Byrne, Alex, 2020. “Inverted Qualia”. The Stanford Encyclopedia of Philosophy (Fall 2020
Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/fall2020/entries/
qualia-inverted/. Accessed August 31, 2020.
Calosi, Claudio and Jessica M. Wilson, 2019. “Quantum Metaphysical Indetermi-
nacy”. Philosophical Studies, 176: 2599–627. First published online in 2018 at
https://doi.org/10.1007/s11098-018-1143-2.
Camazine, Scott, Jean-Louis Deneubourg, Nigel R. Franks, James Sneyd, Guy Theraulaz,
and Eric Bonabeau, 2018. Self-organization in Biological Systems. Princeton, NJ: Prince-
ton University Press.
Campbell, Keith, 1990. Abstract Particulars. Oxford: Blackwell.
Carere, Claudio, Simona Montanino, Flavia Moreschini, Francesca Zoratto, Flavia
Chiarotti, Daniela Santucci, and Enrico Alleva, 2009. “Aerial Flocking Patterns of Win-
tering Starlings, Sturnus vulgaris, under Different Predation Risk”. Animal Behaviour,
77:101–7.
Carruth, Alexander, 2018. “Emergence, Reduction and the Identity and Individuation of
Powers”. Topoi, 39:1021–30.
Caruso, Gregg, 2012. Free Will and Consciousness: A Determinist Account of the Illusion of
Free Will. Washington, D.C.: Lexington Books.
Chalmers, David J., 1996. The Conscious Mind. Oxford: Oxford University Press.
290 bibliography
Chalmers, David J., 1999. “Materialism and the Metaphysics of Modality”. Philosophy and
Phenomenological Research, 59:473–96.
Chalmers, David J., 2002. “Does Conceivability Entail Possibility?” In Tamar Gendler and
John Hawthorne, editors, University Press Conceivability and Possibility, 145–200. New
York: Oxford University Press.
Chalmers, David J., 2003. “Consciousness and its Place in Nature”. In Stephen P. Stich and
Ted A. Warfield, editors, Blackwell Guide to the Philosophy of Mind, 102–42. Malden, MA:
Blackwell.
Chalmers, David J., 2004. “The Representational Character of Experience”. In Brian Leiter,
editor, The Future for Philosophy, 153–81. New York: Oxford University Press.
Chalmers, David J., 2006a. “Strong and Weak Emergence”. In Philip Clayton and Paul
Davies, editors, The Re-Emergence of Emergence 244–56. New York: Oxford University
Press.
Chalmers, David J., 2006b. “The Foundations of Two-Dimensional Semantics”. In Manuel
Garcia-Carpintero and Josep Macia, editors, Two-Dimensional Semantics: Foundations
and Applications, 55–140. Oxford: Clarendon Press.
Chalmers, David J., 2009. “The Two-Dimensional Argument Against Materialism”. In
Brian P. McLaughlin and Sven Walter, editors, Oxford Handbook to the Philosophy of
Mind, 1–26. Oxford: Oxford University Press.
Chalmers, David J., 2012. Constructing the World. Oxford University Press.
Chalmers, David J. and Frank Jackson, 2001. “Conceptual Analysis and Reductive Expla-
nation”. The Philosophical Review, 110:315–60.
Churchland, Patricia S., 1986. Neurophilosophy: Toward A Unified Science of the Mind-Brain.
Cambridge, MA: MIT Press.
Churchland, Paul M., 1981. “Eliminative Materialism and the Propositional Attitudes”.
Journal of Philosophy, 78:67–90.
Churchland, Paul M., 1984. Matter and Consciousness. Cambridge, MA: MIT Press.
Clapp, Lenny, 2001. “Disjunctive Properties: Multiple Realizations”. Journal of Philosophy,
98:111–36.
Clarke, Randolph and Justin Capes, 2017. “Incompatibilist (Nondeterministic) Theories
of Free Will”. The Stanford Encyclopedia of Philosophy (Spring 2017 Edition), Edward
N. Zalta, editor. https://plato.stanford.edu/archives/spr2017/entries/incompatibilism-
theories/. Accessed August 31, 2020.
Clarke, Randolph, 1992. “Free Will and the Conditions of Moral Responsibility”. Philosoph-
ical Studies, 66:53–72.
Clarke, Randolph, 1999. “Nonreductive Physicalism and the Causal Powers of the Mental”.
Synthese, 51:295–322.
Clarke, Randolph, 2003. Libertarian Accounts of Free Will. New York: Oxford University
Press.
Cohen, Stewart, 2010. “Bootstrapping, Defeasible Reasoning, and a Priori Justification”.
Philosophical Perspectives, 24:141–59.
Cohen-Tannoudji, Claude, Bernard Diu, and Franck Laloë, 1977. Quantum Mechanics,
volume 1. Paris: John Wiley & Sons, 2nd edition.
Conee, Earl, 1994. “Phenomenal Knowledge”. Australasian Journal of Philosophy, 72:
136–50.
Couzin, I. D., 2007. “Collective Minds”. Nature, 445:4.
Crane, Tim, 2001. “The Significance of Emergence”. In Carl Gillett and Barry M. Loewer,
editors, Physicalism and Its Discontents, 207–24. Cambridge: Cambridge University
Press.
bibliography 291
Crane, Tim and D. H. Mellor, 1990. “There Is No Question of Physicalism”. Mind, 99:
185–206.
Craver, Carl F., 2001. “Mechanisms, and Hierarchy”. Philosophy of Science, 68:53–74.
Cucu, Alin C. and J. Brian Pitts, 2019. “How Dualists Should (Not) Respond to the Objection
From Energy Conservation”. Mind and Matter, 17:95–121.
Cummins, Robert, 1975. “Functional Analysis”. Journal of Philosophy, 72:741–64.
Cummins, Robert, 1983. “Analysis and Subsumption in the Behaviorism of Hull”, Philoso-
phy of Science, 50:96–111.
Cunningham, Bryon, 2001. “The Reemergence of Emergence”. Philosophy of Science,
68:S62–S75. Supplement: Proceedings of the 2000 Biennial Meeting of the Philosophy
of Science Association.
Damper, R. I., 2000. “Editorial for the Special Issue on ‘Emergent Properties of Complex
Systems’ ”. International Journal of Systems Science, 31:811–18.
Dasgupta, Shamik, 2014. “The Possibility of Physicalism”. Journal of Philosophy, 111:557–92.
Davidson, Donald, 1963. “Actions, Reasons, and Causes”. Journal of Philosophy, 60:685–700.
Davidson, Donald, 1970. “Mental Events”. In L. Foster and J. Swanson, editors, Experi-
ence and Theory, 207–224. Amherst, MA: Massachusetts University Press. Reprinted in
Davidson 1980/2001.
Davidson, Donald, 1980/2001. Essays on Actions and Events. Oxford: Oxford University
Press.
deRosset, Louis, 2011. “What Is the Grounding Problem?” Philosophical Studies, 156:
173–97.
Descartes, René, 1641–7/1984. “Meditations”. In Robert Stoothoff, John Cottingham, and
Dugald Murdoch, editors, The Philosophical Writings of Descartes, volume II. Cambridge:
Cambridge University Press.
Dosanjh, Ranpal, 2014. A Defense of Reductive Physicalism. Ph.D. thesis, University of
Toronto.
Dosanjh, Ranpal, 2019. “Token-Distinctness and the Disjunctive Strategy”. Erkenntnis,
1–18. Published online at https://doi.org/10.1007/s10670-019-00127-0.
Douven, Igor, 2017. “Abduction”. The Stanford Encyclopedia of Philosophy (Summer
2017 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/sum2017/
entries/abduction/. Accessed August 31, 2020.
Dowell, Janice, 2006. “Formulating the Thesis of Physicalism: An Introduction”. Philosoph-
Dretske, Fred I., 1995. Naturalizing the Mind. Cambridge, MA: MIT Press.
Dummett, Michael, 1975. “Wang’s Paradox”. Synthese, 30:201–32.
Earman, John, 1976. “Causation: A Matter of Life and Death”. The Journal of Philosophy,
73:5–25.
Ehring, Douglas, 1996. “Mental Causation, Determinables, and Property Instances”. Noûs,
30:461–80.
Ehring, Douglas, 2003. “Part-whole Physicalism and Mental Causation”. Synthese,
136:359–88.
Ekstrom, Laura W., 2001. Agency and Responsibility: Essays on the Metaphysics of Freedom.
Boulder, CO: Westview Press.
Emmeche, Claus, Simo Koppe, and Frederick Stjernfelt, 1997. “Explaining Emergence:
Toward an Ontology of Levels”. Journal for General Philosophy of Science/Zeitschrift für
Allgemeine Wissenschaftstheorie, 28:83–119.
Evans, Gareth, 1983. “Can There Be Vague Objects?” Analysis, 38:208.
Ewing, Alfred C., 1951. The Fundamental Questions of Philosophy. Oxford: Routledge.
Fales, Evan, 1990. Causation and Universals. New York: Routledge.
292 bibliography
Farrell, B. A., 1950. “Experience”. Mind, 59:170–98.

Feigl, Herbert, 1959. “The ‘Mental’ and the ‘Physical’ ”. In Grover Maxwell, Herbert Feigl,
and Michael Scriven, editors, Minnesota Studies in the Philosophy of Science, volume 2.
Minneapolis, MN: University of Minnesota Press.
Feigl, Herbert, 1967. The “Mental” and the “Physical”: The Essay and a Postscript. Minneapo-
lis, MN: University of Minnesota Press.
Feyerabend, Paul K., 1963. “Mental Events and the Brain”. Journal of Philosophy, 40:295–6.
Feynman, Richard, 1963. The Feynman Lectures on Physics, Volumes I-III. Boston, MA:
Addison Wesley.
Feynman, Richard, 1965. The Character of Physical Law. Cambridge, MA: MIT Press.
Field, Hartry, 2003. “Causation in a Physical World”. In Michael Loux and Dean W.
Zimmerman, editors, The Oxford Handbook of Metaphysics, 435–460. Oxford: Oxford
University Press.
Fine, Kit, 2001. “The Question of Realism”. Philosophers’ Imprint, 1:1–30.
Fine, Kit, 2003. “The Non-Identity of a Material Thing and its Matter”. Mind, 112:195–234.
Flanagan, Owen, 1992. Consciousness Reconsidered. Cambridge, MA: MIT Press.
Fodor, Jerry, 1974. “Special Sciences (Or, The Disunity of Science as a Working Hypothesis)”.
Synthese, 28:77–115.
Forster, Malcolm and Alexey Kryukov, 2003. “The Emergence of the Macroworld: A Study
of Intertheory Relations in Classical and Quantum Mechanics”. Philosophy of Science,
70:1039–51.
Francescotti, Robert, 2007. “Emergence”. Erkenntnis, 67:47–63.
Frankfurt, Harry G., 1971. “Freedom of the Will and the Concept of a Person”. The Journal
of Philosophy, 68:5–20.
Franklin, Christopher Evan, 2018. A Minimal Libertarianism: Free Will and the Promise of
Reduction. New York: Oxford University Press.
Frege, Gottlob, 1892/2010. “On Sense and Reference”. In Darragh Byrne and Max Kölbel,
editors, Arguing About Language, 36–56. London: Routledge.
Frigg, Roman, 2009. “GRW Theory: Ghirardi, Rimini, Weber Model of Quantum Mechan-
ics”. In Klaus Hentschel Daniel Greenberger, Brigitte Falkenburg, and Friedel Weinert,
editors, Compendium of Quantum Physics: Concepts, Experiments, History and Philoso-
phy, 266–70. Berlin: Springer.
Funkhouser, Eric, 2006. “The Determinable-Determinate Relation”. Noûs, 40:548–69.
Garcia, Robert K., 2014. “Closing in on Causal Closure”. Journal of Consciousness Studies,
21:96–109.
Gertler, Brie, 2002. “Explanatory Reduction, Conceptual Analysis, and Conceivability
Arguments About the Mind”. Noûs, 36:22–49.
Gertler, Brie, 2006. “Consciousness and Qualia Cannot Be Reduced”. In Robert J. Stainton,
editor, Contemporary Debates in Cognitive Science, 202–16. Malden, MA: Blackwell.
Ghirardi, G., A. Rimini, and T. Weber, 1986. “Unified Dynamics for Microscopic and
Macroscopic Systems”. The Physical Review D, 34:470–9.
Gibb, Sophie C., Robin Findlay Hendry, and Tom Lancaster, 2018. The Routledge Handbook
of Emergence. Oxford: Routledge.
Gibb, Sophie C., 2013. “The Entailment Problem and the Subset Account of Property
Realization”. Australasian Journal of Philosophy, 92:551–66.
Gibbard, Allan, 1975. “Contingent Identity”. Journal of Philosophical Logic, 4:187–222.
Gillett, Carl, 2002a. “The Dimensions of Realization: A Critique of the Standard View”.
Analysis, 62:316–23.
bibliography 293
Gillett, Carl, 2002b. “The Varieties of Emergence: Their Purposes, Obligations and Impor-
tance”. Grazer Philosophische Studien, 65:95–121.
Gillett, Carl, 2010. “Moving Beyond the Subset Model of Realization: The Problem of
Qualitative Distinctness in the Metaphysics of Science”. Synthese, 177:165–92.
Gillett, Carl, 2016. Reduction and Emergence in Science and Philosophy. Cambridge:
Gillett, Carl and Barry M. Loewer, editors, 2001. Physicalism and Its Discontents. Cambridge:
Ginet, Carl, 1990. On Action. Cambridge: Cambridge University Press.
Ginet, Carl, 2002. “Reasons Explanations of Action: Causalist Versus Noncausalist
Accounts”. In Robert H. Kane, editor, The Oxford Handbook on Free Will, 386–405.
Oxford: Oxford University Press.
Grandy, Richard E., 2016. “Sortals”. The Stanford Encyclopedia of Philosophy (Winter
2016 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/win2016/
entries/sortals/. Accessed August 31, 2020.
Grandy, Richard E., 2007. “Artifacts: Parts and Principles”. In Eric Margolis and Stephen
Laurence, editors, Creations of the Mind: Theories of Artifacts and Their Representation,
18–32. New York: Oxford University Press.
Greiner, Walter, 1996. “The Discovery of the Weak Interaction”. In Walter Greiner and
Berndt Müller, editors, Gauge Theory of Weak Interactions, 1–24. Berlin, Heidelberg:
Springer.
Guay, Alexandre and Olivier Sartenaer, 2016. “A New Look at Emergence. Or When After
is Different”. European Journal for Philosophy of Science, 6:297–322.
Haggard, Patrick and Benjamin W. Libet, 2001. “Conscious Intention and Brain Activity”.
Journal of Consciousness Studies, 8:47–63.
Hall, Ned, 2004. “Two Concepts of Causation”. In John Collins, Ned Hall, and Laurie Paul,
editors, Causation and Counterfactuals, 225–76. Stanford, MA: MIT Press.
Harman, Gilbert H., 1965. “The Inference to the Best Explanation”. Philosophical Review,
74:88–95.
Harman, Gilbert H., 1990. “The Intrinsic Quality of Experience”. In James Tomberlin, editor,
Action Theory and the Philosophy of Mind, volume 4 of Philosophical Perspectives, 31–52.
Atascadero, CA: Ridgeview. Reprinted in Harman 1999.
Harman, Gilbert H., 1999. Reasoning, Meaning, and Mind. Oxford: Oxford University Press.
Haug, Matthew C., 2010. “Realization, Determination, and Mechanisms”. Philosophical
Studies, 150:313–30.
Hawthorne, John, 2001. “Causal Structuralism”. Noûs, 35:361–78.
Hawthorne, John, 2002. “Deeply Contingent a Priori Knowledge”. Philosophy and Phe-
nomenological Research, 65:247–69.
Hawthorne, John and Philip Pettit, 1996. “Strategies for Free Will Compatibilists”. Analysis,
56.4:191–201. Hawthorne originally published this article under the name ‘John O’Leary-
Hawthorne’.
Heil, John, 1992. The Nature of True Minds. Cambridge: Cambridge University Press.
Heil, John, 2003a. From an Ontological Point of View. Oxford: Clarendon Press.
Heil, John, 2003b. “Levels of Reality”. Ratio, 16:205–21.
Hellie, Benj, 2007. “ ‘There’s Something It’s Like’ and the Structure of Consciousness”.
Philosophical Review, 116:441–63.
294 bibliography
Hellie, Benj, 2019a. “An Analytic-Hermeneutic History of Consciousness”. In Kelly Michael

Becker and Iain Thompson, editors, Cambridge History of Philosophy: 1945-2015, 74–89.
Cambridge: Cambridge University Press.
Hellie, Benj, 2019b. “Semantic Gaps and Protosemantics”. In Acacio de Barros and Carlos
Montemayor, editors, Mind and Quanta, 201–22. Berlin: Springer.
Hellie, Benj, forthcoming. Out of This World: Logical Mentalism and the Philosophy of Mind.
New York: Oxford University Press.
Hellman, Geoffrey and Frank Thompson, 1975. “Physicalism: Ontology, Determination,
and Reduction”. The Journal of Philosophy, 72:551–64.
Hempel, Carl G., 1979. “Comment at a Symposium on Nelson Goodman’s Ways of World-
making”. Presented at the 76th Annual Meeting of the American Philosophical Association.
Hempel, Carl G. and Paul Oppenheim, 1948. “Studies in the Logic of Explanation”. Philos-
ophy of Science, 15:135–75.
Hendry, Robin Findlay, 2010. “Emergence vs. Reduction in Chemistry”. In Cynthia F.
Macdonald and Graham Macdonald, editors, Emergence in Mind, 205–21. New York:
Hendry, Robin Findlay, 2017. “Prospects for Strong Emergence in Chemistry”. In Michele
Paolini Paoletti and Francesco Orilia, editors, Philosophical and Scientific Perspectives on
Downward Causation, 146–63. New York: Routledge.
Hill, Christopher S., 2009. Consciousness. New York: Cambridge University Press.
Hitchcock, Christopher Read, 1996. “The Role of Contrast in Causal and Explanatory
Claims”. Synthese, 107:395–419.
Hodgeson, Shadworth, 1962. The Theory of Practice: An Ethical Enquiry. London: Long-
mans, Green, Reader, and Dyer.
Hooker, C. A., 2004. “Asymptotics, Reduction and Emergence”. British Journal for the
Horgan, Terence, 1982. “Supervenience and Microphysics”. Pacific Philosophical Quarterly,
63:29–43.
Horgan, Terence, 1989. “Mental Quausation”. Philosophical Perspectives 3: Philosophy of
Mind and Action Theory, 47–76.
Horgan, Terence, 1993. “From Supervenience to Superdupervenience: Meeting the
Demands of a Material World”. Mind, 102:555–86.
Howell, Robert J., 2009. “Emergentism and Supervenience Physicalism”. Australasian Jour-
nal of Philosophy, 87:83–98.
Huggett, Nick and Christian Wüthrich, 2013. “Emergent Spacetime and Empirical
(in)Coherence”. Studies in History and Philosophy of Modern Physics, 44:276–85.
Humphreys, Paul, 1996. “Aspects of Emergence”. Philosophical Topics, 24:53–70.
Humphreys, Paul, 1997. “How Properties Emerge”. Philosophy of Science, 64:1–17.
Humphreys, Paul, 2016. Emergence: A Philosophical Account. Oxford: Oxford University
Press.
Huxley, Thomas, 1874. “On the Hypothesis That Animals Are Automata, and its History”.
Fortnightly Review, 95:555–80.
Jackson, Frank, 1982. “Epiphenomenal Qualia”. Philosophical Quarterly, 32:127–36.
Jackson, Frank, 1986. “What Mary Didn’t Know”. The Journal of Philosophy, 83:291–5.
Jackson, Frank, 1994. “Armchair Metaphysics”. In John O’Leary-Hawthorne and Michaelis
Michael, editors, Philosophy in Mind, 23–42. Dordrecht: Kluwer.
Jaworski, William, 2002. “Multiple-Realizability, Explanation and the Disjunctive Move”.
Philosophical Studies, 108:289–308.
bibliography 295
Jenkins, C. S., 2013. “Explanation and Fundamentality”. In Miguel Hoeltje, Benjamin

Schnieder, and Alex Steinberg, editors, Varieties of Dependence (Basic Philosophical
Concepts Series), 211–42. Munich: Philosophia Verlag.
Judd, D. and G. Wyszecki, 1975. Color in Business, Science, and Industry. New York: Wiley.
Kane, Gordon, 1993. Modern Elementary Particle Physics: the Fundamental Particles and
Forces. Boulder, CO: Westview Press.
Kauffman, Stuart, 1993. The Origins of Order. Oxford: Oxford University Press.
Kauffman, Stuart, 1995. At Home in the Universe: The Search for Laws of Self-organization
and Complexity. New York: Oxford University Press.
Kim, Jaegwon, 1984. “Concepts of Supervenience”. Philosophy and Phenomenological
Research, 45:153–76.
Kim, Jaegwon, 1989. “The Myth of Nonreductive Materialism”. Proceedings and Addresses
of the American Philosophical Association, 63:31–47. Reprinted in Kim 1993b.
Kim, Jaegwon, 1990. “Supervenience as a Philosophical Concept”. Metaphilosophy, 21:1–27.
Reprinted in Kim 1993b.
Kim, Jaegwon, 1992a. “ “Downward Causation” in Emergentism and Nonreductive Phys-
icalism”. In Ansgar Beckermann, Hans Flohr, and Jaegwon Kim, editors, Emergence or
Reduction?: Prospects for Nonreductive Physicalism, 119–38. Berlin: De Gruyter.
Kim, Jaegwon, 1992b. “Multiple Realization and the Metaphysics of Reduction”. Philosophy
and Phenomenological Research, 52:1–26. Reprinted in Kim 1993b.
Kim, Jaegwon, 1993a. “The Non-Reductivist’s Troubles with Mental Causation”. In John Heil
and Alfred Mele, editors, Mental Causation, 189–210. Oxford: Oxford University Press.
Reprinted in Kim 1993b.
Kim, Jaegwon, 1993b. Supervenience and Mind: Selected Philosophical Essays. Cambridge:
Kim, Jaegwon, 1998. Mind in a Physical World. Cambridge, MA: MIT Press.
Kim, Jaegwon, 1999. “Making Sense of Emergence”. Philosophical Studies, 95:3–36.
Kim, Jaegwon, 2006. “Emergence: Core ideas and issues”. Synthese, 151:547–59.
Kim, Jaegwon, 2010. “Thoughts on Sydney Shoemaker’s Physical Realization”. Philosophical
Studies, 148:101–12.
Kim, Jaegwon, 2015. “What Could Pair a Nonphysical Soul to a Physical Body?” In Keith
Augustine and Michael Martin, editors, The Myth of an Afterlife: The Case against Life
After Death, 335–47. Lanham, MD: Rowman & Littlefield.
Kirchhoff, Michael, 2014. “In Search of Ontological Emergence: Diachronic, But Non-
Supervenient”. Axiomathes, 24:89–116.
Kirk, Robert, 2019. “Zombies”. The Stanford Encyclopedia of Philosophy (Spring 2019
Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/spr2019/
entries/zombies/. Accessed August 31, 2020.
Klee, Robert, 1984. “Micro-Determinism and Concepts of Emergence”. Philosophy of Sci-
ence, 51:44–63.
Kleman, Maurice and Oleg Lavrentovich, 2003. Soft Matter Physics: An Introduction.
New York: Springer.
Knox, Eleanor, 2013. “Effective Spacetime Geometry”. Studies in History and Philosophy of
Science Part B: Studies in History and Philosophy of Modern Physics, 44:346–56.
Knox, Eleanor, 2018. “Physical Relativity From a Functionalist Perspective”. Studies in
History and Philosophy of Science Part B: Studies in History and Philosophy of Modern
Physics, 67:118–24.
296 bibliography
Korman, Daniel Z., 2020. “Ordinary Objects”. The Stanford Encyclopedia of Philoso-
phy (Fall 2020 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/
fall2020/entries/ordinary-objects/. Accessed August 31, 2020.
Koslicki, Kathrin, 2008. The Structure of Objects. New York: Oxford University Press.
Koslicki, Kathrin, 2012. “Varieties of Ontological Dependence”. In Fabrice Correia and
Benjamin Schnieder, editors, Metaphysical Grounding: Understanding the Structure of
Reality, 186. Cambridge: Cambridge University Press.
Koslicki, Kathrin, 2016. “Where Grounding and Causation Part Ways: Comments on
Jonathan Schaffer”. Philosophical Studies, 173:101–12.
Kovacs, David, 2019. “The Myth of the Myth of Supervenience”. Philosophical Studies,
176:1967–1989.
Kripke, Saul, 1972/80. Naming and Necessity. Cambridge, MA: Harvard University Press.
Ladyman, James, James Lambert, and Karoline Wiesner, 2013. “What is a Complex Sys-
tem?” European Journal for Philosophy of Science, 3:33–67.
Ladyman, James and Don Ross, 2007. Every Thing Must Go: Metaphysics Naturalized.
Lam, Vincent and Christian Wüthrich, 2018. “Spacetime is as Spacetime Does”. Studies in
History and Philosophy of Science Part B: Studies in History and Philosophy of Modern
Physics, 64:39–51.
Lamb, Maurice, 2015. Characteristics of Non-reductive Explanations in Complex Dynamical
Systems Research. Ph.D. thesis, University of Cincinnati.
Lancaster, Tom and Mark Pexton, 2015. “Reduction and Emergence in the Fractional
Quantum Hall State”. Studies in History and Philosophy of Science Part B: Studies in
History and Philosophy of Modern Physics, 52:343–57.
Langsam, Harold, 2011. The Wonder of Consciousness: Understanding the Mind Through
Philosophical Reflection. Cambridge, MA: MIT Press.
Leibniz, Gottfried Wilhelm, 1714. The Monadology.
Lepore, Ernest and Barry M. Loewer, 1987. “Mind Matters”. The Journal of Philosophy,
84:630–42.
Lepore, Ernest and Barry M. Loewer, 1989. “More on Making Mind Matter”. Philosophical
Topics, 17:175–91.
Levin, Janet, 2016. “Review of Thomas Polger and Stuart Shapiro, The Multiple Realiza-
tion Book”. Notre Dame Philosophical Reviews. Published online at https://ndpr.nd.edu/
news/the-multiple-realizat.
Levin, Simon A., 2005. “Self-organization and the Emergence of Complexity in Ecological
Systems”. BioScience, 55:1075–9.
Levine, Joseph M., 1983. “Materialism and Qualia: The Explanatory Gap”. Pacific Philosoph-
ical Quarterly, 64:354–61.
Levine, Joseph M., 2001. Purple Haze: The Puzzle of Consciousness. New York: Oxford
University Press.
Lewes, G. H., 1875. Problems of Life and Mind. London: Kegan Paul, Trench, Trübner & Co.
Lewis, David, 1966. “An Argument for the Identity Theory”. The Journal of Philosophy,
63:17–25. Reprinted in Lewis 1983b.
Lewis, David, 1981. “Are We Free to Break the Laws?” Theoria, 47:113–21.
Lewis, David, 1983a. “New Work for a Theory of Universals”. Australasian Journal of
Lewis, David, 1983b. Philosophical Papers, volume I. Oxford: Oxford University Press.
Lewis, David, 1986. On the Plurality of Worlds. London: Blackwell.
bibliography 297
Lewis, David, 1990. “What Experience Teaches”. In William G. Lycan, editor, Mind and
Cognition: An Anthology, 29–57. Oxford: Blackwell. Reprinted in Lewis 1999.
Lewis, David, 1999. Papers in Metaphysics and Epistemology. Cambridge: Cambridge Uni-
versity Press.
Lewtas, Patrick, 2013. “Emergence and Consciousness”. Philosophy, 88:527–53.
Libet, Benjamin W., 2002. “The Timing of Mental Events: Libet’s Experimental Findings
and Their Implications”. Consciousness and Cognition, 11:291–9.
Libet, Benjamin W., 1999. “Do We Have Free Will?” Journal of Consciousness Studies, 6:
47–57.
List, Christian and Peter Menzies, 2009. “Nonreductive Physicalism and the Limits of the
Exclusion Principle”. Journal of Philosophy, 106:475–502.
Locke, John, 1690. An Essay Concerning Human Understanding.
Loewer, Barry M., 2001. “From Physics to Physicalism”. In Carl Gillett and Barry M. Loewer,
editors, Physicalism and Its Discontents, 37–56. Cambridge: Cambridge University Press.
Lowe, E. J., 1989. “What Is a Criterion of Identity?” Philosophical Quarterly, 39:1–21.
Lowe, E. J., 1998. The Possibility of Metaphysics: Substance, Identity, and Time. New York:
Lowe, E. J., 2007. “Sortals and the Individuation of Objects”. Mind and Language, 22:514–33.
MacBride, Fraser, 1999. “Could Armstrong have been a Universal?” Mind, 108:471–501.
Macdonald, Cynthia and Graham Macdonald 1986. “Mental Causes and Explanation of
Action”. In L. Stevenson, R. Squires, and J. Haldane, editors, Mind, Causation, and Action,
145–58. Oxford: Basil Blackwell.
Macdonald, Cynthia and Graham Macdonald, 1995. “How to be Psychologically Rele-
vant”. In Philosophy of Psychology: Debates on Psychological Explanation, 60–77. Oxford:
Blackwell.
Maddy, Penelope, 1997. Naturalism in Mathematics. New York: Oxford University Press.
Malcolm, Norman, 1968. “The Conceivability of Mechanism”. The Philosophical Review,
77:45–72.
Margolis, Eric and Stephen Laurence, editors, 2007. Creations of the Mind: Theories of
Artifacts and Their Representation. Oxford: Oxford University Press.
Martin, C. B., 1996. “Properties and Dispositions”. In Tim Crane, editor, Dispositions: A
Debate, 71–81. London: Routledge.
Mayoral, E. and A. Robledo, 1970. “A Recent Appreciation of the Singular Dynamics at the
Edge of Chaos”. In Marcel Ausloos and Michel Dirickx, editors, The Logistic Map and the
Route to Chaos: From the Beginnings to Modern Applications, 339–54. Berlin: Springer.
McCann, Hugh, 1998. The Works of Agency: On Human Action, Will, and Freedom. Ithaca,
NY: Cornell University Press.
McDaniel, Kris, 2001. “Tropes and Ordinary Physical Objects”. Philosophical Studies,
104:269–90.
McKenna, Michael and D. Justin Coates, 2020. “Compatibilism”. In The Stanford Encyclope-
dia of Philosophy (Summer 2020 Edition), Edward N. Zalta, editor. https://plato.stanford.
edu/archives/sum2020/entries/compatibilism/. Accessed August 31, 2020.
McLaughlin, Brian P., 1992. “The Rise and Fall of British Emergentism”. In Ansgar Beck-
erman, Hans Flohr, and Jaegwon Kim, editors, Emergence or Reduction? Essays on the
Prospects of Non-reductive Physicalism, 49–93. Berlin: De Gruyter.
McLaughlin, Brian P. and Karen Bennett, 2018. “Supervenience”. In The Stanford Encyclope-
dia of Philosophy (Winter 2018 Edition), Edward N. Zalta, editor. https://plato.stanford.
edu/archives/win2018/entries/supervenience/. Accessed August 31, 2020.
McLaughlin, Brian P., 2007. “Mental Causation and Shoemaker-Realization”. Erkenntnis,
67:149–72.
298 bibliography
McLeish, Tom C. B., Mark Pexton, and Tom Lancaster, 2019. “Emergence and Topological
Order in Classical and Quantum Systems”. Studies in History and Philosophy of Science
Part B: Studies in History and Philosophy of Modern Physics, 66:155–69.
McLeish, Tom C. B., 2017. “Strong Emergence and Downward Causation in Biological
Physics”. Philosophica, 92:113–38.
Megill, Jason, 2013. “A Defense of Emergence”. Axiomathes, 23:597–615.
Mele, Alfred, 2009. Effective Intentions: The Power of Conscious Will. Oxford: Oxford
University Press.
Melnyk, Andrew, 1997. “How to Keep the ‘Physical’ in Physicalism”. The Journal of Philos-
ophy, 94:622–37.
Melnyk, Andrew, 1999. “Supercalifragilisticexpialidocious”. Noûs, 33:144–54.
Melnyk, Andrew, 2003. A Physicalist Manifesto: Thoroughly Modern Materialism. New York:
Melnyk, Andrew, 2006. “Realization-based Formulations of Physicalism”. Philosophical
Studies, 131:127–55.
Menon, Tarun and Craig Callender, 2013. “Ch-Ch-Changes Philosophical Questions
Raised by Phase Transitions”. In Robert W. Batterman, editor, The Oxford Handbook of
Philosophy of Physics, 189–212. New York: Oxford University Press.
Merricks, Trenton, 2003. Objects and Persons. Oxford: Clarendon Press.
Messiah, Albert, 1970. Quantum Mechanics. Amsterdam: North-Holland.
Mill, John S., 1843/1973. A System of Logic. Toronto: University of Toronto Press.
Mitchell, Sandra D., 2012. “Emergence: Logical, Functional and Dynamical”. Synthese,
185:171–86.
Montero, Barbara, 2003. “Varieties of Causal Closure”. In Sven Walter and Heinz-Dieter
Heckmann, editors, Physicalism and Mental Causation, 173–87. Exeter: Imprint Aca-
demic.
Montero, Barbara, 2006. “Physicalism in an Infinitely Decomposable World”. Erkenntnis,
64:177–91.
Monton, Bradley, 2004. “The Problem of Ontology on Spontaneous Collapse Theories”.
Studies in History and Philosophy of Modern Physics, 35:407–21.
Moore, G. E., 1903. Principia Ethica. Cambridge: Cambridge University Press.
Mørch, Hedda Hassel, 2018. “The Evolutionary Argument for Phenomenal Powers”. Philo-
sophical Perspectives, 31: 293–316.
Mørch, Hedda Hassel (2020). “The Phenomenal Powers View and the Meta-Problem of
Consciousness”. Journal of Consciousness Studies, 27:131–42.
Morgan, C. Lloyd, 1923. Emergent Evolution. London: Williams & Norgate.
Morris, Kevin, 2011. “Subset Realization, Parthood, and Causal Overdetermination”. Pacific
Philosophical Quarterly, 92:363–79.
Morris, Kevin, 2013. “On Two Arguments for Subset Inheritance”. Philosophical Studies,
163:197–211.
Morris, Kevin, 2014. “Supervenience Physicalism, Emergentism, and the Polluted Superve-
nience Base”. Erkenntnis, 79:351–65.
Morris, Kevin, 2018. Physicalism Deconstructed: Levels of Reality and the Mind–Body
Problem. Cambridge: Cambridge University Press.
Morrison, Margaret, 2006. “Emergence, Reduction, and Theoretical Principles: Rethinking
Fundamentalism”. Philosophy of Science, 73:876–87.
bibliography 299
Morrison, Margaret, 2012. “Emergent Physics and Micro-Ontology”. Philosophy of Science,

79:141–66.
Mossio, Matteo, Leonardo Bich, and Alvaro Moreno, 2013. “Emergence, Closure and Inter-
Level Causation in Biological Systems”. Erkenntnis, 78:153–78.
Nagel, Ernest, 1961. The Structure of Science. London: Routledge & Kegan Paul.
Nagel, Thomas, 1974. “What is it Like to be a Bat?” The Philosophical Review, 83:435–50.
Nagel, Thomas, 1979. “Panpsychism”. In Thomas Nagel, editor, Mortal Questions, 181–95.
Binghamton, NY: Cambridge University Press.
Nemirow, Laürence, 2006. “So This is What It’s Like: A Defense of the Ability Hypothesis”. In
Torin Alter and Sven Walter, editors, Phenomenal Concepts and Phenomenal Knowledge:
New Essays on Consciousness and Physicalism, 32–51. New York: Oxford University Press.
Newman, David, 1996. “Emergence and Strange Attractors”. Philosophy of Science, 63:
245–61.
Ney, Alyssa, 2008. “Defining Physicalism”. Philosophy Compass, 3:1033–48.
Ney, Alyssa, 2010. “Convergence on the Problem of Mental Causation: Shoemaker’s Strategy
for (Nonreductive?) Physicalists”. Philosophical Issues, 20:438–45.
Nida-Rümelin, Martine and Donnchadh O Conaill, 2019. “Qualia: The Knowledge
Argument”. The Stanford Encyclopedia of Philosophy (Winter 2019 Edition), Edward
N. Zalta, editor. https://plato.stanford.edu/archives/win2019/entries/qualia-knowledge/.
Accessed August 31, 2020.
Noordhof, Paul, 2010. “Emergent Causation and Property Causation”. In Cynthia
Macdonald and Graham F. Macdonald, editors, Emergence in Mind, 69–99. Oxford:
Nozick, Robert, 1995. “Choice and Indeterminism”. In Timothy O’Connor, editor, Agents,
Causes, and Events: Essays on Indeterminism and Free Will, 101–14. New York: Oxford
University Press.
O’Connor, Timothy, 1994. “Emergent properties”. American Philosophical Quarterly,
31: 91–104.
O’Connor, Timothy, 2000. Persons and Causes: The Metaphysics of Free Will. New York:
O’Connor, Timothy, 2005. “Free Will”. In John Martin Fischer, editor, Free Will: Critical
Concepts in Philosophy, 7–32. New York: Routledge. Originally published in The Stanford
Encyclopedia of Philosophy (2002 Edition).
O’Connor, Timothy, 2009a. “Agent-Causal Power”. In Toby Handfield, editor, Dispositions
and Causes, 189–214. New York: Oxford University Press.
O’Connor, Timothy, 2009b. “Conscious Willing and the Emerging Sciences of Brain and
Behavior”. In Nancey Murphy, George Ellis, and Timothy O’Connor, editors, Downward
Causation and the Neurobiology of Free Will, 173–86. Berlin: Springer.
O’Connor, Timothy and Hong Yu Wong, 2005. “The Metaphysics of Emergence”. Noûs,
39:658–78.
O’Connor, Timothy and Hong Yu Wong, 2015. “Emergent Properties”, The Stanford
Encyclopedia of Philosophy (Fall 2020 Edition), Edward N. Zalta, editor. https://plato.
stanford.edu/archives/fall2020/entries/properties-emergent/. Accessed August 31, 2020.
Owens, David, 1989. “Levels of Explanation”. Mind, 98:59–79.
Page, R. E. and Sandra D. Mitchell, 1991. “Self Organization and Adaptation in Insect
Societies”. Proceedings of the Philosophy of Science Association, 2:289–98.
Page, R. E. and Sandra D. Mitchell, 1998. “Self Organization and the Evolution of Division
of Labor”. Apidologie, 29:101–20.
300 bibliography
Paolini Paoletti, Michele, 2017. The Quest for Emergence. Munich: Philosophia.
Paolini Paoletti, Michele, 2018a. “Emergent Powers”. Topoi, 1–14.
Paolini Paoletti, Michele, 2018b. “Formulating Emergence”. Ratio, 31:1–18.
Papineau, David, 1993. Philosophical Naturalism. Oxford: Basil Blackwell.
Papineau, David, 2001. “The Rise of Physicalism”. In Carl Gillett and Barry M. Loewer,
editors, Physicalism and Its Discontents, 3–36. Cambridge: Cambridge University Press.
Paul, L. A., 2002. “Logical Parts”. Noûs, 36:578–96.
Paul, L. A., 2010. “The Puzzles of Material Constitution”. Philosophy Compass, 5:579–90.
Paull, C. P. and Theodore Sider, 1992. “In Defense of Global Supervenience”. Philosophy and
Peacocke, Christopher, 1993. “How Are A Priori Truths Possible?” European Journal of
Peacocke, Christopher, 2003. The Realm of Reason. New York: Oxford University Press.
Pearson, Olley, 2018. “Emergence, Dependence, and Fundamentality”. Erkenntnis, 83:
391–402.
Pereboom, Derk, 2002. “Robust Non-reductive Materialism”. Journal of Philosophy,
99: 499–531.
Pereboom, Derk, 2011. Consciousness and the Prospects of Physicalism. New York: Oxford
University Press.
Pereboom, Derk and Hilary Kornblith, 1991. “The Metaphysics of Irreducibility”. Philosoph-
Perry, John, 2001. Knowledge, Possibility, and Consciousness. Cambridge, MA: MIT Press,
second edition.
Pettit, Philip, 1995. “Microphysicalism, Dottism, and Reduction”. Analysis, 55:141–6.
Pitts, J. Brian, 2020. “Conservation Laws and the Philosophy of Mind: Opening the Black
Box, Finding a Mirror”. Philosophia, 48:673–707.
Poland, Jeffrey, 1994. Physicalism; the Philosophical Foundations. Oxford: Clarendon Press.
Polger, Thomas W., 2007. “Realization and the Metaphysics of Mind”. Australasian Journal
of Philosophy, 85:233–59.
Polger, Thomas W. and Lawrence A. Shapiro, 2016. The Multiple Realization Book. Oxford:
Popper, Karl R. and John C. Eccles, 1977. The Self and Its Brain: An Argument for Interac-
tionism. Oxfordshire: Springer.
Putnam, Hilary, 1967. “Psychological Predicates”. In Art, Mind, and Religion, 37–48. Pitts-
burgh, PA: University of Pittsburgh Press.
Raffman, Diana, 1994. “Vagueness Without Paradox”. Philosophical Review, 103:41–74.
Raven, Michael J., 2015. “Ground”. Philosophy Compass, 10:322–33.
Raven, Michael J., 2016. “Fundamentality Without Foundations”. Philosophy and Phe-
nomenological Research, 93:607–26.
Ravenscroft, Ian, 1997. “Physical Properties”. Southern Journal of Philosophy, 35:419–31.
Robb, David, 1997. “The Properties of Mental Causation”. Philosophical Quarterly, 47:
178–94.
Robinson, William, 2019. “Epiphenomenalism”. The Stanford Encyclopedia of Philos-
ophy (Summer 2019 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/
archives/sum2019/entries/epiphenomenalism/. Accessed August 31, 2020.
Rosen, Gideon, 2010. “Metaphysical Dependence: Grounding and Reduction”. In B. Hale
and A. Hoffmann, editors, Modality: Metaphysics, Logic, and Epistemology, 109–36.
bibliography 301
Rosen, J., 2007 (April 22). “Flight Patterns”. New York Times, 255–73.
Rosenberg, Alexander, 1994. Instrumental Biology, or, the Disunity of Science. Chicago, IL:
University of Chicago Press.
Rueger, Alexander, 2000. “Physical Emergence, Diachronic and Synchronic”. Synthese,
124:297–322.
Rueger, Alexander and Patrick McGivern, 2010. “Hierarchies and Levels of Reality”. Syn-
these, 176:379–97.
Russell, Bertrand, 1912. “On the Notion of Cause”. Proceedings of the Aristotelian Society,
13:1–26.
Russell, Bruce, 2020. “A Priori Justification and Knowledge”. The Stanford
Encyclopedia of Philosophy (Summer 2020 Edition), Edward N. Zalta, editor.
https://plato.stanford.edu/archives/sum2020/entries/apriori/. Accessed August 31,
2020.
Salmon, Nathan, 1989. “The Logic of What Might Have Been”. Philosophical Review, 98:
3–34.
Santos, Gil C., 2020. “Integrated-Structure Emergence and Its Mechanistic Explanation”.
Synthese. https://doi.org/10.1007/s11229-020-02594-3.
Santos, Gil C., 2015. “Upward and Downward Causation From a Relational-Horizontal
Ontological Perspective”. Axiomathes, 25:23–40.
Schaffer, Jonathan, 2003. “Is There a Fundamental Level?” Noûs, 37:498–517.
Schaffer, Jonathan, 2004. “Quiddistic Knowledge”. In Frank Jackson and Graham Priest,
editors, Lewisian Themes, 210–30. Oxford: Oxford University Press.
Schaffer, Jonathan, 2009. “On What Grounds What”. In David J. Chalmers, David Manley,
and Ryan Wasserman, editors, Metametaphysics: New Essays on the Foundations of
Ontology, 347–83. Oxford: Oxford University Press.
Schaffer, Jonathan, 2016a. “Ground Rules: Lessons from Wilson”. In Ken Aizawa and
Carl Gillet, editors, Scientific Composition and Metaphysical Ground, 143–69. London:
Palgrave-Macmillan.
Schaffer, Jonathan, 2016b. “Grounding in the Image of Causation”. Philosophical Studies,
173:49–100.
Schroder, Jurgen, 1998. “Emergence: Non-Deducibility or Downward Causation?” The
Seager, William, 1999/2016. Theories of Consciousness: An Introduction and Assessment.
Abingdon: Routledge.
Seager, William, 2012. “Emergentist Panpsychism”. Journal of Consciousness Studies, 19:
9–10.
Searle, John R., 1992. The Rediscovery of the Mind. Cambridge, MA: MIT Press.
Searle, John R., 2007. “Social Ontology and the Philosophy of Society”. In Eric Margolis
and Stephen Laurence, editors, Creations of the MInd: Theories of Artifacts and Their
Representation, 3–17. New York: Oxford University Press.
Shoemaker, Sydney, 1975. “Functionalism and Qualia”. Philosophical Studies, 27:292–315.
Shoemaker, Sydney, 1980. “Causality and Properties”. In Peter van Inwagen, editor, Time
and Cause, 109–35. Dordrecht: D. Reidel.
Shoemaker, Sydney, 1998. “Causal and Metaphysical Necessity”. Pacific Philosophical Quar-
terly, 79:59–77.
Shoemaker, Sydney, 2000/2001. “Realization and Mental Causation”. In Proceedings of the
20th World Congress in Philosophy, 23–33. Cambridge, MA: Philosophy Documentation
Center. Published in revised form in Gillett and Loewer 2001, 74–98.
302 bibliography
Shoemaker, Sydney, 2003. “Realization, Micro-Realization, and Coincidence”. Philosophy

and Phenomenological Research, 67:1–23.
Shoemaker, Sydney, 2007. Physical Realization. Oxford: Oxford University Press.
Sider, Theodore, 2009. “Ontological Realism”. In David J. Chalmers, David Manley, and
Ryan Wasserman, editors, Metametaphysics: New Essays on the Foundations of Ontology,
384–423. Oxford: Oxford University Press.
Sider, Theodore, 2011. Writing the Book of the World. Oxford: Oxford University Press.
Silberstein, Michael, 2009. “Emergence”. In Tim Bayne, Axel Cleeremans, and Patrick
Wilken, editors, The Oxford Companion to Consciousness, volume 50, 254–7. New York:
Silberstein, Michael and J. McGeever, 1999. “The Search for Ontological Emergence”.
Simons, Peter M., 1994. “Particulars in Particular Clothing: Three Trope Theories of
Substance”. Philosophy and Phenomenological Research, 54:553–75.
Simons, Peter M., 1987. Parts: A Study in Ontology, volume 100. New York: Oxford
University Press.
Smart, J. J. C., 1958. “Sensations and Brain Processes”. The Philosophical Review, 68:141–56.
Smith, John A., 2006. Qualitative Complexity: Ecology, Cognitive Processes and the Re-
Emergence of Structures in Post-Humanist Social Theory. London: Routledge.
Sperry, Roger, 1986. “Discussion: Macro- Versus Micro-Determination”. Philosophy of
Science, 53: 265–70.
Stalnaker, Robert C., 1996. “Varieties of Supervenience”. Philosophical Perspectives, 10:
221–42.
Stalnaker, Robert C., 1999. Context and Content. Oxford: Oxford University Press.
Stephan, Achim, 2002. “Emergentism, irreducibility, and downward causation”. Grazer
Philosophische Studien, 65:77–93.
Steward, Helen, 2015. “What is Determinism”. Flickers of Freedom blog. https://philo
sophycommons.typepad.com/flickers_of_freedom/2015/07/what-is-determinism.html.
Stoljar, Daniel, 2007. “Distinctions in Distinction”. In Jesper Kallestrup and Jakob Hohwy,
editors, Being Reduced: New Essays on Causation and Explanation in the Special Sciences,
263–79. Oxford: Oxford University Press.
Stoljar, Daniel, 2010. Physicalism. Abingdon: Routledge.
Stoljar, Daniel, 2017. “Physicalism”. The Stanford Encyclopedia of Philosophy (Winter
2017 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/win2017/
entries/physicalism/. Accessed August 31, 2020.
Strawson, Peter F., 1962. “Freedom and Resentment”. Proceedings of the British Academy,
48:1–25.
Stump, Eleonore, 1999. “Dust, Determinism, and Frankfurt”. Faith and Philosophy, 16:
413–22.
Sturgeon, Scott, 1998. “Physicalism and Overdetermination”. Mind, 107:411–32.
Swinburne, Richard, 2001. Epistemic Justification. New York: Oxford University Press.
Swoyer, Chris, 1982. “The Nature of Natural Laws”. Australasian Journal of Philosophy,
60:203–23.
Tahko, Tuomas E., 2014. “Boring Infinite Descent”. Metaphilosophy, 45:257–69.
Tahko, Tuomas E., 2018. “Fundamentality”. The Stanford Encyclopedia of Philoso-
phy (Fall 2018 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/
fall2018/entries/fundamentality/. Accessed August 31, 2020.
bibliography 303
Taylor, Elanor, 2015. “Collapsing Emergence”. Philosophical Quarterly, 65:732–53.

Thomasson, Amie L., 2007. Ordinary Objects. New York: Oxford University Press.
Thomasson, Amie L., 2010. “The Controversy Over the Existence of Ordinary Objects”.
Philosophy Compass, 5:591–601.
Thompson, Evan, 2007. Mind in Life: Biology, Phenomenology, and the Sciences of Mind.
Cambridge, MA: Harvard University Press.
Thompson, Evan and Francisco J. Varela, 2001. “Radical Embodiment: Neural Dynamics
and Consciousness”. Trends in Cognitive Sciences, 5:418–25.
Tro, Nivaldo J., 2016. Chemistry: A Molecular Approach. New York: Pearson.
Tye, Michael, 1990. “Vague Objects”. Mind, 99:535–57.
Tye, Michael, 1995. Ten Problems of Consciousness: A Representational Theory of the Phe-
nomenal Mind. Cambridge, MA: MIT Press.
Unger, Peter, 1980. “The Problem of the Many”. Midwest Studies in Philosophy, 5:411–68.
van Cleve, James, 1990. “Mind-dust or Magic? Panpsychism versus Emergence”. Philosoph-
ical Perspectives, 4:215–26.
Van Gulick, Robert, 2001. “Reduction, Emergence and Other Recent Options on the
Mind/Body Problem: A Philosophic Overview”. Synthese, 8:1–34.
van Inwagen, Peter, 1983. An Essay on Free Will. Oxford: Clarendon Press.
van Inwagen, Peter, 1990. Material Beings. Ithaca: Cornell University Press.
Varzi, Achille, 2019. “Mereology”. The Stanford Encyclopedia of Philosophy (Spring 2019
Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/spr2019/entries/
mereology. Accessed August 31, 2020.
Walter, Sven, 2006. “Determinates, Determinables, and Causal Relevance”. The Canadian
Journal of Philosophy, 37:217–43.
Walter, Sven, 2010. “Taking Realization Seriously: No Cure for Epiphobia”. Philosophical
Studies, 151:207–26.
Wandell, Brian, 1993. “Color Appearance: The Effects of Illumination and Spatial Pattern”.
Proceedings of the National Academy of Sciences, 90:9778–84.
Wasserman, Ryan, 2010. “Material Constitution”. The Stanford Encyclopedia of Philosophy
(Fall 2018 Edition), Edward N. Zalta, editor. https://plato.stanford.edu/archives/fall2018/
entries/material-constitution/. Accessed August 31, 2020.
Watson, Gary, editor, 1982. Free Will. Oxford: Oxford University Press.
Wedgwood, Ralph, 2013. “A Priori Bootstrapping”. In Albert Casullo and Joshua Thurow,
editors, The A Priori In Philosophy, 226–46. New York: Oxford University Press.
Wiggins, David, 2001. Sameness and Substance Renewed. Cambridge: Cambridge University
Press.
Williamson, Timothy, 1994. Vagueness. London: Routledge.
Wilson, Jessica M., 1999. “How Superduper Does a Physicalist Supervenience need to Be?”
The Philosophical Quarterly, 49:33–52.
Wilson, Jessica M., 2002. Physicalism, Emergentism, and Fundamental Forces. Ph.D. thesis,
Cornell University, Ithaca, NY.
Wilson, Jessica M., 2002a. “Causal Powers, Forces, and Superdupervenience”. Grazer
Philosophische-Studien, 63:53–78.
Wilson, Jessica M., 2002b. “Review of John Perry, Knowledge, Possibility, and Consciousness”.
Philosophical Review, 111:598–601.
Wilson, Jessica M., 2005. “Supervenience-based Formulations of Physicalism”. Noûs,
39:426–59.
304 bibliography
Wilson, Jessica M., 2006. “On Characterizing the Physical”. Philosophical Studies, 131:
61–99.
Wilson, Jessica M., 2007. “Newtonian Forces”. British Journal for the Philosophy of Science,
58:173–205.
Wilson, Jessica M., 2009. “Determination, Realization, and Mental Causation”. Philosophical
Studies, 145:149–69.
Wilson, Jessica M., 2010a. “From Constitutional Necessities to Causal Necessities”. In
Helen Beebee and Nigel Sabbarton-Leary, editors, Classifying Nature: The Semantics and
Metaphysics of Natural Kinds, 192–211. New York: Routledge.
Wilson, Jessica M., 2010b. “Non-reductive Physicalism and Degrees of Freedom”. British
Journal for the Philosophy of Science, 61:279–311.
Wilson, Jessica M., 2010c. “What Is Hume’s Dictum, and Why Believe It?” Philosophy and
Wilson, Jessica M., 2011a. “Much Ado About ’Something’: Critical Notice of Chalmers,
Manley, Wasserman, Metametaphysics”. Analysis, 71:172–88.
Wilson, Jessica M., 2011b. “Non-reductive Realization and the Powers-based Subset Strat-
egy”. British Journal for the Philosophy of Science, 94:121–54.
Wilson, Jessica M., 2012. “Fundamental Determinables”. Philosophers’ Imprint, 12:1–17.
Wilson, Jessica M., 2013a. “A Determinable-based Account of Metaphysical Indetermi-
nacy”. Inquiry, 56:359–85.
Wilson, Jessica M., 2013b. “Nonlinearity and Metaphysical Emergence”. In Stephen Mum-
ford and Matthew Tugby, editors, Metaphysics and Science, 201–35. Oxford: Oxford
University Press.
Wilson, Jessica M., 2014a. “No Work for a Theory of Grounding”. Inquiry, 57:1–45.
Wilson, Jessica M., 2014b. “Hume’s Dictum and the Asymmetry of Counterfactual Depen-
dence”. In Alastair Wilson, editor, Chance and Temporal Asymmetry, 258–79. Oxford:
Wilson, Jessica M., 2015a. “Hume’s Dictum and Metaphysical Modality: Lewis’s Combina-
torialism”. In Barry M. Loewer and Jonathan Schaffer, editors, The Blackwell Companion
to David Lewis, 138–58. Chichester: Blackwell.
Wilson, Jessica M., 2015b. “Metaphysical Emergence: Weak and Strong”. In Tomasz Bigaj
and Christian Wüthrich, editors, Metaphysical Emergence in Contemporary Physics;
Poznan Studies in the Philosophy of the Sciences and the Humanities, 251–306. Brill:
Leiden.
Wilson, Jessica M., 2016a. “Are There Indeterminate States of Affairs? Yes”. In Elizabeth
Barnes, editor, Current Controversies in Metaphysics, 105–125. New York: Routledge.
Wilson, Jessica M., 2016b. “The Question of Metaphysics”. The Philosophers’ Magazine,
74:90–6.
Wilson, Jessica M., 2016c. “Three Barriers to Philosophical Progress”. In Damien Broderick
and Russell Blackford, editors, Philosophy’s Future: The Problem of Philosophical Progress,
91–104. Hoboken, NJ: Wiley-Blackwell.
Wilson, Jessica M., 2016d. “The Unity and Priority Arguments for Grounding”. In Ken
Aizawa and Carl Gillet, editors, Scientific Composition and Metaphysical Ground,
171–204. London: Palgrave-Macmillan.
Wilson, Jessica M., 2017. “Determinables and Determinates”. The Stanford
Encyclopedia of Philosophy (Spring 2017 Edition), Edward N. Zalta, editor.
https://plato.stanford.edu/archives/spr2017/entries/determinate-determinables/.
bibliography 305
Wilson, Jessica M., 2018. “Grounding-based Formulations of Physicalism”. Topoi, 37:

495–512. (First published online in 2016 at https://doi.org/10.1007/s11245-016-9435-7).
Wilson, Jessica M., 2019. “Between Scientism and Abstractionism in the Metaphysics of
Emergence”. In Sophie C. Gibb, Robin Findlay Hendry, and Tom Lancaster, editors, The
Routledge Handbook of Emergence, 157–76. Oxford: Routledge.
Wilson, Jessica M., 2020. “Essence and Dependence”. In Mircea Dumitru, editor,
Metaphysics, Meaning and Modality: Themes from Kit Fine, 283–300. Oxford: Oxford
University Press.
Wilson, Jessica M., in progressa. “On the Notion of Diachronic Emergence”.
Wilson, Jessica M., in progressb. “Primitivist Fundamentality”.
Wilson, Mark, 1982. “Predicate Meets Property”. The Philosophical Review, 91:
549–89.
Wilson, Mark, 2006. Wandering Significance: An Essay on Conceptual Behaviour. Oxford:
Clarendon Press.
Wimsatt, William C., 1994. “The Ontology of Complex Systems: Levels of Organization,
Perspectives, and Causal Thickets”. Canadian Journal of Philosophy, 24:207–74.
Wimsatt, William C., 1996. “Aggregativity: Reductive Heuristics for Finding Emergence”.
Wimsatt, William C., 2007. “On Building Reliable Pictures with Unreliable Data: An Evo-
lutionary and Developmental Coda for the New Systems Biology”. In Fred C. Boogerd,
Frank J. Bruggeman, Jan-Hendrik S. Hofmeyr, and Hans V. Westerhoff, editors, Systems
Biology: Philosophical Foundations, 103–20. Amsterdam: Elsevier.
Witmer, D. Gene, 2003. “Functionalism and Causal Exclusion”. Pacific Philosophical Quar-
terly, 84:198–214.
Wolff, Johanna, 2015. “Spin as a Determinable”. Topoi, 34:379–86.
Worley, Sara, 1997. “Determination and Mental Causation”. Erkenntnis, 46:281–304.
Wyszecki, G. and W. S. Styles, 1982. Color Science: Concepts and Methods, Quantitative Data
and Formulae. New York: John Wiley and Sons, Inc., second edition.
Yablo, Stephen, 1992. “Mental Causation”. The Philosophical Review, 101:245–80.
Yates, David, 2012. “Functionalism and the Metaphysics of Causal Exclusion”. Philosophers’
Imprint, 12:1–25.
Yates, David, 2016. “Demystifying Emergence”. Ergo: An Open Access Journal of Philosophy,
3:809–41.
Zimmerman, Dean W., 1995. “Theories of Masses and Problems of Constitution”. Philo-
sophical Review, 104:53–110.
Index of Names
Aizawa, Kenneth 14 Camazine, Scott 170

Akiba, Ken 131, 202 Campbell, Keith 191
Alexander, Samuel 14, 46, 48, 158–9 Capes, Justin 265–70, 272
Alter, Torin 24, 221 Carere, Claudio 171
Antony, Louise M. 14, 55, 68, 87, 89 Caruso, Gregg 275, 278
Armstrong, David M. 28, 88 Carnap, Rudolf 235
Audi, Paul 31, 113, 115 Carruth, Alexander 33, 124
Auyang, Sunny 131 Chalmers, David J. 10, 14, 21, 43, 98–100, 140,
Aquinas, Thomas 273 142, 152, 214, 217, 226–31, 233, 236–41
Christensen, Jonas 97–8
Baker, Lynne Rudder 41, 197 Churchland, Patricia 43, 84, 243
Baltimore, Joseph A. 104–5 Churchland, Paul 43, 84, 243
Barnes, Elizabeth 14, 131, 149–51, 202 Clapp, Lenny 14, 18, 56–7, 62–4, 86, 89, 108
Batterman, Robert W. 14, 18–19, 156–7, 163, Clarke, Randolph 60, 265–70, 272–6
166, 172–7, 185–6 Coates, D. Justin 253
Baysan, Umut 85–6, 125, 129, 135–6 Cohen, Stewart 235
Bealer, George 72, 227 Cohen-Tannoudji, Claude 3, 192
Beebe, James R. 231 Conee, Earl 223
Bedau, Mark A. 9, 10, 14, 19–20, 137, 156–7, Conway, John 155, 161, 166–7, 187
161–2, 166–70, 176–7, 185, 187–8, 215 Couzin, I.D. 170
Bennett, Karen 11, 14, 27, 31, 82, 86, 141 Crane, Tim 14, 22–3
Bergandi, Donato 283 Craver, Carl F. 14, 56, 60–1, 110–11
Berkeley, George 77 Cucu, Alin C. 122
Berker, Selim 116 Cummins, Robert 56, 60
Bernstein, Sara 21, 252–3, 255–61, 263, 281 Cunningham, Bryon 14
Biggs, Stephen 100, 227–9, 231–8, 240
Bird, Alexander 95, 143 Dasgupta, Shamik 14, 31, 115
Blandin, Patrick 283 Davidson, Donald 108, 261
Bliss, Ricki 115 deRosset, Louis 284
Blitz, David 26 Descartes, René 43, 122, 240
Block, Ned 99, 239 Dosanjh, Ranpal 88–90
Boghossian, Paul 229 Douven, Igor 35
Bohm, David 254 Dowell, Janice 23
Bokulich, Alisa 131, 202, 204 Dretske, Fred 98
Bonjour, Lawrence 235 Dummett, Michael 201
Bouchard, Frederic 283
Boyd, Richard 14, 153 Earman, John 34
Broad, C. D. 11, 14, 19, 46, 48–9, 123, 127–9, Ehrenfest, Paul 195
151–3, 158–9, 163–4, 215, 217 Ehring, Douglas 14, 65, 72, 108–9, 244–6
Bryant, Amanda 117 Ekstrom, Laura W. 268
Byrne, Alex 13, 98, 100 Emmeche, Claus 13
Euler, Leonhard 122
Callender, Craig 163 Evans, Gareth 201
Calosi, Claudio 131, 202, 204 Ewing, Alfred C. 216
308 index of names
Fales, Evan 242 Jackson, Frank 21, 144, 217–21, 224, 229, 239
Farrell, B.A. 218 Jaworski, William 88
Feigl, Herbert 218 Jenkins, C. S. 31
Feynman, Richard 2, 193–4 Judd, D. 248
Field, Hartry 34–5
Fine, Kit 31, 115, 191 Kane, Gordon 2, 34, 121, 266–9
Flanagan, Owen 221 Kant, Immanuel 235
Fodor, Jerry 14, 55, 88, 99 Kauffman, Stuart 170
Forster, Malcolm 195 Kim, Jaegwon 9, 11–14, 16, 28, 34, 39–40, 42–4,
Francescotti, Robert 124 56, 59, 61, 69, 79, 81–2, 88, 91–2, 103, 118,
Franklin, Christopher Evan 268 124–6, 143, 150, 213, 240
Frankfurt, Harry G. 261–2 Kirchhoff, Michael 10
Frege, Gottlob 229 Kirk, Robert 100
Frigg, Roman 195 Klee, Robert 14–15, 91, 166
Funkhouser, Eric 203, 244–5, 248–9 Kleman, Maurice 184
Knox, Eleanor 283
Garcia, Robert K. 41 Kornblith, Hilary 88, 112
Gertler, Brie 229–30 Korman, Daniel Z. 191
Ghirardi, Giancarlo 195 Koslicki, Kathrin 27, 116, 191
Gibb, Sophie C. 15, 85, 102–4 Kovacs, David 143
Gibbs, Josiah Willard 175, 186 Kripke, Saul 100, 198, 228, 230, 235
Gibbard, Allan 284 Kryukov, Alexey 195
Gillett, Carl 11, 14–15, 61, 110–12
Ginet, Carl 126, 266, 272–3 Ladyman, James 13, 123, 156
Grandy, Richard E. 191, 197 Lamb, Maurice 14, 20, 106, 174, 182–4
Greiner, Walter 123, 132 Lancaster, Tom 282
Guay, Alexandre 8 Landau, Lev 184
Langsam, Harold 99
Haggard, Patrick 275 Lavrentovich, Oleg 184
Hall, Ned 125–6 Leibniz, G.W.F. 217, 254
Harman, Gilbert H. 35, 98, 221 Lepore, Ernest 10, 14, 57, 65, 67, 82,
Haug, Matthew C. 14, 56, 61–2 88, 118
Hawthorne, John 95–6, 235, 258–62 Levin, Janet 88, 283
Heil, John 25, 76–9, 88 Levine, Joseph M. 14, 55, 68, 143, 226
Heisenberg, Werner 195 Lewes, G.H. 46
Hellie, Benj 6, 100, 214, 218 Lewis, David 44, 97, 198, 201, 221, 253
Hellman, Geoffrey 24, 26–7 Lewtas, Pat 142, 144, 146
Hempel, Carl G. 14, 22–3, 105 Libet, Benjamin 275, 278–80
Hendry, Robin Findlay 50, 123–4, 139, 282 List, Christian 81–2
Hill, Christopher S. 98, 153 Locke, John 191
Hitchcock, Christopher 82 Loewer, Barry M. 10, 14, 22, 24, 57, 65, 67, 82,
Hodgeson, Shadworth 44 88, 118
Hooker, C.A. 174, 175 Lowe, E. J. 136, 197, 199–200
Horgan, Terence 57, 65–7, 82, 103, 118, 121,
143, 152–3, 253–4 MacBride, Fraser 144
Howell, Robert J. 124, 142, 144–8 Macdonald, Cynthia 14, 18, 57, 64–5, 69,
Huggett, Nick 283 108–9, 209
Hume, David 143–4, 254 Macdonald, Graham 14, 18, 57, 64–5, 69,
Humphreys, Paul 8, 14, 123, 282 108–9, 209
Huneman, Philippe 283 Maddy, Penelope 283
Huxley, Thomas 44 Malcolm, Norman 39
index of names 309
Martin, C. B. 129 Pearson, Olley 150–1

Mayoral, E. 175 Pereboom, Derk 14, 88, 112–14
McCann, Hugh 272–3 Perry, John 100
McDaniel, Kris 27 Pettit, Philip 66, 258–62
McGeever, J. 14, 49, 123–4, Pexton, Mark 282
163, 282 Pitts, J. Brian 122
McGivern, Patrick 14, 24–5, 56, 64 Poland, Jeffrey 14
McKenna, Michael 253 Polger, Thomas W. 55, 88, 96
McLaughlin, Brian 11, 19, 46–9, 102–3, 121–3, Princess Elizabeth of Bohemia 240
133–4, 141, 158, 164 Putnam, Hilary 14, 55, 59, 88
McLeish, Tom C. B. 282
Megill, Jason 14 Raffman, Diana 237–8
Mele, Alfred 280 Raven, Michael J. 31, 115
Mellor, D.H. 22, 23 Ravenscroft, Ian 24
Melnyk, Andrew 14, 23, 27, 55, 64, 94–6, 101, Rimini, Alberto 195
103–6, 119, 143, 153, 172 Robinson, William 44
Menon, Tarun 163 Robb, David 108
Menzies, Peter 81–2 Robledo, A. 175
Merricks, Trenton 43, 191, 199, 211, 269–70 Rosen, Gideon 115
Messiah, Albert 195 Rosen, J. 170
Mill, John S. 8, 9, 14, 46–8, 157–8, 160, 163, 165 Rosenberg, Alexander 198–200
Mitchell, Sandra D. 10, 14, 19, 20, 156–7, 166, Ross, Don 13, 123
170–2, 177, 185, 187–8, 196 Rueger, Alexander 10, 14, 24–5, 56, 64, 166
Monton, Bradley 195 Russell, Bertrand 4, 34–5
Montero, Barbara 24, 41 Russell, Bruce 232–3
Moore, G.E. 72
Mørch, Hedda 99 Salmon, Nathan 198
Morgan, C. Lloyd 8, 14, 158 Santos, Gil C. 8, 283
Morris, Kevin 10, 60, 70, 78, 80, 82–3, 92, 95–6, Sartenaer, Olivier 8
118, 140, 146–7, 153 Schaffer, Jonathan 14, 24, 31, 97, 115–16
Morrison, Margaret 14, 20, 176, 182–5, 282 Schroder, Jurgen 14
Mossio, Matteo 282 Seager, William 14, 43
Searle, John R. 14, 197
Nagel, Thomas 21, 91, 217–9, 221, 224 Shapiro, Lawrence A. 88
Nemirow, Lawrence 221 Shoemaker, Sydney 14, 18, 28, 56, 62–4, 76,
Newton, Isaac 121–2, 192–3, 195 80–2, 85–6, 91–2, 95, 97, 99, 102–3, 105, 108,
Newman, David 14, 162, 166–7 128, 130, 143, 209
Ney, Alyssa 23, 76, 78, 80–3 Sider, Theodore 31, 35, 103
Nida-Rümelin, Martine 220–1, 223 Silberstein, Michael 13–14, 49, 123–4, 163, 282
Noordhof, Paul 14, 144 Simons, Peter M. 27, 86, 191
Nozick, Robert 268 Smart, J. J. C. 44
Smith, John A. 283
O’Connor, Timothy 9, 13–15, 19, 50, 124, Sperry, Roger 14
126–9, 254, 261, 266, 271–3, 278–9 Stalnaker, Robert C. 144, 230, 239
Oppenheim, Paul 14 Stephan, Achim 8, 14–15
Owens, David 91 Steward, Helen 254
Stoljar, Daniel 23, 143, 144
Page, R.E. 171 Strawson, Peter F. 262
Paolini Paoletti, Michele 8, 9, 14, 49, 150 Stump, Eleanore 272, 274
Papineau, David 22–4, 55, 121 Sturgeon, Scott 41
Paul, L. A. 27, 31, 63, 191, 284 Styles, W. S. 248
Paull, C. P. 103 Swinburne, Richard 235
Peacocke, Christopher 229, 235 Swoyer, Chris 95, 143
310 index of names
Tahko, Tuomas E. 30, 31 Wiggins, David 197

Taylor, Elanor 124, 127–129, 134 Williams, J.R.G. 131, 202
Teller, Paul 163 Williamson, Timothy 206
Thomasson, Amie L. 20, 191–2, 199, 212–13, Wilson, Jessica M. 8, 14–15, 17–19, 21, 23,
283 31–2, 35, 37, 39–40, 43, 50, 57–8, 60, 63–4, 66,
Thompson, Evan 14, 27, 163, 170, 283 72, 81, 86–7, 92, 94, 96, 100, 104–5, 107, 111,
Thompson, Frank 24, 26 114, 116–17, 122–5, 129, 131, 134–5, 139,
Trogdon, Kelly 115 143–4, 147, 149, 157, 160, 164, 168, 175,
Tye, Michael 98, 143, 201 177–8, 186, 193–4, 201–4, 209, 224, 227–9,
231–8, 240–4, 246, 248–9, 252–3, 255–61,
Unger, Peter 206, 284 263, 281
Wilson, Mark 236–8
van Cleve, James 14, 124, 127–8, Wimsatt, William C. 14, 24–7, 170
142, 217 Witmer, D. Gene 55, 59
Van Gulick, Robert 14–15, 49 Wolff, Johanna 131, 204
van Inwagen, Peter 27, 255 Wong, Hong Yu 9, 14–15, 50
Varela, Francisco J. 14, 163, 170, 283 Worley, Sara 244
Varzi, Achille 27 Wüthrich, Christian 283
Wyszecki, G. 248
Walter, Sven 97, 244–5
Wandell, Brian 248 Yablo, Stephen 14, 18, 57, 64, 67, 81–2,
Warhol, Andy 210 209, 245
Wasserman, Ryan 136 Yates, David 41, 55, 59, 84, 137–40
Weber, Tullio 195
Wedgewood, Ralph 235 Zimmerman, Dean W. 24
Subject Index
a posteriori brain dynamics 283

abduction as 231–5 British Emergentism 8, 14, 17, 46–7, 50, 122,
characterization of compositionally 134, 140, 156–65, 189, 210, 217, 270
basic/physical goings-on 22–3 butterfly effect (see also complex systems,
identification of realizers of causal roles 152 sensitivity to initial conditions of) 153,
necessities 228–30, 237 160
a priori
access to meanings/intensions 226, 229–31 causal
access to modal truth 100, 228–31, 235, 237 account of properties 63–4, 95–6
abduction as 231–6 closure of the physical (see Physical Causal
characterization of compositionally Closure)
basic/physical goings-on 22 efficacy 14, 67, 80, 82, 170, 240
conceiving as 231, 233 interactions 4, 56, 61, 171–2
identification of causal roles 152 joints/levels of grain (see also laws,
synthetic 235 comparatively abstract systems of) 5,
abduction/inference to the best 68–9, 71, 80, 82–3, 90, 118, 200, 258,
explanation 35–6, 228, 231–6, 239–40, 264, 277, 281
241 mechanisms 56, 61, 110
Ability Hypothesis 221, 223 overdetermination 15–17, 36, 40–53, 55, 57,
abstractionism (see anti-realism) 62, 66, 69–70, 72, 79–82, 108–9, 112–13,
accommodation of the appearances (see also 117–18, 149, 212–13, 282–3
Criterion of Appropriate powers (see powers)
Accommodation; Criterion of proportionality (see also difference-making
Illuminating Accommodation) 78–9 considerations) 14, 67, 72, 81–2, 258
Acquaintance Hypothesis 221, 223 realization 43, 152
algorithmic incompressibility 19, 156, 166–9 relations 32, 34–5, 69, 80, 82, 98, 111, 126,
anti-realism 18, 25–6, 29, 37, 75–80, 84, 118, 158–9
194, 243, 257, 283 role 56, 59, 62, 96–7, 110, 152, 198–200, 255
artifacts (see also functional role; sortal Causal Antecedents Condition 259–61
feature) 1–3, 5–9, 13, 20, 24, 32, 40, 54, Causal Determinism (see also
74, 88, 96, 105, 191–2, 196–201, 204–6, determinism) 255–7, 259, 265, 267
209–13, 282 Causal Realization Principle 43
asymptote/asymptotic singularities 166, causation (see also causal relations; quausation)
172–7 agent 266–7, 270–2, 277
autonomy difference-making/counterfactual account
causal 1, 6, 7, 14, 36, 67–9, 73–4, 82–3, 86, of 68, 82–3, 126
90, 108–10, 112, 124, 168–9, 172, 176, downward 14, 43, 121, 124
180, 266, 283 event 266–70, 277
emergent 128, 159, 169 higher-level 21, 70, 82, 117, 256
ontological 1, 6, 7, 36, 73, 81, 86, 108, 110, problem of (see problem of
112, 124, 169, 172, 176, 180, 283 mental/higher-level causation)
personal and moral 281 indeterministic 268, 273
ceteris paribus conditions
biological systems 282 in abductive principles 231–3
Boolean combinations 27, 90 in causal relations 35
312 subject index
characteristic states 107, 177–81, 184, 188, 194, entailment, lack of 151–3
196 underdetermination 236–40
classical conceptualism (see antirealism)
limit 181–2, 194–5, 211 conceivability argument (see also epistemic
(‘Newtonian’) mechanics 20, 93, 173, 179, two-dimensionalism; zombies) 21,
191–4, 196, 211 100–1, 217, 226–7, 230, 236, 240–1, 251
collapse objection 33, 124–9, 132, 134–7, conditionals
149–50, 153–4, 164 counterfactual 14, 33, 57, 66
collective individuals 283 modal 235
color 64, 203–4, 206, 214, 216, 219–23, 237, subjunctive 65–6, 118
241–2, 245, 247–9 configuration 1, 4, 7, 12, 17, 25–6, 29, 48–9, 55,
compatibilism 21, 252–4, 256–68, 272, 274–7, 99, 108, 111, 117, 121, 124, 128, 134, 136,
281–2 141, 148, 158, 162, 164–5, 170–2, 197–9,
Complete Physical Knowledge without 201, 205–6, 222, 242, 282
Knowledge of all the Physical Facts configuration state (see also characteristic state;
strategy 221, 223 degrees of freedom) 5, 177–8
complex configuration (see configuration) conjunct/conjunction relation 84–7, 102, 118,
complex systems (see also 244, 246–7
nonlinear/nonlinearity) 91, 129, conjunctive feature 85–6, 102, 104, 118
155–90, 190, 192–3, 214–5 277, 281–2 consciousness/conscious experience (see also
algorithmic incompressibility of 19, 156, mentality) 123, 209, 212, 214–51, 269,
166–9 271, 275
chaotic 127, 156, 160–3, 166, 170, 173, 175, as epiphenomenal 97–8, 101, 140–1, 217,
182, 186–7 240–1
criticality/critical points/fixed points in 5, 10, as involving a novel fundamental
20, 173–6, 183, 186 interaction 189, 24–1, 277
elimination in degrees of freedom of 19–20, in free will 275–6, 277–80
185–90 in perception/perceptual 5–6, 77–8, 189, 191,
nonchaotic 166–7 210, 214, 217–18, 241–3, 245, 248, 281
nonlinearity in 155, 157–66 intention-involving 266, 175–6, 276,
phase transitions in 14, 155, 160, 163, 166, 278–9
172–7, 184–5, 190 qualitative/phenomenal aspects/features
self-organization in 10, 19, 156, 166, 170–2, of 20–1, 97–101, 140–1, 145, 165, 215–8
188, 190 222–3, 225, 241–51, 264, 273, 281–2
sensitivity to initial conditions of 153, 156, self- 72
160–2, 167, 186, 215 subjective aspects of 215–16, 219, 222, 225
stability under perturbation of 175–6, 185, consequence argument 254–5
190 conservation law 19, 122–3, 132, 156, 165, 190,
universality of 5, 14, 19, 156, 166, 173–6, 182, 277
185, 190 constitution
unpredictability of 157, 215, 225 condition 94–6
compositional material 136, 284
complexity 17, 50 of functional roles 212
flexibility/plasticity (see also complex of powers 108, 112–15, 268
systems, universality; multiple self 268
realizability) 6, 14, 197–8, 200, 211 constitutive mechanism (see realization,
laws/principles 47–8, 158, 160, 165 constitutive mechanism)
structures 170 constraints 105–7, 118–19, 178–83, 185, 188–9,
compositionally basic entities (see also physical 193
entities) 9, 12, 22–5, 30, 39, 224–5 Cooper pairs 184
computational properties 96 counterfactual considerations (see
conceptual difference-making considerations)
subject index 313
Criterion of Appropriate Accommodation 36, 78, determinable/determinate relation 14, 18, 57,
155, 282 64, 72, 82, 85–7, 109, 113, 116–17, 150,
Criterion of Illuminating Accommodation 37–8, 209, 241–2, 244–7, 251, 284
117, 155, 182 powers-based account of (see Powers-based
Determination)
debunking explanation 78–9 Determinable-based MI 203–6, 208
deducibility/explainability/predictability 14, 34, Determinable-based MI (ordinary object
84, 90–1, 94, 113, 144, 159, 165, 179, 195 boundaries) 205
as entailing ontological reducibility 94 determinism 238, 253–5, 265, 275
failure of (see also British Emergentism; hard 256–7, 267
entailment, lack of conceptual or soft (see compatibilism)
representational; explanatory Determinism (see also determinism) 254–6,
gap) 13–14, 49, 127–8, 145, 147, 151, 260, 265–6
153, 156–61, 163–7, 170–1, 189–90, 215, difference-making considerations 14, 46, 57,
217 66–9, 71–2, 80–3, 118, 181, 258, 264, 277
lack of as characteristic of complex direct vs. indirect powers 129–31
systems 156, 163, 165–6, 190 disjunction/disjunct relation 89, 246–7
only by simulation 14, 167, 168 disjunctive
deflationism (see antirealism) types/properties 88–9
degrees of freedom (see also Degrees of Freedom strategy 88–90, 169
(DOF)) dispositions/dispositional properties 127–30,
as entering into individuation of scientific 145, 147–9
entities, features, laws 30, 92–3 of fields (see fields, second-order dispositions
associated with order parameters 20, 157, of)
182–5 distinctness/distinctive efficacy (see autonomy,
-based account of realization (see realization, ontological; autonomy, causal)
degrees of freedom-based; DOF-based DOF-based Weak emergence 107, 179–81, 183,
Weak Emergence) 188–9
elimination of 5, 14, 18–19, 20, 67, 87, 93, double-rock-throw cases 41–2, 44, 55, 69,
174–7, 182–3, 185–90, 194–6, 211 70, 117
increase in 20, 183–4 downwards exclusion 81–2
quantum 20, 93 dualism
reduction/restriction of 174, 178, Cartesian/substance 3, 8, 12, 43–4, 136–7,
188–9 150, 256, 271
Degrees of Freedom (DOF) 177 interactionist 122, 140–1
dependence (see also minimal nomological naturalistic 140
supervenience; realization) 14
cotemporal material 1–18, 36–7, 39, 41–7, Ehrenfest’s equations 195
51, 53–5, 58–9, 69, 71–5, 86–7, 94, 102, electromagnetic/electroweak interaction 50,
105, 110, 120, 125–6, 130, 133–4, 136, 131–2, 160, 184, 192
143, 153, 155, 164, 167–8, 170, 176–7, eliminativism 43–4, 84, 99, 243, 256–7
180–5, 196, 208, 218, 253, 257, 263, 266, entailment, lack of conceptual or
280–4 representational 14, 151–3
primitivist 150 epiphenomenalism 10, 44, 50, 59, 97–8, 101,
determinable 140–1, 149, 217, 240–1, 256
-based account of metaphysical epistemic two-dimensionalism (E2D) (see also
indeterminacy 20, 131, 191, 202–8, 211 extensions; intensions) 100, 226–31,
(see also Determinable-based MI) 236–41
-based account of realization (see realization, access-based objections to 236–40
determinable-based) argument against materialism 227
-based boundary indeterminacy 205–9 epistemically inaccessible identity 166–7
determination dimensions 244–5, 249–51 exclusion problem (see problem of
perceptual/qualitative features 216, 241–51 mental/higher-level causation)
314 subject index
explanation (see also deducibility/explainabili- configurational 133, 165

ty/predictability) new/novel 19, 50, 123, 132, 134, 156, 158,
inference to the best (see abduction) 160, 163–5, 210, 241, 272
reductive vs. non-reductive 183 non physical 143, 210, 241
explanatory gap (see also conceivability powers relativized to and based in 131–5,
argument; knowledge argument) 20–1, 164–5
84, 215–16, 225–8 -relative account (see also Interaction-relative
explosion objection 136 Strong Emergence) 19, 121, 132, 134,
extensions 102, 203, 222, 229–31, 148–9, 153–4, 210
236–41 theory of 179, 182, 193
fundamentality 2, 30–2, 117, 139, 149, 151
fields 2, 27, 133–4, 179, 184, 224 complete minimal basis account of 31–2
second-order powers of 134 independence-based account of 30–2
first-person perspective 222 primitivist account of 30–2, 149, 151
flocking of birds 20, 166, 170–2, 177, 188–90
forces 14, 27, 46–7, 50, 121–3, 131–2, 157–61, Game of Life 20, 155, 161, 166–7, 187–8
163–4, 186, 192–4, 225 gliders in 20, 166–7, 169, 176–7, 187–8, 190
configurational 121–3, 134, 158, 165 genus/species relation 86, 244, 246
Newtonian 134, 160, 193–4 Grounding (see realization, Grounding-based)
fractional quantum Hall state 282 gunk/gunky world (see worlds, gunky/infinitely
free choice/choosing/will/agency 6, 19, 21, 123, decomposable)
155, 165, 216, 252–82
freedom haecceity 95
as nomologically transcendent 267, 271, Hamiltonian 50, 121–2, 174–5, 182
276–82 Hempel’s Dilemma 22–3, 105
as ownership 259, 261–2 heteropathic effect 9, 47–8, 157–8
as responsibility 259, 262–3 homopathic effect 47, 157
as underdetermination 259–62 Horgan’s Constraint 152
Full Inheritance principle 60, 118 Hume’s Dictum 143–4
function Humean 18, 33, 45–6, 144
additive 159 contingentist categoricalist 18, 33, 45–6
analytic 173
asymptotic (see asymptote) identity (type, token) 6, 26, 45, 55, 80, 88, 95,
composition 27 108–10, 112–14, 116, 137, 167, 197, 205,
control 280 216, 222
deterministic 195 incompatibilism (see libertarianism)
linear/nonlinear 158–60, 171 Incomplete Physical Knowledge strategy 220–3
wave 184, 195, 224 indeterminacy 23, 131
functional epistemic 201
dependence/independence 5, 177–9, 194, conceptual (see conceptual
196 underdetermination)
duplicate 99, 101, 226, 240 metaphysical/worldly (see metaphysical
realization (see realization, functional) indeterminacy)
reduction 59 of free choice 267, 269, 271
role/characterization 20, 59, 62, 85, 172, semantic 201–3
191–2, 196–200, 210–12, 219–21, 283 sentential/propositional 203, 206
types/features/properties 59–60, 68, 226 indeterminism 266, 270, 275
functionalism 56, 59–60, 221 (see also individuation conditions (see sortal feature)
realization, functional) intensions 229–30, 236, 239
second-order 146 abduction-based epistemology of 228,
fundamental force/interaction 19, 47, 50, 123, 231–40
131–5, 143, 148–9, 154, 156, 159–61, conceiving-based epistemology of 227–8,
164–5, 189, 210, 241, 283 230–1, 236, 238, 240–1
subject index 315
primary 229–30, 236, 238–40 trans-ordinal 49

secondary 229–30, 236 trans-physical 152
intentionality 271–3 Laplacian demon 161
intentions 199–200, 255–6, 261, 266, 272, levels
275–6, 280 ‘higher-’ not synonymous with
Interaction-relative Strong emergence 133–4, 164 ‘emergent’ 12, 39
inverted spectrum 100 individuation of 12–13, 22, 24–30, 48, 92,
iridescent feather case 203, 206–7 171
law-consequence approach to
knowledge individuating 29, 30
argument 21, 217–28, 241, 251 lightweight combination approach to
by acquaintance 218, 221–3 individuating 26–8
of abstraction/specificity 68, 249
laws 18, 33, 49, 252 of causal/grain (see causal joints/levels of
as individuating features 147 grain)
as individuating levels (see levels, of compositional
law-consequence approach to complexity/organization 17, 50, 91, 93
individuating) of metaphysical grain 249
as metaphysical 29 of natural reality 12–13, 15, 40, 68, 79,
as structural 224 160, 218
causal 10, 49 libertarianism 21, 256–7, 265–82
comparatively abstract systems of 5, 30, agent-causal 266–7, 270–2, 277
68–9, 71, 80, 82–3, 90, 118, 181, 258, event-causal 266–70, 277
264, 277, 281 noncausal 266, 272–4, 277
composition principles in 47–8, 157–8, 160 Libet cases 275, 278–80
configurational 49 linear coupling 183
conservation 19, 122–3, 132, 156, 165, 190, linearity/linear combination 27–8, 47, 157–60,
277 162, 165, 171, 183
deterministic 166, 195, 254–6, 267, 271
emergence/emergent 91, 142, 145 macro-patterns 169, 190
higher-level/macro-level system of (see also Malebranchean occasionalism case 72, 143,
laws, special-science) 168–9, 275 147, 150
indeterministic/quantum (see also laws, of meaning (see extensions; intensions)
physics/physical) 30, 254, 256–7, 271, mental
276 causation (see also problem of
intra-ordinal 49 mental/higher-level causation) 80,
lower-level/micro-level system of (see also 252–4, 256–7
laws, of physics/physical) 29, 90, 168–9 multiple realizability, objection from 88,
metaphysical consequences of 29, 90–1, 93 240–5, 247–5
moral 262–3 superdeterminates, objection from 244–6,
Newton’s/classical 121–2, 192–3, 195 250
novel 14, 17, 46, 48–50, 123, 152–3, 156–8, mentality
160, 163 as implicated in functional roles of
of nature 11–12, 32–3, 41, 45, 97, 141, 143, artifacts 198, 210–13
187, 193–4, 210, 227, 254–5, 263, 269 as fundamental 104, 145
of physics/physical 4–6, 12, 29–30, 34, 84, as not fundamental (see also No Fundamental
90–91, 93–4, 96, 106, 113, 121, 127, Mentality (NFM) constraint) 23–4,
142–3, 153, 181–2, 193, 195, 227, 240, 43–4, 99, 147–8, 190, 192, 198, 209–13,
249, 254, 256 222, 225–6, 241
requiring certain information 30, 93 conscious (see consciousness/conscious
special-science 4, 5, 12, 29–30, 34, 37, 48, experience)
68–9, 78, 84, 90–1, 93–4, 121, 168, mereology/mereological 285
181–2, 193, 195, 249 classical theory of 27
316 subject index
mereology/mereological (cont.) 63, 67–8, 78, 87–9, 93, 108, 114–15, 126,
combination 11, 12, 27 137, 156, 166, 173–6, 182, 190, 198, 206,
determination 14 208, 244–5, 249
part-whole relation 108, 116–17 reductive strategies for accommodating (see
parts 63 reduction/reductionist strategies)
realization (see realization, mereological) multiple relativized determination 204–6, 208
sums/wholes/fusions 27, 63, 117
weak supplementation 86 natural kind 88, 228
metamerism 247–48 expressions/predicates/terms 229–30, 236–9
metaphysical indeterminacy 131 necessary connections (see also modality;
determinable-based account of (see also supervenience; superdupervenience)
Determinable-based MI) 20, 131, 191, in supervenience 126, 143–4
202–9, 211, 242 ontologically explanatory 57, 65
gappy/glutty 202, 205–6 necessitation conditions 94–6
of boundaries 20, 191, 201–2, 204–5, 207–9, New Object Strategy 135–7
211, 242 New Power Condition 51–4, 58, 71–2, 133–4,
of properties 202 140, 281
quantum value 201, 267 No Fundamental Mentality (NFM)
supervaluationist/primitivist account of 131, constraint 23–4, 92, 99, 105, 147
202–4, 206–7 No Low-Level Mentality constraint 24
methodology nomic role (see also causal role) 95
metaphysical 22, 35–8, 82, 212–13, 283–4 nomological possibility/necessity/supervenience
naturalistic 123 (see modality, nomological;
scientific 212 supervenience with nomological but not
scientistic 212 metaphysical necessity)
micro-configuration (see configuration) nonadditivity (see nonlinear/nonlinearity)
micro-structural feature 28, 50, 56, 61–2, non-aggregative interaction 170–1
110–11 nonlinear/nonlinearity (see also complex
mill argument 217 systems) 14, 155
modal as characteristic of Weak emergence 166–77,
correlations (see supervenience) 185–90
intuitions 240 as indicative of Strong emergence 19–20,
truth/knowledge 100, 226, 228–31, 235–7 156, 157–61, 163–5, 183, 189, 215
modality 226, 230 automata 176
epistemology of 230, 236 chaotic systems 127
link with meaning 231, 236 dynamics 161–2
metaphysical (see also supervenience with logistic map 161, 175
metaphysical necessity) 12, 99–100, nonchaotic map 167
142–5, 147–8, 151, 226–7
nomological (see also supervenience with Only Structural-Dynamic Mentality
nomological but not metaphysical constraint 24
necessity) 11–12, 14, 17, 41–2, 52, 54, Ontological Dependence (OD) 150
62, 68, 73, 86, 94, 99, 102–4, 121, 124–6, order parameters 20, 157, 182–5
129–30, 134, 137, 141–3, 148, 150 ordinary (inanimate) objects 19–20, 133, 155,
molecular structure 123, 138–9, 282 191–214, 242, 277, 282–4
moral artifactual (see artifacts)
autonomy 281 classical 192, 194, 211, 281
notions/values/goods 4, 252, 266–7, 269 natural 200, 209
reasons 266, 275 overdetermination
responsibility 266–7, 269–70, 273, 276 causal (see causal overdetermination)
multiple realizability (see also complex systems, problem (see problem of mental/higher-level
universality; compositional causation)
flexibility/plasticity) 14, 19, 28, 59–60, part-whole 62, 70
subject index 317
panpsychism 3, 22, 43–4, 104–5, 145, distinctive/new/novel 19, 21, 46, 48–54,
148, 217 66–7, 69, 71, 73, 80, 82, 118, 120–1,
parsimony, ontological (see also methodology, 125–7, 129–31, 133–7, 139–40, 152, 156,
metaphysical) 15, 36–7, 78–9, 207, 158, 161, 163–5, 183, 196, 209–11, 253,
231–2, 239–40 258, 264–6, 268–9, 271, 274, 277, 281
perception/perceptual features (see distinctive profile of 17, 67–9, 72, 80, 82, 106,
consciousness/conscious experience, in 118–19, 258, 277
perception/perceptual) forward-facing 102, 104
persistence conditions (see also sortal inheritance of 14, 60, 138
feature) 197–8, 200, 211, 284 operative notion of 17, 32–3
phase transitions (see complex systems, phase second-order 133
transitions in) Powers-based Determination 249-51
phenomenal aspects/features (see consciousness, (first pass) 246
qualitative/phenomenal (second pass) 247
aspects/features) precisifications 203–4, 207
Phenomenal Incorporation thesis 99–101, 141 predictability (see deducibility/explainabili-
physical ty/predictability)
acceptability 54, 71, 94–7, 101, 103–5, problem of
119–20, 137, 140, 147–8, 153, 162, 185, free will (and determinism) 254–9
190, 217, 223, 227 mental/higher-level causation 12, 16, 34,
entities 1–4, 6–9, 17, 22–5, 27, 29–30, 51, 76, 39–45, 51, 53, 55, 58, 69, 73, 81–3,
90, 92–3, 99, 105, 121, 136, 147–8, 189, 117–18, 124, 138, 213, 266, 282–4
201, 225 mental quausation 253–6, 258, 281
features 11, 20, 41, 56–7, 84, 90–1, 99–100, the many 206, 284
108, 128–9, 142, 147–8, 215, 222, 225, proper parthood relation between properties
246–8, 250 and behaviors 14
monism 8, 10, 22, 136, 180 proper subset
Physical Causal Closure 34, 41–2, 44, 46, 49, relation between powers 14, 60, 63, 89, 109,
51–3, 55, 67, 71, 81–2, 92, 124, 130, 138, 182, 208, 246, 263
180, 253–6, 264–5 strategy 21, 253, 257, 261–3, 274 (see also
physicalism 15–16, 22–4, 34, 39, 46–7, 50, 82, Weak Emergence)
93–4, 96, 99–101, 105–6, 109, 113, Proper Subset of Causal Antecedents
116–17, 135, 138, 142, 183, 190, 218–19, Condition 259–60
220–3, 227–8, 264 Proper Subset of Powers Condition (see also Weak
eliminativist 43–4, 84, 99, 243, 256–7 Emergence, schema for) 58–9, 62–70,
Grounding-based (see also Physicalism 72–3, 75–6, 78, 80–90, 93, 95–6, 98, 100,
(Grounding)) 115–17 102–4, 106–7, 109–13, 115–16, 118, 146,
nonreductive (see also realization; weak 152, 180–2, 194, 208, 258, 260, 281
emergence; Weak Emergence) 17, 29, proportionality considerations (see
39–40, 44, 55, 58, 67, 69, 71–3, 75, differencemaking considerations)
83, 86–7, 109, 112–13, 115, 252,
257, 265 qualia/qualitative/qualitativity (see
reductive 12–3, 20, 28, 40, 43–5, 48, consciousness, qualitative/phenomenal
58, 78, 91, 100, 115, 124, 127, 222, 253, aspects/features)
256–7 quantum
schematically expressed 15 degrees of freedom (see degrees of freedom,
supervenience-based 145–6 quantum)
Physicalism (Grounding) 115 entanglement 103, 123–4, 282
Picture Theory 77, 79 features 93, 194
powers indeterminacy 267
backwards-facing 102, 104 mechanics 2–3, 122, 173, 193–5, 254
direct vs. indirect 129–31, 266 spin 5, 30, 93, 150, 179, 181–2, 194
318 subject index
quausation/quausal relevance 57, 65–6, 253–8, of determinables 116, 243

263, 281 of dispositions/powers 18, 33
quiddities 33, 94–8, 149 of entities 76, 172
of features 75, 77, 169
realism of interactions 132
about forces/interactions of metaphysical indeterminacy 207
134 ontological/metaphysical (see also reductive
about higher-level features 76 physicalism) 12–13, 18, 20, 24–9, 37, 40,
about theories 29 43–6 48, 75–8, 84–94, 109, 118, 168–72,
ontic structural 31 190, 196, 216, 219, 283
realization (see also Weak Emergence) 14, 39, strategies for 28, 33, 44, 84–94, 116, 156, 169,
55, 58, 66, 69, 75, 107, 169, 172, 190, 205, 172, 176, 190, 208
253 reductive
constitutive mechanism-based 14, 56, 60–2, explanation 183
110–12, 116–17, 169 physicalism (see physicalism, reductive; see
degrees-of-freedom (DOF)-based (see also also reduction,
DOF-based Weak Emergence) 19–20, 75, ontological/metaphysical)
87, 94, 107, 118, 157, 169, 172, 177–82, representationalism 98
183, 185, 189–91, 194, 196, 211 relational holism 163, 165
determinable-based 21, 57, 64–7, 72, 75, 82, Renormalization Group (RG) method 174, 176,
86–7, 118, 169, 192, 202, 209, 211, 241, 185–7, 190
243–50
dot-shape analogy-based 66 Schrödinger’s cat 194
functional 14, 20, 55, 59–60, 62–3, 68, 72, 85, Schrödinger’s equation 122
116–17, 137, 140, 146, 152, 169, 172, second-order desires 262
190–1, 196, 211, 227 singularities (see asymptotes/asymptotic
Grounding-based 14, 108, 113, 115–18, 139, singularities; complex systems, phase
149 transitions in)
mereological 56, 62–4, 169 soft determinism (see compatibilism)
multiple (see multiple realizability) Sorites series 237, 242
qualitative 137, 139–40 sortal feature 20, 191, 197–200, 211
-indifferent regularities 68, 82 spacetime/spacetime points 133, 283
small-‘g’ (pluralist) understanding of options spatiotemporally coincident objects
for 116–17, 139 136, 235
subset of powers-based or condition on (see special-science
also Proper Subset of Powers entities (objects, events, processes, things) 1,
Condition) 18, 45, 59, 65–7, 69, 78, 80, 3–6, 10, 20, 40, 54, 67, 92–3, 136, 172,
85–6, 91–2, 95, 97, 102, 104, 111–12, 176 176, 179, 192, 211, 213, 284
superdupervenience-based/ontologically features (behaviors, properties, types) 3–4,
explanatory 57 40–2, 45–6, 50, 52–3, 55, 58, 67–71, 81,
supervenience-based 103, 144–6, 153 84, 88, 92–3, 95, 106
token identity-based 108–10 laws (see laws, special science)
reduction predicates 77
as entailed by theoretical deducibility taxonomy 30, 37
94 spectral inversion 99–101
causal 168–72, 190 spherical conductor 179
explanatory 138 spontaneous collapse theory 195
functional 59 stability under perturbation (see complex
intertheoretic 176 systems, stability under perturbation)
of consciousness/phenomenal aspects 20, 98, strange attractors 162–3, 166
214, 216, 219, 222, 243 Strong emergence (see also Strong Emergence)
of degrees of freedom 107, 174, 178–80, as compatible with physicalism
185–6, 188–9, 194 137–40
subject index 319
as involving fundamentally novel with metaphysical necessity 108, 124, 143,

powers/properties/forces/laws 14, 151
46–53 with nomological but not metaphysical
as involving new objects 135 necessity 14, 41, 54, 73, 124, 134, 141–9,
as epiphenomenal 140–1 151, 153
epistemic criteria for 140, 151–3
fundamental interaction-relative account of token identity 108, 116
(see also Interaction-relative Strong in realization (see realization, token
emergence) 19, 134, 148, 153 identity-based)
new power approach to/condition in (see also of powers 80, 113–14 (see also Token Identity
New Power Condition, Strong of Powers Condition)
Emergence) 17, 19, 21, 46–58, 69, 71–2, of features 108–110
118, 121, 125, 133–4, 152, 163–4, 183, Token Identity of Powers Condition 58–9, 65,
258, 265, 281 112, 130–1, 181
primitivism about 140 Token Power Constitution 112–13
Strong Emergence 17, 53–4, 74, 120, statistical mechanics 173, 175, 186
140, 155 symmetry breaking 14, 124, 183–4
strong nuclear interaction 131–2 symmetry concern 82
subjective character (see thermodynamic
consciousness/conscious experience, limit 173
subjective aspects of) systems (see also complex systems) 3, 5, 93,
subset of antecedents (see also Subset of Causal 163, 166, 173, 186, 193, 224
Antecedents Condition) 260–1, 264 Thomist understanding of free will 272, 274
Subset of Causal Antecedents Condition 259, tropes 33, 108–9, 191
261–3 truthmaking condition 94–6
Subset View 110 two ways (New Knowledge; Old Fact)
substance 137, 270–1 strategy 221, 223
dualism 3, 12, 43, 136–7, 150, 271
immaterial 136
universality (see complex systems, universality;
independent existence of 137
multiple realizability; compositional
material 8
flexibility/plasticity)
monism, material or physical 9, 54, 73, 134,
unpredictability (see deducibility/explainabili-
136, 170
ty/predictability,
physical 2, 8–9, 73, 134, 136, 170
failure of)
superconductivity 185
superdupervenience 57, 151–2
supervaluationism 202, 204, 206–7 vague predicates 207, 237–8
metaphysical 131, 203, 206–7 vitalism 8, 12, 39
semantic 203, 207
Supervaluationist MI (ordinary object Weak emergence (see realization; Weak
boundaries) 204 Emergence)
supervenience Weak Emergence 17–18, 58, 64, 72–5, 87, 95–7,
as compatible with epiphenomenalism 140 101, 104–7, 141, 147, 155, 257
as insufficient for complete metaphysical weak nuclear interaction 123, 132, 165, 283
dependence 115–16, 143–50 Weak Ontological Emergence (DOF) 107
as ruling out Strong emergence (see also worlds (see also supervenience)
collapse objection) 124, 126, 127 centered 229
as unexplanatory 143 considered as actual vs.
minimal nomological 11–12, 17, 41, 54, counterfactual 229–30, 237
73–4, 102, 134, 141 gunky/infinitely decomposable 24, 99
on a spatiotemporally extended base 103, metaphysically possible 12, 143, 147, 150,
127 227
ontological/robustly explanatory (see physically and/or functionally identical 99,
superdupervenience) 151–2 101, 141–2, 145, 227, 240
320 subject index
worlds (see also supervenience) (cont.) zombie 226

precisificationally possible 203–4, 207 argument (see also conceivability
with different laws of nature 94, 97, 99, 227 argument) 217
with the same laws of nature 11–12, 68, 94, as metaphysically possible 228, 230, 236,
120, 141–3, 227 240–1
with the same physical laws 94, 142–3, 227 conceivability of 99–101, 217, 226–7, 230–1,
zombie 227, 230–1, 240–1 241

Jessica M Wilson - Metaphysical Emergence (2021, Oxford University Press)

Uploaded by

Copyright:

Available Formats

Jessica M Wilson - Metaphysical Emergence (2021, Oxford University Press)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jessica M Wilson - Metaphysical Emergence (2021, Oxford University Press)

Uploaded by

Copyright:

Available Formats

OUP CORRECTED PROOF – FINAL, 2/2/2021, SPi

For my mother and my husband

2.1. Case 1 of the problem of higher-level causation: S causes S* 42

Epistemology at Saint Andrews (as Visiting Fellow), the Department of Philosophy

Philosophy of Science; ‘Non-Reductive Realization and the Powers-based Subset

1.1 Metaphysical emergence: dependence with autonomy

1.1.1 Dependence with autonomy

dependence, distinctness, and distinctive efficacy seemingly at issue in all these

1.1.2 The prima facie motivations

Cotemporal material dependence

Everything is made of atoms. That is the key hypothesis.

key issues and questions 3

Ontological and causal autonomy

• Distinctive special-scientific taxonomy: Special-science entities are classified

key issues and questions 5

abstract from details relevant to characterizing physical (more generally: depen-

• Universal properties and behavior: Many special-science entities, including

• Perceptual unity: Though the macro-entities of our acquaintance are, sci-

particles or other physical fundamenta; rather, a tree perceptually appears as a

Summing up: many considerations, drawn from science, perception, language,

key issues and questions 7

ontologically and causally autonomous with respect to these underlying micro-

1.2 Two key questions

1.2.1 What, more specifically, is metaphysical emergence?

One source of unclarity as regards the question ‘What is metaphysical emergence?’

certain ways in which accounts of metaphysical emergence directed at the target

• Accounts of metaphysical emergence typically agree in taking the only sub-

key issues and questions 9

Moreover, on contemporary accounts of emergence, it is also typically sup-

• Accounts of metaphysical emergence typically agree that the emergence of

key issues and questions 11

What it is for a feature S to cotemporally materially depend on a feature P

key issues and questions 13

candidate physicalist reduction base for macro-entities such as tables, planets,

So there are important core points of agreement among accounts of meta-

to accounts of emergence intended to have such metaphysical consequences, there

key issues and questions 15

of options, it’s no surprise that many discussions of specifically metaphysical

1.2.2 Is there actually any metaphysical emergence?

1.2.3 My aim: to provide clear, compelling, systematic

1.3 Outline of the book

The plan for carrying out this project is as follows.

key issues and questions 17

generally, no controversial theses pertaining to the nature of powers, properties,

key issues and questions 19

In Chapter 4 (‘The viability of Strong emergence’), I consider and respond to

key issues and questions 21

most promising forms of explanatory gap argument: knowledge arguments of the

1.4 Operative notions

1.4.1 The physical

As above, debates over the status of some goings-on as metaphysically emergent

key issues and questions 23

1.4.2 The individuation of levels

It is natural to think of metaphysical emergence in the target cases, whether natural

Talk of levels or layers of reality is ubiquitous in science and in philosophy. It

Still, as Wimsatt (1994) notes, “the notion of a compositional level of organization

key issues and questions 25

⁴⁶ As Heil (2003a) puts it:

sense of Wimsatt’s reductionist-friendly claim that “Because any complex material

The ‘lightweight combination’ approach