Académique Documents
Professionnel Documents
Culture Documents
Managing Editor
Matthew Benton (Rutgers University)
OXFORD STUDIES
IN METAPHYSICS
Volume 7
Edited by
Karen Bennett
and
Dean W. Zimmerman
1
1
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© The several contributors 2012
The moral rights of the authors have been asserted
First Edition published in 2012
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
ISBN 978–0–19–965908–1 (Hbk.)
978–0–19–965907–4 (Pbk.)
Printed in Great Britain by
MPG Books Group, Bodmin and King’s Lynn
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
PREFACE
D.W.Z.
New Brunswick, NJ
This page intentionally left blank
CONTENTS
I. RELATIVE SAMENESS
dwzimmer@rci.rutgers.edu
I
RELATIVE SAMENESS
This page intentionally left blank
1. Possibility Relative to a Sortal
Delia Graff Fara
they were when they were a child because their interests and values
have changed drastically. In my view, such people are not thereby
making a category mistake. They have not thereby confused qualita-
tive identity with numerical identity. They have not made the mistake
of thinking that significant qualitative change over time precludes
numerical identity over time. Rather, they have made a mistake
about what conditions are required for a particular sort-relativized
sameness relation to hold over time. They mistakenly think that sig-
nificant change with respect to interests and values precludes same-
ness of person over time. That latter relation, sameness of person, is
not, on my view, the relation of numerical identity—even when
restricted just to persons.
In addition to the distinction between sameness relativized to a
quality and sameness relativized to a sort, there is a distinction to be
made even within the category of sort-relative sameness relations.
There is sort-relative sameness of types and also sort-relative same-
ness of tokens. If you and I are both wearing blue platform wedges
then there is a sense in which we are wearing the same shoes even
though the ones on my feet are not on your feet. We are wearing the
same type of shoes, but not the same individual (“token”) shoes
Peter Geach would have thought that in the case of our shoes, not
only are you and I wearing the same type of shoes, my shoes and
your shoes are the same relative to a certain sort, namely, the sort
shoe-type (Geach, 1962). Of course Geach did think that my shoes and
your shoes were not the same shoe-tokens. I disagree with the former
claim. We do not have a case here of things being the same relative to one
sort of thing while not being the same relative to another sort of thing.
Our shoes are not the same shoe-type while being different shoe-
tokens. Although we are wearing the same type of shoes on our feet,
neither of us is wearing a shoe-type on our feet. In other words,
although we are wearing the same type of shoes, there is nothing on
my feet that is the same, relative to any sort, as anything on your feet.
On this point, I agree with William Alston and Jonathan Bennett
(1984) and with John Perry (1970), as against Geach. From here
onwards, I will ignore sort-relative sameness of types.
I will also from here onwards ignore relations of sameness rela-
tivized to a quality. So as shorthand I will use the terms “relative
sameness” and “sort-relative sameness”—even sometime just
“sameness”—to mean sort-relative sameness of tokens.
Possibility Relative to a Sortal | 5
1
The view is most associated with Geach (1962).
2
A number of philosophers do allow for the space: Peter Geach (1962, 1967, 1972);
Roderick Chisholm (1973, 1976); and Anil Gupta (1980) are examples.
3
With respect to identity over time, I here agree with Roderick Chisholm (1976),
who thought that this ship here now could be the same ship as that one there then
even though they are not numerically identical. Like Chisholm, I also think that the
same-ship relation, rather than the numerically-identical relation is the one that is rele-
vant for counting ships. Following Bishop Butler, though, Chisholm accepted that it
is only in a “loose and popular sense” that identity may hold between numerically
distinct ships. On that view, for two objects to be the same ship they must each be a
ship and must be identical in the loose and popular sense. But for the two objects to
be numerically the same thing is for them to be identical in the “strict and philosophical”
sense. We can speak of identity “in the loose and popular sense” if we like, but there
is really no need to once we recognize that being the same F over time does not
require numerical identity (identity in the strict and philosophical sense) over time.
6 | Delia Graff Fara
4
Such a relative-sameness counterpart theory is discussed briefly by Theodore
Sider (1999), who attributes it to Michael Jubien (1993), although Jubien would disa-
vow it. One finds a kernel of the same theory in Jubien (2001).
Possibility Relative to a Sortal | 7
Granting this much, what do we say about the truth of the fol-
lowing modal and temporal claims? What conditions must be met
in order for them to be true?
(1) Angel, Michael’s boat, might have been rowed ashore.
(2) Angel will be rowed ashore.
The less popular answer, derived from Lewis (1968, 1971, 1986) for
the modal case and from Katherine Hawley (2001) and Theodore
Sider (1996, 2001) for the temporal case, is that these sentences are
true just in case there is a possible world (in the case of (1)), or a
future time (in the case of (2)), at which something that is a counter-
part of Angel is rowed ashore. For the moment, let us leave behind
the question of just what this counterpart relation is. The important
point for us now is that it may hold between things that are not
identical to each other.
The more popular answer, derived from Kripke (1963, 1972), is
that these sentences are true just in case:
(A) There is a possible world, or future time, at which some-
thing identical to Angel is rowed ashore.
Do we accept the popular answer, so phrased? For Kripke, and
those who follow him on this question, there is no distinction
between that condition and another:
(B) There is a possible world, or future time, at which some-
thing that is the same boat as Angel is rowed ashore.
But for us, the two conditions come apart. The leading idea for us is
that (A) and (B) are different views. We may accordingly say that
(B), the second phrasing of the popular view, is the correct one;
while (A), the first phrasing, is incorrect. A de re modal claim “a
might have been Φ” is true when there is something in some pos-
sible world to which a bears a certain relative-sameness relation
and which is Φ in that world—whether or not there is something
identical to a that is Φ in that world. Likewise, a de re future claim “a
will be Φ” is true when there is something in the future to which a
bears a certain relative-sameness relation and which is Φ at that
future time—whether or not there is something identical to a that is
Φ at that time. Which sameness relation is involved (same person, same
Possibility Relative to a Sortal | 9
body, same boat, et cetera) varies with the context.5 We will have to
revise and sharpen this up in a bit. But before we do, I want to point
out that this view about the truth conditions for de re predications
(temporal or modal) has something in common with both the popu-
lar view and the Lewisian view.
It shares with the popular view the idea that future and coun-
terfactual possibilities for a thing a are determined by the ways
that a thing b is at a future time, or possible world, whenever a
bears a certain relative-sameness relation to b. Because the the-
ory has this feature, I call it a relative-sameness theory of de re
predication.
It shares with the Lewisian view the idea that future and counter-
factual possibilities for a can be determined by the ways that a thing
b is at a future time, or possible world, even if a is not identical to b.
Because the theory has this feature, I call it a counterpart theory of
de re predication.
Let me emphasize two things before moving on. One, I intend for
relative-sameness counterpart theory to cover temporal predication
(de re) as well as modal predication (de re). It simplifies matters,
however, if I focus on just one of these at any given time. For the
most part I focus on modal predication, but it should be somewhat
clear how the view extends to temporal predication.
Two, I use the term ‘counterpart theory’ liberally. To be a counter-
part theorist is not to have any particular view about what it takes
for an individual in one possible world to be a counterpart of some
individual in some different possible world. Any theory of de re
modality is a counterpart theory in my sense if it takes the trans-
world relation that is important for the analysis of de re possibility
to be a relation other than identity. For example, someone who held
the ridiculous view that it is true that a certain woman might have
worn green shoes today just in case there’s a possible world in
which she has a sister who is wearing green shoes today would be a
counterpart theorist in my sense.
5
This appeal to contextual variation can be found in Lewis (1971) and Allan
Gibbard (1975). Gibbard’s views on contingent identity inspired me to develop the
view presented here. I differ from both Lewis and Gibbard on a number of important
issues—among them, the question whether objects are extended in time.
10 | Delia Graff Fara
6
Mark Johnston has argued that there is no plausible argument for the identity
thesis (Johnston 1992). Harold Noonan (1993) argued, in response to Johnston, that
not all of Johnston’s dismissals of arguments for the identity thesis are justified.
That’s as may be. I do not take myself to be providing a direct argument for the iden-
tity thesis, nor even a rebuttal to Johnston’s arguments that there is no plausible such
argument. Rather I take myself in what follows, to be giving a defense of the identity
thesis—since it’s a thesis that I find attractive—by explaining how it does not lead to
a contradiction.
7
I appreciate comments from Dean Zimmerman that led to this formulation of the
stipulation.
Possibility Relative to a Sortal | 13
I use one or the other phrase according to what sounds more natural
to my ear.)
The leading idea of this paper enables us to coherently maintain
The Sensible View. The leading idea is that sort-relative sameness
does not require identity. This boat here now might be the same
boat as that one there later even though they are not identical.
Clearly, anyone who thinks this—that ordinary material objects,
like boats, are identical to the matter that they are made up of even
though the objects, unlike the matter itself, could be made up of dif-
ferent matter—must accept some statements of contingent identity.
Angel is identical to Lumber, but she might not have been.
The leading idea allows us to make sense of this (merely) seem-
ing contingency of identity. It allows us to say that it is true that
Angel might not have been identical to a thing x when there’s a
possible world in which there’s some boat, Angel*, that is the same
boat as Angel but which is not identical to x. Let us introduce a piece
of terminology. We will call Angel* Angel’s “boat-mate” in that
world.
Since Angel is not identical to her otherworldly boat-mate, there
is no contradiction in her being identical to Lumber even though
Lumber is not identical to the boat-mate. The following are
compatible:
(8) Angel = Lumber;
(9) Angel*, in the counterfactual world β, is the same boat as
Angel;
(10) Lumber ≠ Angel*.
These are compatible since being a boat-mate does not require iden-
tity. The joint truth of (8)–(10) is what on my view renders the fol-
lowing true:
(11) Angel is identical to Lumber, but Angel might not have
been identical to Lumber.
None of this really commits us to the contingency of identity, how-
ever. Identity is a relation that holds necessarily (and perma-
nently). By this we mean that there are no things a and b which are
identical in one possible world but not identical in some other pos-
sible world. (Similarly, there are no things a and b which are identi-
14 | Delia Graff Fara
cal at one time but not identical at some different time.) Sort-relative
sameness, in contrast, is not a relation that holds necessarily (or
permanently). Sort-relative sameness does not coincide with iden-
tity, but it, rather than identity, is the important relation for analy-
ses of de re modal (and temporal) claims—even when these are
modal (or temporal) identity claims. That is why the truth of par-
ticular modal statements of contingent identity does not commit us
to the contingency of identity.8 This thought will be developed
and sharpened in what follows.
8
This interpretation of what acceptance of contingent identity amounts to can be
attributed to Gibbard (1975).
Possibility Relative to a Sortal | 15
In the actual world, be is the boat that Michael embarked upon and
it is constituted by some wood d. Where we have the boat and the
wood that constitutes it we have just one thing, so in α, be = d. In β,
br is the boat that Michael rows ashore. It is not constituted by the
wood d. In β, d does not constitute any boat at all. So in β, d ≠ br. In
the actual world, α, Michael does not row the boat ashore, since the
boat he embarked on capsizes. Nevertheless, Michael might have
rowed the boat ashore. In the part of modal space that I’m describ-
ing, that is because be , the boat embarked-upon in α, and br, the boat
rowed ashore in β, are the same boat. We can see that be and br are
nevertheless not identical, since be = d but d ≠ br.
Now is anything wrong with this picture? We can see that it allows
for transworld identities, which were proscribed by Lewis’s P2, since d
exists in the actual world as well as in the counterfactual world β. But
does be , the embarked-upon boat, exist in the counterfactual world β as
well in the actual world α? Since it is identical with d it has to exist in β,
since d exists there. And so it does. Identity is a necessary relation. So
since the wood d, which in fact constitutes the embarked upon boat be ,
is not identical to the boat rowed ashore in β, be ≠ br. Although be and br
are the same boat, they are not identical—not even in β.
In view of the above, it might seem that we have incurred the fol-
lowing commitments, which cannot all be right:
(12) be = d.
(13) be is the same boat as br in β.
(14) d is not a boat in β and therefore not the same boat as
anything in β.
These cannot all be right since it follows from these that be is the
same boat as br in β but also not the same boat as anything in β. As
we sharpen the idea, we will see that we are not in fact committed
to this inconsistent triad.
9
Jubien makes roughly this same point (Jubien 2001). He, however, makes it as a
psychological point about how we are inclined to think about the cross-time same-
boat relation rather than as a metaphysical point about the cross-time same-boat rela-
tion itself.
18 | Delia Graff Fara
What those conditions are is not per se part of the theory presented
here.10 I am interested primarily in the formal properties of relative
sameness and in applying these to certain metaphysical views
about constitution that I happen to hold. I want to develop a rela-
tive-sameness counterpart theory whose formal properties allow
for the combination of the Identity Thesis, the Accidental Constitution
Thesis, and the Necessary Matter Thesis while providing a plausible
analysis of de re modal (and temporal) predication. Someone with
different metaphysical views about essence and accident—whether
that be as it relates to material constitution or to something else,
such as personal “identity”—might well find the formal analysis
here congenial for her purposes. That would not be to accept a dif-
ferent view but rather to combine the view here with a different
metaphysics.11
This is like saying that someone could accept Kripke’s identity
analysis of de re possibility but reject his view that sameness of ori-
gin is required for transworld identity. Similarly, it would be like
saying that someone could accept Lewis’s counterpart theory of de
re possibility but think that similarity of origin is more important
for transworld similarity than is similarity in any other respect.
10
Sider (1999, 289–90) makes essentially the same point in partial defense of
Jubien’s (1993) version of counterpart theory.
11
Chisholm offered us criteria for cross-time sameness of table (Chisholm 1976,
218–21). But he need not have. I’m pointing out that we can present a theory of the
kind presented here while having no obligation to analyze what’s required for cross-
time or transworld sameness for each sort of thing that there is.
Possibility Relative to a Sortal | 19
12
We understand ‘Φ(rFw,w’ (x))’ to be the result of replacing each occurrence of ‘x’
in ‘Φ’ with ‘rFw,w’ (x)’. We assume that Φ has no unnecessarily repeated uses of ‘x’.
24 | Delia Graff Fara
13
We understand ‘Φ(rFxw,w’(x),rFyw,w’ (y))’ to be the result of replacing each occur-
rence of ‘x’ in ‘Φ’ with ‘rFxw,w’ (x)’ and each occurrence of ‘y’ in ‘Φ’ with ‘rFyw,w’ (y)’.
We assume that Φ contains no unnecessarily repeated uses of ‘x’ or ‘y’.
26 | Delia Graff Fara
The idea here is that Fx might be the sortal boat while Fy is the sortal
wood. This application to two variables can be straightforwardly
extended to explicitly cover indefinitely many variables.
The matter of how to extend this sort of generalization to qualita-
tive-similarity counterpart theory is not quite so straightforward.
(If there is a world in which I and my mother both have multiple
personal counterparts, do we have our various counterparts inde-
pendently? Or do we have our counterparts only as a pair? When
Lewis originally presented his counterpart theory, he chose the
former.) But we will come to this soon enough.
The relevant instance of (RSC◊2) for us right now is this:
Instance: ‘◊ Angel ≠ Lumber’ is true just in case there is a
possible world w’ in which Angel has a boat-mate a
in w’, Lumber has a wood-mate l in w’, and ‘a ≠ l’ is
true at w’.
This condition is witnessed in any world in which Angel has a boat-
mate that is not made up of some wood-mate of Lumber.
tical objects are distinct, but rather because the context depend-
ency of modal predicates prevents co-referential names from being
intersubstitutable in modal contexts—no matter whether they
flank the identity symbol or not. Despite this, we can coherently
affirm the necessity of identity by giving it a nonmodal, exten-
sional expression.
(NI) x = y ⇒ there is no possible world in which x ≠ y.
While both of the counterpart theories we have discussed have the
advantage of preserving (it), (act), and (nmt) there are nonetheless
important differences between them. We will proceed, then, with
our defense of relative-sameness counterpart theory by explaining
its advantages over qualitative-similarity counterpart theory. We
will subsequently turn to a discussion of Leibniz’s Law and explain
why, although one version of Leibniz’s Law is not validated by our
theory, there is a superior version of Leibniz’s Law that is. The paper
concludes with a summary and in the end suggests why relative-
sameness counterpart theory provides for a more satisfactory
response than qualitative-similarity counterpart theory does to
Kripke’s famous Humphrey objection.
14
For simplicity of exposition, we are temporarily prescinding from the need for
distinct variables to have distinct sortal-relativizations.
28 | Delia Graff Fara
(19) ◽∃xRax.
Let us consider a concrete example.
(20) ◽ Adam is Abel’s father, therefore
(21) ◽ Adam is someone’s father.
Lewis analyses the first as true just in case every possible world is
one in which every counterpart <Adam’, Abel’> of <Adam, Abel> is
such that Adam’ is father of Abel’. This is to say that there is no pos-
sible world in which the pair of Adam and Abel have a counterpart
pair with the first member of the pair not being father of the second.
But there are possible worlds in which Adam has a counterpart but
Abel does not, because there are possible worlds in which Adam has
no children at all. But these possible worlds are effectively ignored
for the purposes of evaluating (20). These are possible worlds in
which the pair of Adam and Abel have no pair as counterpart. So
these are possible worlds in which it is vacuously true that every
counterpart of the pair of Adam and Abel have a first member that
is father of the second member. But for the purposes of evaluating
(21), we do not ignore such possible worlds. ‘Adam is someone’s
father’ is analyzed as true on Lewis’s analysis just in case every pos-
sible world is one in which all of Adam’s counterparts have some
child or other. This is false on Lewis’s analysis, since there are worlds
in which Adam has a counterpart who has no child at all.
But the inference should hold good, at least in the case of logi-
cally simple relations, as in (18)–(21) above.15 If it is necessary that a
bear a certain logically simple relation to b, then surely it is neces-
sary that a bear that relation to something or other. The failure of
this inference to hold good within similarity-based counterpart the-
ory stems at bottom from its failure to distinguish essence and
necessity. Since (21) is false and the inference from (20) to (21) is a
good one, (20) must be false. Here’s why it is false. There are possi-
15
Given that the atomic predication ‘F( f (x))’ is not true—discounting identity—
whenever ‘f (x)’ has no denotation, the negation ‘¬F( f (x))’ is true whenever f (x) has
no denotation. For this reason, the truth of ‘◽¬R(x, y)’ does not guarantee the truth of
‘◽∃y¬R(x, y)’. This is as it should be. It is necessary that I am not related to you, but it
does not follow that it is necessary that there is someone to whom I'm not related.
There are possible worlds in which I and my ancestors are the only humans that have
ever existed.
32 | Delia Graff Fara
ble worlds in which Abel was never born. So, contra (20), there are
possible worlds that don’t contain both counterparts of Adam and
Abel with the first being father of the second.
It is, though, essential to Abel that he have Adam as a father.
Nothing could be Abel if it didn’t have Adam as a father. For the
similarity-based counterpart theorist, this is equivalent to its being
necessary that Abel have Adam as a father. This in turn is equiva-
lent to its being essential to Adam that he be father of Abel. But
because there are possible worlds in which Adam has a counterpart
who has no children at all, it is not necessary that Adam be father to
some child or other.
We want to be able to make the distinctions—and we can—
between the pairs of sentences (22)–(23) through (26)–(27).
(22) ◽(Abel exists → Abel’s father is Adam), (true)
therefore
(23) ◽(Abel exists → ∃x Abel’s father is x). (true)
This is the sound argument from ‘Abel essentially has Adam as a
father’ to ‘Abel essentially has someone as a father’. (22) is true on
our theory because there’s no possible world in which there’s some-
one who is the same person as Abel but who does not have some-
one who’s the same person as Adam for a father. (23) is then
guaranteed to be true on our theory since it is then guaranteed that
there is no possible world in which there’s someone that is the same
person as Abel but who does not have (in that world) anyone for a
father.16
(24) ◽(Abel’s father is Adam), (false)
therefore,
(25) ◽(∃x Abel’s father is x). (false)
This is the valid but unsound argument from ‘It is necessary that
Abel’s father be Adam’ to ‘It is necessary that Abel have someone as
a father’. The latter sentence, (25), is false since there are possible
16
We have not given truth clauses for all of the logical symbols. The details are
slightly complex once spelled out fully. I refer readers to Fara (2008) for details.
Possibility Relative to a Sortal | 33
any sentence that results if ‘b’ is put in for ‘a’ in any or all occur-
rences of ‘a’; and (II) the property version.
a
Leibniz Law I (a = b → (Φ → Φb)),
Leibniz Law II ∀x∀y(x = y → ∀z(z is a property → (x has z →
y has z))).
These principles can be easily restated to cover relations of
indefinitely large adicity.
The first version of Leibniz’s Law is properly called “the principle
of Substitutivity of Identicals” rather than Leibniz’s Law. The
two may come apart since it may be that ‘a’ does not occur in a ref-
erential position in ‘Φa’. If we say “a is Φ,” we may not be attribut-
ing a genuine property to the thing named by ‘a’. Substitutivity of
Identicals simply is not a valid principle. A classic counterexample
is Quine’s: ‘Giorgione was so-called because of his size’. Whether a
sentence ‘____was so-called because of his size’ is true once we put
in a name for the blank depends not only on what the referent of
that name is like, but also on which name we have used to refer to
it. That relative-sameness counterpart theory invalidates the substi-
tution principle is therefore not per se a problem with it. Moreover,
we have an explanation for why it fails in cases of attributions using
modal predicates. Our case is ‘____might have been constituted by
different wood’. Whether this sentence is true when we put in
a name for the thing that is Michael’s boat—e.g., ‘Angel’ or
‘Lumber’—depends on which sortal is associated with the name
that is used to refer to the thing in question.
What does it mean to say that different names for a thing can be
associated with different sortals? Let me give an example. We have
a woman, Larissa Seiger, who is both a ballerina and an elementary-
school teacher. Sometimes people ask, “What is Mrs. Seiger?” In
those situations it is typically appropriate to answer by saying that
she’s a teacher (my former teacher, my daughter’s teacher, et cetera).
Sometimes people ask “What is Madame Larissa?” In those situa-
tions it is typically more appropriate to answer by saying that she’s
a ballerina (the principle ballerina in company X, Odette in Swan
Lake, et cetera). This is to say that different ones of her names are
associated with different sortals—teacher or ballerina.
So why is it that ‘Angel’ cannot always be substituted with ‘Lum-
ber’ in the modal context in question? Why does this failure not
Possibility Relative to a Sortal | 35
falsify the property version of Leibniz’s law? One could say, with
Allan Gibbard, that ‘____might have been constituted by differ-
ent wood’ does not express a property (Gibbard 1975, 201). Or one
could say instead, with Harold Noonan, that the predicate
‘____might have been constituted by different wood’ does express
a property, but it is context dependent. Which property it expresses
depends on which name is substituted for the blank (Noonan 1991,
188). I am inclined to side with Noonan, but his view would have to
be modified so as to account for substitution failures with variables
rather than with names. The satisfiability of the following illustrates
this point:
(30) There is a boat x and some wood y such that x = y and x,
but not y, is possibly not made up of z.
7. CONCLUSION
The leading idea here was that relative sameness does not require
identity. This was subsequently qualified by saying that transworld
relative sameness does not require transworld identity. Michael’s
boat actually had one third of its planks replaced by the time it was
rowed ashore, but if one half of its planks had been replaced instead,
it would still have been the same boat that was rowed ashore in that
event, even though the boat rowed ashore in the counterfactual
world is not (absolutely) identical to the boat rowed ashore in the
actual world. Nonetheless, intraworld sameness—equivalent to
what I called sameness simpliciter—does require identity. Whenever
x is the same boat as y, x and y are identical.
This was not to say that we never have identity between things in
different possible worlds. In other words, it was not to say that indi-
viduals are “world bound.” Much less was it to deny that there is
such a relation as transworld identity. In that sense, I rejected radical
modal realism since I think that things in one possible world typi-
cally are identical to something in some other possible world. The
very thing that is actually Michael’s boat, and also some wood,
exists in many possible worlds in which it is not a boat at all.
Given the distinction between transworld sameness and trans-
world identity, the following analyses of de re possibility come
apart.
36 | Delia Graff Fara
the same horse as y than z is. Furthermore, sameness relations are not
purely qualitative. The boat that Michael rows ashore in the actual
world might be the same boat as the one he rows ashore in some
counterfactual world even if in that counterfactual world the boat
that Gabriel rows ashore is qualitatively just like the one that Michael
actually rows ashore—and even if Michael is, in that counterfactual
world, qualitatively just like Gabriel is in the actual one.
But like Lewis, I am able to maintain the sensible view that
Michael’s boat and the matter that composes it are one, even though
they might not have been, since the boat might have been com-
posed of different matter, while the matter could not have been
composed of different matter. In extensional possible-worlds talk in
terms of relativized counterpart relations, Michael’s boat and the
matter it is composed of are identical but they may have something
in another world as a boat-counterpart (a “boat-mate”) without
having that thing as a matter-counterpart (a “matter-mate”).
We saw that a counterpart theorist gets into trouble if his counter-
part relations allow for more than one counterpart in some one other
possible world. This rendered relative-sameness counterpart theory
far more attractive than superlative-qualitative-similarity counter-
part theory. Unlike superlative qualitative similarity, sameness is a
weak equivalence relation. Moreover, it is functional, albeit partial.
Making explicit the relativization to sortal F and double world rela-
tivizations to w and w’, we expressed this latter fact by saying that
x has at most one Fw,w’-mate. This allowed us to sharpen up (B)
above as follows.
RSC◊: When relativized to a sortal F, ‘◊Φ(x)’ is true at a
world w just in case there is a possible world w’ in
which ‘Φ(x’s Fw, w′-mate)’ is true.
Treating the functional expression on the right-hand side of this
truth clause as a singular term gave us this statement of truth condi-
tions for claims of de re necessity.
RSC□: When relativized to a sortal F, ‘□Φ(x)’ is true at a
world w just in case every possible world w’ is one in
which ‘Φ(x’s Fw,w’-mate)’ is true.
A counterpart theorist without a functional counterpart relation is
not entitled to the above analyses of de re modal claims. Instead,
38 | Delia Graff Fara
17
The qualitative-similarity counterpart theorist who gives up world-boundedness
needs to relativize his counterpart relation to pairs of worlds, just as we have done.
18
Murali Ramachandran (1998) has formalized Lewis’s version of sortal-relative
counterpart theory by devising a “translation scheme” of the kind we find in Lewis
(1968).
19
On this see Hazen (1979, 330) and also Michael Fara and Timothy Williamson
(Fara and Williamson 2005).
Possibility Relative to a Sortal | 39
REFERENCES
Alston, William and Bennett, Jonathan (1984), ‘Identity and Cardinality:
Geach and Frege’, Philosophical Review 93(4): 553–68.
Butler, Joseph, D.C.L. (1844), The Analogy of Religion, Samuel Hallifax, edi-
tor, Robert Carter & Brothers, New York.
Chisholm, Roderick M. (1973), ‘Parts as Essential to their Wholes’, Review
of Metaphysics 26: 581–603.
—— (1976), Person and Object, Allen and Unwin, London.
Fara, Delia Graff (2008), ‘Relative-Sameness Counterpart Theory’, Review
of Symbolic Logic 1: 167–89.
Fara, Michael and Williamson, Timothy (2005), ‘Counterparts and Actu-
ality’, Mind 114: 1–30.
Geach, Peter (1962), Reference and Generality, Cornell University Press, Ithaca.
——, ‘Identity’, Review of Metaphysics 21: 3–12. Reprinted in Geach 1972,
pp. 238–47.
——, Peter (1972), Logic Matters, Blackwell, Oxford.
Gibbard, Allan (1975), ‘Contingent Identity’, Journal of Philosophical Logic
4: 187–221.
40 | Delia Graff Fara
Gupta, Anil (1980), The Logic of Common Nouns, Yale University Press,
New Haven.
Hawley, Katherine (2001), How Things Persist, Clarendon, Oxford.
Hazen, Allen (1979), ‘Counterpart-Theoretic Semantics for Modal Logic’,
Journal of Philosophy 76(6): 319–38.
Johnston, Mark (1992), ‘Constitution is Not Identity’, Mind 101(401): 89–106.
Jubien, Michael (1993), Ontology, Modality, and the Fallacy of Reference,
Cambridge University Press, Cambridge.
—— (2001), ‘Thinking About Things’, Philosophical Perspectives 15: Meta-
physics pp. 1–15.
Kripke, Saul (1963), ‘Semantical Considerations on Modal Logic’, in Pro-
ceedings of a Colloquium on Modal and Many-Valued Logics, Acta
Philosophica Fennica, Helsinki, pp. 83–94.
—— (1972), ‘Naming and Necessity’, in D. Davidson and G. Harman,
eds., Semantics of Natural Language, Reidel, Dordrecht, pp. 253–355.
Reprinted as Kripke (1980).
—— (1980), Naming and Necessity, Harvard University Press, Cambridge,
MA. Reprinted, with an added preface, from Kripke (1972).
Lewis, David (1968), ‘Counterpart Theory and Quantified Modal Logic’,
Journal of Philosophy 65(5): 113–26.
—— (1971), ‘Counterparts of Persons and Their Bodies’, Journal of Philoso-
phy 68(7): 203–11.
—— (1986), On the Plurality of Worlds, Blackwell.
Noonan, Harold (1991), ‘Indeterminate Identity, Contingent Identity and
Abelardian Predicates’, The Philosophical Quarterly 41(163): 183–193.
—— (1993), ‘Constitution is Identity’, Mind 102: 133–46.
Perry, John (1970), ‘The Same F’, Philosophical Review 79: 181–200.
Ramachandran, Murali (1998), ‘Sortal Modal Logic and Counterpart The-
ory’, Australasian Journal of Philosophy 76: 553–65.
Sider Theodore (1996), ‘All the World’s a Stage’, Australasian Journal of
Philosophy 74(3): 433–53.
—— (1999), ‘Critical Study of Michael Jubien’s Ontology, Modality and the
Fallacy of Reference’, NOÛS 33: 284–94.
——(2001), Four-Dimensionalism: An Ontology of Persistence and Time,
Clarenden Press, Oxford.
Stalnaker, Robert (1986), ‘Counterparts and Identity’, Midwest Studies in
Philosophy 11: 121–40. Page references are to reprinted version (Stal-
naker 2003).
—— (2003), Ways a World Might Be, Oxford University Press, Oxford.
2. Reflections on Counterpart Theory
Allen Hazen
II
many similar objects, with the result that anything resembling one
of them in some specified way would resemble many. And, indeed,
we want to allow an object to be transworld identical to an object
that closely resembles some other object in its world: I might have
had an identical twin is intuitively plausible.)
The logical point can be made without assuming “world-bound”
individuals. Suppose, for the sake of argument, that individuals
literally exist in multiple worlds. We can still define a domain of
world-bound pseudo-individuals as, say, ordered pairs of a world
and an individual existing at it. Identity of individuals induces a
pseudo-identity relation over these pseudo-individuals, and this
relation is an equivalence with no cell containing two distinct
ordered pairs with the same world as a component. We can also
define pseudo-similarity relations over them: relations hold between
pairs <w,i> and <v,j> according to the manner and degree of simi-
larity and difference between what i is like at w and what j is like at
v. (Note that whether a pseudo-similarity holds between <w,i> and
<v,j> is independent of whether i=j: our ability to say that someone
“has changed,” or that he “looks just like his father at that age,”
depends in effect on our understanding of the analogous intertem-
poral similarity relations as independent of the identity or distinct-
ness of the terms.) Stated in this framework, the problem is that no
relation between world-individual pairs defined in terms of pseudo-
similarity relations will coincide with the relation induced by iden-
tity of individuals. All this should be acceptable even by the
staunchest defender of transworld identity. For Lewis it was a start-
ing point, providing (what must have been for him and certainly is
for me conclusive) grounds for scepticism about literal transworld
identity.
So Lewis went ersatz. Instead of postulating the sort of multi-
verse our intuitive modal thinking seems to be about—one in which
the same object often inhabits, and has different characteristics in,
multiple worlds—he postulated a replacement multiverse in which
no object exists in more than one world. (Those of us who prefer to
think of worlds as set-theoretic entities can enforce the disjointness
of worlds by stipulation: let non-actual worlds be models of the sort
usual in model theory, but take the “objects” existing in them to be,
not the entities in the domains of the models, but ordered pairs of
these with the models themselves. Assorted ordinal numbers, for
48 | Allen Hazen
counterpart in that world who is not the daughter of all the counter-
parts of George in that world. One can tinker with the truth condi-
tions (substituting, say, existential for universal quantifiers, etc.),
but nothing seems to work in general.
What is needed, then, is something which yields what Fara, in
her paper, has assumed: some way of picking out one of the per-
haps many counterparts an object has in some world, and allowing
the chosen counterpart, but not the others, to figure in our semantic
account of de re modalities.
III
IV
then necessarily, if the first and third both exist, it relates the first
to the third: crossworld transitivity, to give it a name) but which
was not derivative from another relation to further objects. I am
not at all confident that there really are such relations (though, if
your biological species is part of your essence, being conspecific
with might be one), but I also don’t feel that I could, with any con-
fidence, declare that there are none. So it seems to me that the
modal semantics of (1) is, at least, exposed to a counterexample of
the structure described here.
Since one of my main criticisms, in both (1) and (5), of Lewis’s
original semantics (and thus one of the main motivations for my
complicated modification of it) was that it didn’t handle examples
involving essential relations, it is acutely embarrassing to me that
my own semantics may have problems with more complicated
examples involving essential relations.
VI
VII
REFERENCES
Austin, J.L., How to Do Things with Words, (Oxford: Oxford University
Press, 1960).
Cohen, P.J., Set Theory and the Continuum Hypothesis, (New York: W.A. Ben-
jamin, 1966).
Gupta, A.K., The Logic of Common Nouns: an investigation of quantified modal
logic (New Haven: Yale University Press, 1980).
Hazen, A.P., “Counterpart theoretic semantics for modal logic,” Journal of
Philosophy 76 (1979), pp. 319–38.
Kripke, S.A., “Naming and necessity,” in G. Harman and D. Davidson,
eds., Semantics of Natural Language (Dordrecht: Reidel, 1972); reprinted
with new preface as S.A. Kripke, Naming and Necessity (Cambridge:
Harvard University press, 1980).
Lewis, D.K, “Counterparts of persons and their bodies,” Journal of Phi-
losophy 68 (1971), pp. 203–11; repr. In D.K. Lewis, Philosophical Papers,
vol. I, (New York: Oxford University Press, 1983).
—— On the Plurality of Worlds (Oxford: Basil Blackwell, 1986).
Parsons, C.,”Sets and classes,” Nous 8 (1974), pp. 1–12; repr. in his Math-
ematics in Philosophy (Ithaca: Cornell University Press, 1983).
Stalnaker, R., “The interaction of modality with quantification and iden-
tity,” in W. Sinnott-Armstrong et al., eds., Modality, Morality and Belief:
essays in honor of Ruth Barcan Marcus (Cambridge: Cambridge Univer-
sity Press, 1995), pp. 12–28.
This page intentionally left blank
II
ABSOLUTE GENERALITY
This page intentionally left blank
3. All Things Must Pass Away
Joshua Spencer
0. INTRODUCTION
The notion of all things must pass away from philosophical theori-
zing. By this, I do not mean that we must reject the set of all things.
Rather, we must reject the notion that some things are such that any
things are amongst them. This is a claim made in a purely plural
language. It may be formulated as follows:
(AT) (∃xs)(∀ys)(ys are amongst xs)
Unfortunately, (AT) is false. The argument against (AT) is quite sim-
ple. It involves three premises:
(1) There are two or more things.
(2) For any things, there is a unique thing that corresponds to
those things.
(3) For any two or more things, there are fewer of them than
there are pluralities of them.
Given (3), if there are some things that are all things and there are
two or more things, then there are fewer things altogether than plu-
ralities of them. But, given (2), there is a unique thing that corre-
sponds to each plurality of things. So, there are at least as many
things as there are pluralities of things. So, either there are no things
that are all things or there are fewer than two things. But, according
to (1), there are two or more things. It follows that there are no
things that are all things. That is (AT) is false.
Before I defend the argument above, I would like to indicate just
a bit of what is and is not at stake in a denial of (AT). Many people
might worry, for example, that one consequence of my thesis is
that unrestricted singular quantification is impossible. In section 1,
I indicate that this worry is, to some extent, legitimate. Once we
acknowledge that there are no things that are all things, we must
68 | Joshua Spencer
1
This thought may simply be an instance of what Uzquiano calls The All-in-Many
Principle. According to the All-in-Many Principle, “quantification over objects satis-
fying a certain condition presupposes that there are some objects which are all and
only those objects that satisfy the condition” (2009, 312).
All Things Must Pass Away | 69
text in which that same statement is false. One might say that the
domain of any putatively unrestricted singular quantifier is always
extensible.2
On the other hand, consider someone who denies (AT). That is,
consider someone who accepts the following:
(~AT) ~(∃xs)(∀ys)(ys are amongst xs)
This person might consistently hold that there is a sentence of the
form (US) which expresses a truth in every context. Here is an
example to show that this is consistent. Suppose we restrict our
attention to things that are finite in number and each one of which
is a positive integer. Now, it is clear that the following three claims
are consistent:
(i) There are no integers that are finite in number and are all
integers.
(ii) For every integer, either it is identical to 1 or identical to 2
or . . .
(iii) There is no context in which the quantifier of (ii) is
expanded to make (ii) false.
Notice that (i) involves the plural quantifier ‘there are no integers’
and expresses a restricted variant of (~AT). Thus, under an appro-
priate interpretation of our plural language, (~AT) expresses (i).
Moreover, (ii) involves the singular quantifier ‘for every integer’
and expresses a restricted variant of (US). Thus, under an appropri-
ate interpretation of our plural language, (US) expresses (ii). Finally,
(iii) is just a denial of the indefinite extensibility of the singular
quantifiers in (ii). But, since (i)–(ii) are consistent and are appropri-
ate restricted interpretations of (~AT) and (US), and since (iii) is a
2
This, though, may be mistaken. On one view, there are certain unrestricted quan-
tifiers that fail to have a domain. Consider someone who believes that domains are
sets of things and yet still believes that unrestricted singular quantification is possi-
ble. On this view, when an unrestricted singular quantifier is employed, it has no
domain. This is because domains are sets and there is no set of all things. Moreover,
since there is no set of all things, domains are indefinitely extensible even though the
unrestricted quantifier is not. So, on this view, the claim that unrestricted quantifiers
are indefinitely extensible comes apart from the view that domains are indefinitely
extensible. The same may be true if domains are pluralities rather than sets and there
are no things that are all things.
70 | Joshua Spencer
3
Here is another way to make the same point. The way that I understand a plural-
ity, a single thing is merely a very sparse plurality. Given this fact, we can simply
introduce singular quantification as a restriction on plural quantification. Singular
quantification is merely plural quantification restricted to those things that are one in
number (McKay 2006). But the argument only shows that there are no things that are
all things. As long as there is more than one thing, the argument doesn’t show any-
thing about things that are one in number. Moreover, since the argument relied on the
premise that there are two or more things, it doesn’t show anything about things that
are one in number even if there is only one thing.
4
For more details see Rayo and Uzquiano (1999) and Uzquiano (2009). Also see
Williamson (2009) for related problems for unrestricted quantification. The limita-
tions placed on plural languages by a denial of (AT) may impede certain attempts to
solve the problems posed in Williamson (2009) using plurals.
All Things Must Pass Away | 71
5
Jonathan Schaffer (2010), for instance, says “Classical mereology—with its axiom
of unrestricted composition—guarantees the existence of a unique fusion of all con-
crete objects. Thus there are gunky models of classical mereology, but no junky mod-
els. Indeed, a mereologically maximal element is the only individual that classical
mereology guarantees on every model.” Einar Duenger Bohn, in both (2009a) and
(2009b), has made similar remarks.
6
Some people might reject this premise if they think it’s possible for something to
be shrunk down and sent back in time to become a part of itself. Although I am a fan
of the possibility of such science fiction examples, I will not be discussing the implica-
tions of such fantastical possibilities here.
72 | Joshua Spencer
7
This principle is entailed by Classical Extensional Mereology.
8
Gustave Flaubert (or so I've heard).
All Things Must Pass Away | 73
9
Literally speaking, something is a cardinal number only if it numbers the mem-
bers of some set. I am going to use the notion of a cardinal number in an extended
sense. On the notion I will be employing, something is a cardinal number if it num-
bers some things, whether or not they form a set.
10
Hazen (1993) introduced a language of perplurals. Hazen, however, tried to use
the sensibility of such a language to argue against plural quantification. However, I
think we can understand the language and need not accept Hazen’s rejection of plu-
ral quantification. Perhaps, we can understand the language through a convenient
fiction according to which pluralities are real entities over which perplural quantifi-
ers range. The fictional interpretation of the language need not lead to inconsistencies
as long as we stipulate that the plural quantifiers of such a language range only over
the non-fictional entities of the universe. The plural quantifiers will have their ordi-
nary meaning whereas the perplurals will have a fictional meaning that merely helps
us in our counting.
11
Something like it follows from a plurals version of Cantor’s theorem. Stewart
Shapiro (1991) has a proof of a second-order version of Cantor’s theorem and George
Boolos (1984) has famously suggested that we can reinterpret second-order state-
ments in a language of plurals. If Boolos’s method of doing so is adequate, then Sha-
piro’s proof and statement of Cantor’s theorem should be reinterpretable in a
language of plurals. Something like the cardinality claim immediately follows. In the
remainder of this section, I present a Cantorian argument for the cardinality princi-
ple. See also Rayo (2002).
74 | Joshua Spencer
12
Thanks to Karen Bennett for this argument.
All Things Must Pass Away | 75
13
The argument of this section is similar to, but distinct from, arguments I presented
in my (2006). However, I am indebted to Greg Fowler for many helpful discussions
about some puzzling aspects of my previous argument. These discussions helped me to
see some of my underlying assumptions and develop the argument that appears here.
76 | Joshua Spencer
14
Many of these worlds will be impossible worlds since many things cannot exist
without other things also existing. For example, I cannot exist without the number
two also existing.
15
One more view that commits us to two or more things and a unique thing associ-
ated with any things is a robust view of properties. According to a robust view of
properties, there are many properties and for any things, there is a property that they
and only they have in common.
16
Moreover, there is also at least one more proposition. Not every proposition has
its truth value necessarily. Some propositions are contingent. But, since each of the
two propositions above have their truth values necessarily, it follows that there is at
least one more proposition. So, on the view that there are propositions, there are at
least three things.
All Things Must Pass Away | 77
Now, for any things, there is a proposition just about those things.
I think it is fairly clear what it means to say of some things that a
proposition is about them. The proposition that Nicholas and Brie
both exist is, for example, about Nicholas and Brie. Some philoso-
phers might use the word ‘about’ in an extended sense. They might
say that, in addition to being about Nicholas and Brie, the proposi-
tion that Nicholas and Brie both exist is also about the property of
existence. But, I do not intend to use the word in this extended sense.
The way I am using ‘about’, the proposition that Nicholas and Brie
both exist is about Nicholas and Brie and is not at all about the prop-
erty of existence. There are, of course, propositions that are about
properties (or at least I think there are). The proposition that exist-
ence is monadic is about the property existence. But, it should be
clear, now, that it is not also about the property of being monadic.
It is obvious that for any things there are some propositions about
them. After all, for any things, it is obvious that there are some truths
about them; that they exist, perhaps, or have some kind of being.
But, I want to make a slightly stronger claim. I believe that for any
things there is a proposition about them. Moreover, I believe that for
any things, there is a proposition just about them. That is to say that
for any things there is at least one proposition about those things and
there are no further things that that proposition is also about. We can
state this Propositions Thesis a bit more formally as follows:
(PT) for any xs, (i) there is a proposition about those xs and
(ii) for any ys which are not the same as the xs, there is a
different proposition about those ys.
If (PT) is true, then statement (2) is true as well.
One way to support (PT) is to indicate a certain type of proposition
that is such that for any things whatsoever, some proposition of that
type is just about those things. That is, indicate some propositions
that correspond, in the appropriate way, to all the various pluralities
in the world. It seems to me that existential propositions are good
candidates for the appropriate correspondence. If that is correct, then
the following Existential Propositions Thesis might be true.
(EPT) For any xs, (i) there is the proposition that those xs exist
and (ii) for any ys which are not the same as the xs, the
proposition that those ys exist is distinct from the prop-
osition that those xs exist.
78 | Joshua Spencer
17
Admittedly, there are some who deny the first tenet of (EPT). Some people think
that existence is not a first order property of either individuals or pluralities. Rather,
existence is a higher order property of properties. Thus, we cannot, for any individ-
ual x, say that x exists. Rather, we can only say, for some property F, that there are Fs.
Furthermore, some people might claim that for all we know, there may be some
things that share no properties. Thus, we cannot say of them that they exist because
we cannot say of some property they all share that some things have that property.
However, this worry should not be too worrisome since any things have the prop-
erty of being just those things. This is a property that they have and any other things
lack. So, even if existence is a second-order property, we can say of any things that
existence applies to the property of being just those things.
18
Even if existence is not a first-order property of plurals, surely the property of
composing something is such a property.
All Things Must Pass Away | 79
are not the same as the xs, the proposition that those
ys are thought about by someone is distinct from the
proposition that those xs are thought about by someone.
If (DPT) is true, then for any things there is a proposition, a doxastic
proposition, just about them.19 Hence, (PT) is true and so is (2).
So, if any of (EPT), (CPT), or (DPT) are true, then (PT) is true as
well. That is, each of the three theses above supports the claim that
for any things there is a proposition just about those things. Of
course, if (PT) is true, then so is (2). But, we need not appeal to the
existence of particular types of propositions to support (PT).
It is easy to show that for any things, there is a proposition just
about them; that is, it is easy to show that (PT) is true. Suppose, for
reductio, that there are some things such that there is no proposition
just about them. Let’s say that a1 . . . an are some such things. Then, it
is true that a1 . . . an are such that there is no proposition just about
them. But, since every truth is a proposition, it is clear that there is
at least one proposition about them, namely the proposition that
a1 . . . an are such that there is no proposition just about them. Moreo-
ver, this proposition is just about them.20 So, there is at least one
proposition just about them. But, this contradicts our claim that
19
Notice that the denial of (DPT) entails that there can be no omniscient being. For,
if there could be an omniscient being (even a non-actual one), then for any things
whatsoever, that being would believe that they possibly exist.
20
One might wonder whether this proposition really is just about a1 . . . an. I have to
admit that I have no criteria for determining whether or not a proposition is just
about some things. However, the following seems intuitively correct to me. P is about
some xs if those xs are some of the things that make P true and they are the only such
things that are such that P logically implies that they have some feature or other. For
example, Nicholas and the property of being tall both help to make the proposition
that Nicholas is tall true. But, the proposition that Nicholas is tall logically implies
that Nicholas has some feature or other. It does not logically imply that the property
of being tall has some feature or other. Moreover, there are no other things that help
to make that proposition true. So, the proposition is about Nicholas and not tallness.
Similarly, being blue and being a color are the only things that help to make it true
that blue is a color. However, the proposition that blue is a color logically implies that
blue has some feature or other but does not logically imply that being a color has
some feature or other. So, that proposition is about the property of being blue and not
about the property of being a color. Now consider the proposition that a1 . . . an are
such that there are no propositions about just them. It seems that a1 . . . an are things
that make that proposition true and that proposition logically implies that a1 . . . an
have some property or other (namely being such that there are no propositions about
them). Moreover, there are no other things that help to make that proposition true
which are also such that that proposition logically implies they have some property
or other.
80 | Joshua Spencer
21
A better view might be that propositional attitudes (and other properties and
relations that seem to take propositions as objects) are really irreducibly plural. The
fundamental belief relation, on this view, is irreducibly plural in its second place and,
hence, it will have the form ‘S believes those xs’ where the plural variable is satisfied
All Things Must Pass Away | 81
by worlds. Consider the sentence ‘grass is green’. On this view, that sentence picks
out, plurally, all those worlds where grass is green. Moreover, if someone believes the
content of that sentence, then she stands in the belief relation to those worlds. This
view might be favored by those who believe that some worlds are too numerous to
form a set. If propositions are sets, then there is no proposition that corresponds to
some worlds that are too numerous to form a set. However, if (speaking vulgarly)
propositions are pluralities, then those worlds that are too numerous to form a set are
still propositions and might still be expressed and believed.
One downside of this view is the following. There are some truths about proposi-
tions that seem to be irreducibly plural. For example, when I say that Nicholas’s
beliefs are consistent, I seem to be saying something irreducibly plural about propo-
sitions. However, if propositions are pluralities, then I must be saying something
irreducibly perplural. As I mentioned before, some people think that there’s no way
to make sense of a perplural language without reifying pluralities (which is exactly
what we want to avoid). The only hope for a defender of this view is to find a way to
paraphrase away those claims that seem to be irreducibly plural; like the claim that
Nicholas’s beliefs are consistent. Gabriel Uzquiano (2004) has a good discussion of
the prospects for paraphrasing seemingly irreducible plural claims in another con-
text. Much of what Uzquiano says will apply in this circumstance as well. Luckily,
nothing that I say in the remainder of this section depends on whether propositions
are sets of worlds or pluralities of worlds. So, I will simply focus on the former
view.
82 | Joshua Spencer
22
Strictly speaking this does not follow from the last two claims. On Lewis’s view,
a merely possible de re modal truth about me may be true because there is some other
actual individual who is my counterpart. However, it is necessary that there are
merely possible de dicto truths. The claim that there is a world distinct from the actual
world follows from the claim that there are merely possible de dicto truths and the
claim that necessarily, for any proposition, P, P is possibly true just in case there is a
possible world, w, and P is true-at-w.
All Things Must Pass Away | 83
order to undermine all hope of support for premise (2), one must
accept a generally sparse and rather radical metaphysics.23
23
To make the darkness of this path more acute, note that one who accepts that
propositions are merely sets of possible worlds should say exactly what a possible
world is. But, the standard views of possible worlds cannot be held given a sparse
metaphysical view. We cannot, for example, say that a possible world is a complex
state of affairs or a complex property or even a complex sentence in a lagadonian
language (where everything is its own name and every property is a predicate that
expresses itself). For if any of these resources are rich enough to use in our construc-
tion of possible worlds, then they are also rich enough to generate support for premise
(2). I see three options available to the defender of a sparse metaphysics. First, one
could, of course, simply endorse Lewis’s unorthodox view according to which pos-
sible worlds are concrete things composed of individuals which are appropriately
related. Second, one could accept Magical Ersatzism according to which there is no
true account of how possible worlds represent ways the world could be (i.e. they just
do). Finally, one could accept a poor world-making language and allow most repre-
sentation to be implicit. However, I think few of us are willing to accept the counter-
intuitive costs of Lewis’s concrete modal realism or Magical Ersatzism (though see
van Inwagen (1986) for a defense of the latter). So, the third option seems most
plausible.
24
A version of this argument is presented by Rayo and McGee (2000).
All Things Must Pass Away | 85
since (PT) was our primary support for premise (2), it looks like our
support for (2) has been undermined.
This is a very powerful argument against (PT). However, I do not
believe the argument is sound. I do not accept (F).25 This may seem
shocking, given that (F) seems like a logical truth. But, it turns out
that propositions of the form P or not-P are not logical truths. This is
one lesson we should take away from the Liar Paradox. Consider
the English sentence, hereby named ‘(LIAR)’: “(LIAR) does not
express a truth”. If we accept that either (LIAR) expresses a truth or
it does not, then contradiction immediately follows. So, we must
not accept that either (LIAR) expresses a truth or it does not. But,
that means that there is at least one proposition of the form p or not-
p which is unacceptable and hence, not a logical truth.
Now, I said that I do not accept (F). But, I do not accept that Con-
fusion is neither about itself nor not about itself either. To move
from not accepting (F) to accepting the negation of (F) is to make the
same mistake as before. Such a move presupposes that (F) or not-(F)
is true. Rather, we must remain silent about whether (F) is true and
we must remain silent about whether Confusion is about itself.26
It turns out that ‘is not about itself’ and ‘does not express a truth’
express partial properties. Let’s call those things that are instances of
a property the ‘metaphysical extension’ of that property.27 Partial
properties are properties that have a definite metaphysical extension
and a definite metaphysical anti-extension. An atomic proposition
of the form a is F is true if a lies within the metaphysical extension of
the property F. The negation of an atomic proposition of the form a
is F is true if a lies within the metaphysical anti-extension of the
property F. However, there are some things that lie outside both the
metaphysical extension and the metaphysical anti-extension of par-
tial properties. If a lies outside both the metaphysical extension and
25
The view that I present is very similar to the view presented by Field (2008) and
is inspired by Soames’s (1999) view of truth.
26
If we were to model a language that behaves this way, then we would assign to
each predicate an extension and an anti-extension. However, some predicates, such
as ‘is not true’ and ‘is not about itself’, will be such that the union of their extension
and anti-extension fails to include the entire universe.
27
Here I am following Salmon (1981 pp. 46) in distinguishing a metaphysical
extension from a semantic extension. Properties have metaphysical extensions
whereas predicates have semantic extensions.
All Things Must Pass Away | 87
Some may have noticed that the argument presented in the last sec-
tion against (PT) is similar in structure to one part of the argument
28
Moreover, given the truth conditions for negation, (LIAR) lies outside of the
metaphysical extension and anti-extension of the property expressed by ‘is true’ as
well. This means that the proposition, hereby named (TRUTH), that this proposition
is true is also outside the metaphysical extension and anti-extension of the property
expressed by ‘is true’. Hence, we must remain silent about whether the truth teller is
true. See Soames (1999) for a discussion of this consequence.
88 | Joshua Spencer
Eve is like Mark Barber. Eve’s existence is not foisted upon us and
we are free to avoid contradiction by denying that Eve exists. This
route leads us to the cardinality principle. Confusion, on the other
hand, is like (LIAR). We know that Confusion exists because we
know that there are some humble propositions. Moreover, I am con-
vinced by the discussion of section 2 that there must be a proposi-
tion just about the humble propositions. The existence of Confusion
forces us into silence. That is, we are forced to remain silent about
whether Confusion is about itself.
There is also an independent reason for avoiding silence about
the claim that either Eve is amongst The Fallen or Eve is not. Sup-
pose the claim that either Eve is amongst The Fallen or Eve is not
amongst The Fallen is unacceptable. That is, suppose that we must
remain silent about that disjunction. If that is the case, then we must
also remain silent about each of the disjuncts. So, the claim that Eve
is amongst The Fallen is a claim that we must remain silent about.
But, we know that the amongst relation is transitive. But, that means
that if Eve is amongst any things that are amongst The Fallen, then
Eve is amongst The Fallen as well. So, if we must remain silent
about whether Eve is amongst The Fallen, then we must remain
silent about whether Eve is amongst any things that are amongst
The Fallen. Some things that are amongst The Fallen are individuals
and individuals are simply pluralities that are one in number. So,
for any things that are one in number and amongst The Fallen, we
must remain silent about whether Eve is amongst those things. One
interesting discovery of the logic of plurals is that there is a relation-
ship between the amongst relation and the identity relation. In par-
ticular, the following Identity Principle is true:
(IP) (∀xs)(∀ys) (xs=ys ↔ (xs are amongst ys & ys are amongst
xs & xs are one in number))
That is, one plurality is identical to a second plurality iff each is
amongst the other and neither is more than one in number.29 But,
supposedly we must remain silent about whether Eve is amongst
any things that are amongst The Fallen (including individuals).
Moreover, Eve is definitely one in number. It follows from these two
claims that for any individual amongst the fallen, we must remain
29
See, for example, McKay (2006, 129).
90 | Joshua Spencer
8. CONCLUSION
Although the claim that there are some things that any things what-
soever are amongst seems intuitively plausible, I believe this thesis
must be rejected. Those who disagree must accept widespread met-
aphysical limitations, not only with respect to propositions, but also
with respect to states of affairs, properties, divine thoughts, and
impossible worlds (to name a few). Moreover, they must accept that
there are some things that are such that there are is no proposition
just about them and there is no state of affairs involving just them
and there is no property had by just them etc. On the other hand,
those who reject all things, must face certain limitations on the use-
fulness of plurals and perhaps accept certain surprising metaphysi-
cal theses. It seems to me that the costs of rejecting all things are less
drastic than the costs of sparse metaphysics.30
University of Wisconsin, Milwaukee
BIBLIOGRAPHY
Bohn, Einar Duenger (2009a) “An Argument Against the Necessity of
Unrestricted composition,” Analysis Vol. 69: 27–31.
—— (2009b) “Must there be a Top Level?” The Philosophical Quarterly
Vol. 59: 193–201.
Boolos, George (1984) “To Be is to be the Value of a Variable (or to be Some
Values of Some Variables),” The Journal of Philosophy Vol. 81: 430–49.
30
Thanks to Karen Bennett, Gregory Fowler, Mark Heller, Hud Hudson, Kris
McDaniel, Tom McKay, Chris Tillman, and Gabriel Uzquiano, for discussing these
issues and reading earlier drafts of this paper. Thanks also to audiences at The Uni-
versity of Manitoba and Syracuse University for helpful comments and discussion.
All Things Must Pass Away | 91
Years ago, when I was young and reckless, I believed that there was
such a thing as an all-inclusive domain.1 Now I have come to see the
error of my ways.
The source of my mistake was a view that might be labeled ‘Trac-
tarianism’. Tractarians believe that language is subject to a meta-
physical constraint. In order for an atomic sentence to be true, there
needs to be a certain kind of correspondence between the semantic
structure of the sentence and the ‘metaphysical structure’ of reality.
The purpose of this paper is to explain why I think Tractarianism is
mistaken, and what I think an anti-Tractarian should say about
absolutely general quantification.
1. THE PLAN
6. Tables
For there to be a table just is for there to be some things
arranged tablewise.
7. Dinosaurs
For the number of the dinosaurs to be Zero just is for there
to be no dinosaurs.
Statement 1 is utterly uncontroversial. Statement 2 should be pretty
uncontroversial too, at least if we ignore certain complications (such
as the possibility of impurities). Statement 3 is not totally uncontro-
versial (Chalmers (1996)), but it seems to be the dominant view
amongst philosophers.
Statements 4–7, on the other hand, are all highly controversial met-
aphysical theses. My own view is that they are all true, but I won’t try
to convince you of that here. The aim of this paper is to argue that
they shouldn’t be rejected on general linguistic or metaphysical
grounds. I will proceed by defending a conception of language I call
compositionalism, and showing that it makes room for Statements 4–7.
I will then argue that a compositionalist who accepts Statements 4–7
is left with an attractive metaphysical picture of the world.
The plan for the paper is as follows. I will start by explaining
how I think the ‘just is’-operator should be understood (section 2).
I will then introduce my foil: Tractarianism. I will explain why I
think Tractarianism is bad philosophy of language (section 3), and
develop compositionalism as an alternative to Tractarianism (sec-
tion 4). Attention will then turn to metaphysics. I will argue that
compositionalism does not lead to untoward metaphyical conse-
quences, even if one accepts ‘just is’ statements such as 4–7 (section
5). I will conclude by addressing the problem of absolute generality
from the perspective of the compositionalist (section 6).
case that she shares a parent, and if she shares a parent it is thereby
the case that she is a sibling. More colorfully: when God created the
world, and made it the case that Susan shared a parent, there was
nothing extra She had to do, or refrain from doing, in order to ensure
that Susan was a sibling. She was already done. And when God cre-
ated the world, and made it the case that Susan was a sibling, there
was nothing extra She had to do, or refrain from doing, in order to
ensure that Susan shared a parent. She was already done.
In the special case in which Susan is, in fact, a sibling, there is an
additional way of clarifying the meaning of Sibling. For Sibling to
be true is for ‘Susan is a sibling’ and ‘Susan shares a parent’ to be
full and accurate descriptions of the same feature of reality.
Other ‘just is’-statements should be understood in the same sort
of way. For Death to be true is for there to be no difference between
someone’s dying and a death’s taking place. When someone dies it
is thereby the case that a death takes place, and when a death takes
place it is thereby the case that someone dies. The feature of reality
that is fully and accurately described by saying ‘A death took place’
is also fully and accurately described by saying ‘Someone died’.
It is useful to compare Sibling and Death with an identity state-
ment such as ‘Hesperus is Phosphorus’. If you accept ‘Hesperus is
Phosphorus’, you believe that there is no difference between
traveling to Hesperus and traveling to Phosphorus. Someone who
travels to Hesperus has thereby traveled to Phosphorus, and some-
one who travels to Phosphorus has thereby travelled to Hesperus.
The feature of reality that is fully and accurately described by say-
ing ‘A Soviet spaceship traveled to Hesperus’ is also fully and accu-
rately described by saying ‘A Soviet spaceship traveled to
Phosphorus’.
Since ‘just is’-statements are treated as equivalent to the corre-
sponding ‘no difference’ statements, the ‘just is’-operator is treated
as symmetric. There is a different reading of ‘just is’ on which it fails
to be symmetric. One could suggest, for example, that a ‘just is’-
statement should only be counted as true if the right-hand-side
‘explains’ the right-hand-side, or if it is in some sense ‘more funda-
mental’. This is not the reading that will be relevant for present pur-
poses. If you find the asymmetric reading more natural than the
symmetric reading, please substitute a suitable ‘no difference’-
statement for each ‘just is’-statement in the text.
96 | Agustín Rayo
3. TRACTARIANISM
2
See Rayo (forthcoming) "Neo-Fregeanism reconsidered". For a different way of
thinking about ‘just is’-statements, see Bennett (2009).
3
For further discussion of Tractarian conceptions of language, see Heil (2003). For
criticism, see Eklund (2009).
Absolute Generality Reconsidered | 97
Metaphysical Structure
Tractarianism is a hybrid of linguistic and metaphysical theses. It
deploys a metaphysical assumption—the existence of metaphysi-
cal structure—to impose a constraint on linguistic theorizing.
I suspect that the notion of metaphysical structure is not in good
order, and I would like to make a few remarks about why I think
this is so.
Let me start by talking about objectivism. Most of us are objectiv-
ists about truth. We believe that it makes sense to speak of what is
objectively the case, as something over and above what is true
according to one person or another. Many philosophers, but not all,
are objectivists about morality. They believe that it makes sense to
speak of what is objectively good, as something over and above
what would be good with respect to some value system or other. Few
philosophers, if any, are objectivists about fashion. You may think
that ascots are fashionable. But it would be preposterous to suggest
that they are objectively fashionable: fashionable over and above the
tastes of some community or other.
Objectivism comes at a cost. An objectivist about fashion, for
example, would be faced with the awkward task of elucidating a
non-trivial connection between what is objectively fashionable and
what various communities take to be fashionable. She would also
have to choose between coming up with an explanation of what it
takes for something to be objectively fashionable and burdening
her picture of the world with the view that there are brute facts
about objective fashion. And the rewards for her efforts would be
Absolute Generality Reconsidered | 99
4
See Lewis (1983) and Lewis (1984).
5
See Fine (2001), Schaffer (2009), and Sider (typescript).
100 | Agustín Rayo
6
See, for instance, Schaffer (2009) and Sider (typescript). For a critique of certain
forms of metaphysical objectivism, see Hofweber (2009).
7
Here I have in mind a traditionalist interpretation of the Tractatus, as in Hacker
(1986) and Pears (1987). See, however, Goldfarb (1997).
Absolute Generality Reconsidered | 101
Moderate Tractarianism
There is a moderate form of Tractarianism according to which the
constraint that there be a correspondence between semantic struc-
ture and metaphysical structure applies only to assertions made by
philosophers in the ‘ontology room’. When I use the term ‘Tractari-
anism’ here, the view I have in mind is always non-moderate Trac-
tarianism. My arguments for the claim that Tractarianism is bad
philosophy of language do not apply to moderate Tractarianism.
For all I know, there is a special convention governing discourse
in the ontology room, which demands correspondence between
semantic and metaphysical structure. If you are sympathetic
towards moderate Tractarianism, that’s fine. Just make sure you
don’t interpret me as a moderate Tractarian.
A moderate Tractarian is free to accept a ‘just is’-statement such
as ‘for a death to take place just is for someone to die’. All she needs
to do is insist that at most one side of the ‘just is’-statement is taken
in an ontology-room spirit. To avoid confusion, moderate Tractari-
ans might consider introducing a syntactic marker for ontology-
room discourse, as in Fine (2001). They could then say:
What it really is for a death to take place is for someone to die
102 | Agustín Rayo
or:
What it is, in fundamental terms, for a death to take place is for
someone to die
to indicate that the feature of reality described by ‘a death takes
place’ gets carved by the world’s metaphysical structure in a way
that corresponds to the semantic structure of ‘someone dies’.
Just to be clear: this is not what I intend when I use ‘just is’-state-
ments here.
4. COMPOSITIONALISM
8
It is worth noting that the nominalizations of open formulas turn out to be open
formulas, and therefore lack truth-conditions. Fortunately, all that is required for
present purposes is an assignment of truth-conditions to well-formed sentences.
104 | Agustín Rayo
9
For closely related views, see Frege (1884), Wright (1983), and Rosen (1993). Dis-
cussion of compositionalism amongst contemporary metaphysicians in the United
States has tended to focus on the work of Eli Hirsch, who draws on earlier work by
Hilary Putnam. (See, for instance, Putnam (1987) and Hirsch (2002); for criticism, see
Eklund (2008) and Bennett (2009).) It is worth keeping in mind, however, that some
of Hirsch’s linguistic theses go significantly beyond anything defended here, and
that his general attitude towards metaphysics is profoundly different from my own
(see, in particular, the discussion in section 4.1 of the present text).
Absolute Generality Reconsidered | 105
10
It is also worth noting that compositionalism is not in tension with the view—
first suggested in Lewis (1983) and Lewis (1984)—that problems of referential inde-
terminacy can sometimes be resolved by attending to metaphysical naturalness.
Compositionalism is a view about what it takes for a singular term to be in good
order, not about the sorts of considerations that might be relevant to fixing the refer-
ence of singular terms. There is room for thinking that Lewis himself was a composi-
tionalist: see Lewis (1980).
106 | Agustín Rayo
11
This is a badly inaccurate statement of the thermodynamic theory of heat. Fortu-
nately, the inaccuracies are harmless in the present context.
Absolute Generality Reconsidered | 107
12
For a more detailed discussion of these matters, see Rayo (forthcoming) "The
Contruction of Logical Space"
13
Here I am indebted to Andrew Graham’s PhD thesis.
108 | Agustín Rayo
14
See Jackson (1982) and Jackson (1986); for a review of more recent literature, see
Byrne (2006).
15
See Rayo (forthcoming) “Neo-Fregeanism reconsidered”, ch. 4.
Absolute Generality Reconsidered | 109
for there to be some things arranged tablewise, one eliminates the need
to address an awkward question: what would it take for a region that
is occupied by some things arranged tablewise to also be occupied by
a table? It is true that one also loses access to a certain amount of theo-
retical space, since one is no longer in a position to work with scenarios
in which there are things arranged tablewise but no tables. It seems to
me, however, that this is not much of a price to pay, since the availabil-
ity of such scenarios is not very likely to lead to fruitful theorizing. (Not
everyone would agree; see, for instance, van Inwagen (1990).)
For similar reasons, it seems to me that Properties, Death, and
Dinosaurs are all eminently sensible ‘just is’-statements. Again,
not everyone will agree. But I hope to have convinced you that
these statements shouldn’t be rejected merely on the basis of syntac-
tic considerations. They should be rejected only if one thinks that
the resulting theoretical space leads to theorizing that is fruitful
enough to pay the price of answering awkward questions.
And the relevant questions can be very awkward indeed. By
rejecting Dinosaurs, for example, one is forced to concede that the
following is a legitimate line of inquiry:
I can see that there are no dinosaurs. What I want to know is whether it is
also true that the number of the dinosaurs is Zero. And I would like to
understand how one could ever be justified in taking a stand on the issue,
given that we have no causal access to the purported realm of abstract
objects.16
If, on the other hand, you accept Dinosaurs you will think that
such queries rest on a false presupposition. They presuppose that
there is a gap between the non-existence of dinosaurs and dino-
saurs’ having Zero as a number—a gap that needs to be plugged
with a philosophical account of mathematical objects. Dinosaurs
entails that the gap is illusory. There is no need to explain how the
non-existence of dinosaurs might be correlated with dinosaurs’
having Zero as a number because there is no difference between the
two: for the number of the dinosaurs to be Zero just is for there to
be no dinosaurs.17
16
That this is a legitimate line of inquiry is famously presupposed by Benacerraf
(1973).
17
For an account of mathematics along these lines, see Rayo (2009) and Rayo
(forthcoming) The Construction of Logical Space.
Absolute Generality Reconsidered | 111
5. LIFE AS AN ANTI-TRACTARIAN
18
I address the second challenge in Rayo (forthcoming) The Construction of Logical
Space, where I develop a semantics for mathematical discourse and an account of
mathematical knowledge.
19
Tractarianism and anti-Tractarianism are incompatible with each other, but they
are not contradictories. A compositionalist who accepts no metaphysically conten-
tious ‘just is’-statements would reject them both. So would someone whose views of
reference fall somewhere between Tractarianism and compositionalism.
112 | Agustín Rayo
For the direction of line a to equal the direction of line b just is for a and b to
be parallel.
Realism
A Tractarian might be tempted to complain that if anti-Tractarian-
ism were correct, there would fail to be a definite fact of the matter
about how the world is. I have sometimes heard arguments such as
the following:
Say you believe that for the number of the dinosaurs to be Zero just is for
there to be no dinosaurs. You believe, in other words, that a single fact can
be described fully and accurately by asserting ‘the number of the dinosaurs
is Zero’ and by asserting ‘there are no dinosaurs’. This presupposes that a
single fact can get carved up into objects and properties in different ways.
When the fact is described by asserting ‘the number of the dinosaurs is
Zero’, it gets carved up into an individual (the number Zero), a first-order
property (the property of being a dinosaur), and a second-order function
(the function taking first-order properties to their numbers); when it is
described as ‘there are no dinosaurs’, it gets carved out into a first-order
property (the property of being a dinosaur) and a second-order property
(non-existence).
But if this is so, there can’t be an objective, language-independent fact of
the matter about whether there are numbers. It all depends on how we
choose to describe the world.
20
Hirsch is a compositionalist (see Hirsch (2002)). But it is not clear to me that he
is also an anti-Tractarian.
Absolute Generality Reconsidered | 113
Comprehensivism
Comprehensivism is that view that it is in principle possible to give a
comprehensive description of the world—a description such that: (1)
there is precisely one way for the world to be that would satisfy the
description, and (2) the world, as it actually is, satisfies the description.
A critic might be tempted to think that anti-Tractarianism is incompat-
ible with comprehensivism. ‘According to anti-Tractarianism’—the critic
might argue—‘the same fact can be described in many different ways.
One can say that there is a table, or that some things are arranged table-
wise, or that the world tableizes, or that tablehood is instantiated, or that
two half-tables are put together in the right sort of way, and so forth,
with no natural end. But one hasn’t given an exhaustive description of
the world until one has described it in all these ways. So the anti-Tractar-
ian could never give a comprehensive description of the world.’
To see where the critic goes wrong, it is useful to consider an
example. Suppose I hand you a box and ask you to give me a com-
prehensive description of its contents. You examine it and say:
‘There is a hydrogen-1 atom in such-and-such a state, and nothing
else.’ It would be inappropriate for me to respond by complaining
that your answer is incomplete on the grounds that it failed to men-
tion at least two objects: a proton and an electron. Such a response
would be guilty of double-counting. Part of what it is for there to be
a (non-ionized) hydrogen-1 atom is for there to be a proton and an
electron. So when you mentioned that there was a hydrogen-1 atom,
the presence of protons and electrons was already included in the
information you gave me. It is true that you never mentioned pro-
tons and electrons explicitly. But that was not required for your
description to be comprehensive. All that comprehensiveness
requires is that there be precisely one way for the contents of the
box to be such that it would satisfy your description.
Moral: Anti-Tractarianism does not entail that comprehensivism
is false. What it entails is that there could be more than one way of
giving a fully comprehensive description of the world.
Paraphrase
It is tempting to think that in accepting a ‘just is’-statement one com-
mits oneself to the availability of a paraphrase-method for translating
Absolute Generality Reconsidered | 115
21
For a related point, see Alston (1957).
116 | Agustín Rayo
6. ABSOLUTE GENERALITY
22
Compare Eklund (2008).
118 | Agustín Rayo
An Analogy
An analogy might be helpful. Suppose you are told that the ordinals
are built up in stages. One starts with a ‘base’ ordinal, and at each
stage one gets a new ordinal by pooling together all the ordinals that
have been constructed so far. The process is to be carried out
indefinitely.
In the absence of further constraints, your understanding of
‘ordinal’ will be hopelessly incomplete. It will be consistent
with taking the ordinals to be isomorphic with the natural
numbers. But it will also be consistent with taking the ordinals
to be isomorphic with the natural numbers followed by an addi-
tional copy of the natural numbers—or two additional copies, or
three, or as many copies of the natural numbers as there are nat-
ural numbers. In fact, one’s understanding of ‘ordinal’ will be
consistent with taking the ordinals to be isomorphic with any
limit von Neumann ordinal.
Absolute Generality Reconsidered | 119
23
For more on this sort of picture, see Parsons (1974).
24
For more on languages of transfinite order, see Linnebo and Rayo (typescript).
120 | Agustín Rayo
A Language-infused World?
‘Wait a minute!’—you might be tempted to complain—‘Are you
setting forth a view according to which the existence of objects is
somehow constituted by language?’
Absolutely not. What is ‘constituted by language’ is the use of
singular terms. If we had no singular terms (or variables taking
singular-term positions) we wouldn’t be able to describe the
Absolute Generality Reconsidered | 121
7. CONCLUSION
REFERENCES
Alston, W. (1957) “Ontological Commitments,” Philosophical Studies
8–17.
Beall, J., ed. (2003) Liars and Heaps, Oxford University Press, Oxford.
Beaney, M., ed. (1997) The Frege Reader, Blackwell, Oxford.
Benacerraf, P. (1973) “Mathematical Truth,” Journal of Philosophy 70,
661–79. Reprinted in Benacerraf and Putnam (1983).
—— and H. Putnam, eds. (1983) Philosophy of Mathematics, Cambridge
University Press, Cambridge, second edition.
Bennett, K. (2009) “Composition, Colocation and Metaontology.” In
Chalmers et al. (2009).
Block, N. (2002) “The Harder Problem of Consciousness,” The Journal of
Philosophy 99, 391–425.
—— and R. Stalnaker (1999) “Conceptual Analysis, Dualism, and the
Explanatory Gap,” Philosophical Review 108, 1–46.
Bueno, O. and Ø. Linnebo (2009) New Waves in Philosophy of Mathematics,
Palgrave Macmillan.
Burgess, J. (2005) “Being Explained Away,” Harvard Review of Philosophy
13, 41–56.
25
For their many helpful comments, I would like to thank Karen Bennett, Ross
Cameron, Matti Eklund, Andrew Graham, Caspar Hare, Jeremy Hartman, Øystein
Linnebo, Ricardo Mena, David Nicolas, Damien Rochford, Brad Skow, Bob Stalnaker,
Pedro Stepanenko, Eduardo Villanueva, Steve Yablo, and Dean Zimmerman. I would
also like to thank seminar participants at MIT, and audiences at the Paris Absolute
Generality Workshop, the MIT Work in Progress Seminar, the First Annual Meeting
of the Latin American Association for Analytic Philosophy, the Institut Jean-Nicod,
the University of Leeds, the University of Connecticut, and Institute for Philosophi-
cal Research at the National Autonomous University of Mexico. Portions of this
paper were written during the tenure of an ACLS Fellowship, for which I am
extremely grateful.
Absolute Generality Reconsidered | 123
SHOEMAKER’S THEORY
they are in fact static, which could be the result of the introduction
of new catalysts, rather than an actual intrinsic change.
Though Shoemaker’s envisioned skeptical scenarios and the ana-
logues that arise from within dispositionalism itself are subtly dif-
ferent, standard epistemologies will not draw a bright line between
them. Shoemaker may have called our attention to a new and dev-
astating species of skeptical argument, but he has failed to make a
convincing case for dispositionalism.
But don’t go thinking that means that here in the US, the joke has
any power to amuse audiences in England!” Bizarre.
I think this is just a general feature of power and disposition talk.
F disposes things to become G in C iff in C, F disposes things to become G.
And it is easiest to see for habituals: Fs G in C iff in C, Fs G.
It is worth asking why the analogous inference is invalid for sub-
junctive conditionals. At least according to the standard semantics
for subjunctives, it does not follow from (x)(Fx >cf (Gx > cf Hx)) that (x)
(Gx > cf (Fx > cf Hx)). It may be true that all soluble things are such that
if they were immersed in water, they would dissolve, but false that all
things actually immersed in water would dissolve if they were solu-
ble. Suppose the only soluble things are not in water, and that noth-
ing would prevent their dissolving if put in water, but that objects in
the water are such that if they were to become soluble, they would be
prevented from dissolving. Perhaps a sorcerer guards over all of the
immersed objects, and would cast a spell instantly removing them
from water, or else in some way interfering with the dissolution proc-
ess in some way at the microphysical level, should they become solu-
ble. This sorcerer, however, may be perfectly content to let soluble
objects that are not actually immersed dissolve, should they be put in
water. He watches over only the immersed objects.
So in general, it would be a bad idea to infer from the fact that if
something were F, then if it were G, it would be H to the fact that
if something were G, then if it were F, it would be H. Given the
undeniably close relation between disposition attributions and sub-
junctive conditionals, we must take care to avoid that fallacious
inference in another guise. Yet it seems undeniable that if in water,
soluble objects are disposed to dissolve, then soluble objects are dis-
posed to dissolve in water, and also that if in water, soluble objects
dissolve, then soluble objects dissolve in water.
Why is this so? I think the cases where the relevant inference fails
for subjunctives are precisely those cases where disposition ascrip-
tions succeed and counterfactual or subjunctive analyses appear to
falter. So-called “finkish” dispositions are dispositions that disap-
pear in their stimulus conditions (C.B. Martin 1994). Lewis imag-
ines a sorcerer guarding a fragile chalice, poised to cast a spell
making it not fragile if it is going to be dropped (1997). The chalice’s
disposition is finkish, vanishing exactly when called upon. It’s frag-
ile, intuitively, but would not break if dropped. Relatedly, Bird notes
Goodbye, Humean Supervenience | 139
that “antidotes” are external factors that disrupt and prevent the
manifestation of dispositions in the stimulus conditions, e.g., the
antidote to a poisonous venom (1998). Here the venom has a
disposition to kill when ingested, even though it may be ingested
without causing death because of the interference of the antidote.
Finally, we can generate the reverse phenomenon, where a certain
subjunctive conditional holds without the associated disposition,
by appealing to what Mark Johnston calls “mimics” (1992). Imagine
a non-fragile piece of clay attached to a motion-sensing machine
that, if it detects jarring or dropping of the clay, will instantly stream
liquid nitrogen over the clay, causing it to shatter.
This is my claim: all counterexamples to the pattern of subjunctive
inference above are also cases of finks, antidotes, or mimics. That is why
the analogous form of reasoning is permissible in the case of pow-
ers, dispositions, and habituals.
I do not have a proof that all of the cases in which the relevant
inference pattern for subjunctive conditionals fails are cases of finks,
antidotes, or mimics, but you’re invited to submit counterexamples.
My only goal is to explain away the threat that swapping the
conditions in and out of the scope of the power operator inherits the
danger of the analogous operation for subjunctives. To dismiss this
threat, it’s enough to have strong initial intuitions about the legiti-
macy of the inference and a plausible explanation of the difference
between the good case (dispositions) and the bad case (subjunc-
tives), and that much, I think we have.
Even on a pure subjunctive conditionals account of dispositions,
Lewis is in trouble. While, in general, the form of inference with
subjunctives noted above is invalid, it may still be valid for a subset
of cases. For instance, suppose something is an F-detector iff it is
disposed to register “1” if in a world with an F. Now, we can make
this disposition to register “1” if in a world with an F as strong as
we like. So let’s make it a full-strength, deterministic, unconditional
power. Lewis might argue that there couldn’t be any such power.
But I don’t see why he would say that. For him, it would just be a
highly non-natural property. Now, isn’t every F such that, if it were
in a world with an F detector, the F detector would register “1”? If so,
then F underwrites a non-trivial subjunctive conditional. So, on a
subjunctive conditional account of dispositions, F underwrites a
disposition.
140 | Troy Cross
ANTI-HYPERINTENSIONALITY
Observe that for Lewis, the perfectly natural properties, like all
other properties, are sets of possibilia: “. . . the property of being
donkey comes out as the set of all donkeys, the donkeys of other
worlds along with the donkeys of ours” (Lewis 2001b). The natural
properties may be specially marked by perfectly resembling tropes,
universals, or simply primitive naturalness, but properties, whether
natural are not, are sets, and sets are, by the axiom of extensionality,
individuated by their membership. Thus if a perfectly natural prop-
erty P and a power P’, which is endowed by P, are necessarily co-
extensive, then P=P’.
Given the argument above for Shoemaker’s thesis, it follows that
for Lewis the perfectly natural properties—all of them—are pow-
ers. And if the perfectly natural properties, the local qualities of
which Lewis speaks, are all powers, then Humean Supervenience
isn’t worthy of its name. Lewis might as well have said that every-
thing supervenes on powers, or that there is just one little disposi-
tion and then another.
142 | Troy Cross
DLEWIS
Dlewis agrees with Lewis about which properties are the perfectly
natural ones. That is, when Lewis and Dlewis survey the pluriverse
from their God-like pedestals, there is never an occasion on which
Lewis says some set of possibilia, S, is natural and Dlewis does not,
or vice versa. They also agree on principles of recombination. They
even agree on counterfactuals.
There are notable differences in their metaphysics, however.
Dlewis says that all of the natural properties are dispositions to have
certain effects, conditional on global states of the world. Which glo-
bal states? Precisely the ones Lewis says are the conditions in which
the perfectly natural properties endow the powers to bring about
those effects. In other words, Dlewis does not fight the general
inference I was trying to draw earlier, but embraces it.
Lewis’s actual views are a bit more complex, but if we think about
the (very nearby) nomic theory of powers discussed earlier, we can
simplify and say that F endows the power to become G just in the
circumstances in which it is nomically necessary that Fs become Gs.
Thus, for Dlewis, F-ness is (in part) the power to become G if in the
following condition: the global state of the world is such that the
simplest and most informative axiomatic description of that world
entails that Fs become Gs (i.e., it’s a Lewis-law that Fs become Gs).
For Dlewis, such global states of the world do not constitute laws.
He thinks, like a typical dispositionalist, that laws reflect the dispo-
sitional essences of properties. He thinks laws are necessary truths
and that (deterministic) dispositions metaphysically necessitate
their effects. Yet, he agrees with Lewis that there are worlds where
Fs do not become Gs. As Dlewis sees it, those are worlds where the
crucial (global) activation conditions, e.g., the supervenience basis
for its being a “Lewis law” that Fs become Gs, are absent, and thus
fail to trigger the disposition into action.
What Lewis sees as contingency in which powers are granted by F,
Dlewis sees as contingency in the existence of the activation condi-
tions for various F-involving dispositions. For Dlewis, that contin-
gency in no way counts against the metaphysical necessity of the
relation between a disposition, its activation conditions, and its
manifestation. And Lewis agrees that each connection Dlewis thinks
is necessary is indeed necessary.
Suppose Lewis somehow resists my earlier argument. Dlewis’s
metaphysics still looks coherent, and, when it comes to the charac-
Goodbye, Humean Supervenience | 145
Well and good. Except that for Lewis to say such a thing, he would
also have to say that properties can be non-identical even if they’re
necessarily co-extensive. That would mean that properties are not
sets of possibilia, that naturalness is neither a primitive feature of
sets, nor a matter of whether sets have members instantiating a sin-
gle universal nor perfectly resembling tropes (Lewis 1983).
Hyperintensionalism, then, is a perfectly acceptable counter to
my argument, and I would be content if it were widely recognized
as a presupposition of categoricalism. But there is reason to assume
that Lewis himself would not abandon his set-based theory of
properties. One reason, though certainly not the only one, is that
the concession would undermine part of Lewis’s argument for
modal realism, namely, his objection to building a semantics for modal
146 | Troy Cross
1. Significance
Some philosophers have granted the central argument of this paper,
but charged it with triviality. According to this objection, the
Humean should simply accept my conclusion—that natural prop-
erties are all powers—without batting an eyelash. After all, the
powers identified with Lewisian natural properties are highly unu-
sual, and easily screened off from the powers endorsed by mystery
mongering dispositionalists: the Lewisian is committed only to per-
fectly local powers, instantiated by point-sized objects, and sensi-
tive only to global states of the world. What’s wrong—so goes the
objection—with saying that qualities are, in fact, identical to these
kinds of powers? Yes, ontologically speaking, Lewis is Dlewis, but
so what?
We can resolve the matter by careful attention to the disposi-
tionalist/categoricalist dialectic. The expressions “quality” or
“categorical property” are typically used by categoricalists to mark
Goodbye, Humean Supervenience | 147
2. Symmetry
Another objection charges me with favoritism. The Lewis/Dlewis
scenario, as depicted, has no obvious asymmetries, so if Lewis is in
trouble, then so, it seems, is Dlewis. In particular, Dlewis, as a fel-
low anti-hyperintensionalist, is just as incapable of distinguishing
himself from Lewis as Lewis is incapable of distinguishing himself
from Dlewis. So, why shouldn’t we conclude by extension that dis-
positionalists generally, at least if they are anti-hyperintensionalists,
are in the same boat as categoricalists?
The objection has a grain of truth: anti-hyperintensionalism can
create a problem for categoricalists and dispositionalists alike.
Dlewis is a case in point. However, actual dispositionalist philoso-
phers differ from Dlewis in at least three respects.
First, they tend not to be anti-hyperintensionalist, at least not
explicitly. Perhaps this is because anti-hyperintensionalism is a
remnant of nominalism and dispositionalists tend to be less squeam-
ish about property realism, but the literature simply does not hold
any loudly anti-hyperintensionalist dispositionalists.
Second, dispositionalists do not attempt to reduce or explain the
categorical in terms of a purely dispositional, non-categorical base.
Rather, they are often friendly to the idea that dispositions are them-
selves categorical. It is categoricalists, historically, who have insisted
on the non-categoricity of dispositions, in an effort to banish them
from ontology along with the rest of the merely possible. If any-
thing, dispositionalists tend to applaud C.B. Martin’s insistence
that dispositions are as actual, as categorical, as anything else (1993,
75; 1996). Though there is a literature on the plausibility of pure
Goodbye, Humean Supervenience | 149
CONCLUSION
WORKS CITED
Achinstein, Peter. 1974. The Identity Conditions of Properties. American
Philosophical Quarterly 11: 257–75.
Armstrong, D. M. 1973. Belief, Truth and Knowledge. Cambridge University
Press.
—— . 1978. A Theory of Universals: Volume 2: Universals and Scientific Real-
ism. Cambridge University Press.
—— . 1980. Identity Through Time. In Time and Cause: Essays Presented to
Richard Taylor, ed., P. van Inwangen, 67–78, Dordrecht: D. Reidel.
Armstrong, D. M., C. B. Martin, U. T. Place, and Tim Crane. 1996. Disposi-
tions: A Debate. Psychology Press.
Ashwell, Lauren. 2010. Superficial Dispositionalism. Australasian Journal
of Philosophy 88: 635–53.
Bird, Alexander. 1998. Dispositions and Antidotes. Philosophical Quarterly
48: 227–34.
—— . 2007. Nature’s Metaphysics. New York: Oxford University Press.
Goodbye, Humean Supervenience | 151
Bird, Alexander and Toby Handfield. 2008. Dispositions, Rules and Finks.
Philosophical Studies 140: 285–98.
Carroll, John. 1994. Laws of Nature. New York: Cambridge University
Press.
Choi, Sungho. 2008. Dispositional Properties and Counterfactual Condi-
tionals. Mind 117: 795–841.
—— . Intrinsic Finks and the Dispositional/Categorical Distinction. Noûs,
forthcoming.
Clarke, Randolph. 2010. Opposing Powers. Philosophical Studies 149: 153–60.
Cross, Troy. 2004. The Nature of Fundamental Properties. Dissertation,
Rutgers, the State University of New Jersey.
—— . 2005. What is a Disposition? Synthese 144: 321–41.
Davidson, Donald. 1963. Actions, Reasons, and Causes. Journal of Philoso-
phy 60: 685–700.
—— . 1969. The Individuation of Events. In Essays in Honor of Carl
G. Hempel, ed. Nicholas Rescher, 216–34.
Davis, Wayne. 1984. A Causal Theory of Intending. American Philosophical
Quarterly 21: 43–54.
Eagle, Antony. 2009. Causal Structuralism, Dispositional Actualism, and
Counterfactual Conditionals. In Dispositions and Causes, 65–99. Oxford
University Press.
Evans, Gareth. 1973. The Causal Theory of Names. Proceedings of the Aris-
totelian Society 47: 187–208.
Everett, Anthony. 2009. Intrinsic Finks, Masks, and Mimics. Erkenntnis 71:
191–203.
Fara, Michael. 2005. Dispositions and Habituals. Noûs 39: 43–82.
Freddoso, Alfred J. 1986. The Nature of Necessity. Midwest Studies in Phi-
losophy 11: 215–42.
Gettier, Edmund L. 1963. Is Justified True Belief Knowledge? Analysis 23:
121–3.
Goldman, Alvin I. 1967. A Causal Theory of Knowing. The Journal of Phi-
losophy 64: 357–72.
Goodman, Nelson. 1983. Fact, Fiction, and Forecast. 4th ed. Harvard Uni-
versity Press.
Grice, H. P. 1961. Symposium: The Causal Theory of Perception. Proceed-
ings of the Aristotelian Society 35: 121–68.
Harre, R., and E. H. Madden. 1975. Causal Powers: A Theory of Natural
Necessity. Oxford: Blackwell.
Hawthorne, John. 2001. Causal Structuralism. Philosophical Perspectives:
361–78.
Heil, John. 2005. From an Ontological Point of View. Oxford University
Press, USA.
152 | Troy Cross
Holton, Richard. 1999. Dispositions all the way round. Analysis 59: 9–14.
Johns, Richard. 2002. A Theory of Physical Probability. University of Toronto
Press, September 28.
Johnston, Mark. 1992. How to Speak of the Colors. Philosophical Studies
68: 221–63.
Kneale, William. 1952. Probability and Induction. Oxford University Press.
Kripke, Saul A. 1980. Naming and Necessity. Harvard University Press.
Lewis, David. 1973. Causation. Journal of Philosophy 70: 556–67.
—— . 1979. Counterfactual Dependence and Time’s Arrow. Noûs 13:
455–76.
Lewis, David K. 1983. New Work for a Theory of Universals. Australasian
Journal of Philosophy 61: 343–77.
—— . 1987. Philosophical Papers: Volume II. Oxford University Press.
—— . 1997. Finkish Dispositions. Philosophical Quarterly 47: 143–58.
—— . 2001a. Counterfactuals. 2nd ed. Wiley-Blackwell.
—— . 2001b. On the Plurality of Worlds. Wiley-Blackwell.
Manley, David, and Ryan Wasserman. 2008. On Linking Dispositions and
Conditionals. Mind 117: 59–84.
Martin, C. B, and John Heil. 1998. Rules and powers. Philosophical Perspec-
tives 12: 283–312.
Martin, C. B. 1993. Power for Realists. In Ontology, Causality and Mind:
Essays in Honour of D M Armstrong, ed. John Bacon. New York: Cam-
bridge University Press.
—— . 1994. Dispositions and Conditionals. Philosophical Quarterly 44: 1–8.
—— . 1997. On the Need for Properties: The Road to Pythagoreanism and
Back. Synthese 112: 193–231.
Maudlin, Tim. 2007. The Metaphysics Within Physics. Oxford University
Press.
McKitrick, Jennifer. 2003. The Bare Metaphysical Possibility of Bare Dis-
positions. Philosophy and Phenomenological Research 66: 349–69.
Nielsen, H. B., and M. Ninomiya. 2008. Search for Effect of Influence from
Future in Large Hadron Collider. International Journal of Modern Physics
A (IJMPA) 23: 919–32.
Nozick, Robert. 1981. Philosophical Explanations. Harvard University
Press.
Prior, Elizabeth W., Robert Pargetter, and Frank Jackson. 1982. Three The-
ses about Dispositions. American Philosophical Quarterly 19: 251–7.
Putnam, Hilary. 1960. Minds and Machines. In Dimensions of Mind : A
Symposium, ed. Sydney Hook: 362–85. New York: New York University
Press.
—— . 1970. On Properties. In Essays in Honor of Carl G. Hempel, ed. Nicho-
las Rescher. Dordrecht: D. Reidel.
Goodbye, Humean Supervenience | 153
1
The force law alone is not enough to entail that the interaction will conserve
energy. The explanation must also appeal to the fundamental dynamical law: the law
relating forces to the motions they cause (in classical physics: Newton’s second law
of motion).
2
I say “important” and “substantially” in order to acknowledge that two compo-
nents of a coincidence may have some explainers in common—as long as they are
beside the point in the context in which an explanation of the two components is
being demanded. For instance, suppose that the two friends both happened to travel
to Chicago on the same airplane flight. Then there would be some common explain-
ers of their both being there (e.g., the flight, the natural laws governing jet engines).
But these are not the sorts of explainers that we would (ordinarily) be asking for in
asking “What brings you to Chicago?” Likewise, although the fundamental dynami-
cal law (see note 1) is common to the explanation that gravitational interactions con-
serve energy and to the explanation that electric interactions conserve energy, it is
incidental; the force laws involved would (typically) be the focus of our explanatory
demand. Hence, it is a coincidence that both forces conserve energy.
"There sweep great general principles" | 157
3
They are not collectively exhaustive. Rather, they are the extremes; there are
intermediate cases. For instance, suppose that some fundamental kinds of interac-
tions have a certain feature (e.g., are capable of both attraction and repulsion)
whereas others (namely, interactions A, B, and C) do not have this feature. Suppose
it is a law that every kind of interaction with that feature conserves energy, and sup-
pose that law is a constraint. Then the fact that every kind of interaction conserves
energy might be explained by this constraint together with A’s force law, B’s force
law, and C’s force law (along with the fundamental dynamical law). In that case,
energy conservation is neither a complete coincidence nor a constraint. As another
kind of intermediate case, the law of energy conservation might follow from exactly
two separate constraints (e.g., that all of the forces capable of both attraction and
repulsion must conserve energy, and that all of the forces capable only of attraction
or only of repulsion must conserve energy). For the sake of simplicity, I shall not
return to these intermediate possibilities, but I believe that it is clear how my remarks
apply to them.
158 | Marc Lange
4
The explanandum is a scientifically significant fact; it is not a fact that only a
philosopher would inquire into. (Don’t pretend that you don’t know what I mean!)
For example, it is central to the reason why Archimedes’ Principle holds.
"There sweep great general principles" | 159
course, still sound) does not explain why the fluid is not circulat-
ing because it would then inaccurately depict the explanandum
as depending on the particular kinds of forces that happen to be
acting on fluid parcels. The only explanation is a top-down expla-
nation: that any circulation-inducing force would violate energy
conservation, which is impossible. Here is an analogy. Consider
Jones and Smith, each convicted in separate trials before separate
judges of possessing (independently) 100 kilograms of marijuana.
Why did each of them receive a sentence exceeding five years?
The reason is not that Smith’s judge passed this sentence because
he believed that Smith’s crime rose to a certain level of serious-
ness because of various factors including . . . and Jones’s judge
passed this sentence because he believed that Jones’s crime rose to
a certain level of seriousness because . . . if the two judges were
constrained by a mandatory minimum sentencing law to pass
sentences of at least five years for the possession of 100 kilograms
of marijuana. If there is such a law, then it is no coincidence that
the two judges handed down sentences that are alike in this
respect. Rather, the law is a common explainer—and any account
is mistaken if it depicts the two sentences as the products of inde-
pendent judicial decisions that weighed the particulars of the
individual cases.
The success of various proposed top-down scientific explana-
tions, then, depends upon the status of energy conservation as a
constraint. Even if energy conservation is a coincidence, the law
that gravitational forces conserve energy could still be used to help
explain why the fluid does not circulate. But this explanation would
simply be a bottom-up account that portrays the fact that the fluid
does not circulate as arising from the coincidence that each of the
particular kinds of forces acting on the fluid conserves energy. In
contrast, if energy conservation is a constraint, then the fact that the
fluid fails to circulate does not depend on the particular kinds of
forces at work on it.
Here is another way to bring out this contrast. Consider a wooden
block (of any shape) sitting on top of a post, and suppose that across
the upper surface of the block is laid part of a uniform loop of rope
(or chain), while the rest of the loop hangs below the block, experi-
encing uniform downward gravity. Why does the rope loop, having
been laid across the block, not spontaneously begin to turn round
"There sweep great general principles" | 161
5
The explanandum is a very important fact. For example, it is central to Simon
Stevin’s 1586 clootcrans explanation of the law of the inclined plane (Stevin, 1955,
Vol. 1, 178).
6
If we believe that energy conservation is a constraint if it is true, then we are
prepared to confirm the hypothesis that energy is conserved very differently than if
we believe it is a coincidence if it is true. Roughly speaking, if we believe that energy
conservation is a coincidence if it is true, then we regard the fact that one fundamen-
tal kind of interaction conserves energy as no evidence that another kind does (just
as we take my being in Chicago as no evidence that you are there, too, if we believe
that our both being there would be coincidental). However, if we believe that energy
conservation might be a constraint, then we may take the fact that one fundamental
kind of interaction conserves energy as some evidence that another kind also does.
Feynman (1967, 76) says that we are “confident that, because we have checked the
energy conservation here, when we get a new phenomenon we can say it has to
satisfy the law of conservation of energy.” A good example of such a new phenom-
enon was radioactive decay which physicists believed to conserve energy before
they had any significant confidence in any theories regarding the particular force(s)
involved.
162 | Marc Lange
Conservation laws are not the only “great general principles” that
have sometimes been reasonably thought to “sweep” across the
various force laws, explaining why all of those laws share certain
features. One candidate proposed by Heinrich Hertz may not turn
out to succeed. But an adequate metaphysical account of natural
law must at least leave room for explanations of the kind Hertz
proposed.
"There sweep great general principles" | 163
7
A “central force” is a force directed along the line joining the body exerting it and
the body feeling it.
166 | Marc Lange
In that case, the kinds of forces and force laws there could have
been go beyond the kinds there actually are. For gravitational forces
to exist and to diminish with the inverse-cube of the distance, for
example, is physically impossible (i.e., is logically inconsistent with
some laws of nature—in this case, the gravitational-force law) but
nevertheless possesses some broader species of possibility in being
logically consistent with the various constraints on the force laws.
In contrast, if energy conservation is a constraint, then energy’s con-
servation is not just physically necessary, but also possesses an even
stronger kind of necessity (that is, one that applies to some but not
all of the physical necessities).
On this view, a top-down explanation may proceed by expressly
considering hypothetical states of affairs that the lower-level laws
rule out (and that are not even approximations to or idealizations of
some physical possibility). A top-down explanation may succeed
even if it appeals to a physical impossibility, as long as that hypo-
thetical state of affairs is not ruled out by the constraints on the
lower-level laws. The top-down explanation exploits this broader
species of possibility since it works by showing the explanandum
to possess the corresponding species of necessity (stronger than
physical necessity).
Here is an example of such an explanation: the standard textbook
explanation (originating with J. Willard Gibbs) of the entropy of a
mixture of two non-interacting ideal gases. The explanation uses
energy conservation to account for the expression for ΔS: the differ-
ence between the mixture’s entropy and the entropy of the gases
when separated. Suppose NA molecules of gas A occupy volume VA
(the left side of a container) and NB molecules of gas B occupy vol-
ume VB (the right side); the container is isolated and the two gases
have the same pressure P and temperature T. Suppose gas A is con-
fined behind a freely moveable membrane permeable to B but not
to A, and gas B is similarly confined behind a membrane permeable
to A but not to B. Initially, the two membranes divide the container
along the same plane, so the gases are entirely separated. Then the
membranes are allowed to move slowly, each gas expanding quasi-
statically, so that ultimately the two membranes reach opposite
ends of the container and both gases fill the entire container
(volume VA + VB ). Each gas’s expansion is a reversible isothermal
process. Let W be the total work done on the system:
"There sweep great general principles" | 167
W = − ∫ VA PdV − ∫ VB P dV = − ∫ VA
VA +VB VA +VB VA +VB dV
N A kT
V
− ∫ VBA
VA +VB dV VA VB
N B kT = N A kT ln + N B kT ln
V +
VA VB VA + VB
By energy conservation (i.e., the first law of thermodynamics), the
change ΔU in internal energy and the heat Q absorbed are related
by
ΔU = Q + W
Since the gases expand isothermically, ΔU = 0 , so Q = −W. Thus
VA + VB V + VB
Q = N A kT ln + N B kT ln A
VA VB
Later he elaborated:
8
Many textbooks dance lightly over the fact that these membranes are generally
physically impossible, characterizing the two gases as separated “conceptually”
(Yourgrau, Van der Merwe, and Raw 2002, 235) or “hypothetically” (Annamalai and
Puri 2002, 145) without elaborating any further. Fermi (1956, 101) is admirably forth-
right: “We should notice . . . that in reality no ideal semipermeable membranes exist.
The best approximation of such a membrane is a hot palladium foil, which behaves
like a semipermeable membrane for hydrogen.”
"There sweep great general principles" | 169
Let’s now try to be more precise about what it would take to make
a conservation law into a constraint rather than a coincidence. As
170 | Marc Lange
I hinted at the close of the previous section, I suggest that this dis-
tinction be elaborated in terms of subjunctive conditionals. Energy
conservation constrains the possible force laws exactly when energy
would still have been conserved even if there had been an additional
kind of force (that is, a force that is not electric or gravitational or any
of the other actual kinds) acting together with the various actual
kinds—that is, even if there had been an additional kind of interac-
tion experienced by some of the same entities undergoing some of
the actual kinds of interaction. (If the additional kinds of force were
uninstantiated, then they would obviously pose no threat to energy
conservation. If forces of the additional kinds were not influencing
any of the actual sorts of entities, then they would pose no threat to
the conservation of quantities possessed exclusively by those enti-
ties.) The subjunctive fact associated with energy conservation as a
constraint is supposed to be roughly that energy’s conservation is
resilient: that energy would still have been conserved even if there
had been additional kinds of force threatening to undermine its con-
servation. On the other hand, to say that energy conservation is a
coincidence of the actual force laws is to say that it is not the case that
energy would still have been conserved, had there been additional
kinds of force. Rather, energy is conserved because as it happens,
each of the actual kinds of force conserves energy as a result of its
own particular force law. So had there been additional kinds of force,
energy might still have been conserved, but then again, it might not
have been, depending upon the force laws of the additional forces.
This means of distinguishing constraints from coincidences por-
trays constraints as like “higher-order” laws. The lawhood of Cou-
lomb’s law is traditionally thought to be associated with the fact
that Coulomb’s law would still have held, had there been additional
charged bodies. Similarly, the accidental character of the fact that
each of the families on my block has exactly two children is associ-
ated with the fact that it is not the case that had there been an addi-
tional family on my block, it would still have been true that each of
the families on my block has exactly two children. My account
draws the same sort of distinction at a “higher order”: energy con-
servation is a constraint exactly when energy would still have been
conserved, had there been additional kinds of forces.
This means of distinguishing constraints from coincidences fits
nicely into my more general account of natural law. I have presented
"There sweep great general principles" | 171
In this proposal and until further notice, I reserve letters like “m”
for “sub-nomic” claims, i.e., for claims such as “The emerald at spa-
tiotemporal location . . . is 5 grams” or “All emeralds are green” as
contrasted with “nomic” claims such as “It is an accident that the
emerald at spatiotemporal location . . . is 5 grams” or “It is a law that
all emeralds are green”. (On my view, a claim is “sub-nomic” exactly
when in any possible world, what makes the claim hold (or fail to
hold) is not that a given fact in that world is a law or that a given
fact in that world is an accident.) Let me also note that the account
of laws I am sketching here presupposes that every logical conse-
quence of laws qualifies as a law and that every broadly logical
172 | Marc Lange
9
For the sake of simplicity, this definition of “sub-nomic stability” omits some
details from my (2009) that will not make any difference here.
"There sweep great general principles" | 173
consistent with all of this set’s members). For instance, the set spanned
by the fact that all gold cubes are smaller than a cubic meter is unsta-
ble because this set’s members are all logically consistent with Bill
Gates wanting a gold cube larger than a cubic meter, yet the set’s
members are not all invariant under this counterfactual supposition.
It is a law that m, then, exactly when m belongs to a (non-maxi-
mal) sub-nomically stable set. Now let’s show that this account
leaves a natural place for the distinction between constraints and
coincidences. Are there any other non-maximal sub-nomically sta-
ble sets besides Λ? The sub-nomic broadly logical truths form a sub-
nomically stable set. I’ll now show that for any two sub-nomically
stable sets, one must be a proper subset of the other:
1. Suppose (for reductio) that Γ and Σ are sub-nomically stable,
t is a member of Γ but not of Σ, and s is a member of Σ but
not of Γ.
2. Then (~s or ~t) is logically consistent with Γ.
3. Since Γ is sub-nomically stable, every member of Γ would
still have been true, had (~s or ~t) been the case.
4. In particular, t would still have been true, had (~s or ~t)
been the case. That is, (~s or ~t) □→ t.
5. So t & (~s or ~t) would have held, had (~s or ~t). Hence, (~s
or ~t) □→ ~s.
6. Since (~s or ~t) is logically consistent with Σ, and Σ is sub-
nomically stable, no member of Σ would have been false
had (~s or ~t) been the case.
7. In particular, s would not have been false, had (~s or ~t)
been the case. That is, ~((~s or ~t) □→ ~s).
8. Contradiction from 5 and 7.
Thus, the sub-nomically stable sets must form a nested hierarchy.
Since no non-maximal superset of Λ is stable (since it would
include an accident), any other stable sets must be among Λ’s proper
subsets. Many of them are clearly unstable. For instance, the set
spanned by a restriction of Coulomb’s law to the past is unstable
since had Coulomb’s law been violated sometime in the future,
then (with Coulomb’s law “out of the way”) it might have been
violated sometime in the past.
However, some of Λ’s proper subsets may be stable, and I sug-
gest that any constraint must belong to at least one such set. Other
174 | Marc Lange
and therefore are constraints. That is because for any nomically sta-
ble set, its sub-nomic members must form a sub-nomically stable
set. Here is the proof:
1. If p (a sub-nomic claim) is logically inconsistent with a
nomically stable set Γ, then Γ must entail ~p (also sub-
nomic), and so p is logically inconsistent with the set Σ con-
taining exactly Γ’s sub-nomic logical consequences.
2. Conversely, if p is logically inconsistent with Σ, then obvi-
ously p is logically inconsistent with Γ.
3. By Γ’s nomic stability, Σ is preserved under every sub-
nomic antecedent p that is logically consistent with
Γ—which (we have just shown) are exactly those sub-
nomic antecedents that are logically consistent with Σ.
Hence, Σ is sub-nomically stable.
Therefore, if the symmetry meta-laws (forming a nomically stable
set) entail that a given conservation law holds under the Hamilto-
nian dynamical framework, then the conservation law’s holding if
the Hamiltonian dynamical law holds belongs to a sub-nomically
stable set that is more exclusive than Λ (since presumably, not all of
Λ’s members follow from the symmetry meta-laws’ nomically stable
set). Hence, if the Hamiltonian dynamical law is not a member of
that set but (in transcending the various particular force laws)
belongs to another sub-nomically stable set that does not include the
force laws, then (since the sub-nomically stable sets form a nested
hierarchy) the fact that the conservation law holds under the Hamil-
tonian framework must also belong to that set, and hence (by the
set’s logical closure) the conservation law must belong, too. So it
constitutes a constraint. In other words, that the conservation law
would still have held, even if the force laws had been different, fol-
lows from the fact that not only the fundamental dynamical law, but
also the symmetry meta-law would still have held had the force laws
been different.
Of course, I cannot do more here than sketch the relevant parts of
this conception of natural law. But it is worth seeing how an account
of lawhood can incorporate the constraint/coincidence distinction
in a natural way. Now I shall conclude by turning to an approach to
natural law that is ill-equipped to have the same success.
"There sweep great general principles" | 177
laws. Even if gravity and the electron have essences, it is not obvi-
ous that ‘the world’—reality—does” (Lange 2009, 83). Bird also
finds this move “somewhat ad hoc” (2007, 213). However, I shall
not pursue this concern here.
This form of the objection may turn out not to pose a terribly
severe challenge for Bird’s view. One option that Bird might explore
is to say simply that although the conservation “law” is not a meta-
physical necessity, the fact that energy is conserved is no accident
either. Rather, it is grounded in the sparse fundamental properties
of physics, i.e., in the world’s repertoire of fundamental causal pow-
ers. But this option would require Bird to allow different repertoires
in different possible worlds, which Bird is reluctant to do (since in
that case the actual laws, though still true in all worlds, are not laws
in some of them). Another option for elaborating Bird’s view accords
metaphysical necessity to the conservation law and allows every
world to have exactly the same laws. As I just explained, energy’s
conservation in electric interactions is a metaphysical necessity aris-
ing from the dispositions essential to various fundamental proper-
ties of physics. So likewise is energy’s conservation in gravitational
interactions, and similarly for any other actual type of interaction.
On Bird’s view, the various fundamental properties of physics
would be different properties if they bestowed additional powers
and susceptibilities—for instance, the power to exert and the sus-
ceptibility to feel some alien force that fails to conserve energy. Pre-
sumably, no metaphysical necessity precludes the instantiation of
alien fundamental properties of physics. But “electric charge” (or
any other non-alien property) would be a different property if it
bestowed susceptibility not just to the influence of other electric
charges, but also to some other alien influence. What if we go fur-
ther and suppose that all of the non-alien properties (or at least
enough of them that all non-alien kinds of entities must possess
one) have as part of their essences that they bestow immunity to
being influenced by any alien property (and so their possession by
an entity precludes its also possessing any alien property that would
bestow susceptibility to being so influenced)? Then the possession
of any alien properties (by entities not possessing any of the non-
alien properties) could not influence the behavior of any entities
possessing non-alien properties and so could not disturb the con-
servation of the quantities that are in fact conserved under each of
180 | Marc Lange
But I have argued (in the opening sections of this paper) that the
price of adopting this view is too high. It precludes top-down expla-
nations of a kind that science has reasonably taken seriously—
indeed, on which science has often placed great importance. Views
like Bird’s rule out scientifically respectable theories. Perhaps Bird’s
prognostication regarding future science will be proved right; per-
haps symmetry principles and conservation laws, along with the
alleged top-down explanations they supply, will ultimately be
eliminated from physics—and their place not be taken by other
top-down explanations (like Hertz’s) employing other sorts of
constraints. Personally, I doubt it. But more importantly, a meta-
physics that cannot do justice to top-down explanations and the
constraints they require is at a serious disadvantage even if as a
matter of fact, there turn out to be no such explanations. Room
should still be left for them; it should be up to science rather than
metaphysics to foreclose them.
Consider, for instance, the conservation of baryon number in
contemporary physics. By energy conservation, an isolated proton
can decay only into particles that have less rest-mass than it does,
and the proton is the lightest baryon (i.e., the lightest particle with
non-zero baryon number). The conservation of baryon number thus
entails that the proton is stable (radioactively, I mean—not “sub-
nomically!). Does baryon-number conservation explain why the
proton is stable? This remains controversial.
It is no explanation if the conservation of baryon number is a so-
called “accidental symmetry” (a term that was introduced by Ste-
ven Weinberg; see Weinberg 1995, 529). An accidental symmetry
reflects merely the particular forces in action at lower-energy
regimes rather than some deeper “symmetry of the underlying the-
ory” (Weinberg 1995, 529). Accordingly, if baryon-number conser-
vation is an accidental symmetry, then it may not even hold at
higher energies (and so the proton may turn out not to be stable, but
"There sweep great general principles" | 185
REFERENCES
Annamalai, K. and Puri, I. (2002), Advanced Thermodynamic Engineering.
Boca Raton, CRC Press.
Bartlett, D.F. and Su, Y. (1994), “What Potentials Permit a Uniqueness
Theorem,” American Journal of Physics 62: 683–6.
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 Philoso-
phy of Science 43: 371–88.
Bird, A. (2007), Nature’s Metaphysics. Oxford: Oxford University Press.
186 | Marc Lange
CONTINGENT OBJECTS
AND COINCIDENT
OBJECTS
This page intentionally left blank
7. Relativized Metaphysical Modality*
Adam Murray and Jessica Wilson†
INTRODUCTION
* Special thanks to Karen Bennett and Dean Zimmerman for extensive comments
which greatly improved this paper. Thanks also to Benj Hellie, Phil Kremer, Dan
Rabinoff, Chris Tillman, and audience members at the University of North Carolina
and the University of Miami, for helpful comments and questions, and to Brent
Cromwell for assistance in constructing the figures.
†
Department of Philosophy, University of Toronto; adam.murray@utoronto.ca,
jessica.m.wilson@utoronto.ca
190 | Adam Murray and Jessica Wilson
operators are briefly introduced in order to respond to concerns about the validity of
axioms of the form Aα→□ Aα on the semantics offered for ‘A’, that ‘“actually a” need
not have been true because another world might have been actual’ (1979, 2); in the
later paper, these new modal operators are applied in service of accommodating
putative cases of the contingent a priori and necessary a posteriori, with the assist-
ance of certain theses about names and natural kind terms. Our motivations and
target applications are different. We aim to show that certain puzzles, having to do
with natures or essences as opposed to scope or semantics (or epistemology), ulti-
mately arise from a failure of standard metaphysical modal logics to incorporate rela-
tivization to indicatively actual worlds, and to argue that proper incorporation of a
relativized structure makes for better resolutions of these puzzles than those cur-
rently on offer. Our suggested ways of making sense of such relativization within a
possible worlds semantics for modal logic do not involve any additional operators,
and though in the course of explicating our view we heuristically appeal to properly
metaphysical interpretations of the notions, familiar from 2-D semantics, of consider-
ing worlds ‘as actual’ or ‘as counterfactual’, unlike Davies and Humberstone we are
officially neutral on both the semantics and epistemology of names and natural kind
terms. These differences aside, in arguing that the traditional modal operators should
be relativized to indicatively actual worlds we are on the same side as these other
authors; our contribution here is, first, to offer distinctively metaphysical reasons for
incorporating such relativization, and second, to show that this can be done in ways
minimally departing from standard modal logic(s).
2
More precisely, modal logics that include the distribution axiom (□(p → q) → (□p
→ □q)) are known as normal modal logics. We assume in what follows that the modal
logics under discussion are normal.
192 | Adam Murray and Jessica Wilson
3
This assumption reflects, in part, that the theorems of S5 coincide with those on
a modal logic where the accessibility relation is ‘total’, such that (as per the standard
conception) every world is accessible to every other. To prefigure a bit: one way of
implementing our transitive relativized conception of metaphysical modality takes
advantage of the fact that, notwithstanding the coincidence of theorems, S5 is com-
patible with modal space being partitioned into non-overlapping equivalence classes
(see S2.2).
Relativized Metaphysical Modality | 193
4
For simplicity we have altered the indexing on the hunks of matter at issue, and
will later do so for associated worlds.
194 | Adam Murray and Jessica Wilson
The last claim expresses that a certain claim is necessary, but not
necessarily necessary, contra axiom (4). Salmon thus concludes that
neither S4 nor S5 is the correct logic for metaphysical modality.5
We find the data that Salmon canvasses in the Woody case to be
intuitively compelling. But do the data really establish, as Salmon
maintains, that axiom (4) should be rejected as a general constraint
on metaphysical modality?
5
Salmon’s discussion focuses on the normal modal logical systems S4 and S5.
However, insofar as Salmon’s target is the transitivity of the accessibility relation, as
characterized by axiom (4), his conclusion plausibly extends to simpler modal logics
such as K4, which also include (4). Thanks to Phil Kremer for discussion of this point.
Relativized Metaphysical Modality | 195
The first pass is not yet sufficiently specified, however, since it fails
to express the sense in which Woody’s actual origins in matter m are
‘held fixed’, notwithstanding that in evaluating the antecedent of
the conditional, it is supposed that Woody actually originates in
some different matter. What is needed is the distinction between a
given state of affairs (or whatever) being indicatively vs. its being
counterfactually actual. To prefigure our heuristic analogy: such a
distinction is operative when we allow that, holding fixed that
‘water is H2O’ is (‘indicatively’) true in our very own actual world,
‘water is H2O’ would remain true were a world where the watery
stuff is XYZ to be (‘counterfactually’) actual. The first reading of
premise 4 should mark this distinction, as follows:
Holding fixed that Woody indicatively actually originates from matter m: if
Woody had originated from matter m’, then it would have been possible for
Woody to originate from matter m”.
6
Note that the claim here is not that premise 4 is false because the consequent of
the embedded conditional is actually false; the claim is rather that, holding fixed that
Woody indicatively actually originates in matter m, the consequent would be false in
a world where the antecedent is (counterfactually) true. In other words: the con-
straints imposed by Woody’s indicatively actual origin are in force even in contexts
(e.g., w2) where Woody counterfactually originates from different matter.
Relativized Metaphysical Modality | 197
4’. If Woody had originated from matter m’, then it would have
been possiblew2 for Woody to originate from matter m”.
Correspondingly, premise 5 should now reflect that the possibilities
at issue are evaluated with respect to different indicatively actual
worlds:
5’. It is possiblew1 that it is possiblew2 that Woody originates
from matter m”. (2’, 4’)
Properly specified, then, Salmon’s argument is as follows:
1’. Woody originates from matter m.
2’. It is possiblew1 that Woody originates from matter m’.
3’. It is not possiblew1 that Woody originates from matter m”.
4’. If Woody had originated from matter m’, then it would have
been possiblew2 for Woody to originate from matter m”.
5’. It is possiblew1 that it is possiblew2 that Woody originates
from matter m”. (2’, 4’)
6’. It is not possiblew1 that Woody originates from matter m”,
but it is possiblew1 that it is possiblew2 that Woody origi-
nates from matter m”. (3’, 5’)
7’. It is necessaryw1 that Woody does not originate from mat-
ter m”, but it is not necessaryw1 that it is necessaryw2 that
Woody does not originate from matter m”. (6’)
Note that the last claim is no longer a clear counter-instance to axiom
(4). As per the standard assumption that metaphysical modality is
modality in the broadest (non-syntactically logical) sense, previous
discussions of the axiom have not incorporated the need for relativi-
zation to an indicatively actual world. How the axiom should be
understood in light of the need for relativization—and in particular,
whether it should be understood to apply to modal claims involving
iterated or (as we’ll put it) ‘in situ’ shifts in which world is held fixed
as indicatively actual (of the sort occurring in 5’–7’) remains to be
seen. This result constitutes our first objection to Salmon’s argument
that the Woody case motivates the rejection of transitive accessibility:
if we appropriately attend to which worlds are held fixed as indica-
tively actual in this case, no clear violation of axiom (4) results.
More generally, the need to incorporate facts about which world is
held fixed as indicatively actual in order to appropriately express the
198 | Adam Murray and Jessica Wilson
Salmon notes two related ways in which his treatment departs from
the standard understanding of metaphysical modality. First, ‘[i]f
worlds include ways things metaphysically cannot be in addition
to ways things metaphysically might have been, then the idea
that metaphysical necessity corresponds to truth in every world
whatsoever is flatly mistaken’ (1989, 15). Second, given that ‘[a] pos-
sible world is a total way for things to be that conforms to metaphysi-
cal constraints concerning what might have been [. . .] metaphysical
modality is definitely not an unrestricted limiting case’ (1989, 12–13).
The conception of metaphysical modal space that is in the first
instance suggested by Salmon’s treatment of the Woody case is of a
single space of (‘maximal’, abstract) worlds, whose status as possi-
ble or impossible is relative to whatever world is supposed to
obtain,7 and which are connected by an intransitive accessibility
relation. In pictorial terms: the standard conception of metaphysical
modal space—the conception that Salmon, and we, reject—has the
following structure:
w2
w1 w3
7
We uniformly interpret Salmon’s talk of metaphysical possibilities and necessi-
ties as relative to which world ‘obtains’ (in our terms: is indicatively actual), such
that, e.g., w3 is possible relative to w2 when the latter obtains in just the same way that
our very own actual world obtains. Though Salmon’s emphasis on impossible worlds
might be thought to suggest that he sees a substantive difference between whatever
world in fact obtains and other worlds (e.g., w2) that merely hypothetically obtain,
Salmon seems to reject such a privileging of our world when disparaging what he
calls the ‘ostrich approach to modality, with its consequent misconstrual of “necessar-
ily” as meaning actual necessity and “possibly” as meaning actual possibility”’ (1989,
29). In any case, a restricted understanding would still have the structure to follow.
Relativized Metaphysical Modality | 201
w1 w2 w3
(Here, and in the figures to follow, dotted lines around a world indicate
that the world is held fixed as indicatively actual. Solid-line ovals around
worlds represent the relativization of accessibility due to shifts in which
world is held fixed as indicatively actual.8) Again, we emphasize that in
our view, Salmon’s conception is on the right track, in recognizing the
need for metaphysical modal deliberation to be sensitive to which world
is supposed to ‘obtain’ (be indicatively actual). Still, Salmon’s approach
to a relativized conception is revisionary, in departing from the usual
assumption of transitive accessibility and associated systems of modal
logic; and many have found the posit of metaphysically impossible
worlds problematic (see Lewis 1986, 7, fn. 3 and 246–8).
8
Our purpose here is merely to sufficiently distinguish Salmon’s intransitive rela-
tivized conception from the standard, transitive conception of metaphysical accessi-
bility (see Figure 1). Plausibly, on Salmon’s intransitive conception, had w2 obtained
(in our terminology, been ‘indicatively actual’) then w1 as well as w3 would have
been accessible/metaphysically possible from w2. In this sense, Figure 2 as it stands
represents the relations of accessibility that obtain between worlds w1, w2, and w3 in
a fairly abbreviated form.
202 | Adam Murray and Jessica Wilson
9
Hence, on this implementation, worlds are not alethically ‘maximal’ prior to rela-
tivization; accordingly, in the pre-relativized space, v(p) (where p is the proposition
that Woody originates in matter m”) is indeterminate.
Relativized Metaphysical Modality | 203
w2
w1 w3
w2
w1 w3
w2
w1 w3
10
In this case it will rather be true at w3 that Woody originates from matter m”:
v(p, w3, w2) = T
Furthermore, given that w2 is indicatively actual (together with the fact that, in w2,
Woody originates in matter m'), v will assign T to the proposition that p is possible in
both w1 and w2:
v(◊p, w1, w2) = T
v(◊p, w2, w2) = T
Moreover, since w2 is accessible to w1, v will assign T to the proposition that p is pos-
sibly possible in w1:
v(◊◊p, w1, w2) = T
Again, no violation of transitivity results.
206 | Adam Murray and Jessica Wilson
11
What if some claims concerning Woody—e.g., that Woody is a table—are neces-
sary relative to each indicatively actual world? We may define, if we like, a notion of
‘absolute’ metaphysical necessity. But given the desirability of avoiding shifts in
indicatively actual worlds, we should see the absolute notion as grounded in the rela-
tivized notion—as tracking a uniform pattern of variation in what is relatively meta-
physically necessary—as opposed to taking ‘absolute’ metaphysical necessity to be
either prior to or distinct from relativized metaphysical necessity.
Relativized Metaphysical Modality | 207
w2 w2⬘ w2⬘⬘
w1 w3 w1⬘ w3 ⬘ w1⬘⬘ w3 ⬘⬘
S1 S2 S3
Figure 4. Non-Overlapping Subspaces. In subspace S1, w1 is indica-
tively actual; in subspace S2, w2’ is indicatively actual; in subspaces
S3, w3” is indicatively actual. Each ‘w1-type’ world contains a table-
shaped hunk of matter m; each ‘w2-type’ world contains a table-
shaped hunk of matter m’; each ‘w3-type’ world contains a
table-shaped hunk of matter m”.
p = the proposition that woody originates from matter m”
v(p, w3, w1) = F v(p, w3’, w2’) = T
12
Even though worlds in different subspaces are not identical on this view, worlds
in different partitions may be taken to be ‘basic’ or ‘canonical’ counterparts of each
other, in that (lifting some of the structure of the overlapping spaces interpretation)
worlds in different spaces may be basically alike; such similarity may serve as the
basis for loose (as opposed to strict) identification of worlds across subspaces. So, for
example, distinct subspaces might each contain worlds that are basically, canonically,
or qualitatively similar, in containing, e.g., a table-shaped hunk of matter m”; such
worlds, we might say, are of type w3. In a world of type w3, is it true or false that
Woody originates from m”? That depends. In a subspace where the indicatively
actual world is (of type) w1, this is false; but in a subspace where the indicatively
actual world is (of type) w2, this is true.
208 | Adam Murray and Jessica Wilson
13
Indeed, given the satisfaction of transitive accessibility within a particular sub-
space (either overlapping or non-overlapping), the evaluation of claims of necessity
at a world w in a subspace in which some distinct world w' is indicatively actual is
straightforward (thanks to Phil Kremer for pressing us on this point). Given transitiv-
ity, if some proposition A is necessary at a world w, A is necessarily necessary at w,
and hence necessary at every world in the subspace accessible from w; thus, in par-
ticular we have:
v(□A,w,w') = T iff v(□A, w',w') = T
(A is necessary at w, given that w' is indicatively actual, if and only if A is necessary
at the indicatively actual world w' of the subspace). Necessary truth at the indica-
tively actual world of a subspace can then be given its familiar analysis in terms of
truth at all accessible worlds:
v(□A, w',w') = T iff ∀w[wRw' → v(A, w) = T]
Combining these results yields the truth-clause for necessary propositions evaluated
at worlds accessible from the indicatively actual world of a particular subspace:
v(□A,w,w') = T iff ∀w[wRw' → v(A, w) = T]
Relativized Metaphysical Modality | 209
H2O, and an XYZ-world is one where the watery stuff is XYZ; and
where it is furthermore assumed that our very own actual world is
an H2O-world:
than our very own actual world ‘considered as actual’. We may also
define generalized horizontal intensions, where a generalized horizon-
tal intension is a function which takes a world considered as actual
as its first argument, and a world considered as counterfactual as its
second argument; we call such secondary intensions ‘generalized’
in that the world considered as actual need not be our very own
actual world. So, for example, consider the generalized horizontal
intension associated with ‘water is XYZ’, when an XYZ-world is
considered as actual:
Does the fact that (as attention to the diagonal and generalized
horizontal intensions suggests) we can make sense of claims involv-
ing shifts in indicatively actual worlds pose a problem for our think-
ing that some such shifts are illegitimate, from the perspective of
metaphysical modal reasoning? No. Certain claims and associated
reasoning involving shifts in which world is (considered as) indica-
tively actual are, on our view, perfectly legitimate—namely, those
which are appropriately seen as ‘meta-modal’, in tracking patterns
of what is the case, non-modally or relatively modally, when differ-
ent worlds are held fixed as indicatively actual. We see nothing
defective in claims like the following, tracking patterns in what is
non-modally the case across different worlds considered (held
fixed) as indicatively actual:
Considering as (indicatively) actual a world where water is H2O: water is
the watery stuff; and considering as (indicatively) actual a world where
water is XYZ: water is the watery stuff.
Again, the sort of claims that we maintain are in some sense defec-
tive, from the point of view of metaphysical modal reasoning, are
those involving iterated or ‘in situ’ shifts in which world is indica-
tively actual, of the sort characteristic of Salmon’s disambiguated
premise 5’, or of the following sort of claim:
Considering as (indicatively) actual a world where water is H2O: consider-
ing as (indicatively) actual a world where water is XYZ, then it is necessary
that water is XYZ.
14
Moreover, qua natural property schmass appears to be on a par with mass: it too
is a property that, in appropriate circumstances, lawfully influences the motion of
entities having the property. Hence considerations rendering it natural to think that
necessities involving mass are grounded in its nature and identity would seem
equally to motivate thinking that necessities involving schmass would be grounded
in its nature and identity.
218 | Adam Murray and Jessica Wilson
REFERENCES
Armstrong, D.M., 1989. A Combinatorial Theory of Possibility. Cambridge:
Cambridge University Press.
Bealer, George, 2000. ‘A Theory of the A Priori’. Pacific Philosophical Quar-
terly 81 (1):1–30.
Bennett, Karen, 2005. ‘Two Axes of Actualism’. Philosophical Review
114(3):297–326.
Bird, Alexander, 2007. Nature’s Metaphysics: Laws and Properties. Oxford:
Oxford University Press.
15
Thanks to Laurie Paul for calling this advantage to our attention.
Relativized Metaphysical Modality | 225
1
The idea that a person could change bodies by way of a brain transplant (see
Shoemaker (1963)) takes its inspiration from Locke’s example in which the soul of a
prince is transferred to the body of a cobbler. Eric Olson (1997) has claimed that since
the whole brain includes the brainstem which is the control center for the biological
life of a person, the claim that the person would go with the brain in a brain transplant
is compatible with the rejection of the psychological view; it is compatible with per-
sonal identity consisting in biological continuity rather than psychological continuity.
If this is right, the psychological account should say that the transplantation of just the
cerebrum could be person preserving. For a case in which the change of bodies is via
brain reprogramming, see Perry (1972).
228 | Sydney Shoemaker
entities. It also seems that on such accounts persons are not identi-
cal with the biologically individuated animals with which they
share their space and matter. If the body of a person can survive
death, as a corpse, and if both the person and the biological animal
cease to exist at death, then wherever there is a person there are (at
least) three different coincident entities – the person, the person’s
body, and the biological animal that shares the space and the matter
of both.2
Advocates of perdurance (four-dimensionalist) views of the per-
sistence of objects are committed to a kind of coincidence. They are
committed to the coincidence of temporally extended objects with
their temporal parts, and to the coincidence during the period of
overlap of temporally extended objects that share temporal parts. For
some perdurance theorists this is the only sort of coincidence that is
possible. And they see it as a benign form – since they think that in
the first instance it is the momentary temporal parts of objects that
have the properties things have at times, and since these parts are
shared in cases of temporal overlap, they are not committed to what
may seem an obnoxious consequence of coincidence, namely there
being multiple instantiations of the same property at the same time
and place. E.g., they are not committed to there being two instances
of weighing one hundred pounds where a coincident statue and
piece of clay, each weighing one hundred pounds, are located, and
the apparent consequence of this that a scale ought to read two hun-
dred pounds when the statue is placed on it. If the only coincidence
allowed is in cases of temporary overlap, the perdurance theorist will
not allow that there can be pairs of objects that coincide throughout
their careers. However, it appears that a perdurance theorist could
allow this for the same reasons that others do, namely modal rea-
sons – if the careers of the statues and the piece of clay could come
apart, even if in fact they don’t, it would appear that the statue and
the piece of clay must be different.
In any case, I will not be concerned here with the sort of coinci-
dence that all perdurance theorists allow – that which consists in
the sharing of temporal parts. I don’t myself believe that persisting
2
By a “biological animal” I mean an animal with biological persistence conditions.
A neo-Lockean can hold that persons are animals, but not that they are biological
animals.
Coincidence Through Thick and Thin | 229
II
3
See Shoemaker (1979).
Coincidence Through Thick and Thin | 231
III
we will see shortly, all of the physical properties that figure in our
ordinary thought and discourse are realized in other physical
properties.
It is useful to distinguish two different sorts of realization.4 One is
the sort that figured in the preceding paragraph, where the realizer
of a property instance is another property instance. Call this prop-
erty realization. I favor an account of this on which one property
realizes another just in case the forward-looking causal features of
the realized property are a subset of the forward-looking causal fea-
tures of the realizer, and, accordingly, one property instance realizes
another just in case the realized properties are so related. In the
other sort of realization what realizes a property instance is a micro-
physical state of affairs – its being the case that certain micro-entities
are propertied and related in a certain way. Here too the notion of
realization involves the notion of a causal profile; microphysical
states of affairs have causal profiles, consisting in their aptness to
cause or contribute to causing certain effects, and the causal profile
of the microphysical state of affairs that realizes a property instance
must match the causal profile of the instantiated property.5 Call this
microphysical realization. In both cases there will be a multipli-
city of ways in which the instantiation of a given property can be
instantiated – in both there will be “multiple realization” – but in
both the existence of a realizer of a property instance constitutes
that property instance.
Realization of the second of these sorts, microphysical realization,
brings with it a special case of realization of the first sort, property
realization. If a thing’s career embeds a microphysical state of affairs
at a given time, the thing thereby has at that time the property of hav-
ing a career that at that time embeds such a state of affairs. Call this a
microphysical-state-of-affairs-embedding property, or, for short, an
MSE property. There is a one–one relation between types of micro-
physical states of affairs and MSE properties. Since every physical
property instance has a microphysical state of affairs realizer, every
physical property instance has a property realizer that is an instance
of an MSE property. MSE properties are not properties that figure in
ordinary thought and discourse. They are too fine-grained for that.
4
See Shoemaker (2007).
5
For a discussion of this sort of matching, see Shoemaker (forthcoming).
Coincidence Through Thick and Thin | 233
III
IV
VI
6
I say “normally” because there is the exceptional case in which the stage is the
final stage in the thing’s career, because its propensity to contribute to a future
career is overridden by the causal efficacy of other states of affairs that leads to the
termination of the career.
Coincidence Through Thick and Thin | 239
VII
VIII
IX
It is natural to say that the piece of clay constitutes the statue, and
that the body, or at any rate the biological animal, constitutes the
person. The account I have given of coincidence suggests an account
of constitution. As a first pass, let’s say that object A constitutes
object B if A and B are coincident and the states of affairs that are
strongly embedded in the career of B, and so realize its thick prop-
erties, are only weakly embedded in the career of A.
This faces a difficulty. I have taken it that when coincident objects
are of different kinds they have different persistence conditions,
and that something’s being of a kind and having the persistence
conditions that go with that kind is a thick property of it. This means
that each member of a pair of coincident objects has a thick prop-
erty the other lacks and is realized by a microphysical state of affairs
that is strongly embedded in its career and only weakly embedded
in the career of the other. So on the proposed account of constitution
each member of a pair of coincident objects would constitute the
other. While the piece of clay would constitute the statue, it would
also be the case that the statue constitutes the piece of clay. Simi-
larly, the person would constitute, as well as being constituted by,
the body and the biological animal. Yet we presumably want the
constitution relation to be asymmetrical.
If we could be sure that the objects we want to count as constitut-
ing other objects have no thick properties other than sortal proper-
ties that they don’t share with the objects they constitute, then we
could modify the account so as to give us the asymmetry we want:
A will constitute B just in case A and B are coincident and B has
thick properties not shared by A that are realized by microphysical
states of affairs strongly embedded in B’s career and only weakly
244 | Sydney Shoemaker
7
Doepke (1982), p. 54 and p. 57.
246 | Sydney Shoemaker
One reply to the weight objection rests on the claim that the coin-
cident objects are composed of the same matter. What determines
the reading on the scale is the weight of the matter placed on it, and
in the example this is one hundred pounds, not two hundred
pounds. I have no objection to this reply, but I doubt if it can be
applied in the case of electric charge.
I think that we can give a better reply by appealing to the fact that
the microphysical states of affairs in the careers of the objects are
identical. The influence of the piece of clay’s presence on the scale is
due to the causal profile of a microphysical state of affairs it con-
tains; and since the very same microphysical state of affairs occurs
in the statue, the influence of the statue’s presence on the scale is
already included in the influence of the piece of clay’s presence. The
same point applies to electric charge; there is a single microphysical
state of affairs that determines the effects of both coincident objects
on our instruments, and its occurring in two objects shouldn’t be
expected to double its effects.
XI
Perhaps the most frequently raised objection to the view that there
are coincident objects is that given that such objects share the same
matter, arranged in the same way, it is impossible to explain how
they can be of different sorts and have different persistence condi-
tions and different modal properties. Michael Burke asks “Given
the qualitative identity of these objects, what explains their alleged
difference in sort?”8 Addressing Alan Gibbard’s Lumpl and Goli-
ath, a supposedly coincident lump and statue that come into exist-
ence and go out of existence at the same time, Karen Bennett asks
“What grounds the difference between Lumpl and Goliath, given
that they are otherwise so alike? They are the same shape, the
same size, made of the same parts, have the same history and
future, are the same distance from the bagel store, and so on and so
forth. So what exactly makes it the case that they could have differ-
ent shapes and sizes, etc.?”9 And Eric Olson poses what he calls
8
Burke (1992), pp. 14–15.
9
Bennett (2004), pp. 339–40. Gibbard (1974). Bennett is addressing here only the
case of coincident objects that come into existence and go out of existence at the same
time. But she clearly does not think that the problem arises only for such cases.
Coincidence Through Thick and Thin | 247
10
Olson (2001), p. 339. Note that Olson takes the constitutionalists to hold that
coincident objects have a qualitative difference while Burke takes it to be agreed that
they are qualitatively identical. I think this is only a terminological difference.
248 | Sydney Shoemaker
XII
What I have said so far may have left the impression that on my
view the coincidence of two objects requires that they be of differ-
ent sorts and have different persistence conditions. That is, in fact,
a common view among believers in coincidence; citing Locke they
hold that it is impossible for two objects of the same kind to share
their matter and occupy the same place at the same time. But that
is not my view. I think it is possible for there to be coincident objects
that are of the same kind and have the same persistence conditions.
We need to see how this can be.
Kit Fine has offered one simple example of such intra-kind super-
venience.11 Someone writes a note on a piece of paper. The recipient
turns it over and writes a reply on the other side. There are two
notes, but just one piece of paper. The notes coincide with the piece
of paper, but they also coincide with each other.
In several place I have offered a different example, in which the
coincident objects are both persons.12 The case is similar to that of
Dr. Jekyll and Mr. Hyde. (I got it from Eric Olson, who puts a differ-
ent interpretation on it.13) We are to imagine that two persons
11
Fine (2000).
12
See my 2003 and my 2007.
13
Olson presented his example at a conference on the Self at the University of
Arkansas in September 1999. What I take to be a case of coincident persons he takes
to be a case of one person with a highly unusual history.
250 | Sydney Shoemaker
XIII
I suspect that one reason for the skepticism many philosophers feel
about the idea that there are coincident objects is the paucity of
plausible examples of coincidence. The most frequently cited exam-
ples are the two I have discussed – the case of the statue and the
piece of clay, and the case of persons and their coincident bodies
and biological animals. Some believers in coincidence cannot avail
themselves of the latter example because they deny that persons
and their bodies are distinct; thus Judith Thomson, a believer in
coincidence, says ”I take cats and people to be their bodies.”14
14
Thomson (1998), p. 169.
Coincidence Through Thick and Thin | 251
XIV
To sum up, I believe that there are cases of coinciding objects and
that such cases are common. I have restricted my attention to how
15
See Unger (2000).
252 | Sydney Shoemaker
REFERENCES
Bennett, K. (2004). “Spatio-Temporal Coincidence and the Grounding
Problem.” Philosophical Studies, 118: 339–71.
16
Thanks to Karen Bennett and Carl Ginet for comments on an earlier draft.
Coincidence Through Thick and Thin | 253
Recent times have been very much focussed on the future. The elec-
tion of Barack Obama in America was accompanied by a wave of
optimism. The Global Financial Crisis and the challenge of climate
change caused many to descend into pessimism. It is not whether
our glass is half-empty or half-full that we worry about, but whether
it will be empty or full. As philosophers, naturally, we want to illu-
minate such concerns, so that we might understand them better
and subject them to rational scrutiny.
According to the Growing-Block view of time, most famously
put forward by C.D. Broad (1923), the flow of time consists in events
coming into existence, so that past things and events are real, while
future things and events are not. While the Growing-Block view is
often considered intuitively appealing, some are concerned it has
little to say about the future it denies the existence of. We show how
Growing-Block theorists can assign meaningful, mind-independent
truth conditions to sentences about the future.
a sentence of the form ϕ(a) for some a ∈ De. (Both these assumptions
are claims about what it is for e to be a complete representation of a
time: e must specify exactly which things exist; and if it says that
something satisfies a particular open sentence, then it also specifies
which thing satisfies that open sentence.) Finally, we assume a ‘one
name, one object’ rule: each name has at most one referent, and
refers to the same object across times and possible worlds, while
each object has at most one name.
Since they are abstract representations, ersatz times contain no
tables and chairs, donkeys and cats, or quarks and electrons. None-
theless, they can be instantiated by concrete times: maximal simul-
taneous hunks of spatially connected stuff that do include tables,
chairs, donkeys, cats, quarks, and electrons as parts. For a concrete
time t to instantiate an ersatz time e, we claim, is simply for e to be
a complete and true description of t. The concrete present instanti-
ates one ersatz time; various concrete past times have instantiated
other ersatz times. Still other ersatz times are not instantiated, and
never have been.
The Growing-Block theorist should claim that concrete times
(past and present) are ordered from earlier to later. We will assume
that this ordering relation is total, transitive, irreflexive, and asym-
metric, and that it has all four properties necessarily. Following
Bourne (2006, 53–5), we can represent possible timelines as sets of
ersatz times ordered by a total, transitive, irreflexive, and asymmet-
ric relation R.1, 2 (The same ersatz time may appear more than once
in the ordering.)
Having defined timelines, we can now explain what it is for a
timeline to be instantiated. The whole of reality instantiates a time-
1
Perceptive readers will note that we have just ruled out a number of apparent
possibilities, including closed time-like curves and branching concrete universes.
We are not at all convinced these apparent possibilities are genuine possibilities. We
believe that our theory could be expanded to accommodate these scenarios, but to do
so here would require adding to an already long and complex paper.
2
In addition to the ordering relation, one might think that there are meaningful
distance relations between ersatz times. Given two times such that one is later than
the other, it makes sense to ask how much later—an hour, a minute, a year? We will
assume that facts about the distances between times supervene on the facts about
which concrete times there are and how those times are ordered, and do not need to
be added to possible timelines as extra information.
The Real Truth about the Unreal Future | 261
1 if veT (f ) = 0
veT (¬f ) =
0 if veT (f ) = 1
future. Where there is no object x such that ϕx, it may be the case
that there will be an object x such that ϕx (once more times have
become instantiated by the Growing-Block), or it may not. In order
to settle the matter, we will need to say more about the future.
We can also fully define the tense operator P, and partially define
the tense operator F.
F is only partially defined for the same reason that Σ and П are only
partially defined. We’ve said nothing yet about how to ascertain
veT(Fϕ) when T fails to contain an e’ such that eRe’ and ve’T(ϕ) = 1. We
do not want to say that veT(Fϕ) is invariably 0 under these circum-
stances, because we want to allow for the possibility that ϕ will
become true in the future that does not yet exist.
We can introduce an additional class of tense operators—one for
each temporal distance d. (We will write the distance from e to e’ as
positive where eRe’, and as negative where e’Re.)
Fd is only partially defined, for the same reason that F was only
partially defined.
Since Σ, П, F, and the Fd operators are only partially defined, the
valuation function as a whole is only partially defined. The fragment
264 | Rachael Briggs and Graeme A. Forbes
Timeline
Relative Absolute
Relative Truth at e in T Truth at t
Time
Absolute Truth in T Truth
The Real Truth about the Unreal Future | 265
3
See Section 3.1.
4
Still other interpretations of validity are possible in the supervaluationist frame-
work we develop in section 2.1; see (Varzi 2007) for a detailed discussion.
266 | Rachael Briggs and Graeme A. Forbes
Whether the concept of absolute truth aligns with any causal con-
cept is a further question, and one that the authors of this essay
disagree about. One of us thinks Forrest’s concept of a causal fris-
son captures something important about the sense in which the
present is ontologically privileged; the other thinks that it is unlikely
to satisfy any of the useful roles carved out for causation in meta-
physics and the philosophy of science.
1.4 Feasibility
So far, timelines have behaved somewhat like the possible worlds
of Robert Merrihew Adams (1974). We should, however, mark sev-
eral important disanalogies between timelines and traditional pos-
sible worlds. First, unlike traditional possible worlds, timelines are
not sets of propositions, but ordered sets of such sets. Second, dis-
cerning what is true at a time (i.e., a set of tensed propositions) in a
timeline, unlike discerning what is true at a traditional possible
world (i.e., a set of tenseless propositions), is not simply a matter of
checking to see which propositions the relevant set contains. Every
proposition contained in a time in a timeline is true at that time in
that timeline, but a proposition may be true (within a timeline) at
times that do not contain it.
Third, unlike traditional possible worlds, timelines stand in inter-
esting parthood relations. In formal terms, we can say that timeline
T is an initial segment of timeline T’ or that T’ is an extension of T, just
in case
Key:
concrete times
ersatz times
earlier later
plete. Finally, there may be timelines that the laws of nature per-
mit, but do not require, to extend into the future. Call these
semi-complete.
We will leave open the question of how to distinguish among the
three types of timelines, the question of whether there are any com-
plete timelines, and the question of whether there are any semi-
complete timelines. We will, however, assume that every timeline
has either a feasible complete extension or a feasible semi-complete
extension—that it is possible to tack on enough time to reach an
acceptable stopping point.
The concepts of completeness and semi-completeness are crucial
because they are linked to the concept of an exhaustive description—
270 | Rachael Briggs and Graeme A. Forbes
5
Even if the actual world is deterministic, we would not want to rule out the pos-
sibility of indeterministic worlds.
272 | Rachael Briggs and Graeme A. Forbes
6
Here and throughout, we ignore the complications caused by unrelated phenom-
ena such as vagueness and the Liar. A theory that addressed these complications
might replace our ersatz times—assumed to be classical—with ersatz times that were
complete and consistent by the lights of some suitable non-classical logic.
The Real Truth about the Unreal Future | 273
(4) seems to say something stronger than (3)—not just that the
future obeys the law of excluded middle, but that it is determined
by the past and the present, at least as far as sea battles are con-
cerned. For (4) is a disjunction, and each of its disjuncts entails that
the future is determined, at least as far as sea battles are concerned.
One could object to the inference from the claim that each of (4)’s
disjuncts entails that the future is determined to the claim that (4)
itself entails that the future is determined. The supervaluationist
view we discuss in section 2.1 takes this objection seriously. Still,
there is at least a prima facie case for thinking that (4) might not be
true, even when (3) is.
(5) is a classical tautology. This suggests that (5) is valid. But to all
three of the proposals we consider, there are situations where the
sentences FdS and ¬FdS are both untrue (that is, not true—either
false or indeterminate). If (5) is valid, then it is sometimes true with-
out having a true disjunct—a puzzling state of affairs!
(6) says simply that a prize will be given out tomorrow. (7)
appears to say something stronger—that the lottery is a fix. The
relationship between (6) and (7) is closely analogous to the relation-
ship between (3) and (4). (3) says that a particular disjunction will
be true tomorrow, while (4) says that one of its disjuncts will be true
tomorrow. Likewise, (6) says that a particular existentially quanti-
fied statement will be true (of something that did, does, or will
exist) tomorrow, while (7) says that there did, does, or will exist
something that will witness that existentially quantified statement
tomorrow. (4) and (7) seem to claim that certain questions (whether
there will be a sea battle, who will win the lottery) are settled, while
(3) and (6) seem not to make such claims.
We have chosen to translate both (6) and (7) using the tenseless
quantifier Σ rather than the tensed quantifier ∃. Our translation of
(6) does not require that the winner of tomorrow’s lottery be some-
one who will exist tomorrow. (For all the logic says, the winner may
exist today but be dead by tomorrow’s drawing, or may not be born
until several years into the future.) Likewise, our translation of (7)
does not require that the future winner of tomorrow’s lottery be
someone who exists today. (For all the logic says, the winner may
be dead already, or not yet born.) The aim of this choice is simply to
avoid the complications that would result from changes in the
domain of quantification.
274 | Rachael Briggs and Graeme A. Forbes
7
Some exceptions will have to be made in cases of self-reference; we ignore these
complications here.
The Real Truth about the Unreal Future | 275
past, the present, and the laws of nature. For Growing-Block theo-
rists, this means that there are (or could be) propositions whose
truth values are not settled at all.
Philosophers who accept (a) must make tradeoffs elsewhere.
First, they must choose between giving up (b) and giving up (c). By
(a), there is some disjunction ϕ ∨ ¬ϕ, neither of whose disjuncts is
true—(5) is such a disjunction under some possible circumstances.
By (b), (5) is sometimes untrue. But since (5) is tautologically valid,
by (c) it is always true. So rejecting either (b) or (c) is mandatory.
The supervaluationist proposal gives up (b) and keeps (c).8 It
treats (3) and (4) as equivalent, and (6) and (7) as equivalent. The
Łukasiewicz and intuitionist proposals, on the other hand, give up
(c) and keep (b). They are able to treat (4) as stronger than (3) and (7)
as stronger than (6).
(d) need not be accepted or denied wholesale, but can be accepted
to a greater or a lesser extent, depending on how important one
thinks it is to assign truth values to as many sentences as possible.
If we treat (d) as a desideratum for theories, the supervaluationist
semantics does better than the intuitionist semantics, which in turn
does better than the Łukasiewicz semantics. All three proposals are
outdone by classical logic, but we think the cost of classical logic for
the Growing-Block theorist—spooky truths without truthmakers—
is excessive.
In addition to the tradeoff between (b) and (c), and the decision
about how seriously to take (d), accepting (a) necessitates a tradeoff
between (e) and (f). In any sufficiently powerful semantics that
allows for truth value gaps, one can formulate propositions for
which Reductio ad Absurdum, Contraposition, and Conditional Proof
fail. (A reminder: these three inferences are as follows:
8
At least, it keeps (c) in the sense that it validates all classically valid single-con-
clusion arguments. Classically valid multiple-conclusion arguments are supervalua-
tionistically invalid.
276 | Rachael Briggs and Graeme A. Forbes
1 if p is true according to e
veh ( p) =
0 else
Notice that so far, the definition of veh looks exactly like the defini-
tion of veT, except that the F operator is fully rather than partially
defined. F(ϕ) is false at e in h whenever it is not true at e in h. Before
we go on defining veh for a larger range of operators and quantifiers,
it is worth pausing to think about how Thomason’s system maps
onto ours.
278 | Rachael Briggs and Graeme A. Forbes
A Timeline and
Two Extensions T1
T
e⬘
such that
Ve⬘T1(S)=1
e
T2
The other connectives can be defined exactly as before, and veT can
be taken to be the smallest fixed point satisfying the recursive defi-
nition of the value function.
This way of defining the connectives yields an extension of the
logic advocated by Jan Łukasiewicz (1970). (Łukasiewicz outlines
definitions for truth-functional connectives and tense operators,
but does not discuss quantifiers.) Unlike the supervaluationist
semantics, the Łukasiewicz proposal treats the connectives as
truth-functional: the truth of a disjunction (existential) will always
be appropriately grounded in the truth of one or more disjuncts
(witnesses).
The benefits of the Łukasiewicz system are not without their
costs, however. In particular, Łukasiewicz’s logic does much worse
than supervaluationism on criterion (d)—it assigns truth values to
far fewer sentences. As a result, it treats fewer sentences as valid—
every classical tautology has untrue instances in the Łukasiewicz
logic, including the law of excluded middle. For instance, (5) is
indeterminate at e in T whenever some but not all semi-complete
extensions of T contain an e’ exactly one day after e at which there is
a sea battle. The situation is not quite as bad as one might expect
from the observation that Łukasiewicz’s truth-functional logic has
no theorems, however. Sentences without F operators or tenseless
quantifiers behave classically, so that every sentence that is both
valid in first-order logic and free of F operators and tenseless quan-
tifiers is a theorem.
Since the Łukasiewicz semantics agrees with the supervalua-
tionist semantics in its treatment of the operator, it agrees with
the supervaluationist semantics in its treatment of (1) and (2).
(1) is indeterminate at T in whenever T‘s feasible semi-complete
extensions disagree about whether there is a sea battle exactly
one day after, while (2) is determinate at every time in every
timeline.
The Łukasiewicz semantics also agrees with the supervaluation-
ist semantics in its treatment of (3) and (6). (3) is true at e in T pro-
vided there will be a future—i.e., provided all of T‘s feasible
complete or semi-complete extensions contain an e’ exactly one
day after e. And (6) is true at e in T whenever a prize will be given
out for the lottery—i.e., whenever all of T’s feasible complete or
The Real Truth about the Unreal Future | 285
The two issues to which we now turn are closely bound up with
the concept of expressive power. In section 3.1, we consider how
Growing-Block theorists might enrich their language to express
the idea of determinate truth. Our results shed further light on the
tradeoffs between (e), (f), (g), and (h) discussed in part 2. In sec-
tion 3.2, we consider the prospects for expressing the Growing-
Block theorist’s ontological commitments in a formal language.
3.1 Determinacy
Our metalinguistic discussion of truth at times in timelines has made
extensive use of the distinction between sentences that are true (at
times in timelines) and sentences that are not true (at times in
timelines)—either because they are false or because they are indeter-
minate. How can the Growing-Block theorist express the thought
that a sentence is true, rather than false or indeterminate, using an
object language expression in one of the idealized languages we
have developed? The usual response to this question is to define a
determinacy operator. How exactly should this be done?
Thomason defines a type of determinacy operator—the inevita-
bility operator L—as follows:
1 if veh’ (f ) = 1 for every h’ in the model structure
veh (Lf ) = containing e
0 else
How should this inevitability operator be understood in terms of
truth at times in timelines?
The question is somewhat thorny. For any sentence not contain-
ing the L operator, we could define truth at a history, and more
290 | Rachael Briggs and Graeme A. Forbes
1 if veT ’ (f ) = 1
veT (Lf ) =
0 else
A Timeline and
Two Extensions T1
T
e⬘
such that
Ve⬘T1(S)=1
e
e* T2
that eRe’ and the atomic sentence S is true at e’; T2 contains no such
e’. Furthermore T has no feasible semi-complete extensions. The
example is depicted in Figure 3.
In this example, veT2 (FS ⊃ LFS) = 0 , so that veT(FS ⊃ LFS) ≠ 1.
Similarly, veT2 (PFS ⊃ PLFS) = 0 , so that veT(PFS ⊃ PLFS) ≠ 1.
But according to our semantics for L, and contrary to Thomason’s
claims, L1–L3 are also invalid. Once again, they are invalid regard-
less of whether we use the intuitionist semantics, the supervalua-
tionist semantics, or the Łukasiewicz semantics. In the example just
considered, veT1 (FS) = 1 , while veT1 (LFS) = 0 ; this constitutes a coun-
terexample to L1 and L2. Similarly, veT1 (FS ⊃ LFS) = 0 ; this consti-
tutes a counterexample to L3.
Although L1–L3 are not valid in general—that is, not valid eT—they
are valid T. Thomason’s model structure runs validity eT and validity
T together. For the fragment of the language without the L operator,
this conflation is harmless: validity eT and validity T coincide. But in a
language containing the L operator, the difference matters.
Despite initial appearances to the contrary, the introduction of L
does not provide counterexamples to Conditional Proof or Contrapo-
sition. Enriching a language with the L operator preserves the hered-
ity property (see Appendix for proof). And in any supervaluationist,
Łukasiewicz, or intuitionist language where all sentences satisfy the
heredity property, Conditional Proof and Contraposition are valid.
Distinguishing between validity T and validity eT also reveals
that L is a poor candidate for a truth operator. Not everything true
at a time is inevitable at that time, as the invalidity of inference L1
292 | Rachael Briggs and Graeme A. Forbes
shows. What is true full stop coincides with what is inevitable at the
absolute present, but it would be a mistake to confuse truth at a
time with inevitability at that time.
We can define a determinacy operator that does a better job of
capturing the idea of determinate truth by making a different choice
about how to pick out the relevant model structure, given a time
and a timeline:
1 if veT (f ) = 1
veT ( Δf ) =
0 else
While Lϕ expresses the thought that ϕ is determinately true at e
from the perspective of the initial segment of the actual timeline
ending at e, Δϕ expresses the thought that ϕ is true at e from the
perspective of the actual timeline. Determinate truth at e and inevi-
tability at e correspond when e is the absolute present, but come
apart when e is in the absolute past.
This definition of the Δ operator creates some trouble for the
supervaluationist and intuitionist proposals, in which certain con-
nectives are defined intensionally rather than truth-functionally.
Sentences containing Δ violate the heredity condition—a sentence
may have a truth value in one timeline without having that truth
value in all of its extensions.
In supervaluationism, the failure of the heredity condition leads
to inconsistent value assignments. In the example shown in Figure 3,
veT1 ( ΔFS ∨ ΔF¬S) = veT2 ( ΔFS ∨ ΔF¬S) = 1 —and so by the supervalu-
ationist inheritance clause, veT(ΔFS ∨ ΔF¬S) should be 1 as well. But
veT(ΔFS) = veT(ΔF¬S) = 0—and so by the semantic rule for disjunc-
tion, veT(ΔFS ∨ ΔF¬S) should be 0.
In intuitionism, the failure of the heredity condition leads to
highly counterintuitive value assignments. In the example shown
in Figure 3, veT(ΔFS) = 0—it is false at e in T that determinately, there
will be a sea battle. Yet veT(¬ΔFS) ≠ 1. This is because at e in T, there
is some feasible timeline—namely T1—according to which ΔFS is
true. Flat-footedly applying the intuitionist semantics yields coun-
terintuitive results.
The trouble can be easily resolved by making the following
adjustment to both the supervaluationist and the intuitionist seman-
tics. To evaluate sentences containing determinacy operators at a
time in T, first set veT ’ ( Δf ) equal to either 1 (if veT(ϕ) = 1) or 0 (other-
The Real Truth about the Unreal Future | 293
what makes ‘It was the case that p‘ true is an actually R-related ordered
triple, whereas according to the tenseless theory, what makes it true is an
actually earlier than-related concrete fact. Now to ask why these ordered
triples are actually R-related is about as fair as asking why the concrete
facts are actually earlier than-related in the tenseless theory, i.e., not at all—
they just are.9
But this misses the point of the objection. Truth should supervene on
being—on the concrete things that tenselessly exist, the properties
and relations those things instantiate, and the laws of nature. Accord-
ing to Bourne’s theory, however, there could be two possible worlds
exactly alike with respect to the concrete things they contained, the
properties and relations those concrete things instantiated, and the
laws of nature, in which the ersatz present was nonetheless R-related
to different ersatz past times. This, we think, grants the abstract
realm of ersatz times far too much spooky autonomy.
The Growing-Block theory does better. Which timeline is actual-
ized is determined by which concrete things tenselessly exist, the
properties and relations those concrete things instantiate, and the
laws of nature. Which complete timelines are feasible are deter-
mined by the whole of concrete reality. There are no brute facts
about which ersatz times are appropriately related to the present:
where concrete reality and the laws of nature are insufficient to
uniquely determine the future, all candidate futures are possible.
9
In the interest of notational uniformity, we have replaced Bourne’s ‘E-related’
with our ‘R-related’.
The Real Truth about the Unreal Future | 297
1. Exactly one day into the future, there will be a sea battle.
(FdS)
We will call the utterer, as he does, Jake. Imagine Jake utters (1) on
a Monday, and the world’s history up to Monday together with the
laws of nature fails to determine whether there will be a sea battle
on Tuesday. MacFarlane thinks that on Monday, we should regard
Jake’s utterance as having an indeterminate truth value. MacFarlane
accepts, as we do, that there can be two possible futures (one on
which there is a sea battle and one on which there isn’t), and no
reason to think that one of these futures is ‘singled out’ as the actual
one. In such a case “symmetry conditions seem to rule out saying
either the utterance is true or that it is false. Thus, it seems, we must
count it as neither true nor false. This is the indeterminacy intui-
tion” (MacFarlane 2003, 323).
Next, imagine it is Tuesday, the day after Jake’s utterance of (1)
on Monday. Imagine also that on this particular Tuesday a sea battle
is raging. MacFarlane argues we should be inclined to agree to the
following:
300 | Rachael Briggs and Graeme A. Forbes
5. CONCLUSION
University of Sydney
University of Sheffield
REFERENCES
Robert Merrihew Adams. Theories of actuality. Noûs, 8(3): 211–31, 1974.
Craig Bourne. A Future for Presentism. Oxford University Press, Oxford,
UK, 2006.
David Braddon-Mitchell. How do we know it is now now? Analysis,
64(3): 199–203, July 2004.
C.D. Broad. Scientific Thought. Kegan Paul, London, 1923.
—— A reply to my critics. In P.A. Schlipp, editor, The Philosophy of C.D.
Broad, pages 711–830. Tudor, New York, 1959.
Michael Dummett. The philosophical basis of intuitionistic logic. In Truth
and Other Enigmas, pages 215–47. Harvard University Press, 1978a.
—— The reality of the past. In Truth and Other Enigmas, pages 358–74.
Harvard University Press, 1978b.
302 | Rachael Briggs and Graeme A. Forbes
0. INTRODUCTION
Presentism is the thesis that everything that exists, exists now. The
view faces a familiar problem. Most are inclined to think that there
are truths about the past. However, if the past does not exist, then
what is it that makes true our talk about the past? To borrow from
Armstrong:
What truthmaker can be provided for the truth <Caesar existed>? The obvi-
ous truthmaker, at least, is Caesar himself. But to allow Caesar as a truthmaker
seems to allow reality to the past, contrary to [presentism]. (2004: 146)1
1. DESIDERATA
* We are very grateful to Karen Bennett, Ross Cameron, Jonathan Curtis, Suki
Finn, Daniel Nolan, Harold Noonan, and Dean Zimmerman for comments and feed-
back on previous drafts.
1
We follow the convention, as Armstrong does, that ‘<p>’ denotes ‘the proposi-
tion that p’.
306 | Jonathan Tallant and David Ingram
1.1 Lucretianism
An adequate solution to the truth-maker problem must, according to
Cameron, satisfy four conditions, which are: (1) presentism; (2) truth-
maker theory (that truths about the past are made true by some exist-
ing element of our ontology); (3) realism about the past – simply, that
there are (evidence-transcendent and objective) truths about the past;
and, (4) we should not posit ‘suspicious’ properties, where a property
is suspicious iff it makes no difference to the present intrinsic nature of
its bearer.2 To illustrate how we might apply these conditions, con-
sider the true proposition <Socrates existed>. The Lucretian claims
that there is a property being-such-as-to-have-contained-Socrates and the
state of affairs of the world instantiating it. This makes true <Socrates
existed>. Thus, the world instantiates present, past-directed proper-
ties under Lucretianism, and these necessitate truths about the past.
Lucretianism therefore satisfies each of (1)–(3). The Lucretian endorses
(1) by affirming that only the present exists. Lucretians endorse (2):
what makes true our talk about the past are present properties, and so
the Lucretian can thereby satisfy (3): there are truths about the past.
But the Lucretian property of being-such-as-to-have-contained-Socrates
tells us nothing about the present intrinsic nature of the bearer, i.e. the
world. According to Cameron, the property being-such-as-to-have-
contained-Socrates makes no difference to the present intrinsic nature
of its bearer. Lucretian properties are therefore suspicious, because
they fall foul of condition (4), and should be rejected.
We are prepared to concede, here, that (4) is plausible enough.
The Lucretian is not obviously compelled to agree. Indeed, so far as
we can tell, nothing in Cameron’s paper compels a Lucretian who
does not feel the force of the ‘suspicion’ intuition to concur with
Cameron’s rejection of Lucretianism.3
2
This is based on Cameron’s view that a property is suspicious iff it fails to satisfy
his principle of Intrinsic Determination: ‘for all objects x and properties F and times t,
if x instantiates F at t, then x has the intrinsic nature at t that it has partly in virtue of
instantiating F at t’ (2011: 61).
3
The seriousness of this charge is discussed in Tallant (2009). Our preferred reason
for rejecting Lucretianism is given by Sanson and Caplan (2010) and Merricks (2007:
136–7): what we might crudely call the ‘aboutness’ objection.
Presentism and Distributional Properties | 307
4
Being-polka-dotted is just one example of an SDP, a colour-distributional property.
Other dialectically respectable SDPs include being-hot-at-one-end-and-cold-at-the-other (a
heat-distributional property), having-a-uniform-density-of-1kg/m3-throughout (a density-
distributional property), and so forth.
5
It is more accurate to say that Flatty is a line at the moment immediately preceding
t*, because at t* Flatty ceases to be. However, for simplicity, we will assume that it is
true of Flatty that he was a line at t*, and then ceases.
308 | Jonathan Tallant and David Ingram
6
Cameron (2011: 76–7) considers this problem as his ‘Objection 5’.
Presentism and Distributional Properties | 309
7
These TDPs are ‘tenseless’ in the way described.
8
This problem is discussed for presentist ersatzism (e.g. Crisp (2007)) by Tallant
(2010).
310 | Jonathan Tallant and David Ingram
9
Cameron might reply that it is only a change in an object that requires the instan-
tiation of the TDP, and once again attempt to run his argument. We do not see that
this will work: age, surely, is a property of objects.
10
There is no obvious contradiction in the above representation of the change of
TDPs over time. Recall, Cameron forced the contradiction by insisting that change
was simply variation in the tenseless TDP. Since that is not the case, there is no longer
any contradiction.
11
Elsewhere, Cameron (2010: fn.14) makes it clear that he thinks possibility is the
‘default mode’.
312 | Jonathan Tallant and David Ingram
12
We owe this objection to Karen Bennett.
Presentism and Distributional Properties | 313
Cameron might then try to deny that it’s possible for objects to
change with respect to which TDPs they exemplify, but we do not
think Cameron should take this line. We take the view that possibil-
ity is the default status for a proposition that seems conceivable,
and that if Cameron wishes to take the view that some element of
our ontology is necessarily unchanging in some regard, then this
generates particular dialectical requirements. Namely, Cameron
must explain why objects cannot gain or lose TDPs.
To see the explanatory burden more clearly, we think it is worth
considering the claim that objects cannot change with respect to which
TDPs they exemplify, in light of claims typically made about concreta
and abstracta. We naturally think of concreta as capable of undergoing
change. Were we to claim of any concrete object that it is impossible for
it to undergo any change, then we think some explanation must be
forthcoming as to why this is the case. For example, were Cameron to
tell us that there was a concrete object that instantiates the property
‘being green’ and, of necessity, cannot lose this property and/or gain
another colour property, despite the fact that this object can gain and
lose other properties, we should want some explanation of why this is
the case. It would not do to simply assert that this is how things are.
Nor, then, should Cameron merely assert that an object cannot change
which TDP it exemplifies, whilst simultaneously allowing that an
object can change which age property it exemplifies.
Now consider abstracta; putative examples of which include
numbers, sets, and propositions. These entities are often thought to
be necessarily unchanging. Certainly, we do not think that ‘2’ can
undergo any change in its properties over time. However, the con-
trast between objects bearing TDPs and abstracta is marked. Abstracta
are usually thought to exist ‘outside’ time, in some sense or other.
Indeed, they lack any location in time. That abstracta lack temporal
location renders them unchanging. Because abstracta do not reside
at any point along the temporal dimension, we have an explanation
of why it is that they cannot change their properties over time. Some
similar explanation is required if we are to endorse Cameron’s posi-
tion, but it is hard to see how the appeal to abstracta can help Cam-
eron here. For, as already noted, the bearers of TDPs, unlike abstract
objects, have a location in time. As a consequence of these argu-
ments we think that Cameron’s view should be rejected.
University of Nottingham
314 | Jonathan Tallant and David Ingram
REFERENCES
Armstrong, David (2004) Truth and Truthmakers (Cambridge: Cambridge
University Press).
Bigelow, John (1996) ‘Presentism and properties’ Philosophical Perspectives
10, 35–52.
Cameron, Ross (2010) ‘From Humean truthmaker theory to priority mon-
ism’ Noûs 44:1, 178–98.
—— (2011) ‘Truthmaking for presentists’ in Karen Bennett and Dean
W. Zimmerman (eds.) Oxford Studies in Metaphysics, vol. 6 (Oxford:
Oxford University Press), 55–100.
Crisp, Thomas (2007) ‘Presentism and the grounding objection’ Noûs
41:1, 90–109.
Merricks, Trenton (2007) Truth and Ontology (Oxford: Clarendon Press).
Parsons, Josh (2004) ‘Distributional properties’ in Frank Jackson and Gra-
ham Priest (eds.) Lewisian Themes (Oxford: Oxford University Press),
173–180.
Sanson, David and Ben Caplan (2010) ‘The way things were’ Philosophy
and Phenomenological Research 81:1, 24–39.
Tallant, Jonathan (2009) ‘Presentism and truth-making’ Erkenntnis 71:3,
407–16.
—— (2010) ‘Time for presence?’ Philosophia 38:2, 271–80.
NAME INDEX