Powers and Properties; Dispositions

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

DISPOSITIONS:
PROPERTIES OF POWERS

by

Jonathan Munich Weil

_____________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHILOSOPHY)

August 2011

Copyright 2011 Jonathan Munich Weil

ii
Epigraph
"I am proposing as a mark to distinguish real things that they are nothing
but power."
-Plato, Sophist (247e)

iii
Dedication
For those loved ones, family and friends, who are no longer available for
conversation. May we all meet again somewhere down the Road...

iv
Acknowledgments
First and foremost I would like to thank my parents Sally and David, who
have seemingly unending reserves of support, guidance, generosity, love and
encouragement. There is no doubt in my mind that without them I would not
have been able to accomplish this. Also I will always be grateful to Laura
Lorentz, whose company and care kept me going through even the worst
moments (along with being there to enjoy the best together).
The rest of my family was extremely understanding of my constant mood
swings throughout the years of this project's completion as well. In particular my
sister Gwen shared her own similar experiences in order to provide much-
needed motivation and much-desired psychological feedback, with the added
benefit of helping me to feel a bit more human each step of the way. Also my
grandmothers Hattie and Vivian offered just the right balance of sensitivity and
criticism while I was working on this project, which kept me from feeling
inhuman while also ensuring it didn't take any longer than it already did.
In addition, a group of amazing friends helped keep me sane while
regularly offering new food for thought (philosophical and otherwise). The entire
Gotcher family has been amazingly generous with all forms of assistance, and
Jim in particular has shown again and again that he is truly the best and most

v
loyal of friends—willing to do anything to help, anytime. Thank you for
everything along the way. Jeff Murad and David Ducommun never failed to be
ready with challenging ideas and a comfortable place to stay, Eddie Zagha
somehow beamed stimulation and stamina into my head from the East Coast,
while others like Chris Cary, Dave Panitz and Ben Richardson made sure I was
having fun along the way (as did Alex Vorobiov, who made sure I kept
surfing…). The same goes for all the unmentioned others too, with whom I have
come into contact who have influenced my ideas and work in countless ways.
Last but certainly not least, I would like to thank Jim Higginbotham and
Jim Van Cleve—two amazingly reliable, insightful and flexible advisors—as well
as Barry Schein, for helping so much at all stages of this project. David Manley,
Jeff King and Kadri Vivhelin also provided invaluable comments and
conversation about these matters, for which I am very appreciative.
Other than that, all that remains is to pay respect to the ocean, which has
the final word on everything.

vi
Table of Contents
Epigraph ii
Dedication iii
Acknowledgments iv
Abstract viii
Chapter 1: Introduction
1.1 The Intuitive Distinction and Some Paradigms 1
1.2 Methodology 2
1.3 Detailed Outline 4
Chapter 2: Difficulty Linking Dispositions and Conditionals
2.1 The Simple Conditional Analysis 12
2.3 The Reformed Conditional Analysis 19
2.3 Other Problems 23
2.4 Moving Forward with (PROP) 27
2.5 Some Perspectives on the Project of Conditional Analysis 33
2.6 Conclusion 37
Chapter 3: The Entailment Problem and Fundamental Properties
3.1 The Prospect of Ubiquitous Entailment 39
3.2 Resistance to Ubiquitous Entailment 41
3.3 The Significance of Ubiquitous Entailment 49
3.4 Assessing the Situation: Ontology in the Background 57
3.5 Conclusion 68
Chapter 4: Brute Powers and Scientific Inquiry
4.1 The Stakes 70
4.2 Framing the Issue: Ungrounded Dispositions 71
4.3 Dispositions and Scientific Autonomy 80
4.4 Conclusion—Empirical Investigation of Brute Powers 87

vii
Chapter 5: Scientific Modalism
5.1 The Scientific Project 89
5.2 The Fundamental Level 90
5.3 Scientific Modalism 95
5.4 Conclusion—Eleatic Ontology 99
Bibliography 102

viii
Abstract
There is an intuitive distinction between dispositional properties—which
involve powers to act in certain ways, and non-dispositional ones—which
comprise strictly categorical qualities. However, this distinction is resistant to
analysis, particularly on the side of dispositions, in ways that to many suggest all
dispositions should be explained away in terms of categorical properties. I argue
that dispositions cannot be entirely accounted for in this manner. Instead,
dispositions are among the most fundamental properties of the world. As such
they are crucial components of any successful ontology.
In particular, I maintain that certain kinds of properties are exclusively
dispositional. To argue for this position I first detail some important conceptual
space in the philosophical literature on dispositions. Doing so involves precisely
elucidating several major points of contention. These include epistemological
and linguistic concerns regarding the necessary and sufficient conditions for
ascriptions of dispositions. In addition certain metaphysical questions about
dispositions are at issue—most important whether fundamental properties can
be purely dispositional. Ultimately I appeal to a characterization of the scientific
project to support the claim that brute powers do exist, and argue that this in
turn counts in favor of a related account of dispositions. Along the way I detail

ix
more general theories of dispositions reflected by various approaches to
answering these contentious questions.

1
Chapter (1): Introduction
(1.1) The Intuitive Distinction and Some Paradigms
It is generally accepted that ascriptions of dispositions are somehow
closely associated with conditionals. As a first pass by way of ostension, the
following concepts are paradigmatically dispositional: fragile, soluble, flexible, and
explosive. Such paradigmatic examples exhibit this intuitive conceptual
connection between dispositionality on one hand, and the entailment of
conditionals on the other: fragile objects are disposed to break when struck
properly in normal circumstances. More generally, for any object O, stimulus S
and manifestation M, O is disposed to M upon undergoing S iff O would M if
subjected to S.
Against this, categorical features of things are thought to be quite
different. According to the traditional understanding of non-dispositional
qualities, on their own they are not closely associated with any such conditionals
or entailment. Instead these qualitative properties
1

1
For now I follow Van Inwagen in identifying the property role "with the role 'thing that can be said of
something'" (2004, pg. 131) to avoid difficulties facing individual views of properties, and issues
surrounding the property/predicate distinction as it relates to dispositions. For more on these matters, see
Mellor (2000) and Shoemaker (1998).
strictly characterize to how
things are, as opposed to dispositions which determine how things may tend to
act.

2
Properties such as square, tall, breaking, dead, and spatially located are
paradigmatically categorical. Such qualities characterize how things are
independent of any behavior or tendencies for action. To be sure there might be
behavioral tendencies displayed by any and all things to which, say, dead is
accurately ascribed—for instance perhaps all dead things tend to become stiff
and tend not to conduct electricity. Regardless qualitative properties intuitively
are not defined by such tendencies, in the manner powers do somehow
intuitively seem to be.

(1.2) Methodology
Throughout what follows I ground my discussion of dispositions in the
examination of certain crucial difficulties involved with the attempt to
understand what powers are, and distinguish them explicitly and conclusively
from non-dispositional qualities.
2

2
Following Molnar (2003) and others I use the terms power and disposition interchangeably, as with
qualitative, categorical and non-dispositional.
Doing so helps bring into focus some
important problems plaguing the philosophical project of specifying necessary
and sufficient conditions for ascribing dispositions, by way of understanding
dispositions.

3
For instance there is trouble on the side of the dispositions: notorious
issues like finks, masks, antidotes, complex multi-track and gradable dispositions
make it extremely difficult to specify any general form of conditional that applies
to all and only those things having the disposition in question, which conditional
also applies in all and only those situations that count as appropriate stimulus
and manifestation conditions.
3
The ultimate goal of my project is to get to the bottom of what is involved
with intuitions about the distinction between powers and qualities. I attempt to
do so in a manner that enables systematic understanding of ascriptions of
dispositions, of what dispositions are, and how they are structured. My basic
approach is to address certain major philosophical obstacles, along the way
keeping track of what is at stake and how various views handle the important
challenges. Taking cues from science, I go on to claim that pure ungrounded
powers are fundamental features of the world, existing at the most basic levels of
And there is trouble on the side of the
qualitative—some argue that categorical and dispositional concepts alike a priori
entail conditionals. For example, because of their intrinsic features square objects
are such that their sides add up to four when counted properly (properly
meaning each side counted once).

3
More on such issues in chapters 2 and 3.

4
physical reality. My scientific realism thus works in favor of a modalist account
of dispositions, according to which certain properties are exclusively
dispositional. I proceed to outline and endorse this picture.
As will become clear along the way, some of the issues (e.g. the problem
of masks) have an epistemic/linguistic focus on the necessary and sufficient
conditions for ascriptions of dispositions. Other concerns target more substantive
metaphysical matters, such as the possibility of ungrounded powers. I argue that
both are key concerns for philosophical accounts of dispositions to address. This
helps narrow the field of viable accounts. Further, I argue that a modalist view of
dispositions, according to which some fundamental properties are pure
ungrounded dispositions, best satisfies all such prerequisites in a manner that
coheres with science. This sets it apart from other contenders.

(1.3) Detailed Outline
I start in chapter two with a review of the most prominent contemporary
philosophical attempts at understanding dispositions—the overall project of
conditional analysis. The origins of this project trace back to Rudolf Carnap's
work with reduction sentences as definitions, which took place prior to rigorous
theories of counterfactuals. Gilbert Ryle, W.V.O. Quine, and Nelson Goodman

5
each followed Carnap, with the goal of demystifying the occult nature of
dispositional predicates and their connection to merely possible occurrences.
Such attempts are striking in their linguistic (as opposed to metaphysical) focus,
and alike in their appeal to subjunctive conditionals.
In other words despite differences between their positions, all have a
common linguistic focus: accounting for correct ascriptions of dispositions, rather
than the metaphysical goal of specifying what the nature of dispositional
properties might be.
4
Notwithstanding such apparent contributions, examples involving finkish
dispositions (for which the stimulus conditions are the exact conditions that
cause acquisition or loss of the disposition itself
Additionally, according to these various versions of
conditional analysis,

accurate ascriptions of dispositions underwrite related
inferences about the behavior of those entities to which a disposition is
accurately ascribed. A sugar cube is soluble: if it were submerged in water it
would dissolve.
5

4
There are two arenas in which the categorical/dispositional distinction might be thought to apply: at the
level of properties, and at the level of concepts or predicates. This point is easily overlooked, leading to the
conflation of points about properties and predicates/their ascriptions. Cf. note 1. What I have to say is
meant to apply to primarily to properties, but incidentally applies to predicates too (as discussed in §3.4).
) reveal grave shortcomings in
the simple conditional analysis: often the relevant conditional is neither sufficient

5
Martin (1994), pp. 2-4. More on these examples in chapter 2.

6
nor necessary for ascription of the associated disposition. Consequently
philosophers such as David Lewis offer refined versions of conditional analysis,
which (for example) get around the problem of finks by way of appeal to
intrinsic causal bases of dispositions. Refined conditional analyses thus result
from the recognition that something more (metaphysically speaking) than a
typical manifestation in proper circumstances is needed.
However Lewis' analysis is susceptible to other related problems such as
masks, which themselves are not due to the kind of intrinsic change exploited by
C.B. Martin's type of case. Other versions of problems with themes similar to
those of finks and masks exist as well: reverse-masks or 'mimicking,' 'antidotes'
(last-minute masks) and the like seem to guarantee any conditional associated
with a given disposition can be falsified even when the disposition is present, or
made true when the disposition is absent. This spells big trouble for refined
versions of conditional analysis.
To avoid these and other difficulties, David Manley and Ryan Wasserman
arrive at a flexible approach. Assuming that this approach successfully handles
the myriad counterexamples designed to pose problems for the very project of
conditional analysis, the prospects for some such form of analysis remain strong.

7
More important, this barrage of attacks on specific forms of conditional
analysis does not defeat the intuitive conceptual connection between dispositions
and conditionals. Even if it is hard to say just how, the intuitive link remains
strong, as does the intuitive distinction between powers and qualities.

In chapter three I go on to examine another problem that arises in
philosophical debate about dispositions. This issue (perhaps misleadingly
dubbed the 'symmetry' problem by Troy Cross), arises from D.H. Mellor's
argument that ascriptions of paradigmatically categorical properties also (along
with ascriptions of dispositional paradigms) apriori entail subjunctive
conditionals. The argument is intended as a hurdle for attempts to specify
necessary and sufficient conditions for ascribing dispositions in terms of
conditional entailment.
Categoricists like Elizabeth Prior resist this claim, insisting that different
types of conditionals are entailed by ascriptions of dispositions than by
ascriptions of categorical qualities. Ultimately they appear unsuccessful.
Regardless, I argue that the entailment problem is important as a foil for
understanding what is at stake in the larger debate, since assessing the

8
entailment problem and its consequences reveals a second and related major
topic about dispositions.
In addition to (1) the necessary and sufficient conditions constraining
dispositional ascriptions, there are important questions about (2) the metaphysics
of dispositions—what they are; how they are structured. Philosophical accounts
of dispositions that address both sorts of concern can be grouped roughly into
two basic categories: modalism and categoricism.
6

Aided by the foil of entailment
concerns, in this chapter I begin to construct the machinery by which to carefully
formulate and assess these competing views of dispositions.
In the fourth chapter I take up the topic of brute powers, or
bare/ungrounded dispositions. The main task involved is simply framing this
topic so as to bring out the basic problem at issue. According to Simon Blackburn
scientific explanation trades exclusively on dispositions, which is problematic
because the notion of a purely dispositional world that results from this feature
of science is incoherent.
However, others argue that such a world is not incoherent, nor is it
impossible, implausible or even objectionable. It turns out that focusing on the

6
The issue of brute powers/bare dispositions, which is the subject of chapter 4, adjudicates between these
two options. See chapters 4 and 5.

9
precise connection between various analyses of the relation between dispositions
and their bases, and associated views on the possibility of bare dispositions,
sheds a great deal of light on this matter. In particular, though science does rely
on categorical grounds to account for most dispositions, it is not required nor is it
clearly even possible to do so with dispositions at the most basic levels.
Different treatments of what in general dispositions are directly result
from different views on the very possibility of ungrounded dispositions.
Therefore I take the issue of brute powers to be of fundamental importance to the
overarching debate surrounding dispositions—or at least to the metaphysical
strain contained therein. As a result I explicitly address several different
perspectives on the possibility of brute powers, in order to ground the
assessment according to science I endorse in the fifth chapter.
Ultimately, the most promising approach to deciding between these views
is to let the world determine theory, as mediated (to the extent possible) by
scientific investigation. If we derive an assessment of brute powers (and an
associated account of dispositions) from metaphysical biases regarding
ontological priority between types of properties, we encounter with concern and
resistance a situation such as the scientific world described by Blackburn. But if
we let what science reveals to exist guide our ontology, we can settle Blackburn's

10
world with equanimity. Note that there are two distinct issues in this vicinity:
whether any dispositions are ungrounded, versus whether all properties are
dispositional. In endorsing my scientific modalism I focus only on the former,
and maintain that, at the least, some fundamental properties are pure
ungrounded powers.

In the fifth and final chapter, I offer an appeal to the operation of science
and scientific inquiry. In the end, it is possible to employ a 'last-theory-standing'
approach to the project of understanding dispositions. Beginning with the
intuitive distinction between qualitative versus dispositional properties, we
move through various philosophical accounts of dispositions. Upon addressing
questions surrounding entailment and, subsequently, the possibility of bare
dispositions, we are left with the choice of two basic approaches.
On one hand are accounts that are at bottom modalist, positing primitive
modality; bare dispositions—at least some properties are ungrounded
dispositions. On the other are views that are essentially categoricist, holding that
no dispositions are ungrounded; causal bases are always distinct, qualitative,
non-dispositional. In the end this matter hinges upon whether the general form

11
and specific content of scientific findings determine ontology, or if instead we
allow our take on what there is to be dictated by metaphysical theory.
If what science reveals to exist shapes metaphysical theory, then we can
accept the existence of pure ungrounded powers at the bottom level. As
mentioned I call this 'scientific modalism' regarding dispositions. Science takes
dispositional tendencies to act and interact as fundamental, and (at least
potentially) to be ungrounded—so brute powers exist at the bottom level, even if
nowhere else. If instead metaphysical theory governs what science may reveal to
exist, then we can reject brute powers in favor of a categoricist account of
dispositions grounded in categorical qualities—in spite of what scientific inquiry
says about the fundamental level.
To conclude I argue that metaphysics at the least must closely heed
scientific findings, if not somehow being directly derived from scientific
investigation. Though this stance might itself be a bias in favor of science, such
preference for methodical empirical inquiry is advisable here given that we can
explore some of these questions about dispositions empirically. In so doing,
modalism emerges as the best remaining viable option. Therefore I close by
arguing for a moderate scientific modalism (SM) grounded in scientific realism.

12
Chapter (2): Difficulty Linking Dispositions and Conditionals
(2.1) The Simple Conditional Analysis
Most philosophers acknowledge some form of intimate connection
between ascriptions of dispositions, and conditionals describing what behavior
can be expected from bearers of the ascribed dispositions in specific
circumstances. Take for example the predicate fragile. When a glass vase is
described as fragile it is natural to assume that, if one were to drop the vase from
high enough up onto a hard surface, the vase would shatter.
7
The instinctive
frantic movement of anyone who drops something they think is fragile provides
evidence of how natural and fundamental this assumption is in our very
behavior.
8
Even with this common ground of sophisticated intuition, the project of
making precise what this link between dispositions and conditionals consists in
has proved to be difficult. Working prior to any well-developed theory of
counterfactuals, with little more at hand than the material conditional, Rudolf
Carnap's use of reduction sentences as definitions is a pioneering effort at linking

7
Or strike it with a hammer, slam it on concrete, etc. such that it breaks, fractures, etc. For now I take it as
uncontroversial that some dispositions like fragility are such that one and the same disposition may have a
range of stimulus conditions and characteristic manifestations. I focus on just one pair (dropping, breaking)
for simplicity. For more on so-called 'multi-track' dispositions, see for example Prior (1985), pp. 96-98. I
address this issue further in chapter 3 as well.

8
This is exactly why people do things like put stickers on packages clearly ascribing fragility to the
contents—we hope to regulate the behavior of anyone involved with the process of delivering them.

13
dispositions with conditionals—though his work leaves serious difficulties to be
addressed.
9
Carnap recognizes that statements ascribing dispositions (e.g. the glass is
fragile) entail some sort of sentence such as if the glass were dropped suitably it
would break. This sentence containing the subjunctive outstrips any associated
indicative conditional such as if the glass is dropped suitably then the glass breaks. To
account for this added component of the statements entailed by disposition
ascriptions with limited logical resources (i.e. no appeal to counterfactuals),
Carnap strengthens the initial conditional by quantifying over times. Thus we get
fortified conditionals such as whenever the glass is dropped suitably it breaks.

Such statements, however, have the undesirable consequence that
anything which is never struck counts as being fragile. Reading this statement
with the material conditional it is true of anything that either is not dropped, or
that breaks whenever it is dropped. This prompts Carnap to offer up a 'reduction
sentence' such as
(g)(t)[D(g,t) ⊃ (F(g) ≡ B(g,t))]
10

9
Carnap (1953), especially pp. 52-53. Mellor (1974), pp. 159-173 provides an informative discussion of
the origins of the conditional analysis.

10
Where "D(g,t)" means the glass is dropped at time t, "F(g)" means the glass is fragile, and "B(g, t) means
the glass breaks at time t. All this in turn gives us the following reduction sentence for g is fragile: If any

14
which entails nothing about the fragility of that which is never dropped, but
which has the result (seen by many as one to be avoided) that dispositions like
fragility are in principle unchangeable.
Imagine a fragile glass that undergoes a strengthening heat treatment,
rendering it such that it no longer breaks when dropped. On Carnap's picture, if
the glass is fragile then it breaks every time it is dropped, while in the current
case the glass will not break when it is dropped after the heat treatment.
Two options are available. Either (1) the glass was never fragile and so
does not break when it is dropped—even before the treatment. This is obviously a
problematic option. Or (2) perhaps the glass somehow stubbornly remains fragile
even after the treatment, and breaks when it is struck. But this can't be right,
either. Dispositions can come and go; their bearers may gain and lose them, and
Carnap's picture cannot account for this key feature without reinforcement.
11

glass g is dropped at any time t, then if g is fragile, g breaks at the time t, and if g is not fragile, then it does
not. This implies [F(g) ⊃ (D(g,t) ⊃ B(g,t)]: anything which breaks upon being dropped is fragile.

11
Carnap might instead have appealed to a different reduction sentence

(g)(t)[D(g,t) ⊃ (F(g,t) ≡ B(g,t))],

which implies

(g)(t)[F(g,t) ⊃ [D(g,t) ⊃ (B(g,t))].

Of course this implication is compatible with the mutability of dispositions, since it says that what is fragile
at a time breaks if struck at that time; there is no implication that if something is fragile at a time, it breaks

15
Ryle, Quine, Goodman and others thus endeavor to improve upon
Carnap's observations, arguing to various ends that ascriptions of dispositional
predicates are alike in being somehow intimately connected to subjunctive
conditionals.
12
It is important to note that despite differences between their
positions, all have a common linguistic focus upon accounting for correct
ascriptions of dispositions in terms of necessary and sufficient conditions, rather
than the metaphysical goal of specifying what the nature of dispositional
properties might be.
13
Crucial to these views is the recognition that dispositions present a
distinct problem in being somehow connected to merely possible occurrences.
Goodman in particular brings this out in his discussion of dispositions such as a
thing's fragility, flexibility, inflammability, or solubility. Though dispositions are
as important to philosophers as qualitative properties, our understanding of

at any other time when it is struck. However, it still cannot account for merely possible manifestations—for
instance fragile glasses that never are dropped nor break.

12
Consider Ryle's cashing out of the sense in which so-called 'dispositional statements' "resemble laws in
being partly 'variable' or 'open' ((1949), pg. 123), Quine's argument that "[t]he subjunctive conditional is
seen at its most respectable in disposition terms ((1960), pg. 222), and Goodman's attempts at grounding
the seemingly mysterious "threats and promises" he takes to be key features of dispositions ((1955) pg. 40).

13
Also important to keep in mind is that there are two arenas in which the categorical/dispositional
distinction might be thought to apply: at the level of properties, and at the level of concepts or predicates.
Cf. note 4. This point is easily overlooked, leading to the conflation of points about properties and
predicates/their ascriptions. For more on the lack of metaphysical focus, see §2.5 below.

16
them is deficient, because of their intuitive connection to possible manifestations
such as breaking when dropped.
With a 'manifest' predicate such as breaks, the predicate cannot apply
without something—a breaking—taking place. But there is a sense in which
nothing need occur in order for a glass to be fragile; for the dispositional
predicate fragile to be correctly attributed to an object.
14
In comparison with overt
behavior such as breaking, being fragile thus appears to be strange. This is why
Goodman couches the matter in terms of "whether we can explain disposition-
terms without any reference to occult powers."
15
Though substantial differences do exist between various versions of
conditional analysis, views falling under this general framework for
understanding ascriptions of dispositions share important common ground.

16

14
To be sure it is possible to resist this claim in a variety of ways, for instance by insisting that the
molecules of which a thing is composed must be bonding in a certain structure—perhaps one that is
multiply realizable by different substances and objects—in order for fragility to be correctly ascribed to that
thing (more on this in chapter 4). Mellor (1974) contains a detailed discussion of the sense in which
Carnap, Quine, Ryle and Goodman all require something to actually take place in order for a disposition to
be ascribed correctly, as part of his criticism of their pictures. (pp. 160-168) Still, I take it that the basic
point here is clear and uncontroversial: intuitively, dispositions involve some associated behavior (breaking
when dropped for being fragile) that need not in fact occur in order for the disposition to be attributed
correctly.
In
particular, the possession of a disposition like fragility is commonly associated
with a characteristic event such as breaking, the occurrence of which in

15
Goodman (1955), pg. 40.

16
See for instance Mellor's (1974) discussion for more on these differences.

17
paradigmatic circumstances (again for now, being suitably dropped) counts as a
manifestation of the disposition. These paradigmatic circumstances are the
stimulus conditions for manifestation of any given associated disposition. So
something qualifies as having the disposition to respond in characteristic fashion
to paradigm stimulus conditions just in case it would respond that way to those
conditions, were the conditions to occur.
In this manner the correct ascription of a disposition to an object licenses
certain inferences
17
(SCA) Something X is disposed to give response R to stimulus S iff, if S
were to occur, then X would give response R.
about how the object will act in various situations: a sugar
cube is soluble; if it were submerged in water it would dissolve. More explicitly,
according to a simple conditional analysis (SCA):
18
Once more, central to (SCA) is the idea that ascriptions of dispositions—unlike
attributions of categorical properties—in some sense extend beyond what

17
I borrow this language from Ryle (1949), pp. 119-122—leaving out the details of his picture, according
to which the inferences licensed by (e.g.) g is fragile are a subset of those licensed by some larger law such
as every (suitable) dropping of certain kinds of glass is followed by breakage. Similarly, for the time being
I do not delve further into Goodman's appeal to counterfactuals as demystifying dispositions inasmuch as
laws are thought to sustain counterfactuals, as well as subsequent difficulties with counterfactuals that
undermine the appeal of such a move. These interesting issues are to a certain extent relevant, but take us
too far afield for the purposes of the current project. Consequently I leave them aside for now.

18
As Manley and Wasserman make explicit ((2008), pg. 60) following Lewis (1997), in cases where a
predicate is in implicitly dispositional form (g is fragile), we first connect this predicate with an overtly
dispositional counterpart (g is fragile just in case g is disposed to break when dropped). Next comes the
crucial move of linking the explicitly dispositional locution with a subjunctive conditional (g is disposed to
break when dropped just in case g would break if it were dropped). In the interest of simplicity, I assume
that implicitly dispositional phrases can be translated into overt locution, and leave out the first step.

18
actually happens and are connected with what could happen; with possible
manifestations. For something can possess a disposition, but never undergo the
associated stimulus, and so never display the response characteristic of this
disposition. Surely in normal cases a delicate vase that is never dropped
nonetheless remains constantly fragile—it would break if it were dropped (from
sufficient height onto a hard surface…).
The same clearly cannot be said about the possession of a categorical
property; it is absurd to claim that a square object would retain its squareness
though it should happen never to manifest the requisite spatial configuration. It
is even conceivable that a disposition could be correctly attributed to an object or
objects, which disposition is never manifested by any of its bearers—though they
nonetheless possess the disposition.
19
Such considerations have led to what
Elizabeth Prior calls the "commonly accepted view"
20

that ascriptions of
dispositions bear special, distinctive relations to subjunctive conditional
statements.

19
This point underscores the shortcomings of Carnap's view, with its reliance upon the material
conditional: no manifestation of characteristic response is required in order for the associated disposition to
be present (assuming stimulus conditions do not obtain).

20
Prior (1982), pg. 93.

19
(2.2) The Reformed Conditional Analysis
Despite the intuitive appeal of this sort of position, however, it faces
severe difficulties as an account of the necessary and sufficient truth conditions
for ascriptions of dispositions. Most notorious are examples made famous by
C.B. Martin involving so-called 'finkish' dispositions. Finks are dispositions with
associated stimulus conditions that are the exact conditions for stimulating
acquisition or loss of the disposition itself.
21
In the first such case, involving Martin's 'electro-fink,' a wire that is dead
and thus not disposed to conduct electricity when touched is attached to a device
that monitors the wire, and makes it live whenever a conductor touches it. In this
instance the analysans is true while the analysandum is false. The wire is not
disposed to conduct electricity before the electro-fink intervenes, but it would
conduct electricity if touched.
Thus because of some external agent,
objects lacking/having the disposition to R in response to S would acquire/lose
this disposition if S were to occur, as the case may be.
In the second sort of case involving a 'reverse-cycle electro-fink,' the
opposite occurs. Here a wire that is live and as such disposed to conduct
electricity whenever it is touched is attached to a device that monitors the wire.

21
Martin (1994), pp. 2-4.

20
Just before the wire is touched, the electro-fink alters the wire's intrinsic structure
by removing any free electrons so that it does not conduct electricity. In this
instance the analysandum holds (the wire is disposed to conduct electricity when
touched), while the analysans does not (the wire would not conduct electricity if
touched by a conductor; the electrofink renders it dead upon being touched).
These examples present a great obstacle for the simple conditional analysis, since
they show that the conditional in question is neither sufficient nor necessary for
ascription of the associated disposition.
Attempts abound to reform (SCA)
22

22
To be sure, philosophers such as Sungho Choi (2006) argue that a revised form of (SCA) can do the
trick; that nothing as complex as (RCA) is needed. Though I do not explore such views in the current
project, I believe such positions are in fact versions of a reformed conditional analysis—as sophisticated by
any other name.
so as to protect it from such
counterexamples, by way of more precision in specifying the stimulus, response,
and underlying properties associated with manifestations of any given
disposition. An especially noteworthy effort is that undertaken by David Lewis
in his article "Finkish Dispositions." Lewis recognizes that Martin's reverse-cycle
electro-fink case, for example, exploits a change in the intrinsic causal basis of the
disposition in question. Such an alteration occurs in that the electro-fink removes
free electrons from a live wire whenever it is touched. This renders the wire dead
by removing the intrinsic grounds of its disposition to conduct electricity when

21
touched.
23
To this end, Lewis' refined conditional analysis (RCA) requires that the
intrinsic causal bases of dispositions remain fixed:
As a result Lewis' revision of (SCA) is formulated with an eye toward
surmounting the problems caused by altering the intrinsic causal basis of
dispositions, in the very manner exploited by Martin's examples.
24
(RCA) Something X is disposed to give response R to stimulus S iff it has some
intrinsic property B such that, if S were to occur and X were to retain B, then X's
possession of B and the occurrence of S would cause X to give response R.

Lewis' main contribution here is the thought that something more than the
characteristic response in stimulus conditions is required in order to correctly
attribute a disposition to a bearer. It is not enough that some fragile object
shatters when dropped; instead it must shatter when dropped because of some
internal structural feature of the object (along with being dropped).
In effect (RCA) pushes the general project of a conditional analysis of
dispositions toward a greater level of causal specificity. For its structure requires
that some intrinsic property qua causal basis remain constant throughout the
stimulus and response for a disposition to be correctly ascribed. This in turn

23
I address questions concerning the intrinsic causal bases of dispositions in subsequent chapters, leaving
the matter aside for the time being in order first to elucidate certain key aspects of the conditional analysis
program.

24
Lewis (1997), pg. 157. For present purposes I simplify Lewis' version somewhat by suppressing the
temporal modifier, and the clause regarding 'x-complete' causes. For more on such complications, see for
example Molnar (2003), pp. 129-130.

22
enables identification of the mechanism responsible for both actual and possible
occurrences.
To illustrate further, return to Martin's example involving the electro-fink.
The dead wire in this example has the intrinsic property of having all its electrons
bound, which grounds the fact that it does not conduct electricity when
touched—so long as the wire possesses this intrinsic property. However as the
case goes the wire would not keep this property when touched, because upon its
being touched the electro-fink frees up some of the electrons within the wire. It
follows that the wire would conduct electricity when touched, though it is not
disposed to conduct electricity when touched prior to being touched. We get the
correct assessment of antecedent and consequent.
Unlike (SCA), (RCA) handles this situation properly by rendering
irrelevant in principled fashion the fact that the wire would conduct electricity.
Even though the wire would conduct electricity were it touched, (RCA) delivers
the correct verdict: the wire is not disposed to conduct electricity when touched,
separate from the electro-fink's operation. In so doing Lewis' conditional analysis
avoids the problem of finks, as well as mutatis mutandis the problem of reverse
finks.

23

(2.3) Other Problems
Despite this persistent philosophical push toward greater precision within
the framework of a conditional analysis, other recalcitrant counterexamples
emerge. For instance (RCA) does not avoid the problem of masks, which as Mark
Johnston indicates
25
do not involve the kind of intrinsic change exploited by
cases of finks and reverse finks. Instead Johnston's masks are extrinsic to the
bearer of the disposition and stimulus conditions. Therefore it is built into these
examples that the disposition remains along with its intrinsic causal basis,
26
To make this point, Johnston draws our attention to a situation in which a
fragile glass cup is reinforced with internal packaging to protect it from severe
jolts during its transportation. The whole point of such packaging is to render the
glass safe from breaking, though the glass remains fragile throughout being
transported wherever it may be shipped. While the cup would not break if

despite the fact that it is not and would not be displayed in response to the
occurrence of stimulus conditions typically associated with its manifestation.

25
(1992) pg. 233.

26
Assuming for now, along with Lewis, that dispositions are intrinsic. For more on whether or not this
assumption is legitimate see McKitrick (2003a). Cf. note 23.

24
dropped, it nonetheless is disposed to break—something extrinsic to the glass
intervenes, causing the glass to endure being dropped without breaking.
Without the packaging to serve as an extrinsic mask of the glass' fragility, it
remains the case that the cup would break were it dropped.
When a disposition is masked along the lines detailed by Johnston it fails
to be manifested. This failure occurs even though the requisite stimulus
conditions obtain, and the disposition's intrinsic causal basis does not alter:
fragile things protected by packaging do not break when dropped, but they
remain fragile inasmuch their intrinsic features are the same throughout. It
follows that despite its increased specificity and subsequent ability to handle
finks and reverse finks successfully, (RCA) does not properly characterize the
necessary and sufficient conditions for ascribing dispositions such as fragility.
For according to Lewis' formulation a fragile glass will break when dropped if it
retains the intrinsic causal basis for the disposition and stimulus conditions
obtain. Yet in cases involving masks, fragile glasses do not display this
characteristic behavior in response to the stimulus, despite the fact that the
glasses retain the intrinsic causal basis of their fragility.

25
Variations on the theme of finks and masks exist as well. For example
there are cases involving reverse-masks or 'mimicking'
27
Or again, Alexander Bird's 'antidotes' resemble in structure, and as such
are in effect last-minute masks. Thus when a person ingests a poison that is
disposed to kill upon being swallowed, some time might remain prior to the
manifestation of this disposition during which an antidote may also be ingested.
The result of taking the antidote is that death is prevented. However, as is the
case with Johnston's scenarios in which a disposition is masked, the antidote
blocks the disposition from being manifested, rather than removing the disposition
wherein something X
would give response R to S, even though X does not possess the disposition to R
in S and would not acquire this disposition were S to occur. In this vein, Johnston
asks us to consider a sturdy chalice that is not fragile. However, a supernatural
being dislikes the vessel, and wills to make the cup break whenever it is
dropped. Indeed the chalice does and would shatter whenever dropped. But it is
not fragile, since fragility is intrinsic while the causal basis of the cup's breaking
is extrinsic. The being's actions provide us with an instance of a reverse mask; the
externally-induced shattering mimics that of something which is disposed to
break when dropped because of an intrinsic property of the glass.

27
In order to simplify the terrain as much as possible I assume cases involving the operation of so-called
deviant processes are also included in this category.

26
entirely (as with Martin's reverse finks).
28
Moreover, such scenarios need not involve anything as far-fetched as
supernatural beings who consistently act upon some deep resentment they
harbor for inanimate objects such as golden chalices. Michael Fara points out that
the presence of masks and mimics is as prevalent as it is ordinary and un-
mysterious. Tired people have their disposition to fall asleep masked by loud
noise in the neighborhood; car tires have their disposition to roll masked by
application of brakes. As he sees things, "the masking of dispositions is such a
humdrum occurrence that any adequate account of disposition ascriptions must
accommodate it."
In so doing antidotes guarantee the
conditional associated with a given disposition is false, even if the disposition is
present nonetheless.
29
In light of these observations it is striking that Lewis responds to such
examples simply by shrugging off the threat to (RCA) posed by masks (and thus
by implication mimics and antidotes), with a vague appeal to increased
specificity. And yet he does just this, attributing such matters to

28
Bird (1998), pg. 228.

29
Fara (2005), pg. 50. Or again: reverse-cycle electro-finks are also known as fuses.

27
the problem of specifying the stimulus and response correctly…We
might offhand define a poison as a substance that is disposed to
cause death if ingested. But that is rough: the specification both of
the response and of the stimulus stand in need of various
corrections.
30
Rather than amounting to a taciturn dismissal of masks, mimics and antidotes, I
follow Manley and Wasserman in taking Lewis' terse appeal to more precise
specifications of the stimulus and response associated with any given disposition
to be indicative of something important in the background. Namely, a
fundamental methodological conviction regarding disposition ascriptions: the
belief that, with an adequate degree of precision, it is possible to rule out the
kinds of problems for (RCA) under investigation in this chapter. If so, it remains
viable as an account of the conditions constraining ascriptions of dispositions—
and indeed there is an even more refined version of conditional analysis on offer.

(2.4) Moving Forward with (PROP)
Unfortunately Lewis did not get around to cashing out the suggestion in
his usual detailed manner prior to his death. As a result we must seek elsewhere
for this move toward more precise accounts of the stimulus and characteristic
response associated with any given disposition. This is the lead Manley and

30
Lewis (1997), pg. 153.

28
Wasserman follow, with their efforts at 'getting specific' within the general
framework of (RCA). To this end, they attempt to take Lewis' suggestion
seriously and apply it: "[g]iven the context-sensitivity of at least some
dispositional ascriptions, [Lewis'] 'getting specific' proposal will be that a given
utterance of 'N is fragile' will express a very precise dispositional property. So in
every context, 'N is fragile' will be equivalent to a highly specific conditional."
31
To generalize, the thought here is that ordinary attributions of
dispositions in fact ascribe hyper-specific dispositions. A precise specification of
the characteristic stimulus and response suffices to rule out conditions involving
finks, masks and the like for any given disposition. Ultimately, however, Manley
and Wasserman conclude that the tactic of getting specific does not succeed
alone. For even in conditions that count as paradigmatic for the manifestation of
a disposition, the disposition can be masked or mimicked. Such conditions
simply cannot be ruled out.

32

31
Manley & Wasserman (2008), pp. 65-66. The authors do acknowledge that which conditional gets
expressed varies with context, since dispositional ascriptions are taken to be context-sensitive. For more on
this point see especially pg. 65.

32
(2008), pg. 66.

29
To illustrate, though we might grant ordinary ascriptions of dispositions
to objects do in fact attribute some highly specific disposition,
33
Manley and
Wasserman point out that the question remains whether this more precise
disposition is associated with an individual stimulus condition, or a range of
such conditions. They go on to argue that the former option is subject to
examples involving so-called 'Achilles' heels,' for example a concrete block which
breaks only when it is dropped onto a weak spot it has, from a precise height at a
determinate angle. In ordinary contexts ascriptions of fragility to the block are
false; as a result of its Achilles' heel it just mimics fragility in a highly specific
situation. Therefore though standard cases of masking such as those raised by
Johnston can be ruled out under the getting-specific strategy, "the circumstance
under which an object with an Achilles' heel would break can be absolutely
paradigmatic for fragility (or the disposition to break if dropped)…For this
reason, Achilles' heels provide a recipe for generating counterexamples."
34
The other option—associating the more precise disposition with a range of
stimulus conditions—does not fare much better according to Manley and
Wasserman. They suggest that three different conditionals, all with different

33
See §2.5 below for more on this assumption.

34
(2008), pg. 68.

30
truth conditions, might be adopted for the task. The first, which requires only
that the object would break if dropped from at least one within a range of highly
specific circumstances, is far too weak:
(1) If O were dropped from some height over .5 meter, O would break.
Many objects will satisfy this conditional (such as the block with an Achilles
heel), though we should not want to deem them fragile in any ordinary sense as
a result.
The second is far too strong, since it requires that the object would break if
dropped for every circumstance within the range specified by the associated
conditional:
(2) Every height over .5 meter is such that, if O were dropped from it, O
would break.
Many fragile objects will violate this—such as a glass with a reverse Achilles'
heel that does not break if dropped at just such a height and angle, but breaks at
all others within the specified stimulus range.
35
Finally, the last conditional on offer requires that the object break when it
is dropped from some height or other within the specified interval:

35
As the authors point out ((2008), pp. 69-70), anyone who has been surprised to find that a glass does not
break when they should happen to drop it can attest to the plausibility of examples involving reverse
Achilles' heels.

31
(3) Some height over .5 meter is such that, if O were dropped from it, O
would break
This conditional is subject to what Manley and Wasserman dub 'the problem of
accidental closeness:' a counterexample in which, on a standard account of
counterfactuals, "we get the result that fragility is not intrinsic, and an object with
an Achilles heel at [e.g.] exactly two metres will count as fragile only when it
happens to be held at exactly two metres up."
36
Manley and Wasserman therefore conclude that certain counterexamples
involving finks defeat (SCA), while others involving masks cause trouble for
(RCA).

37

36
(2008), pp. 70-71. We are asked to consider an ordinary concrete block that is not ordinarily disposed to
break when dropped, but which is disposed to break when dropped from a height of 20 meters or more, and
which is actually on a windowsill 20 meters high. On a standard possible-worlds reading of counterfactuals,
the closest world in which the block drops from a height within the interval of paradigm cases (over .5
meters) is a world in which it would break (it drops from 20 meters). In this instance the third option
therefore wrongly delivers the verdict that the block is fragile. 'Capping' the interval with an upper limit
merely iterates the difficulty (70-71).
Attempts to reinforce the latter by more precisely specifying the
stimulus and response are overcome by examples exploiting Achilles' heels and
accidental closeness. Setting aside such cases, they go on to consider objections
seen as more fundamental in that they target the very form of conditional
analysis itself rather than specific versions of it. These objections are said to be
decisive.

37
I move quickly here through the details of Manley and Wasserman's careful arguments in order to focus
instead on their significance for the larger debate about linking dispositions with conditionals. For more see
especially (2008), pp. 68-82, as well as §2.5 below.

32
To start, there is the problem of comparative dispositional ascriptions—
namely, that extant conditional analyses ignore comparative locutions such as
glass A is more disposed to break than glass B. "But these are clearly problematic for
any analysis of a disposition in terms of a single conditional. A single conditional
can provide no scale along which objects can be compared, but such a scale is
required by a semantics of comparatives."
38
Next comes the problem of
specifying a mechanism for context dependence, which mechanism is required to
account for the fact that standards for dispositions such as fragility vary across
contexts. Last, there is the problem of absent stimulus conditions: some
dispositions need not be associated with any particular stimulus conditions at
all.
39
To surmount all these difficulties, Manley and Wasserman arrive at the
following flexible approach:

(PROP) X is disposed to give response R to stimulus S iff X would give
response R in some suitable proportion of S-cases
40

38
(2008), pg. 71.

39
For instance, one should certainly count as irascible if they become angry for no apparent reason
whatsoever—indeed, they are all the more irascible for doing so.

40
I translate the terms of (PROP) in order to adapt it to my earlier formulations of (SCA) and (RCA). For
Manley and Wasserman's own version, as well as the careful arguments for and treatment of it I skip for
now, see (2008) pp. 74-82.

33
where 'S-cases' cover every precise combination of values for properties such as
heights, densities and the like (relative to fragile for example). After running
(PROP) through the gauntlet of standard counterexamples and fundamental
problems for conditional analyses as well as some concerns unique to their new
analysis, they hold that (PROP) remains viable.
41

As a result they conclude with
cautious optimism that the outlook for (PROP)—and thus for underwriting the
intuitive link between dispositions and conditionals—remains hopeful.
(2.5) Some Perspective on the Project of Conditional Analysis
With (PROP) in place as the latest sophisticated version of conditional
analysis on offer, this is a good point at which to take a step back to see where
the path pioneered by Carnap has led. To start, on one hand advocates of
conditional analysis (regardless of particular formulation) defend their position
against counterexamples such as finks, masks and Achilles' heels by modifying
specific details of their preferred analysis. The hope for such maneuvers is that
what remains will be immune to further counterexample, thereby vindicating the

41
(2008), pp. 82-83. For example, the original fink case is overcome by (PROP) inasmuch as most S-cases
do not involve the presence of an electrofink: the wire fails to conduct electricity in a suitable proportion of
S-cases to be deemed dead, though it would conduct if touched in those S-cases (if there are any) involving
the presence of an electro-fink.

34
proposed conditions for ascribing dispositions. Against this, critics of conditional
analysis (as well as some proponents) construct further counterexamples
resistant to their opponents' proposals.
To the eyes of an observer unsympathetic to the cause of conditional
analysis, it starts to look as though the proponents' side of the debate includes
many who are readily willing to hack away at the conditional analysis they favor
when counterexamples demand. Such advocates of conditional analysis are
prepared if necessary to whittle down a seemingly viable proposal until it is so
precise as to be nearly unrecognizable as an analysis of ordinary disposition
ascriptions.
42
Additional counterexamples such as those discussed above tend to make
trouble for each new and improved conditional analysis. The better of these
counterexamples proceed not by concocting far-fetched scenarios begetting
murky intuitions, but rather by appealing to commonplace situations. This
improved strategy for resistance has far more bite to it than before, since the
appeal to paradigm cases enables more robust intuitions regarding the presence
or absence of any disposition under consideration.

42
Viz. Lewis' original formulation of (RCA), (1997) pg. 157.

35
Such desultory progress is neither uncommon to philosophical discourse,
nor on its own decisive for any one position on this particular matter. To their
credit Manley and Wasserman recognize this, and after diving into a careful
treatment of several counterexamples to proposed analyses and subsequent
revisions, attempt to move the debate into more substantive considerations. Thus
they address concerns that threaten to undermine the very form of conditional
analyses, and respond by proposing a version which they argue is immune to
these problems in addition to all the counterexamples they discuss.
Again, let us assume that they are right and (PROP) or something very
much like it does succeed in linking dispositions with conditionals as per
intuitions about the connection. Recall the ordinary-language origins of the
discussion in Carnap, Goodman and the like. The project with respect to
dispositions began as a primarily linguistic one, focused on conditions for
predication within commonplace ascriptions of dispositions in everyday
language.
Even if we grant that (SCA) is a reasonably approximate translation of
such commonplace attributions it is significantly more tendentious to assert this
of (RCA), and without question controversial to make the claim regarding

36
(PROP).
43
The accounts we are forced toward involve conditionals that are less
and less intuitively connected with our ordinary ascriptions of dispositions.
Instead of demystifying the occult, we seem to be adding complexity to it. More
important, something more than conditions constraining the ascription of
dispositions seems to be at stake in philosophical debate about dispositions.
44
Manley and Wasserman offer a state-of-the-art sophisticated analysis qua
one more instance of the same pattern: the general project of characterizing
necessary and sufficient conditions for dispositional ascriptions is not to be
abandoned, even in the face of particularly recalcitrant problems. Note that one
can endorse (PROP) and still ask what it is about dispositions that makes (PROP)

This is suggested by the fact that proponents do not abandon the project of
conditional analysis. The intuitive connection between dispositions and
conditionals which was a point of origin remains intact, in spite of any remaining
difficulties that arise for specific versions of conditional analysis.

43
Query whether non-philosophers have the intuition that X is disposed to break if dropped entails anything
about X's behavior in a suitable proportion of highly specific stimulus conditions—whether or not such
entailments in fact hold. Regardless, even if it is granted that most ordinary ascriptions of dispositions do
intuitively involve such entailments, it remains implausible to think that every ordinary attribution of a
disposition does. There are hints that Manley and Wasserman are sympathetic to this concern, for instance
in stating "[o]ne might be concerned that ordinary languages and ordinary contexts are not rich enough to
pick out" such precise dispositions or conditionals. ((2008), pg. 66) However such concerns and the appeal
to a theory of vagueness to get around them are somewhat orthogonal to my worry here: that a view which
is grounded in and driven by ordinary intuitions pushes us more and more toward a highly counterintuitive
translation of 'everyday' dispositional ascriptions.

44
This point receives substantial emphasis in contrast to the opposite point I make regarding the 'entailment
problem' in the next chapter. See especially §3.4.

37
apply. The Lewisian reply will be that this is the case in virtue of some intrinsic
categorical base. But it is also possible to claim that at bottom some dispositions
just are ungrounded modal properties—that being disposed to repel other
negatively-charged subatomic particles at bottom is all there is to being
negatively charged, for example. If so then at least some dispositions might not
be grounded in an underlying structure explicable in terms of categorical
properties.
Despite the failure of attempts leading up to (PROP) and its own
counterintuitive articulation, the underlying intuitive connection between
ascriptions of dispositions and conditional statements remains safely intact. This
is why the conditional analysis is not abandoned even in the face of
counterexamples. At the same time, despite this undeniable intuitive link
between dispositions and conditionals it remains to be shown that (and exactly
how) (PROP) applies to all ordinary dispositional predicates.

(2.6) Conclusion
Ultimately, I accept that something like (PROP) gives the necessary and
sufficient conditions for ascribing a disposition to an object. However, while

38
(PROP) remains neutral regarding whether a categorical ground is necessary for
every disposition, I take this matter of grounds to be of crucial importance to our
understanding of dispositions.
45

45
More on this point in chapters 4 and 5.
This suggests a separate strand beneath the
project of specifying the conditions constraining ascriptions—metaphysical
questions about what it is to actually be a disposition. In the next chapter I
discuss one route by which to shift focus to this other strand of the analysis of
dispositions.

39
Chapter (3): The Entailment Problem and Fundamental Properties
(3.1) The Prospect of Ubiquitous Entailment
Suppose it turns out that according to our best understanding of
dispositions, each ascription of a power entails some sort of subjunctive
conditional.
46
Mellor famously argues against views that take some entailment of
subjunctive conditionals to suffice for distinguishing between ascriptions of
dispositions and other sorts of properties.

Still, certain examples suggest that ascriptions of paradigmatically
categorical properties also appear to entail subjunctive conditionals. If in fact this
is so then it represents a major setback for attempts to parlay the intuitive
distinction between properties like fragility or solubility and properties like being
square or being spatially located—the basic distinction for which the notion of
'disposition' is invoked—into a conditional analysis of the necessary and
sufficient conditions on ascriptions of dispositions. For even if some such
entailment is necessary for ascriptions of dispositionality, it would not be
sufficient.
47

46
Note that this is importantly different from it turning out to be the case that each power is analyzable in
terms of some subjunctive conditional.
Mellor asks us to consider a triangle.

47
Again for present purposes I lump all such conditional accounts together. Which particular counterfactual
fact holds isn't as important as whether, in addition to being necessary to dispositions, some account is
sufficient for identifying dispositional properties (assuming some such counterfactual fact holds). For more
see chapter 2.

40
He maintains that "[t]o be triangular is at least to be a [planar] figure such that, if
its corners were (correctly) counted, the result would be three."
48
His point is that
paradigmatically categorical properties—such as the spatial configurations
comprising various shapes—share with dispositions the entailment of certain
subjunctive conditionals. Hence the 'Entailment Problem.'
49
The implications of this seemingly simple example are far-reaching. All
molecular structures are spatial (geometrical) configurations of inertial masses,
and having inertial mass entails subjunctive conditionals that specify behavior
(acceleration) relative to other inertial masses. Therefore it appears that any and
all qualitative properties—including quintessentially non-dispositional ones such
as shape, solidity, and the like—entail subjunctive conditionals. If conditional
analyses are not augmented somehow, they risk 'incorrectly' characterizing
properties that intuitively seem categorical as dispositions.

48
Mellor (1974), pg. 171. Mellor states in a footnote that the parenthetical 'correctly' refers to the method
rather than the result of counting, a point that Prior attempts to exploit in her reply (discussed in the next
section).

49
Namely, that ascriptions of dispositions and non-dispositional qualities alike entail subjunctive
conditionals. The standard name for this problem misleadingly invokes 'symmetry' (Cross (2005)), whereas
the real trouble is from ubiquitous entailment of conditionals by ascriptions of any property.

41
(3.2) Resistance to the Entailment Problem
Prior replies to this reasoning by arguing that being a triangle does not
entail the specific associated conditional Mellor adduces. If correct this means the
entailment first raised by Mellor's example is not actually a problem. She points
out that Mellor takes 'proper' counting to be a matter of the correct counting
process, rather than obtaining a specific result. She then goes on to maintain that
the exact manner of counting a triangle's corners cannot guarantee a certain
outcome. This is so according to Prior since the result of counting correctly
would not be three in a possible world with strange laws of nature that cause
triangles to become squares whenever the figures are counted. Essentially, she
insists there is a distinction to be made between the types of conditionals entailed
by dispositions versus those entailed by categorical qualities.
She offers two possibilities for drawing this distinction. (1) On one hand,
ascriptions of dispositions and categorical features alike entail subjunctive
conditionals. But the conditionals entailed by ascriptions of categorical properties
such as triangularity cannot be explicated without circularity (unlike those
entailed by dispositions). For instance 'X is triangular' entails 'If the corners of X
were (correctly) counted, the result would be three.' And she insists 'being
counted correctly' requires reference back to the results of counting—'if the

42
number is three (four, … , ) then the result of counting will be three (four, … , )'
and so on. Against this, the same is said not to be so for ascriptions of
dispositions such as fragility—'would break if (suitably) dropped' can be
explicated without the problematic circularity engendered by reference back to
the disposition (fragile) being defined.
50
Or (2) ascriptions of dispositions and categorical features alike entail
subjunctive conditionals, but for categorical properties such as triangular the
antecedent of the conditional refers back to the actual world. Anything X
triangular is such that if one were to count the corners of X and get the actually
correct answer the answer would be three.

51

50
Prior (1982), pp. 94-95. The latter part of this point is controversial, but for now in the interest of
discussion I ignore the controversy. Since Mellor's reply works we end up on safe ground regardless.
To summarize Prior's response:
either no relevant conditional is entailed at all, or the conditional holds but is of a
different type with categorical qualities than with dispositions. The second
option itself has two possibilities: (a) the type of conditional that does involves a
problematic circularity with categorical properties but not with dispositions. Or
(b) the type of conditional that does hold refers to the actual world with
categorical properties but not with dispositions. On any of these construals,
ubiquitous entailment is held to be at best misleading.

51
Prior (1982), pp. 95-96. Again, this formulation is designed to skirt the possibility of worlds with bizarre
laws of nature where the number of sides of an object changes whenever someone starts to count them.

43
Notice that short of denying some such conditional holds, Prior is not
denying an important common element of the entailments that result from
ascriptions of both dispositions and categorical features. Contrary to the stance
she favors, each type of property entails quite similar sorts of subjunctive
conditionals. In addition Mellor's definition of correct counting—placing corners
in one-one correspondence with the initial segment of the number line—
guarantees the result in this instance will be three without requiring circularity
or reference back to the actual world.
52
Contra Prior, the entailment holds.
53
As it turns out there is no real need to appeal to abstract things like
triangles or to debate about competing schemas for proper counting in the actual
world or any other. The entailment problem arises in much more familiar ways
since myriad categorical properties, when accurately ascribed to an object, entail
subjunctive conditionals as a matter of course. Take for example the property of
being serrated.

54

52
Mellor (1982), pg. 97.
When correctly ascribed to a steel knife, something along the
lines of 'if moved back and forth across suitably tender pieces of meat with
downward pressure, would slice meat' is entailed. Here a categorical property

53
To be sure Prior's definition of counting is clearly suspect, as shown by the ease as well as the tone with
which Mellor overcomes the objection.

54
This example is borrowed from conversation with Jim Higginbotham.

44
ascription entails a related subjunctive conditional, of the same basic type as
those entailed by ascriptions of dispositions.
In addition there are examples of entailments involving objects to which
precise sizes and shapes are ascribed, which ascriptions enable the apriori
deduction of counterfactuals about their fitting into certain regions of space. Or
again there are "counterfactuals that can be apriori known to hold of everything,
like 'If X were at least two feet away from Y, X would be more than one foot
away from Y.'"
55
Picking up on a suggestion from Stephen Mumford, Hawthorne and
Manley offer that the further distinction between what is entailed apriori and
what is constitutive of concept mastery might help. For even if categorical
ascriptions do apriori entail subjunctive conditionals, one could argue that
grasping such entailments is not constitutive of concept mastery for categorical
features of the world, while it is at the bottom of concept mastery for
dispositions.
Ascriptions of dispositions and categorical facts apriori entail
and are entailed by counterfactual facts, thus apriority alone cannot confer
adequate constraints on the associated ascriptions.

55
Hawthorne and Manley (2005), pg. 185.

45
Consider the earlier example. Appreciating the entailment from 'X is
fragile' to 'X would break if suitably dropped' is arguably indicative of concept
mastery for the dispositional predicate 'fragile' in a way that recognizing the
entailment from 'X is (suitably) dropped' to 'X would break if fragile' is not for
the categorical predicate '(suitably) dropped.' And it does seem possible that one
could effectively master the concept spherical without realizing that spherical
objects of a certain size fill spherical regions of properly corresponding sizes. But
it does not similarly seem that one could master the concept soluble without
making the proper inferences regarding behavior when placed in water.
This approach packs a great deal into the distinction between what does
and does not count as concept mastery for the concept of a given property.
According to the authors this is "as it should be."
56

56
Hawthorne and Manley (2005), pp. 185-186.
But such dependence upon a
prior grasp of what constitutes mastery for each type of concept in question
comes as cause for legitimate concern. For our starting point was confusion about
how precisely to draw out the intuitive distinction between categorical
properties and dispositions so as to properly classify all and only dispositions as
such.

46
Yet it is difficult to see how the proposed criteria for concept mastery with
dispositions like fragile, versus closely related categorical qualities like dropped
(suitably), could even be established without first having a firm distinction in
mind between the types of concepts (dispositional and categorical) involved—
not just an intuitive distinction. What else could account for the thought that
grasping apriori entailments indeed is required for concept mastery with
dispositions, yet is unnecessary for mastery of categorical qualities? As a result it
is unclear how appealing to concept mastery could help ground this distinction,
or properly constrain dispositional ascriptions.
Even if we suppose that proper criteria could be established without
circularity, appeal to concept mastery still runs into trouble rather than
conclusively sufficing for ascribing dispositions. There are many situations in
which we are familiar with an object's dispositions, but not its 'nature.' For
example consider a fragile china vase which a child knows will break if he isn't
careful with it. This child can be familiar with key dispositions of the vase, even
though he knows nothing about the role its internal makeup plays in its
disposition to break when dropped. If conceptual mastery for dispositions
requires making the proper inferences, then this child has mastery of the concept
fragile as applied to the vase. And yet, this mastery in a substantive sense falls

47
short of really comprehending fragility as manifested by the vase—something
crucial is missing.
57
In particular we have achieved concept mastery for dispositions without
getting at what it is to actually be a disposition, or at how dispositions are
structured, in any fundamental or metaphysical sense. Perhaps some believe that
all it takes to characterize dispositions is the understanding that with ascriptions
of dispositions but not of qualitative properties, making the proper inferences
apriori amounts to concept mastery. This might even turn out to be true for
certain dispositions, such as irascibility. Still, this approach is not in keeping with
the project for which it is adopted—for which a key point of departure was the
acknowledgement that something about dispositions underwrites their special
connection with subjunctive conditionals.

Consider the following statements from Quine: "[d]ispositions are, we see,
a better behaved lot than the general run of subjunctive conditionals; and the
reason is that they are conceived as built-in, enduring structural traits."
58

57
Thanks to Jim Higginbotham for suggesting something like this line of argument to me in conversation.
Quine's
thought is that something about the bearer of a given disposition is responsible
for the possession and manifestation of the disposition. The presence of this

58
Word and Object, pg. 223 (emphasis added). I address any concern that this picture of dispositions
conflicts with my subsequent endorsement of ungrounded, unstructured dispositions in the final chapter.

48
feature (the discovery of which happens often to be essential to scientific
progress) grounds ascriptions of the disposition, and underwrites the entailment
of any associated counterfactual facts. We have just seen that conceptual mastery
(as provisionally defined) does not require any knowledge of the presence of
these enduring structural traits. Nor does it require any understanding of the
connection between these enduring structural traits and the subjunctive
conditionals entailed by dispositional ascriptions.
Consequently even if conceptual mastery and apriority could be used to
constrain ascriptions of dispositions in terms of necessary and sufficient
conditions, doing so does not directly address the more fundamental question
about what dispositions in fact are.
59

If the entailment problem pushes us to
identify dispositions exclusively in terms of understanding the conditions on their
ascription, then it does not directly address the heart of the matter at hand.
Furthermore, appealing to conceptual mastery and apriority does not provide
much help on this front. The question about what dispositions really are remains
a perplexing issue, even if one is encouraged by recent attempts to conclusively
distinguish proper constraints on their ascription.

59
I elaborate on this line of thought in the next sections (esp. §3.3).

49
(3.3) The Significance of Ubiquitous Entailment
The prospect of ubiquitous entailment is therefore extremely revealing of
what is at stake in the philosophical investigation of dispositions. This is the case
much more so than with the basic project of linking dispositions with
conditionals.
60
Mellor's acceptance of the entailment problem as a problem and his very
formulation of the problem at first glance appear inextricably intertwined with
his basic picture of what it is to be a disposition. But there is more than one way
to read his view of dispositions as it relates to ubiquitous entailment.
Unfortunately the dialectical importance of the entailment
problem is often lost in the wake of debate about conditional analyses, at the
expense of more fundamental questions about what dispositions actually are.
61
At times it
appears that to be a dispositional property for him is just to be a property that
supports subjunctive conditionals.
62

60
Even Prior acknowledges that there is some intimate connection between dispositions and the entailment
of subjunctive conditionals (see chapter 2).
Again examples like the one involving
triangularity show that even paradigms of categoricity such as spatial
configuration entail subjunctive conditionals. From this he seems to conclude

61
In fact Mellor changes views between his (1982) and (2000).

62
Mellor (1982), pg. 96.

50
that there are no non-dispositional features of things—there are no strictly
categorical properties.
This reading is somewhat uncharitable however, as there are no strictly
dispositional properties according to Mellor either—the distinction does not
apply to properties. Instead it is a distinction between types of predicates.
63
Full endorsement of the position that all properties are dispositional,
however, would be an unfortunate move to make by way of response to
entailment concerns. For one thing the ease with which subjunctive conditionals
can be generated trivializes the ascription of dispositions on such a view. No
doubt examples such as those discussed above, and the fact that they are so
readily available in such wide variety, represent a great challenge for conditional
analyses of dispositional ascriptions. But this in turn again underscores that there
is another big question—what it is to be a disposition (rather than simply how to
ascribe one and what results from doing so). Surely it is not that easy to be a

Again at points he does indeed edge toward the stance that all properties are
dispositional. This conceptual movement is due to entailment concerns combined
with implementation of the austere view that to entail the requisite subjunctive
conditionals is to be a disposition.

63
Mellor (2000).

51
disposition. Even if intuitively categorical properties like squareness entail the
type of conditionals associated with dispositions, this does not overpower the
intuition that being square is not a disposition. This suggests that an account of
the conditions on dispositional ascriptions is only one important piece of the
puzzle.
It would be quite odd to take the lesson of ubiquitous entailment instead
to be that all properties are dispositions, given the very nature of the project from
which the problem of entailment stems. Again, most acknowledge the strong
intuition that properties such as fragile are importantly different from properties
like squareness. The modern disposition/categorical distinction has arisen as a
means by which to mark that difference by way of systematizing ascriptions of
dispositions. The debate continues about competing possibilities for exactly
which type of entailment/conditionals,
64
Taking this idea as a basis for the claim that all properties are dispositions
is therefore a drastic move. For to do so is to take a standard (conditional
entailment) called on to mark the intuitive distinction between dispositions and
but they all share the idea that we
should appeal to some such entailment in order to distinguish dispositions from
other properties.

64
Such as the familiar ones discussed in the previous chapter.

52
categorical properties so far that it dissolves the distinction—against
sophisticated intuition. This is an overzealous application that misses the original
purpose for appealing to the distinction in terms of entailing counterfactual facts,
as well as the deeper project regarding what dispositions actually are.
This fundamental concern is reflected by the earlier statement quoted
from Quine.
65
To ascribe a disposition is not just to make certain inferences, but also to
think there is explanation in terms of features of the disposition's bearer for why
the inferences go through. This idea is an important part of Quine's point
regarding the role played by 'enduring structural traits' in ascriptions of
Intuitively dispositions' nature outstrips their ascription
conditions—there is something else at work, in virtue of which the entailments
go through, whose identity is not exhausted by the constraints on dispositional
ascriptions. In particular, it is quite natural to think that applicable counterfactual
facts are grounded in some feature(s) of the bearer of the disposition with which they
are associated. This is the case regardless of what the exact entailments involve,
or just how the relevant counterfactual facts are related to their objects. Mellor
and Prior alike are able to acknowledge this without compromising or violating
their respective positions.

65
Ryle, Goodman, and others echo this general idea in similar statements of their own Ryle (1949);
Goodman (1954).

53
dispositions and associated entailment of counterfactual facts. These internal
features of bearers serve to ground the modal profiles associated with the
dispositions they bear.
Reading Mellor more charitably then, the entailment problem reveals two
central issues regarding dispositions, distinctions between which are orthogonal
to one another. One is the link between ascriptions of dispositions and the
entailment of associated conditionals; the necessary and sufficient conditions
constraining ascriptions of dispositions. The other is the more fundamental
matter of what actual dispositions are and how they are structured; the
metaphysics of dispositions.
Unfortunately, these two sorts of concern are often obscured in discussion
of the problem of entailment, by participants who conflate the distinct questions.
There is no doubt important common ground, namely the assumption that some
form of entailment is necessary for dispositionality (that all dispositions similarly
entail a certain type of conditional). Along with this common ground though
there are at least three different issues at play in the background of treatments of
the entailment problem. These separate matters usually are not explicitly teased
apart and stated as such, to the detriment of the debate and our understanding of
its significance for analyzing dispositions.

54
The issues are (I) What form the conditionals entailed by ascriptions of
dispositions take; what such conditionals look like;
66
To illustrate Mellor takes the main upshot of entailment concerns to be
that no appeal to entailment suffices for distinguishing dispositions from other
kinds of properties. But ultimately for him this does not amount to the
ontological claim that all properties are dispositions. Instead he concludes that
(II) Whether or not there
really is ubiquitous entailment by dispositions and categorical properties;
whether the form identified in addressing (I) suffices for tracking dispositional
properties; (III) assuming—in denial of the entailment problem—that some
entailment of counterfactual facts suffices for tracking dispositions, whether or
not we can account for why; what it is to be a disposition such that the
entailments sufficient for tracking dispositions do in fact suffice. Of these, (III) is
the most important to a substantive metaphysical understanding of dispositions
and how they are different from categorical properties, since only it targets the
fundamental nature of dispositionality in the sense evoked by Quine. The
matters involved therefore are importantly distinct from one another, at least
inasmuch as (III) addresses what dispositions are fundamentally, in a manner
that the others do not.

66
Strictly speaking this issue should not figure into responses to entailment concerns, but as discussed in
the previous section it does (for example in the appeal to apriority).

55
the distinction is not a metaphysical one applying to properties but rather an
epistemic/linguistic one that applies to predicates. This indicates his focus in
exploring ubiquitous entailment, and that his primary concern is not about what
it is to be a disposition in any metaphysical sense—which actually makes it easier
for him to acknowledge ubiquitous entailment.
Prior's resistance to the entailment problem as a problem on the other
hand belies an emphasis on a certain picture of what it is to be a disposition,
which in turn determines her denial of ubiquitous entailment. For her the
structure of dispositional properties is such that dispositions are necessarily
distinct from their grounds. These grounds, she maintains, are fundamentally
categorical.
67

67
I elaborate upon Prior's position below in the next section, as well as in chapter 4 on brute powers.
According to her categoricist picture, there is at bottom a clear
distinction between ascriptions of dispositions versus those of categorical
properties. A disposition is a higher-level property of possessing some lower-
level categorical property, the causal features of which ground the entailment of
counterfactual conditionals. Even though possessing the categorical property
grounds the entailment and disposition, it remains distinct from the higher-level
property of having some lower-level property qua causal grounds. An important
distinction between the two kinds of properties remains.

56
Thus Prior feels compelled to dig in and resist ubiquitous entailment, even
if it is with the two minor (and flawed) attempts at identifying asymmetries
discussed in the previous section. Since in her account categorical properties
ground all dispositions, there must be some difference between the two types of
features. All this is a sign that Prior is focused on why possession of dispositions
entails counterfactuals; on what dispositions are metaphysically speaking, as
opposed to Mellor's linguistic/epistemic focus.
This divergence of emphasis in addressing the issue of ubiquitous
entailment once more signals the presence of multiple dimensions of the
philosophical discussion of dispositions. Since those involved do not explicitly
acknowledge these various dimensions of the project, they often work at cross
purposes, with important metaphysical biases often affecting things from the
background. Whether directly or indirectly, this means that more hangs upon the
entailment problem than it might seem—much more than simply whether a
certain type of entailment holds for ascriptions of categorical qualities and
dispositions alike. Instead, along with the linguistic/epistemic project, addressing
entailment concerns leads to more fundamental metaphysical questions about
what dispositions actually are.

57
(3.4) Assessing the Situation: Ontology in the Background
Fortunately there is the potential to return a unified focus to philosophical
debate about dispositions. What is required is a firm grasp of the strong ties
between various analyses of dispositions on one hand, and associated views of
what types of properties exist at the fundamental level on the other.
68
To illustrate this point briefly, assume that only categorical properties
exist at the lowest level. This assumption is compatible with either a nominalist
or a realist picture of properties. Separate from any overarching view of
properties, at issue for present purposes is the somewhat more modest matter of
what kinds of fundamental properties exist. However if forced to commit, again
at a minimum I take it that something like Peter van Inwagen's theory of
properties as assertibles holds.

Furthermore it is significant that this approach does not depend upon any
particular theory of what properties are. Instead it homes in upon what types of
properties populate the bottom level—on whichever view of properties one
favors.
69

68
Assuming for now that there is such a thing as 'the fundamental level.' I drop this caveat in subsequent
discussion throughout this chapter, and define 'fundamental' to mean roughly 'partless.' More on this below.
Cf. note 87.
While it may do little else, this theory
accommodates my narrowing the focus to the fundamental level—as van

69
Cf. note 1.

58
Inwagen maintains though it "does have the virtue of truth…it is very nearly
vacuous."
70
That said there are four relevant possibilities for what kinds of properties
might be thought to exist at the fundamental level. First there are three monist
views:

Categoricism (C) fundamental properties are exclusively categorical;

Strong Dispositionalism (SD) fundamental properties are exclusively
dispositional;

Identity Dispositionalism (IDD) fundamental properties are powerful
particulars meaning they are qualities that are simultaneously
categorical and dispositional.

In addition there is a dualist view:
Moderate Dispositionalism (MD): entirely categorical properties exist
at the fundamental level, and so do entirely dispositional properties.

Note that only (C) forbids dispositional properties at the fundamental level; the
others permit dispositions at bottom.
71
Prior (1982) holds (C), since although she believes dispositional properties
exist at higher levels they all have categorical grounds, which hierarchy bottoms

70
(2004), pg. 131

71
This will become important in particular with respect to the distinction I draw later in this section
between categoricism and 'modalism' regarding dispositions—the latter of which includes (SD), (IDT) and
(MD). I ignore views according to which some properties at the fundamental level are exclusively one or
the other while others are mixtures of both.

59
out in a level of exclusively categorical properties. Shoemaker (1980) famously
argues that all properties are essentially causal, so all fundamental properties are
too (SD). Heil's (2004) so-called 'dual aspect' (IDD) has it that the
dispositional/categorical distinction represents different aspects of one and the
same property—in a metaphysical sense rather than at the level of predicates.
Ellis (2010) maintains that all causation (and therefore dispositions) require
categorical properties, but that both types of properties can be fundamental, pure
and ungrounded (MD).
It is striking that even if it were interpreted somehow as raising
metaphysical flags, the entailment problem would only be a true concern for the
categoricist (C). Again Prior (for instance) is committed to the view that
dispositions, and for that matter all other things, necessarily are grounded in
distinct categorical qualities of the world.
72

72
Lewis' (RCA) represents a similar conception—especially in that neither version allows ungrounded
powers anywhere in their ontology. In fact (RCA) serves equally as a characterization of Prior's
categoricism. Technically though, she and Lewis argue their case for different reasons: the metaphysics of
dispositions necessarly involving categorical grounds in Prior's case versus Humean supervenience (all
there is to the world is local matters of particular fact; everything supervenes on such categorical particulars
and so at the bottom level there is nothing but the local matters of qualitative fact) in Lewis.' More on this
distinction in chapter 5.
With this picture she has to resist any
dissolution of a clear distinction between dispositions and non-dispositional
qualitative features. Another option available to such an advocate of the
exclusive ontological fundamentality of categorical features is to deny the

60
existence of dispositions full stop. But that is clearly going too far by flying in the
face of intuition. Thus Prior's handling of ubiquitous entailment is firmly
grounded in this distinction between types of properties qua basis for her
categoricist picture of dispositions, and her view that only categorical properties
exist at bottom (C).
73
Against this, ubiquitous entailment does not similarly threaten the other
three views. If all properties are dispositional at the fundamental level (SD) and
everywhere else then it is no problem that every property ascription entails some
counterfactual conditional. For if there are only dispositional properties there is
no risk of incorrectly identifying a seemingly categorical property ascription as
dispositional. Nor is ubiquitous entailment a concern for (IDD), for similar
reasons—all categorical properties are also dispositional, so no categorical
property at any level can be incorrectly ascribed and identified as dispositional.
Even if entirely dispositional properties and entirely categorical properties both
exist at the fundamental level (MD), ubiquitous entailment is not a serious threat.
For even if every ascription of a categorical property involves the entailment of
some conditional, dispositions are by nature a different sort of property from
categorical qualities—though there may be nothing that grounds this difference.

73
This point does hinge upon the specific definition of 'fundamental' introduced earlier (cf. note 68). More
on this below in the remainder of the section.

61
This amounts to an asymmetry: entailments resulting from ascriptions of
dispositions describe and hold in virtue of something different for fundamental
dispositions than for categorical properties at the lowest level. Any apparent
ubiquitous entailment at the bottom level thus can be diffused by advocates of
(MD). All that is required is the intuitive appeal to causal behavior as central to
ascriptions of dispositional properties, versus the ascription of categorical
properties which do not intuitively involve causal activity in the same intuitive
sense.
74

More important, we can divide these views into two main types of
account regarding whether or not dispositions are always grounded in non-
dispositional qualities: Categoricism (C) and Modalism (M). Again according to
the categoricist, dispositions always have distinct, lower-level categorical
grounds. Borrowing from (RCA) to represent the general categoricist picture,
75

74
There are no doubt challenges to the notion that such a distinction can be made in terms of causation as
essential to the identity of different types of properties. For now I leave such matters aside.

any property D is a disposition iff for some φ and ψ, necessarily D is the property
of possessing some underlying intrinsic categorical property P that would cause

75
Cf. note 72.

62
one to φ in ψ.
76
Against this, according to a general modalist perspective, the fact
that property D is a disposition does not necessarily mean it has distinct or
underlying categorical grounds. Instead, there are dispositional properties D
such that, for some φ and ψ, necessarily D is the very property being such that one
would φ in ψ.
77
As discussed in the previous chapter with reference to Prior versus
Manley and Wasserman (respectively), both of these general sorts of theory offer
a version of why the relevant entailments go through, and both account for the
necessary connection that the others do not explain. For (C) the entailments
always hold because of the possession of some categorical property underlying
the disposition qua necessary causal base; for (M) they at least sometimes hold
because dispositions simply are modal properties of their bearers. With such
explanations on offer each of these positions involves handling dispositions in a
It is crucial to note that (SD), (IDD) and (MD) are all compatible
with modalism, since they all countenance the possibility of dispositions at the
fundamental level without distinct, underlying categorical grounds.

76
This picture of dispositions, which represents the default account for many philosophers, is often called
'functionalism' regarding dispositions. This is so because the multi-level operation of dispositions allows
for one and the same disposition to be realized by multiple underlying categorical properties. I use the
name 'categoricism' to avoid conflation with functionalism as it arises in the philosophy of mind. For more
on this distinction between the two types of functionalism, see Heil (1999) pp. 191-193.

77
Modal accounts differ from those deemed 'simplistic' above in that the modal accounts explain why that
which is entailed in fact is entailed, in terms of modal facts grounded in the nature of the bearers of
dispositions. As opposed to this the simplistic entailment picture merely has it that entailment of the
appropriate condtionals suffices for dispositionality, without weighing in on why these counterfactual facts
hold or thereby on what it is to be a disposition.

63
fashion that indeed addresses what dispositions actually are in a fundamental
way.
Despite these significant similiarities between the categoricist and
modalist positions, however, there is as indicated a decisive issue that helps
adjudicate between them. Each of them handles the matter of brute powers—
dispositions that are ungrounded in categorical properties or in other
dispositions—in crucially different ways. This difference in turn illuminates
some important aspects of the larger issues involved with the analysis of
dispositions.
Without question the majority of cases involve dispositions that are
somehow grounded in other categorical properties, for instance the fragility of a
vase grounded in its crystalline molecular structure. However, primitive powers
do seem to exist, or at least intuitively to be possible.
For example Manley has devised scenarios that closely parallel what
science seems to suggest about properties of micro-level particles such as
protons, electrons and quarks.
78

78
The following example borrows from one suggested in personal communication.
Imagine a toy world populated exclusively by
particles. Only some of the toy-world particles possess categorical property P,
and among the non-P ones some particles are such that they would repel a P-

64
particle if they were near it (we can say these particles possess property D) while
others are not. Beyond this counterfactual fact, there is no difference between
particles that possess D and those that do not. Intuitively, D is a disposition—
particles possessing D are disposed to repel P-particles in close proximity. And
the disposition has no underlying categorical causal basis, by stipulation.
The intuitive force of this example is closely connected with an associated
argument that is not restricted exclusively to toy worlds.
79
(P1) All categorical properties are structural
Keep in mind I take
fundamental particles to be particles with no proper parts, which enables an
argument that runs as follows:
(P2) x is a fundamental particle only if x has no parts
(C) Therefore, no dispositional property of a fundamental particle has a
categorical (non-dispositional) ground.

To elaborate, the first premise borrows from Armstrong, in that property P is
structural iff x's having P is analyzable in terms of the parts of x having certain
properties and standing in certain relations.
80

79
A similar argument was suggested to me by Jim Van Cleve, which I modified somewhat into its current
form.
The second premise comes directly
from the above definition of 'fundamental.' And the conclusion follows since if
any dispositional property of a fundamental particle did have a categorical

80
I state (P1) in these terms in order to allow for the possibility that some structural properties are neither
dispositional nor categorical—for example perhaps the properties of numbers.

65
ground, the ground would have to be structural—but given x's lack of parts it
cannot be.
This argument has the consequence that all properties of fundamental
particles are by definition dispositional (SD). For if all non-dispositional
properties are structural (including but not necessarily limited to categorical
properties), then because all fundamental particles are non-structural (since
again the particles have no proper parts), all of their properties are as well. So if
(P1) is true then all properties of fundamental particles are non-structural,
dispositional properties.
81
Indeed this outcome is stronger than what is required
for the claim that it is plausible that at least some dispositions are ungrounded.
However, it serves just as well to elucidate this line of reasoning that can be
associated with the claim that at least some powers are brute or ungrounded.
82
On the other hand categoricists simply cannot permit the possibility of
bare dispositions. Instead they must refuse to accept that any brute property
could count as a disposition even if its ascription entails counterfactual facts. This

81
For present purposes I take (P1) to be uncontroversial; my goal here is simply to motivate the claim that
some (if not all) dispositions are ungrounded at the fundamental level.

82
A possibility remains for categoricist resistance on this front: one could appeal to Lewisian intrinsic,
point-sized categorical properties as providing grounds for dispositions at the fundamental level that do not
violate the operative notion of fundamentality defined above. I find this picture of fundamental categorical
grounds for dispositions at the fundamental level to be at best extremely puzzling, and at worst incoherent.
As a result I postpone a more precise formulation and treatment of this possible stance for a subsequent
project.

66
is the case because for them, dispositions necessarily are higher-level properties
of possessing some lower-level intrinsic realizer property. Yet at the fundamental
level there are no lower levels for such realizers to populate. There are only
ungrounded properties—and thus, by definition for categorical realists such as
Prior, no grounds for dispositions that exist at the bottom level. They must deny
the possibility that dispositions be bare.
Modalistic views of dispositions, however, countenance the intuitive
possibility of brute powers at the fundamental level. Again, for modalists at least
at the bottom level dispositions just are identified with purely modal properties of
a particular sort; a certain sort of counterfactual fact. These counterfactuals do
not necessarily need to be reduced to anything else, at least not at the
fundamental level. Similarly, it can be a primitive explanatory fact that Manley's
D-particles are disposed to repel P-particles. There is no trouble for modalism per
se from primitive counterfactuals.
83
If the relevant modal properties instead are not treated as primitive
counterfactuals, then at any non-fundamental level the property identified with a
given disposition by modalists will be co-intensional with the one identified by

83
There are some other troubles with accepting explanatorily primitive counterfactual features. But these
troubles are not specific to the role played by such counterfactuals in modalist accounts of dispositionality.
As such I do not address the worries here.

67
categoricists. Namely, the disposition will be the non-primitive property (a
modal or causal property respectively) responsible for the relevant
counterfactual fact. In other words (C) and (M) can agree about all levels other
than the fundamental.
Indeed an acceptance of distinct co-intensional properties is required in
order to keep modalism from collapsing into a version of categoricism at levels
above the fundamental. But this is not such a stretch,
84
and more important
modalism is better equipped to handle the intuitions regarding ungrounded
powers that are pumped by toy cases such as Manley's. For categoricism
prohibits bare dispositions, while modalism makes the lesser sacrifice of positing
primitive modality to account for them.
85

This gives it the upper hand over
categoricist analyses of dispositions, especially given the intuitively compelling
possibility that bare powers can exist—categoricist strictures notwithstanding.
The take-home point here is that the issue of brute powers is the decisive
matter in debate about dispositions—particularly as it pertains to the
modalist/categoricist standoff. For the most important (read: metaphysical)

84
Consider for example the distinct but co-intensional properties being Barack Obama and being a member
of singleton-Barack Obama.

85
In the remaining chapters I support the claim that the modalist acceptance of primitive modality is indeed
a lesser sacrifice than the categoricist rejection of ungrounded dispositions.

68
questions about what dispositions are can all be consolidated into a focus on
issues surrounding properties at the fundamental level and bare dispositions.
This framing of the matter holds much more promise than grappling with the
problem of entailment discussed in previous sections.

(3.5) Conclusion
Throughout this chapter we have explored the consolidating impact of
entailment concerns with respect to various stances in philosophical debate
about dispositions. Making precise the problem of ubiquitous entailment also
brings out two different strands of philosophical discourse on dispositions. One
is epistemic/linguistic, the other is metaphysical. Both are key aspects of the
debate.
Moreover the second, metaphysical strand makes precise the strong sense
in which various pictures of dispositions are closely tied to corresponding
approaches to the ontology of fundamental properties. One's preferred ontology
of properties at the fundamental level plays a pivotal role in one's take on the
debate about dispositions, and vice versa. Unlike the linguistic strand, the
various relevant pictures of fundamental properties and associated perspectives

69
on how dispositions are structured are therefore brought out forcefully by
questions about bare dispositions.
Given that this is so, the issue of brute powers lies at the true heart of the
matter with respect to philosophical approaches to understanding dispositions.
In the next chapter I go on to explore precisely how this is the case, and just what
the exact implications of the more specific topic of brute powers are for the larger
ontological project of understanding dispositions.

70
Chapter (4): Brute Powers and Scientific Inquiry
(4.1) The Stakes
I have argued that two different major issues concerning dispositions
come into relief as a result of examining the prospect of ubiquitous entailment.
First there are epistemic and linguistic questions about what the necessary and
sufficient conditions might be for ascriptions of dispositions, which the project of
conditional analysis dating back to Carnap is designed to address. Second there
are more fundamental questions about what dispositions actually are, in terms of
how their bearers are structured. Having distinguished between these projects
explicitly and explored the former in detail it is time to take up the latter, more
metaphysically incisive matter in earnest.
To do so I focus on fundamental properties—in particular the possibility
of ungrounded powers and other primitive modal features at the fundamental
level.
8687

86
For present purposes and simplicity I lump bare dispositions, primitive counterfactuals and the like
together as closely related guises of 'primitive modality.'
This is a highly effective approach to questions about what dispositions
actually are. For different responses to the possibility of bare dispositions at the
bottom level directly reflect a preference between modalism versus categoricism
about dispositions. Of the various competing views on dispositions, only these

87
Again, if there is a fundamental level. Though I drop this extra phrase in subsequent discussion it remains
applicable throughout, unless otherwise indicated. I take it the answer is still a long way off. Cf. note 68.

71
two types of account address both sorts of major concern regarding dispositions.
Therefore the matter of brute powers is an absolutely fundamental aspect of the
larger project regarding dispositions.

(4.2) Framing the Issue: Ungrounded Dispositions
Despite the central part this issue of ungrounded powers can play in
reaching an overall philosophical understanding of dispositions, it has proved
difficult to frame the matter in precise terms. Simon Blackburn's succinct attempt
has become a sort of default starting point in the literature. But Blackburn
broaches the subject only to more or less bracket it as an oddity—he does not
weigh in with his own take in clear, conclusive terms.
88
In fact Blackburn does not himself use the topic as a method for
addressing what dispositions actually are, nor does he explicitly favor any
particular view of dispositions. In this spirit I remain agnostic until the final
chapter as to the modalism/categoricism dichotomy, and in what follows assume
only that some uncontroversial account of dispositions is operative. Though the
specifics of such a game-ending picture are not likely right around the corner, as
a vague placeholder it is uncontroversial enough to enable discussion.

88
It is possible to read him as taking a stand on the side of categoricism, or to be avowing some sort of
modalism, or to be simply puzzling at how science seems to proceed.

72
To set things up, Blackburn maintains that the entities investigated by the
natural sciences are identified in terms of dispositions: "science finds
dispositional properties, all the way down."
89
Furthermore according to him this
is how science operates, by definition. Thus "as the molecular theory [of gases]
gives us only things with dispositions, so any conceivable improvement in
science will give us only a better pattern of dispositions and powers. That's the
way physics works."
90
Even if we assume Blackburn is correct in his analysis of scientific
findings,

91

89
Blackburn (1990), pg. 255.
it remains to be said just what the problem is supposed to be with a
world of dispositions 'all the way down.' According to him the issue involves a
world in which everything gains its identity from how it impacts other things. In
such a world, there would be nothing but effects of effects of effects—if all there
is to things is their powers to impact other things, then there is nothing
qualitative for any impact to manifest itself in. This is viewed as a problematic
regress.

90
Ibid, pg. 256.

91
I say more about this below in the next sections of this chapter and in the final chapter.

73
In such a world there is nothing qualitative in which a causal power can
inhere in order to cause things; nor to be caused by things (for passive powers
92
The problem is underscored by a possible-worlds reading of
counterfactuals and a counterfactual analysis of dispositions. On such accounts,
if all truths about constituents of the actual world are dispositional then they are
made true by what is true of constituents of other possible worlds. It follows that
the actual world is exhaustively characterized by truths about neighboring
worlds. To make matters worse this situation obtains in all possible worlds, with
the result that there is no truth anywhere.
).
In other words there is nothing qualitative on either side of causation: nothing
qualitative for powers to exert their causal influence through; nothing qualitative
for powers to exert their causal influence on/to be influenced by powers. If the
qualities of objects are reduced to powers with no categorical features,
substantial reality fades away. According to this line of thought all that is left is
an incoherent reality—again, a problematic regress.
93

92
For now I do not make much of the distinction between active and passive causal powers, nor do I
address related issues like reciprocal disposition partners. What I have to say applies similarly to active as
well as to passive participants in reciprocal disposition pairs. For more on such matters see for example
Heil (2004).
Science thus appears to
systematically eliminate categoricity from nature, yet without categorical

93
Ibid, pg. 257. This is a compelling reductio if indeed it is sound.

74
qualities the world seems to lack qualitative qua tangible characteristics in some
ontologically significant sense.
Not all philosophers agree with this assessment that the notion of an
entirely relational world is incoherent. For example Richard Holton targets this
very assumption, incorporating the aforementioned relational model of
dispositionality. First he argues that the description of a world whose objects are
entirely characterized in terms of their relations to other objects amounts to a
description of a world entirely composed of pure powers. Then he offers what he
takes to be an example of just such a description.
Consider four points (A, B, C, D) about which the following is known:
(i) A is to the left of B and above C
(ii) B is to the right of A and above D
(iii) C is to the left of D and below A
(iv) D is to the right of C and below B

This schema yields the following visual model:

A ● ● B

C ● ● D

Crucially, however, Holton insists that this model is only a visual aid; that "there
is really nothing more to A, B, C, and D than that given by the descriptions. So do
not think that in describing them I have helped myself to the non-dispositional

75
notion of a point."
94
There are still things populating this world,
95
To be sure it is difficult to subtract all non-relational properties of the
points while retaining the points' relations. Separate from this mental challenge
many find relations without relata, or more precisely whose relata are exhausted
by their relational properties, highly dubious. In particular it can be argued that
it is unclear how purely relational worlds would differ from those consisting of
nothing but empty space—since the two sorts of world are qualitatively identical.
This challenge of addressing how to distinguish the two is taken to be a
necessary prerequisite for the success of Holton's example. As John Heil
maintains, "If we regard bodies as nothing more than relations or as nothing
more than powers to affect other bodies, it is not clear that we have left ourselves
with sufficient conceptual resources to make this distinction" between a purely
but they have
exclusively relational characteristics. Thus it is a coherent conception of a purely
relational world, and therefore on Holton's assumption of a relational model of
dispositionality a coherent conception of a purely dispositional world.

94
Holton (1999), pp. 9-10.

95
As opposed to the stronger thesis that there aren't even things in such a world to have relational
characteristics; there are only relational characteristics. In its strong form the thesis is both unnecessary to
motivate the concern and undesirable.

76
relational world and a non-relational world consisting entirely of empty space.
96
I take this to be primarily an epistemic line of thought since it involves an
argument rooted in our abilities to conceive and individuate extremely abstract
concepts, properties and situations. Consequently it runs afoul of the
metaphysical focus that was Blackburn's starting point—on what dispositions
actually are, and on whether the world could be like the exclusively dispositional
world depicted by natural science. But if science tells us the world works a
certain way, the fact that we have trouble conceiving or visualizing it to be such
is a decidedly weak objection. In fact, this happens all the time and yet we trust
that science gets at the actual nature of things. Indeed the entire history of science
is brimming with examples of experience diverging significantly from objective
reality—quantum mechanics is just the latest and most obscure example in a long
series.

The burden of explanation is said to be Holton's.
97
Even if it is difficult to visualize precisely, the world Holton characterizes
does not strike me prima facie as incoherent. Nor is it even implausible. Consider

96
Heil (2004), pg. 239.

97
Two rudimentary examples that immediately come to mind are the discovery that the solar system is
helio- rather than geo-centric, and the discovery that the earth is round rather than flat. Keep in mind that
those who favor a theory of powers will read Holton much more charitably than those who do not. I take up
a similar line of thought with respect to Manley's example below.

77
the earlier example from discussion of ubiquitous entailment, involving a toy-
world populated exclusively by particles. The main upshot of this scenario is that
an intuitively compelling case can be made for primitive modal features of
objects. If Holton is correct then entirely relational worlds are at least coherent.
With a relational account of dispositions it follows that entirely dispositional
worlds are as well. If so, then bare dispositions—along with counterfactuals and
other modal features—are too, which simultaneously goes a long way toward
making them seem much more plausible.
98

If indeed some fundamental modal feature could plausibly be thought to
exist despite being ungrounded in any categorical quality, this would count
against a denial of the possibility of a world of ungrounded modal features. The
question how a purely dispositional/relational property could affect anything
else arises either way. Even if there are categorical qualities in a world to provide
pure powers with something in which their effects can inhere, the main question

98
§3.4. Recall that we are asked to imagine a toy world populated exclusively by particles. Only some of
these particles possess categorical property P, and among the non-P ones some particles are such that they
would repel a P-particle if they were near it (we can say these particles possess property D). Beyond this
counterfactual fact, there is no difference between particles that possess d and those that do not. Intuitively,
D is a disposition—particles possessing D are disposed to repel P-particles in close proximity. And there
are no grounds for the disposition.

78
about how a power with no substantial qualities in which its causal force can
inhere could affect anything remains.
All that is needed in order to cause trouble on this front is a lack of
inherence in anything qualitative on one side of the operative force. Therefore
one might resist the idea that some dispositions, counterfactuals, relations, or
similar modal features can truly be ungrounded (entirely defined relationally or
modally). But, it is hard to see what would be worse about a whole world of
ungrounded modal features (of which a purely relational world is just one type).
The same conceptual difficulty remains for those who resist the idea—albeit with
a smaller magnitude than for an entire world comprised exclusively of
ungrounded properties.
Without a doubt there are those who will not accept the possibility of the
scenario. For instance any categoricist who is a proponent of Humean
supervenience—and so takes dispositional features to supervene on local
distributions of particular matters of fact—will not allow a difference in the
behavior of particles with no other difference between them. Appealing to the
fundamental levels of other possible worlds is not an option in this scenario
either, on pain of the now familiar sort of regress that bothers Blackburn.
And so such a counterfactual difference, not grounded in any other qualitative

79
feature qua causal basis, strikes categoricists as impossible. To a modalist who
favors a view that allows for pure powers, however, this response to the
possibility of ungrounded modality begs the question—just as much as accepting
the possibility of brute powers does according to categoricists.
99
Categoricists therefore object to the very notion of ungrounded
dispositions and take the idea of a purely dispositional world to be incoherent,
since they take all dispositions to have distinct categorical grounds necessarily.
Against this modalists in general accept the possibility of ungrounded modal
features, with different versions of the specifics ((SD), (IDD) & (MD)). This means
that the modalism/categoricism distinction cuts across the monist/dualist
dinstinction with respect to fundamental properties, since (for example) one can
be a dualistic modalist (MD).

Regardless of the reaction one has to the possibility of ungrounded
powers, the earlier example helps bring to light what really is at issue with
Blackburn's attempt at playing out the consequences of science. The problem

99
The same point can be made, mutatis mutandis, regarding other forms of categoricism—for example the
so-called 'functionalist' version of Prior, Pargetter and Jackson (1982). Proponents of such an account tend
to assume rather than argue that ungrounded powers are impossible; vice versa for their opponents. See for
example McKitrick (2003) pp. 358-362 and Heil (2004), pp. 232-234. As with the toy-world example, this
unfortunate methodological situation is part of the motivation for my appeal to science in subsequent
sections.

80
trades not so much on an entire world of pure powers, but on the existence of any
ungrounded modal feature like a brute power.
100

(4.3) Dispositions and Scientific Autonomy
Addressing the matter of bare dispositions at the fundamental level—
where ungrounded powers are pure powers because there is no underlying level
to provide grounds for the powers—thus forces one's hand regarding a general
take on the metaphysics of dispositions. Again, the entailment problem reveals
that the two most serious contenders for a viable account of dispositions are
those that are generally modalist, versus those that are generally categoricist.
These two sorts of view address both strands of major philosophical concern
about dispositions.
101
Despite an ambiguous framing of the matter, Blackburn sheds some light
(perhaps inadvertently) on the tensions involved in this dichotomy between
modalism and categoricism regarding dispositions. On one hand the hard
sciences rely almost exclusively on dispositions in order to explain natural
phenomena, which indicates the material world consists entirely of dispositions

100
To be sure, if one has a problem with even a single bare disposition then an entire world of pure powers
is that much worse.

101
Manley and Wasserman represent the former; Prior, Pargetter and Jackson represent one form of the
latter, Lewis another.

81
for all scientific intents and purposes. At the same time there is something about
the notion of a purely dispositional world that remains markedly objectionable
for some, and at least intuitively unsettling to many others.

Earlier I hinted that appealing to science for help is a promising method
for shedding light on how to solve some of these potential problems. To be sure I
strongly advocate a close tempering of theory by science here—to the full extent
possible. In other words, scientific realism about dispositions is a promising
route through the issues in which philosophical understanding of dispositions is
mired. Similarly it is well worth noting that, at least with respect to our
understanding of dispositions as they figure into the various sciences, there is a
strong sense of relativity to a context provided by the field and purpose of
inquiry. In particular, when focusing on one specific discipline it is always
possible and often necessary to restrict investigation by ignoring considerations
from other disciplines ('levels' or sciences).
Take chemistry for example: though it is ultimately grounded in physics,
it is not the case that in order to do chemistry we first have to work out the

82
physics and specify how the reduction goes.
102
This basic notion also gains support from a wide variety of examples
spanning a broad range of disciplines involved with the overall project of human
inquiry (i.e. from the hard/natural sciences and the social sciences). On any
account of dispositions, the ability to reduce a disposition D to a causal base C
alone simply does not eradicate D's relevance or autonomy, when restricted to a
field of inquiry outside that which contains the means for the reduction. More
strongly it is often not the case that identifying a disposition's causal grounds, or
Instead, the proper focus of
chemistry is restricted to the properties of various chemicals as revealed by
analyzing and describing their interactions. Similarly, though physicists speak of
chemical properties they do not speak of them qua chemical properties. Suppose
there is an important and undisputed sense in which chemical interactions are
fundamentally reducible to the types of properties and interactions that are the
proper domain of physics. That still does not divert the focus of chemists away
from the dispositions of chemical agents in favor of studying their underlying
bases as would a physicist. Nor vice versa.

102
There is an important distinction in this vicinity between reduction in practice versus reduction in
principle. Though my claims do not explicitly trade on the matter, I am not making the cagey move of
attributing a lack of reduction in practice to disciplines that display a lack of reducibility in principle.
Instead I am targeting a lack of reduction in science for either reason. If this causes discomfort, the same
claims can be motivated with a restricted appeal to a lack of reduction strictly in practice, or strictly in
principle.

83
indeed reducing the disposition to an underlying causal base, falls within the
proper jurisdiction (so to speak) of a given field of inquiry in question.
For further illustration take the social science of psychology, and consider
a psychologist whose patient is disposed to shy away from relationship intimacy
when she starts to become increasingly romantically involved with another
person. Assume this dispositional fear of intimacy results from childhood
experiences of a bad marriage between her parents, and manifests itself in that
the patient consistently stops courtship after a fourth good date with a would-be
significant other. Excluding treatment by way of medication, reducing this fear
of commitment qua disposition to some neural firing (described more precisely
by neurology and subsequently physics) moves the matter outside of
psychology. For instance, treatment (again excluding drugs) doesn't benefit from
reducing the relevant mental events to the firing of neural synapses. If anything,
the psychologist instead is better served by focusing on the memory at the level
of the client's experience, rather than its underlying causal base (neural synapse
activity).
The main point here is that virtually any scientific discipline other than
physics will in effect be pointless if required to focus myopically on the
underlying causal bases of those dispositions that are their proper domain. There

84
is an important sense in which science does proceed via identification of the
causal mechanisms behind observed phenomena qua categorical grounds of
dispositions. At the same time, it is often equally important to ignore the
presence of underlying causal bases from the perspective of specific fields of
inquiry. The useful operation of scientific disciplines in which dispositions are
featured other than particle physics sometimes simply requires a deliberate
restriction of focus, such that dispositions are not reduced in practice (though
they may undeniably be reducible in principle).
Thus even if it could be shown to obtain, the fact that there is always some
distinct underlying base is in some sense insignificant when inquiry is
undertaken in non-fundamental disciplines (outside of particle physics). It may
even be important to ignore such possibilities for reductive explanation, in order
to work efficaciously within a specific field (such as chemistry) without moving
into a different one (such as physics).
To be sure explanation by way of appealing to underlying bases is often
available, and is a cornerstone of the entire human scientific project. As Quine's
general approach to dispositions suggests, there is always (or at least almost
always) some such explanation in the offing. Instead, the point is that such
appeal to underlying bases can move us from one discipline into other separate

85
(though perhaps importantly related) ones, which are not always central to the
discipline in which the investigation in question occurs.
In other words, perhaps it really is always the case that there is some
scientific discipline or set of sciences capable of analyzing any given disposition
in terms of a distinct underlying causal base. Regardless, this does not
automatically preclude focusing exclusively in other 'non-reducing' disciplines
on the disposition in question, while ignoring the role played by any associated
base. Instead, this is exactly why there are different scientific disciplines.
103
The real source of categoricist concern regarding brute powers is not an
entire world of ungrounded dispositions, but rather the existence of any
ungrounded disposition. Science remains fully intact even if it is actually the case
that the world is purely dispositional—this discovery would not really affect the
vast majority of our scientific understanding of the world and its contents. At the
same time most dispositions (e.g. fragile) do seem to have causal bases that are in

103
Fodor (1974) frames a similar point more forcefully: "I am suggesting, roughly, that there are special
sciences not because of the nature of our epistemic relation to the world, but because of the way the world
is put together: not all natural kinds (not all the classes of things and events about which there are
important, counterfactual supporting generalizations to make) are, or correspond to, physical natural kinds."
pg. 113. Putnam (1977) hints at something similar in less severe terms, despite having argued previously
with Oppenheim (1958) that the unity of science remains a working hypothesis. Putnam (1975) also frames
something like this claim as an appeal to intuition: we are asked to consider whether a micro-level physical
description or a mid-level geometrical description does a better job of explaining why a mid-sized square
peg will not pass through a circular hole with a diameter slightly less than the square's width. He ends up
appealing to the greater generality of the higher-level geometrical description in support of the claim that
the microphysical reduction is for all intents and purposes irrelevant to the proper explanation. (pp. 130-
131)

86
an important sense not dispositional, even if they are themselves reducible at
some level to dispositions.
Why not then just accept that the vast majority of dispositions have some
distinct underlying non-dispositional causal base, but acknowledge that some
ungrounded dispositions exist—or, at least that the fundamental level alone
might be purely dispositional? This way the problem (if it is a problem) of a
purely dispositional world goes away, at least relative to the field(s) of inquiry in
question. Such an account, which combines categoricism about all levels above
the fundamental with a strong dispositionalist modalism about the bottom level,
looks initially promising.
Separate from this matter, it may turn out to be the case that accepting
bare dispositions at the fundamental level is tantamount to accepting a purely
dispositional world, since everything in some sense reduces to the bottom level.
However, even if this ends up occurring, it would not impact the paradigm for
scientific explanation displayed by the various sciences at various levels of
inquiry.

87
(4.4) Conclusion: Empirical Investigation of Brute Powers
Brute powers are not incoherent, nor are they intuitively impossible or
even explicitly distasteful—at least not for clearly principled reasons. Instead, as
I go on to argue in the final chapter they play a crucial role in the scientific
picture of reality, in ways that do not disrupt the rest of science (chemistry,
psychology, etc.).
104
Ultimately this is not a matter to be decided by philosophical analysis.
Instead it is a matter for rigorous science and rational intuition to determine. It is
an entirely open empirical question that cannot be resolved purely apriori. To
claim otherwise is to think that some contrived example like Manley's will have
such intuitive force that it will conclude the matter definitively.
So even if it turns out the world indeed is purely
dispositional, the real problem for categoricists seems to be the possibility of any
ungrounded disposition.
Consequently, arriving at a preference for modalism versus categoricism
regarding dispositions in a certain way amounts to a methodological choice. On
one side, modalist pictures draw an ontology of brute modality from science and
other systematic metaphysical considerations. On the other, categoricists often
relying upon an intransigent (Humean) metaphysical framework in which things

104
§5.1; 5.2.

88
act as they do because of the categorical properties they possess (along with the
laws of nature) combined with common patterns of scientific explanation. In the
final chapter I explore what science actually seems to say about the matter, and
endorse a corresponding account of dispositions.

89
Chapter (5): Scientific Modalism
(5.1) The Scientific Project
In addressing the opposition between modalism and categoricism
regarding dispositions, I have maintained that the entire debate can be cast so as
to turn on differing opinions regarding brute powers. Not perspectives on an
entirely dispositional world such as those articulated by Blackburn, however.
Instead the concern is a fundamental level of brute powers and primitive
modality—for simplicity the existence of any strictly ungrounded disposition.
Still, scientific inquiry writ large remains fully intact even if it is actually
the case that the world is purely dispositional—this discovery would not really
affect the vast majority of our empirical understanding of the world and its
contents. At the same time most dispositions (e.g. fragility) do seem to have
causal bases that are in an important sense not dispositions. This is the case even
if these bases are themselves reducible at some level to dispositions—scientific
explanation relative to specific higher-level disciplines proceeds regardless. Thus
the basic pattern of scientific explanation has indeed yielded considerable data
suggesting that most dispositions have categorical grounds: glasses break when
struck because they are fragile; they are fragile because of their underlying
microphysical structure, and so on.

90
Ultimately however this same pattern implies that some dispositions—at
the bottom level—must be ungrounded. The alternative is a troublesome regress.
In light of this consideration it is a small concession to react by attempting to
accept both the prevalence of categorical grounds in nature and pure powers at
bottom. Opinions to the contrary require a strong pre-theoretic intuition either
that the fundamental level simply cannot contain ungrounded modal features, or
that the bottom rung of reality alone includes primitive modal attributes like
brute powers.
105

Yet it seems unlikely we have such pre-theoretic intuitions
about the bottom levels of reality at all. Absent such prejudices the proposed
approach avoids the problem (if it is a one) of a purely dispositional world, at the
least relative to any given non-fundamental field of inquiry. In what follows I
explore one possibility already suggested for this kind of conciliatory approach.
(5.2) The Fundamental Level
What does our best empirical inquiry tell us about the nature of the
fundamental level and its constituents? For one thing, properties of occupants of
the fundamental level are much different from higher-level features. Take inertial

105
For more on pre-theoretic modal intuitions, see Peacocke (1999), esp. pp. 155-156.

91
mass in classical Newtonian mechanics, which is defined as a particle's tendency
to resist acceleration by an outside force:
(IM) Particle Q has determinate inertial mass m iff Q has a tendency to be
accelerated by F/m when acted upon by force F
106

Such a relational property of Q involves outside forces like gravitational mass as
relata. These forces themselves are characterized in relational terms;
107
Assuming that any particles populating the fundamental level are simple
(an assumption spelled out further below), the preceding characterization is
strikingly reminiscent of the relationally-characterized scenario in Holton's
example.
the
systematic understanding enabled by such characterizations historically are
among the greatest scientific discoveries (hence 'Newtonian' mechanics…).
108
The simple particles populating the bottom level, which are bearers of
these properties, do not have parts. This follows from the earlier definition of
One force is characterized in terms of its bearers' relations to relata
that are relationally characterized in terms of relations to other relationally-
characterized relata.

106
Latham (2004), pg. 4

107
(GM) Particle Q has determinate gravitational mass iff it has a tendency to attract and be attracted by
other particles possessing determinate gravitational mass with a force quantified by the familiar Gmm/r
2
.

108
§4.2. To those versed in the history and philosophy of science this should not actually come as such a
revelation. For instance Poincaré is one among many credible proponents of such a characterization of
science, e.g. (1905), pp. 151-152.

92
their populating the fundamental level: if they had parts it would not be the
fundamental level. No underlying structural properties are available to serve as
causal bases in explanations of the forces manifested by constituents of this level.
And the only non-structural intrinsic properties of the particles are ungrounded
as well. Even if these properties were not thus dispositional by definition, it
would still be a long way toward connecting any non-structural categorical
property to a dispositional property at its same level, as its causal basis in
categoricist terms. Classical mechanics therefore seems to stick us with brute
powers at the fundamental level.
Paradigmatically categorical qualities such as spatio-temporal properties
to be sure can conceivably be characterized relationally, and as relating things
with no intrinsic features of their own. For example if space-time points exist
then the qualitative properties of any one such point can entirely consist of
relations to others, without exhibiting any other intrinsic properties. Consider a
world of physical properties that are grounded in the geometrical properties of
space-time points. The types of physical properties populating such a world are
relational, and knowable in principle absent any intrinsic properties—the relata
are all space-time points, with no intrinsic features.
109

109
Esfeld (2001), pp. 401-402.

93
However far-fetched such a proposal might seem initially, the more
modern physicists discover about the most basic levels of reality the more
applicable it becomes. Without question it is hard to come by agreement amongst
theorists about things like the metaphysical implications of quantum mechanics.
Regardless of specific interpretation, it is clear that the way the world
intrinsically is differs greatly from the way it is in our experiences. Moreover,
most interpretations have in common some spin on contemporary findings to the
effect that that sub-atomic particles—the basic building blocks of matter—are
essentially ungrounded tendencies to exist. The question is not whether signs
point to this feature of reality, but what to make of the fact that signs do point to
it.
For instance, one interpretation of quantum reality (the behavior
displayed by the most basic particles known to humanity) takes the world to be
"thoroughly relational, containing no privileged entities of any kind; a world
entirely thingless, full of quons like rainbows which are themselves made of
rainbows."
110

110
Herbert (1985), pg. 247.
A competing reading takes the ultimate constituents of matter to be
basic tendencies or possibilities: "not a world of actual events like our own but a

94
world of unrealized tendencies for action."
111
Physics at least since Newton has therefore continually produced findings
that suggest certain occupants of the fundamental level of reality are much
different from how reality appears to us to be. Increasingly there is evidence of
ungrounded modal features such as dispositions too. The extent to which our
experiences of the world's middle levels diverge from what we have learned
scientifically about its bottom levels might surprise many. But in a sense this is
an intuitively appealing outcome of scientific inquiry. Science has as its target a
complete taxonomy of what there is in the world, with emphasis on the
fundamental things on which all else is based. Part of this basic project is a
specification of what properties and relations the world's fundamental
constituents have and enter into.
Such formulations might appear too
footloose and fanciful for science. But they are explicit attempts at articulating
the basic quantum frameworks of John von Neumann and Werner Heisenberg,
respectively!
112
Understood in this way, the fact that science determines the fundamental
level to be populated by ungrounded modal features should be welcomed as an

111
Ibid, pg. 27.

112
Armstrong (1989), pg. 22.

95
indication of progress toward accomplishment of the overarching goal of
systematic empirical inquiry. Excessive resistance on the basis of epistemic
worries or metaphysical prejudice just doesn't make sense in the face of the
guiding purpose of the scientific project. Along with less obscure strands of
scientific investigation, cutting-edge contemporary particle physics undeniably
(though in a manner open to different specific interpretations) paints a picture of
ungrounded modal features such as tendencies to act—brute powers—existing at
the foundations of the world.

(5.3) Scientific Modalism
I propose to temper a modalist view of dispositions along the lines of
(PROP)
113

113
§2.5.
with scientific findings about ungrounded modal features at the
fundamental level. In light of the legwork on dispositions others have already
performed, this can be accomplished in short order. Since (PROP) is agnostic
about what kinds of properties play the relevant roles with respect to any given
disposition, all that is needed is a combination of categoricism about non-
fundamental dispositions with the modalistic (SD) about the fundamental level,

96
grounded in appeal to the actual results of scientific investigation into the nature
of these ultimate constituents of reality.
Again scientific explanation in general does operate as though there is
always (or at least almost always) some distinct underlying base causally
responsible for the presence of any given disposition. But at the same time it also
looks as though all we get at the bottom level from science (if not on the entire
way down to the bottom) is dispositions. At bottom reality is indeed defined
entirely in terms of brute tendencies to act in certain ways—evidence that there
are bare dispositions (or at least that physics posits them).
This means to conclude based on patterns of scientific explanation that
there is always a distinct causal base underlying any disposition is as well to fly
in the face of the purest form of modern science we have (contemporary particle
physics). It also goes against the contextualized mode of inquiry (relative to the
perspective of any given discipline of inquiry) exhibited in virtually all rigorous
fields of empirical investigation.
Now consider briefly the basics, unladen by theory (to the extent
possible). In order for anything to exist as a part of the actual world, it must
impact the world somehow—which in turn requires some form of interaction
with the world. Without some power or tendency to interact with the world,

97
such interactions would not occur. Therefore these tendencies qua dispositions
have intuitive appeal as a prima facie mark of existence. This appeal is supported
by the fact that subatomic particles are defined by modern physics as primitive
tendencies to exist and interact, rather than as categorical aspects of reality.
Such scientific findings in turn give (SD) modalists a major leg up in their
debate with categoricists, because the data suggest that modal features such as
dispositions and counterfactuals can indeed be primitive.
114
Categoricism does not suffer from this weakness—categoricist accounts
successfully mirror the pattern of scientific explanation, accounting for
dispositions in terms of underlying categorical grounds. This method works at
all non-fundamental levels. However, extant versions of categoricism fly in the
This result comes as
a major benefit of letting science help determine ontology, rather than the other
way around. Unfortunately on its own (PROP) has the shortcoming that it is
agnostic about what the intrinsic nature of non-fundamental properties might be.
This amounts to reticence regarding causal grounds at levels higher than the
fundamental, contra science.

114
Swoyer (1982), McKitrick (2005) and others at times sounds as though they take primitive modal
features to be a required feature of successful scientific theories. For example, Swoyer considers that the
ungrounded tendencies of fundamental particles ground necessary laws of nature, sustain counterfactuals
and the like. I leave questions about this (and any other) version of metaphysically necessary laws qua
brute modality for a later project.

98
face of science regarding the bottom level, since they reject pure ungrounded
powers in principle—regardless of what science suggests to the contrary.
115
To revive each view all that is needed is a reversal in the order of priority:
allowing science to constrain ontology. Hence my moderate Scientific Modalism
regarding dispositions—a combination of cateogoricist (PROP) with modalist
(SD):

Scientific Modalism (SM) X is disposed to give response R to stimulus
S iff X would give response R in some suitable proportion of S-cases
(PROP)

Dispositions such as the one X has above are grounded in underlying
intrinsic categorical qualities qua causal basis at levels above the
fundamental, while at the fundamental level science reveals only primitive
modal features ungrounded in anything qualitative.

Such an approach avoids running afoul of science and metaphysics in the
manner each individual component falls victim to on its own—yet it gets the
causal basis right all the way up and down the ontological spectrum. It handles
the other main branch of the general project with respect to dispositions too—it
provides a viable account of the necessary and sufficient conditions for

115
As discussed previously in §3.4. Ultimately, identity dispositionalism (IDD) founders on the topic of
pure brute powers too: since all properties are dispositional and qualitative, even if powers at the
fundamental level are ungrounded in any occupant of a lower level, they are not pure powers because the
very same brute power is also a qualitative property—which could count as a categorical ground for the
disposition. This contradicts what science says about pure ungrounded powers, but does not qualify as
categoricist since the grounds are not on a lower, underlying level.

99
ascriptions of dispositions.
116

In addition it does so with the extra benefit of a
rigorous (read: scientific) metaphysical basis for why the account works, in terms
of those dispositions that actually are in the world—at all relevant levels of
reality.
(5.4) Conclusion—Eleatic Ontology
At the core of this scientific modalism (SM) regarding dispositions is the
notion that interaction is basic—at least in that a primitive, ungrounded tendency
for action is the most fundamental mark of reality. This is not a new idea. It
traces back at least to a suggestion voiced by Plato's Eleatic stranger, if not
earlier. Nor is it a recent finding that science corroborates this basic intuition—
certain fundamental properties in Newtonian mechanics receive entirely
relational characterizations; modern physics has added to the evidence of
ungrounded modal features at the fundamental level. Albeit in mystifying ways,
our best methods of empirical inquiry reveal ungrounded powers at the
foundations of reality.
At the same time, many things in the world are better characterized in
terms of the higher-level properties they display than through their underlying

116
Viable in the sense of being immune to masks, finks, etc. §2.4.

100
structural features—even if there is some sense in which the underlying traits
explain the higher-level ones. Clearly most things populating any non-
fundamental level have various potentialities for different actions associated
with them. But such tendencies can almost always be readily accounted for in
terms of underlying categorical bases, rather than possessing a necessarily
ungrounded modal profile.
Some philosophical accounts of dispositions dispense with the project of
specifying necessary and sufficient conditions for ascriptions of dispositions with
what seems like comparative ease.
117
Ultimately both strands (epistemic/linguistic and metaphysical) are key
aspects of the philosophical analysis of dispositions—separately each is
necessary but neither suffices for an analysis; only together are they each
necessary and sufficient. The scientific brand of modalism I have proposed
However they remain silent regarding the
project of accounting for what dispositions are in the deeper, metaphysical sense
of how they are structured and the like. Other philosophical accounts such as
Prior's 'functionalist' categoricism address both issues, but do so in a manner that
amounts to resisting the findings of science—a bad policy for any metaphysical
project, with respect to dispositions or otherwise.

117
For instance Nominalism (N) about dispositions, as discussed in §3.4.

101
regarding dispositions is a unifying view: it retains a viable set of necessary and
sufficient conditions for ascribing dispositions even in the face of
counterexamples, and addresses what dispositions fundamentally are. What's
more, it does so in a manner that is compatible with the empirical results of
scientific investigation. Such considerations distinguish (SM) from and
recommend it over other, less versatile philosophical accounts of dispositions.
These benefits all result from a pragmatic scientific realism, with a
grounding in the Eleatic intuition that nothing could be more basic than an
ungrounded tendency for action. This intuition turns out to be an extremely
fruitful point of departure in the hunt for a philosophical account of dispositions.

102
Bibliography
Armstrong, D. M. (1978) Universals and Scientific Realism. Cambridge University
Press.
–––. (1983) What is a Law of Nature? Cambridge University Press.
–––. (1989) Universals: An Opinionated Introduction. Westview Press.
–––. (1997) A World of States of Affairs. Cambridge University Press.
–––. (1999) "The Causal Theory of Properties: Properties According to
Shoemaker, Ellis, and Others." Philosophical Topics, Vol. 26. (25-37)
Armstrong, D. and Heathcote, A. (1991) "Causes and Laws." Noûs, Vol. 25. (63-73)
Bigelow, J., Ellis, B. and Lierse, C. (1992) "The World as One of a Kind: Natural
Necessity and Laws of Nature." British Journal for the Philosophy of Science, Vol. 43.
(371-388)
Bird, A. (1998) "Dispositions and Antidotes." The Philosophical Quarterly, Vol. 48
(227–234)
–––. (2005) "The Dispositionalist Conception of Laws." Foundations of Science, Vol.
10. (353-370)
–––. (2007) Nature's Metaphysics: Laws and Properties. Oxford University Press.
Blackburn, S. (1990) "Filling in Space." Analysis, Vol. 50. (62-65)
Bonevac, D., Dever, J., and Sosa, D. (2006) "The Conditional Fallacy." Philosophical
Review, Vol. 115 No. 3. (273-316)
–––."The Counterexample Fallacy." Mind. (forthcoming)
Carnap, R. (1936) "Testability and Meaning." Philosophy of Science, Vol. 3. (420-
468)

103
Chalmers, D. (1996) The Conscious Mind: In Search of a Fundamental Theory. Oxford
University Press.

–––. (2002) "Does Conceivability Entail Possibility?" Conceivability and Possibility,
eds. T. Gendler and J. Hawthorne, Oxford University Press. (145-200)
Choi, S. (2005) "Do Categorical Ascriptions Entail Counterfactual Conditionals?"
The Philosophical Quarterly, Vol. 55. (495–503)
–––. (2006) "The Simple vs. Reformed Conditional Analysis of Dispositions."
Synthese, Vol. 148. (369–379)
Cross, T. (2005) "What is a Disposition?" Synthese, Vol. 144. (321-341)
Davidson, D. (1967) "Causal Relations." Journal of Philosophy, Vol. 64 No. 21. (691-
703)
Dretske, F. (1977) "Laws of Nature." Philosophy of Science, Vol. 44. (248-268)
Drewery, A. (2005) "Essentialism and the Necessity of the Laws of Nature."
Synthese, Vol. 144. (381-396)
Edgington, D. (2004) "Two Kinds of Possibility." Proceedings of the Aristotelian
Society, Vol. 8. (1-22)
Ellis, B. (2001) Scientific Essentialism. Cambridge University Press.
–––. (2002) The Philosophy of Nature: A Guide to the New Essentialism. Queen's
University Press.
–––. (2005) "Physical Realism." Ratio, Vol. 18. (371-384)
–––. (2005) "Universals, the Essential Problem and Categorical Properties." Ratio,
Vol. 18. (462-472)
–––. (2010) "Causal Powers and Categorical Properties." The Metaphysics of Powers,
ed. A. Marmadoro, Routledge. (133-142)

104
Ellis, B. and Lierse, C. (1994) "Dispositional Essentialism." Australasian Journal of
Philosophy, Vol. 72 No. 1. (27-45)
Esfeld, M. (2001) "Review of Langton’s Kantian Humility." Erkenntnis, Vol. 54.
(399-403)
Fales, E. (1979) "Relative Essentialism." British Journal for the Philosophy of Science,
Vol. 30. (349-370)
–––. (1982) "Natural Kinds and Freaks of Nature." Philosophy of Science, Vol. 49.
(67-90)
Fara, M. (2005) "Dispositions and Habituals." Noûs, Vol. 39. (43-82)
Fine, K. (1994) "Essence and Modality." Philosophical Perspectives, Vol. 8. (1-15)
Fodor, J. (1974) "Special Sciences (Or: The Disunity of Science as a Working
Hypothesis)." Synthese, Vol. 28. (97-115)
–––. (1997) "Special Sciences: Still Autonomous After All These Years."
Philosophical Perspectives, Vol. 11 (Mind, Causation, and World). (149-163)
Gibbard, A. (1981) "Two Recent theories of Conditionals." Ifs: Conditionals, Beliefs,
Decision, Chance, and Time, eds. W.L. Harper, R. Stalnaker and G. Pearce, D.
Reidel (211-247).
Goodman, N. (1955) Fact, Fiction and Forecast. Harvard University Press. (40-49)
Harré, R. and Madden, E. (1975) Causal Powers. Basil Blackwell Publishing.
Hawthorne, J. (2002) "Advice for Physicalists." Philosophical Studies, Vol. 108. (17-
52)
–––. (2002) "Causal Structuralism." Philosophical Perspectives, Vol. 15, reprinted in
Metaphysical Essays, Oxford University Press (2006).
Hawthorne, J. and Manley, D. (2005). "Stephen Mumford's Dispositions." Noûs,
Vol. 39 No. 1. (179-195)

105
Heil, J. (1999) "Multiple Realizability." American Philosophical Quarterly, Vol. 36
No. 3. (189 -208)
–––. (2004) "Properties and Powers." Oxford Studies in Metaphysics, ed. D.
Zimmerman, Clarendon Press. (223-254)
–––. (2005) "Dispositions." Synthese, Vol. 144. (343-356)
–––. (2005) "Kinds and Essences." Ratio, Vol. 8. (405-419)
–––. (2010) "Powerful Qualities." The Metaphysics of Powers, ed. A. Marmadoro,
Routledge. (58-72)
Heller, M. (1998) "Property Counterparts in Ersatz Worlds." Journal of Philosophy,
Vol. 95 No. 6. (293-316)

Herbert, N. (1987) Quantum Reality: Beyond the New Physics. Anchor Press.
Holton, R. (1999) "Dispositions All the Way Round." Analysis, Vol. 59. (9-14)
Hume, D. (1978)[1739] A Treatise of Human Nature. Oxford: Clarendon Press.
Johnston, M. (1992) "How to Speak of the Colors." Philosophical Studies, Vol. 68
No. 3. (221-263).
Kim, J. (1993) Supervenience and Mind: Selected Philosophical Essays. Cambridge
University Press.
Kripke, S. (1980) Naming and Necessity. Harvard University Press.
Lange, M. (2004) "A Note on Scientific Essentialism, Laws of Nature, and
Counterfactual Conditionals." Australasian Journal of Philosophy, Vol. 82 No. 2.
(227-241)
Langton, R. (1998) Kantian Humility: Our Ignorance of Things in Themselves. Oxford
University Press.
Langton, R. and Lewis, D. (1998) "Defining Intrinsic." Philosophy and
Phenomenological Research, Vol. 58 No. 2. (333-345)

106
Lewis, C.I. (1923) "A Pragmatic Conception of the A Priori." Journal of Philosophy,
Vol. 20 No. 7. (169-177)
–––. (1924) Mind and the World Order. Dover Publications.
Lewis, D. (1970) "How to Define Theoretical Terms." Journal of Philosophy Vol. 67
No. 13. (427-446)
–––. (1983) "New Work for a Theory of Universals." Australasian Journal of
Philosophy, Vol. 61. (343-377)
–––. (1986) On the Plurality of Worlds. Blackwell Publishing.
–––. (1994) "Finkish Dispositions." The Philosophical Quarterly, Vol. 47 No. 187.
(143-158)
–––. (2001) "Ramseyan Humility." Unpublished manuscript.
Loewer, B. (1996) "Humean Supervenience." Philosophical Topics, Vol. 24. (101-
126)
Manley, D. (2002) "Properties and Resemblance Classes. Noûs, Vol. 36 No. 1. (75-
96)
–––. (2009) "When Best Theories Go Bad." Philosophy and Phenomenological
Research, Vol. 78 No. 2 (392-405)
–––. (forthcoming) "Dispositions: Beyond the Biconditionals." (under review)
Manley, D. and Wasserman, R. (2007) "A Gradable Approach to Dispositions."
Philosophical Quarterly, Vol. 57. (68-75)
–––. (2008) "On Linking Dispositions with Conditionals." Mind, Vol. 117 No. 465.
(59-84)
Martin, C. B. (1993) "Power for Realists." In Ontology, Causality, and Mind: Essays
on the Philosophy of D.M. Armstrong, Cambridge University Press. (175-193)

107
–––. (1994) "Dispositions and Conditionals." The Philosophical Quarterly Vol. 44.
(1–8.)
–––. 1997, "On the Need for Properties: The Road to Pythagoreanism and Back."
Synthese, Vol. 112. (193–231)
Martin, C. B. & Heil, J. (1998) "Rules and Powers." Philosophical Perspectives, Vol.
12. (283-312)

–––. (1999) "The Ontological Turn." Midwest Studies in Philosophy, Vol. 23 (34–60)
Maudlin, T. (2007), "A Modest Proposal Concerning Laws, Counterfactuals and
Explanations," Metaphysics Within Physics. Oxford University Press (5-50)
McKitrick, J. (2003a) "A Case for Extrinsic Dispositions." Australasian Journal of
Philosophy, Vol. 81. (155–174)
–––. (2003b) "The Bare Metaphysical Possibility of Bare Dispositions." Philosophy
and Phenomenological Research, Vol. 66 (349–369.)
–––. (2005) "Are Dispositions Causally Relevant?" Synthese, Vol. 144 (357–371)
–––. (2010) "Manifestations as Effects." The Metaphysics of Powers, ed. A.
Marmadoro, Routledge. (73-83)
Mellor, D.H. (1982) "Counting Corners Correctly." Analysis, Vol. 42 No. 2 (96-97)
–––. (1974) "In Defense of Dispositions." Philosophical Review, Vol. 83. (157-181)
–––. (2000) "The Semantics and Ontology of Dispositions." Mind, Vol. 109 No.
436. (757-780)
Molnar, G. (1999) "Are Dispositions Reducible?" The Philosophical Quarterly, Vol.
49. (1–17)
–––. (2003) Powers: A Study in Dispositions. Oxford University Press.
Mumford, S. (1995) "Ellis and Lierse on Dispositional Essentialism." Australasian
Journal of Philosophy, Vol. 73 No. 4. (606-612)

108
–––. (2004) Laws in Nature. Routledge.
–––. (2005) "Kinds, Essences, Powers." Ratio, Vol 62. (420-436)
Nagel, E. (1979) The Structure of Science. Hackett Publishing Company.
Oddie, G. (1982) "Armstrong on the Eleatic Principle and Abstract Entities."
Philosophical Studies, Vol. 41. (285-295)
Oppenheim, P. and Putnam, H. (1958) "The Unity of Science as a Working
Hypothesis." Minnesota Studies in the Philosophy of Science, Vol. 2. (3-34)
Peacocke, C. (1999) Being Known. Oxford University Press.
Poincaré, H. (1905) Science and Hypothesis. Walter Scott Publishing.
Prior, E. (1982) "The Dispositional/Categorical Distinction." Analysis, Vol. 42 No.
2. (93-96)
–––. (1985) Dispositions. Aberdeen University Press.
Prior, E., Pargetter, R. and Jackson, F. (1982) "Three Theses about Dispositions."
American Philosophical Quarterly, Vol. 19 No. 3. (251-257)
Psillos, S. (2005) "Scientific Realism and Metaphysics." Ratio, Vol 18. (385-404)
Putnam, H. (1975) "Philosophy and Our Mental Life." In Mind, Language and
Reality, Cambridge University Press. (291-303)
–––. (1977) “Realism and Reason.” Proceedings and Addresses of the American
Philosophical Association, Vol. 50, Reprinted in Meaning and the Moral Sciences,
Routledge (1978). (123-138)
Quine, W.V.O. (1948) "On What There Is." Review of Metaphysics. Reprinted in
From a Logical Point of View, Harvard University Press (1953). (1-19)
–––. (1951) "Two Dogmas of Empiricism." Philosophical Review, Vol. 60. Reprinted
in From a Logical Point of View, Harvard University Press (1953). (20-46)

109
–––. (1953) "Reference and Modality." From a Logical Point of View, Harvard
University Press. (139-159)

–––. (1960) Word and Object. MIT Press. (222-227)

–––. (1969) "Natural Kinds." Ontological Relativity and Other Essays, Columbia
University Press. (114-138)

Ryle, G. (1949) The Concept of Mind. Hutchinson & Co. (116-153)
Salmon, N. (2005) Reference and Essence. Prometheus Books.
Schaffer, J. (2004) "Two Conceptions of Sparse Properties." Pacific Philosophical
Quarterly, Vol. 85. (92-102)
–––. (2005) "Quiddistic Knowledge." Philosophical Studies, Vol. 123. (1-32)
Scheffler, I. (1963) The Anatomy of Inquiry. Knopf.
Shoemaker, S. (1975)"Functionalism and Qualia." Philosophical Studies, Vol. 27.
(291-315)
–––. (1980) "Causality and Properties." Time and Cause. D. Reidel, reprinted in
Properties, eds. D.H. Mellor and A. Oliver, Oxford University Press (1999). (228-
254)
–––. (1998) "Causal and Metaphysical Necessity." Pacific Philosophical Quarterly,
Vol. 79. (59-77)

Sidelle, A. (2002) "On the Metaphysical Contingency of Laws of Nature."
Conceivability and Possibility, eds. T. Gendler and J. Hawthorne, Oxford
University Press. (309-336)

Sober, E. (1982) "Dispositions and Subjunctive Conditionals, or, Dormitive
Virtues Are No Laughing Matter." Philosophical Review, Vol. 41 No. 4 (591-596)

–––. (1984) "Two Concepts of Cause." Philosophy of Science Association, Vol. 2 (405-
424)

110

–––. (1988) "Confirmation and Lawlikeness." Philosophical Review, Vol. 97 No. 1.
(93-98)

–––. (1997) "Two Outbreaks of Lawlessness in Recent Philosophy of Biology."
Philosophy of Science, Vol. 64 (proceedings) (S458-S467)

Sosa, E. (1975) Causation and Conditionals. Oxford University Press.

Stalnaker, R. (1968) "A Theory of Conditionals." Studies in Logical Theory, ed. N.
Rescher, Basil Blackwell. (98-112)

–––. (1976) "Possible Worlds." Noûs, Vol. 10. (65-75)

Swoyer, C. (1982) "The Nature of Natural Laws." Australasian Journal of
Philosophy, Vol. 60 No. 3. (203-223)

Taylor, B. (1993) "On Natural Properties in Metaphysics." Mind, Vol. 102 No. 405.
(81-100)

Tooley, M. (1977) "The Nature of Laws." Canadian Journal of Philosophy, Vol. 7 No.
4. (667-698)

–––. (1999) Necessity and Possibility: The Metaphysics of Modality. Garland
Publications.

–––. (1999) Laws of Nature, Causation, and Supervenience. Garland Publications.

Van Fraassen, B. (1989) Laws and Symmetry. Clarendon Press.

Van Inwagen, P. (2004) "A Theory of Properties." Oxford Studies in Metaphysics,
ed. Dean Zimmerman, Oxford University Press.

Wiggins, D. (2001) Sameness and Substance Renewed. Cambridge University Press.

Linked assets