inclined to do. Despite the protestations of some philosophers, there isn’t much to be
said in favor of absolutely unrestricted composition.^53 The restriction that things
have a mereological sum only if they share a mode of being is neither vague nor
unmotivated, for example.
Another, more radical possibility, is to hold that even an ordinary object can have a
part that differs in mode of being from the whole. Consider an arbitrary way of
dividing me in two. (Merely consider!) Does this arbitrary way of dividing me
correspond to arbitrary proper parts of me? Some philosophers would say“yes.”^54
But perhaps the right thing to say is“yes and no”: this arbitrary division corresponds
topotentialbut notactualparts, and the distinction between potential and actual
objects is a distinction in way of existence.^55 An arbitrary undetached part of me is a
part of me, but it does not exist in the same way as, for example, my hands or my
cells. If this is the right way of thinking of things, then even ordinary wholes do not
always have the same mode of being as their parts.
Let’s return to the problem of persistence over time. Suppose that we reject
perdurantism, perhaps because of the above considerations. As far as I can see, PEP
raises no new difficulties for endurantism. We’ll discuss three endurantist views
about how things change in what follows: the relationalist account, adverbialism,
and a hybrid of the two.
One endurantist account of change, therelationalist account, holds that what we
took to be intrinsic properties are really relations to times.xbears thered-atrelation
tot1, theblue-atrelation tot2, and thegreen-atrelation tot3. Some take the
replacement of intrinsic properties with relations to times to be unacceptable.^56
One variant of relationalism meets this complaint not via replacing intrinsic prop-
erties with relations to times but rather bysupplementingintrinsic properties to
relations to times. On this variant, for any predicate F that something can satisfy at
one time and fail to satisfy at a later time, there are two ways of being F. If one
is presently F, then one is F by virtue of having an intrinsic property of Fness. If one is
merely pastly F, then one was F by virtue of bearing the is-F-at relation to a past time.
On this view, nearly every predicate applicable to temporal beings is analogous.
Another view,adverbialism, claims that, it is not the property that should
be“relativized”but rather the having of it. On this view,x is-in-a-t1ly-wayred,
is-in-a-t2ly-wayblue, andis-in-a-t3ly-waygreen.^57 In some ways, adverbialismfits
(^53) For a defense of unrestricted composition, see Lewis (1991).
(^54) See van Inwagen (1981) for extensive discussion and criticism of what he calls theDoctrine of
Arbitrary Undetached Parts 55. McDaniel (2007) is also relevant.
Holden (2007) provides a masterful account of the history of the metaphysics of actual and potential
parts. The particular way of understanding the distinction suggested here is not one of the ones discussed,
however. Stein (2009: 53 56 – 4) could be interpreted as holding the sort of view I suggest here.
57 See Lewis (1986: 202–4).
See Haslanger (2003) for a discussion of this view. Haslanger distinguishes two theories, which she
callscopula-tensingandadverbialism. Both views take adverbial modification of the copula metaphysically
seriously, but only the former view is explicitly committed to there being an instantiation relation that