Professional Documents
Culture Documents
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.
The Analysis Committee, Oxford University Press are collaborating with JSTOR to digitize,
preserve and extend access to Analysis
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
There might be nothing
THOMAS BALDWIN
1. In recent papers Peter van Inwagen (1996) and Jonathan Lowe (1996)
have discussed the 'fundamental'1 question of metaphysics 'Why is there
anything at all?'. In different ways they argue that the nihilist hypothesis
that there might be just nothing can be set aside, either because it is impos-
sible for there to be nothing (Lowe 1996: 118) or because this hypothesis
is 'as improbable as anything can be' (van Inwagen 1996: 99). By contrast
I shall here defend the nihilist hypothesis.
The point at issue does not simply concern the metaphysics of existence.
It also connects with debates concerning modal concepts. David Lewis
explicitly declares 'there isn't any world where there's nothing at all. That
makes it necessary that there is something' (1986: 73). The reason for this,
as Lewis explains, is that because he conceives a world as a maximal
mereological sum of spatiotemporally interrelated things, there cannot be
an empty world, since mereology does not permit 'empty sums'. A little
surprisingly, David Armstrong, whose combinatorial theory of possibility
is in many respects opposed to that of Lewis, also embraces this conclusion,
because 'the empty world is not a construction from our given elements
(actual individuals, properties and relations)' (1989: 93). Armstrong takes
this view despite the fact that his theory permits the construction of repre-
sentations of 'contracted worlds' which lack actual individuals, properties
and relations because he conceives of worlds as maximal states of affairs
and holds that where there is nothing at all, there is no state of affairs. Thus
for both Armstrong and Lewis the nihilist hypothesis is to be rejected
because the conception of a possibility (or world) has sufficient substance,
as a mereological sum or a state of affairs, to demand the existence of
something as a part or constituent.
Despite the agreement of the two Davids on this issue, Peter van Inwagen
himself shows how a contrary position can be developed: he writes 'By a
possible world, we mean simply a complete specification of a way the
World might have been.' (1993: 82) At first this sounds all too reminiscent
of Lewis, and when he adds that by the phrase 'the World' he means 'the
totality of everything there is' it is not so clear that this allows for the possi-
bility of there being nothing at all. But he then proceeds to write of the
World as 'a mere collection' so that 'any use of the phrase "the World" is
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
232 THOMAS BALDWIN
a mere manner of speaking; use of this phrase is no more than a device for
speaking collectively about all individuals' (1993: 112),2 and these
remarks are precisely intended to head off the implication that because the
World has some substance it couldn't be that there is nothing at all. With-
out entering into the metaphysics of modality in any detail, therefore, I
shall assume that a position of the kind adopted by van Inwagen is prima
facie defensible, and thus that the debate on the nihilist hypothesis should
not be regarded as one to be settled outright by the immediate implications
of the concept of possibility.
world wniI
by (A3), the just like Wminofwhose
non-existence these domain lacksnot
things does all require
these objects; and since,
the existence of
anything else, wnil is a world in which there is no concrete object at all.
If one now allows that accessibility between worlds is transitive (the
characteristic S4 assumption), it follows that Wnil is accessible from, or
possible relative to, the actual world. I take this conclusion to be the nihil-
ist hypothesis that was to be established by the subtraction argument. It
may be objected that, at least as far as the subtraction argument goes, the
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
THERE MIGHT BE NOTHING 23 3
domain of wil still includes plenty of abstract objects, such as the natural
numbers, so that its existence cannot properly be regarded as the possibil-
ity of there being nothing at all. This must indeed be conceded; but
whatever view one takes about the existence or not of such objects as the
natural numbers, the focus of most interest (and certainly that of van
Inwagen and Lowe) is on the possibility of there being no 'concrete'
objects, in a sense still to be specified. So it is, I think, legitimate to concen-
trate primarily on this case while remaining, so far as possible, detached
from presumptions concerning the existence or not of abstract objects.
ical objects, x, and x2, the unit sets {xi) and {x2) are not in the same way
exactly similar since they have different intrinsic properties.
With concreteness understood in this way, the claim that there might be
only a finite number of concrete objects looks reasonable. To reject it, one
has to hold, not only that there is in fact an infinite number of individuals
(Russell's Axiom of Infinity), but that there has to be. One might try to
substantiate this thought by invoking the infinite divisibility of space-time;
but regions of space-time do not count as concrete objects by the identity
of indiscernibles test, since although otherwise indistinguishable objects
can be distinguished by their space-time location, space-time regions them-
selves cannot be thus distinguished. So the first premiss of the subtraction
argument appears a reasonable presumption in this context: if the only
way to deny the nihilist hypothesis that there might be nothing was by
Lewis takes this view (1986: 84).
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
234 THOMAS BALDWIN
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
THERE MIGHT BE NOTHING 23 5
whose essence does not determine who in particular she is is less perfect
than a God whose essence does determine this. After all, in the former case,
the God has to be conceived as distinguished by further extrinsic properties
which do not belong to her essence; and a God of this kind is surely less
perfect than a God whose essence ensures her uniqueness and thus rules
out the need for these extra extrinsic properties - at least by the standard
of reasoning which is invoked to argue for the greater perfection of an
existent over a non-existent God. I confess that I do not at present see how
to generalise from this case to a general defence of (B3), mainly because
other plausible grounds for necessary existence are not easy to find. None-
theless, given the familiar deep connections between existence and identity
(as instanced by the Fregean slogan 'no entity without identity'), it does
appear to me that (B3) is a reasonable hypothesis. Hence my Counter-
Ontological Argument has some substance in defence of the claim that
there cannot be a concrete object whose existence is necessary, and thus in
support of the second premiss of the subtraction argument (A2) that
concrete objects are things whose existence is only contingent.
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
236 THOMAS BALDWIN
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
THERE MIGHT BE NOTHING 237
It seems to me, however, that if one holds that the truth of arithmetic
depends on the existence of concrete truth-makers whose existence one
takes to be contingent, one cannot just invoke standard intuitions concern-
ing the necessity of arithmetic in order to dismiss the nihilist hypothesis.
On the contrary it looks more reasonable to reject either the necessity of
number theory or the 'Aristotelian' conception of numbers or the denial of
the empty set. Indeed this last point appears especially pressing: on the
Aristotelian theory of number, the existence of the number 0 demands that
there be at least one 0-membered set, i.e. the null set. But, for Lowe, there
is no such set: so his arithmetic is not that familiar theory to whose intui-
tive necessity he appeals, but a radical revision of it which makes no
reference to the number 0. If Lowe wants to invoke the intuitive necessity
of arithmetic, then the arithmetic should be the standard theory; but, given
his Aristotelian theory of number, he then requires the null set after all and
his argument against the nihilist hypothesis will have collapsed: for given
the standard iterative hierarchy of pure sets, there will be no need for any
concrete objects.
6. So far as I can judge, therefore, there are good enough reasons to accept
the three premisses of the subtraction argument, and, therefore, its nihilist
conclusion. Although van Inwagen also (though not for the same reasons)
accepts that there might be nothing, he attempts to deflate this conclusion
by showing that the possibility that there is nothing has a probability of
zero, because this possible world is merely one among an infinite number
of such worlds all of which are equally probable or, rather, improbable.
It is, perhaps, of some small comfort to the nihilist to note (as Lowe
observes, 1996: 113-4) that, on van Inwagen's reasoning, the actual world
is just as (im)probable as the null world. But, although I have no strong
views about the probability of the null world, which appears to me a ques-
tion for physicists, if anyone, van Inwagen's argument for his conclusion
that this probability is zero is defective. As he acknowledges, it rests on the
thesis that all worlds have the same probability of being actual; and this
thesis is one which we can find good reason not to accept once we consider
the existence of properties which affect the probability of outcomes. There
is, for example, a familiar relationship between age and the chance of
death, arising from the limited durability of such essential organs as the
human respiratory and circulatory systems. Hence, if we compare the
actual world, in which G. E. Moore died in 1958 at the respectable age of
85 with a possible world which is just like the actual world except that
Moore is still alive (in 1996) at the extraordinary age of 123, then we
surely have good reason to think that it is much less probable that this
possible world should have been actual than the world which is, in fact,
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms
238 THOMAS BALDWIN
actual. Of course, within this possible world Moore's physiology will differ
in some respects from that of the actual G. E. Moore; but these differences
will themselves involve improbable properties for a human constitution, so
introducing essential detail of this kind will not diminish the improbability
of the outcome. It may also be said that there is not just one world which
is like the actual world except that it contains a 123-year old G. E. Moore:
there are infinitely many such worlds differentiated by different specifica-
tions of the details of Moore's life in the period 1958-1996, and since these
are all equiprobable, they can have only zero probability, just like the
actual world. But this equiprobability thesis is just a version of that which
is here in question and, intuitively, one can easily think up more or less
probable ways of filling out the details of Moore's life in this period (e.g.
residing in Cambridge versus travelling to Mars and back).
Van Inwagen's basic thought is, I think, that once one grasps that worlds
are maximally determinate states of affairs, so that there is an infinite
number of alternatives to any given world, one should then acknowledge
that all worlds have the same probability, namely zero. It is not clear to me
what kind of non-theological basis there is for any judgment that one
might make concerning the absolute probability of a particular world's
being actual. But even if one were to acknowledge that the actuality of any
one world in particular is, for van Inwagen's kind of reason, improbable,
it still does not follow that one should hold that the probability is the same
for all worlds. For some worlds will involve combinations of properties
and outcomes, such as a 123 year-old Moore, which we know to be much
less probable than other combinations, such as an 84 year-old Moore.4
University of York,
York, Y01 5DD, UK
trb2@york.ac.uk
References
Armstrong, D. 1989. A Combinatorial Theory of Possibility. Cambridge: Cambridge
University Press.
Heidegger, M. 1959. Introduction to Metaphysics. Trans. Manheim, New Haven: Yale
University Press.
Lewis, D. 1986. On the Plurality of Worlds. Oxford: Blackwell.
Lowe, E. J. 1996. Why is there anything at all? Aristotelian Society Supplementary
Volume 70: 111-120.
van Inwagen, P. 1993. Metaphysics. Oxford: Oxford University Press.
van Inwagen, P. 1996. Why is there anything at all? Aristotelian Society Supplementary
Volume 70: 95-110.
4 This is a revised version of my Chairman's comments at the 1996 Joint Session; many
thanks to Timothy Williamson for advice on how to improve them.
This content downloaded from 128.223.86.31 on Tue, 02 Aug 2016 06:44:14 UTC
All use subject to http://about.jstor.org/terms