Persistence 9783110327052, 9783110326673

Citation preview

Christian Kanzian (Ed.) Persistence

Philosophische Analyse Philosophical Analysis Herausgegeben von / Edited by Herbert Hochberg • Rafael Hüntelmann • Christian Kanzian Richard Schantz • Erwin Tegtmeier Band 21 / Volume 21

Christian Kanzian

Persistence

ontos verlag Frankfurt I Paris I Ebikon I Lancaster I New Brunswick

Bibliographic information published by Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliographie; detailed bibliographic data is available in the Internet at http://dnb.ddb.de

North and South America by Transaction Books Rutgers University Piscataway, NJ 08854-8042 [email protected]

United Kingdom, Eire, Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited White Cross Mills Hightown LANCASTER, LA1 4XS [email protected]

Livraison pour la France et la Belgique: Librairie Philosophique J.Vrin 6, place de la Sorbonne; F-75005 PARIS Tel. +33 (0)1 43 54 03 47; Fax +33 (0)1 43 54 48 18 www.vrin.fr

2008 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm www.ontosverlag.com ISBN 13: 978-3-938793-74-9

2008 No part of this book may be reproduced, stored in retrieval systems or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use of the purchaser of the work

Printed on acid-free paper FSC-certified (Forest Stewardship Council) This hardcover binding meets the International Library standard Printed in Germany by buch bücher dd ag

Table of Contents CHRISTIAN KANZIAN Preface of the Editor

1

STEVE BARKER & PHIL DOWE What is Endurance?

5

ARKADIUSZ CHRUDZIMSKI Enduring States

19

PIERRE GRENON & BARRY SMITH Persistance and Ontological Pluralism

33

LUDGER JANSEN The Diachronic Identity of Social Entities

49

JONATHAN LOWE How Are Identity Conditions Grounded?

73

UWE MEIXNER A Brief Ontology of Spatial and Temporal Localization

91

KRISTIE MILLER Ontology, “Existence” and The Role of Intuition

103

EDMUND RUNGGALDIER Sortal Dependence of Persistence

119

JOHANNA SEIBT Beyond Endurance and Perdurance: Recurrent Dynamics

133

PETER SIMONS The Thread of Persistence

165

ERWIN TEGTMEIER Persistence

185

List of authors

197

Preface The problem of persistence is as old as the tradition of systematic ontology: How can we explain that the middle-sized standard objects of everyday’s life like tables, cats or human beings are regarded normally as remaining “the same”, even if they change their properties and their material constituents? The most radical solutions are that either change is considered as illusionary or that the things stay the same. The first solution is labelled sometimes as “Parmenidean”, the latter as the “Heraclitean” position. (I leave it aside whether this labelling is historically appropriate or not.) The Aristotelean assumption of substances as diachronic identical bearers of properties and of changes may be understood as a kind of “middle” position avoiding the mentioned extremes. The reality of “remaining the same” is guaranteed by substances, the primary beings. They survive the gain and the loss of “accidents”, what grounds ontologically the phenomenon of change. The historical success of Aristotle’s ontology can be explained by its fitting into the conceptual framework of our everyday’s language. We speak about things as if they would remain the same through time, and as if they would be able to change. And we act in everyday’s life as if that would be really the case. Some philosophers say that Aristotle’s ontology is perfectly “intuitive”, alternatives to it “counter-intuitive”. But is intuitivity really a necessary condition of truth for ontological theories? (That it cannot be sufficient seems to be undeniable, also amongst the greatest optimists concerning intuitivity.) Is it not the case that 21st century-ontologists must also take into account the results of modern natural sciences? And these results are very clear: the physical basis of our world cannot be covered by the framework of Aristotle’s, or any other Aristotelean substance-ontology. The supporters of Aristotle reply that it is not the task of ontology to provide a physical theory of the world in a more general terminology. Ontology is no “inductive” discipline. It is a philosophical description of the structures of everyday’s world as it is. The investigation into its physical constituents is another enterprise. The “physical-

ists” fight back by insisting that the object of ontology is reality on all its levels, also on the level of the physical basics. The acceptance respectively the denial of substances is in most cases crucial for the understanding of persistence. The supporters of substanceontologies normally commit themselves to the interpretation of persistence which is called “endurance”. Tables, cats and human beings are identical through time in a strict sense. The relation of diachronic identity is primitive, not analysable and thus not derivable from underlying relations. Things are continuants existing in three spatial dimensions, without any temporal extension. They exist at each moment “as a whole”. Physicalists and other sceptics concerning substances deny that. They concede that in everyday’s life we speak about tables, cats and human beings as the same through time. But ontologically considered their identity is “gen-identity”, reducible to more basic relations like spatio-temporal or causal contiguity. Things “perdure” as the sum of numerical different four-dimensional spatio-temporal “phases”. The aim of this edition should be that this standard picture of the current debate about persistence is perhaps informative for the beginning, but shorthanded if we really want to grasp the actual ontological discussion. There is much more than the endurance-perdurance alternative, the acceptance and the denial of continuants. And there are new arguments, new perspectives, and new theoretical backgrounds worth to be considered in the context of our ontological problem. New aspects are brought into the debate (to start alphabetically) by Arkadiusz Chrudzimski who discusses the phenomenon of endurance not only with regard to the category of things, but also to states. Ludger Jansen enlarges the scope of persistent objects to social entities, and discusses their characteristic mode of remaining the same in time. Uwe Meixner provides a discussion of the problem of persistence starting from the topic of spatial and temporal location, arguing for four kinds of objects distinguishable by their being in space and time. Kristie Miller’s article attends to intuitions to which ontologists frequently appeal to. But when is that appeal useful? An answer is given by focusing on our “existential intuitions”. Erwin Tegtmeier’s intention is to reflect the problem of persistence from the perspective of an adequate understanding of temporal relations and the existence of entities in time.

2

That there is more than the simple acceptance respectively the denial of continuants is documented in the articles of Johanna Seibt and Peter Simons. Seibt is reflecting the hidden presuppositions of the debate about persistence and develops, from the perspective of a new process ontological framework, a new account of persistence that aims to avoid the notorious problems of endurance and perdurance. Simons gives a new interpretation of continuants as “invariants” based on underlying gen-identical occurrences. Pierre Grenon & Barry Smith are suggesting a moderate form of ontological pluralism concerning persistence: a combination of threedimensionalism (for substances) and four-dimensionalism (for occurrences). That the persistence of a thing has to do with it’s kind or sort may be seen as a common theme in the articles of Steve Barker & Phil Dowe, Jonathan Lowe, and Edmund Runggaldier. Barker/Dowe give a new account on endurantism referring to the classical concept of “haecceitas”, the individual “whatness” of a thing. Jonathan Lowe asks for the ontological grounds of identity conditions. The answer is that the identity conditions of entities of any kind K are grounded in the essence of K’s, what it is to be a K. Edmund Runggaldier argues against conventionalist, attempts to solve the problem of persistence, and sticks to the sortal dependence of identity to defend an endurantist position. I express my thanks to all of “my” authors here in this volume. I am very happy to have contributors from all over the world, who are leading ontologists representing different positions concerning our topic. I am also grateful to Dr. Hans Burkhardt for the invitation to edit this collection, and, as always, to our publisher ontos represented by Dr. Rafael Hüntelmann. May this book advance the discussion of one of the core problems of ontology, and persist happily in the libraries with as many readers as possible.

Christian Kanzian, February 2007

3

What is Endurance? STEVE BARKER, Nottingham / PHIL DOWE, Brisbane

1. Introduction A long-standing picture about persistence is as follows: There are broadly speaking two rival accounts of the sense in which objects or people persist through time. One account says when a thing persist it exhibits ‘genuine identity’, being ‘strictly’ the self-same entity throughout its life, the other (‘loose and popular identity’) says persistence just involves some stand-in relation, sometimes called ‘genidentity’. We will call these two positions ‘endurantism’ and ‘perdurantism’ respectively (albeit that these terms were introduced as specific, mereological definitions- we ask the reader to bear with us in our use of these terms). Of course there is actually a variety of extant theories, yet these theories can be taken as fitting under one or other of these two headings. We think that, despite whatever weaknesses they may or may not have, perdurantist accounts at least have been clearly defined, provided the definition does not attempt to distinguish perdurantism from endurantism. We shall argue however that (1) this is not the case for endurantism, and that a range of attempts fails to show how endurantism provides a theory of persistence, but that (2) that such an account can be given in terms of haecceities. For the sake of exposition, our discussion assumes fourdimensionalism about time, although we believe the same considerations apply under other conceptions of time. 2. Semantic Definitions For the purposes of this discussion we shall take the task in offering a theory of persistence to be one of saying what persistence amounts to such that people and common sense objects do indeed persist as common sense says they do. If common sense is right then terms such ‘Fred’ or ‘this apple’ refer to things that have determinate persistence conditions. To start with some examples, here are three extant semantic accounts of what the

name of a persisting object or person (or more generally, singular term) refers to: A. the four-dimensional sum of distinct temporal parts connected by a relation of genidentity (Lewis 1986, in part). B. successive distinct three-dimensional items connected by a relation of genidentity (Sider 1997, described by van Inwagen 1982 and 1990). C. successive identical three-dimensional items (Merricks 19951, Van Inwagen 2001). Theories A and B are normally taken to be perdurantist theories, while C is taken to be endurantist. These are semantic theories, telling us what singular terms refer to. We still need to know what metaphysical account of persistence makes sense of these. In particular, we need to know what metaphysical commitments one might need to make sense of theory C. If it turns out the metaphysical commitments underlying each of these accounts is the same, then we would say the distinction between endurantism and perdurantism is merely semantic. On the other hand, if the commitments are different, then the distinction is metaphysical. The debate over genidentity is well known and needs no rehearsal here. The putative components include spatiotemporal continuity, qualitative similarity, and causal relations. We shall refer (admittedly a little loosely) to the range of putative identity-bestowing genidentity relations or components thereof as ‘genidentity relations’. We think that provided claims about ‘distinctness’ are left aside, accounts of persistence in terms of some genidentity relation are at least clear enough in what they claim persistence consists in, whatever their faults may be. Our questions will primarily concern endurantism. We shall argue in the next two sections that standard formal ways of defining endurantism vis a vis perdurantism, namely the relational and mereological definitions, fail to yield an account of what it means to endure.

1

“If a three-dimensional object lasts from one time to another, then there is a threedimensional object existing at one of those times which is literally identical with a three-dimensional object existing at the other.” (Merricks 1995, 525)

6

3. Relational Definitions The relational definitions represent the traditional way to characterise the debate over persistence. They focus on a relation between the appearances of an object at different times in virtue of which the object persists, claiming that the distinction between endurance and perdurance turns on the existence or otherwise of relations of identity. (Our label ‘relational’ applies to accounts that take the relation to be one of identity as well as those that take it to be genidentity, in contrast to Armstrong (1980), who uses the term to apply just to the genidentity account because he doesn’t count identity as a relation.) PR: An object (or person) a at time t1 perdures from t1 to t2 just if a at time t1 is distinct from but genidentical to some object or object part (or person or person part) b at time t2. ER: An object (or person) a at time t1 endures from t1 to t2 just if a at time t1 is identical to some object (or person) b at time t2. (Merricks 1994).2 (Of course if we want to say the object persists through the interval t1 to t2 then we need some attention to intermediate times, but we may leave this point aside.) The term ‘distinct’ just means ‘not identical’, so the intended contrast is between ‘genuine identity’ and the stand-in relation called ‘loose and popular’ identity (Butler), ‘feigned’ identity (Hume) or ‘genidentity’ (Reichenbach). Since the endurantist need not, and in fact we shall argue below should not, deny that some relation of genidentity holds for enduring objects, the metaphysical side of the debate turns on whether or not the relation of genuine identity obtains, and hence what it is. Firstly, the relational definitions miscategorise our three theories (A-C). Theories A and B of Lewis and Sider respectively are usually taken to be perdurantist, while theory C of Merricks/van Inwagen is usually taken to be endurantist. But the relationalist definitions categorise A, and hence David Lewis, as endurantist. For the name ‘Lewis’ refers on Monday through the Monday Lewis-part to the four dimensional Lewis, and on

2

“(I) An enduring object that exists at one time is identical to itself existing at another.” (Merricks 1994, 166).

7

Tuesday through the Tuesday Lewis-part to the same four-dimensional Lewis (van Inwagen 2001, ch 8). So the relation between the thing the name ‘Lewis’ refers to on Monday and the thing it refers to on Tuesday is genuine identity. It’s true that the four-dimensional Lewis exists on Monday and Tuesday “by having temporal parts” at those times, but it exists at those times nonetheless. Hence by ER theory A is endurantist. But more importantly, what is genuine identity? Synonyms such as ‘strict’, ‘literal’ (Loux 1998, 204), ‘deeper’(Armstrong 1980, 68) ‘true’(Sider 1996, 433, n3) and ‘genuine’(van Inwagen 1990, 247) identity clearly do nothing to explain what identity is. Sometimes the term ‘numerical identity’ is used as if that settles what is meant (Loux 1998, 204). It doesn’t, if numerical identity means the sense of identity used for counting, since the “loose” notion of identity cashed out in terms of genidentity is numerical identity in this sense. To see this, suppose John goes in and out of his house several times in a given day, and that no-one else does, and suppose the police ask how many people entered the house that day. The appropriate answer is one, because we take it that the various manifestations of John at different times throughout the day are not to be counted as distinct persons. Perdurantist accounts are expected to give this result. According to David Lewis identity is entirely unproblematic and uncontroversial: identity is the relation that something has to itself and to nothing else: Identity is utterly simple and unproblematic. Everything is identical to itself; nothing is ever identical to anything else except itself. There never is any problem about what makes something identical to itself; nothing can ever fail to be. And there is never any problem about what makes two things identical; two things never can be identical. (Lewis 1996, 192-3)

That this is true of identity we can no doubt grant, but as an account of identity it seems circular and unilluminating, since it appeals to the concepts ‘itself’ and ‘something else’, both of which seem to require a prior notion of identity. Further, as an account of the distinction between endurance and perdurance it seems inadequate. The second half is something all agree to; both identity and appropriate genidentity are relations that nothing has to something else, in their respective persistence-denying senses of ‘something else’. The first half amounts to self-reflexivity. The problem with this is that some versions of the genidentity relation are reflexive. Causation probably isn’t reflexive, but similarity is, and there are also spatiotemporal conditions that are reflexive. So there are genidentity relations

8

that all things—even momentary things— have to themselves and to nothing else. So some genidentity theories at least will count as defining genuine identity. Thus Lewis’ characterisation fails to tell us the meaning of ‘identical’ in ER. Sider writes in the same vein: Yet another poor characterization of the dispute is the claim one sometimes hears that the disagreement is over whether an object at one time is ever "strictly" identical to an object at another. This claim about "strict identity" isn't at all controversial: since everyone agrees that every object is strictly identical with itself, everyone who accepts the basic phenomenon of persistence over time accepts the existence of objects that exist at one time and are strictly identical with objects that exist at other times. (Sider 1997, 198)

But here it seems that ‘strict identity’ has become a stand-in for the general notion of identity through time, especially given that Theory B is Sider’s own view. So our question remains, what is strict identity? Endurantists sometimes characterise identity in terms of Leibniz’ Law: If a and b are identical then they are exactly similar (i.e. share all properties). Can this supply a definition of identity adequate for an endurantist account of persistence? First, Leibniz’ Law is not an analysis of identity, since it gives only a necessary condition for identity. No contemporary endurantist that we know of endorses the stronger version of Leibniz’ Law, that a and b are identical if and only if they are exactly similar, a fact which in itself could be seen as an admission that Leibniz’ Law alone does not characterise endurantism. In any case, as is well known, Leibniz’ Law only causes trouble for endurantists, since as soon as they state it they immediately have to explain how it can be true of persisting things which apparently change, the ‘problem of temporary intrinsics’ (Lewis 1986). One popular answer is that properties are to be relativised to times: there’s no property ‘being bent’ as such, but many related properties such as ‘being-bent-at-t1’, ‘being-bent-att2’ and so on. A major focus of the debate between endurantists and perdurantists concerns whether this or parallel strategies can work. The having of a property-at-a-time is a tenseless matter. But we may well ask what properties does a have at t? The property relativizer cannot say, just those properties relativised to t, because then at t a will have a different set of properties to the set it has at t*, in violation of Leibniz’ Law. The relativizer must say that at t a has all its properties regardless of what time they contain as an argument. On Monday Lewis has the properties ‘bent-on-Monday’ and ‘straight-on-Tuesday’, the same properties he has

9

on Tuesday. But now, if endurantism is taken to involve the notion of identity defined by Leibniz’ Law, it will fail to rule out time-wise gerrymanders (a term introduced in Dowe 2000). Say the name ‘Sp’ refers to Steve Barker at t1, and Phil Dowe at t2. Sp satisfies Leibniz’ Law on a time relativised theory of properties: although Sp is tall at some times and short at others, short-at-t1 and tall-at-t2 are perfectly compatible properties all attaching to Sp timelessly. Sp has the properties short-at-t1 and tall-at-t2 at t1 and at t2 Therefore Leibniz’ Law together with the thesis that properties are properties at times cannot characterise a notion of identity adequate for defining a theory of persistence. (This argument will no doubt seem too quick for relativizers, but all space allows us to say here is: see Barker and Dowe unpublished.) As far as we can see, the only story left about endurantism’s genuine identity relation is to say that it is primitive. We have nothing to say about this option, except that below we will offer an alternative account according to which it is possible to say what persistence-bestowing identity amounts to. 4. Mereological Definitions On reading certain points in the last section contemporary philosophers may feel that the problem is that we really should be characterising the distinction mereologically. After all that’s how the terms were introduced by Lewis (Lewis 1996 following Johnston 1984). Starting with endurance, according to Lewis “[Something] endures iff it persists by being wholly present at more than one time” (1996, 202). This raises an immediate problem concerning ‘persists’. Is this a standalone necessary condition for endurance? If so, as an account of persistence, endurance is unacceptably circular. ‘Persists’ means ‘exists at various times’, and the problem is: what makes it the case that what exists at these times is the same thing? Whether or not Lewis intends it as a standalone necessary condition for endurance, many (eg Merricks 1999, 424, see below) have formulated endurance so that it clearly is, eg: something endures over a time interval iff it persists over that time interval and it is wholly present at every time that it exists. We acknowledge the need to avoid implying that something endures at times when it doesn’t exist, nevertheless, the account is blatantly circular. It may be replied that endurance is not, and never was a theory of persistence. Rather, it a thesis about persisting things, that they are three di-

10

mensional or similar (Sider 2000). We have no plan to speak to such a thesis about persisting objects (although our comments below may be relevant). Our question is, is endurance an adequate theory of the persistence of common sense objects? In any case, enough people have construed endurance as such an account, including perhaps David Lewis, depending how you read this sentence: “The principle and decisive objection against endurance, as an account of the persistence of ordinary things such as people or puddles, is the problem of temporary intrinsics” (1996, 203, our italics) Perhaps ‘persists’ is not intended as a stand-alone necessary condition for endurance. Perhaps ‘persists by’ is superfluous, or arises (somehow) from being wholly present at more than one time, so that in fact Lewis’ definition is: something endures iff it is wholly present at more than one time. Consider then the following definitions of endurance/perdurance: PM: An object (or person) perdures over a time interval iff it has temporal parts at each time on that interval. EM: An object (or person) endures over a time interval iff it is wholly present at each time on that interval. However, this way of characterising the debate about identity through time immediately faces three problems. First, the mereological definitions are inadequate because they miscategorise accounts of persistence. On the mereological definitions, A is perdurantist while B and C are endurantist. In particular picture B problematically counts as a version of endurantism because it takes objects to be three dimensional and therefore as lacking temporal parts. Proponents of B (e.g. Sider) may or may not accept that the four-dimensional mereological sum of these three dimensional objects exists and indeed is an object in some sense, but if they do they will deny that common sense objects are such four-dimensional sums. Thus the proponent of B denies PM, that common sense objects persist by having temporal parts. While this will not be unexpected news, we note that there is a strong intuition that such accounts should be categorised as perdurantist, presumably because theory B does not appeal to identity. At least, we have that intuition, and it’s worth noting that so too does Lewis, as evidenced by his assertion:

11

... perdurance through time is analogous to the ‘trans-world identity’... of a transworld individual composed of distinct parts in non-overlapping worlds. Perdurance, which I favour for the temporal case, is closer to the counterpart theory which I favour for the modal case; the difference is that counterpart theory concentrates on parts and ignores the trans-world individual composed of them. (Lewis 1996, 203)

Lewis says he subscribes to A, but his counterpart theory is analogous to B. This is clear evidence that he, like us, is inclined to call B a perdurantist theory. Second, by themselves neither EM nor PM can be seriously counted as an account of the everyday concept of persistence of objects or persons. Both leave unanswered the question what makes me now the same person that was here last week. Take perdurance. To give an account of the common sense notion of persistence we need to distinguish persisting objects from time-wise gerrymanders: an account should tell us that the four dimensional object consisting of Steve Barker on Monday and Phil Dowe on Tuesday is not a persisting person. PM does not, and so is at best a necessary condition for persistence. Endurance fares no better. Let the name ‘Sp’ now refer to the three dimensional object Steve Barker at t1 and the three dimensional object Phil Dowe at t2, Sp, clearly a time-wise gerrymander, is wholly present at each time it exists in virtue of the fact that it is a three-dimensional object thereby lacking temporal parts. It’s not just a problem with gerrymanders, for genuine persistence one still wants to ask in virtue of what does genuine persistence obtain. If so, EM also is at best just a necessary condition for persistence. In general, we need to ask what facts of the matter could make it the case that what is wholly present at one time is the same thing that is wholly present at another time. Therefore the mereological definitions, by themselves, are not serious accounts of the everyday concept of persistence. Of course a perdurantist as in theory A claims object parts are connected by the appropriate relation of genidentity, which is lacking in the case of gerrymanders. However, given the difficulties we have encountered for the notion of identity, the corresponding move is not available to the endurantist. Those problems notwithstanding, it seems there is a strong case for saying that the mereological approach to persistence, in the case of endurantism, requires in addition to mereological considerations, the relational notion of strict identity through time. (Lewis seems to take his mereological definition of endurance to involve strict identity, as immedi-

12

ately following the definition of ‘endurance’ he says “Endurance involves overlap: the content of two different times has the enduring thing as a common part.” (Lewis 1996, 202).) Thirdly, and most significantly, the mereological definition of endurance cries out for an account of ‘wholly present’, as has often been pointed out. What does it mean to be ‘wholly present’? A common attempt to explicate this is in terms of ‘having all its parts’ at the relevant time. The endurantist could be either a relativizer, taking ‘having x as a part’ as incomplete requiring specification of the time, or a non-relativizer, taking ‘having x as a part’ simpliciter to be complete. We will take the latter first, by considering an attempt to define wholly present in terms of parts simpliciter. According to Merricks the appropriate way to draw the distinction between endurance and perdurance is as follows (Merricks 1999, 424), formulated so as to be neutral between presentists (who believe only what is present exists) and non-presentists: a) For any presently existing object O, O endures if and only if O persists and all of O’s parts simpliciter exist at the present time. b) For any presently existing object O, O perdures if and only if O persists and some of O’s parts do not exist at the present time. This has three problems. First, it is not a theory of persistence as it appeals to persistence, as discussed above. Second, it runs into a problem with parts that I lose or are yet to gain. Suppose Russell lost his nose at some time in the past. Then there is a part of Russell that does not presently exist. But to acknowledge that is to commit to Merricks’ b, perdurantism. Third, commitment to parts simpliciter hands endurantists a problem similar to the problem of temporary intrinsics, call it the ‘Russell-minus” problem. Suppose at t Russell had a nose, and at t* he doesn’t, and for simplicity that he doesn’t change in any other way. At t Russell is composed of a bunch of parts including a nose, and t* Russell is composed of that same bunch of parts minus the nose. Then the bunch of parts which compose Russell at t* is a proper part of the Russell composed at t. For the endurantist, who holds Russell at t is identical to Russell at t*, this is a contradiction. For reasons such as these Merricks argues that his characterisation of the distinction in terms of (a) and (b) can only work for the presentist, who simply denies that any past thing exists, and this of course includes Russell's nose, and that endurantism therefore entails presentism. This solution

13

cannot work. Take the sentence "Russell was composed of the things which now compose him together with his nose". Is this sentence now true, according to a presentist? Is yes (Bigelow 1996), then we have the same contradiction. ("Russell had Russell as a proper part" where 'Russell' refers to the same thing in both occurrences.) If no, then presumably we have the unusual thesis of logical indeterminacy of the future and past: that sentences about the future and past lack truth values. The immediate problem with this is that it entails that there is no truth of the matter about claims of persistence, which contradicts the premise of the project at hand. So far we’ve considered the non-relativizing endurantist. What if we replace parts simpliciter with parts-at-times, composing-at-times objects or persons? ‘Nose’ is a part-at-t of Russell, ie helps compose-at-t Russell. Then we no longer have the Russell-minus problem. Sure, the parts which compose-at-t* Russell compose-at-t a proper part-at-t of Russell. But the idea would be that a is a proper part-at-t of O is compatible with a composes-at-t* O. But the relativising endurantist has a dilemma (Sider 1997, sec 2). When we say ‘wholly present at t’ means ‘has all its parts at t’ the relativizer believes ‘has its parts’ must be relative to some time. If that time is t then the statement of ‘wholly present’ becomes trivial: ‘wholly present’ at t means ‘has all its parts-at-t at t’. In addition, this is something a perdurantist can agree to given only that they can understand a part-at-t, which Sider argues they can do. But in any case the definition simply fails to rule out something having a temporal part. On the other hand, if ‘wholly present’ at t means ‘has all its parts-atany-times at t’, then it is simply false for common sense objects, which typically lose and gain parts. Here’s some attempts (ours) to get around these problems: a*. O is wholly present at t if and only if all of O’s proper parts-at-t compose-at-t O. Unfortunately this makes four dimensionalist objects wholly present. Suppose O is four dimensional. Divide O in half along the time dimension, and call the earlier half ‘O1’. Similarly, call the earliest 3/4 of O ‘O2’, 4/5 O ‘O3’, etc. There are times at which all these parts of O exist. Consider the infinite series O1, O2, O3, … This series approaches but does not include the non-proper part, O itself. Now sum these proper parts of O: they compose O. Thus we still lack an account of what it is to wholly exist. What about:

14

a**. O wholly present at t if and only if all of O’s non-overlapping proper parts-at-t compose-at-t O. This will not work for the following reason. Consider the infinite set (fusion) of non-overlapping objects as in Figure 1. (t is the time we are concerned with, X is a 4-D object that begins at t* and ends at t**. The horizontal axis is time, the vertical is space.) The first object is 1, which is rectangular. The second object 2 is rectangular with a rectangular hole in the middle, which is taken up by 1. The third object 3 is rectangular in outer perimeter but has a 1+2 sized hole in it. 4 is the object with a 1+2 +3 sized hole in it, and so on. (Translating this into a concrete example, object 1 might be my head from 2nd May to 7th, object 2 my head on the 1st and 8th of May plus my neck from 1st to 8th, etc.)

Figure 1 This is a convergent set: With increasing number n, we get objects whose outer perimeter approach the outer edge of X, but none in itself ever gets there. But there is an infinite set of such objects. Their fusion is just X. And they do not overlap. So they are indeed non-overlapping parts that fuse to form X—they compose X at t. Thus a** fails for reasons similar to the reasons for which a* fails. Then we still have no account of “wholly present”.

5. Haecceities Neither of the concepts we have considered— the mereological or identity relations—either separately or jointly supply sufficient resources for the endurantist to give a theory of persistence of common sense objects or

15

people. We are yet to see an account of how to endure. We conclude that there are at least prima facie reasons for doubting that standard formal modes are sufficient to characterise the endurantist’s position. The problems turn on difficulties in saying what ‘identity’ and ‘wholly present’ amount to. We will show how endurantism can be understood in terms of haecceities – individuating properties. What wholly exists at more than one time, being strictly identity across space and time? There is at least one metaphysical entity that, if one grants its existence, uncontroversially meets these desiderata: immanent universals. Suppose any object n has its unique individuating property Tn, its haecceity. To be object n is to instantiate this property. To instantiate this property at a given time is sufficient for being this object, hence we can say the object wholly exists at that time. We have a straightforward sense in which the object is strictly identical across time: the same property is instantiated. We indicated above that our arguments about identity could be circumvented if one holds that identity is primitive. Properties, if understood as immanent universals, involve a primitive identity (Armstrong 1980). (We leave aside the option of platonic universals.) If understood as tropes properties involve a primitive similarity relation. However, having a primitive identity through time does not exempt one from giving an account of properties, hence our view simplifies the number of primitives. It is sometimes said that a bundle tropist can achieve a simplification of the number of primitives by avoiding primitive identities of properties and individuals. But bundle tropists still must give an account of persistence. In the unlikely event that they want to be endurantist, we hold that the arguments above should lead them to haecceities. Haecceities are often thought of as non-humean. We wish to remain neutral on whether they are or aren’t. We will show how they can be completely humean, but even if they are construed as non-humean, primitive entities, we can show how they might supervene on actual distributions of properties and space-time points. Assume for simplicity 4-dimensionalism about time. We can define worms, space–time regions in which objects are judged to persist and to be spatially located as sub-regions of space-time with the particular propertydistribution holding at that region, plus causal relations. We can admit that worms have temporal stages; they are simple certain space-time regions of the form xyzt, where t is a point, with the property distribution – or part thereof – holding in that region. Once we have worms and stages, we have

16

enough to define the genidentity relation that holds between temporal stages of worms. The genidentity relation – from the philosophically neutral standpoint – is simply a way of pointing to the empirical conditions that speakers search for in making judgements of persistence. Neither perdurantists nor endurantists, insofar as they are analysing the metaphysical nature of persistence, are in the business of analysing, accepting or rejecting, this relation. They differ, of course, on what its role might be in the metaphysics of persistence, but that is another matter. We claim that the debate between perdurantists and endurantists is about is the ontological commitment of such statements, not the empirical warrants for the judgements of persistence. Restrict consideration to only those worms Wi whose stages are related by the right genidentity relations. Take the property T, being a temporal stage of a worm Wn ∈ Wi. Each worm Wi is associated with a unique Ti. Further, T is an individuating property of an object, its haecceity, which wholly exists at each time that the object exists, and in virtue of which the object displays strict identity across time. Here then is our characterisation of the distinction between endurantists and perdurantists. Endurantists are committed to the view that property t is what an object is. Perdurantists deny this. For an endurantist, an object is 3-dimensional. For a perdurantist, an object is either 4-dimensional, identified with the worm, or three dimensional, identified with a time slice of the worm. All views, in order to be taken as an account of the persistence of ordinary objects, must rely on genidentity relations. If one takes the endurantist haecceity to be humean, one takes it to be nothing over and above the worm. If one takes it to be non-humean, it merely supervenes on the worm. In the former case, endurantists and perdurantists agree on what there is, and the distinction is strictly semantic. In the latter case, endurantists and perdurantists agree on what there is, and the distinction is not strictly semantic, turning on the existence or otherwise of the metaphysical, primitive haecceity.

REFERENCES Armstrong, D. 1980 “Identity Through Time”, in P. van Inwagen (ed.), Time and Cause: Essays Presented to Richard Taylor, Dordrecht: Reidel, 67-78. Barker, S. and Dowe, P. unpublished “How to Change” unpublished ms. Bigelow, J. 1996 “Presentism and properties”, Philosophical Perspectives, 10, 35–52. Dowe, P. 2000 Physical Causation, Cambridge: Cambridge University Press.

17

Johnston, M. 1984 Particulars and Persistence, Ph.D.: Princeton. Lewis, D. 1986 On the Plurality of Worlds, Oxford: Blackwell. Loux, M. 1998 Metaphysics: a contemporary introduction, London: Routledge. Merricks, T. 1994 “Endurance and Indiscernibility”, Journal of Philosophy, 91, 165– 84. - 1995 “On the Incompatibility of Enduring and Perduring Entities”, Mind, 104, 523531. Sider, T. 1990 “Four Dimensional Objects” Nous, 24, 245-255. - 1996 “All the World’s A Stage”, Australasian Journal of Philosophy, 74, 433-453. - 1997 “Four Dimensionalism”, Philosophical Review, 106, 197-231. - 2000 “Recent Work on Identity Over Time”, Philosophical Books, 41, 81-89. Van Inwagen, P. 1981 “The Doctrine of Arbitrary Undetached Parts”, Pacific Philosophy Quarterly, 62, 123-37. - 1990 “Four Dimensional Objects”, Nous, 24, 245-255. - 2001 Ontology, Identity and Modality: Essays in Metaphysics, Cambridge: Cambridge University Press.

18

Enduring States ARKADIUSZ CHRUDZIMSKI, Szczecin and Salzburg You know that I love you, And what love endures… (‘Don’t Explain’ by Billie Holliday)

The problem of how a concrete individual survives changes of its properties has long divided the philosophical community into ‘enduratists’ and ‘perduratists’. Enduratists take the idea of a surviving individual ontologically seriously. They claim that many objects we encounter in our everyday (and for that matter also scientific) life endure in time, which means that these entities are wholly present at any time at which they exist. For those who are in principle happy with the conceptual framework of our ‘everyday’ or ‘folk’ ontology it is common to assume that such things as human beings, animals, and plants are endurants in this sense, and the most famous articulation of this view is to be found in Aristotle’s concept of substance. Enduring entities are to be contrasted with perdurants, such as a life of a human being or a process of growing of a plant. Think of the process of writing this very paper. The beginning of the writing, the actual phase of it and the final completion of the paper are not points at which the process of writing could be wholly present. Rather they constitute phases or parts of the process in question. This means that perduring entities have a temporal dimension whereas enduring ones do not. Most of us are prepared to accept that in the world around us there are many entities of this kind, but perduratists try to defend a far stronger thesis. They claim that in fact there are only perdurants. In our everyday language it is equally common to speak of enduring objects and enduring states. But it was the first idiom which mainly attracted philosophers’ attention. Yet in this paper I want to concentrate on the second figure of speech. I will investigate, whether it is ontologically legitimate to distinguish between enduratist and perduratist perspectives with respect to states.

1. What is the problem of enduring states? The concept of state that is relevant for my paper could be defined as the having or exemplifying a property by an individual. On this reading, when we speak of a pain or a love, we mean in fact an individual’s being in pain or in love. A state is thus always a state of a certain individual x and it consists in x’s having a certain property F. Of course, beside ‘monadic’ states there are also states consisting of several individuals standing in a relation. But in the following I will restrict myself to the monadic variety. Everything I say about monadic states can be easily generalised to the relational ones. What it means to have a property will of course depend on the general ontological scheme within which we formulate our theory. It is therefore of no surprise that the answer to our main question, namely whether it is legitimate and important to distinguish enduring states as a distinctive ontological category, will also depend on this scheme. In the following sections I will survey several such schemes, and we will see that only some of them make the talk of enduring states ontologically meaningful, but in this section I have first to state the very problem of enduring states more precisely. Historically, the observation that prompted the whole enduratistperduratist debate was that an (apparently) enduring entity (like a cat or a human being) can have different properties at different times. The difficulty is well known. If it is really the same entity (say a tomato), which is green on Monday and red (and a fortiori not-green) on Tuesday, then this concept of identity seems to violate the so-called Leibniz’s law according to which two entities are identical only if they have all their properties in common, symbolically: ∀x∀y[(x = y) ⊃ ∀φ(φx ≡ φy)]. Two classical solutions of this difficulty are also well known. Enduratists claim that it makes generally no sense to speak of a having a property without reference to the time at which the property is had. Each monadic property becomes thus, in a sense, a relation between an individual and a time. Granted this, the fact that a tomato is green on Monday and red on Tuesday is no more mysterious then the fact that I am stronger then my wife but weaker than Arnold Schwarzenegger. (Cf. Mellor 1998, 90 ff.; Van Inwagen 1990; Howley 2001, 16 ff.)

20

According to the perduratist picture, on the other hand, there is, strictly speaking, no change. Neither complete individuals (construed as temporally extended, four-dimensional ‘worms’), nor their temporal parts change in any understandable sense. The only meaning of change is this: a fourdimensional individual a ‘changes’ in time if it is composed of temporal parts which are qualitatively different. (Cf. e.g. Sider 2001) The dialectical situation which is crucial for the enduratist-perduratist debate in its classical form thus involves a (supposed) identity of a subject with changing properties. It is tacitly assumed that if no property changes then there is no change at all and consequently no problem of identity either. This seems to have rather bad consequences for the concept of an enduring state. If the paradigm situation is really a (supposed) identity in spite of a qualitative change then it seems initially very puzzling what it could mean to say of a state which lasts for a certain time-interval that it does or does not endure. Remember, by supposition we are speaking here of the same property (say F) belonging the same subject (say a). (For if the subject of the state changed then it would definitely make no sense to speak of the identity of the state involving it; and if the involved property changed then we would have another state too.) Explain the identity of the subject in question as you like (in enduratist or perduratist way). Once you grant it and assume that the state under discussion consists of this subject’s having the same property F for some time-interval, it seems that you cannot help but assert that the state in question endures in this time-interval. To be more precise, the only consideration which could eventually incline us to say that the state in question does not endure (but instead perdures), seems to be the perduratist picture of the underlying subject. If you assume that the subjects of the states in question are perdurants, you can argue that it makes no sense to think of such states as enduring, but the reason for this claim lies entirely in the assumed theory of the subject, not in any consideration concerning the property in question. What the above remarks seem to point at is that the only metaphysical point of the distinction between enduring and perduring states seems to be an indirect hint at the nature of the underlying subjects. It seems that perdurance or endurance with respect to states could be at most a characteristic which states inherit from their subjects. Consequently it seems that if we solve the issue of endurance-perdurance for the subjects of states, then

21

there will remain nothing more to say about it with respect to the states themselves. However, the solution won’t be so simple. We will see that under certain metaphysical assumptions it makes a perfectly good sense to differentiate between enduring and perduring states. But before proceeding to the next sections, where this will become apparent, I want to stress the importance of the point made above. The problem of enduring states presented itself as in a sense a mirror image of the problem of enduring individuals. While the problem of endurance for individuals was historically the problem of a possible identity in spite of a qualitative change, the problem of endurance for states appears instead as the problem of a possible diversity in spite of a qualitative constancy. The crucial question will be, therefore, whether or not we are able to make sense of such a ‘qualitatively undetectable’ diversity. 2. Enduring subjects and universal properties The claim that a perduratist with respect to underlying individuals cannot sensibly speak of the endurance of their states seems at first sight very plausible. In the later sections we will see that in point of fact it needs important restrictions, but for the time being let’s begin with this enduratist assumption. So, for the sake of my analysis I temporally take it for granted that there are concrete individuals (most importantly human beings) which are wholly present at any time at which they exist. Imagine such an enduring individual a. For a to be in a state of F-ing means nothing other than to exemplify property F. In this section I assume also that the properties in question are universals and try to formulate the problem of enduring states anew. First consider an individual a changing in time. Within the time-interval from t1 to t2 it is F, and within the time-interval from t3 to t4 it is G. According to the enduratist scheme this means that a stands in relation F to the series of times beginning with t1 and ending with t2 and in relation G to the series of times beginning with t3 and ending with t4. (Remember, within the enduratist framework monadic properties become dyadic relations to times.) Now consider a time interval during which a does not change (e.g. t1– t2). Should we say that we have here an enduring state (of a’s F-ing)? Of course we could say that. After all, we have here not only the same individual, but also a strictly identical property (identical in the strongest

22

metaphysical sense reserved for universals) which are fully present at any time at which they exist. But the problem is that we were in a serious trouble if we had to contrast this enduring state with a possible non-enduring one. Surely we can contrast it with an exemplification of another property, which would be of course a different state. Unfortunately this is all we can do. It seems that within the ontological framework involving endurants and universals there could indeed be only enduring states, which in turn means that within this framework the very adjective ‘enduring’ becomes uninformative when applied to states. This seems to confirm our initial impression that the only thing to say about the enduratist-perduratist status of states is that they automatically inherit it from their bearers. 3. Do we need a distinction between enduring and non-enduring states? But do we really need to differentiate between enduring and non-enduring states? What would be the metaphysical point of a theory which allows us to split the time interval t1–t2, during which a is F, into two (or more) states in spite of their absolute qualitative homogeneity? At first sight this idea looks artificial indeed, but I believe that we sometimes encounter evidence for such a split. Consider an exactly determined kind of pain (P). Assume that it is normally caused by a certain malfunction of a lever which happily could be easily removed within two hours by a certain powerful drug, aptly called ‘P-killer’. Assume that you often suffer from P-pains and for this reason always have some P-killers in your bathroom. Now imagine the following situation: One morning you awake with an excruciating P-pain. You hurry to the bathroom to take your P-killer only to realize that there are no more of them there. What you find is only some multivitamins. Well, you take a multivitamin, as you do every morning, but you know that it cannot help against P-pains. And indeed, after two hours you are still in P-pain. Now, this is not the end of the story. An important fact is that, unknown to you and for absolutely mysterious reasons, your wife put yesterday all remaining P-killers in the empty multivitamin box. The drug you took was thus in fact a P-killer! Your lever is now absolutely in order, but, as you are strongly convinced not having taken the appropriate medicine you still feel a pain of P-type. Needless to say, this P-pain isn’t now caused by your lever but instead by your creative mind.

23

It is clear that similar situations are not only logically possible but also often happen to us. The question is now, what should we say about the ontology of the above story, and most importantly: what exactly happened two hours after the (unconscious) application of a P-killer? The natural cause of the P-pain in question disappeared at this point. The pain itself seems instead to have remained as an ‘imaginary’ (but unfortunately quite really felt) pain. But is it really so? Should we really say that this imaginary pain is the same pain as the lever-caused pain felt before? Or should we rather say that we have here another pain, which in its ‘felt quality’ (i.e. considered as cogitatio in Descartes’ sense) is exactly similar to the preceding ‘real’ pain? In other words: Should we take seriously the kind of transition that Ehring (1997, 95) called a ‘non-salient qualitative change’ – the transition in which one state is replaced by a qualitatively undistinguishable, yet numerically different one? The answer to this question is not quite clear, but what is indubitable is the fact that the ontological framework considered so far (including enduring individuals and universals) forces us to accept the first answer. Within this framework all qualitatively homogenous states are automatically treated as enduring ones. 4. Endurants and tropes An ontological framework that promises more flexibility replaces universals by tropes. If the qualitative content of P-pains is not considered as a repeatable, always the same universal, but instead as an individual property which can be replaced by a qualitatively strictly similar but numerically different one, then we can stipulate that in our case there is a kind of exchange of qualitatively strictly similar tropes at the time-point at which the P-killer has done its job. (Cf. Ehring 1997, 95ff.) Consider a four hour long time interval (t1–t2) beginning with the application of the P-killer and ending four hours later. Exactly in the middle of it there is the point at which the P-killer has removed the physical cause of the pain. Let’s call this point tkill. Now consider two successive intervals t1– tkill and tkill–t2, and think of them as both open at the point tkill. A trope theorist can now assume that there are two exactly similar pain-tropes P1 and P2 occupying intervals t1–tkill and tkill–t2 accordingly. The point tkill should be thought of as a boundary, at which the exchange of tropes takes place. For our purposes we need not care about the details of this picture (and in particular about the exact nature of the boundary point tkill). All that is

24

important is the fact that we have here two enduring pain-states instead of the single state, which was the necessary consequence of the ontology of universals. Within the framework of trope ontology, to speak of enduring states is therefore far from being uninformative. Even if I have a qualitatively homogenous P-pain from t1 to t2, it still makes sense to ask, whether I am during this time in one enduring state or not. 5. Tropes and perduring states A trope theorist can thus distinguish several successive enduring states in cases in which a partisan of universals is forced to postulate only one such state. But this is not the only advantage. For a trope theorist it is possible to formulate a theory to the effect that there are only perduring states, even if he simultaneously accepts enduring individuals as their bearers. An important metaphysical assumption involved in the considerations of the previous section was that tropes are entities that can endure in time. It is, however, not clear, if it is a very natural move for a trope theorist. What usually makes tropes philosophically attractive is the possibility of dispensing with universals. And what makes tropes ontologically less problematic than universals is nothing other than their individuality. What is metaphysically appealing about tropes is that a single trope can be present only ‘in’ one individual. More precisely: a single trope cannot be present in more than one individual, if they do not overlap mereologically. (Cf. Lehrer/McGee 1992, 43) In this fashion tropes are contrasted with universals, considered as entities capable of a mysterious multiple localisation. Now, this multiple localisation, which a trope theorist takes great pains to avoid, is normally construed in spatial terms. Two spatially disjoint tomatoes cannot, according to the tropist picture, have their redness in common. They must have two individually distinguished red-tropes. But the fear of multiple localisation can also be extended to the temporal dimension. Why should we take it to be more problematic to speak of the same colour of two tomatoes than of the same colour that one and the same tomato is supposed to have at various times? Provided, we have already tropes at our disposal, wouldn’t it be wise to suppose that they are momentary entities, non-capable of repetition at different times? To flesh out this idea, let’s assume that the time is an ordering of momentary instants. The details of this conception (in particular the question of whether the ordering of the instants is (i) discrete, (ii) dense, or (iii) continuous) are irrelevant here. In the preceding section we contrasted one

25

continuous P-pain lasting from t1 to t2 with a series of two pains with the same P-quality. Now imagine, still within the same time-interval (t1–t2), a series of many extremely short pains. The idea is that no singular pain of this series lasts longer than one instant. This means in turn that none of the pain-states belonging to this series is an entity which, as numerically the same, is wholly present at more then one time. (To be sure, if we want, we can treat such an instantaneous state as a border-case of an enduring entity. Such a state is of course wholly present at any time at which it exists, but only because it exists only at one singular time-point. But this issue is irrelevant here.) The interesting point is that in many cases we will have good reasons for treating a certain series of such instantaneous states as one singular state. In particular if there are appropriate (causal or other) relations connecting subsequent instantaneous pains (cf. Hawley 2001, 85ff.), then it seems reasonable to say that we have here not a series of mutually independent events, but rather a homogenous singular state. The states of the kind we are talking about last for a certain time interval and consist of appropriately connected momentary states. Consequently they have temporal dimension and the momentary states they consist of can be construed as their temporal parts or phases. In short: the states we are talking about are to be construed as perduring entities. That any reasonable ontology must include a category of such perduring states is relatively clear. A trivial case would be any series of suitably causally connected qualitatively different states. A good example is any pain which increases, decreases, or otherwise changes its intensity or qualitative character in time. But, as we see, a trope theorist can propose a far les trivial theory. If he has strong reasons for assuming that no singular trope can last longer than one instant, he is free to formulate a theory to the effect that each state which lasts longer than one instant is a perduring state (even if its temporal parts are absolutely qualitatively homogenous). 6. Enduring states of perduring individuals In the preceding sections I made a working assumption that it only makes sense to speak of enduring states if we already have enduring bearers at our disposal. But upon further reflection this restriction cannot be accepted without reservation. In this section I will show that it is quite possible to develop an interesting metaphysics assuming perduratism with respect to

26

the concrete individuals, while simultaneously allowing for enduring states. Assume, thus, that we operate within the framework of perduring individuals and tropes, and that tropes are here construed as entities capable of endurance in time. Perduring objects are temporally extended. It means that in general they have temporal parts, but there is still a variety of ways in which they can be divided into such parts. One of these ways (probably the most systematic one) that was discussed in the preceding section divides every four-dimensional worm into ‘equally short’ momentary instants (instantaneous stages). But another way of dividing them is furnished by the intuition that we need to introduce a new temporal part in a spatio-temporal worm only if we have a change in the object. As long as the analysed object remains unchanged, we can assume that we are dealing with still the same temporal part. The difference is this: According to the first picture, even an absolutely unchangeable object has many momentary temporal parts (provided it lasts longer then one moment). According to the second picture a temporally extended object consists of no more temporal parts then the changes it undergoes during its existence. To be exact: n changes generate n+1 temporal parts. An object that never changes thus has only one temporal part. Now take the second perspective, but instead of focusing on the ‘whole’ perduring individual, concentrate on its properties, which are tropes. It will be generally the case, that some of these tropes last longer than the others. If we imagine that a temporal slice of such an individual is simply a bundle of tropes, then our actual metaphysical model can be illustrated by the following diagram:

four-dimensional individual a C

B

A

D

t1

t2

t3

t4

t5

t6

27

We have here a four-dimensional individual a involving four tropes (A, B, C, D). As each change generates a new temporal part and there are no temporal parts without change, a consists of exactly five temporal parts occupying the corresponding time-intervals: t1–t2, t2–t3, t3–t4, t4–t5, and t5–t6. All this is fairly obvious, but an important point is that we are in no way forced to assume that the tropes A, B, C, D, which individual a consists of, do not last longer than a singular temporal part of this individual. We can also assume that a given trope can span several parts of a perduring individual. In particular, it is very natural to assume that the trope D lasts from t1 to t5. If we allow this, we obtain an interesting situation: From the perspective of the whole perduring individual a the time interval t1 to t5 contains no less than four temporal parts. From the perspective of the isolated trope D, the time interval t1 to t5 builds only one temporal part. Yet D is, by supposition, an abstract part of a. This means that the trope in question can be construed as an individual entity (an abstract individual) which as identical the same is fully present in several temporal parts of its individual bearer. We have here an enduring (abstract) part of a perduring individual a. Needless to say, within this framework we can also distinguish the case of a continuous pain from the case of a series of several qualitatively strictly similar pains. Consider the trope C. It is natural to think of it as a single trope enduring over the time-interval t4–t6. But if we need to, we can also divide it into parts, which can even generate new temporal parts of the whole perduring individual a. The ontological resources of the trope theory thus allow us to speak of enduring states, even under the assumption of perduratism with respect to their bearers. 7. States enduring several bearers But with help of the ontological resources of tropes ontology we can still do something more. If we need to, we can also speak of a state which in a sense endures several bearers. Of course this will be possible only ‘in a sense’. As a state has been defined as having a property by a subject, an exchange of the bearer will automatically mean that we are dealing with another state. The sense in which a state can endure several subjects is that the same enduring trope can be thought of as belonging to two or more individuals (of course, not at the same time, since this would make tropes

28

definitely into universals). The situation under discussion can be illustrated by the following diagram:

four-dimensional individual a

four-dimensional individual b

C

A

t1

B

t2

t3

We have here two four-dimensional individuals (a and b). They are very simple, each consisting only of two tropes (tropes A and C, and tropes B and C correspondingly). We assume that the tropes A and B are solely responsible for the identity of the individuals in question. To be sure, this is a simplification; but that is of no relevance for my present purposes. If you have problems with that, you can imagine that the lines A and B symbolise complicated bundles of tropes which constitute the ‘natures’ of individuals a and b, e.g. according to the picture sketched in Simons 1994. All that is important is the fact that the enduring trope C belongs both to a and b. I’m not quite sure if we really need such possibilities, but possibly some science-fiction stories involving brain transplantations could be a good testing field for them. Imagine Smith driving his Ford while being in Ppain. Plagued by this pain he overlooks a red light and causes an accident in which his brain has been totally damaged. In a state hospital a courageous medical student replaces his useless brain by a functioning brain of Jones, who suffered an irreparable cardiac infraction while hunting with Dick Cheney. Assume further that Jones never suffered from P-pains. After the successful operation the patient awakes. How should we refer to him? Surely, it is a matter of controversy, but I believe that the better answer would be: Jones. We can expect that he will have many new experiences due to his new body, but the one that is particularly interesting

29

for us is a new unpleasant feeling of a P-pain produced without interruption by Smith’s still malfunctioning lever. Now, what is the ontology of this story? We have here (by supposition) two succeeding subjects who seem to feel the same pain. Sadly enough, the pain in question seems to have endured despite the death of its initial bearer… I don’t want to claim that this is the only possible analysis of this case and I am not sure if we can find more serious cases in which we can exploit the fascinating possibilities of states enduring several subjects, but in any event, it is nice to have ontological devices enabling us to do this, in case we should need this one day. 8. Conclusion In an earlier article (Chrudzimski 2002) I tried to show that many hopes which are traditionally associated with trope ontology, in particular the hope to use them as truth-makers (cf. Mulligan/Simons/Smith 1984) are highly exaggerated. But there are still some merits to the ontology of individual properties and in this paper I have explored some of them. It seems that the problem of enduring states is a genuine case where trope ontology offers us a far more flexible framework than an ontology operating with universals. I should like to thank the Austrian Foundation for the Promotion of Scientific Research (FWF) for the financial support.

REFERENCES Chrudzimski, A. 2002 “Two Concepts of Trope”, Grazer Philosophische Studien, 64, 137–155. Ehring, D. 1997 Causation and Persistence, New York, Oxford: Oxford University Press. Hawley, K. 2001 How Things Persist, Oxford: Clarendon Press. Lehrer, K./McGee, V. 1992 “Particulars, Individual Qualities, and Universals”, in: Mulligan (ed.) 1992, 37–47. Mellor, D.H. 1998 Real Time II, London: Routledge. Mulligan, K., Simons, P. and Smith, B. 1984 “Truth-makers”, Philosophy and Phenomenological Research, 44, 287–321. Mulligan, K. (ed.) 1992 Language, Truth and Ontology, Dordrecht/Boston/London: Kluwer Academic Publishers.

30

Sider, T. 2001 Four-Dimensionalism. An Ontology of Persistence and Time, Oxford: Clarendon Press. Simons, P. 1994 “Particulars in Particular Clothing: Three Trope Theories of Substance”, Philosophy and Phenomenological Research, 54, No 3, 553–75. Van Inwagen, P. 1990 “Four Dimensional Objects”, Noûs, 24, 245–255.

31

Persistence and Ontological Pluralism PIERRE GRENON / BARRY SMITH Saarbrücken, Genève / Saarbrücken, Buffalo

1. Introduction Consider Fritz, who is nowadays a happy resident of Saarbrücken. On sunny Saturdays, Fritz has the habit of barbecuing Bratwurst with his friends in his yard and going thereafter to his local, the Nächstestrasse Biergarten, ‘NB’ for short. Fritz may have a couple of beers while eating his Bratwurst at home, but nothing which compares to the amount he will drink thereafter at the beer garden. That is why he always goes by foot: he knows that his Blood Alcohol Content, ‘BAC’ hereafter, will not allow him to drive his car back home when the time comes to leave. We can break down Fritz’s typical Saturday in the following way: i) Fritz starts his day with a very low BAC. ii) Fritz stays at home and eats and drinks reasonably during the afternoon. His BAC rises significantly but not dramatically. iii) Fritz goes to NB, where he drinks a less reasonable amount of beer, increasing dramatically his BAC. iv) Fritz goes back home, sleeps and lets his BAC level fall back to a normal level. What have we done here? We have divided up a day according to Fritz’s activities and the changes in Fritz and his BAC. We have mentioned several entities, either by name or by description; for instance, Fritz, the beer garden, Fritz’s BAC, Fritz’s home, Fritz’s activities on this Saturday, which break down into an afternoon at home, a walk to the NB, a sojourn at NB, a walk back. These activities themselves break down into various sub-activities of Bratwurst eating, beer drinking, step-taking, and many others. Not all of these entities are on a par. Surely, all those we have mentioned exist in time. It seems also that they are all entities which last for more than a mere instant. They are temporally extended. But there is a way in which some differ from and a way some resemble each other.

By the end of the day, Fritz has changed in many ways. He has changed because at the start of the day he had neither yet eaten any Bratwurst nor drunk any beer nor walked to the NB, etc. In particular, he has changed because yet another entity, namely his BAC, has changed. In fact the description of Fritz’s typical Saturday was in part an account of the changes in his BAC, and hence in himself, over such a day. The Bratwurst, too, has changed: each sausage went from raw to cooked to going out of existence as a result of being eaten. There have been many qualitative changes (from raw to cooked, from ajeun to repus, from sober to intoxicated) and many spatial changes also (from the fridge to the barbecue to the bun, from the kitchen to the garden, from home to NB, from NB to home). Fritz, his BAC and his tiredness are entities which persist in time and remain numerically identical through the various changes which they undergo. Fritz is the same entity before, during, and after eating Bratwurst and over the whole day, although he has changed during this time. Similarly, his BAC has changed, but it is the same entity which has gone from low to mild and from mild to high on the alcoholemic scale. Fritz has gone from rested to tired, but it is still always Fritz, through and through, one and identical. Some of the other entities which we have mentioned are in fact some of these very changes. For instance, Fritz’s walk from home to NB is one of Fritz’s changes in location. Changes are processes. Taking a step, drinking a beer, spending an evening at the beer garden are examples of entities which are temporally extended. But in contradistinction to Fritz and his BAC, they do not change. It is not that Fritz’s spending an evening at the beergarden is first a sitting, then a beer drinking, and so on and so forth. The spending of the evening lasts the whole evening, the beer drinking lasts a beer, so to speak. The spending of the evening at NB is a complex entity made of many other entities which exist only during some part of the period during which the whole sojourn exists. Thus there is that part of the evening during which Fritz is drinking his sixth beer at the NB. Entities of this second kind do not change, because they are themselves changes, and they are made up of successive (temporal) parts which are also changes. Fritz himself, however, is not made up of temporal parts. He exists throughout the whole evening (and before and after, of course), and there is no entity which would be the transient proxy of Fritz during which he is involved in this or that activity, or during which he or his BAC sustains this or that change. There is no shorter-lived part of Fritz whose existence

34

is restricted to these periods (Inwagen 2000). Fritz’s evening is not a part of Fritz, any more than is Fritz’s youth or Fritz’s passage down the birth canal. Rather, these are parts of Fritz’s life. *** In this paper we aim to provide the ontological grounds for an adequate account of persistence. We defend a perspectivalist, or moderate pluralist, position, according to which some aspects of reality can be accounted for in ontological terms only via partial and mutually complementary ontologies, each one of which captures some relevant aspect of reality. Our thesis here is that this is precisely the sort of ontological account that is needed for the understanding of persistence. Our brand of moderate pluralism is a generic doctrine, which has been applied to the problem of granularity (Smith and Brogaard 2003; Bittner and Smith 2003). Here we will present reasons for adopting the doctrine also as providing a solution to the problem of persistence. We sketch the main categories of possible ontologies for endurance and perdurance without, however, arguing for these in any detail. In particular we do not address the topic of substantivalism versus relationalism about space, time or spacetime in detail, because we think that our approach is compatible with each major variant position on such topics. What we do address in detail is the need for two independent ontologies in any account of persistence. 2. Two Sorts of Entities According to our account of Fritz and of certain portions of his life, there are entities that have continuous existence and a capacity to persist selfidentically through time even while undergoing different sorts of changes. These entities, sometimes called ‘continuants’, come in several kinds; but for the purposes of this paper we will give only a sketch of the most relevant types. On the one hand there are substances and other substantial entities, which are all bearers of property particulars, and on the other hand there are these property particulars themselves – Fritz’s present thirst, for example, or his BAC (Smith 1990). Substances are those material entities that enjoy a certain rounded-offness or natural completeness (Smith and Broogard 2003). In addition to artefacts such as mugs and sausages, organisms are the prototypical examples of substances. The wider category of substantial entities embraces not only substances in this sense, but also

35

their parts and aggregates, the parts and aggregates of these parts and aggregates, and so forth. Substantial entities are not dependent for their continued existence upon other entities; they are the bearers of qualities, and they are subject to qualitative change. Property particulars are endurants requiring a basis in substantial entities in order to exist; they inhere in the latter and this inherence is a form of existential dependence (Husserl 1913/21; Simons 1987). Examples of property particulars include: particularized qualities (the colour of this sausage, the ambient temperature in the NG), functions (for example, the function of this barbecue: to cook food), roles (as customer, as waiter), dispositions (of vegetables, that they are prone to rotting), and so on. Finally, for the sake of expository simplicity, we embrace here a so-called absolutist view of space, sometimes called a container view. According to this view, spatial regions are entities in their own right, which are endurants of such a sort that other endurants can be located at or in them. In addition, the world contains entities traditionally called ‘occurrents’ – entities which occur, which happen to continuants, more familiarly referred to as processes, events, activities, changes, happenings. Occurrents are entities which unfold themselves through a period of time or they are the temporal boundaries of such entities. The former are all bound in time in the way described by Zemach (1970). This means that each portion of the time during which an occurrent occurs can be associated with a corresponding temporal portion of that occurrent. This is because occurrents exist only in their successive temporal parts or phases. Some of the latter – for example, beginnings and endings (the initial and terminal boundaries of processes) – are instantaneous. Following Ingarden (1964), we shall reserve the term ‘process’ for temporally extended occurrents, and ‘event’ for their instantaneous boundaries (which we take to include also the crossings of transition thresholds within processes). Examples are: the beginning of Fritz’s walk; Fritz’s becoming 19 years of age; the detaching of a portion of sausage as a result of Fritz taking a bite out of it; the ceasing to exist of an electron as a result of collision with a positron. On the occurrent side of our ontology we include also the temporal and spatiotemporal regions at which other occurrents are located. Prima facie, the view which seems hereby implied, a view which we believe is in close conformity with common sense, is that of a bicategorial ontology, embracing both continuants and occurrents as entities with equal standing. Common sense, however, is not for us an ultimate justification. Indeed, although we will defend a view that is close to common sense in

36

that it recognizes virtually the same entities, we think nonetheless that a bicategorial ontology is untenable. In mainstream analytical philosophy, the name of the game when it comes to the problem of persistence in time is eliminativism. This forces us to choose between either an ontology of continuants, so-called ‘threedimensionalism’, or an ontology of occurrents, so-called ‘fourdimensionalism’. Following Lewis (1986), the view that continuants are threedimensional is sometimes called ‘endurantism’ (and the way they exist in time is to endure), while the view that continuants are fourdimensional is sometimes called ‘perdurantism’ (and the way they exist in time is to perdure). From our perspective, however, there are in fact four available alternatives from which to choose: 1) threedimensionalism, according to which the correct ontology for handling persistence is a monocategorial ontology of continuants; 2) fourdimensionalism, according to which the correct ontology for handling persistence is a monocategorial ontology of occurrents; 3) bicategorialism, according to which the correct ontology for handling persistence is a bicategorial ontology of continuants and occurrents; 4) two-ontology moderate pluralism, according to which there is no correct and adequate way of handling persistence within a single ontology, but rather two ontologies are needed, which must somehow be made to work in tandem. While each of the eliminativist strategies is found wanting not only when measured against common sense but also when measured against the increasingly important needs of ontology applications, for example in the field of biomedicine (Grenon, Smith and Goldberg 2004), bicategorialism leads to contradictions. These problems do not, however, arise on the fourth option, which is thus the option we favor.

3. Rejection of Solutions based on Single Ontologies 3.1 Rejection of Bicategorialism The reasons for rejecting bicategorialism have to do with the fact that each of the two categories of continuant and of occurrent requires an ontology with a characteristic which conflicts with a characteristic of the ontology required by the other category. There is no single ontology countenancing

37

both continuants and occurrents because the proper treatment of these entities requires different and conflicting treatments of time (Merricks 1995). Reasons for this have to do with the fact that under the view of time required for the proper handling of occurrents, continuants possess incompatible determinations, the countenancing of which leads to contradictions. On the other hand, as we shall see shortly, under the view of time required by an ontology which recognizes continuants, occurrents simply cannot be countenanced. Strictly speaking, we should not say that continuants exist in time, because this expression suggests that continuants are temporal entities which belong to an ontology in which time is an entity in its own right. However, time itself is not a continuant, and thus it lies beyond the scope of the continuant ontology. Persisting continuants exist at multiple successive times, and they persist from one time to the next. From the point of view of ontology, however, they have to be taken, so to speak, one time at a time. We could say that they exist over a time in the sense just explained. But because time does not belong to an ontology of continuants, the contrast between existing in time and over a time or at multiple times cannot be expressed using the terms of such an ontology. We can (at the risk of some metaphor) look upon time as a dimension of location along which continuants can move; but that dimension is ipso facto extraneous to the realm of continuants. From the standpoint of an ontology of continuants, to appeal to time would be at best to appeal to an abstraction. Taking temporality into account on the side of continuants thus requires a multiplicity of time-indexed ontologies of continuants corresponding to the points along this temporal dimension. There is only one Fritz; but to do justice to his persistence through time requires a multiplicity of continuant ontologies. Suppose Fritz now1 has a low BAC, and now2 a high one. It is not possible to account for the change in Fritz if the Fritz who exists now1 is seen as an entity existentially on a par with the Fritz who exists now2. This is because the two entities would have contradictory properties in virtue of the fact that Fritz has changed (Lowe 2002). It is in order to do justice to this fact that continuant ontologies like the one we favour see continuants as existing always at a time. The ontology of continuants is in a sense always exemplified in a synchronic fashion – it is exemplified now at this moment, now at this moment. There is therefore a need for a multiplicity of ontologies of a sort commonly called ‘presentist’. Each such ontology applies to a particular time. These temporally specified continuant ontologies recognize no diachronic relations. Diachronic facts about contin-

38

uants are rather accounted through relations between entities in ontologies with different time indices, and thus temporal considerations concerning continuants are relegated to the meta-ontological level (Smith and Grenon 2004). Continuants gain and lose parts over time. It is thus not determinate, given two continuants, whether one is part of the other, because the relations between the two can change from one time to the next. Parthood as a relation between continuants must, therefore, always be indexed to time. Similarly, continuants change in their properties, and it is not determinate across time whether a given continuant has a given property particular inhering in it unless this inherence itself is indexed by time. This implies, too, that there is no diachronic summation of continuants into larger wholes. Suppose, for example that a sausage has a property particular of being straight at one time and another of being bent at a second time. Summing those entities would make for an aggregate, S, which contains the sausage, s, its straight shape, u, and its bent shape, v. But these particular straight and bent shape qualities are entities which do not, because they cannot, co-exist within a single whole – for when would this whole exist? The problem of accounting for an entity having incompatible properties has been dubbed the problem of temporary intrinsics by Lewis (1986). The fourdimensionalist has a solution to this problem: S itself is seen as having two temporal parts, one which is made of s and u and one which is made of s and v. S itself is therefore an occurrent. The only way to countenance continuants within a single ontology in which they are allowed to exist at different times is therefore to transform them into occurrents. The friend of genuine continuants might be tempted by the idea that property particulars are in fact relations to time. It is true that property particulars inhere in their bearers at a time; but the reconstruction of such facts of inherence in terms of relations to time is not satisfactory from an ontological standpoint. Certainly it provides a solution to the problem of logical contradiction by creating pairs of distinct consistent statements – u inheresnow1 in s, v inheres-now2 in s – but since we know of no ontological ground to these statements we shall persist in our assumption that this ‘solution’ is a mere syntactic trick. In contrast to continuants, occurrents are mereologically determinate over time (indeed over worlds): they have the parts they have as a matter of necessity (Olson, in press). Occurrents exist in time in a way that allows them to be apprehended by a single diachronic ontology (an ontology

39

whose purview is extended in time): they exist in time by having parts located at different times. Their existence in time is mediated by that of their temporal parts, which deputize for them. Occurrents have their properties not merely constantly (i.e. at all times at which they occur); they (or their respective temporal parts) have their properties essentially. Thus occurrents do not change in the way continuants change (Mellor 1981). Here again time is a dimension of location. But in contrast to continuants, whose existence is not specific to a time, occurrents have no degree of freedom with respect to time, they are tied essentially to it – this is so whether time itself is an entity or some sort of abstraction over the order of occurrents. For what is distinctive of the realm of occurrents is that there is a temporal order, built out of temporal relations such as earlier than, whose relata exist on a par in spite of the fact that they exist at different times. If time is an abstraction, it is an abstraction over the order of occurrents. And whether or not it is an abstraction, it still allows the occurrents which are located along it to be systematic segmented into arbitrarily large and small temporal parts. Only entities which are existentially on a par may be summed or aggregated into a whole. We can take, for instance, the sum of the beer and the mug into which it as been poured before Fritz drinks it. But once Fritz has drunk and digested his beer and there remains only the mug, then we may not sum the mug at that time with the no longer existing beer from the earlier time. Surely, we can still sum up Fritz and what remains, if anything, from the beer after it has been digested (some aggregate of scattered molecules that is genidentical to the beer (Lewin 1922)). But it is no longer the beer which we sum with Fritz. Fourdimensionalism goes hand in hand with a view according to which all the entities which exist co-exist; they all exist on a par. For occurrents, this view is, unproblematic. We can create the sum of all the beer drinking done by Fritz in his life, which is an aggregate of temporally separated entities. In contrast, the foregoing considerations to the effect that there is no single diachronic continuant ontology mean that it is some version of the presentist doctrine propounded by Brentano (1976) and Chisholm (1989) which is appropriate for dealing with continuants. The presentist holds that all entities which exist exist at the present time. Presentism and fourdimensionalism are clearly incompatible. The presentist cannot accept the fourdimensionalist view according to which time is an entity that is stretched out, and that there are many times all of which co-exist in the same way that the many marks on a measuring rod co-exist. Times, both

40

instants and intervals, are for the fourdimensionalist parts of time and reality as a whole is a fourdimensional spatiotemporal block universe (Alexander 1920). It is sometimes said that, when tied to presentism, the doctrine of threedimensionalism is not even formulable (Hawley 2001) since in order to formulate temporal assertions (for example to the effect that continuants persist through time), we need to appeal to time itself, which does not belong to the presentist ontology. This means that in order to be formulated threedimensionalism needs a form of eternalism, which allows the successive tagging of specific times as ‘the present’ on the basis of some sort of indexicalist view. This means adding to our presentist ontology a further piece – one that does not fit with the rest of the ontology – in order to accommodate time. This is not in itself an objection against realism about continuants, however; it is merely an objection to standard reductionistic versions of this doctrine. The correlation between two ways of persisting and two incompatible ontologies of time is enough to pose the fundamental problem of the ontology of persistence. Drawing on an intuition of Zemach (1970), one way of rationalizing our view of the inconsistency of the bicategorial view which melds the two ontologies together into one would be as follows: distinct ways of existing generate distinct ontologies. Continuants and occurrents are not simply two categories of entities – or two species of existence – but two radically distinct ways of existing in time. And, if both are to be taken ontologically seriously, then they require distinct and incompatible ontologies. 3.2 Rejection of Monocategorial Eliminativism Contemporary debates on the problem of persistence are standardly formulated in terms of the need for an adjudication between two alternative views of temporal existence. These debates grew out of attempts to account for the persistence of different types of entities through time, but they have turned into a fight over what is the unique way in which all temporal entities persist. This in turn has contributed to the obliteration of the distinction between continuants and occurrents, yielding either eliminativism or one or other form of reconstruction (for example, of continuants as more or less gerrymandered varieties of occurrents). The background motivation is, of course, the widespread tendency on the part of contemporary analytical philosophers to believe that they must maximalize ontological parsimony

41

at all costs. We, in contrast, are not concerned by a motivation of this sort. We are interested not just in the theoretical elegance of ontologies but also in their practical utility. We thus prefer to conceive debates about reduction and reconstruction as belonging to the domain of logical representation rather than to ontology properly conceived. When it comes to ontology, and not least to the application of ontology in various practical domains, we see the need to have both ways of persisting. (The medical ontologist needs both Fritz, and Fritz’s BAC, and the risings and fallings of Fritz’s BAC.) Moreover, in addition to these practical considerations, there is also a theoretical argument to the effect that each of the two just-mentioned views in fact requires the other. Exclusivistic fourdimensionalism holds (in our terms) that only processes exist, and that continuants are special kinds of process-things (Quine 1960). Time is just another dimension, analogous to the three spatial dimensions, and substantial (threedimensional) entities are analyzed away as worms or fibres within the fourdimensional plenum. Strictly speaking there is for the fourdimensionalist no such thing as Fritz. Rather, there exists within the fourdimensional plenum a continuous succession of processes, which are similar in a Fritzing sort of way. Exclusivistic threedimensionalist holds (in our terms) that only continuants exist. This view seeks to eliminate or reduce events and processes by conceiving all reference to such entities as a mere façon de parler about what is acceptable threedimensionalistically. Talk of occurrents is cashed out, logically, as talk of continuants and the properties (‘states’) inhering in them at successive points in time. Time itself is an abstraction over statements concerning these. For us, both varieties of reduction are unacceptable: for processes to stand in the mentioned similarity relation is precisely for them to be a part of Fritz’s life. Conversely, there is more to dynamic reality than a mere succession of snapshots along a temporal continuum and more to change than mere discrepancies along this continuum. At the same time, however, we accept that both varieties of reduction bring important contributions. Fourdimensionalism brings the signal benefit, from the point of view of rigorous ontology, that its ontology is mereologically determinate. This means that, given any a and any b, it is always determinate whether a is part of, or overlaps with, or is disjoint from, or is itself a whole including b. Indeed, we hold that fourdimensionalism is right in everything it says, but only when it is making assertions

42

about those entities – precisely the fourdimensional ones – which fall properly within its purview. The fourdimensionalist thus does indeed offer, from the standpoint of realist perspectivalism, one veridical perspective on reality: the perspective which apprehends processes and the spatiotemporal regions at which they are located. But other, complementary perspectives are required also: namely those threedimensional perspectives which capture, for given instants of time, the substances (and other continuant entities) which exist at and endure through those times. And then not only is it the case that each such instantaneous threedimensionalist ontology also enjoys the same nice feature that it is mereologically determinate, but also that this same feature is preserved when we embrace both the four- and the threedimensional ontologies simultaneously. The main problem with threedimensionalism, when rigorously (which means, roughly, presentistically) conceived, is that it leaves no room for time. The main problem with fourdimensionalism is that it provides no account of those demarcations in reality which we discover to be so important when we need to understand, for example, the facts of biology (associative mating, disease transmission, host-pathogen interaction) – it provides no counterpart of the joints we seem to find in the world around us as we encounter organisms and cells and molecules. Occurrents merge with each other in unconstrainedly kaleidoscopic ways to make larger and smaller perduring wholes. Again: putative unitary occurrents are from the fourdimensionalist perspective nothing more than fiat products of the sort of gerrymandering alluded to by Quine (1960, §36). Not only can occurrents be segmented at will into temporal parts; they can also be summed without restriction to form wholes which ride roughshod over the natural joints in reality (Varzi 2003). Neither of the two views, then, is sufficient by itself. But when merged together in a single ontology, they are inconsistent. The solution to this quandary consists in the recognition of the fact that the two views are complementary. Occurrents are precisely the entities we need in order to obtain a temporal line which can be used to index, and provide an ordering for, our plurality of time-specific presentistic ontologies of continuants. The occurrentist view, on the other hand, is in need of continuants as the bearers of its processes, and as that which (through their spatial and temporal boundaries) provide the support for those segmentations of reality at its joints which yield up Fritz and his grandfather.

43

4. Ontological Pluralism What is the picture so far? There is a need for ontologies of two distinct types, one for continuants, the other for occurrents. For the former, in fact, we need a multiplicity of time-specific ontologies, for the latter one ontology suffices, which can be assumed to encompass all of time. That continuants need to be dealt with at a time, and thus through a plurality of timeindexed ontologies, does not mean that their existence is confined to one time only – as if they would be replaced by a succession of instantaneous slices. Ontologies of continuants are like snapshots. They yield views of continuants at a time, but these continuants pre- and post-exist such snapshots. We thus reject a strong presentist solution, and also reject approaches such as Chisholm’s view of continuants as entia successiva and constituted by a succession of momentary entities (entia per se). It is ontologies in the realm of continuants which are tied to time, not the continuant entities they recognize. The very same continuants reappear in a multiplicity of numerically distinct ontologies, as the very same Fritz can appear in a succession of distinct snapshots. The question then remains as to how, given such a framework, we are to go about unifying the two sorts of ontologies. The first clue is that changes in continuants are themselves occurrents. They are processes which happen to continuants. There is, then, a relation between a continuant and an occurrent when the latter is happening in or to the former. Each continuant participates in a multiplicity of occurrents and each occurrent involves or has as participants one or more continuants. For continuants, there is no temporal mereology (no mereology across time), but only a temporalized mereology (a mereology at a time). Occurrents, on the other hand, can be summed across time. This brings us to the major link between continuants and occurrents. We can define the life of a continuant as the aggregate of the occurrents in which it participates. This definition respects our use of this term in our discussion of Fritz, above. It might be thought that this is a dangerous strategy, because the demarcation of occurrents is wild and unconstrained and so is their aggregation. It is here that we most clearly see at work the interdependency and complementarity of the two sorts of ontologies. It is continuants that, by persisting identically through time and through their participating relations to occurrents, provide for the needed individuation of occurrents (a walk to the beer garden is a walk of somebody in particular to somewhere in particular). Clean joints in the ontology of processes are provided through the well-

44

boundedness of the temporally extended entities which are their participants. But continuants change; in particular they change mereologically. How, then, can the continuant ontologies provide for the needed demarcations in the realm of occurrents? Consider, for instance, the changes that occur in an organism such as Fritz, who is constantly exchanging matter with the outside world. It is precisely in order to do justice to this fact that a framework of numerically distinct ontologies is needed, in each of which we once again enjoy the single benefit that the entities recognized are mereologically determinate. Temporal indexation of statements expressing mereological relations among continuants is achieved via reference to these ontologies. Qualitative changes in continuants, too, can yield non-gerrymandered demarcations within the realm of occurrents: property particulars come into being and they go out of existence. But property particulars may also themselves change; thus Fritz’s BAC changes from low to high and his nose changes from pink to red. It is then not simply the substance or substantial entity (Fritz, his nose) which changes, but rather the corresponding quality. That is, it is not simply that the nose has successively the quality of being pink and then of being red. Rather, the nose has all along its quality of being coloured and it is this quality which changes. Fritz’s BAC, in the same way, endures through time, but it changes over time in its specific determinates. Colour and BAC are determinables (Johnson 1921). Changes in such determinables come out as successions of shorter-lived determinate qualities. The determinable endures through these changes. Qualities too have a life and participate directly in processes and indirectly in the processes in which their bearers participate (Smith and Grenon 2004). What about instantaneous entities? Do they belong to an ontology of continuants or to an ontology of occurrents? The answer is that there are instantaneous entities of different categories on both sides of this divide, which participate in the corresponding ways of being in time. Instantaneous entities which fall within an ontology of continuants might be property particulars which, per accidens, endure only instantaneously – for instance, the property-particular enjoyed momentarily by Fritz, whose temperature is rising continuously, of having a temperature of 93°. Instantaneous entities which fall within an ontology of occurrents are, for example, the beginnings and endings of occurrents. Mereological relations are, at a certain level of generality, category preserving. This means that the parts of an entity in the category of substantials are always themselves substantial enti-

45

ties and similarly for entities in other categories. Similarly, each of our two sorts of ontologies is mereologically cumulative and dissective (Varzi, to appear). In other words, belonging to an ontology of a given sort is a property which is preserved under mereological relations. The parts of occurrents are occurrents; the parts of continuants are continuants. Thus no instantaneous temporal part of any occurrent falls within the scope of an ontology of continuants and no instantaneous continuant falls within the scope of an ontology of occurrents. In particular, space is not an instantaneous slice of spacetime. Finally, processes are dependent on their bearers. But in which sense are they dependent? If processes are dependent on continuants in the way that qualities are dependent on their substantial bearers, then this contradicts the claim that eternalism is true for what concerns the realm of occurrents. For if this were the case, then occurrents would cease to exist with their participants. The solution to this puzzle lies, again, in ontological pluralism. There would be a contradiction if processes and their bearers were in the same ontology. Processes co-exist (and thus allow unrestricted mereological summation) over the largest span of time there is. To each of the points of time within this span there can correspond a temporally indexed ontology of continuants. Dependence, too, is temporally indexed and may thus hold simpliciter at these times. This allows us to solve a problem that arises for the non-pluralist presentist, namely the problem of truthmakers for propositions about the past or for propositions involving putative reference to objects no longer existing, such as the proposition that Fritz is taller than Socrates. For even though Socrates no longer exists when Fritz exists and thus there may be no direct relations between them, their respective lives, on the side of occurrents, do indeed co-exist in the relevant sense (Grenon and Smith 2004). The height of Fritz and the height of Socrates are in the same boat as their respective bearers: they do not coexist. But their lives – the sums of all processes in which these heights, respectively, participate – do co-exist in the relevant sense. We may thus account for the alleged facts of comparison between Socrates and Fritz and for a variety of other problematic transtemporal relations on the continuant side via links between ontologies of the two sorts.

46

5. Conclusion We started out from an account of persistence in time which fits common sense in recognizing both continuants and occurrents. We claimed that a simple treatment in terms of a bicategorial ontology does not work because it employs contradictory accounts of the nature of being in time. At the same time we rejected eliminativism, since both continuantism and occurrentism suffer from critical shortcomings, which become especially conspicious when we attempt to apply ontologies to real-world problems. We suggested instead a form of ontological pluralism: a combination of threedimensionalist and fourdimensionalist perspectives, which are mutually incompatible views of reality only when embraced together within a single ontology. The ontological perspectivalism that is implied here consists in the view that any given domain of reality can be viewed from a number of different ontological perspectives, all of which can have equal claim to veridicality. This does not mean that any and every perspective is equally veridical. For our moderate pluralism is of course constrained by realism – by the view that there is a way reality is, independent of the perspectives (languages, theories, cultures, measuring instruments, ontologies) which we human beings adopt. Acknowledgements This paper was written with financial support from the Wolfgang Paul Program of the Alexander von Humboldt Foundation, the European Union Network of Excellence on Medical Informatics and Semantic Data Mining, the Volkswagen Foundation project Forms of Life, and the National Institutes of Health through the NIH Roadmap for Medical Research Grant 1 U 54 HG004028. Thanks go to Robert Arp and Frederic Tremblay for their helpful comments.

REFERENCES Alexander, S. 1920 Space, Time and Deity, London: Macmillan. Bittner, T. and Smith, B 2003 “A Theory of Granular Partitions”, in Duckham, Goodchild and Worboys (eds.), Foundations of Geographic Information Science, London: Taylor & Francis, 117–151. Brentano, F. 1976 Philosophical Investigations on Space, Time and the Continuum, trans. by B. Smith, London: Croom Helm, 1987.

47

Chisholm, R.M. 1989 On Metaphysics, Minneapolis: University of Minnesota Press. Grenon, P., Smith B., and Goldberg, L. 2004 “Biodynamic Ontology”, in D.M. Pisanelli (ed.), Ontologies in Medicine, Amsterdam: IOS Press, 20-38. Hawley, K. 2001 How Things Persist, Oxford: Clarendon Press. Ingarden, R. 1964 Time and Modes of Being, Springfield, IL: Charles C. Thomas. Johnson, W.E. 1921 Logic, Part I, Cambridge: University Press. Lewin, K. 1922 Der Begriff der Genese in Physik, Biologie und Entwicklungsgeschichte. Eine Untersuchung zur vergleichenden Wissenschaftslehre, Berlin: Borntraeger. Lewis, D. 1986 On the Plurality of Worlds, Oxford: Blackwell. Lowe, E.J. 2002 A Survey of Metaphysics, Oxford: Oxford University Press. Mellor, D.H. 1981 Real Time, Cambridge: Cambridge University Press. Merricks, T. 1995 “On the Incompatibility of Enduring and Perduring Entities”, Mind, 104, 523-31. Olson, E.T. (in press) “The Paradox of Increase”, The Monist. Quine, W.V.O. 1960 Word and Object, Cambridge, MA: The MIT Press. Sider, T. 2001 Four-Dimensionalism: An Ontology of Persistence and Time, Oxford: Clarendon Press. Smith, B. 1990 “Characteristica Universalis”, in K. Mulligan (ed.), Language, Truth and Ontology (Philosophical Studies Series), Dordrecht/Boston/Lancaster: Kluwer, 50-81. - 1997 “On Substances, Accidents and Universals: In Defence of a Constituent Ontology”, Philosophical Papers, 26, 105–127. Smith, B. and Brogaard, B. 2003 “A Unified Theory of Truth and Reference”, Logique et Analyse, 169-170, 49-93. Smith, B. and Grenon, P. 2004 “The Cornucopia of Formal-Ontological Relations”, Dialectica, 58:3, 279-296. van Inwagen, P. 2000 “Temporal Parts and Identity Across Time”, The Monist, 83:3, 43759. Varzi, A. 2003 “Perdurantism, Universalism and Quantifiers”, Australasian Journal of Philosophy, 81, 208-15. - (to appear) “Spatial Reasoning and Ontology: Parts, Wholes, and Locations”, in Aiello, Pratt-Hartmann, and van Benthem (eds.), The Logic of Space, Dordrecht: Kluwer. Zemach, E. 1970 “Four Ontologies”, The Journal of Philosophy, 47, 231-247.

48

The Diachronic Identity of Social Entities LUDGER JANSEN, Rostock

There are entities whose diachronic identity essentially hinges on human 1 decisions. This is the claim I want to argue for in this paper. Cases in question are social collective entities like states or associations. In such cases, questions of identity over time are very much practical questions. To argue for this, I first delineate persistence or diachronic identity from other meanings that the term “identity” often has in social contexts (§ 1). Secondly I point out that, as a matter of fact, the world in which we live does not only contain natural entities, but also social entities (§ 2). Then I use John Searle’s account of social facts as a starting point to answer the question of how social entities come about in the first place and of how they are constituted (§ 3). Based on my discussion of Searle’s analysis, I present a general account of the diachronic identity of social entities (§ 4). Having done this, I turn to consideration of those kinds of social entities to which human beings can belong as members, such as collectives, groups and institutions (§ 5). After having pointed out some peculiar features of the diachronic identity of such entities, I finally argue that there are cases where the diachronic identity of groups depends on decisions of the group. In a way, groups are self-determining with respect to the question of what will count as their past, though there are also external constraints for the diachronic identity of groups. 1. Social Identities and the Identity of the Social The present paper is concerned with the persistence or diachronic identity of social entities. Logically, identity is a dyadic relation that holds between any thing and itself and only between a thing and itself. Identity thus conceived is the ontic counterpart of (and the necessary precondition for) the 1

To my surprise, some have objected that this claim is a truism. To sharpen my case I may add that for the entities in question identity is a matter of decision even with a given sortal and in a linear time order (for example with respect to the already determined past).

2

co-referentiality of terms. Just as co-referentiality is an equivalence relation, so identity is reflexive, symmetric and transitive, too. Within the context of the social, however, the word “identity” is often used to denote a 3 range of different phenomena, which are often not clearly delineated. The most important of these phenomena are feelings of belonging and ascriptions of belonging. In social contexts, the word “identity” is often used for answers to the question: “Where do I belong?” This is not a mere factual question, because one’s identity is often connected with strong emotional reactions, with allegiances to groups, and with the bias to favour co-members of a 4 group to which one belongs. If we talk about identity in this way, we are obviously dealing with a psychological concept. What is relevant for deciding about identities conceived of in this way is the inner perspective of the individual. Seen in this way, a person’s identity is a complex individual property: Everybody has his own identity. Thus we can use the plural here and talk about “identities”, even while there is exactly one equivalence relation of identity. In fact there are as many identities of this kind in the world as there are persons. Another use of the word “identity” fits with a sociological conception of identity. Conceived in this way, an identity answers the question: Who belongs to us? What is relevant here is the perspective of the group on the individual, and often such a perspective is connected with the postulation of a set of typical properties of a member, i.e. with a conception of what the “real” or prototypical member of the group is like. The perspective of the group can, of course, differ from the perspective of the individual itself: On the one hand, a group can ascribe membership to an individual who does not feel connected to the group. On the other hand, a person can feel like she belongs to a group that does not recognise her as a member. Seen in this way, different persons can share the same identity. Thus such “group identity” is a generic property. Nevertheless, the plural “identities” can be applied here, too: The identity of a member of a scientific community is quite different from the identity of, say, a Bavarian. Although these different meanings of the word “identity” are to be kept apart, they are not totally unrelated to each other. The feeling of belonging, 2

The classic text still is Frege 1892. Cf. Niethammer 2000 in connection with Jansen 2005. 4 Cf. Maalouf, 2001; Appiah 2005. 3

50

the ascription of belonging, and the persistence of social entities are strongly connected through causal interdependence. On the one hand, if someone feels like he belongs to a group, he is likely to act in such a way that his belonging to this group will also be ascribed to him by others. On the other hand, if nobody else counts a person to be among the members of a group, it is unlikely (though not impossible) that this person will have or develop a feeling of belonging to this group. But note that these are only 5 probability claims. They are not conceptual necessities. Quite the contrary: our world is full of exceptions to these rules. A closet homosexual, for example, may feel like he belongs to the gay community, even though no one else ever ascribes this affiliation to him. He may, in fact, try hard to act in such a way that no one will ever notice his allegiance to this group. In turn, the persistence of a group is often connected with feelings of belonging and ascriptions of membership. In many cases the lack of such feelings and such ascriptions will lead to the group ceasing to exist. There is, thus, an intimate causal connection between the different brands of identity. Nevertheless, it is important to keep them apart on the conceptual level. 2. The Variety of Entities There are many different entities in the world. Here are two lists of examples: • List 1: cats, oak trees, amoebae, electrons, earthquakes, the morning star, elementary charge, solidity, siblings, the size of a proton. • List 2: European Commissioners, universities, laws, poems, borders, Poland, the European Union, ordinations of priests, Wednesdays, a sabbatical, Manchester United, superiors. According to common-sense, all of these things exist, and many of them are important in our daily life. The examples in these lists differ however in various respects: • Some of the examples in the two lists are individuals: the morning star is an individual as well as Poland, the European Union and Manchester United. • Other items in the lists are named by general terms. They name classes or kinds of things, and may correspond to universals. Cats form a 5

Witness my use of the probability adverbs “likely” and “unlikely”.

51

natural kind, whereas universities and European Commissioners might be said to form social kinds. • Some of the entries in the lists are things that are present as wholes at any time at which they exist. They are called “continuants”, a term introduced by W. E. Johnson, who defines a continuant as “that which continues 6 to exist while its states or relations may be changing” . Similarly, the term “endurer” has been coined for a thing which persists by being wholly pre7 sent at more than one time. • Other entities do not exist as wholes at different times, but rather stretch out, or develop, over time. Following Johnson, these entities are called occurrents. Examples of this kind are events, like earthquakes and ordinations, or periods of time, like Wednesdays or a sabbatical. In another terminology, such an entity is called a “perdurer”. They persist by having different temporal parts, although none of its parts exist at more than one 8 time. • The entities in the list belong to different Aristotelian categories: Solidity is a quality; the size of a proton is a quantity. Superiors and siblings, however, identify relations: nobody is a superior or a sibling for herself, but only with respect to someone else. And for every superior (and for every sibling) there is some person of whom she is not the superior (or the 9 sibling, respectively). There is another crucial distinction, in fact, that separates the examples in list 1 from those in list 2. The examples collected in list 1 are natural entities; they exist independently of any human activity or human intention. The examples in list 2, however, are “social entities”. Some philosophers would deny the existence of social entities, and others would grant 10 existence to certain categories of social entities. The world in which we live, however, obviously consists of both natural and social entities. But what would the world look like if we did not inhabit it? How would the world look without sentient rational beings? What was it like when, long ago in natural history, no human beings existed? It should be clear that the absence of human beings should not affect the entities in list 1. 6

Johnson 1921, 199. Lewis 1986, 202. Lewis credits this distinction to Mark Johnston. 8 Lewis 1986, 202. 9 Cf. Jansen 2005c. 10 E.g., Searle 2003 affirms the existence of social facts but denies the existence of social objects. 7

52

Earthquakes happen and the morning star exists whether there is anyone to take notice of them or not. The entities in list 2, however, would be seriously affected if there were no sentient rational beings. There would be no European Commissioners, no universities, no laws, no poems, no borders, no Poland, no European Union, no ordinations of priests, no sabbatical, no Manchester United, and no superiors. Now, if, in the long course of natural history, there was once a time without social entities, consisting only of natural entities like those in list 1, how could social entities possibly arise within a natural world? 3. The Constitution of Social Entities 3.1 Searle’s account of social reality One answer to this question has been provided by John Searle in his book The Construction of Social Reality.11 Roughly, Searle’s account consists of three elements: brute facts, collective intentionality, and constitutive rules. According to Searle, in constructing social reality we can start from what 12 he calls “brute facts” – from nature as it is, independent of our conceptions of it. The generative factor of social reality within the realm of nature is, in particular, the social human nature: our ability to say “we”, to form thoughts and intentions in the first person plural. According to Searle, our collective abilities are on one and the same scale with, say, the ability of 13 wolves to hunt in packs or the ability of birds to fly in flocks. It is this ability which allows human communities to accept constitutive rules. A constitutive rule does not “regulate an antecedently existing activ14 ity”, but creates “the very possibility of certain activities”. The rule “Ladies first” regulates the activity of passing through doorways, which we can be engaged in without this rule as well, though maybe not as smoothly. The game of chess, however, could not exist without its rules. Any activity that is not governed by (amongst other things) the rule that the king may move only one field at a time would not be chess, but some other game, say pseudo-chess. Social entities, or so Searle claims, need such constitu-

11

Searle 1995. Searle takes this term from Anscombe 1958. 13 Cf. Searle 1995, 37-38. 14 Searle 1995, 27. 12

53

tive rules in order to come into existence. 3.2 Constituting things and things constituted According to Searle, the general form of a constitutive rule is the following “counts as” scheme: X counts as Y in context C. The “Y” in the “counts as” scheme represents the constituted object, i.e. the social thing that comes into existence through this rule. While Searle 15 himself is primarily interested in a theory of social facts, there is no reason why Y should not represent social things like social acts, social events, 16 social qualities or social relations. The “X” represents the constituting object – the object that functions as the bearer of the new status Y. The constituting object can be a natural or a social entity. If it is itself a social entity, it comes into existence through the collective acceptance of another constitutive rule, whose Y-term can again be a social entity. We have, for example, a constitutive rule that establishes that uttering an affirmative speech act like “I do” in front of the registrar counts as a marriage vow. The utterance of “I do” is, of course, itself a social entity governed by constitutive rules: It is a constitutive rule that establishes the social fact that a certain pattern of sound waves, i.e. the sound “I do”, counts as an affirmative speech act. We can thus have a whole cascade of nested constitutive rules, which is a good mirror of the factual complexity of our social sphere. In some cases the X-term refers to tokens, whereas in other cases it refers to types. When we declare a certain river in the countryside to be a border between two states, we refer to a particular river and not to all rivers of a certain type. The status of money, however, is not transferred individually to each bank note, but to a certain type of pieces of paper printed by the issuing bank. Searle’s argument for this is that a bank note that falls behind the printing machine would also be money, even if it would never 17 as a particular be the object of anybody’s intentional attitude.

15

Cf. Searle 2003. Searle himself uses plenty of examples that are not facts. Cf. also Searle 1995, 36 (“social acts”, “social objects”). 17 Cf. Searle 1995, 32. 16

54

Note that the “counts as” scheme does not account for all social entities. For there are, so to speak, “free-standing” social entities Y without a corre18 sponding X that counts as Y. The money in my purse consists of coins and bank notes; pieces of metal and paper, that is, that count as money. But I also have money in my bank account. There are a lot of physical traces of this money: There are numbers on my bank statement, memory traces on the bank’s computers and hard disks, and so on. But neither the numbers on my bank statement nor the hard disk tracks count as my money. Rather, they are signs of it. They are not constituents of my money but representations of it. Thus Searle’s original “counts as” formula cannot account for all social entities. There are also social entities that are constituted by representations of them. Things are even more complicated with my credit card. It neither counts 19 as my money nor represents my money. If you need an argument for this claim, just ask yourself, how much money the card is constituting or representing. Maybe there is an upper limit for the amount of money you can transfer with its help, but that still leaves open the exact amount that will be transferred on a certain occasion. And the card will allow this amount to be transferred again and again. Thus even with an upper limit the card would cumulatively constitute or represent a potentially infinite amount of money. Thus rather than constituting or representing money, the credit card is a document that gives evidence of an agreement between me and the credit card agency to the following effect: When I owe 20 Euro to my petrol station, I can use the credit card to transfer the right to claim 20 Euro from my credit card company. The credit card company, in turn, has the right to claim this amount from me. Thus the credit card itself does not constitute or represent any money. Nor does the swiping of a credit card through a card reader constitute or represent money. It is, however, part of a process that counts as the imposition of the aforementioned deontic powers: the right to reimburse a certain amount of money and my obligation to pay this sum.

18

Cf. Smith 2003a, Smith 2003b, Johansson 2005. Contrary to Searle 2003, 307: “I say that one form that money takes is magnetic traces on computer disks, and another form is credit cards. Strictly speaking, neither of these is money, rather both are different representations of money.”

19

55

3.3 Collective acceptance We can, of course, think of many different constitutive rules and come up with arbitrary formulations of “counts as” phrases. But in order to bring 20 about social entities, it is not formulation that counts, but acceptance: We need to accept the constitutive rules in question. There are two elements here that need discussion: the acceptance-element and the we-element. I will discuss them in turn. First, what exactly is the intentional attitude involved in acceptance of constitutive rules? In the opening passage of The Construction of Social Reality, Searle is a bit sloppy about this, writing that “there are things that 21 exist only because we believe them to exist”. However, further along in his book, Searle says that the attitude in question is the attitude of accep22 tance. Acceptance is a propositional attitude similar to, but distinct from, belief. It differs from belief in various respects: Belief can come in degree, acceptance cannot. Belief is based on evidence, whereas acceptance is rather based on the goals we pursue. Belief aims at truth, while acceptance aims at success. The classical tradition of the ens morale considered the 23 will to be the intentional stance involved in creating a social entity. And indeed there is a voluntaristic element connected with acceptance. We can 24 decide to accept that p, but we cannot decide to believe that p. Second, what is needed to build up a social entity is collective acceptance, i.e. an acceptance that can be expressed using the pronoun of the first person plural, “we”: “We accept that X counts as Y in context C.” In his own account, Searle does not say anything about who actually has to accept the constitutive rules in order to bring social things into exis-

20

In fact, explicit formulation seems neither to be necessary nor sufficient: It is not sufficient, because we can formulate arbitrary rules. It is not necessary, because an implicit acceptance of rules is possible (cf. Searle 1995, ch. 6). 21 Searle 1995, 1. 22 E.g. Searle 1995, 104-106. 23 Cf. Kobusch 1993. 24 Cf. Hakli 2006.

56

tence. In fact, Searle is a methodological individualist about collective in25 tentions: For him, even brains in a vat can have we-intentions. It might well be the case that brains in a vat form such thoughts of collective acceptance, but does this bring about a social entity? More seems to be required for this: Interaction between a plurality of actual persons seems to be at the 26 bottom of anything social. An interaction between imagined persons is not enough. Thus there needs to be a real “we” to be connected with such collective intentions. Thus social entities are mind-dependent in a peculiar way: They come into existence through appropriate intentional attitudes in 27 a plurality of minds. Such a real “we” comprising several minds can either be an informal we-group or a formal we-group. If it is a formal we-group, then there are usually fixed procedures for establishing group-acceptance. Such procedural rules are laid down, e.g., in the constitution of a state or in the by28 laws of an association. But if the “we” in question refers to an informal we-group, no such rules exist. However, there is still a possibility to establish group-acceptance: Group-members can enter mutual obligations (or at 29 least commitments) to accept jointly some proposition p. By means of this procedure, they form a mini-size social contract to accept something collectively, “as a group”. 3.4 Constituents of social entities We have started our investigation on the constitution of social entities with Searle’s “counts as” scheme. In Searle’s picture, a physically localisable entity counts as a simultaneously existing social entity. Paradigmatic cases of this structure are sound waves that count as sentences and pieces of paper that count as bank notes. From the problem of free-standing Y-terms we learned, however, that this is not the whole story. Some social entities exist without there being simultaneously existing physically localisable entities that count as these social entities. The money on my bank account exists, but it is not constituted by the ink on my bank account or by the 25

Cf. Searle 1990, Jansen 2004, Schmid 2005. On Schmid’s book cf. Jansen 2007. Cf. Jansen 2005e. 27 Because several minds are involved in the constitution of social entities, it is sometimes said that they are a “conspirational” matter. Cf. Mumford 2004. 28 Cf. Jansen 2004a. 29 Cf. Gilbert 1989, 1996, 2000. 26

57

magnetic properties of my bank’s hard disk. The ink on my bank account and the magnetic properties of the hard disk do, however, represent my money. In these cases, we are confronted with physically localisable entities that are signs for simultaneously existing social entities. But for the existence of a social entity it is not even necessary to have a simultaneously existing representation. A contract may exist without any documentation or representations of the contract. It comes into existence through a mutual promise of the parties involved. The speech acts of mutual promises bring about a reciprocal obligation of the parties, and after 30 the speech acts are completed, these obligations continue to exist. It is thus possible that physically localisable events constitute social entities that continue to exist in the present even when the constituting events belong to the past. 4. The Persistence of Social Entities In the socio-psychological and sociological literature, as well as in history and cultural studies, it is the psychological or sociological concept of identity that is discussed. My present concern, however, is not primarily with these but with the ontological concept of diachronic identity. I will first give a general account of the diachronic identity of social continuants (§ 4.1) and their ability to persist through change (§ 4.2). I will then shortly turn to social occurrents in order to distinguish diachronic identity from the phenomena of resumption and re-enactment of social events. After this, I will turn to groups, i.e. to social continuants that have human beings as members (§ 5). 4.1 A general account My account above presented social entities as being, in simple cases, constituted by a social status imposed either on natural entities or on lower level social entities. In other cases, social entities are only represented by other entities or constituted by past events that brought about certain obligations. Social entities are thus ontologically dependent on several things which, taken together, are constitutive of them. In the simple cases, they 30

Cf. Reinach 1913. The example of obligations as a free-standing social entity is already mentioned (though not explicitly discussed) in Smith 2003a.

58

depend on the constituting thing in the X-position, and on the status imposed on this X by collective acceptance of the respective constitutive rule. They cease to exist as soon as one of the constitutive elements ceases to exist. Correspondingly, there are two ways to destroy such social things: Either you destroy the constituting X or you destroy the status. For example, in order to diminish the amount of legal tender in the EU, you could collect a certain number of Euro notes and burn them. Alternatively, you could equally well pass a law that withdraws the status of being legal tender from, say, all 20 Euro notes. By contraposition, a bank note continues to be legal tender as long as both the constituting X and its status continue to exist. In general, a social entity continues to exist, as long as its constitutive elements continue to exist. If the constituting X is a natural entity, it might still be difficult enough to account for its diachronic identity. This problem is, however, no problem that is specific to the ontology of social entities, and hence I will just presuppose that we have a good working account of the identity of natural entities in order to allow me to focus on the question of how long the status conferred upon X persists. A question that typically arises in this context is: Can a piece of metal on the moon or at some other remote place (e.g., between the cushions of my settee) still be money? To answer this question, we first have to remember that money typically comes into existence through type-related attitudes. We do not have constitutive rules for every single coin, but rather for certain types of coins. Thus, even if we bring a Euro-coin to the moon and subsequently forget about this particular coin, it still retains its status as money, because it did not come about through a token-related attitude in the first place. Secondly, suppose the piece of metal would lose its status of being a coin under such circumstances. What would happen were it accidentally to be recovered from the moon? Surely it would then regain its status. Though losing and gaining status are not intrinsic changes, it is quite strange to assume so many changes of status to be going on with respect to something of whose existence nobody is any no longer aware. Third, if the coin on the moon loses its status after falling into oblivion, what would happen to our money were we suddenly all to fall asleep, simultaneously, for an hour? Would any money exist during this hour? It would be strange to deny this. What we need to uphold a status is not necessarily an active and conscious contemplation of some collective accep-

59

tance. According to Searle, a collective acceptance can even be implicit within a group, without any member of the group knowing explicitly the 31 rule that is accepted within the group. Moreover, once a group’s intention has been established, it can exist even if the members do not share this 32 intention individually. If this is the case, then why should the group’s intention not persist, even if the members of the group do not remember their decision to establish the group-intention? Think about another example. Imagine that one day the loving couple Adam and Eve take their marriage vows. Only a few seconds after the ceremony a very strange cosmic incident wipes out all memories of things that happened the hour before and all relevant documents related to them. In the case of such a global amnesia, would the marriage still exist even if no one remembers it? In an appropriate context, a marriage comes about through a mutual promise of mutual love and care. What happens, from an ontological point of view, when Adam promises Eve that he will care for her? First, Adam produces some sound waves. These sound waves count as uttering a sentence. This sentence, then, constitutes a speech act of promising. Granted that Eve accepts the promise, an obligation comes into exis33 tence: Adam’s obligation to care for Eve. Adam’s obligation survives even if Eve forgets about it. It makes perfect sense for Adam to explain his actions by saying: “I promised to care for you, did you forget?” And Adam’s obligation survives even if Adam forgets about it. Indeed, we would doubly chide him: first for not caring for Eve and second for forget34 ting about his promise. Obligations, or so it seems, are not ontologically dependent on memories of them: The existence of an obligation neither presupposes actually remembering nor the respective memory, i.e. the potential to remember it. They can exist even if no one has any memory of 35 them. And if obligations can survive oblivion, the bonds of marriage can survive oblivion, too.

31

Cf. again Searle 1995, ch. 6. Cf., e.g., Gilbert 2000, 22. 33 Cf. again Reinach 1913. 34 This demonstrates that (contrary to Margalit 2000, 39) even on the individual level remembering or forgetting may be subject to moral evaluation. 35 After the end of all human consciousness, however, all obligations cease to exist because there will no longer be anyone who could be obliged by them. 32

60

4.2 Persistence through change Some social entities persist not only through time, but also through change. This is a feature commonly attributed to substances as opposed to things of the other Aristotelian categories. Socrates survives his hairs becoming white; while the blackness of these hairs does not survive this change. Some, though not all, social entities exhibit this feature as well. An association may, e.g., transfer one further duty to one of its committees. Here, one and the same committee first did not have a certain duty and then does have it. Hence the committee survived a change of its duties. 36 Thus the committee is substance-like in that it can survive changes. In principle, every social entity can function as an X-term in a constitutive rule, i.e. can be a constitutive element in the constitution of some further social status. But not all changes are as innocent with regard to diachronic identity as the imposition of a new duty. If we look at both bearerentity and status transferred, there are the following four possibilities: (C0) (C1) (C2) (C3)

The same bearer-entity continues to have the same status. The same bearer-entity gets a new status. The same status is transferred to a new bearer-entity. A new bearer-entity gets a new status.

Of these four possibilities, (C0) does not describe a change at all. (C3), on the other hand, does describe a change, but it is a clear case of diversity rather than identity. Thus two possibilities remain: (C1) and (C2). (C1) describes cases like our committee-example: The same bearer-entity, the committee, gets a new status, a new duty. This is, as I said, a case of diachronic identity. (C2), however, is a case of diversity. If the same function of, e.g., presidency is transferred to a distinct bearer, we get a new president. If the same status of being the border between France and Germany is being transferred to another place, we do have a new border. If, however, the same line on the map used to represent a border between states, but now, as the two states have merged, as a border between two counties, we can still speak about the same border that now serves in a new function. 36

I argue more extensively for the substance-likeness of some social entities in Jansen 2005b.

61

4.3 A word on social events So far I dealt with social continuants only. There is a reason for this: A social continuant is a social whole that exists at different times. That is why a social continuant can exist through time, even if it may change some of its accidental properties. However, social events like the ordination of a priest or a soccer match or social times like Wednesdays or a sabbatical are not continuants. They are social occurrents. They need time to happen; a soccer match, for example, needs two 45 minute periods, while a chess game consists of an indeterminate number of successive moves. They have different temporal parts, for example a first and a second half. Of course, plenty of soccer matches happen at different times all over the world. But none of these soccer matches is identical with any other soccer match that happens at a different time and a different place. Once a soccer match is over, it will never happen again. Therefore, there is no application for the concept of diachronic identity with respect to events. There are, however, two other phenomena related to social occurrents that may be worth mentioning: resumption and re-enactment. It may happen that a chess game is not completed on its original day. The game can then be adjourned to the next day, when the two players can meet to resume that very game. What happens on this second day is a resumption: we can, after some interval, go on with a previously begun social event and eventually complete it. The original chess game on the first day together with its completion on the next day can be said to form a single token of a chess game. Resumption, however, differs from diachronic identity. It is non-reflexive, as no event is its own resumption. And due to the arrow of time and its directedness, it is asymmetric: If x is a resumption of y, y is not a resumption of x. Hence it is not an equivalence relation. Rather it is a concatenation of a series of events which together form a single token of a type of social event. Re-enactment, on the other hand, occurs when we begin a new token of an event type with the intention of somehow repeating the original token. E.g., a historical battle might have been so important that people decide to celebrate this event each year through a re-enactment of this famous encounter. Often religious or even secular feasts are re-enactments of some important event, like Christmas is the re-enactment of the birth of Christ

62

and the Jewish Passover feast the re-enactment of Israel’s flight from 37 Egypt. Re-enactment differs from diachronic identity, too. It is nonreflexive, as no event is its own re-enactment. And it is asymmetric: If x is a re-enactment of y, y is no re-enactment of x. Finally, it is not transitive: If x is a re-enactment of y, and y a re-enactment of z, then it does not follow that x is a re-enactment of z, but only that x is a re-enactment of a reenactment of z. Hence, re-enactment is not an equivalence relation either. It rather has the structure of having one original event and many “copies”, with the original event being numerically distinct from each copy and the copies being pairwise numerically distinct from each other as well. With re-enactment, we thus have (possibly many) distinct tokens of the same type of social events. 5. The Persistence of Groups After this general account of the diachronic identity of social entities, I will now turn to a special brand of social entities: those which have human beings as members, like collectives, groups and institutions. They are of special importance, not only because we human beings are members of them, but also because, according to our analysis, all social entities ontologically depend on them: Social entities, or so I said, are ontologically dependent on the acceptance of their status by some collective or group at some time. Groups have a very special feature: Like natural persons, they have an “inner perspective” on their identity. Groups and their members are able to give an account of their identity themselves, while coins, borders and Wednesday cannot do this. I will now survey some peculiar features of the diachronic identity of such entities and then make a suggestion how to deal with these. 5.1 Surviving the exchange of members The first peculiar feature I want to point out is that some groups can survive the exchange of their members. Here is some linguistic evidence for this:

37

Cf. Assmann 1997.

63

(E1) “After 1934, 1938 and 1982, the Italian national team won the world championship for the forth time in 2006.” (E2) “For many centuries, the House of Lords has met regularly in the House of Parliament in Westminster.” Both of these sentences are meaningful and true statements. But not one of the team members of the 1934 match participated in the 2006 championship, and it goes without saying that even the House of Lords has different members in different centuries. Such examples demonstrate that groups can exist much longer than individual human beings, because they can survive the exchange of their members. In this, groups are unlike sets. Sets are defined alone in terms of their elements. If you could change the elements, you would not get the same set with different members, but a wholly different set. In this respect, groups are rather like organisms or artefacts: Organisms can survive the exchange of atoms, molecules, cells or even organs, in the same way that Theseus’ ship can survive the exchange of its planks. Similarly, groups can survive the exchange of their members and are thus able to exist much longer than individual human people. That groups can span over a much longer time than individuals makes it possible for us to speak of “our” history with respect to times way before the 38 beginning of our personal existence. 5.2 Surviving non-existence Even more curious is the fact that groups can survive their own nonexistence. They can have, that is, a gappy existence. A social entity like a 39 university or a soccer team can cease to exist and later be re-established. An association may be abolished at one time and later be re-founded again. A famous example of such a gappy existence is Poland, which ceased to exist as an independent state in 1795 after the Polish territory had been distributed among Austria, Prussia and Russia. Poland regained independence in 1918, only to be carved up for a second time by Germany and the Soviet Union in 1939. Since 1945, there is an independent Polish state again. Historically, it is important that there was a continuous patriotic 38 39

Cf. Jansen, 2005a. Mumford 2004 mentions several examples from British soccer.

64

th

tradition from the Napoleonic times onwards, holding alive during the 19 century the hope for Polish independence. Witness thereof is the mazurka that later became the Polish national anthem, which begins with the words “Poland has not yet succumbed, as long as we remain”. In terms of the identity concepts distinguished at the beginning of this paper, it was always the case that there have been individuals who psychologically retained a Polish identity, who felt they belonged to Poland, although it no longer existed as a state. But as important as this may be historically, it is not an ontological necessity. It could well be the case that there is also a gap concerning the psychological identity, i.e. that there is a time when nobody feels as if he belongs to this passed-away entity, till some renaissance or romantic movement revives such feelings. Even in this case, a social entity can be re-established. To be sure, the possibility of a gappy existence like in the case of Poland is not restricted to social entities. When my clock, during a repair, is disintegrated into its parts, there is good reason to say that it does not exist at this time. When the parts are then integrated again, it would be strange to say that now I am in possession of some new clock that consists of the very same parts as my old one. Rather my old clock regained existence. Similarly, my right fist regains existence every time I move my fingers appropriately, and it loses its existence again when I open my hand. 5.3 Groups decide on their past … How do we know that what happened in Poland in 1918 or 1945 was a reestablishing of the same Poland which had ceased to exist in 1795 or 1939? In the case of Poland, on all four occasions it was a different territory, a different population and a different constitution. Still, in each case we speak about Poland and think of the events as re-establishments rather than the founding of something new. Why? Simply because the Polish say so. It is a re-establishment because they want it to be a re-establishment. How is this possible? Well, not only states are social entities, but also their histories. Even if we “objectively” know what happened in a certain historical period, it is in many cases an open question whether or not this period counts as a part of “our” history. It is not determined by the facts of the past that the events of 1918 and 1945 were both re-establishments of an independent Poland. Each time the Polish could as well have decided not to connect the new state with the old ones. Here, diachronic identity becomes a matter of decision. To be sure, even for groups, diachronic identity is in most cases a fac65

tual matter. If an informal group retains all their members or if a formal group retains its constitution, there will rarely be a dispute about whether it is still the same group as before. In other cases, however, such a dispute can arise. What if an informal group does exchange some members? Is it a new group? Or is it the old group with new members? What if a formal group does change its constitution? Is it the old thing with a new constitution? Or is it a new entity, like in the case of France, which now has its “Fifth Republic”? And what if the members of a collective are elected in regular periods? Do such elections decide on new members for a perpetually existing (say: parliamentary) collective, like the United States Senate? Or does the election constitute a new collective, as in the case of the United States Congress or the German parliament: Currently, we have the th th 16 Bundestag, and the year 2007 will see the establishment of the 110 United States Congress. It is in such moments that groups have a principal choice between two options: One option is to regard two historical periods as an earlier and a later phase of the same group. Another option is to regard the earlier group as a mere precursor to the later group. In many cases, both options can do justice to the facts of history. But if the group chooses the first option, we have a case of diachronic identity, whereas the second option yields a case of non-identity. The reason for this is that a (direct) precursor and a (direct) successor are by definition numerically distinct from that of which they are the precursor or successor. A special case of these relations is the legal institution of the successor in title who takes over rights and liabilities from his precursor. Although the successor-relation is quite similar to the relation of diachronic identity, it is nevertheless distinct from identity and is, in fact, a case of non-identity. This can be seen if we consider the individualistic equivalent of the successor in title. Between organisations, the successor in title plays the role that the heir plays between individuals. Of course, nobody can be her own heir. The heir is of necessity distinct from the person whose heir she is. Analogously, the direct successor in title is of necessity distinct from his precursor. Thus, the direct-successor-relation is irreflexive: Whoever wants to be the next head of department after Jones has to be distinct from Jones. Nor is it symmetric, although it is conceivable that Jones is first followed by Smith as the head of department, but then returns into office, which leads to a situation where both Smith is successor of Jones and Jones is successor of Smith. In fact, this situation is such that Jones is also a successor of himself, albeit not a direct successor. Let us summarise this: Even if it is indisputable that a certain period in 66

history is crucially important for the development of a group, the group can regard such a period of the past either as a part of their own history or as (a part of) the history of some precursor. In deciding between these two options, groups decide on matters of diachronic identity. A good example for groups’ range of choice is German history after the 40 Second World War. At this time the state called Deutsches Reich ceased to exist. After some years of interregnum under the administration of the allied forces, two new states were established: the German Federal Republic in the west and the German Democratic Republic in the east. Neither of these states claimed to be a re-establishment of the Reich. Both established themselves as successor states. But only the western Bundesrepublik accepted to act as the successor of title of the Reich, and thus accepted some continuity with the Reich, regarding the Nazi crimes as its moral heritage. The eastern German Democratic Republic, however, regarded itself to be a new state of the hitherto oppressed workers and peasants, defining itself as an antifascist state, thus by definition all the criminal Nazis were in the west. Here, clearly, the rules of identity (in the sociological sense) have been shaped differently. The two German states formed totally different ways to regard people of the past and of the present as belonging to the respective state or not. This shows how differently collectives can form their successorship and their transtemporal allegiances. 5.4 … but there are external constraints Up to now I have argued for what could be dubbed “the autonomy of groups regarding their past”: I have stressed the moment of arbitrariness and the decisionistic element in the diachronic identity of collective social entities. But, as in general not all our decisions are successful, thus also in this particular case, decisions about the diachronic identity of groups can fail. Not all such decisions are successful. If we would decide to reestablish, say, an ancient kingdom and consequently claim its former territory, we would most likely end up in a mental hospital rather than in a palace. Thus it is time now to talk about the external constraints for the success of such decisions. 40

My apologies to the historians for the following sketchy and imprecise remarks. A more detailed and accurate account can be found, e.g., in Kleßmann 1986 and Herf 1997.

67

First, there are what I want to call ontological constraints. In order to reestablish a social entity, it must be the case that this entity previously ceased to exist in the first place. In addition, the formal axioms of identity must be taken care of. Because of the symmetry of identity, if x is a reestablishment of y and x is thus to be identical with y, y must also be identical with x. And because of the transitivity of identity, x must not only be identical with y, but with all z, with which y is identical. From this it follows that there cannot be more than one re-establishment of the same social entity at the same time. Let us assume for refutation that two distinct things x1 and x2 are re-establishments of the same y. Then y is identical with both x1 and x2. By symmetry, x1 is also identical with y, and thus, by transitivity, x1 is identical with x2. But this is against our assumption that x1 and x2 are two distinct entities. Hence only exactly one entity can rightly claim to be diachronically identical with a past entity. Finally, in order to be diachronically identical with a group of a certain kind, often additional material criteria must be met. For a family, for example, bonds of relationships are constitutive of diachronic identity; these cannot be established through a decision alone. Second, there are sociological constraints; after all, groups do not judge about their past alone and on their own. Other groups do so, too, and, of course, also individuals both inside and outside the groups in question. All these judgements, together with a network of dependencies and powerrelations, lead to a pattern of social control. We will not be successful with the re-establishment of an ancient kingdom that we contemplated at the beginning of this section, because others would not agree with us. They would not recognise what we do as a re-establishing, and especially the state on whose territory the kingdom was once upon a time situated would most likely see what we are doing as an act of aggression. Or, to use a historical example: The two German post-war states would never have been successful in shaping their successor-relations as they did, were it not with the consent of the victorious allied nations. Thus the diachronic identity of groups is in a very interesting respect similar to the diachronic identity of individual persons. There, too, we have both an external and an internal perspective on diachronic identity, which do not necessarily coincide in their judgements. If some madman considers himself to be identical with Napoleon, we would simply disagree, just as others disagree with regard to our re-establishment of that ancient kingdom. There is, however, an important difference between the diachronic identity of groups and that of individual persons: While we can, at least sometimes, revive past groups, we cannot at will revive past individual 68

persons. Individual persons and their identity are much more bound to the natural realm and are far less a matter of decision than the realm of the social. 6. Conclusion Social entities, or so I have argued, come about either by collective acceptance that a certain status is imposed on some bearer-entities, or by collective acceptance that some entities count as representations of social entities without that status being imposed on any bearer-entity. They exist as long as their constitutive elements exist, most prominently their bearer-entity (if there is any) and their status. I argued that those entities to which human beings can belong exhibit some peculiar features in their diachronic existence: They can survive not only the exchange of their members, but also their own non-existence. Like natural persons, they have an inner perspective on their identity, but unlike natural persons, they decide autonomously about their past, i.e. about what is to count as an earlier phase of them or as a precursor. There are, however, some restrictions to the success of their decisions: Other groups and individuals judge about such diachronic identity as well, and these judgments matter. Thus we have both a judgement from the inside of the group and judgements from the outside. We thus have two perspectives from which we can make judgements about the diachronic identity of such social entities. We do have two generating factors, the interplay of which will eventually lead to a decision for one of these two perspectives. The diachronic identity at least of social entities is thus very much both a contingent and a practical matter. It depends crucially on 41 decisions.

REFERENCES Anscombe, G.E.M. 1958 “On Brute Facts”, Analysis, 18, 69-72. Appiah, K.A. 2005 The Ethics of Identity, Princeton NJ: Princeton University Press.

41

I would like to thank audiences in Rostock, Saarbrücken and Munich for stimulating discussions of previous versions of (bits of) this paper. Special thanks for comments on earlier versions of this paper are due to Ingvar Johansson, Christian Kanzian, Bertram Kienzle, Katherine Munn, Barry Smith, and Andrew D. Spear.

69

Assmann, J. 1997 Das kulturelle Gedächtnis. Schrift, Erinnerung und politische Identität in frühen Hochkulturen, München: Beck. Frege, G. 1892 “Über Sinn und Bedeutung”, Zeitschrift für Philosophie und philosophische Kritik, new series, 100, 25-50. Gilbert, M. 1989 On Social Facts, London-New York: Routledge. - 1996 Living Together. Rationality, Sociality, and Obligation, Lanham MD: Rowman and Littlefield. - 2000 Sociality and Responsibility. New Essays in Plural Subject Theory, Lanham MD: Rowman and Littlefield. Hakli, R. 2006 “Group beliefs and the distinction between belief and acceptance”, Cognitive Systems Research, 7, 286-297. Herf, J. 1997 Divided Memory. The Nazi Past in the Two Germanies, Harvard University Press, Cambridge MA. Jansen, L. 2004a “Who has got our Group-Intentions?”, in J.C. Marek, M.E. Reicher (eds.), Experience and Analysis. Papers of the 27th International Wittgenstein Symposium, Kirchberg am Wechsel: ILWG, 151-153. - 2004b “We, They, You. Persons in the Plural”, in R. Bluhm, Chr. Nimtz (eds.), Selected Papers Contributed to the Sections of GAP.5, Fifth International Congress of the Society for Analytical Philosophy (CD-ROM), Mentis: Paderborn 2004, 479491. Online available at http://www.gap5.de/proceedings/pdf/479-491_jansen.pdf. - 2005a “Zeit und Gemeinschaft. Soziale Geschichtlichkeit und geschichtliche Sozialität”, in F. Stadler, M. Stöltzner (eds.), Zeit und Geschichte. Beiträge des 28. Internationalen Wittgenstein-Symposiums, Kirchberg am Wechsel: ILWG, 117-120. - 2005b “Institutionen und die kategoriale Ontologie“, in G. Schönrich (ed.), Institutionen und ihre Ontologie, Frankfurt: Ontos 2005, 45-57. - 2005c “katêgoria/Kategorie”, in O. Höffe (ed.), Aristoteles-Lexikon, Stuttgart: Kröner, 299-304. - 2005d “Identität und Gemeinschaft. Neuere Beiträge zur Ontologie des Sozialen”, Zeitschrift für philosophische Forschung, 60, 118-132. - 2005e “Was ist Sozialontologie?”, in O. Neumaier, C. Sedmak, M. Zichy (eds.), Philosophische Perspektiven. Beiträge zum VII. Internationalen Kongress der ÖGP, Frankfurt-Lancaster: Ontos, 279-284. - 2007 “Being a Community and Being in Community. On Hans Bernhard Schmid’s study of we-intentionality”, Metaphysica, 8, 122-130. Johansson, I. 2005 “Money and Fictions”, in Kapten Mnemos Kolumbarium, Festschrift Helge Malmgren, http://www.phil.gu.se/posters/festskrift2/mnemo_johansson.pdf. Johnson, W.E. 1921 Logic. Part I, Cambridge: Cambridge University Press. Kleßmann, C. 1986, Die doppelte Staatsgründung. Deutsche Geschichte 1945-1955, 4th rev. ed., Göttingen: Vandenhoek und Ruprecht. Kobusch, T. 1993 Die Entdeckung der Person. Metaphysik der Freiheit und modernes Menschenbild, Freiburg-Basel-Wien: Herder. Lewis, D. 1986 On the Plurality of Worlds, Oxford: Blackwell. Maalouf, A. 2001 In the Name of Identity. Violence and the Need to Belong, New York: Arcade Publishing.

70

Margalit, A. 2000 Ethik der Erinnerung. Max Horkheimer Vorlesungen, Frankfurt: Fischer. Mumford, St. 2004 “Allegiance and Identity”, Journal of the Philosophy of Sport, 32, 184-195. Niethammer, L. 2000 Kollektive Identität. Heimliche Quellen einer unheimlichen Konjunktur, Reinbeck bei Hamburg: Rowohlt. Reinach, A. 1913 Die apriorischen Grundlagen des bürgerlichen Rechtes, repr. in: Sämtliche Werke. Textkritische Ausgabe, vol. 1: Werke, München: Philosophia 1989, 141-278. Schmid, H.B. 2005 Wir-Intentionalität. Kritik des ontologischen Individualismus und Rekonstruktion der Gesellschaft, Freiburg-München: Alber. Searle, J.R. 1995 The Construction of Social Reality, New York: Free Press. - 2003 “Reply to Barry Smith”, American Journal of Economics and Sociology, 62, 299309; also in D. Koepsell, L.S. Moss (eds.), John Searle’s Ideas About Social Reality: Extensions, Criticisms and Reconstructions, Oxford: Blackwell 2003. Smith, B. 2003a “The Ontology of Social Reality”, American Journal of Economics and Sociology, 62, 285-298; also in D. Koepsell, L. Moss (eds.), John Searle’s Ideas About Social Reality: Extensions, Criticisms and Reconstructions, Oxford: Blackwell 2003. Smith, B. 2003b “John Searle – From Speech Acts to Social Reality”, in B. Smith (ed.), John Searle, Cambridge: Cambridge University Press, 1-33.

71

How Are Identity Conditions Grounded? E. J. LOWE, Durham

In this essay, I examine the notion of persistence, as it applies to material objects, in the context of a general investigation into the role played by criteria of identity in the individuation of objects, both concrete and abstract, and the ontological source or ground of the identity conditions that such criteria assign to objects. 1. Criteria of identity Many philosophers have expressed scepticism about the very idea of a criterion of identity, conceived as a metaphysical principle. Some think that any role that such a principle could be supposed to play is already played by Leibniz’s law, so that nothing further needs to be said on the matter. Others accuse advocates of the idea of supposing, mistakenly, that there is anything substantive to be said about the relation of identity, whose nature is in fact perfectly simple and straightforward. Yet others accuse them of the even grosser error of imagining that identity must be differently defined for different kinds of entity and thus failing to understand that all talk of identity is perfectly univocal. However, all such criticisms are wide of the mark, misconstruing what a criterion of identity is properly conceived to be. A criterion of identity for entities of kind K is simply a principle which specifies, in an informative and non-circular way, the conditions that are logically necessary and sufficient for an entity x and an entity y of kind K to be numerically the same K (see Lowe 1989a). It is manifestly the case that there are such criteria and that these criteria can differ for different kinds K. For example, if x and y are sets, then x and y are the same set if and only if x and y possess the same members. By contrast, if x and y are ordered sets, then they are the same ordered set if and only if they possess the same members in the same order. As a consequence, no set can be identified with any ordered set. Why not? Well, consider the ordered sets and — where a ≠ b — which possess the same members but in a different order. By the criterion of identity for ordered sets, these ordered sets are not identical with each other. However, if each of them were a set, they would be the same set according to the criterion of identity for

sets, because they possess the same members. But then we would have the intolerable consequence that these entities are both identical and distinct — a contradiction (cf. Lowe 1989b, 17). Someone may worry that here I am neglecting the well-known method of ‘reducing’ ordered sets to sets, proposed by Wiener and Kuratowski (see Suppes 1972, 32). According to this proposal, the ordered set may be ‘reduced’ to the set {a, {a, b}}, while the ordered set may similarly be ‘reduced’ to the set {b, {b, a}}. This proposal preserves, as it should, the non-identity of the ordered sets and , by reducing them to different sets. However, the ordered set certainly cannot be identified with the set {a, {a, b}}, because the former has only a and b as its members, whereas the latter has as its members a and the pair {a, b}. The fact is that nothing in the concept of the ordered set connects it ontologically in any significant way with the pair {a, b}. The only connection between these entities was forged, entirely arbitrarily, by the Wiener– Kuratowski proposal itself. In other words, it is no part of the essence of the ordered set that it bears any ontological relationship to the pair {a, b}. By contrast, it is very much a part of the essence of the set {a, {a, b}} that it bears such a relationship to that pair, for the very identity of that set is determined by its members, one of which is this pair. In short, the Wiener–Kuratowski proposal, construed as maintaining that ordered sets may be identified with sets of a certain kind, does not respect the essential differences between sets and ordered sets. 2. Concepts and essences I have dwelt on the foregoing example because it is both simple and yet illustrates some important general points concerning objects and criteria of identity. One lesson of the example is that the Wiener–Kuratowski proposal does not provide an adequate conception of an ordered set. It does not, because it does not accurately reflect the essence of such an entity. By the concept of a K, I shall henceforth mean a conception of a K that is adequate, inasmuch as it accurately reflects the essence of such an entity. Concepts and conceptions are ways of thinking of entities but, in the usage that I am now adopting, conceptions may be more or less adequate ways of thinking of the relevant entities, whereas concepts are to be understood precisely as adequate conceptions of them — adequacy in a conception of a K consisting in the conception’s accurately reflecting the essence of Ks.

74

But what is the ‘essence’ of a kind of entities, K? Here I agree with John Locke, who remarked that in the ‘proper original signification’ of the word essence it denotes ‘the very being of any thing, whereby it is, what it is’ (Locke 1975, III, III, 15). However, we need to distinguish here between general and individual essences. The general essence of a K is ‘what it is’ to be a K — and is therefore shared by all particular Ks — whereas the individual essence of a particular K is ‘what it is’ to be this particular K, as opposed to some other particular K. So, for example, it is part of the general essence of an ordered set to possess some entities in a certain order as its members and it is part of the individual essence of a particular ordered set to have its particular members in a certain specific order — for instance, it is part of the individual essence of the ordered set to have just the entities a and b, in that order, as its members. To the extent that we grasp such facts about the general and individual essences of ordered sets, we possess the concept of that kind of entity and of particular entities of that kind, because we possess adequate conceptions of them, or adequate ways of thinking of them. Such a concept I shall call an individuative concept, because different particulars may be said to fall under it, these being individually distinguished from one another in accordance with a criterion of identity that is built into the concept and is grounded in the general essence of that kind of entity. But, it may asked, how do we know that we do in fact possess an adequate conception of entities of some putative kind? After all, the history of set theory notoriously reveals that early set theorists mistakenly thought that they possessed an adequate conception of sets. The inadequacy was revealed by Russell’s famous paradox and consisted in the fact that a set was conceived to be, in effect, the extension of any possible predicate that is meaningful — a conception which rested on the false presupposition that every such predicate has an extension. A minimum condition on the adequacy of a conception of a K is that it should at least be possible — that is, really, or ‘metaphysically’, possible — that Ks, thus conceived, should exist. An adequate conception of Ks must be a coherent way of thinking of Ks — but there is no coherent way of thinking of something whose existence is metaphysically impossible. Yet, it may now be asked, how can we know whether Ks, conceived in a certain way as entities of a certain kind, are metaphysically possible? If what is being asked here is how we can know this for certain, without any risk of error, then the answer, I think, is that we cannot ever know this. However, I nevertheless think that we are very often entitled to feel confident that we do in fact know that something

75

of this sort is the case, at least in certain instances. To suppose otherwise is to abandon all hope that we can engage in rational thought and discourse — a counsel of despair whose own coherence is dubious in the extreme. What I am saying, in effect, is that a minimum condition of our own rationality is that we can, at least sometimes, grasp the essences of at least some possible kinds of entities. This is consistent with acknowledging that we may be mistaken in thinking that we have grasped the essence of a certain kind of entity, but is also consistent with acknowledging that such errors may be corrected through further thought and reflection on our own part. It should surely be agreed that, if one can grasp essences at all, one may grasp the general essence of a possible kind of entity without yet knowing whether or not entities of that kind actually exist. In this sense, at least, ‘essence precedes existence’, for many possible kinds of entity. That seems clearly to be the case with many entities of kinds whose instances, if they exist, are necessarily concrete and contingent beings. For example, we seem to have a perfectly clear concept of an entity of the following kind: a particular quantity or mass of homogeneous and infinitely divisible matter or stuff. But whether there actually are any such entities in existence would appear to be a question that could at best be settled only empirically. As it turns out, the actual world, it would seem, contains no such quantities of matter, but instead — what might be mistaken for them — aggregates of ‘fundamental’ particles that are either punctiform or else spatially extended but nonetheless non-composite. Such, at any rate, is what modern theoretical physics suggests. Superficial observation of the natural world does not reveal to us which, if either, of these views of physical reality is correct, but the results of controlled experimentation do appear to favour the second hypothesis over the first. The most important point, however, is that scientific experimentation couldn’t be undertaken with a view to supporting either hypothesis in the absence of clear concepts, on the part of the scientists concerned, of the kinds of entities whose existence is in dispute. Unless one already grasps the essence of entities of a putative kind K, one cannot appeal to empirical evidence in support of — or, indeed, against — the claim that entities of that kind actually exist. For we need to know what it is whose existence we are affirming or denying before we can appeal to evidence in support of our claim.

76

3. Macroscopic material objects and their identity conditions Let us now focus on some putative concrete entities of less esoteric kinds than those postulated by fundamental physics — such as a bronze statue and the lump of bronze of which it is made. What, in essence, are entities of these putative kinds supposed to be? Well, to a first approximation, a lump of bronze is supposed to be a maximal connected aggregate of bronze particles — what Locke called a ‘parcel of matter’. By a connected aggregate of certain particles I mean a whole consisting of those particles, with each particle in the whole standing to any other in the ancestral of the adherence relation, so that between any two such particles there is a chain of particles adjacent members of which adhere to one another. And by a maximal connected aggregate of particles I mean a connected aggregate of particles which is not a proper part of a larger such aggregate. The criterion of identity for lumps of bronze is something like this: if x and y are lumps of bronze, then x and y are the same lump of bronze if and only if x and y are maximal connected aggregates of exactly the same bronze particles. For present purposes, we need not inquire too closely into the nature of bronze particles: we may treat them as minimal parts of bronze, that is, as parts of bronze which cannot be divided into entities that are themselves parts of bronze. Of course, modern scientific chemistry will give a more informative account, in terms of the atomic theory of matter. Now, it is evidently a consequence of the criterion of identity for lumps of bronze just proposed that no lump of bronze can survive the loss or replacement of any of the bronze particles composing it, since it cannot be identical with the lump of bronze resulting from any such loss or replacement. But what about a bronze statue? What kind of entity is that supposed to be? Well, clearly, is it is supposed to be a material object made of bronze and possessing a certain specific shape, fit to represent whatever it is supposed to be a statue of. Of course, a lump of bronze is made of bronze and will inevitably possess, at any given time at which it exists, some shape or other. But that in itself doesn’t suffice to show that a bronze statue may be identified with a lump of bronze — the lump of bronze of which, at any given time, it is made. And, indeed, reflection reveals that a bronze statue is conceived to be an entity of such a kind that it can, in principle, survive the loss or replacement of some of the bronze parts of which it is made. If, for instance, some bronze particles from the interior of a statue were to be lost or replaced — indeed, even if they were to be replaced by particles of a different kind, such as lead — this would not be regarded as

77

implying that the statue itself had ceased to exist. And yet, as we know, the implication would be that the lump of bronze of which the statue had formerly been made had ceased to exist. This suffices to show that the criterion of identity for bronze statues differs from that for lumps of bronze and consequently that no bronze statue can be identified with a lump of bronze. There really is no reasonable way to avoid this conclusion and with it the conclusion that sometimes two distinct material objects — to wit, a bronze statue and the lump of bronze of which, at any given time, it is made — exist in exactly the same place at the same time. At least, there is no reasonable way to avoid the latter conclusion short of denying the very existence of either lumps of bronze or bronze statues. For, thus far, I have not been inquiring into the reasons we may have for thinking that entities of either of these kinds actually exist, only into their natures or essences. As for the question of what the criterion of identity for bronze statues is, given that it differs from that for lumps of bronze, that is not quite so easy to answer, precisely because the criterion must allow for the possibility of such a statue’s surviving a loss or replacement of some of the bronze particles composing it at any given time, but not such a loss as to alter significantly its outward shape. It would be a useful exercise to try to formulate such a criterion in a satisfactory fashion, but that is a task that I shall set aside for present purposes, as the desiderata that need to be met are clear enough. Meanwhile, we can learn something from the fact that such a task is not a trivial one. We learn, first, that we were right to insist that criteria of identity are substantive metaphysical principles, by no means rendered redundant by Leibniz’s law. We also learn that we may have a practical grasp of such principles before we are capable of formulating them explicitly — and for many purposes the practical grasp is all that is needed. 4. The (pseudo-)problem of material coincidence This is a convenient point at which to expose more fully some of the misunderstandings of those who are sceptical about criteria of identity. Consider, in particular, those who object to the foregoing line of argument for the conclusion that two distinct material objects may exist in exactly the same place at the same time, such as a bronze statue and the lump of bronze of which it is made. Typically, they raise an objection of the following sort (see, e.g., Olson 2001). Consider the bronze statue, S, and the lump of bronze, L, at a certain moment of time, t. Ex hypothesi, S and L are composed at t of exactly the same bronze particles, arranged in exactly the

78

same way. But, surely — so the objectors urge — all of S’s properties at t are determined by, or supervene upon, the properties and relations of the bronze particles at t, and the same applies to all of L’s properties at t. In that case, however, how can those particles simultaneously compose two distinct objects, which differ in certain of their properties? After all, we have been supposing that S and L differ in their identity conditions and thus in their persistence conditions, so that S, for example, has the property of being able to survive a change in its component particles, while L does not. Again, L supposedly has the property of being able to survive a significant change in its shape, but S does not. But, the objector asks, what can possibly be the ontological ground of these supposed differences in their properties at t, given that S and L are, at t, indistinguishable in respect of their material composition? The property of being able to survive a certain sort of change is a modal property, which is akin — it might be thought — to an object’s powers or dispositions. But an object’s powers or dispositions are seemingly grounded in properties and relations of its material constituents. For instance, if an object is soluble, or brittle, or elastic, this will be because it is composed of particles possessing certain properties or arranged in a certain way. Hence, two objects that were composed of exactly the same particles arranged in exactly the same way couldn’t possibly differ in respect of the powers or dispositions that they possessed. How, then, could S and L, which are composed of exactly the same particles arranged in exactly the same way, differ in that L can survive a significant change of shape but S cannot? Isn’t this as absurd as supposing that L can dissolve, or break, or stretch — that is, that L is soluble, or brittle, or elastic — but S cannot? What would make L and S differ in this supposed way, given their identical material composition? Here it may be replied that L and S differ in this way precisely because L is a lump of bronze, whereas S is instead a bronze statue — and consequently they possess different persistence conditions. But the objector will respond that this answer merely pushes back the problem one step. For now he will ask how it is that S and L, despite having an identical material composition, come to differ in respect of what kind of thing they are — S having the property of being a bronze statue, but not the property of being a lump of bronze, while L has the property of being a lump of bronze, but not the property of being a bronze statue. Even granted that such properties exist and are had by objects composed by material particles arranged in a certain way, how can two objects composed by exactly the same particles arranged in exactly the same way differ in that only one of them has one of

79

these properties, while only the other has the other property? In virtue of what do the two properties get assigned to different objects? How do we even have two objects here, enabling the previous question to be so much as raised? The first thing to be said about this supposed conundrum is that it is patently mistaken to see any sort of parallel between that fact that a bronze statue cannot survive a significant change of shape and the fact, say, that a brittle object cannot be bent without breaking. Brittleness in an object is indeed grounded in the nature of the object’s material constitution, in a way that detailed scientific investigation could reveal. A statue made of glass would be brittle for this reason. But the fact that such a statue could not survive a significant change of shape has nothing, essentially, to do with its brittleness. True enough, if a brittle statue were to break owing to its brittleness, it could well undergo a significant change of shape and thereby cease to exist. But the reason why a statue cannot survive a significant change of shape and why, hence, a statue that broke in that way would cease to exist, is simply that having (approximately) the shape it does is part of the essence of any statue and thus one of its essential properties. What it is for something to be a statue is, in part, for it to have a certain shape. But the essence of a material object is not conferred upon it by its material constitution, in the way that its powers or dispositions may be. On the contrary, it is part of the essence of any material object that it should have a certain sort of material composition. For instance, it is part of the essence of a bronze statue that, at any time at which it exists, it should be composed of bronze particles arranged in a certain overall shape. Similarly, it is part of the essence of a living organism that, at any time at which it exists, it should be composed of organic tissue capable of sustaining the characteristic metabolic processes of life. So, far from its being the case that a bronze statue’s essential properties are conferred upon it by its material constitution, it is the statue’s essence that determines what sort of material composition it can have. To believe that bronze statues exist is to believe that the world includes entities whose general essence is that of a bronze statue and therefore that bronze particles sometimes compose such entities, with the persistence conditions that are implied by that essence. Trying to derive the existence of bronze statues and their persistence conditions from facts about how bronze particles are sometimes arranged, without appeal to the essence of entities of that kind, would be putting the cart before the horse. The same applies with regard to lumps of bronze and their persistence conditions. And yet, providing pre-

80

cisely such derivations is what the envisaged objector is, in effect, insisting that we should be able to do. It is an entirely unreasonable demand to make and the fact that it cannot be met is indicative of no failing in the position that the objector is trying to oppose. The objector is simply labouring under a misapprehension concerning the relationship between facts about essence, persistence conditions and material constitution. Facts about essence explain facts about persistence conditions and impose constraints on facts about material constitution. The objector mistakenly — indeed, incoherently — supposes that the proper direction of explanation and determination is the reverse of this. 5. The incoherence of conceptualism However, seeing off the objector in the foregoing fashion does not leave us without difficult questions on our hands. To be sure, it would be misguided to try to explain the essence and persistence conditions of composite material objects like bronze statues and lumps of bronze in terms of facts about their material constitution. Facts of the latter sort are simply not suited to be the ground or source of facts about the essence and persistence conditions of such objects. But what then is their ground or source? Here we need to ask what, if anything, essences are — what their ontological status is. One very common view would be that, to that extent that talk about essences is legitimate at all, essences are conceptual in character. On this view, the essence of a kind of entities K is simply constituted by ‘our’ concept of a K — or, if not by ‘our’ concept, then at least by some thinking being’s concept. But what is a concept? As I indicated earlier, I myself regard a concept as a way of thinking of something — but I have also stipulated that a concept is an ‘adequate’ conception of something, which accurately reflects the nature or essence of the entity or entities in question. Clearly, then, it is not open to me to say that the essence of Ks is constituted by our concept of a K, because this would leave no room for the notion that the concept of a K is a conception of Ks that reflects the essence of Ks. Essences, on my account, must be mind-independent, if the question can sensibly be put as to whether or not a conception of Ks adequately reflects the essence of Ks. But before I defend and explicate my own view, what can be said of the rival view, that essences are always constituted by concepts? Let us call this view conceptualism. First, conceptualism is a strongly anti-realist view. Second, it is a view that is doubtfully coherent. Let us deal with the first point first.

81

In virtue of what, according to conceptualism, can it be truly said that there exist entities that fall under, or satisfy, our concepts — including, most centrally, our individuative concepts? That is to say, what does it take for there to be Ks, on this view? This is simple enough, it may be said: there must be entities that possess whatever features they are that we have built into our concept of a K. So, for example, if K is lump of bronze, conceived as a maximal connected aggregate of bronze particles, then there must be just such things. This will be the case if, sometimes, some bronze particles adhere to one another so as to form a maximal connected whole. Well and good: but remember that conceptualism is the doctrine that all essences are constituted by concepts. So, in particular, the doctrine must be taken to extend to the essence of bronze particles — what it is to be a bronze particle. Bronze particles, on this view, exist just in case there are some things that possess whatever features we have built into our concept of a bronze particle. However, either the concept of a bronze particle is relevantly similar to that of a lump of bronze, in that it characterizes the nature of such an entity in terms of properties and relations of entities of other kinds, or it is not. If it is, then the next question is just pushed back one stage. If it is not — and this is the next question — then what does it take for the world to contain entities falling under the concept? What, in this case, must the world contribute to the fact that entities of this putative kind exist? Since, according to conceptualism, all essences are constituted by concepts, the conceptualist cannot suppose that how the world is, in respect of what kinds of entities it contains, is something that is the case independently of what concepts thinkers deploy. On this view, what it is for the world to contain entities of a kind K just is for the concept of a K to have application, or be applicable. Consequently, an adherent of this view cannot cash out what it is for such a concept to have application in terms of there being in the world entities answering to the concept. For, as I say, on this view, there being in the world such entities just is a matter of the concept’s ‘having application’. So a different understanding of ‘having application’ must at least implicitly be in play. What is this alternative understanding? I think that it can only be something like this: the concept of a K ‘has application’ just in case thinkers find it useful, or convenient, to conceive of the world as containing Ks. This may require the concept in question to be logically consistent, but otherwise the constraints would seem to be purely pragmatic.

82

This, it seems clear, is a deeply anti-realist view. It is a view according to which, in Hilary Putnam’s well-known words, there isn’t a ‘ready-made world’ — or, if you prefer, there isn’t any truth about ‘what is there anyway’, to use Bernard Williams’s equally familiar phrase (see Putnam 1983, Williams 1978, 64). Or, yet again, it is view according to which, to employ Michael Dummett’s somewhat less felicitous metaphor, reality is an ‘amorphous lump’ — one that can be ‘sliced up’ in indefinitely many different but equally legitimate ways, depending on what ‘conceptual scheme’ we or other thinkers happen to deploy (see Dummett 1981, 563). It may also be the view to which David Wiggins is committed, willy-nilly, by the doctrine that he calls ‘conceptualist realism’ — committed in virtue of the fact that the only notion of individuation that he admits is a cognitive one, whereby individuation is a singling out of objects by thinkers (see Wiggins 2001, 6). Not only is this view deeply anti-realist: it is also, as I have said, doubtfully coherent. For those who philosophize in these terms rarely stop to think about how their doctrine is supposed to accommodate thinkers, their thoughts, and the concepts that they deploy. For these, too, are putative kinds of entities, whose essences, according the conceptualist doctrine, must like all others be constituted by ‘our’ concepts of them. It is at this point that the conceptualist manifestly paints himself into a corner from which there is no escape. There simply is no coherent position to be adopted according to which all essences are constituted by concepts, because concepts themselves are either something or else nothing — they either exist or they do not. If they don’t, then conceptualism is out of business. But if they do, then they themselves have an essence — what it is to be a concept. The conceptualist, to be consistent, must say that the essence of concepts is constituted our concept of a concept. But what could this mean? And what could it mean, according to conceptualism, to say that the concept of a concept ‘has application’ — that there are concepts? I don’t believe that conceptualism has any intelligible answer to such questions. The lesson, I take it, is that at least some essences must be mindindependent, in a way that conceptualism denies. The next task is to try to understand what this entails and how it can be possible.

83

6. What are essences? So what are essences, if they are not — or, at least, not always, ‘constituted by concepts’? The temptation is to say that they are abstract or ‘Platonic’ entities that can be grasped only by the intellect. This temptation should be resisted, in my view — though not for the reason that I am hostile to abstract entities quite generally, for I am not. But it would be equally disastrous, I think, to try to treat essences as entities located in the world of concrete things existing in space and time. This, in effect, was the error of the objector who challenged us to explain how the essence of a material object of any kind is determined by, or supervenes upon, its material constitution. Or, rather, it was the error of any essentialist who was tempted to try to meet that challenge in the terms in which it was posed. Paradoxical though it might superficially seem to be, the well-advised essentialist will say that essences are not entities at all — are not be included amongst what there is in the world, whether abstract or concrete. Rich though the world is in respect of the many kinds of entities that it contains, the essences of entities of those kinds are not further entities to be included in an inventory of what there is. A fortiori, essences are not ‘constituted’ by anything — neither by concepts nor by entities of other types, such as properties or propositions. Part of the point here is that all entities have essences, so that if we took essences themselves just to be entities of a certain kind, then they too would have to have essences, and so on ad infinitum. The essence of an entity of kind K can certainly involve other entities of various kinds — as, for example, the essence of a bronze statue involves bronze particles and a certain characteristic shape. But the essence is not itself an entity, nor is it somehow ‘composed’ or ‘constituted’ by the entities, or kinds of entity, which it ‘involves’. Indeed, we cannot really do much better than to repeat that the essence of a K is what it is to be a K — without, of course, any implication that ‘what it is to be a K’ is itself something, an entity of another kind. If there are, or even merely could be, Ks, then there is a fact of the matter as to what it is to be a K, which obtains — at least in many cases — quite independently of thinkers and their concepts. It is such a fact that a thinker must grasp in order to be able to possess, and evaluate, the belief that Ks exist. For, as I have said before, one is in no position to be able to judge whether or not there is evidence in favour of the existence of Ks unless one knows what Ks are supposed to be — what their essence is. But facts about essence are not made true by the existence of anything what-

84

ever. Essential truths are truths without truthmakers. And any rational thinker must be able to grasp at least some essential truths, simply in order to be able to know what it is that he or she is thinking and reasoning about. Moreover, essences are the ultimate grounds of all modal truths — at least, where the kind of modality that we are concerned with is so-called metaphysical possibility and necessity. If something is possibly or necessarily the case in this sense, it is so in virtue of the essences or natures of the entities, or kinds of entities, that are implicated in the modal fact in question. For example, that it is metaphysically impossible for a statue to survive the loss of its shape is a fact that obtains in virtue of the essence of a statue — in virtue of what it is to be a statue. Similarly, that it is metaphysically possible for a bronze statue and a lump of bronze to exist in exactly the same place at the same time is a fact that obtains in virtue of the essences of a bronze statue and a lump of bronze. 7. Material objects and causal powers I imagine that essentialism of the sort that I am recommending will be attacked for a number of predictable reasons, one of which will be this. It will be complained that such essentialism implies excessive liberality about the kinds of entity that could possibly exist. For instance, it may not seem to exclude the possibility of such entia non grata as something of the following supposed kind: a material object consisting of a lump of bronze and a live cat situated ten metres away from that lump. The thought is that such an object would supposedly be one which, in virtue of its essence, would come into existence whenever a live cat moved to a position ten metres away from a lump of bronze and would go out of existence whenever the cat moved to a different distance from the lump — and that it is absurd to suppose that anything can come into or go out of existence for this sort of reason. However, it would be rash to suppose, quite generally, that no entity of any genuine kind can be such that it is part of its essence that it comes into existence when two entities of other kinds move to a certain distance from each other and goes out of existence when they move further apart. After all, a hydrogen atom is just such an entity: it comes into existence when a proton and an electron are close enough to each other for the latter to be captured by the former and goes out of existence when the electron is subsequently ejected from its orbit. Why do we say that a new entity — a hydrogen atom — is first created and later destroyed in these circumstances? Because new causal powers are

85

brought into being when a proton and an electron stand in the specified relation to one another — causal powers that are not possessed by either free protons or free electrons individually. These powers must be attributed to something as their bearer and this thing is a hydrogen atom. It is — or so I am strongly inclined to urge — part of the essence of any material object to have certain causal powers that are not possessed individually by any of its material constituents (if it has any). That, indeed, is why I am very doubtful whether a mere ‘mereological sum’ of material objects should be considered to be a further material object in its own right. For all that is supposedly required for the existence of a certain mereological sum is the coexistence of the various objects of which it is said to be the sum. I cannot, however, see how such a sum differs from the bare plurality of those objects. In short, I don’t see how it can be supposed that they together compose it — and hence that it exists — merely in virtue of their coexistence. A genuine principle of composition for material objects of any kind must, I believe, be one in virtue of which an object composed according to that principle acquires causal powers additional to any that are possessed by its material components individually. Returning, then, to our example of the putative object composed by a lump of bronze and a live cat situated ten metres away from it: if I could be persuaded that new causal powers were brought into being by this quite specific arrangement, analogous to those acquired by a hydrogen atom upon its formation from a proton and an electron, then I would be happy to recognize this as an object of a genuine kind. But I shall need a lot of persuasion! Notice, in this connection, that lumps of bronze do not fare badly by this criterion, for in virtue of the mutual adherence of their constituent particles, they do acquire causal powers that are not possessed by the particles individually — powers such as a certain moment of inertia that is different from any possessed by any of those particles individually. Notice, too, that what I have just said about ‘mereological sums’ of material objects does not carry over to sets of material objects, because sets are abstract objects and are not, in any case, composed by their members — the members of a set are not parts of it. Finally, notice that the requirement that material objects should have causal powers over and above those individually possessed by their component parts (if they have any) does not present problems for such objects as bronze statues, which are ‘made’ of, but not identical with, lumps of matter. For even if a bronze statue has the same causal powers as the lump of bronze of which it is made, it still has, therefore, causal powers over and above those individually possessed by

86

the bronze particles composing both it and the lump — for the lump itself, as I have just remarked, has such additional causal powers. Moreover, it is at least arguable that the statue has some causal powers not even possessed by the lump, such as the power to evoke an aesthetic response in someone viewing it. As I have already indicated, it is one thing to establish that entities of a putative kind K genuinely could exist, but another to establish that they actually do exist. The former task is at least partially an a priori one, which is completed once we succeed in grasping the essence of entities of kind K. Consider again the putative kind lump of bronze. A lump of bronze, if there can be such a thing, is in essence a maximal connected aggregate of bronze particles and, supposedly, a material object. Provided that there is such a species of matter as bronze and provided that it is particulate in nature — rather than, say, homogeneous and infinitely divisible — then, provided also that bronze particles can adhere together in the way necessary for them to compose such an aggregate, such an aggregate can exist. But there is a further requirement to be met if such an aggregate is to qualify as a material object, namely, that in virtue of its component bronze particles adhering together in the specified way, additional causal powers are generated which can be attributed to the aggregate as a whole but not individually to its component particles. Only empirical inquiry, it seems, can ultimately reveal to us whether bronze is a distinctive species of matter that is particulate in nature, whether its particles can adhere together and whether, in virtue of their adhering together, new causal powers are generated. Then, however, the further empirical discovery that sometimes bronze particles do adhere together in the required way will suffice to establish that some lumps of bronze do actually exist. What is clear is that, inasmuch as a grasp of essences is necessary in order to establish empirically that entities of any putative kind actually exist, the task of establishing such existential truths cannot be a wholly a posteriori one, but includes, rather, ineliminable a priori elements. 8. Conclusion I am now in a position to give a summary answer to the question raised in the title of this essay: How are identity conditions grounded? The answer is that the identity conditions of entities of any kind K are grounded in the essence of Ks — what it is to be a K. The essence of a kind K is not, however, a further entity of any kind, neither abstract nor concrete. As rational

87

beings, we must be able to grasp the essences of at least some kinds of entities. Indeed, it is part of the essence of a rational being that it has such a grasp. In grasping the essence of entities of a kind K, we come to grasp their identity conditions and may — albeit, very often, only with some effort — be able to formulate an explicit criterion of identity for Ks, which specifies their identity conditions in an informative and non-circular way. Such a criterion will make reference to, or quantify over, entities of other kinds involved in the essence of Ks, unless K is a fundamental kind of entity — in which case, its essence will not involve entities of any other kinds and no informative and non-circular criterion of identity for Ks will be forthcoming. It is knowledge of essences that grounds all of our modal knowledge of what is metaphysically possible or necessary, including our knowledge of what sorts of changes an entity can or cannot survive or persist through, if indeed it is an entity in whose nature it is to persist. The persistence conditions of such an entity are not grounded in its material constitution, if it is a material object, but in its general essence. This is why it is a mistake to find it at all puzzling that two distinct material objects, of different kinds and possessing different identity conditions, may simultaneously be composed by the same material constituents, as in the case of the bronze statue and the lump of bronze of which it is made. Our individuative concepts — our concepts of those kinds that have individual instances, distinguishable in principle from one another — are ways of thinking of entities of those kinds that accurately reflect their general essences, thus involving a grasp of their identity conditions and thereby an understanding of what makes one particular instance of such a kind distinct from another. The essences of such kinds are not constituted by our individuative concepts, but are in general entirely mindindependent. Conceptualism — the doctrine that all such essences are constituted by our individuative concepts — leads inexorably to an incoherent global anti-realism and must be rejected. However, it is not enough, in general, to grasp the essence of entities of a kind K in order to know whether or not entities of that kind actually exist. For that purpose, we must appeal to empirical evidence — at least in the case of concrete, as opposed to abstract, entities. At the same time, to the extent that such empirical inquiry always depends upon a grasp of essences, there is an un-

88

avoidable a priori element in our knowledge of the existence of entities of any kind whatever.1

REFERENCES Dummett, M. 1981 Frege: Philosophy of Language, 2nd edn, London: Duckworth. Locke, J. 1975 [1690] An Essay Concerning Human Understanding, ed. P.H. Nidditch, Oxford: Clarendon Press. Lowe, E.J. 1989a “What is a Criterion of Identity?”, The Philosophical Quarterly, 39, 1–21. - 1989b Kinds of Being: A Study of Individuation, Identity and the Logic of Sortal Terms, Oxford: Blackwell. Olson, E.T. 2001 “Material Coincidence and the Indiscernibility Problem”, The Philosophical Quarterly, 51, 337–55. Putnam, H. 1983 “Why There isn’t a Ready-Made World”, in his Realism and Reason: Philosophical Papers, Volume 3, Cambridge: Cambridge University Press, 205– 28. Suppes, P. 1972 Axiomatic Set Theory, New York: Dover. Wiggins, D. 2001 Sameness and Substance Renewed, Cambridge: Cambridge University Press. Williams, B. 1978 Descartes: The Project of Pure Enquiry, Harmondsworth: Penguin, 1978.

1

This is an abbreviated version of a paper delivered at a conference on Criteria of Identity at the Katholieke Universiteit Leuven in December 2004. I am grateful to members of the audience on that occasion for their critical comments.

89

A Brief Ontology of Spatial and Temporal Localization UWE MEIXNER, Saarbrücken

1. Preliminaries A location in space or spatial location will here be not a point in space, but what is produced by enclosing a (in three dimensions positively extended) part of space in an outward boundary (the boundary, too, is to belong to the location). A location in time or temporal location will here be not a point in time, but a closed interval in time (the boundary-points, too, are to belong to the location). L, L´, L´´, … will be used to designate locations in space; T, T´, T´´, … will be used to designate locations in time; t, t´, t´´, … will be used to designate points in time; l, l´, l´´, … will be used to designate points in space; X, X´, X´´, … will be used to designate objects of any kind (including temporal and spatial locations, points in time and points in space). The word “object” is not used as a synonym for “entity” here. Rather, it designates any individual-like entity that is necessarily of finite extension in all the dimensions it may happen to have. Spatial locations are special sets of points in space, temporal locations special sets of points in time. Accordingly, “t is in T,” “l is in L” simply means that t is an element of T, that l is an element of L. Note that, presupposing the continuity of space and time, every spatial location is an infinite set of points in space, and every temporal location an infinite set of points in time. The fundamental localization-predicate for space is “X is at t in L”; the fundamental localization-predicate for time is “X is in T”. We will also be using the fundamental mereological predicates “X´ is at t a spatial part of X,” “X´ is a temporal part of X,” and “X´ is a part of X.” 2. Two fundamental principles P1 For all X, t, and L: X is at t in L ≡ some X´ is at t a spatial part of X, and every spatial part of X at t is at t in L.

P2 For all X and T: X is in T ≡ some X´ is a temporal part of X, and every temporal part of X is in T. 3. Two consequences of these principles C1 For all X and t: X has no spatial part at t ⊃ X is in no spatial location at t. C2 For all X: X has no temporal part ⊃ X is in no temporal location. Note that an object’s being in no spatial location does not preclude that a spatial location can be analogically associated with the object. Likewise, an object’s being in no temporal location does not preclude that a temporal location can be analogically associated with it. Important examples will be presented below. 4. The definitions of exact localization and exact location D1 X is at t exactly in L := X is at t in L, and there is no part L´ of L which is such that there is at t no spatial part X´ of X with X´ being at t in L´. D2 L is the spatial location of X at t := X is at t exactly in L. D3 X is exactly in T := X is in T, and there is no part T´ of T which is such that there is no temporal part X´ of X with X´ being in T´. D4 T is the temporal location of X := X is exactly in T. 5. Two questions, leading to further questions If X is at t in L, does is follow that there is a spatial location L´ such that X is at t exactly in L´? If X is in T, does it follow that there is a temporal location T´ such that X is exactly in T´? The answer to the second question is “No.” Suppose X is an object with temporal parts and with an intermittent existence: first it exists for a while, then it does not exist for a while, then again it exists for a while. Then there is nevertheless a temporal location T such that X is in T, but obviously there cannot be a temporal location T´ such that X is exactly in T´. (Remember that T´, as a temporal location, must be a closed interval in time.) Taking the idea from the 92

must be a closed interval in time.) Taking the idea from the temporal counterexample, it is easily seen that one cannot conclude from X being in L at t that X is at t exactly in any spatial location L´. While being in some spatial location will often be (we may take it) without being exactly in any spatial location, it would (come to think of it) be a very strange universe indeed if being in some spatial location were in each and every case, of a point in time t and an object X that is not itself a spatial location, without being exactly in any spatial location. X is at t a spatially full object if, and only if, there is some spatial location in which, at t, X exactly is. Trivially, every spatial location is a spatially full object at any point in time. But is any object ever spatially full that is not itself a spatial location? Yes, a finite part of a gravitational field seems to be an example. But, so far, spatially full material objects have never been observed. Indeed, spatially full material objects seem to be impossible, since it seems such objects would have to be of infinite density (and therefore of infinite mass). But if there is no material object that is ever spatially full (which entails that there are no Democritean atoms), what, then, is the nature of the composition of material objects? If one delves into this question, all initial clarity quickly disappears. Thus, what many still consider to be the most lucid – least obscure – kind of entity, material object, is by no means as lucid as it seems at first sight to be. 6. Definitions concerning physical objects D5 X is a physical object := it is (metaphysically) possible that there is a point in time, t, and a spatial location, L, such that X is at t in L. Since metaphysical possibility is an S5-modality – which implies that the principle what is possible is necessarily possible is valid for it – D5 entails that every physical object is a physical object with metaphysical necessity. The relevant logical structure here is not the trivial □(A ⊃ A), but rather this: (A ⊃ □A).

Note also that being a physical object does not entail being purely a physical object. This remains true if “physical” (in “physical object”) is augmented as follows: “at t existing physical,” “existent physical,” “persistent physical,” “enduring physical,” “perduring physical” (concerning these augmentations of “physical,” see below). However, while it is true that

93

every physical object is a physical object with metaphysical necessity, it is not true that every at t existing physical object is an at t existing physical object with metaphysical necessity. And what results if “at t existing” is, in the previous sentence, replaced by “existent” is also not true. The simple reason for this is that there are certainly at t existing – respectively, (simpliciter) existent – physical objects for which it is not metaphysically necessary that they exist at t – respectively, that they are existent. D6 X is at t an existing physical object := there is a spatial location, L, such that X is at t in L. D7 X is an existent physical object := there is a point in time, t, such that X is at t an existing physical object. D8 X is a persistent physical object := there is a location in time, T, such that for every t in T there is a spatial location L such that X is at t in L. Note that even though it seems certain that there are existent physical objects, it does not automatically follow that there are also persistent physical objects. D9 X is an enduring physical object := X is a persistent physical object, and there is no X´ such that X´ is a temporal part of X. D10 X is a perduring physical object := X is a persistent physical object, and there is an X´ such that X´ is a temporal part of X. 7. An ontological classification (I)

Objects that do not have, at any point in time, any spatial part and that do not have any temporal part. (II) Objects that have some temporal part, but do not have, at any point in time, any spatial part. (III) Objects that have, at some point in time, some spatial part, but do not have any temporal part. (IV) Objects that have, at some point in time, some spatial part, and that also have some temporal part.

94

For purely logical reasons, no object can be subsumed under two of the above four headings, and for purely logical reasons, every object must be subsumed under one of the four headings. It is a curious fact of the history of metaphysics that most modern metaphysicians are more or less eager to argue that there is in fact no object that falls under the headings (I), (II), and (III). The opposition to (I)-objects is very old, the opposition to (II)objects more recent; the most recent opposition is the opposition to (III)objects. Before discussing each of the above four categories of objects, we assume the following two additional, highly plausible principles: P3 For all X and t: X has at t some spatial part ⊃ there is some L such that X is at t in L. P4 For all X: X has some temporal part ⊃ there is some T such that X is in T. These principles, taken together with C1 and C2, have the following logical consequences: C3 For all X and t: X has at t some spatial part ≡ X is at t in some spatial location. The left-to-right part of the biconditional C3 is just P3. The right-to-left part of the biconditional C3 is the contraposition of C1 (which itself is logical consequence of P1).

C4 For all X: X has some temporal part ≡ X is in some temporal location. The left-to-right part of the biconditional C4 is just P4. The right-to-left part of the biconditional C4 is the contraposition of C2 (which is itself a logical consequence of P2).

Now, concerning (I)-objects: According to C3 and C4, the objects that do not have, at any point in time, any spatial part and that do not have any temporal part are precisely the objects that are in no temporal location and never in any spatial location. In other words, (I)-objects are outside of time and space. Putative examples of such objects are not only numbers and other type-objects (i.e., non-predicative universals), like the lion and the homo sapiens, but also non-physical substances, like human souls. Though

95

(I)-objects are outside of space, a spatial location can be analogically associated with (I)-objects if they are exemplified by objects that are, literally, in a spatial location. If an exemplum X´ of a (I)-object X is at t in L, then X, too, can be said – analogically – to be at t in L. Thus, the type-object homo sapiens is at t in any spatial location in which there is at t a human being. The analogical being-in a spatial location is quite different from the literal being-in a spatial location, since, obviously, homo sapiens is at one and the same point in time analogically-in separated spatial locations – while no human being can be literally-in separated spatial locations at one and the same point in time. Exemplification is not the only way to locate (I)-objects analogically in space. Spatial locations can also be analogically associated with, say, human souls, which are (I)-objects of a quite different sort than type-objects (the latter being individual-like universals, whereas souls are true individuals), via the living human bodies to which these souls are connected: a soul is at t (analogically) in every spatial location in which its body is at t (literally). It should not go unmentioned that temporal locations, too, can easily be analogically associated with souls: Consider the non-empty set of points in time at which a given soul (of a body) is connected to its body. Every temporal location that includes this set is a temporal location in which that soul is (analogically speaking, not literally). Concerning (II)-objects: According to C3 and C4, (II)-objects are outside of space, but in time. The putative examples of (II)-objects that immediately come to mind are conscious mental occurrences, (conscious) experiences in the broadest sense. But in these times of the predominance of materialism, one is constrained to offer further specifications: experiences as they are traditionally conceived of, i.e., experiences as they are conceived of by psychophysical dualists, or by idealists. It goes without saying that experiences thus conceived of are nowadays almost as much out of favor as souls are (and as universals would still be, I dare say, if the authoritative materialist David Armstrong had not had pity on them). Concerning (III)-objects: According to C3, D6, and D7, all (III)-objects are existent physical objects. This may make it seem easy to find examples of (III)-objects that are not only putative examples of them. Unfortunately, according to C3 and C4, (III)-objects are not only in space, but also outside of time, which for many metaphysicians makes them very dubious entities indeed. Their being outside of time can easily be eclipsed, since a temporal

96

location can very naturally be analogically associated with any (III)object: Take the non-empty set of all points in time at which a given (III)object exists, in short: its duration; a (III)-object is analogically in every temporal location that includes its duration. But literally – because of its lack of temporal parts – it is in no temporal location. Frequently, the absence of a (literal) temporal location for (III)-objects is expressed in the following highly suggestive, but somewhat misleading way: (III)-objects are said to be “wholly present at each point in time at which they exist.” The same can be said of (I)-objects: they, too, are “wholly present at each point in time at which they exist.” The formulation is somewhat misleading since, contrary to its real intention, it suggests that (III)-objects and (I)-objects are (literally) in time, albeit in a strange way. Despite their recent dubiousness to many metaphysicians, (III)-objects constitute, in the history of ideas, the standard conception of existent physical objects. Note that so-called Aristotelian material continuants are (III)-objects. Note also that one could define (III)-objects as existent physical objects that have no temporal parts. Above, (III)-objects have been defined as “objects that have, at some point in time, some spatial part, but do not have any temporal part.” But the first part of that definition – “objects that have, at some point in time, some spatial part” – is synonymous to “existent physical objects,” according to D6 and D7.

Concerning (IV)-objects: According to C3 and C4, (IV)-objects are both in space and in time. According to D6 and D7, they are the existent physical objects that have some temporal part. For most modern metaphysicians, the spatio-temporal double locatedness of (IV)-objects is a mark of ontological distinction that confers to them, so to speak, the right of there-being – given that those metaphysicians tend to deny this right to the three other categories of objects in the above fourfold classification. Direct examples of (IV)-objects are physical occurrences, goings-on in time and space; in fact, physical occurrences do seem to exhaust the (IV)-objects. 8. Reductionism and eliminativism regarding objects other than (IV)objects If one believes, as modern metaphysicians tend to believe, that there just aren’t any other objects than physical occurrences, one still has a choice with respect to objects other than physical occurrences: either one thinks it worthwhile to try to model, on the basis of (IV)-objects, at least some of

97

those objects that one considers to be, in fact, non-entities (i.e., the (I)objects, (II)-objects, and (III)-objects), or one does not think it worthwhile. Usually, the first alternative has seemed much more attractive than the second (and this bespeaks the considerable intuitive force which is retained by the assumption that there are other objects than physical occurrences even after this assumption has been discarded). Adopting the first alternative with regard to (I)-objects, has led to reconstruction-theories for abstract objects. With regard to (I)-objects, however, that are not abstract objects – like souls, God, and non-existent physical objects –, one has, as a rule, not gone to so much trouble; simple denial has usually been thought good enough for them. Non-existent physical objects are (I)-objects? – Yes, they are. According to D6 and D7, a non-existent physical object is a physical object such that there is no point in time, t, such that there is a spatial location, L, such that X is at t in L. In short, a nonexistent physical object never is in a spatial location; it is, therefore, not in space. But if a physical object is not in space, it is not in time either, because if it never is in a spatial location, it does not exist at any point in time – according to D6 – and hence cannot be in a temporal location, since the following statement certainly is a true general principle: P5 For all X and T: X is in T ⊃ there is a t in T at which X exists. Thus, non-existent physical objects are indeed (I)-objects.

The said reconstruction-theories for abstract object have not been entirely successful, since one is forced to postulate (besides (IV)-objects) certain non-reconstructed, hence non-reduced, abstract objects – for example, sets – as forming the basis for the reconstructions. It’s not only true: de nihilo nihil, it’s also true: de nihilo abstracto nihil abstractum. Adopting, in turn, the above first alternative with regard to (II)-objects, has led to physicalist reductionism in the philosophy of mind, one of its manifestations being the notorious token-identity theory. Finally, to adopt the said alternative also with regard to (III)-objects may seem an easy thing to do, because to all putative (III)-objects there correspond one-toone certain (IV)-objects, certain physical occurrences: Clearly, each putative (III)-object has exactly one course of existence, and different putative (III)-objects have different courses of existence; and if X is a putative (III)object, then its course of existence is no doubt a physical occurrence. Therefore, why not simply replace the putative (III)-objects – while retaining their names – by their courses of existence, which are (IV)-objects?

98

Here is a reason why not: It might have been otherwise with a putative (III)-object than it is in fact; for example, this stone might now have lain over there instead of over here. It seems that there is no way of reconstructing this way of speaking truly about this stone if this stone is no longer regarded as a certain (III)-object but as a course of existence (namely, the initially assumed (III)-object’s course of existence); for if the stone had lain over there, its course of existence would have been a course of existence that is numerically different from its actual course of existence, while the stone itself, even if it had lain over there, would still be numerically identical to the stone that actually lies over here. But, in fact, there is a way of reconstructing the above true assertion of a specific counterfactual possibility within a – broadened – (IV)-object ontology. It is another question, however, whether that way of reconstruction is plausible. Suppose this stone is identified with a certain physical occurrence. Then we can say some things truly about the stone that we could not say formerly, when we still regarded the stone as a (III)-object; for example, that an earlier stage of this stone lay over there. That’s not a bad thing. But we would certainly also like to still say some things truly about the stone that we could already say truly about it formerly, when we still regarded the stone as a (III)-object; one of these things is that this stone might now have lain over there instead of over here. Now, one who is inspired by David Lewis will argue that asserting that this stone (qua (IV)object, qua physical occurrence) might now have lain over there instead of over here amounts to asserting that, while this stone is now actually lying over here, there is a counterpart of it, X, that now lies over there and not over here – but of course not in the actual word but in some other, nonactual possible world. This counterpart, X, of the stone is, indeed, a nonexistent physical object. Hence (see above in this section) it is neither in space nor in time (in other words, it does not have, at any point in time, any spatial part, and it does not have any temporal part, either), which, however, does not prevent it (1) from being (simpliciter) a physical object (see D5), and (2) from being a (IV)-object in another possible world – that is, an in another possible world existent physical object that has some temporal part (compare the end of the previous section), in other words: a physical occurrence in another possible world. The Lewis-inspired proposal just described should not be criticized, in my eyes (others will think otherwise), on account of relying on nonexistent objects (or non-actual objects, if you prefer; Lewis, in fact, would have very much preferred the latter verbalization, since he did not believe

99

that existence and actuality could be legitimately identified). It also should not be criticized on account of introducing a notion of existence which is relative in what may seem to be a weird way: existence in a possible world. The real problem with this proposal is that it leaves one with nothing but arbitrary stipulation regarding the question what counts as a counterpart of a given (IV)-object in another possible world, once unreduced reference to (III)-objects is forbidden (and such reference must be forbidden if all putative (III)-objects are to be reducible to (IV)-objects). If one does accept (III)-objects, then the counterpart-relation need not be an arbitrary concoction even when it straddles possible worlds: the possible course of existence X´1 of the (III)-object X1 is a counterpart of the possible course of existence X´2 of the (III)-object X2 if, and only if, X1 = X2. This is certainly a non-arbitrary definition of X´1 being a counterpart of X´2 (one that makes the counterpart-relation an equivalence relation if that relation holds precisely between the possible courses of existence and if each such course is the possible course of existence of some (III)-object). But if one does not accept (III)-objects, what will then be the necessary and sufficient condition for X´1 being a counterpart of X´2? As is well known, Lewis relied on overall-similarity to define the counterpart-relation: a counterpart X´ of a (IV)-object X is maximally overallsimilar to X among all the (IV)-objects in X´’s world (in the space-time to which X´ belongs). But like love, similarity is a fickle thing, and like beauty, it is in the eye of the beholder. Thus, a person may at a time legitimately judge that X´ – a (IV)-object in another possible world – is a counterpart of X, and just as legitimately at another time that it is not (having shifted her emphasis from certain similarity-aspects to others); and one person may at a time legitimately judge that X´ is a counterpart of X, but another person at the same time just as legitimately that it is not. This is certainly not an ontologically satisfactory situation. 9. The strangeness of ontological denial A classification of objects is easily made. Some object-classifications are uninteresting; others are interesting. I submit that the above classification of objects into (I)-objects, (II)-objects, (III)-objects and (IV)-objects is a rather interesting classification. The intriguing thing is that the ingredients of that classification – (1) having some temporal part and the negation of this: not having any temporal part, and (2) having, at some point in time, some spatial part and the negation of this: not having, at any point in time,

100

any spatial part – seem harmless enough, in themselves and in relation to each other. There seems to be no a priori reason why any of the four categories that can be formed by consistently combining the said classificationingredients should be empty. Why, then, is it nevertheless assumed by so many that three of the categories are in fact empty? That seems strange. Do they have straight empirical reasons for this assumption? Most certainly not; for the questions whether there are (I)-objects, (II)-objects, (III)objects, (IV)-objects are quite far removed from, say, the questions whether there are mammals with neither tail nor beak, mammals with beak but without tail, mammals with tail but without beak, mammals with both tail and beak. The former questions are, so to speak, on a plane of inquiry that is quite different from the plane of inquiry to which the latter questions belong. The latter questions are empirical, the former metaphysical. The distinction between empirical and metaphysical inquiry does, however, not imply that metaphysical inquiry is bound to be a priori (to assume that the distinction does imply the apriority of metaphysical inquiry would just amount to making all over again the old, Kantian mistake regarding the conception of metaphysics, which has been so detrimental to the subject). Metaphysical inquiry is not empirical inquiry, but it is nevertheless strongly related to human experience, and therefore certainly not a priori. Metaphysical inquiry aims at a basic interpretation of experience; it is, so to speak, the basic hermeneutics of experience. In this light, what position should one adopt regarding (I)-objects, (II)objects, (III)-objects, and (IV)-objects? The answer is quite clear: all four kinds of objects have, for a very long time, been objects of reference in that interpretation of experience that is manifest in everyday, and also in scientific, human discourse. Contrary to what eliminativists of various stripes are wont to claim, it is not likely that this situation will change in any nonnegligible way. The best explanation of this, in turn, is that not only (IV)objects but also (I)-objects, (II)-objects and (III)-objects are indeed needed for a basic interpretation of experience. Therefore, it is reasonable to assume that there are not only (IV)-objects but also (I)-objects, (II)-objects, and (III)-objects; it is reasonable to assume that all four categories of objects are non-empty, it is not reasonable – or at least less reasonable – to assume the contrary. I am well aware that a huge amount of literature has been dedicated to arguing the contrary – literature that I cannot even begin to do justice to in this short paper –, but it seems to me that all that is shown by this remarkable phenomenon is the staggering extent to which the ontological community has been ruled by a mentality that is by no

101

means rationally necessary for it to adopt: a mentality which one might term the Democritean Ockham’s Razor Mentality. (or in short: Quineianism). It is time for a change. Ontological tolerance is better than ontological intolerance, and Ockham’s Razor – Entia non sunt multiplicanda praeter necessitatem, no more than an ontological rule of thumb – needs to find its partner and correcting principle in Husserl’s Boat (as the following quite different ontological rule of thumb may with good reason be dubbed): Intra finem possibilitatis entia sunt salvanda.

REFERENCES Lewis, D. 1986 On the Plurality of Worlds, Oxford: Blackwell. - 1999 “Noneism or Allism?”, in D. Lewis, Papers in Metaphysics and Epistemology, Cambridge: Cambridge University Press, 152-163. Meixner, U. 2006 David Lewis, Paderborn: mentis. - 2002 “Change and Change-Ersatz”, in A. Bottani, M. Carrara, P. Giaretta (eds.), Individuals, Essence and Identity. Themes of Analytic Metaphysics, Dordrecht: Kluwer, 427-449.

102

Ontology, ‘Existence’ and The Role of Intuition* KRISTIE MILLER, Brisbane

1. Introduction Metaphysicians frequently appeal to intuition. But when is that appeal useful? I consider that question by focusing on our existential intuitions. In particular, I want to go some way to answering the question of whether, and when, appeal to existential intuitions is useful, by considering the issue in the light of an argument for unrestricted composition. This argument appeals to a difference in the extent to which restricted and unrestricted compositionalists appeal to existential intuitions, and concludes that at the very least, restricted compositionalists are on shakier epistemic ground than unrestricted compositionalists. Having outlined this argument for unrestricted composition, I consider the nature of the unrestricted existential quantifier in the context of three different metaphysical scenarios about the way the world might be: scenarios that differ with respect to whether the metaphysics of a world completely or uniquely settles the division of that world into objects. I argue that appeal to intuition is not only useful but necessary if we are to determine the intension of ‘exists’ and hence its extension at any given world. But this appeal is not simply an appeal to our brute existential intuitions— our dispositions actually to quantify over a certain domain. Rather, it requires equilibration over a set of worlds in which we consider our ‘quantificational’ dispositions in each world, where, importantly, included in this set of worlds are worlds in which the metaphysics differently settles the division of those worlds into objects. Ultimately, I argue, our dispositions yield various conditional claims about our concept . These are conditional claims of the most abstract kind: they tell us, for instance, that if there is some natural, unique division of our world into objects, then the unrestricted quantifier ranges over all and only the objects in that domain. *

Thanks to David Braddon-Mitchell and David Chalmers for helpful discussion and to the Australian Research Council for funding.

But, I will argue, our untutored, world-bound dispositions are no guide at all to what actually exists even on the supposition that metaphysics alone does not uniquely or completely settle a natural division of our world into objects. Hence they are no guide to the nature of composition. And to the extent that certain restricted compositionalist views depend on such intuitions, so much the worse for them. Or so I shall argue. 2. An Argument for Unrestricted Composition1 According to some metaphysicians, composition is restricted: only some sets of particulars are such that the members of that set compose something. Most restricted compositionalists think that composition is restricted in a way that preserves most of our folk ontology—cars, cats and canapés exist, odd gerrymandered fusions of miscellaneous particulars do not.2 Call this a folk friendly (FF) restriction. The task is then to define a relation of composition such that the relation holds between all and only the particulars that are parts of some whole. In the case of FF restrictions, accomplishing this relies on an appeal to a certain class of what I will call brute intuitions: intuitions about the way the world is, as opposed to say, semantic intuitions about the meanings of words. Brute intuitions include intuitions such as that I have a hand, that we have free will, that some acts are wrong, and so on. The class of brute intuition in question are brute existential intuitions: intuitions about what actually exists. If a restriction is to be folk friendly, then the composition relation must be defined with the folk ontology in mind. The composition relation just is the complex multi-grade relation, whatever it is, that holds between all and only the particulars that are parts of some whole that is posited by the folk ontology. FF restricted compositionalists presuppose that our brute existential intuitions are right, and build the constraints imposed by those intuitions into their account of composition. Those who embrace unrestricted composition, however, tend to be motivated by various theoretical considerations, such as the idea that FF restricted composition entails vague exis-

1

This is also an argument for compositional nihilism—the view that there is no composition. For simplicity I will talk only of unrestricted composition, but the parallels should be clear. 2 Notable exceptions are van Inwagen (1990), and Merricks (2001), who reject the existence of the gerrymandered fusions, but also reject the existence of canapés, cars and much else that the folk admit into their ontology.

104

tence, and vague existence is incoherent.3 The unrestricted compositionalist need not appeal to any brute existential intuitions. This suggests the following argument. The FF restricted compositionalist must appeal to brute existential intuitions. Unrestricted compositionalists appeal to intuitions about theoretical and conceptual coherence. There are good reasons to think that intuitions about what is incoherent are significantly better guides to what is incoherent, than brute existential intuitions are guides to what does exist. For these latter intuitions are not good guides to what exists. So at the very least we should prefer unrestricted composition to FF restricted composition since the latter rests on such shaky foundations. This all rests on the idea that brute existential intuitions are a poor guide to what exists. Are they? 3. When are Intuitions a Guide? Until recently there has been a sustained backlash against conceptual analysis and a priori knowledge generally, and hence against the use of intuition. Lately, however, so-called ‘armchair philosophy’ has been revived by Frank Jackson and David Chalmers through the apparatus of twodimensional semantics. In essence, they claim that conceptual analysis can yield certain sorts of a priori knowledge. According to two-dimensional semantics, intensions are functions from worlds to extensions. Natural kind terms have two intensions, an A and C intension: the former are functions from worlds considered as actual to extensions, the latter are functions from worlds considered as counterfactual to extensions. The idea is that the a posteriori part of an a posteriori necessity is tracked by the C intension, while the A intension tracks the referencing fixing part—the part that tells us which actual properties we rigidify on—and hence, when combined with various empirical data, tell us the C intension. We gain access to the A intension of a term by considering our dispositions to pick out extensions at different worlds considered as actual. In the case of non-natural kind terms, the A and C intension collapse into a single intension that is a function from worlds considered as actual to extensions. In this sense, the A intension of our terms is revealed by our dispositions to react to situations under the supposition that those situations are actual. This does not mean that we have infallible introspective access to our meanings, for we might have imperfect access to our dispositions. But we have good enough ac3

Sider 2003.

105

cess. Our best guide to our concepts and meanings thus lies with our intuitions. If that is so, and it is not uncontroversial, what does it tell us about the usefulness of brute existential intuitions? In part, that depends on what we think about the metaphysics of our world. 4. Existence 4.1 The Absolute Quantifier Suppose that our world ‘naturally’ divides into objects, such that either metaphysics, or our best science, or the intersection of both,4 settle completely this division. That is to say, in some sense what objects there are is given to us by the world. Let us say that there is a complete division of a world into objects just if: Completeness: A division of a world into objects is complete just if (i) for any occupied space-time region R, there is a complete decomposition of R into parts5, such that each part of R is wholly occupied by some object and (ii) for any occupied space-time region R, it is determinate whether R is wholly occupied6 by an object and determinate whether R is wholly occupied by a proper part of an object.7 Intuitively, a division is complete just if it does not leave out any occupied regions, and there are no regions where it is unsettled what exists in that region. Then let us say that a division is metaphysically complete just if the division is both complete and natural.8 Since there might be more than one

4

Henceforth I will simply talk about what metaphysics settles, where I intend this to incorporate what science settles, or indeed, to be limited to what science settles if that is all that does the settling. 5 Where S is a complete decomposition of x iff every member of S is part of x, no members of S have any parts in common, and every part of x not in S has a part in common with some member of S. 6 Where an object wholly occupies a region just if all and only that region is occupied by that object. 7 I take it that (b) is redundant, since if the occupant of R is a proper part of some object, then it is itself an object. But not everyone agrees. 8 Choose your own account of naturalness.

106

natural way of dividing up a world, let us say that there is a unique division just if: Uniqueness: A division is unique just if either (i) all other divisions are non-natural or (ii) any other division is significantly less natural. Combining these two notions, we can say that a division is absolute just if it is both metaphysically complete and unique. If a division is absolute, then there is an absolute domain of objects. I think the following is a plausible claim about the semantics of ‘exists’: if there is an absolute division of a world, then the unrestricted existential quantifier is the operator that quantifies over all and only the objects in that division. What undergirds this analysis is a claim about eligibility, namely that if one of the potential meanings of some term is more eligible than the others because its extension is more natural, then the more eligible meaning wins out. In general I am not committed to the claim that naturalness guarantees greater eligibility, but I think it is a plausible view about the semantics of ‘existence’. Then if there is an absolute division, all and only the things in that division have the natural property of existence, and the unrestricted quantifier quantifies over this absolute domain of objects. Let us call such a quantifier an absolute (unrestricted) quantifier. The absolute quantifier has a univocal meaning: it is the operator that quantifies over the absolute domain. So any disagreement about what exists cannot be construed as a semantic dispute about the meaning of ‘exists’ (construed unrestrictedly). Rather, the disagreement must be a metaphysical disagreement about what is in the absolute domain. Now, armchair philosophers have never claimed that conceptual analysis is a good guide to the way the world is. Jackson explicitly distinguishes modest from immodest uses of conceptual analysis, favouring only the former. Modest conceptual analysis involves consideration of our dispositions to respond to various worlds considered as actual, revealing a function from worlds to extensions. Abstracting away from this set of worlds we develop something that might be called the conceptual analysis of the term—the reference fixing description in virtue of which we pick out an extension in each of these worlds. Given this analysis, we look to the empirical facts about the actual world to determine what, if anything, is the extension of the term in the actual world. Immodest conceptual analysis, on the other hand, combines conceptual analysis with a brute intuition to yield a claim about the world. For instance, it combines the conceptual

107

claim that an action is free just if one could have done otherwise, with the brute intuition that sometimes we are free, to conclude that determinism is false. The distinction between modest and immodest use of analysis is important. It is modest conceptual analysis that allows us to say that if there is an absolute domain of objects, then the unrestricted quantifier quantifies over that domain. We consider what we are disposed to say exists at worlds considered as actual, given that we discover there is such an absolute domain in these worlds. It is, barely, epistemically possible that were we to discover that there is such a domain, we would nevertheless maintain that the existential quantifier does not quantify over all and only those things in that domain. Then we have discovered that naturalness is not a constraint on the meaning of the quantifier. I think this is monstrously unlikely, but only on the basis of my firm views about our dispositions. None of this suggests that we are justified in using brute existential intuitions to determine what actually exists: it does not suggest that we are justified in using immodest analysis either to affirm that there is an absolute domain or to determine what is in that domain. Absent some reason to suppose that we have a reliable faculty for detecting objects in an absolute domain, there seems little reason to suppose that our existential intuitions are any guide to what exists. And there seems no reason to suppose that we have such a faculty. Williamson, however, disagrees, at least on one interpretation.9 According to Williamson, if we were sceptics about judgement—sceptics about whether we should think we are right in our application of concepts—than we would also have to be sceptics about perception. Consider the mereological nihilist, who rejects the existence of mountains. She concedes that it seems as though there are mountains and that it is, on occasion, a useful fiction to talk as though there are mountains. According to Williamson, in conceding that it seems as though there are mountains, the nihilist allows her scepticism about judgement—scepticism about the correct application of the concept ‘mountain’—to spread into scepticism about perception. For this scepticism infects standard perceptual judgements about the existence or not of mountains. The thought seems to be that if, under normal epistemological constraints, perceptual judgements are not subject to such scepticism (at least to the extent that they are good guides to the external world) then the same will be true about our judgements about concept application. Williamson concludes that our intuition that there exist 9

Williamson 2004.

108

mountains is prima facie good reason to think that there do exist mountains, and thus reason to reject compositional nihilism. Interpreted this way, it looks as though Williamson thinks that brute existential intuitions are a good (though defeasible) guide to what exists. If this is what Williamson intends, then, at least considered in the light of the absolute quantifier, the argument is unconvincing. This is not a case where scepticism about judgement infects our perceptual judgements. Everyone can agree that there are particulars arranged in a certain way, mountain-wise, or that a filled region of space-time has certain ‘mountainy’ properties. That is what we perceive. Whether those particulars compose something is not a matter for perception. We do not mistakenly perceive a mountain where there is none; we perceive a plurality, or a filled region of space-time, and mistakenly infer that there is a single composite entity. But that is not a perceptual error. So while we have faculties for detecting filled volumes of space-time that have certain distributions of properties, that is no reason to think that there are objects wholly occupying those regions that instantiate those properties. What then, is the role of conceptual analysis and intuition? Their role is to tell us what we would say about existence, under different metaphysical scenarios considered as actual. One of these scenarios is one according to which there is an absolute division of our world. Consideration of that scenario gives us part of the meaning of ‘exists’, it tells us that in such a world, existence is a unique natural kind property. Consideration of multiple such scenarios goes some way to giving us the intension of ‘exists’. But it does not go any way to giving us its actual extension. For intuitions cannot tell us either that there is an absolute domain of objects, and hence a unique natural kind property of existence, nor, if there is, what is in that domain. That is an a posteriori matter. Our dispositions to respond in worlds considered as actual reveals the intension of ‘water’, telling us it is, rigidly, whatever actually fills the lakes etc…. but these dispositions cannot tell us that what actually does that is H20, and hence that necessarily, water is H20. So too, we should not expect our brute existential intuitions to yield information about a posteriori existential truths. 4.2 The Non-absoluteN Quantifier It might be, however, that although there is a metaphysically complete division of our world, there is no unique division. Then there is no absolute domain of objects, and hence no absolute quantifier. So far I have sug-

109

gested that our concept of existence gives conceptual priority to an absolute quantifier: if there is an absolute domain of objects, then the existential quantifier quantifies over all and only that domain. But there is further work for some conceptual analysis here. If we were to discover that there are a number of natural ways of dividing up our world, what would we say? One possibility is that this would be the discovery that the unrestricted quantifier has, or can have, multiple meanings. The idea that the quantifier could, or does have, multiple meanings probably owes its origin to Putnam,10 and has recently been defended in more detail by Hirsch.11 The proposal currently under consideration differs substantially from that of Hirsch, however, who does not think that there is any metaphysically complete division of our world.12 What do we mean when we say that the quantifier could have different meanings? Suppose there are two metaphysically complete divisions of our world and that actually we quantify over all and only the objects in one of these divisions. Suppose there is a physical duplicate13 of our world— W*—in which agents quantify over all and only the objects in the alternative division. What are we disposed to say about the agents in W*? Are the agents in that world right, by their own lights, about what exists? That question is not whether they are right given that they use the term ‘exists’ to mean something completely different from us— zebra; going fast; spatially located etc. The question is whether we think that they are using an unrestricted existential quantifier, but using it to mean something different from us. In virtue of what, though, might we think that they have, in some good sense, the same concept of (unrestricted) existence as us. One way of cashing out conceptual similarity across agents in different worlds is in terms of shared A intensions. In a world where the watery stuff is XYZ, the agents share our concept of water if they reference fix on XYZ as ‘water’ via the same A intension as us: they think that water is the actual watery stuff, and the actual watery stuff turns out to be XYZ. But we cannot appeal to a shared A intension, for the worlds in question are physically identical in the relevant respects. All that differs are the semantic practices themselves.

10

Putnam 1988; 1987. Hirsch 2002. 12 As Hirsch would have it, ontology is shallow. 13 Or near enough, not a complete duplicate since these agents have different beliefs and semantic practices, but near enough to a duplicate for our purposes. 11

110

One way to cash out the sense in which the W* agents are genuinely using an unrestricted quantifier, is to hold that ‘(unrestricted) existence’ is vague, and the operator that quantifies over all and only the objects in the domain of one natural division represents one admissible precisification of ‘exists’, and mutatis mutandis for the other operator. So both we and the W* agents are using an unrestricted quantifier, for we are both using admissible precisifications of ‘exists’, and we are both right about what exists given what we mean by the quantifier. This seems like a reasonable proposal given what we have said so far about the meaning of ‘exists’. If we suppose that the existential quantifier quantifies over an absolute domain of objects if there is one—that the absolute domain is the most eligible candidate to be the meaning of ‘exists’—then it is plausible to suppose that if there is no unique most eligible candidate because there is no absolute domain, then ‘exists’ is vague between referring to multiple operators, each of which quantifies over all and only the things in each of the metaphysically complete domains. That, of course, is a matter for modest conceptual analysis. It depends on what we are disposed to say about the W* agents under the supposition that there are two metaphysically complete but nonunique ways of dividing up our world. We could hold that they are simply wrong: they are not using the unrestricted quantifier, because there is no alternative admissible precisification of ‘exists’. But that looks fairly implausible. The agents in W* could have been in our world (there could actually be such sub-populations). Indeed, our semantic practices could have developed differently, so that we all had the practices of the W* agents. But it seems unlikely that we would say that if our practices had developed that way, we would not have been using the unrestricted quantifier at all, any more than it seems right for the agents in the W* world to conclude that we are not using the unrestricted quantifier. That is to say, it seems odd to think that the semantics of ‘exists’ should so heavily rely on contingencies regarding which natural division we happen to track. If that is right, then in part the intension of ‘exists’ is: if there are metaphysically complete but not unique divisions of our world into objects, then it is vague which of these metaphysically complete domains the unrestricted quantifier quantifies over, that is, it is vague which of the precise quantifiers ‘exists’ picks out. Let us say that if there are a number of metaphysically complete divisions of our world, then the unrestricted quantifier is a non-absoluteN quantifier. It is non-absolute because it does not quantify over an absolute domain of objects. It is non-absoluteN because each of its admissible precisi-

111

fications are such that they quantify over all and only the objects in some natural complete division. So modest analysis tells us that if there are nonunique metaphysically complete divisions, then ‘exists’ is ambiguous between quantifying over each of those domains. But it is nature that determines which are the admissible precisifications for the quantifier, since nature determines which metaphysically complete divisions there are. Once again then, it is modest analysis that gives us part of the intension of ‘exists’—it tells us what we would be disposed to say if we discovered that actually, there are metaphysically complete but non-unique divisions of our world. But equally, none of this goes any way to telling us whether there are such natural divisions, or what they are. The considerations that render an appeal to brute existential intuitions unhelpful given an absolute quantifier, yield the same results if the quantifier is non-absoluteN. 4.3 The Non-absoluteS Quantifier We might deny, however, that there is any metaphysically complete division of our world. We could think that metaphysics fail to settle anything, but that view is surely so absurd in its far-reaching anti-realism as be repugnant even to most anti-realists. One way for the metaphysics to incompletely divide up a world would be if it completely divided up only a proper sub-set of space-time regions. More likely though, in an incomplete division there will be regions such that it is not settled whether that region is wholly occupied by some object or proper part thereof.14 For instance, a division might settle which fundamental particulars there are, but not which macro-objects there are. It would then be consistent with what is settled, that composition is restricted, unrestricted or nonexistent. Then, we might say, the metaphysics settles the base15 facts, and semantics settles the remainder by determining the best way in which our terms map onto the base facts. Let us say that under these conditions the quantifier is a non-absoluteS quantifier. Consistent with what is settled, whether the quantifier has multiple meanings and what those meanings are will be settled entirely by our semantics not by any properties of the world.

14

I take it this is what shallow ontologists such as Hirsch think. This is not to imply that this need be a bottom-up approach, ie that the facts in question are facts about the fundamental particulars in the sense of being micro fundamental particulars.

15

112

Let us say that a semantics is consistent just if it is consistent with what is metaphysically settled. And let us say that a semantics is complete just if there is some complete division of our world into objects, such that there is a one to one mapping of the terms in the semantics onto the objects in the division. If a semantics is complete, then it is determinate, for each spacetime region, whether it is wholly occupied by an object or proper part thereof: each division is perfectly precise. Now, it might be that our folk semantics is incomplete. If we try to map our semantics onto the world, we might discover that it maps onto an incomplete division. That would be the case if given the mapping, there are regions such that it is indeterminate whether an object or part thereof wholly occupies that region: it is vague what the borders of objects in the division are—or if there are regions such that it is vague whether any object occupies that region: there is vague existence. In what follows I consider only complete semantics, and assume for the sake of argument that there is at least some complete semantics that is close to our actual semantics, if the latter is incomplete. This allows us to abstract away from worries about ontic vagueness and vague existence, while still considering the role of existential intuitions in the context of something a lot like a FF restricted composition. At the end of the paper I will return to consider incomplete semantics and FF restrictions. Where the quantifier is non-absoluteS, for any world there will be many complete consistent semantics. Here is one possibility; each complete consistent semantics represents an admissible precisification of ‘exists’. Here is another possibility, only one, or some of those semantics represent an admissible precisification of ‘exists’. Consideration of the non-absoluteS quantifier reveals a dual role for conceptual analysis. First, does the semantics uniquely specify a meaning for the quantifier? If we think that agents in duplicate worlds with different complete consistent semantics are speaking truly about what exists (unrestrictedly), then we think they are using an admissible precisification of the unrestricted quantifier, and we think that semantics does not uniquely specify a meaning for the quantifier. If we think the agents in these worlds are speaking falsely about what exists (unrestrictedly) then we think they are not using an admissible precisification of the quantifier, and we think that the semantics uniquely specifies a meaning for the quantifier. For instance, if actually we adopt a complete consistent semantics according to which composition is unrestricted, then we might be disposed to say that agents in duplicate worlds are mistaken. The unrestricted quantifier is the operator that quantifies over the largest possible domain. If it

113

turns out that there are multiple complete divisions of our world, then our semantics uniquely settles that the operator that quantifies over the maximal complete division is the unrestricted quantifier. Agents using other complete consistent semantics are using perfectly metaphysically respectable divisions, but they are mistaken if they think they are using the unrestricted quantifier. Rather, we can interpret their utterances as true only so long as we are prepared to hold that these agents are tacitly restricting the quantifier in some way. So here are three competing analyses of ‘exists’: (E1) If there is an absolute division of our world, then the unrestricted existential quantifier is the operator that quantifies over all and only the objects in that domain; if there are metaphysically complete but not unique divisions of our world, then it is vague which of these natural domains is the one quantified over by the unrestricted quantifier; if there is no metaphysically complete division of our world, then it is vague which of the complete divisions is the one over which the quantifier ranges. Or, more straightforwardly, a priori conceptual analysis tells us that if there is an absolute division, then that division is the domain over which the quantifier ranges, if there is no absolute division, then the quantifier is vague between ranging over the objects in the domains of the most natural set of divisions. Conceptual analysis tells us that this is so if we are disposed to say that under the relevant metaphysical scenarios, the agents in duplicate worlds are using an admissible precisification of the quantifier. Conceptual analysis tell us that this is not the correct analysis of ‘exists’ if we are not disposed to think that those agents are using an admissible precisification. In that case, we might be inclined to embrace (E2), where the first two conditional clauses of the analysis remain the same, but the third clause reads: (E2) ….if there is no metaphysically complete division of our world, then the quantifier ranges over the maximal complete division. Alternatively, (E3) retains the same first two conditional clauses, but changes the third to read:

114

(E3) ….if there is no metaphysically complete division of our world, then the quantifier ranges over the complete division that is closest to the division over which actually, we are disposed to quantify. Notice that our brute existential intuitions have been of little value in constructing these analyses. What matters for the first two clauses of the analyses is not what we are actually disposed to say exists, but what we are disposed to say exists given various metaphysical hypotheses, and given what we would say about agents in duplicate worlds under these different hypotheses. Nor are they helpful in telling us what actually exists given that either of the antecedents of those first two clauses are true, since brute intuitions are no guide to a posteriori matters. So too the formulation of clause three of (E1) owes little to brute existential intuition. What matters is what one is disposed to say about agents in duplicate worlds: both a restricted and an unrestricted compositionalist might agree that duplicate agents are using an admissible precisification of the quantifier, while each might still have firm but competing brute existential intuitions. Similarly, someone who embraces (E2) may well be an unrestricted compositionalist on the basis that she holds the antecedent of the third clause to be true. But here the relevant intuitions are semantic intuitions about the admissible precisification of the unrestricted quantifier: namely that the unrestricted quantifier is maximal. Even the restricted compositionalist might agree with this, though she might reject unrestricted composition because she thinks that the antecedent of clause one or two is true, but that the relevant metaphysically complete division or divisions are not the unrestricted compositionalist’s division. This means that no existential intuitions of the unrestricted compositionalist kind tell one that (E2) is true, since (E2) need not vindicate unrestricted composition. (E3) is an attempt to explicate a view according to which if the antecedent of the third clause is true, then ‘exists’ in some sense rigidifies on our actual use. Of course, that’s not quite the right way to express the view, because everyone agrees that if the antecedent is true, then the meaning of ‘exists’ is settled by our use. But there are two senses in which it might be settled by our use. It might, as (E3) suggests, be settled solely by our dispositions to quantify over some actual domain. In that case, here is another way of interpreting Williamson’s argument. Our brute existential intuitions are a good guide to what exists, given a non-absoluteS quantifier, because they are a good guide to the meanings of the relevant terms, and hence to determining which complete semantics is closest to our folk semantics.

115

Then that we have brute existential intuitions that there are mountains is good reason to suppose that there are, insofar as it is good reason to prefer a complete semantics according to which there are mountains, than one according to which there are not. Another possibility, however is that our meanings are not settled only by our dispositions actually to quantify over objects—our brute existential intuitions—but rather, are settled by our broader dispositions to quantify over objects in other worlds considered as actual, and our dispositions to respond to agents in duplicate worlds. We might think of these postequilibration dispositions as existential intuitions, but not brute existential intuitions. (E3) tries to limit what semantic information we can avail ourselves of: it tells us that if the quantifier is non-absoluteS, then only our brute existential intuitions are relevant in determining the extension of ‘exists’. But this is either based on a general view about the meaning of our terms, and, by my lights, an implausible one: for this is to deny that the intensions of our terms are functions from worlds, plural, to extensions. Or it is an utterly unjustified brute claim about the meaning of ‘exists’. Failing that, if (E3) is true, it is because when we evaluate agents such as those in W*, we are disposed to say that they are straightforwardly wrong in their existential claims. And that requires considering our dispositions in such worlds. Determining the intension of a term crucially involves equilibrating over a domain of worlds. What we are disposed to quantify over actually, in the absence of such equilibration, might be very different to what we would be disposed to quantify over actually, in the light of our dispositions to quantify in other worlds considered as actual. So our untutored, brute existential intuitions are likely a poor guide even to determining which complete semantics is closest to our actual semantics. So if the quantifier is non-absoluteS, then the first role for conceptual analysis lies in formulating the third clause of an analysis of ‘exists’: it lies in determining whether our semantics uniquely settles the meaning of the quantifier or not, and if not, which are the admissible precisifications of the quantifier. In both cases, it is existential but not brute existential intuitions to which we turn. The second role for conceptual analysis lies in determining which possible meaning of the quantifier best captures actual use. If either (E1) or (E3) are the correct analyses, there is still an open question. Those who embrace (E2) have already answered this question: they hold that regardless of any brute existential intuitions we might have, given a nonabsoluteS quantifier, sufficient equilibration across duplicate worlds and

116

agents with different complete semantics will result in us being disposed to claim that the unrestricted quantifier ranges over only the maximal division. 5. Conclusion What does this mean for the argument for unrestricted composition? First we should note that to be an argument for unrestricted composition, rather than merely against FF restricted composition, we would need an additional premise to the effect that if there is no FF restriction, then composition is unrestricted. But there is no good reason to think that is so. If there is an absolute division of our world, or even some non-unique metaphysically complete divisions, there is no particular reason to suppose that one of those divisions corresponds to the division posited by the unrestricted compositionalist. So is this at least a good argument against FF restricted composition? There are really two threads to the argument. The first is the familiar thread according to which any FF restriction will entail that existence is vague, and that is incoherent. The second thread is that the FF restricted compositionalist must appeal illegitimately to brute existential intuitions. Let us suppose that our folk semantics is incomplete, and hence any FF restriction is one according to which the relevant division of our world into objects that maps onto our folk semantics, is an incomplete division. It is incomplete in the sense not that it leaves out regions (though perhaps it does) but in the sense that there are occupied regions of space-time such that it is either indeterminate whether some object wholly occupies that region (it is indeterminate what the borders of the object are) or indeterminate whether there is any object occupying the region (it is vague whether the object exists). I am inclined to say that there is no ontic vagueness in this sense, and there is no vague existence either: that is, talk of divisions that are incomplete in this way is talk of an incoherent possibility. If that is true, then this is reason to reject a FF restricted composition only on the supposition that there is no complete semantics that is sufficiently close to our incomplete semantics as to capture a sufficient amount of that semantics. That is an open question. I take it traditional unrestricted compositionalists think there is no such complete semantics, and if they are right, then FF restricted composition is incoherent. The second thread of the argument alleges that the FF restricted compositionalist illicitly appeals to brute existential intuitions. Something about

117

that allegation seems right. What is unclear is at what point the FF restricted compositionalist avails herself of the intuitions. At worst, she appeals to such intuitions as a way of building in constraints on composition where she takes it that there is an absolute or metaphysically complete domain of objects. Even considering our existential intuitions after the relevant equilibration across worlds will be no guide to composition under these circumstances. At best, we can think of the FF restricted compositionalist as assuming that there is a non-absoluteS quantifier. Given that assumption, appeals to intuition are par for the course: we need them to tell us if semantics specifies a unique meaning to the quantifier, and, perhaps, which of the complete semantics is closest to our actual incomplete semantics. Nevertheless, even under these circumstances it is not an appeal to brute existential intuition that does the work: it is an appeal to our postequilibration broader quantificational dispositions in a variety of worlds considered as actual. One suspects that these are not the intuitions that undergird the restrictions on composition that the FF restrictionist has in mind, and if that is so, then that appeal too is, if not unjustified, then at least fraught with danger.

REFERENCES Chalmers, D. 1996 The Conscious Mind, Oxford: OUP. - 2004 “Epistemic Two Dimensional Semantics”, Philosophical Studies, 118(1-2), 153-226. Jackson, F. 1998 From Metaphysics to Ethics: A Defence of Conceptual Analysis, Oxford: OUP. - 2004 “Why we need A-intensions”, Philosophical Studies, 118(1-2), 257-277. Hirsch, E. 2002 “Quantifier Variance and Realism”, Philosophical Issues, 12, Realism and Relativism, 51-73. Merricks, T. 2001 Objects and Persons, Oxford: OUP. Putnam, H. 1988 “After Metaphysics, What?” in D. Zimmerman and P. van Inwagen (eds.), Metaphysics: The Big Questions, Oxford: Blackwell, 388-392. - 1987) “Truth and Convention.” in D. Zimmerman and P. van Inwagen (eds.), Metaphysics: The Big Questions, Oxford: Blackwell, 392-398. Sider, T. 2003 “Against Vague Existence”, Philosophical Studies, 114, 135-146. Van Inwagen, P. 1990 Material Beings, Ithaca and London: Cornell University Press. Williamson, T. 2004 “Philosophical ‘Intuitions’ and Scepticism about Judgement”, Dialectica, 58(1),109-153.

118

Sortal Dependence of Persistence EDMUND RUNGGALDIER, Innsbruck

1. Introduction The main diverging accounts of persistence are due to two different interpretations of the so-called four-dimensional space-time-system. According to the position called “perdurantism”, every real entity is spread out in time, i.e., composed of temporal stages/parts. Everything is like a process or event and, thus, there cannot be any diachronic identity. Perdurantists have to reduce persistence to a kind of continuity relation between adjacent temporal stages. The continuity relation is weaker than identity and is neither reflexive nor transitive. It allows for differences of degree. According to the other position, called “endurantism”, besides events there are also endurers, i.e., things―human beings included―that go along with time and are always in the “now”. They are fully present at each moment of their existence and therefore neither extended in time nor composed of temporal stages/parts. Perdurantism conforms to conventionalist empiricism. In this tradition, it is commonplace to assume that individuals―people and other living beings included―have to be constituted/constructed from “the given”, i.e., the data of experience, ultimately from sense data. The constitution of individuals is conventional, its constraints being only pragmatic. Consistent conventionalism assumes this for the temporal dimension, too. There is no “fact of the matter” about diachronic identity. In this essay I want to argue for an endurantist account of persistence, first on the basis of the shortcuts or the deadlock of a consequent conventionalist approach und then on the basis of the presuppositions of acting and feeling, i.e., of the subjective experience of being agents. The wide range of the actual debate between perdurantists and endurantists shows that mere theoretic arguments balance each other. What tips the scale towards endurantism is the consideration of our indexical beliefs in acting and feeling and of our planning the future and remembering the past. If we as agents are endurers, so are our organs, and if our organs are endurers, so are our last constituent parts. The scientific view and account of the four-dimensional space-time systems is no obstacle for this onto-

logical position, since it imposes solely methodological restrictions. The four-dimensional system is useful for scientific purposes and for the representation of reality; its application does not, however, imply that the world itself is four-dimensional. It does not force us to accept perdurantism. I favor the thesis that all objects (simple objects included) are endurers/continuants, not only beings endowed with intentionality. An endurantist account of persistence as such, however, does not solve the problem of the identification and re-identification of one and the same continuant, viz. of tracing it through its existence. Furthermore, nothing exonerates the endurantists from elucidating the thesis that persistence as diachronic identity is given or primitive. In this essay I shall stress the crucial role of sortals and sortal expressions for the identification of individuals as continuants/endurers, i.e., as temporal units. There are no objects as such. Nonetheless, the sortal determination alone is not sufficient for tracing one and the same individual. For this purpose we need the individual concrete sortal, i.e. the sortaltoken. 2. Conventionalist deadlocks I will consider conventionalism first in the empiricist version of the early Carnap and then in the realistic version of Quine. The problems and deadlocks both of them face reveal the week spots of perdurantism, too. Due not only to empiricism but to Kantianism as well, many philosophers stick to the conventionalist thesis that diachronic identity is not “given” but has to be constructed or constituted. It seems to be common opinion―at least in the German-speaking world―that neither experience nor empirical knowledge forces us to accept transtemporal identity as real. The various versions of traditional conventionalism converge in the view that all the individuals we speak of in everyday life are conventional posits. The individuals as units in the spatial and temporal dimension are the result of linguistic and cultural conventions. This, however, does not imply that there are no constraints in forming temporal units; it only implies that these constraints are merely pragmatic and lack any cognitive or ontological basis. All knowledge of the world is due to experience, but in the tradition of empiricism this experience is exhausted by sense experience. Individuals, persons included, are not experienced, but constructed as sets of sense impressions.

120

Carnap was the first to develop a detailed constitutional system sketching how everyday objects are derived viz. constructed from a basis referring only to sense experience. The basis does not convey any knowledge of alleged temporal or tensed facts, since the phenomenal “given” does not contain any information on persistence relations. Carnap, however, cannot exclude the temporal dimension of reality altogether: it has an impact on memory and memory allows us to distinguish between similar and dissimilar experiences. He therefore chooses the recollection of similarity (Ähnlichkeitserinnerung) between fundamental or elementary experiences as the basic relation of his constructional system. (Carnap 1928, §61) By forming similarity classes of elementary experiences based on the recollection of similarity, he derives or constructs qualities via complicated steps. He then sketches how to construct visual objects, i.e., the spatial and temporal units of our everyday life, by applying the four-dimensional space-time system of Minkowski. He conceives of objects as bundles of qualities that remain in a stable relation to each other throughout the duration of their existence. “If, in a bundle of world lines which have been constructed … the neighbourhood relations remain at least approximately the same during a protracted stretch (of time), then the class of the corresponding world-points is called a visual thing.” (Carnap 1928, § 128) Due to the mentioned conventionalist assumptions, the qualities have to be assigned to the world points, i.e. lines of the four-dimensional system, solely on practical pragmatic constraints. Quine hints at the problem of this decisive step: “…qualities should be assigned to point-instances in such a way as to achieve the laziest world compatible with our experience.” (Quine1963, 40) Quine sees the weak spot of Carnap’s constructional procedure precisely in the connective “quality q is at x, y, z, t”, i.e., at a given point in the four-dimensional space-time system. “The connective ‘is at’ remains an added, undefined connective; the canons counsel us in its use but not in its elimination.” (Quine 1963, 40) Of course, if somebody is empirically minded, she must be tolerant as to the construction of objects, especially in their temporal dimension. Thus, she must accept many different alternative constitutions of objects as equally valid. But let us ask: are the factual constraints excluding alternative objects really only pragmatic? Aren’t they limits due to our experience and knowledge of the world? The conventionalist principles of simplicity and utility alone are not sufficient to exclude the construction of the most fanciful objects which have nothing in common with those of ‘our world’

121

except their emergence from a set of experiences form which only the visual qualities that we attach to them are derived. (Runggaldier 1984, 140) In spite of his critique of Carnap’s conventionalist connective that a quality is at a particular space-time-point, Quine too is conventionalistically minded in the account of the individuation of objects, i.e. of the spatial and temporal units of everyday life. The difference is due to his realistic conception of the distribution of stuff in the four-dimensional spacetime. It is a matter of fact, and thus “given”, that the basic elements of reality are spread out in space-time as they are. But what regions of the spacetime matter we choose as individuals, that is conventional. Thus, also for Quine our everyday objects are conventional posits. We can form scattered objects as well, i.e., objects interrupted in time. An object “comprises simply the content, however heterogeneous, of some portion of space-time, however disconnected and gerrymandered.” (Quine 1960, § 36) Each persisting concrete object should be thought of as a “space-time worm”, a thing spread out in all four dimensions. “Physical objects, conceived thus four-dimensionally in space time, are not to be distinguished from events or, in the concrete sense of the term, processes.” (Quine 1960, § 36) Thus, there can be no diachronic identity: Nothing is destroyed when stuff is rearranged, neither by putting things together nor by destroying them. For Quine there are no facts of the matter to identity. The last consequence of the conventionalist approach to the problem of persistence concerns our own personal identity: nothing cognitively relevant forces us to think of the identity of our own self in the way we do. For Quine there are no perplexities of personal identity: “…there is no reason why my first and fifth decades should not, like my head and feet, count as parts of the same man, however dissimilar.” (Quine 1960, § 36) However, a strict conventionalist account of persistence does cause perplexities. One doubts whether really nothing factual constrains our statements on the persistence of objects. The following thought experiment by Eli Hirsch will help outline some counterintuitive consequences of conventionalism. On the conventionalist assumptions, it should be possible to form alternative persons combining “temporal stages” in a different way. Hirsch has, thus, tried to see what such a possibility looks like by putting forward an alternative language differing from English solely in the formation rules of persons through time. The idea behind the alternative language that Hirsch calls “Contacti” is this: when two persons come into physical contact, each takes over the physical and mental characteristics of the other, and then when they cease

122

to be in contact they again regain their characteristics. They thus have to use personal pronouns in a different way, even though the rule for the use of “I” is the same, i.e., that a person must use it to refer to herself. (Hirsch 1982, 287ff.) In Contacti the term “person” will not refer to persons in our sense but to persons who swap their “temporal stages” during their contact. Let us suppose that A and B come into physical contact with each other, then A becomes B and B becomes A. The histories, therefore, of persons in our everyday assumptions do not coincide with the histories of the persons in Contacti forming sums of different temporal stages. The history of the alternative A will not contain the stages of our A when it is in contact with B, but instead of these, the stages of B. The same applies analogously to the history of the alternative B. If A is a man and B a woman, during the embrace the man would have to say in Contacti “before the embrace I was a woman” and the woman could say in turn “I was a man”. When A felt pain before the embrace, during the embrace he would have to say “the woman whom I am embracing was feeling pain and I can well remember what that feeling was like”. And she would have to say “I was feeling pain a few moments ago, or so I’m told”. If she were asked whether she could remember it, she would have to say, “Of course not! How could one conceivably remember while touching somebody what one felt when not touching anybody?” (Hirsch 1982, 289f.) In Contacti, the causal relations responsible for our memory and the connections between successive temporal stages in the life of persons have to be described in a radically different way. The peculiar concept of personal persistence presupposed by Contacti requires that the causal chains constitutive of memory can cross the boundaries between one person and another. All this sounds counterintuitive. Hirsch himself takes it as a fact that our way of conceiving personal identity excludes the possibility of speaking Contacti. He even grants that it is factually impossible to stick to a perdurantist view of personal persistence as defended by Parfit (Hirsch 1982, 311). But for Hirsch these are exclusively psychological facts; there are no ontological facts causing us to think in a particular way on persistence. For Hirsch, the difference between English and Contacti has no ontological but only psychological significance: We are innately disposed to think in the way we do. Against the background of everyday assumptions, however, I think we can and should

123

ask: Does this psychological fact not uncover something about the very nature of personal persistence through time? I have hinted at some of the consequences of a conventionalist account of persistence. One strategy for avoiding them consists in opting for a richer ontology allowing more ontological facts, viz. an ontology with endurers. 3. Reasons for opting for endurers The decisive reasons for postulating endurers, and thus genuine diachronic identity, are neither scientific nor the result of mere philosophical speculations. They are due to our subjective experience of being intentional agents programming our future and memorizing our pasts. The mere theoretic arguments in the debate between endurantism and perdurantism seem to balance each other. But the presuppositions of practical rationality, agency, subjectivity, indexicality support endurantism. In science we do not need endurers. In order to make scientific predictions and explain the various empirical phenomena associated with the apparent generation, existence, and destruction of living organisms, perdurantism would suffice. Accepting only stuff that constitutes them, along with its various properties, would do just as well. Thus, for scientific purposes perdurantist ontologies―be they ontologies of stuff, of tropes, or energy fields―might suffice. For an account of the ontological presuppositions of practical philosophy, however, we need more. The consideration of everyday talk about ourselves, especially talk about responsibility, supports the view that we need ontologies with endurers remaining identical throughout their existence. They are present at each moment of time as long as they exist. In positive science we do not only have to exclude subjectivity and indexicality, but diachronic identity as well. It should be easier to do scientific research by assuming that everything is extended temporally as well. This, however, can be interpreted as a methodologically convenient assumption. We are not bound to infer from this that there is no subjectivity and that there are no continuants. One of the main objections against endurantism is, however, due to the successful application of the four-dimensional space-time system. In fact, for representing processes and changes in science, it has proven extremely helpful to work with four dimensions: In addition to the three spatial dimensions—depth, length, and height—scientists use the temporal dimen-

124

sion, which permits the representation of various states of the world at different time instances. Scientifically it should be impossible to work without temporal vectors or without the four coordinates of the aforementioned system. It seems natural, therefore, to suppose that reality too is four-dimensional and that everything is spread out not only in space but in time as well. Accordingly, many philosophers are convinced four-dimensionalists and are bound by this to perdurantism. The perdurantist thesis that the world itself is fourdimensional is however not implied by the successful application of the system. The first interpretations of the system within the Vienna Circle were not yet four-dimensional in the ontological Quinean sense. For Carnap, for instance, the bundles of world lines of the system do not represent the temporal extension of objects but their ways or trajectories through time. The world points do not stand for the smallest “temporal parts” of particles, but for temporal moments of particles. And the world lines stand for the distances covered in time by particles. They indicate the relative position in space of a particle. A world line thus represents the “eternal destiny”—so to say—of a particle. (Minkowski 1911, 432) All the world points at one time instant represent the state of the universe at that time, and the world lines between two different time instants the changes in the distribution of particles from one time instant to the other. This interpretation of the system is endurantist, the particles being very small endurers going with time. The success of the application of the system in positive science does not force us to accept perdurantism instead of endurantism. Endurantism is at least compatible with it. 4. Disanalogies between space and time One common strategy to counter the thesis that everything is temporally extended and that, therefore, there cannot be diachronic identity is to resort to the disanalogies between space and time. However, the decisive disanalogies supporting endurantism again stem from practical philosophy, viz. from our experience of acting and feeling. In science we exclude indexical beliefs, but in everyday life based on our actions and feelings, the disanalogies between being located in space and being located in time are enormous. Indexical beliefs, especially tensed ones, trigger our actions. In order to understand why people do what they do in certain circumstances at a par-

125

ticular time, we try to find out what their indexical beliefs are. It makes a big difference whether the beliefs are third- or first-person beliefs, whether they are impersonal or indexical. As soon as King David realized that the prophet Naatan was talking about him and that he was the one, he was horrified about, he regretted his scandalous action. Indexical beliefs and the very concept of agency entail disanalogies between the ways in which an agent exists in space and in time. As agents, we are in a sense prisoners of time but not of space: We can choose where to live but not when, i.e., in what period of time. For an agent, time necessarily has a direction, space does not. Our use of the indexicals “now” and “here” is disanalogous. An agent can now make choices, deliberate about, and have intentions in respect to her conduct later but not earlier than now. She can however make choices, deliberate about, and have intentions in respect to her conduct in front of and behind her. (Gale 2002, 70) Further disanalogies stem from the axiological asymmetry between the past and the future of agents. An agent’s attitudes toward her past and future limitations are different. For instance, we regret that our existence does not extend beyond some time later than now but do not regret in the same way that our existence does not extend beyond some place behind here. Agents prefer their painful experiences to be in the past and their pleasant ones in the future. However, they do not have any analogous spatial preference. It doesn’t matter whether their pains and pleasures occur in any particular direction from here. Especially our emotional attitudes reveal disanalogies between space and time. In our practical lives we offer ‘the fact’ that something has happened, is happening, or will happen as reason not only for what we do, but also for how we feel: I hear that a loved one has died. I feel deep grief on hearing that he in fact has died. The problem for the four-dimensionalist is to show how anything like our current emotional life might be consistent with the claim that ‘past’, ‘present’, and ‘future’ events all have the same kind of reality and so, in themselves, have exactly the same significance. (Cockburn 1998, 84f.) We communicate about tensed events as reasons for feelings even if we do not know their objective dates: via indexicals we succeed in referring to the actual moment of time even if we do not know what time it is. We do not need any objective time-references for that purpose. In a sense, by using indexicals we even say more than by using their substitutes from the objective language of dates. Prior pointed out that it is impossible to say in

126

a tenseless way what we mean when we say, “Thank goodness that’s over!” My relief is not due to the fact that that thing takes place at a certain date, but that it is over. It is the overness or pastness of the thing that I am thankful for. If I were not convinced that it is a fact that it is over, I would not be relieved! The belief that something has happened in the past or will happen in the future provides us with reasons to act in a certain way, and to feel certain emotions. The mentioned disanalogies support the thesis that the temporal, especially the tensed dimension of reality is not on a par with the spatial dimensions. Our indexical beliefs and the experience of our acting and feeling favor the thesis that at least we as persons are continuants and are thus diachronically identical with ourselves. If there are good reasons for assuming that we as intentional agents are continuants, we have good reasons for assuming the same for our bodies as well. If our bodies are endurers, the main organs and their constituent parts must be endurers as well. Furthermore, by acting and feeling we presuppose that the others and the objects we deal with go with us in the course of time. We treat them as endurers and are convinced that that is the condition for getting acquainted with them. We conceive of and remember objects as lasting in time. This experience strengthens our everyday conception that not only we as agents having first person access to ourselves endure, but objects do so as well. We are firmly convinced that they continue to exist at least as long as we keep on perceiving them. In real life we do not construct objects in their temporal continuity out of momentary stages but, because of the mentioned experience, we presuppose that we can observe how they go on existing. We make our judgements on the identity of the surrounding animals and artefacts spontaneously and do not infer them from the temporal distribution of their alleged stages. If it is possible―as we have seen―to interpret the application of the four-dimensional space-time system as compatible with an ontology of continuants, overall we have good reasons to accept diachronic identity. The fact that the objects of empirical scientific research, the objects of natural science, are methodologically taken to be four-dimensional does not imply that they are extended in time. The thesis, however, that we and everyday objects are continuants does not spare us from asking how we can trace them as the same throughout their existence. Qualitative and spatio-temporal continuity conditions alone are neither sufficient nor necessary for tracing one and the same object

127

through time. Drastic deviations are possible without interrupting the career of an object. A minimal condition of adequacy for any analysis of our identity concept is that it implies the wrongness of drastic whole-part tracing confusions. I’ll try to argue for the thesis that the individual sortal, the individual sortal determination, guarantees the identity through time. 5. Sortals When we are asked what a given thing or animal is, we answer with a sortal expression. With sortal expressions we do not describe things but state what they are. Once a speaker understands the meaning of a sortal expression, she normally is in the position to identify the different instances of the sortal in question. This applies to the temporal aspect of the individuals as well, since sortal expressions convey the continuity- and persistenceconditions and thus, in a certain sense, the existence conditions in time too. Sortal concepts delimit what falls under them in a definite way and do not permit any arbitrary division of them into parts. If one grasps their sense, one is able to count the individuals that fall under them. It is however not easy to elucidate the requirements for the determination of a definite answer to the question: How many fs are there in region r at time t? The question is even more difficult if we take longer intervals for time t. It is linked to the question of when one individual ceases to exist and another begins. (Wiggins 2006, 75) Sortal expressions play a special role in realist as well as in conventionalist philosophies. In conventionalism, the rules governing the use of these expressions and the continuity conditions linked to them are taken to be conventional. In realist philosophies based on ontologies with endurers, however, the sortally determined individuals are centrs of action, powers, capacities, potencies, and functions. The persistence conditions of objects, humans and other organisms included, do not result from conventional posits or linguistic conventions; they are “given”. One has, thus, to distinguish between the stuff making up a substance and the sortally and functionally determined thing or living being itself. The tree and the aggregate of cellulose molecules that constitute that tree are not the same thing, even though they occupy the same region. Objects are not identical with the material they are made of, or with the sum of their parts. They are not exclusively determined by their matter. The kind of composition, the mode of activity, their way to act and react,

128

their basic capacities, their dispositional properties and functions are constitutive for them as well. Sortal expressions refer to these basic traits of the instances falling under them and thus convey their persistence- and continuity-conditions too. They express these individuals’ norms of coming into being and passing away. Of course, the concrete history of the development through time of an individual is contingent upon the environment. But the basic tendencies and the boundaries of the development of an individual depend on what it is, i.e., its sortal. It is only in virtue of grasping the sortal of an object that we can keep track of it through change and distinguish between doing this correctly and making a mistake in doing it. (Wiggins 2001, 59) We can neither identify nor re-identify an object as the same without a sortal rule or the command of a sortal expression. Even ordinary countability depends on the sortals of things: in order to be able to count objects, one has to count one and the same object only once over time. Without an understanding of diachronic identity, there would be no reason not to count the same thing twice. This is only excluded because of the ability to reidentify an object over time as being one and the same . We have to distinguish between sortally determined objects and parcels of matter. The latter are not objects in a strict sense. Even where they cohabit, the ways in which they persist through time differ widely. (Runggaldier 1998, 366f.) For example, what destroys one may entail only a slight alteration in the other. Objects can go out of existence without their parts ceasing to exist; on the other hand, their parts may go out of existence without thereby destroying the objects. The conditions for alteration and identity in time for objects are different from those that apply to mere parcels of matter. Identity in time is more vulnerable for the former than for the latter. The sameness of material composition is not necessary for the identity through time of the objects it makes up. Organisms renew their cells and artefacts are often gradually altered in material parts. On the other hand, retention of material composition, or sameness of material, is not sufficient for the identity of an object, even though it is such for chunks of matter. Melting down a statue destroys it despite the fact that its material composition remains the same. 6. Sortal dependence of the identity of objects For endurantists, diachronic identity is primitive. It cannot be reduced to a more basic relation. By contrast, for perdurantists the relation that in eve-

129

ryday speech is called diachronic identity has to be explained away or reduced to a different type of relation, some weaker continuity relation among adjacent temporal stages. However, endurantists too are confronted with the question of explaining the notion of diachronic identity. That a concept is primitive does not imply that one cannot explain or elucidate what its content is. If an entity is an endurer, it endures throughout its history from the beginning to the end. That it is an endurer implies that it can change and evolve during the whole period of its existence. For this to be possible, something has to guarantee its identity. It is the sortal or the essence of the entity to guarantee and determine an object’s identity through time: a and b are identical iff there exists a sortal f such that 1) a and b fall under f; 2) a is the same f as b, that is, coincides with b under f in the manner of coincidence required for members of f. (Wiggins 2001, 56) It is necessary to distinguish between the claims that for any continuant there is at each moment of its existence a sortal under which it falls and that there is one sortal such that at each moment of its existence a continuant falls under it: 1) (x) (t) [(x exists at t) → (∃g) (gx at t)] 2) (x) (∃g) (t) [(x exists at t) → (gx at t)] The first version does not guarantee the availability of one sortal for the whole duration of a continuant. We need the second version to explicate the thesis that the sortal of an object guarantees its diachronic identity. We have seen that there are limits to the conventionalist freedom of finding sortals under which continuants fall. Indeed, if we could invent sortal concepts simply at will, “then the real content of the assertion that something lasted till t and then ceased to exist would be trivialized completely.” (Wiggins 2006, 65) We can neither prolong nor abbreviate the duration of the existence of continuants merely by applying different sortal concepts. For instance, ―as Wiggins has pointed out―we certainly cannot prolong the existence of Lot’s wife who was transformed into a salt pillar simply by inventing the new sortal “woman-pillar”. (Wiggins 2006, 64f.) If, on the other hand, the range of possible answers to the question what is x? is limited, we have to state what makes the difference between genuine sortal expressions referring to the basic sortal constitution or essence of an object and the countless expressions which only look like sortal expressions. Genuine sortal expressions mark “what a thing is, what matters turn on with regard to its persistence and what matters turn on with regard to identity claims relating to it”. (Wiggins 2006, 69)

130

A genuine sortal must, at any rate, be relevant for determining when a continuant falling under it really ceases to exist. It allows not only to single out the object and distinguish it from other objects of the same kind, but to trace it through time as well. It allows determining―at least to a certain extent―what can and cannot befall it and thus what changes it can tolerate without ceasing to exist. Genuine sortal concepts, first of all, refer to natural kinds. They express a principle of activity or of operation of members of their extension. Decisive for the persistence conditions of natural kinds is the Aristotelian intuition that living beings have within them a principle of change and of staying unchanged, with respect to place or with respect to quantitative change, as in growth and decay, or with respect to qualitative change. (Aristotle, Physics II, 1) There is a link between the relevant questions of identity and how they behave and what they can do. When we talk of natural-kind terms or expressions, we presuppose that they do not function as terms referring to sorts on the basis of some nominal essence or some previously fixed necessary and sufficient conditions. They are not fixed a priori but have to be revealed by study of the typical exemplars of the kind in question. For the purpose of this essay, it is important to notice that if there are law-like principles determining the extension and exemplification of natural kinds, these must determine the development and the history of the exemplars of the kind as well, “they must also determine directly or indirectly the characteristic development, the typical history, the limits of any possible development or history … of anything that instantiates the kind”. (Wiggins 2006, 84) If we want to abide by the thesis that the sortal sameness is a necessary and sufficient condition for diachronic identity of an object we have to add that this is appropriate only for sortal-tokens. In analogy to the distinction between properties as universals and concrete properties, we have to distinguish between sortal-types and spatio-temporally determined sortaltokens. The sortal as type is not enough to guarantee individuality and identity through time. Every individual of a given sortal/kind shares with every other all the features that together characterize this kind. The supplementation with spatio-temporal continuity is also not sufficient. The reason for this is that succeeding portions of this spatio-temporal path under the same sortal can be occupied by different particulars. “A continuous object may become a different one, losing its identity without changing its kind.”

131

(Denkel 1996, 208) The identity of living beings allows for far more alterations than the identity of artefacts. But, in order to maintain their identity, they have to fall under the same sortal-token. Living beings may have disproportionate growth, loose some parts of their body, which may become dysfunctional, they may even generate altogether new parts for the individual. But even in these cases they would maintain the same sortal token as long as they live. The sortal, however, indicates limits of change beyond which the form-token will be lost.

REFERENCES

Carnap, R. 1928 Der logische Aufbau der Welt, Wien: Felix Meiner Verlag. Cockburne, D. 1998 „Tense and Emotion” in R. LePoidevin (ed.), Questions of Time and Tense, Oxford: Clarendon Press, 77-91. Denkel, A. 1996 Object and Property, Cambridge: Cambridge University Press. Gale, R.M. 2002 „Time, Temporality, and Paradox“, in R. M. Gale (ed.), The Blackwell Guide to Metaphysics, Oxford: Blackwell, 66-86. Hirsch, E. 1982 The Concept of Identity, Oxford: Oxford University Press. Minkowski, H. 1911 Raum und Zeit, Gesammelte Abhandlungen. Band 2. Leipzig. Quine, W.V.O. 1960 Word and Object, Cambridge/Mass.: M.I.T. Press. - 1963 From a logical point of view, N. York: Harper. Runggaldier, E. 1984 Carnap’s Early Conventionalism, Amsterdam: Rodopi. - 1998 “Sortal continuity of material things”, Erkenntnis, 48, 359-369. Wiggins, D. 2001 Sameness and Substance Renewed, Cambridge: Cambridge University Press.

132

Beyond Endurance and Perdurance: Recurrent Dynamics JOHANNA SEIBT, Aarhus

1. Introduction During the last decades the so-called ‘problem of persistence’ has been continuously—persistently, one is tempted to say—in the focus of the research debate in analytical ontology. There are impressive advances in finessing positions and arguments, but essentially the debate seems caught in a deadlock between, on the one hand, identity-based accounts suitable for the ontological interpretation of common sense (but not science), and, on the other hand, relation-based accounts suitable for the ontological interpretation of science (but not common sense). Since in philosophy a theoretical impasse of this kind is commonly generated by a joint implicit assumption, one might try to step out of the debate, identify questionable presuppositions, and approach the topic with fresh eyes. This is the task of this paper. In fact, as it appears, participants of the debate about persistence share not only one problematic assumption, but overlapping parts of an entire network of presuppositions. Elsewhere I have shown this network of presuppositions to be operative not only in the debate about transtemporal identity (‘persistence’), but also in the discussion of numerical identity (‘individuation’) and qualitative identity (the debate about ‘universals’). As I have argued, these presuppositions delimitate a research paradigm that has informed the ontological discussion from Aristotle to the present day, the ‘substance paradigm.’1 Ontological presuppositions are hidden axioms of inquiry that in most cases cannot be shown to be false in any straightforward sense. At best one can show that certain theoretical set-ups in ontology are not well suited to fulfilling their explanatory tasks. In fact, the presuppositions of the substance paradigm that drive the debate about persistence present a particularly striking case in this regard—they make it particularly difficult to devise a coherent ontological interpretation of persistence and change 1

Seibt 1990; 1996 a,b,c; 2005.

while taking the ‘logic’ (inferential role) of our statements about persistence and change at face value. Fortunately the presuppositions of the substance paradigm are by no means ‘laws of thought’ but a matter of theoretical habituation only. We can give up on the substance paradigm, on any and even on all of its constitutive presuppositions, and adopt new principles of theory construction in ontology. The theoretical choices in analytical ontology are more variegated than the current debate suggests. The purpose of this paper is to introduce the reader to a new account of persistence developed within such a new ontological framework, General Process Theory (GPT), based on a new ontological category or type of entities called general processes, or, less ‘technically,’ dynamics. Unlike other attempts to promote process-based descriptive frameworks, GPT is firmly committed to the methodology of analytical ontology. Within GPT persistence is the strict identity or sameness in time of an entity that occupies a spatiotemporally extended region. Thus the new account straddles the traditional juncture between endurance and perdurance theories, but in the sense of leaving both of them behind, since it operates with a different notion of individuality that is not tied to particularity (i.e., necessary unique locatedness). To motivate the new approach I begin by highlighting some of the presuppositions that constrain the solution space for the ‘problem of persistence’ in questionable ways. Then I offer a sketch of the core ideas of General Process Theory, and finally I adumbrate how ontological counterparts (‘truth-makers’) for assertions about persistence and change can be formulated within this framework, highlighting some of advantages vis-à-vis extant endurance and perdurance accounts. 2. The problem of persistence Contemporary analytical ontology follows a general methodology that can be reconstructed as follows. An ontology is a theory of truth-makers for the sentences of a certain natural or scientific language L.2 An ontology 2

The term 'truth-maker,' originating from Husserl, has been brought again into currency by Mulligan/Simons/Smith (1984). A 'truth-maker 'is any kind of entity "in virtue of which sentences and/or propositions are true" (ibd. 287), e.g. objects, properties, events, state of affairs; truth-makers contrast with 'truth-bearers,' which are any kind of entity which 'has' the properties of being true or false, e.g., a sentence or proposition. As I use the term, the truth-maker of a sentence p is described by the right-hand side of the Tarski-biconditional for p, where p is a logically atomic

134

describes what there is in the world-of-L-speakers—i.e., in the world as speakers of L do or can conceive of it—that makes the true sentences of L true. The structural descriptions of truth-makers provide explanations for why L-speakers may justifiedly draw certain, highly general, ‘material’— call them ‘categorial’—inferences. For instance, an ontological theory aims to explain, by means of a structural description of a truth-maker, why speakers of common sense English are entitled to infer from the true sentence ‘this tree is tall’ that the denotation of ‘tall’ but not the denotation of ‘this tree’ may occur multiply in space, while this is not so for the denotation of 'this tree.' An ontology thus provides structural descriptions of the domain of a language L in terms of which truth-makers for atomic L-sentences are defined that can justify the categorial inferential role of these sentences—or even more briefly, an ontology is an explanatory theory of the categorial inferences licensed within a certain language L.3 Ontological investigations of persistence are typically focused on the interpretation of statements about the existence of things in time, i.e., assertions about change and sameness such as: [1.] The Statue of Liberty was first copper-colored, but turned green over the course of the years. [2.] The statue you see now is the same as the statue that was being dedicated in 1886. One may suspect that the traditional focus onto things amounts to a problematic heuristic restriction, and below it shall become clear that this is indeed so. But since things do play an important (if arguably not primary) role in our interaction with the world, ontological concern with the persistence of things can easily be motivated. In fact, the successful interpretation of statements about the persistence of things (and persons) might well be considered a necessary condition for the acceptability of any ontological framework, or so I shall assume in the following. Statements like [1] and [2] invite philosophical puzzlement. A statement about change like [1] seems to assert that some thing is different now from the one it was then, and yet is the same one thing. A statement sentence or proposition; negated or logically complex sentences do not have truthmakers. 3 More precisely, the categorial inferences explained by an ontology are typically licensed not by one single language but by a class of intertranslatable languages, i.e., by a conceptual scheme in Wilfrid Sellars’ sense. For a detailed characterization of the methodology of ontology as “theory of categorial inference” see Seibt 2004, part 1, where I also explain why and in which sense ontological theories are metaphysically neutral.

135

about sameness in time like [2] seems to assert that what appears to be ‘two things’—the statue that was dedicated in the past and the statue you are looking at now—somehow are nevertheless one and the same thing. The task of an ontological theory of persistence of things is to remove these apparent tensions. As a first step one might distinguish between two senses of ‘sameness’: ‘qualitative identity’ versus ‘numerical identity’. An assertion about change, as in [1], states a difference that pertains merely to the features of a thing, to its qualitative identity, while the numerical identity of that which said to change remains unaffected. Similarly, an assertion about sameness in time as in [2] is a claim about numerical identity of something referred to with two definite descriptions highlighting differences in qualitative identity only. However, this first step of ontological analysis leads into trouble as soon as the concepts of numerical and qualitative identity are definitionally connected, as postulated in Leibniz’ “principle of the indiscernibility of identicals”, in standard systems of predicate logic known as “the Leibniz Law”. The basic problem for an ontological account of the persistence of things thus is frequently stated in terms of the incompatibility of the following principles: (Endurance: Persistence is identity) If a thing α is said to persist through the temporal interval from t1 to t2, then α as it exists at t1 is numerically identical with α as it exists at t2, (and the same holds for all intervening points in time). (P1: State analysis of change) Change is the qualitative difference of particulars. That is, if a thing α is said to change during the temporal interval from t1 to t2, then α as it exists at t1 is qualitatively different from α as it exists at t2. In other words, predicates that truly apply to α as it exists at t1 do not apply to α as it exists at t2, and vice versa. (Leibniz Law: numerical identity implies qualitative identity) If two names or definite descriptions ‘A’ and ‘B’ denote one thing, i.e., if the denotations of ‘A’ and ‘B’ are numerically identical, then all predicates true of the denotation of ‘A’ are true of the denotation of ‘B’. Reversing it by contraposition, the Leibniz Law says that qualitative difference implies numerical difference: if different predicates can be truly

136

applied to the denotations of some two names or definite descriptions, these cannot denote one entity. The difficulty is by now so familiar that participants in the debate about persistence are content quickly to gesture towards (P1) through (P3) before proceeding to the discussion of a solution strategy. However, upon closer analysis the three principles are in conflict only if they are combined with the following principles, which are commonly not mentioned but presupposed: (P2: The subjects of changes persist) If a thing α persists through a change C, then that to which we attribute persistence is the same entity as that to which we attribute change. In other words, (P2) ensures that when we assert of the statue in [1] that it has changed, and in [2] that it has persisted, the grammatical subjects of these two statements denote one and the same entity. Thus with (P2) we cannot say that what is said to persist is numerically identical in time while what is said to have changed is not. Moreover, with some exceptions, participants in the debate quietly assume that numerical identity is not something an entity could lose or gain over time: (P3: Numerical identity is temporally non-variant) Numerical identity is a relation that applies both at and across time; but it is not relative to time.4 This principle rules out the possibility that the numerical identity of an entity might change over time, such that at times two denotations of names for things might be said to be identical but not at other times. More importantly, it undermines the two most straightforward strategies of eliminating the conflict between the endurance interpretation of persistence, the Leibniz Law, and (P1), the state analysis of change. The first strategy is to restrict the Leibniz Law to judgments about numerical 4

Some authors discuss the question of “temporary identity” but this pertains to the temporary identity of items that are traced under different sortal concepts, such as the temporary identity of a statue-shaped item and a lump of clay (cf. e.g. Gallois 1990). In contrast, (P3) is a principle about numerical identity under one sortal concept—it says that if some x is one F then this holds for all times at which x exists. Note also that a more precise formulation of (P3) should formulate identity as coreference rather than as a relation (two names or definite descriptions denote one and the same entity).

137

identity at a time. But given (P3) it would be an odd move indeed to claim that even though numerical identity at a time and numerical identity across time are the same relation, the axiomatic characterization of that relationship afforded by the Leibniz Law, the only distinctive axiom of identity we have in logic, does not apply to all of its instances. The second strategy consists in restricting the quantification within the Leibniz Law to predicates denoting the essential features of an entity. This restriction makes—prima facie at least—good sense for numerical identity across time: an entity persisting through a change can be said to retain all its essential features, while the change pertains to contingent features. But then the Leibniz Law could no longer, by itself, serve as an axiomatic characterization of numerical identity at a time, for it would allow for one and the same entity to have contradictory contingent features at the same time. The obvious remedy of adding an axiom for numerical identity at a time would amount to a diversification of numerical identity, which is discouraged by presupposition (P3). In sum, in so far as participants in the debate about persistence have felt a more profound tension between the first principles listed above, this must be attributed to implicit commitments to presuppositions (P2) and (P3). In combination the five principles just considered generate what one might call the ‘basic apory of persistence’, a constellation of incompatible yet prima facie plausible tenets. The basic task for a theory of persistence thus consists in deciding whether the initial conflict between these principles can be removed by reinterpreting one of the principles or, alternatively, which one(s) of the conflicting principles ought to be rejected. These two options coincide with the main juncture in contemporary approaches to persistence. So-called ‘endurantists’ hold on to the thesis that persistence is numerical identity but insists on a clear analysis of temporal predications. Once these are in place, so runs the argument, statements about an entity’s changing are no longer generating any difficulty for the identity reading of persistence. For example, endurantists have suggested that the predicates in terms of which we describe a change (‘was green but is now red’) denote time-indexed features: A ‘complete’ property of an individual is a compound temporalized (or dated) property, like being green in August 1965 or being born in 1769,

138

and there is no question of an individual changing in the sense of once having and later lacking these compound properties.5 Since the thing that is said to persist through a change has all of its timeindexed properties at all times of its existence, persistence can be interpreted as numerical identity across times without getting into conflict with the (unrestricted) Leibniz Law. The only difficulty with this move is that to the extent to which it succeeds in ‘rescuing’ identity, it fails in preserving the phenomenon of change. We might try and say that change is a matter of time-indexed properties ‘manifesting’ itself at certain times but not at others—a leaf’s being green on August 8 is manifest exactly on August 8 while it’s being red on September 24 is manifest exactly on September 24. and so on. But such second-order properties of being manifest at t surely should also be covered by the Leibniz Law and thus we are back at the original ‘apory’. To sacrifice change for the sake of the identity interpretation of persistence does not seem an attractive choice. A successful ontological account of persistence should preferably not have any metaphysical implications in the sense of declaring as false sentences about change or transtemporal difference that we take to be true. A successful account of persistence should have a way of explaining what is meant by a true sentence like [1] which says that the properties of the denotation of ‘the roof’ at t1 differ from the properties of the denotation of ‘the roof’ at t2. Endurantists commonly—and often with surprising historical nonchalance—hold that the entities said to be numerically identical through time are “substances in the old sense”,6 or postulate a category of “continuants” of concrete, particular ‘thing-like’ constituents that are the logical subjects of our predications. But there is no reason to believe that only thing-like ontological constituents could be numerically identical in time. Peter Simons has argued that within an ontology of tropes or property instances the traditional notion of a substratum can be refurbished in new and productive ways.7 On Simons’ account things are ontologically constituted by a collection of tropes that can be divided into a “nucleus” and a “periphery”. That which remains identical when a thing is said to change is a collection of tropes forming the “nucleus”, while the “periphery” may at different times hold different tropes. Simons’ trope5

Cf. Wilson 1956, 48. Prior 1965, 63. 7 Cf. Simons 1994. 6

139

based endurantism most helpfully highlights that the linkage between substance ontology and endurantism is a matter of theoretical habit and not of conceptual necessity. However, to discharge its explanatory tasks, tropebased endurantism still owes us a clear account of the transtemporal identity of tropes themselves—if such property particulars are individuated by their spatiotemporal location, how could they ever move? Furthermore, given that the “nucleus” of trope bundle X is to contain all that which is essential to the thing associated with X, how should we account for essential developments in living things?8 The other main road out of the ‘apory of persistence’ is to deny that persistence is a matter of numerical identity, i.e., to reject the endurance view of persistence. Instead, so-called perdurantists claim that persistence is a mereological relationship: assertions like [1] and [2] above are made true by relations between the temporal parts of a fourdimensional entity. The terminological contrast between the endurance and the perdurance view of persistence goes back to Mark Johnston, but is commonly cited in D. Lewis’ pithy formulation. [S]omething persists iff, somehow or other, it exists at various times; this is the neutral world. Something perdures iff it persists by having different temporal...stages at different times...it endures iff it persists by being wholly present at more than one time.9 Philosophers are attracted to the perdurance view and the ontology of socalled “fourdimensionalism”, since it derives its original motivation from the scientific description of natural happenings in special and general relativity theory as taking place in, or constituting, a fourdimensional structure of spacetime.10 Perdurantists differ, however, on the kind of conceptual adjustments that are to be made to the fourdimensionalist description of nature in relativity theory. For example, some perdurantist 8

Cf. Seibt 2005. Lewis 1986, 202. 10 Note, though, while the compatibility with contemporary science certainly should count as an important asset, the ontological intimations contained in scientist’s socalled ‘instrumental glosses’ cannot be turned directly into an ontological theory. A fourdimensional manifold of point events fulfilling certain topological and metrical conditions might suit the interpretation of physics statements about idealized pointmasses, but it does not as such lend itself to the ontological interpretation of common sense statements about the persistence of things through time and change. 9

140

have suggested that statements about change and persistence as in [1] and [2] above should be taken to be about fourdimensional wholes: (“Hunk” View): The expression ‘the Statue of Liberty’ in sentence [1] denotes a fourdimensional entity L that has temporal parts l1 and l2, standing in a relation R of temporal distance, and l1 is copper-colored while l2 is green. The definite descriptions in sentence [2], viz. ‘the statue you are seeing now’ and ‘the statue dedicated in 1886’ denote two different temporal parts li and lj of L and state that these are temporal parts of L.11 The three main difficulties arising for this suggestion come out most palpably in Quine’s early formulation of this version of fourdimensionalism: A physical thing—whether a river or a human body or a stone—is…a sum of the temporally small parts which are its successive stages.12 If we identify things with their de facto spatial and temporal extension sub specie eternitate, i.e., with the material fillings of certain spatiotemporal regions, or with materially filled spatiotemporal regions, we are—in the first instance at least—committed to assuming that (a) all statements about things changing are analytic, (b) that the denotations of most names for things are inscrutable from the speaker’s point of view, and (c) that the predicates for things can be said to express properties of stages.13 The first two of these problematic commitments can be avoided once the dependence relations between temporal wholes and temporal parts are reversed. On the ‘hunk view’ the denotation of ‘the Statue of Liberty’, i.e., a fourdimensional whole (“hunk of matter”) is taken as basic; from this the notion of a stage or temporal part is derived, yielding denotations of temporally relativized definite descriptions such as ‘the statue you are looking at now’. In contrast, proponents of the so-called ‘stage view’ postulate as basic entities temporal stages, the denotations of temporally relativized definite descriptions, and then define the denotations of simple names such as ‘the Statue of Liberty’ in terms of sequences of such stages.

11

Cf. Heller 1990. Quine 1952, 210; emphasis supplied. 13 Cf. Lewis 1988, 66: “When I sit and then stand, bent stages are followed by straight stages. Each stage has its shape simpliciter.” 12

141

(Stage View) The definite descriptions ‘the statue you are looking at now’ and ‘the statue that was being dedicated in 1886’ in sentence [2] denote stages s1 and s2 of a thing that has the essential properties Pi of a statue, and there is a spatiotemporally continuous sequence S of stages all of which have the essential properties Pi that connect s1 and s2. Statement [1] says that there are two stages si and sj in this sequence S that stand in temporal relation R, and si is copper-colored, while sj is green.14 However, even though proponents of the stage view can dodge the mentioned problematic commitments (a) and (b), they still endorse assumption (c), namely, that predicates for things can be said to express properties of stages. As I will argue now, this assumption renders the stage view ultimately incoherent. Let us first note that assumption (c) follows directly from presupposition (P1), the ‘state analysis of change.’ More carefully elaborated, (P1) combines actually two tenets: (i) the assumption that assertions about changes in things are best analyzed in terms of pairs of contrary or contradictory properties exemplified by a particular, and (ii) the assumption that that which is said to undergo change—the subject of the change—is the logical subject of the predicates in terms of which the change can be analyzed. In other words, whether championing the hunk view or the stage view, proponents of perdurantism are still committed to the idea that if the Statue of Liberty is said to turn green, and this change can be analyzed in terms of the properties of being-copper-colored and being-green, then at the beginning of this change there is a particular which exemplifies being-copper-colored and at the end of this change there is a particular (i.e., an entity that necessarily occurs at any time in one spatial location only) which is the logical subject of being-green. While endurantists subscribe to the state analysis of change in its strong reading: ‘∃x(at t1 F(x) and at t2 ¬ F(x))’, perdurantist operate with the weaker reading: ‘at t1 ∃x F(x) and at t2 ∃x ¬ F(x)’, but in each case the variable ranges over a particular that can serve as the logical subject of the predicates (exemplify the properties) in terms of which the change is expressed (can be analyzed). 14

Cf. e.g. Hirsch 1982, Lewis 1983. More recently the “stage theory” has been contrasted—somewhat confusingly perhaps—with “perdurantism”, as the combined thesis that (a) all temporal predications are about stages, but (b) only present stages exist, i.e., a tensed theory of time: cf. Hawley 2001.

142

The fact that perdurantists presuppose the core intuition of the traditional ‘state analysis of change’ generates a profound difficulty, however. For in order to render plausible the claim that those properties in terms of which we describe changes in things can be ascribed also to temporal parts, we need to assign to temporal parts temporal extension. Consider the following characterization of a ‘stage’. [A person stage] does many of the same things that a person does: it talks and walks and thinks, it has beliefs and desires, it has size and shape and location. …[A person stage] even has temporal duration. But only a brief one for it does not last long. It begins to exist abruptly, and it abruptly ceases to exist soon after. Hence a stage cannot do everything that a person can do, for it cannot do those things that a person does over a longish interval.15 If temporal parts are momentary or temporally extended entities, they persist not only through time but also through change; in particular, a stage which “thinks” and “walks” certainly undergoes genuine alteration. But persistence through change is precisely the explanandum of the perdurance theory and thus should not appear in the explanans. Thus, in so far as stages are conceived of as proper candidates for logical subjects of predicates in terms of which we express changes in things, they must be taken to be temporally extended, in which case the perdurantist seems to be caught in a question-begging maneuver. On the other hand, if stages are taken to be instantaneous, almost none of the properties in terms of which we describe the change of a thing, could be ascribed to a stage (the only exception may be locational properties). To throw the difficulty into sharper relief, consider the following central postulate of the stage view: A sufficient condition for the succession S of object stages to correspond to states in the career of a single persisting object is that: (1) S is spatio-temporally continuous; and (2) S is qualitatively continuous; and (3) there is a sortal term F such that S is successsion of F-stages.16 15

Lewis 1983, 76. The difference between thing stages and person stages does not matter in the context of the present argument. 16 Hirsch 1982, 36.

143

According to this postulate stages must have the same sortal properties of the things they compose – stages of a horse, a rose, or a rock must exemplify the properties of being a horse, a rose, or a rock. But being a horse, rose, or rock arguably takes a certain amount of time--it requires certain physiological, chemical, or quantum-physical processes. Thus the properties of being-a-horse, being-a-rose, and even being-a-rock cannot be exemplified by an instantaneous entity. To be sure, one might, in analogy to the concept of 'speed at an instant,' derive from properties of things new properties for instantaneous entities. The latter would not be the properties corresponding to the predicates in terms of which the change is described, but they would be properties in terms of which the change in question can be analyzed. That is, one might postulate that stages of a thing α all have properties {fi} such that for any property G of α, G(α) ↔ . But while this strategy is open to the hunk view, proponents of the stage view cannot resort to it on pain of definitional circularity. As little as one can define a thing in terms of the sum of its temporal stages, one cannot define the properties of stages in terms of the properties of the entities (things) that are said to be ontologically composed of stages. Whether endurantists and perdurantists can overcome the basic difficulties just adumbrated, remains to be seen. The few critical remarks I have offered in this section are certainly not sufficient to show that either of these two paths is unviable. My aim here was merely to draw attention to the fact that proponents of both views share a common preconception of their task, and that the problematic nature of this task derives from a number of additional premises which have largely have gone unnoticed and unquestioned so far. Let me add to these observations thetically a number of claims that are argued for elsewhere.17 Principles (P2) and (P3) are elements of a more comprehensive network of presuppositions constituting the dominant research paradigm in ontology. At the very heart of this network of implicit assumptions or the ‘substance paradigm’ is the following principle: (P4: The primacy of particulars:) All concrete individuals are particulars.

17

Cf. Seibt 1990, 2005, and 200+c.

144

A so-called ‘particular’ is an entity that (i) has a determinate unique spacetime location and (ii) (has this location necessarily since it) is individuated in terms of its location.18 Traditionally the category of a particular has been complemented with a ‘dual category,’ the category of ‘general’ entities. General entities (being white, being a cat) do not have unique locations but may be multiply located in space-and-time. Thus they are individuated not in terms of a unique location but either by the ‘distinctness simpiciter’ of the ‘suchnesses’ that they are or in terms of functional or causal aspects associated with these suchnesses. Whether general entities are indeed needed in ontology, and how one might conceive of their presence in spacetime, notoriously has been the subject of the longstanding debate about ‘universals’. In contrast, apart from Leibniz’ attempt to defined concrete individuals in terms of universals, the commitment to the ontological primacy of particulars virtually has gone unchallenged. Also participants in the debate about persistence proceed without further ado from the theoretical juncture that persistence is to be accounted for in terms of the transtemporal identity of a particular continuant (endurance view) or else in terms of the particular temporal composites of a transtemporally extended particular (perdurance view). Even though only endurantists are still prepared to invoke explicitly the traditional label ‘substances’, endurantists and perdurantists alike proceed from the core assumption of the substance-ontological tradition: that discrete, bounded, concrete particulars are ‘primary’ in being. In the next section I will briefly sketch an ontological scheme that operates outside of the strictures of the substance paradigm and thus, as any paradigm-transcending proposal, at first sight will appear as ‘unorthodox’ in the sense of ‘unnatural’. But cooks and musicians, or chemists and biologists, should find the scheme quite intuitive, since it ties in closely with the stuff and function geared component of our common sense reasoning, where we individuate entities in terms of what they do, know that two portions combined are another portion and that mixing 18

‘Spacetime’ is used vaguely here, without commitment to relativity theory. An entity has a determinate unique spacetime location iff it holds that for any point or moment in time at which it exists, it is located in precisely one bounded, connected spatial region. (Note that a region that is ‘connected’ in the topological sense might be ‘multiply connected’, i.e., include holes.) Furthermore, that x and y have different locations is sufficient and commonly also necessary for their being different particulars. Some trope ontologists allow for particulars (“abstract particulars”) to be spatially superposed at a time. But location is always a necessary component of the identity conditions of particulars.

145

individuals may generate new functional properties. As I will try to show individuals that abide by the logic of activities or spatiotemporal dynamic stuffs provide us with better ontological explanations than the traditional description of our world as assembly of (three- or fourdimensional) “starren Wirklichkeitsklötzchen” (F. Paulsen). 2. The theory of general processes The ‘theory of general processes’ (hereafter ‘GPT’) is a new ontology in the sense that it introduces a new ontological category and a new formal framework (a non-standard mereology) to describe the basic relationships obtaining between general processes. However, in GPT the general methodological commitments of analytical ontology are retained and thus the approach differs markedly from Whitehead’s “philosophy of organism,” where even basic methodological parameters are modified. Let me preface the following remarks on GPT with a quick reminder of the theoretical character of ontological descriptions, which Strawson’s distinction between “descriptive” and “revisionary metaphysics” has obfuscated in most unhappy ways. With respect to descriptive theoreticity, ontology resembles science. A scientific theory furnishes theoretical descriptions of a domain using theoretical terms (e.g. “harmonic oscillator”, “spin”) that—assuming here a non-instrumentalist interpretation—denote theoretical entities and features. These theoretical entities and features are introduced together with ‘models’ (e.g., ideal springs, rotation), which help us (scientists and non-scientists alike) to understand the theory’s theoretical terms and ultimately account for the explanatory function of the theory’s domain descriptions. Analogously, ontological descriptions use theoretical terms, such as ‘substance,’ ‘property’, ‘trope,’ ‘occasion,’ ‘mode,’ ‘relation,’ ‘state of affairs,’ etc. Such theoretical terms (ontological category terms) label theoretical entities (ontological categories), which we (ontologists and non-ontologists alike) understand in terms of a model. The model for substances are things, for example, and much of the longstanding popularity of substance ontology derives from the central function of things in practical contexts, in particular their transtemporal persistence. The new category label ‘general process’19 thus denotes theoretical entities of a certain sort, and should in the first instance be understood in 19

Alias “dynamic masses” and “free processes” in earlier versions of the theory, cf. Seibt 1990, 1995, 1996, 1999, 2000, 2001, 2003b.

146

terms of (but not be confused with!) their model, namely, “subjectless” activities. A subjectless activity is an occurrence that might involve, but is not, the doing or acting of a person or thing (or of a collection of persons or things). For example, snowing, blowing, thundering, erosion, light or electromagnetic radiation are subjectless activities in this sense—they are presences of dynamic features that have no logical subject.20 Subjectless activities have seven characteristic aspects (which apart from the first one also hold for activities in general).21 (i) Subjectless activities are occurrences in their own right rather than modifications of persons or things—like things, and unlike properties and relations, they are independent in the sense that their occurrence in space and time does not necessarily require the existence of a different sort of entity they occur in or qualify (they may of course be constituted or caused by other entities). (ii) They are temporally extended—there are no instantaneous activities.22 (iii) However, quite unlike things, and much like stuffs (water, wood, etc.) subjectless activities occur in space and time both with indeterminate and with determinate locations (cf. ‘there is lots of rain in Denmark’ vs. ‘on Oct. 12 it rained in Aarhus between 8am and 1pm). Most importantly, subjectless activities do not necessarily occur in a unique spatiotemporal location, i.e., in ontological terms, they are not particulars. While things are located at any time in one place only, subjectless activities are multiply locatable like properties and stuffs—they can, and mostly do, occur in many places at the same time. ‘It is snowing’ can be true of many different scattered regions at the same time. (iv) Subjectless activities also resemble stuffs in that they do not come in ‘natural units’ and are not countable, but 20

Sellars, following C.D. Broad, takes “subjectless” or “pure” activities to be expressed by sentences with “dummy subject”, cf., ‘it is snowing,’ ‘it is lightening’ (Sellars 1981). Even though this might be helpful for illustrational purposes, it cannot serve as a criterion since many activities that cannot be understood as the ‘doings’ of a thing (or a collection of things) are expressed by nouns. 21 To simplify I here characterize subjectless activities directly in terms of ontological features such as being independent, concrete, non-particular etc. Note that the proper methodological procedure would be to show how the logical role of activity sentences dovetails with category features of the entities in terms of which statement about activities are interpreted. For a detailed discussion of activity statements cf. Seibt 2003b and 2004, ch. 2 and 3. 22 To be elaborated presently below: there are not instantaneous activities in the sense of stages constituting temporally extended activities. We do, however, assume that activities exist continuously in time, and thus are dynamic features that can be ascribed to any point in the time period during which they are going on.

147

only measurable in portions or amounts (e.g., an hour of snowing, 1000 lumens of light), which then may be counted. (v) Like stuffs and properties, subjectless activities are not necessarily determinate in all of their qualitative or functional aspects—ontologically speaking, they are determinables. (vi) Subjectless activities are individuated in terms of their functional features, rather than by their location in space and time, just like stuffs are. (vii) Subjectless activities are dynamic but they are not changes—constitutive ‘phases’ of their dynamicity (for example: the change of place of every single flake constituting the dynamicity of the snowing) enter the functionality of the activity not as temporal stages or phases. In contrast to developments, activities do not involve internal temporal difference. These seven aspects of subjectless activities illustrate the main features of the new category or entity type called ‘general process’. General processes are independent, individual, concrete, spatiotemporally extended, non-particular, non-countable, determinable, and dynamic entities, in the sense of these terms just sketched. These category features are well-known from the ontological debate, but their combination is new and, in fact, inconceivable as long as one chooses to remain spellbound by the presuppositions of the substance paradigm. There is one additional, novel category feature of general processes, however, which I call selfpartedness or ‘automerity’ and will introduce separately in the following section. The category features of general processes receive a more precisely formulation, explicitly or implicitly, from within the context of a formal framework, a mereology on general processes (GPM). Thus GPT is the theory of general processes as axiomatically defined within the mereology GPM. For present purposes a quick look at the main observation underlying the construction of GPM will suffice. Ontologists commonly assume that the meaning of the part-whole relation is formally articulated by the axioms of “classical mereology” (however variously implemented). Upon a closer look, however, the relation modeled by systems of classical mereology is the relation ‘is a part of’ which holds for particular (locationindividuated) and countable entities only. (That this clandestine restriction to countable particulars has gone virtually unobserved is yet a powerful documentation of the theoretical bias effected by the substance paradigm.) In contrast, the relation ‘is part of’ can be used with all sorts of arguments, both location-individuated (particular) individuals and functionally individuated (non-particular) individuals. Here are some real life examples

148

taken from the web23: “blogging is part of life,” “Russia is part of the West”, “Music is part of God’s universe,” “All I See Is Part Of Me” (book title), “Learning to negotiate is part of the advocacy process,” “my heritage is part of who I am,” “ Is part of your college education missing?”, “Boston is part of the Islam,” “Loss is part of living,”, “Syria is part of the solution,” “looking immaculate is part of what I do,” “having a good signed contract is part of running a professional business”, “Pain is part of running marathon,” “Hopping too is part of running.” As usages like these convey, the expression ‘is part of‘ states a highly general relation of functional ‘belonging with’ (both as contributory or consequence!), without implications concerning spatial or temporal containment. Most importantly, in many cases this relation of functional belonging with is not transitive. For example, (i) through (iii) in the following do not entail (iv): (i) Changing diapers is part of caring for your baby.— (ii) Opening the box with wipes is part of changing diapers.— (iii) Pressing your thumb upwards is part of opening the box with wipes. — *(iv) Pressing your thumb upwards is part of preserving the baby’s health. Thus this highly general sense of ‘is part of’ is not adequately captured by the transitive relation ‘ 1 then x 1n y ↔ ∃z (x 1 z & z 1n-1 y) We can then take the transitive closure over a sequence of length N of iterated applications of '1', to define a part-relation across partition levels with transitivity restricted to level N: (D1b) x 1N y ↔ x 1 y or ∀ n > 1 in N, ∃z (x 1 n-1 z & z 1n y) For a given entity α, any entity y in the field of the relation ‘α1Ny’ belongs to the so-called N-partition of α. In contrast, any entity y in the field of ‘α1ny’ (the n-th iteration of ‘1’) belongs to the so-called the partition-at-level-n of α . That is, the N-partition of α comprises entities at all partition levels up to and including level m, where m is the last member of N, while the partition-at-level-n of α denotes entities at one level only. The n-parts of an entity α are those entities y for which ‘y 1n α’ holds: (D2a) x is n-part of y =df x 1n y In order to refer to processes at any or all partition levels of a process α below or above a certain ‘threshold’ we can either use existential or universal quantification on the i-parts of α, with i ≤ n, or i≥ n. On the basis of (D2a) we can now state that ‘1’ is irrelative and antisymmetric relative to a partion level: (Ax1) ¬ x 1n x (Ax2) x 1n y → ¬ y 1n x This makes room for the possibility that a whole α can make a functional contribution to one of its own n-parts. Similarly, identity in GPM is defined relative to a partition level with the following two axiom schemas, as identity at a level n (Ax3a) and identity across partition level up to depth n (Ax3b): (Ax3a) ∀z (z 1n x ↔ z 1n y) ↔ x =n y. (Ax3b) ∀z (z 1≤n x ↔ z 1≤n y) ↔ x =≤n y. Two terms are coreferential (in common ontological parlance: ‘two entities are identical’) if they have the same n-parts or ≤n-parts. These two identity principles account for the context-relativity of our judgements of sameness; in some contexts coarse-grained comparisons suffice, while others require an in-depth investigation. In this way GPM can be used to

150

reconstruct the insights on either side of the various debates about, e.g., the ‘extensional’ vs. ‘intensional’ identity conditions of events.24 The new technical devices of parthood at level n and parthood up to level n have many fruitful applications in ontology, both with respect to traditional tasks such as the analysis of persistence and change, as we shall see presently, as well as with respect to new tasks in the ontological interpretation of scientific core concepts. For example, in GPM we can define various types of process interaction (complex processes with more parts or less parts than their parts) and in terms of these many different types of interactive process organizations, such as linear mechanisms, autocatalysis, and self-maintenance.25 GPT, the theory of general processes, uses the formal framework of GPM to explicate the core thesis of GPT that, in Wilfrid Sellars’ happy phrase, the world is “the ongoing tissue of goings-on”26—all there is, is a general process. In line with Aristotle’s intuition that being is linked to individuality, GPT is to show that this intuition can be coherently developed if we—pace Aristotle—separate individuality from particularity. Whatever we talk about in common sense and science is a functionally individuated individual, a dynamics, a “tode ti on the go”, which may be, but does not need to be, uniquely located. An amusing illustration of the ‘general spirit’ of GPT-interpretations can be found in Quine’s playful remark that process ontologists read the sentence “a white cat faces a dog and bristles” as “it's catting whitely, bristlingly, and dogwardly.”27 The main task in GPT is to devise a comprehensive typology of general processes or, less technically speaking, ‘dynamics.’ To this effect the dynamic individuals are classified in terms of parameters of five basic 24

Since (Ax3a, b) are versions of the classical Proper Part Principle, GPM is a socalled ‘extensional mereology’, even though the notion of parthood is ‘intensional’ (i.e., not the classical notion of an extent-part drawn from geometric intuitions). 25 Cf. Seibt 200+b. 26 Sellars 1981, 57. 27 Cf. Quine 1985, 169. I owe this illustration to L.B. Puntel, who occasionally uses it in similar fashion to indicate the flavour of his ontology of primary states of affairs (Verhalte), cf. e.g. 1993. Many years ago, in the early 1980s, Puntel identifed an “object-ontological dogma” in contemporary ontology and shared these research explorations with his students in lectures and seminars. I am deeply indebted to him for these early pointers, providing the core inspiration for a common search of ‘ontology beyond the myth of substance’, even though along different paths of implementation. For the currently most sophisticated and advanced philosophical system in this vein see his 2006.

151

evaluative dimensions, namely, participant structure, dynamic composition, dynamic shape, dynamic context, and their “mereological signature” or automerity pattern, which the following section shall explain.28 3. Selfpartedness and selfcontainment One of the category features of general processes is novel and requires special elucidation. Let us look thus again at (subjectless) activities, the model of general processes. Aristotle refers to common stuffs such as water and flesh frequently as “like-parted (homoeomerous) bodies”: they are “substances that are composed of parts uniform with themselves.”29 As various contemporary authors observed, there is an analogous mereological condition for activities, holding with respect to time—they are ‘monotonous’ or ‘homogenous’ occurrences where beginning, middle, and end of the interval of their duration are ‘of the same nature as the whole’.30 Just as any spoon of a puddle of water is like the whole, namely, an expanse of water, so any minute of an hour of snowing is like the whole, namely, a period of snowing. Thus we can formulate the following general mereological condition: Likepartedness or homoeomerity: An entity of kind K is homoeomerous iff all of its spatial or temporal parts are of kind K. That activities are homoeomerous in time is a claim that contemporary authors frequently connect with Aristotle’s observation that activities such as seeing and knowing are “complete” at any time at which they occur, in contrast to directed developments such as walking-to and build-a-house: 28

To give some hints at least: Participant structure: number and type of causal agents and patients. Dynamic constitution: linear vs. nonlinear additivity of n-parts of dynamics (’emergence’). Dynamic shape: basic types of (parts of) trajectories in phase space, some of which correspond with distinctions familiar from linguistic theories of ”Aktionsarten” and verbal aspects. (telic/atelic, ingressive, egressive, repetitive, conative etc.). Dynamic context: linear vs. non-linear causal consequences, such as the air flow around a kite, vs. non-linear consequences in the generative environment of the process, such as changes in ecosystems that alter selection pressures (cf. Bickhard/Campbell 2000, 343). 29 Cf. History of Animals, 487a2. 30 Cf. Vendler 1957; Gill 1993, 369, 370; Dowty 1977; Verkuyl 1978, 224; Mourelatos 1978, 431.

152

“For one cannot in the same moment both be taking a walk and have taken it, nor be house-building and have house-built…But at the same moment the same thing has seen and is seeing and is thinking, and has thought” (Metaphysics 1048b30-33) Upon a closer look, however, Aristotle’s observation on what I call the ‘temporal distributivity’ of activities expresses an even more remarkable mereological condition than likepartedness.31 Since activities are purely functionally individuated, it does not make sense to distinguish between an activity and its nature. Of course we may say that every minute of an hour of snowing is of the same nature as the whole, but then we are talking not about the activity of snowing but about a (or any) particular spatiotemporal amount of snowing. For the activity itself, the following holds: Selfpartedness or automerity: An entity E is automerous iff for any spatiotemporal region R in which all of E occurs it holds that any smallest region r of R is also a region in which all of E occurs. This condition states in effect that activities are temporally extended yet do not have temporal parts in the sense in which developments have temporal parts, namely, as stages or phases that differ in their nature from the whole or, equivalently, that are different types of occurrences, differing from each other and the whole. Rather, activities (subjectless and subjectbound) occur in all of the temporal parts of the region in which they occur. Of course, on the reading of ‘part’ as defined in classical mereology the very idea of selfpartedness or automerity is incoherent.32 No entity can have itself as a part, since this would be in conflict with the idea that parts are always supplemented by another part to form a whole (“supplementation principle”). On the basis of GPM, however, selfpartedness can be coherently articulated, since in GPM the part-relation (a) has the broad sense of ‘belongs with’ rather than extensional containment, and (b) is not transitive.

31

In my view this passage should not be taken to suggest, though, that Aristotle’s notion of energeia refers to our common sense notion of an activity; for a detailed discussion cf. Seibt 2004, ch. 3. 32 Of course, in mereology ‘x is a part of y’ is sometimes defined as ‘x is a proper part of y or x = y’. But this is not at issue here.

153

For example, the dynamics snowing, call it α, is located in some spatiotemporal region(s), since it is a concrete entity. Thus it is part of what α is that it is located in some region R.33 Any such amount of α in R is a dynamics in its own right, however—it is a specification of α for that particular region (snowing on Mount Everest in region R’ differs from snowing at the Danish coast in region R’’ etc.). Let β be the name for some amount of α in R. Part of what it is to be β is to be α (compare: part of what it is to be my cat Ally is to be a cat). Another part of what it is to be β is to be the sum of regions R1,…Rn, (which are the temporal parts of R). As before, due to automerity of α it holds that being an amount of α is part of being any of the regions R1,…Rn. Call these amounts γ1,… γn. As before for β, for any γi it holds, that being α is part of what they are. Graphically the feature of selfpartedness has the following representation:

33

Vice versa, for any such region R locating α it holds that R’s being an amount of α is part of what R is in interaction with α. (This claim can be connected with both absolutist and relational conceptions of spacetime). Note that an amount of α is, more precisely, a 3-part of α (roughly: part of α’s being located in a certain region, which is an interaction product between α and spacetime), but the contracted claim holds as well.

154

Thus, without getting into conflict with the Supplementation Principle in GPM an entity (snowing, water, light) can contain itself as part of a part, as something that belongs with something that belongs with it. Very roughly, the many determinately located snowings are part of what snowing is, and it is part of what the many determinately located snowings are that snowing is part of what their parts are. Selfcontainment in this sense is, I submit, the sense of temporal continuity at issue in our common sense claims about activities existing continuously throughout the period at which they exist. Temporal continuity is the literal ubiquitous recurrence of one and the same feature, down to the smallest conceivable regions or ‘points in time’.34 Let us return then to the features of likepartedness and selfpartedness as stated above. Even assuming that both can coherently be formulated (the latter for functionally individuated entities), one might argue that claims about the likepartedness or selfpartedness of stuffs and activities are empirically false. Surely not every spatial part of a (spatiotemporal) region of water is again a region of water— regions smaller than size a water molecule would not do—and not every temporal part of a (spatiotemporal) region of running is a again a region of running—periods shorter than the lifting of a foot will not qualify.35 In other words, while the features of likepartedness and selfpartedness as formulated above stipulate complete ‘homogeneity’, homogeneity only exists down to a certain ‘grain-size.’ This objection is a fallacy, equivocating on the terms ‘water’ and ‘running.’ It involves a clandestine switch in perspective: the stuff water is here conceptualized in analogy to fruit salad, as a mixture of things (molecules), and the activity running is considered here as an iterative sequence of developments (lifting the right foot, shifting weight, lifting the left foot etc.).36 But the fact that such equivocations on common nouns and verbs are possible is very significant in its own right, in two respects. First, the objection demonstrates that we switch with facility from a function-geared to an extent-geared perspective, and suggests that we abandon the 34

This notion of ‘concrete continuity’ differs from, yet can be connected to, the notion of mathematical continuity as represented by the real numbers’. 35 Cf. for example Gill 1993. 36 Such reconceptualizations are almost forced upon us once we focus on the identity (within some coordinate system) of the spatiotemporal region occupied by a certain amount of water or running; since the parts of spatiotemporal regions are different, we search for differences in their fillings.

155

traditional strategy of associating ontological interpretations with single linguistic expressions; rather, the ontological interpretation of a noun or verb very much depends on the sentential context, even on the discourse context (compare: ‘Kim ate oats’ (activity) vs. ‘Kim ate an apple’ (development)). Second, the objection points us to a possible way to turn a dichotomy into a distinction within a unified classification system. The conceptual shift from stuffs to mixtures and from activities to sequences of developments is to show that only some, not all parts of the entity are like the entity or contain the entity itself. Interestingly, there are also entities for which it holds that there are no parts like them or containing them, namely, things and events (developments). For example, no spatial part of my computer is a computer, and no temporal part of a baptism or an explosion of a bomb is again a baptism or an explosion of a bomb (and similarly for selfpartedness). Thus in GPT general processes can be sorted into different types according to their ‘mereological signature’, i. e., in terms of their likepartedness (homeomerity) and selfpartedness (automerity) in various dimensions: Maximal, normal, minimal homomerity: An entity α of kind K is maximally / normally / minimally likeparted iff all / some / none of the parts of the spatiotemporal extent of α are of kind K. Maximal, normal, minimal automerity: An entity α is maximally / normally / minimally self-contained iff a spatiotemporal region in which α exists has only / some / no spatial or temporal parts in which α exists. To illustrate, consider the following graphics of normal, maximal, and minimal automerity in space, with ‘(MA)’ denoting the minimal amount of ‘a’:

156

a

a

MP(a) Normal automerity

MP(a) Maximal automerity

a

MP(a) Minimal automerity

For example, the ontological correlates of our talk about activities (‘activity-processes’) are temporally maximally automerous, the ontological correlates of our talk about stuffs (‘stuff processes’) are spatially normally automerous, general processes denoted by talk about amounts of stuffs are spatially normally homomerous but minimally automerous, and temporally maximally automerous, general processes denoted be talk about developments are temporally minimally homo- and automerous, and so forth. All general processes are thus selfparted (and by implication: likeparted), but to different degrees in different dimensions. What matters for our purposes here, however, is the fact that maximal temporal automerity, i.e., selfcontainment in time, provides a new way of making ontological sense of our common sense talk about persistence as temporally continuous existence of the same individual. 4. Persistence as recurrence On the basis of the foregoing definitions and considerations the new account of persistence is already in sight. Briefly put, while the occurrence of a development is a sequence of different dynamic features, the occurrence of an activity is the continuous recurrence of one and the same dynamic feature, namely the activity itself. Activities are maximally selfparted in the temporal dimension. Maximal selfpartedness, however, implies sameness in time in the sense of endurance as commonly defined (‘at any time at which an entity exists, all of it exists’). Precisely this sense

157

of sameness in time is also at issue when we talk about the persistence of things and stuffs. When we speak about stuffs and things persisting we can take ourselves to denote dynamics that are temporally maximally selfparted. In a nutshell, then, the persistence of things and stuffs is indeed transtemporal identity, and the best sense we can make of transtemporal identity is in terms of the recurrence of the same dynamic individual. This is the core idea of the so-called reccurrence account of persistence. Let me add now some of the details working along the desiderata for an account of persistence as we encountered them in our discussion of endurance and perdurance accounts in section 2. A successful account of persistence should: (D1) cover the data for a theory of persistence, i.e., explain the inferential role of statements about persistence and chance; (D2) avoid a commitment to individual essences; (D3) avoid postulating that temporally extended entities of kind K are ontologically constituted by instantaneous entities of kind K; (D4) avoid a commitment to a specific theory of time, i.e., either ‘presentism’ or ‘eternalism’. As I will argue now, the recurrence account of persistence holds out the prospect of combining the virtues of endurance and perdurance approaches, while avoiding their vices. Let us begin with a closer look at (D1), the coverage of inferential data. As pointed out above, there are two main inferential data to observe. First, a statement about the alteration of a thing such as [3] implies a statement about transtemporal sameness such as [4]: [3.] The Statue of Liberty was first copper-colored, but turned green over the course of the years. [4.] The green statue you are seeing now is the same as the coppercolored statue dedicated in 1886. Second, a statement which assigns mutually exclusive features to a thing such as [5]: [5.] In 1886 it was true to say: ‘The Statue of Liberty is coppercolored’, now it is true to say: ‘The Statue of Liberty is green.’ entails a statement about an alteration, such as [3], but also a statement about difference, such as [6]: [6.] The Statue of Liberty as it was then (as it exists in 1886) is not the same statue as the Statue of Liberty as it is now (as it exists in 2006).

158

Focusing first on the first inferential datum, [3] and [4], we can endorse the basic motivation that drives the endurance approach and confirm that statements of transtemporal sameness such as [4] have the inferential import of identity statements: (D1-a) Statements about transtemporal sameness are to be interpreted in terms of identity, not unity. On the other hand, in view of the second inference from [5] to [6], we must not follow endurantists in the assumption that statements of transtemporal sameness and statements about alterations have the same logical subject. One of the profound insights of perdurantism—even though proponents of perdurantism do not make anything of it—is that perdurantism takes a statement about a change such as [3] to be about one temporal part in comparison to another, while a statement about transtemporal sameness, such as [4], is about a temporal whole, i.e., about another entity. Perdurantists de facto reject the substance-ontological presupposition (P2) that what persists in an alteration is the very same entity which is said to have changed. This enables perdurantists to take statements about transtemporal difference like [6] at face value, as statements about numerical difference, while endurantists are forced to read [6] as stating a qualitative difference only. Here we need to side with the perdurantists’ interpretation of the data, i.e., we need to hold: (D1-b) Statements about change entail statements of numerical difference (and statement of qualitative difference). In short, while endurantism does not acknowledge that [6] expresses numerical difference, it fails with respect to (D1-b); perdurantism does not acknowledge that [4] expresses identity and thus falls short with respect to (D1-a). Consider now the GPT- interpretations of [3] through [6], stated here in an ‘intuitive’ version where definitions are replaced by informal explications. [3-GPT]: Sentence [3] above states that a complex dynamics α denoted by ‘the Statue of Liberty’ is part of an interaction dynamics δ1 (involving other dynamics such as light, air, sensory processes) which occurs in time period t1 and yields (among others) as interaction product the dynamics β denoted by ‘x is copper-colored’, and is part of interaction dynamics δ2

159

(involving other dynamics such as light, air, sensory processes) which occurs at time period t2 and yields (among others) as interaction product the dynamics γ denoted by ‘x is green’.37 Essentially, the sentence says of that which is going on in a certain region that it is involved in two different interactions at two different times. From this we may conclude that the ways it was/is going on in these two interactions still count as amounts of it, as in [4]: [4-GPT]: Sentence [4] above states that there are two dynamics x1 and x2, and x1 is part of the interaction dynamics δ1 and x2 is part of the interaction dynamics δ2, and the dynamics that x1 is an amount of is identical to the dynamics that x2 is an amount of, namely, α. [5-GPT]: Sentence [5] above highlights that δ1 and δ2 have different parts (β and γ, respectively) and thus are different dynamic individuals. [6-GPT]: Sentence [6] says that there are two dynamics x1 and x2 that are part of the interaction dynamics δ1 and δ2, respectively, and x1 is an amount of α1 and x2 is an amount of α2, and α1 and α2 are different specifications of α. Without the formal details of the predicates ‘is an amount of’ and ‘is a specification of’ some aspects of these process-ontological interpretations of our common sense discourse are bound to remain vague, but the main ideas, I hope, will stand out nevertheless: (i) Statements like [4] and [5] are comparisons of ‘what is going on’ in two different spatiotemporal regions. (ii) A ‘coarse-grained’ or ‘functionally abstractive’ comparison results in the affirmation of sameness or numerical identity of that which is going on in each case, as in [4]. Such a comparison is ‘mereologically shallow’, taking into account only a few levels in the functional partition of the dynamics involved. (iii) A more ‘fine-grained’ comparison will reach into ‘deeper’ partition levels and reveal difference, as stated in [5]—what is going on in one region functionally resembles what is going on in the other region in that the parts are the same, yet the specifics (what is part of being such a part of what is going on) are different. In short, statements about persistence and change involve different calibrations of the ‘lens of specificity’ that guides our comparisons of functionally individuated individuals: persistence statements express the 37

A dynamics is always located in an extended spatiotemporal region; the expression ‘x occurs at t’ is derived.

160

transtemporal identity of more generic individuals (dynamics), statements about change express the difference between more specific individuals (dynamics). Turning to (D2), then: does the recurrence theory of persistence avoid a commitment to individual essences? The account is committed to partitions of dynamics and thus in a sense to ‘functional’ essences of more or less generic individuals, but not to essences of particular individuals, i.e., entities which are necessarily unique, and only the latter ones are problematic. Moreover, in GPT generic or determinable individuals are concrete. The idea of explaining persistence via the transtemporal identity of a generic entity has been considered occasionally, following an idea of But retaining traditional substance-ontological Whitehead’s.38 presuppositions —and in this one respect we need to reckon even Whitehead among the ‘substance ontologists’—these authors immediately linked genericity to abstractness and were thus forced to explain the persistence of concrete individuals indirectly, in terms of their ‘instantiating’ the same abstract patterns or universal, rather than in terms of transtemporal identity. In contrast, the recurrence account of persistence can take our claims about the persistence of concrete entities at face value as claims about the transtemporal identity of concrete generic individuals. This special conceptual feat, I submit, can only be pulled off if we switch to activities as our models for individuals. Universals are ‘repeatables’ at best in the sense of being repeated by being instantiated—the idea of anything repeating in time or recurring can only be spelled out in terms of the multiple occurrence of a dynamics, i.e., only in sense of a dynamic’s selfpartedness of selfcontainment. Consider (D3): while on the stage view the horses and tornados we claim to persist are constituted by instantaneous entities, on the recurrence view of persistence common sense predications ‘horse’ or ‘tornado’ apply to temporally extended regions and only derivatively, in consequence to the selfpartedness of these dynamics, to any temporal part of these. Finally, as to (D4), all GPT-definitions are neutral with respect to the interpretation of time—to speak of temporally extended entities does not entail a commitment to eternalism, i.e., the claim that future and past exist just like the present. Since the data of a theory of persistence, the inferential roles of statements of persistence and chance, do not imply any specific position as to the ontology of time, (D4) is a significant desideratum—theories of persistence that can cover the data only at the 38

Cf. Carter/Hestevold 1994, Simons 2000.

161

expense of introducing presentist or eternalist postulates are profoundly weaker than a scheme that is compatible with both.39 In conclusion, the recurrence view of persistence can prove Heraclitus both wrong and right, quite in line with Heraclitean metaphilosophy. “You cannot step into the same river twice, since ever and ever new waters are flowing in upon you.” But the many new waters are amounts of the same stuff water, a non-particular individual—the sentence thus is false. Such individuals come best in sight, however, if we dare to think that activities, such as the river’s flowing, is all we are left with.

REFERENCES Bickhard, M. / D. Campbell 2000 „Emergence“, in: Andersen et al. Downward Causation: Mind, Bodies, and Matter, Aarhus: University of Aarhus Press, 322349. Browning, D. 1988 “Sameness Through Change and the Coincidence of Properties”, Philosophy and Phenomenological Research, 49, 103-121. Carter, W./ Hestevold, H. 1994 “On Passage and Persistence,” American Philosophical Quarterly, 31, 269-283. Dowty, D. 1977 “Toward a Semantic Analysis of Verb Aspect and the English ‘Imperfective’ Progressive”, Linguistics and Philosophy, 1.1, 45-78. Gallois, A. 1990 “Occasional Identity”, Philosophical Studies, 58, 203-224. Gill, K. 1993 “On the Metaphysical Distinction Between Processes and Events”, Canadian Journal of Philosophy, 23, 365-84. Haslanger, S. 1989 “Persistence, Change, and Explanation”, Philosophical Studies, 56, 1-28. Hawley, K. 2001 How Things Persist, Oxford: Oxford University Press. Heller, M. 1984 “Temporal Parts of Four-Dimensional Objects”, Philosophical Studies, 46, 323-334. - 2002 Hirsch, E. 1982 The Concept of Identity, Oxford University Press, Oxford. Johnston, M. 1984 Particulars and Persistence, Ph.D.Dissertation Princeton. - 1987 “Is There A Problem About Persistence?”, Proceedings of the Aristotelian Society, Suppl. Vol. 61,107-135. Lewis, D. 1983 Philosophical Papers, Oxford: Oxford University Press. - 1986 On the Plurality of Worlds, Blackwell: Oxford. Lowe, E.J. 1988 “The Problems of Intrinsic Change: A Rejoinder to Lewis”, Analysis, 48, 72-77.

39

Note that in GPT even dynamicity is defined in such a way that it does not carry any commitments to the A-series.

162

Mourelatos, A. 1978 “Events, Processes, and States”, Linguistics and Philosophy, 2, 415-434. - 1988 Mulligan, K. / Simons, P. / Smith, B. 1984 “Truth-makers”, Philosophy and Phenomenological Research, 66, 287-321. Prior, A. 1965 “Time, Existence and Identity”, Proceedings of the Aristotelian Society, 66, 183-192. Puntel, L.B.1993 “The Context Principle, Universals and Primary States of Affairs”, American Philosophical Quarterly, 30, 123-135. - 2006 Struktur und Sein, Tübingen: Mohr-Siebeck, 2006. Quine, W.V.O. 1952 Methods of Logic, London: Routledge & Kegan. - 1985 “Events and Reification”, in: E. LePore / P.M.McLaughlin (eds.), Actions and Events. Perspectives on the Philosophy of D. Davidson, Oxford: Blackwell, 162171. Sellars, W. 1981 “Foundations for a Metaphysics of Pure Process”, The Monist, 64, 390. Seibt, J.1990 Towards Process Ontology: A Critical Study in Substance-Ontological Premises, Ph.D. Dissertation: University of Pittsburgh, UMI-Publication: Michigan. - 1996a „Individuen als Prozesse: Zur prozeß-ontologischen Revision des Substanzparadigmas“, Logos, 5, 303-343. - 1996b “The Myth of Substance and the Fallacy of Misplaced Concreteness”, Acta Analytica, 15, 61-76. - 1996c “Non-countable Individuals: Why ‘One’ and the ‘Same’ Is Not One and the Same”, Southwest Philosophy Review, 12, 225-237. - 1996d “Existence in Time: From Substance to Process”, in J. Faye, U. Scheffler, M. Urs (eds.), Perspectives on Time, Boston Studies in Philosophy of Science, Dordrecht: Kluwer, 143-182. - 2000a “The Dynamic Constitution of Things”, in J. Faye et al. (eds.), Facts and Events, Poznan Studies in Philosophy of Science, 241-279. - 2000b “Ontological Categories: The Explanation of Categorial Inference”, in D. Greimann / C. Peres (eds.), Wahrheit-Sein-Struktur. Auseinandersetzungen mit Metaphysik; Hildesheim u.a.: Olms Verlag, 272-297. - 2003a (ed.) Process Theories: Analysis and Applications of Dynamic Categories, Kluwer: Dordrecht. - 2003b “Free Process Theory: Towards a Typology of Occurrences”, in Seibt 2003a, 23-57. - 2004 General Processes—A Study in Ontological Category Construction, Habilitationsschrift Univ. Konstanz. - 2005c “Der Mythos der Substanz”, in K. Trettin (ed.), Substanz—Neue Überlegungen zu einer klassischen Kategorie des Seienden, Frankfurt: Klostermann, 197-229. - 200+a “Mereology on Processes”, submitted. - 200+b “Forms of Emergence in General Process Theory”, Synthese, forthcoming. - 200+c The Myth of Substance, in preparation.

163

Simons, P. 1987 Parts. A Study in Ontology, Clarendon: Oxford. - 1993 “On Being Spread Out in Time: Temporal Parts and the Problem of Change”, in W. Spohn (ed.), Existence and Explanation, Kluwer: Dordrecht, 134-151. - 1994 “Particulars in Particular Clothing: Three Trope Theories of Substance”, Philosophy and Phenomenological Research, 54, 553-575. - 2000 “How to Exist at a Time When You Have No Temporal Parts”, The Monist, 83, 419-436. Vendler, Z. 1967 Linguistics and Philosophy, Ithaca: Cornell University Press, (ch. 4 revised from “Verbs and Times”, The Philosophical Review, 66, 143-160. Verkuyl, L. 1978 “Thematic Relations and the Semantic Representation of Verbs Expressing Change”, Studies in Language, 2. Wilson, N.L. 1956 “The Indestructibility and Immutibility of Substances”, Philosophical Studies, 7, 46-48.

164

The Thread of Persistence PETER SIMONS, Leeds Argo, their thread of life is spun. Shakespeare, 2 Henry VI

1. Introduction According to the folk conceptions endemic in human language and thought, physical things such as stones and stars, boats and goats all belong to the vast metaphysical phylum called continuants: things which persist through time by enduring, being present as a whole at any time at which they exist, and lacking in phases or temporal parts. In this they contrast with the other great phylum of occurrents, such as eruptions, births and weddings, flowings and growings which persist through time with different temporal parts or phases at different times. As we standardly conceptualize continuants and occurrents, they persist in different ways: continuants by enduring, occurrents by perduring. Modern scientific metaphysics offers two major general accounts of how continuants persist. One is the endurance theory, according to which the folk conception of continuants as endurants is literally correct, to be taken at face value, and enduring is a primitive, sui generis metaphysical concept. The other is the perdurance theory, according to which continuants are in fact occurrents, and have temporal parts, folk conceptions notwithstanding. Perdurance theories come in several different flavours, but they all basically deny conceptual appearances and insist that a proper metaphysical account of continuants requires a conceptual revision to accord stones and other physical things temporal parts. I think that both the endurance theory and the perdurance theory are wrong, and that the metaphysical truth is to be found in a third theory which I call the invariance theory, which is distinct from either of the others while preserving the attractive features of each.1

1

For previous attempts to enunciate this theory see my three papers of 2000 listed among the references.

2. The Endurance Theory and Why it is Wrong The endurance theory was first clearly formulated by Aristotle, and it dominated Western metaphysics until the twentieth century. The central concept of the endurance theory is that of a substance, which is a continuant able to exist relatively independently, and to survive many kinds of change, such as changes in properties, size, position, and relations to other things. Modern Aristotelians of all stamps consider the concept of substance to retain its metaphysical relevance and to cover a large portion of the things we encounter in the world, including, notably, ourselves. A typical physical substance comes into being by some change or reconfiguration of some pre-existing matter, which comes to constitute the substance by this change. The substance then exists for a period of time, during which it may change in various ways: in its position, size, qualitative and quantitative properties, its relations to other things, and the matter of which it is made. Depending on the kind of substance in question, it may survive or endure through such changes more or less unscathed for a greater or lesser period of time. Changes of sufficient violence or saliency may however terminate the existence of the substance, so it ceases to exist. In the case of a living organism we call this its dying. Typically, though perhaps not invariantly, substances, once in existence, exist uninterruptedly for a period, but when they cease to exist they do so permanently. The longevity of substances varies hugely: some exotic particles exist for a minute fraction of a second, while other particles and things such as stars survive for billions of years. In between are many of the most familiar to us, including ourselves. What makes substances continuants is that they do not persist by accretion of phases or temporal parts. A tree may start as a seedling and grow until it becomes a forest giant, but each of the parts it gains (and loses) during its growth, and indeed during its whole life, is itself another smaller continuant, sometimes of greater, sometimes of lesser longevity than the tree itself. At each time at which the tree exists, from its inception as a fertilized seed to its final death, the tree itself exists then, even though it fluctuates in the parts it has, and these vary over time in their number, kind and organization. All of this is wholly familiar from the world around us. The only sense it makes to say that not all of a tree is present at a certain time at which it exists is if it either loses a part, whether naturally, by accident or by disease, as when a high wind tears off a bough, or when the tree is somehow deformed or defective in lacking a part normally found among

166

members of its kind. But a tree does not have earlier, middle and later parts in the way that a play or a football game has. This conception of physical continuants as substances is thoroughly wrought into our everyday language and thinking, to such an extent that it takes some effort to even consider that it might not be ultimate metaphysical truth. It has common sense on its side. For Peter Strawson it is one of the cornerstones of the unvarying and transcendentally necessary conceptual scheme he calls descriptive metaphysics.2 Endurantists, whether of Aristotelian or Kantian persuasion, consider that appearances tell the truth here. But there are several nagging difficulties attendant on the endurantist picture, and I shall summarize those I consider the most important. Given the cut and thrust of philosophical debate, no single objection is alone sufficient to scupper the theory, but taken together they motivate looking for an alternative. They are: • the change problem • the passivity problem • the involvement problem • the relativity problem • the truthmaker problem The change problem was notably highlighted by David Lewis,3 but in many respects goes back to worries about substance raised by Hume, and may even have forbears among the presocratics. When a continuant changes intrinsically, i.e., as it is in itself, what happens? The endurantists say: first it has one property, then it has another property inconsistent with the first. But how can it have inconsistent properties and yet exist? One option, the most natural one, is to say that the properties are not monadic but dyadic, relative to times as well as bearers. But that makes the properties into relations and the change into a relational one. Another option is to deny that the thing has both properties, by insisting that only the present is real. This is presentism: according to it, when the thing has one property, the supposedly inconsistent thing with the other property doesn’t exist. The third option is to accord the thing temporal parts and ascribe the contrary properties to distinct temporal parts of the now temporally extended thing. It is this third way that is taken by perdurantists. Endurantists typically embrace one of the alternatives and deny that doing so is disadvantageous,

2 3

Strawson 1959, 9. Lewis 1986, 203 ff.

167

or they look to build up a fourth position without the disadvantages of the other three.4 The passivity problem concerns the idea that enduring things simply passively survive from one moment to the next. Since endurance is a sui generis notion, it is not to be explained by anything more basic, so a continuant’s survival is simply a brute fact with no further explanation. When a continuant ceases to exist, it has to be because some cause or agency intervenes to prevent its survival. But in many cases, such as those of organisms, survival appears to be a much more dynamic and delicate affair. In a generally dynamic universe the idea of something continuing to exist by a kind of brute metaphysical inertia is anomalous. Continuants undergo changes, and such changes are events and processes, in a word, occurrents, in the life of the continuant. What is the relationship between a continuant and the events and changes in its life? The endurantist can give it a name, either ‘participation’ or ‘involvement’, but cannot say anything enlightening about it unless occurrents are downgraded to a metaphysically secondary or non-existent status, say as the instantiation of properties by substances. But not only is this conception suspect in several ways, it is no longer plausible that occurrents are metaphysically second-class citizens, since they play such a fundamental role in science. While no fundamental metaphysical theory can avoid postulating primitive relations, and the relationship between a continuant and the events in its life has the benefit of familiarity, it would be preferable for a theory not to have to postulate involvement as a primitive. The relativity problem affects the emphasis placed on the lack of temporal parts for continuants. According to the special theory of relativity there is no absolute distinction between spatial separation and temporal separation, and hence no absolute distinction between spatial parts and temporal parts. The spatial parts of a continuant at a time at which it exists are supposed to be its parts then, all at the same time. However the concept of same time has no global significance. Hence what counts as the spatial extent and content of a continuant which is extended over a wide spatial volume, a galaxy for example, is relative to a frame of reference. This suggests that the opposition between temporal and non-temporal parts is not hard and fast, as the endurantists would have us believe. The truthmaker problem concerns the question what makes it true that a certain continuant exists at a certain time. Even if not all true propositions have or require truthmakers, existential propositions appear to be just the 4

I attempted such a fourth way in Simons 1991; cf. also Haslanger 1989.

168

sort of proposition that require truthmakers to be true.5 A truthmaker for a proposition is an entity which, in virtue of existing, makes the proposition true. The most obvious and uncontroversial cases would therefore appear to be singular existential propositions. An object a itself is the truthmaker for the proposition that it, a, exists. Consider however not an absolute existential proposition but one stating the existence of a continuant at a time, for example the proposition that John Lennon existed in 1980. The natural truthmaker for this would appear to be John Lennon, but there is a catch. It is normally held that the truthmaker necessitates the truth of the proposition, so that the proposition that the truthmaker exists logically entails the proposition made true. But the proposition that John Lennon exists (absolutely) does not entail logically that he existed in 1980, for John might have lived a shorter life than he did, and have died before 1980, yet he would still have existed and have been the same person. In the case of occurrents, we can find an idiosyncratic object making it true that the occurrent exists (goes on, continues, takes place) at a time, namely any temporal part of the occurrent that covers the time in question, and any such part exists at that time essentially if the occurrent exists at all. Such modal strictness is not forthcoming for continuants, whose careers are modally much more flexible and contingency-prone than those of occurrents, which arguably have their parts and times of existence essentially. So the question as to what made it true that John Lennon existed in 1980 remains unanswered by the endurantist natural candidate. Most of these problems have been debated back and forth at considerable length in the literature on persistence, three- and four-dimensionalism, and there is no consensus. Together the complaints constitute a formidable if not obviously insuperable obstacle to the acceptance of endurantism, such that if we can find a theory of persistence which avoids all or some of them, and which does not bring its own worse problems, it is to be preferred. 3. The Perdurance Theory and Why it is Wrong Although the perdurance theory has gradually and quite reasonably imposed itself on the metaphysical community as a serious rival to the endurance theory, and may indeed be now the dominant position among metaphysicians, it is not without its own problems. Among the most important are the following: 5

Simons 1998, § 3.

169

• the language problem • the paucity of entities problem • the modal rigidity problem The first two problems turn on the fact that the perdurance solution flies in the face of our standard, entrenched conceptualization of continuants and what concerns them. The language problem is that there is no developed language within which to talk about continuants construed as entities with a temporal dimension. This is to be expected since the dualist conception of a world containing both continuants and occurrents is precisely the standard view. However, the metaphysical proponents of the revisionary perdurantist picture, while making their case for its correctness, have never followed through and actually constructed such a language, even in outline, let alone in the sort of detail that would be required to convince sceptics that our attachment to the endurantist paradigm is anything other than habituation or the product of a cultural-historical accident. While this is more a practical problem than a theoretical checkmate, it adds considerably to the balance of considerations inclining one against the perdurantist picture. The paucity of entities problem consists in the fact that for the endurantist there exist both the continuant and the multifarious events and processes that befall it, while for the perdurantist there is only one entity. For the endurantist, when Rebecca skis down a black run, there is both Rebecca who is doing the skiing and the event of her skiing down it, each of which of course has innumerable parts. For the perdurantist there is only the Rebecca-skiing-down-the-black-run event, which is a part of the whole Rebecca-event and a part of the scattered sum of all skiings. Perdurantists do not and cannot make any distinction between a continuant and its life, whereas endurantists can and do. From a folk point of view, the perdurantist comes up short of entities. This means that whereas the endurantist has a suite of predicates predicable of continuants and another suite predicable of occurrents, the perdurantist has only the second. Clearly this impacts the first problem, that of language. The modal rigidity problem is the most theoretical of the three problems facing perdurantism. Consider a continuant such as Rebecca. Like other continuants, it is highly contingent what befalls her. Simply skiing down the run is laced with contingency, as she chooses to turn here rather than there, now rather than half a second later, to take the left rather than right fork, to go left round a mogul instead of right or over it. She could have done the run in a host of different ways, or not done it at all then, or not been there then, and still be the same Rebecca. The event of her skiing

170

down the run however is not so laced with contingency. While there might have been many different events of Rebecca skiing down this slope, dependent on her choices and on other contingent circumstances, the actual event which takes place has its time, its parts and its location essentially. Had any of these been different, it would not have been this event but another that took place. Perhaps it would not have been a wholly different event: two possible events of her skiing down this slope may start out alike and only diverge after a certain stage. Their initial segments would be the same, but their later segments would diverge. This points to a potential advantage of that version of four-dimensionalism known as stage theory which stresses the ontological priority of the momentary temporal stages of an occurrent over the whole “worm”, as in worm theory.6 What is to follow a given stage appears to be contingent: it could be the same stage and yet be followed by different stages. As wholes then, different events of skiing down the slope are distinct, but modally they share the initial part. But either way, if overall we cannot distinguish between Rebecca’s skiing event and the temporally extended part of Rebecca that perdurantists acknowledge, then we cannot separate the contingency affecting Rebecca from the lack of contingency affecting the whole event. Either we have to go against appearances on one of these, or the story has to be fully recast, for example in terms of different counterpart relations, or by another ruse of the many that modal metaphysicians have developed. I am not saying that the problem cannot be addressed, but it will be by epicycles. If a theory can be found which does at least as much and which can avoid complex and implausible epicycles, then it is to be preferred. 4. Abstraction and Invariants To prepare the way for the invariance theory, let us consider the concept of abstraction. In the metaphysics of universals and, relatedly, in the philosophy of mathematics, abstraction plays a key role in many theories. It may do so irrespective of whether one is a realist or an anti-realist about universals or mathematical objects. To take a simple example, consider the universal, having a mass of 1 kg. According to realists, this exists independently of mind and language and is multiply present, instantiated in all material objects having just this mass. According to anti-realists, there is no independently existing universal, but there are many individual objects alike in having a mass of 1 kg, and by abstracting from their differences we 6

See Sider 2001 for the distinction.

171

arrive at the general concept of having a mass of 1 kg, under which they all fall. This concept, or its associated general term, may be considered the universal by conceptualist or nominalist anti-realists. For a realist the concept is not the universal, nor does the procedure of abstraction create a dependent universal, but abstraction is the key step in the cognition of an independently existing universal. For present purposes I wish to leave aside the issue of realism/anti-realism and be neutral. How in general does abstraction work? The following account is not neutral with respect to the different accounts of abstraction that are available, but it represents what I consider the best and metaphysically least committing viewpoint. Consider a collection of objects over which an equivalence relation is defined. For massive objects for example, the relation might be has the same mass as, which, in order not to prejudice the issue of whether there are such things as masses, we may reformulate as is equimassive to. Any massive object is equimassive to some other objects, so taking these all together we get an equivalence class of all the objects equimassive to the given object. But this equivalence class is not the mass in question, since the mass necessarily is the mass it is, whereas all the objects of one mass might have all had a different mass. Also the class is a collection of concrete things and the mass is an abstract universal. A more direct approach to the abstract was put forward by Frege.7 In considering a proposition to the effect that two objects stand in an equivalence relation, the same proposition may be conceptually reparsed (“recarved”) as an identity between two abstract objects. So we may trade in the form of expression a is equimassive to b for the form of expression the mass of a = the mass of b and while both sentences express the same sense, the second ostensibly makes reference to an abstract object, the mass of a, while the first does not. If E is any equivalence relation over some objects, we may express the scheme of such a recarving as a biconditional aEb §E(a) = §E(b) where the double arrow indicates not just logical equivalence but synonymy between the two sides. The new expression ‘§E’ is functional in form and in conjunction with suitable arguments serves to signify a suite of abstract objects internally related to the original equivalence. By ringing the changes on the equivalence relation we may in this way introduce a wide variety of abstract entities, for example 7

Frege 1884, § 64 f.

172

EQUIVALENCE

ENTITY TYPE

… is parallel to … … is equinumerous to … … is geometrically similar to … … is equimassive to … … is equihued to … … exerts the same force per unit area as … … travels as far per unit time as … … is as far away from … as … is from … … is as far round from … as … is from … … is in the same ratio to … as … is to …

directions numbers shapes masses colours pressures speeds distances angles proportions

As the examples show, the notion of an equivalence may be generalized from two to four places, and also to larger even numbers of places. Frege himself rejected recarving equivalences as a way of introducing abstract objects, because they do not suffice unambiguously to identify the object, which remains indeterminate with regard to its identity or difference in respect of objects not given via the equivalence.8 However, this shortcoming is not now universally regarded as a defect so much as a diagnostic feature of abstractness.9 Simply introducing abstracta of whatever kind does not yet constitute science: for that we need the abstracta to have properties and stand in relations. The properties and relations of abstracta are derived from or at least internally related to those of their associated concreta. For example, everything of mass 2 kg is more massive than everything of mass 1kg, so the two abstract masses 2 kg and 1 kg stand in an internal relation of greater to lesser. We say a property P is invariant with respect to an equivalence E if 8

Frege 1884, § 65. The indeterminacy is nowadays known as the ‘Caesar Problem’, since Frege claims the recarving equivalences cannot answer the question whether the number 9 is Julius Caesar. Cf. § 56. 9 The so-called neo-Fregeans accept certain recarving principles as giving us cognitive access to such abstracta as natural and real numbers despite the Caesar Problem. Cf. Hale and Wright 2001.

173

whenever something with P stands in E to something else, the second thing also has P: inv(E)(P) ↔ Def ∀xy(Px ∧ xEy → Py) The property P being thus invariant, we may ascribe a derived property *P to the associated abstracta: inv(E)(P) → ∀x(Px ↔ *P(§E(x)) The derived property *P cannot except on pain of absurdity be the same as the original property P.10 For example an object may be longer than another but its length is not longer than the other length, rather its length exceeds the other. The modificatory contributions * and § effectively cancel each other out. Once again, as the last example intimates, the notion of invariance of a property under an equivalence may be extended to the invariance of a relation of any arity under an equivalence, by the obvious definition. There are several different notions which are more or less confusedly signified by the term ‘abstract’. One is that of a thing which is not in space and time, and which lacks any causal role. This is what most people think of when talking of ‘abstract objects’. Another is that an object is abstract when it is necessarily or typically cognized, introduced or constructed by a process of abstraction, conforming to the schema outlined above. In order not to prejudice later discussion, I shall use the term ‘abstract’ for the former only, and call objects introduced by abstraction ‘invariants’ of their respective equivalences. Thus speeds, directions and so on are invariants. In general the kinds of invariants most readily thought of are also abstract. For example directions, angles, numbers and proportions clearly are not located in space and time and have no causal role. Sometimes the invariants are only a single abstractive step away from location and causality, so locational and causal terminology is used with an adjusted sense. For example the Grey Wolf, Canis lupus, can be ascribed a location in so far as its exemplars are distributed at any time over certain terrains, and it may be ascribed causal powers such as predation in so far as its exemplars hunt. Clearly however the sense in which the Wolf hunts is generic, while the sense in which any individual wolf hunts is individual.

10

I owe this insight to Künne 1983. Cf. Simons 1991.

174

5. The Invariance Theory of Continuants Let us assume that perdurantists are right to the extent that fourdimensionally extended occurrents are metaphysically more basic than continuants. This does not mean that we have to deny the existence of continuants as they do. Rather we can use the occurrents as a basis, showing both what continuants are and how their natures relate to their founding occurrents. The idea is that occurrents stand in certain suitable equivalence relations, and that the invariants under these equivalences are the continuants. Let us for a moment suppose that we are given all the occurrents in the life of a continuant, and that we are also granted a frame with respect to which we may dissect these into their temporal parts. Given any temporal interval which does not extend outside the lifetime of the continuant, there is a temporal segment of the life of the individual continuant which comprises all the occurrents and occurrent-segments of the continuant’s life going on at or within that interval, and nothing besides. Call these total lifesegments of the continuant in question. Now consider that each of these total life-segments stands in an obvious equivalence relation to all the others, and to nothing else besides, namely, being a total life-segment of the same thing as …. The invariant under this relation has a number of pretty obvious derived properties and relations. For a start, it is one–one correlated with its continuant. Secondly, it exists at all and only times at which some total life-segment exists, and therefore its times of existence coincide with those of the continuant. Thirdly, its spatial location at any time can be taken as that of its total life-segment at that time. Finally, it has no temporal extent or temporal part, for such a temporal extent would have to be invariant among the temporal extents of all the total life-segments of the continuant. Assuming the continuant does not exist for a single extensionless instant, which we can rule out by fiat, this means that no such invariant under the relevant equivalence has any temporal extent or any temporal parts. It is spread out in space but, unlike its non-instantaneous life-segments, including its total life, not spread out in time. As I have told this story, it would be blatantly circular to specify that the continuant is the invariant under this equivalence, since the equivalence was defined precisely using the idea of a total life-segment of the continuant. The circularity would disappear however if we could discover an equivalence relation which coincided extensionally with this one but which

175

did not specify and was not defined via the continuant. Suppose we could come up with just such an equivalence, the existence and specification of which was not dependent on a prior specification or identification of the continuant. Then we could have an invariant for this equivalence which would in all its inherited characteristics coincide with that given before, and which would then allow us to identify the continuant as the appropriate invariant. It is my contention, the contention of the invariance theory, that this is precisely the case, and that continuants are invariants over certain occurrents under certain equivalences. The question is what equivalence relations can plausibly be said to perform this function, being such that they can be specified independently of a prior specification of a constituted continuant, yet which yield invariants with the familiar and intuitively correct properties and relations. It may be that there is more than one such equivalence: indeed, given the wide variety of kinds of continuant that we recognize, it would be surprising if there were not. But such equivalences will have to deliver the right kinds of continuant as invariants, and that constrains them quite considerably. It seems to me that in all the cases that I can think of, the equivalence in question involves certain recurrent features. Firstly, it has to have the requisite formal properties of an equivalence relation. At least, it has to in uncontroversial and unproblematic cases. There are certain well-known aporiai concerning the continued identity of continuants through time, such as fission cases, problematic personal identity cases, intermittent existence cases, matter and constitution cases, indiscernibility cases, and Ship of Theseus type cases, where any constituting equivalence ought by rights to turn out to be as problematic as the continuants thereby constituted. Interestingly however, when the literature about such cases is examined, the problems almost all turn on the fact that the standard features required of identity, viz. equivalence with indiscernibility, are not obviously guaranteed by the fundamental description of the case. This suggests that what may be problematic about all the problems, or a large portion of them, is that we are precisely not guaranteed a smooth underlying equivalence on which to base an identity. To take one example, total amnesia is considered a problem for memory-continuity accounts of personal identity. The amnesiac block from later times to earlier precisely inhibits the symmetrical memory-accessibility relation from crossing earlier to later memories. The problem of partial forgetting of earlier events in a person’s life threatens the transitivity of their identity because retrospective remembering is not transitive. So by reflecting on problematic features of constituting equivalences

176

we should be able to predict when identity across times becomes problematic. A second feature that the required equivalences seem to share is a strong causal component. Not all cases of causation give rise to the right kind of equivalence however. For example, an explosion is a rapid and energetic expansion of material with associated high temperatures and an outward travelling shock wave. As the explosion proceeds, energy is naturally dissipated until effects of the explosion which continue the causal succession are mild enough no longer to constitute parts of the explosion itself. The causal process of an explosion is, like all causality, monodirectional, but it does not normally have the kind of stability that would be required to constitute a continuant. Similarly, a collision between two objects dissipates their combined kinetic energy in a typically chaotic and uncontrolled way. For a causal process to be of such a kind as to give rise to an equivalence it will typically be more or less sustained and its successive phases similar to one another in the right sort of way. A process of propagation of a constrained wavefront in an electromagnetic field with relevant isolation of the process from others may be said to constitute the motion of a constituted invariant, a photon, which is a continuant. Likewise a suitably stable resonance in a quantum field may constitute a fundamental particle. The more complex and repetitive or ongoing process of gluon exchange constitutes the nuclei of chemical elements, while the process of photon exchange constitutes the molecules of chemical compounds. The highly complex and energetically expensive continuation of numerous chemical process types with maintenance of process parameters constituting a hierarchy of stable structures may make up the life of an organism. In all these cases we see causal continuity alongside a certain stability, which combines the often small and short-lived processes into a more extended one which sustains an equivalence relation among the temporal segments of the whole. To give such an equivalence a name, we may call it genidentity. Having a name solves no problems, except that of not having a name. With the name however we can concisely state the invariance theory: continuants are the invariants of genidentity relations: Ctt(a) = Ctt(b) a genid b The thread of persistence which maintains a continuant in existence is that of genidentity among the temporal segments of the compounds of its constituent underlying occurrents.11 11

The metaphor of life as a thread is very old. Ancient Indo-European myth tells of the three Norns or Fates who decide an individual’s span of life: one spins the thread of

177

6. Properties and Advantages of Continuants as Invariants Continuants do not have temporal extent or temporal parts, since particular temporal extent or parts are not invariant under genidentity. Continuants have a spatial location and extent which at any one time is the sum of that of their then constituent occurrents. Continuants exist when and as long as their constituent occurrents have the kind of character required to maintain the genidentity. What this character is varies from kind to kind. Whether a certain kind of continuant exists depends on what sorts of suitable causal equivalence relations obtain among occurrents and their temporal parts. Bearing in mind that occurrents may be parts of other occurrents, and that genidentity may obtain at different levels of scale, it follows that continuants may have other continuants as parts, whether permanently or temporarily. The relationship of involvement is no longer a problem: it is that intimate but explicable type of relation between invariants and what they are invariant over. This situates involvement as one of a broader kind and so constitutes an explanatory advance over the view that involvement is sui generis. A universal is an invariant under a relation of exact resemblance. Wherever and whenever it is instantiated, the universal itself is present, “as a whole and not as a part”, but not as an individual is present, since the same universal itself is present at multiple locations, but in its own way. Likewise, a continuant is present whenever it exists, “as a whole and not as a part”, because one of the sustaining or vital phases of its life is then going on. It is this phase of its life which indirectly makes it true that the continuant exists at that time. By the invariance theory we are able at one stroke to provide a plausible solution to the truthmaker problem and to make sense of the otherwise mysterious and much-ridiculed idea that the continuant is “present as a whole” at every time at which it exists. All we need to do is to recognize that the sense of presence differs as between an occurrent and a continuant, as indeed between an individual and a universal, by dint of the sense-adjustment in the notion of presence incurred by abstractlife, the second measures it, the third cuts it. The spinner is Urd (Clotho), who provides the materials and symbolizes the past; the measurer is Skuld (Lachesis), who determines the span and symbolizes the future; the cutter is Verdandi (Atropos), who symbolizes the present. Together the three sisters weave the tapestry of destiny (Wyrd) from the threads of individual lives. Just as a long thread may be made of many shorter filaments twisted together in the right way, so a long life may be made of many shorter-lived events composed and causally connected in the right way.

178

ing under the relevant equivalence. The change in sense is different in each case because the equivalence is different. Both exact similarity and genidentity abstract away from temporal location, but exact resemblance abstracts away from spatial location as well, whereas genidentity does not. Thus a continuant genuinely occupies space, but not in the way that an occurrent does. Occurrents are typically terms in causal relationships: one thing happening causes other things to happen. A continuant constituted by a complex of occurrents thus acts through these: by its constituent events and processes it impinges causally on its surroundings. A continuant is thus both spatially located and causally active, and hence, while an invariant, it is not abstract in the way a number is abstract. Not all continuants are substances in the classical sense. An occurrent can sustain a continuant which depends on other things in a way unacceptable for a substance. The best examples of this are disturbances in a physical medium, such as a wave or a vortex. When a stone is thrown into a still pond, the event of the stone impacting the water causes a series of ripples to propagate outwards from the site of impact. Each ripple is a distinguishable individual continuant expanding from the centre. The processes which constitute the ripple are large collections of circular motions by small parts of the water which transfer energy and motion to neighbouring parts. Each such circular motion is small and short-lived, but together they constitute an impressively coordinated suite of motions which transport energy away from the centre and cause the surface of the water to rise up in the characteristic way of waves. Thus many short-lived events can compose at once and over a longer period a sustained process which has the requisite features of stability and causal succession to generate a genidentity relation. The invariant under the genidentity is the water wave, yet the wave is no substance, being dependent on its medium. Another and more dramatic example is a storm, such as a hurricane in the Atlantic or a vortex in the atmosphere of Jupiter, which consists in a huge number of motions by gas particles fed by an exogenous energy source. The storm is a disturbance, and so a continuant, but the myriad constituent events are generally of much shorter duration than the life of the storm. Once the disturbance has passed a particular location, the constituent events are elsewhere, but events do not move, so it is later and new events which then sustain the storm. The processes which constitute waves or storms are motions among the particles, but the particles themselves are continuants constituted by other and more basic processes as investigated by chemistry and physics. These processes, while go-

179

ing on within the region of the storm, are not themselves parts of it. The complexities which naturally arise are thus mirrored in the way the invariance theory approaches its work, whereas perdurantists typically cut straight to the massive event and ignore the hierarchy of processes of different types. The invariance theory is thus able to distinguish a continuant from its life, and it is also able to explain the divergences in modal rigidity of continuants and occurrents. The contingency that affects continuants turns on the fact that the conditions for genidentical continuation of a continuant’s life are modally quite flexible. Of two events, one consisting of skiing down the left of a slope at a certain time, and the other consisting of skiing down the right of the slope at that same time, at most an initial part can be the same. As wholes, each one is necessarily as it is and has the parts it has. On the other hand, the skier could equally well be involved in the one or the other without prejudice to her identity, because all that is required for her life to continue is that suitable life-segments follow upon those to date, “suitable” here meaning “such as to constitute the prolonging of her life”. The change problem is handled in a similar way to the way perdurantists handle it: when a continuant changes, what is going on at the metaphysically deeper level is that one segment of its life is succeeded by a segment with contrary properties. The invariance theory also explains precisely why many predications about continuants have to be tensed, or relativized to time. When a property or relation of the continuant’s underlying processes is not invariant under genidentity, but applies to only a part of the life, the corresponding predication concerning the continuant cannot be left unqualified as to time. Just as a sweater may be both red and green at one time, provided it is red at one place and green at another, or the Grey Wolf may be both red and white, provided it is red in Arizona and white in the Arctic, so a leaf may be both red and green at the same place, provided it is red at one time and green at another. Because genidentity abstracts away from time, any temporary states of the continuant can only be captured predicatively by thus relativizing to the time. Relativizing is a general possibility with abstraction, allowing non-invariant properties of concreta to relate to their abstracta, but what a predication is relativized to depends on the equivalence relation. The difference between the perdurantist and the invariantist solution to the change problem is that invariantists recognize genuine change among genuine continuants, but unlike endurantists, they can say more about it, and they recognize that the fundamental proc-

180

ess in change is one of replacement of one thing by another thing with contrary properties, the “things” being parts or aspects of parts of the underlying processes making up the continuant’s life. The relativity problem is not a problem for the invariantist, because she does not need to insist on a privileged division between space and time. Like the perdurantist, she can afford to be relaxed about simultaneity. However as an additional bonus, the invariantist can explain the following apparent puzzle. If a continuant is an invariant under genidentity, and genidentity is a relation among temporal segments of a four-dimensional occurrent, why is there not a different continuant for each way to carve up spacetime into space and time? Why does the space-traveller streaking past the Sun at nearly the speed of light see the very same star as the Earthling orbiting it at a snail’s pace? Surely their different reference frames mean their respective genidentity relations over the Sun’s constitutive stellar processes pick out different segments, which only overlap obliquely with those cut at another angle. Surely then there must be many Suns, not just one, despite the unity of the underlying processes. That this is not so is already prepared for in the invariantist’s account of genidentity. Recall that genidentity is in large part a causal notion. What makes one thing later than another is that there is or may be a causal link from the one to the other but not vice versa. Time turns on causality, not the other way round. Causal influences cannot propagate faster than the speed of light, so in fact the operative distinction for genidentity, as for the theory of relativity, is between those occurrents or parts of an occurrent between which there is or can be a causal connection, which are then in timelike separation, with one absolutely earlier and the other absolutely later, and those for which there is not and cannot be a causal connection, which are absolutely elsewhere. Since this is a single distinction, there is a single continuant generated by the one genidentity relation: only one Sun after all. Of course which events in the life of the Sun are considered earlier than which other ones and which ones as simultaneous will depend on the frame, so likewise temporally relative predication regarding continuants is framely relative, not absolutely relative. The proper way to relativize is to sufficiently small spatiotemporal regions or parts. As was mentioned earlier, abstraction is a tool of choice for nominalists as well as for realists in the battle of universals. So someone of a nominalist persuasion may well feel that the invariance theory of continuants should by rights afford continuants a similar shadowy or secondary status to that of universals. They would be seen as cultural or linguistic constructs

181

rather than as something existing independently and awaiting discovery. This might then be held against the invariance theory, because to most of us, the paradigmatically real and concrete objects are continuants such as rocks, mountains and planets. For it to undercut their reality would count heavily against the invariance theory. I think this is a delicately balanced issue. The invariance theory accords metaphysical priority to occurrents over continuants. This priority appears to be in broad agreement with the thrust of modern physical science, which is also behind the philosophical popularity of perdurantism. But, unlike the perdurance theory, it affirms the existence of continuants, and despite the suggestion that the use of abstraction renders continuants dependent on mind and language, the theory does not of itself deny their reality or independence. It could be held, in conformity with realism about universals or mathematical objects, that abstraction does not serve to construct but to disclose continuants. Or on the contrary it might be held, in conformity with nominalism, that the role of abstraction in invariantism shows continuants to be constructs rather than part of the fundamental furniture of the universe. In support of this we can point at the training and maturity required to attain the cognitive capacity to recognize individual continuants as the same in their successive appearances. In acquiring and exercising this capacity we are in a way retracing the mature abstractionist route. However I do not think this necessarily tells against the reality of continuants, albeit that they are generically existentially dependent on their underlying constituent occurrents. Most of the nominalist’s spleen against abstract objects concerns their causal inertness. As we have seen, in the case of material continuants, whether substances or disturbances, or indeed continuant tropes, a causal role is guaranteed by their intimate involvement with occurrents. That there may be many ways in which different kinds of continuant, from quarks to galaxies, from hurricanes to humans, are causally constituted, causally sustained and causally active, simply reflects the heterogeneity of the phylum of continuants.

REFERENCES Frege, G. 1884 Die Grundlagen der Arithmetik, Breslau: Koebner. Hale, B. and Wright, C. 2001 The Reason's Proper Study. Essays towards a NeoFregean Philosophy of Mathematics, Oxford: Oxford University Press. Haslanger, S. 1989 Persistence, Change, and Explanation, Philosophical Studies, 56, 1-28.

182

Künne, W. 1983 Abstrakte Gegenstände: Semantik und Ontologie, Frankfurt am Main: Suhrkamp. Lewis, D.K. 1986 On the Plurality of Worlds, Oxford: Blackwell. - 1989b “Endurance and Temporary Intrinsics”, Analysis, 49, 119-125. Sider, T. 2001 Four-Dimensionalism, Oxford: Oxford University Press. Simons, P.M. 1990 “What is Abstraction and What is it Good for?”, in: A. Irvine (ed.), Physicalism in Mathematics, Dordrecht: Kluwer, 17–40. - 1991 “On Being Spread Out in Time: Temporal Parts and the Problem of Change”, in W. Spohn, B.C. van Fraassen, and B. Skyrms (eds.), Existence and Explanation: Essays in Honor of Karel Lambert, Dordrecht: Kluwer, 131–147. - 1998 “How the World Can Make Propositions True: A Celebration of Logical Atomism”, in M. Omyla (ed.), Skłonności Metafizyczna, Warsaw: Uniwersytet Warszawski, 113–135. Transl. as “Wie die Welt Propositionen wahr machen kann. Ein Lob des logischen Atomismus“, in O. Neumaier (ed.), Satz und Sachverhalt. St. Augustin: Academia,191–228. - 2000a “Continuants and Occurrents”, The Aristotelian Society, Supplementary Volume LXXIV, 78–101. - 2000b “How to exist at a time when you have no temporal parts”, The Monist, 83, 419–436. - 2000c “Identity through Time and Trope Bundles”, Topoi, 19, 147–155. Strawson, P.F. 1959 Individuals. An Essay in Descriptive Metaphysics, London: Methuen.

183

Persistence ERWIN TEGTMEIER, Mannheim Persistence is nothing but duration. The ontological problem of persistence is therefore just the task of analysing duration ontologically. The temporal absolutist does that by an interval on the continuum of time points, the temporal relationist by a relation such as 'last longer than'. The presentist who has to be also an absolutist analyses duration as a movement over a certain number of time points. Lewis' distinction between endurance and perdurance is based on the latter view, although Lewis explicitly rejects presentism. The assay of duration is closely related to the assay of the ordinary thing and of its change. Persistence means literally nothing but a longer duration. What has a longer duration, persists. However, duration is ambiguous. 2 years is a duration but also that from August 14, 1998 to August 14, 2000. Call the latter “date” and the former “duration”. An object which has a duration of two year persists for two years. An object which has the date given as an example persists (persisted) from August 14, 1998 to August 14, 2000. However, different ontologists analyse duration and date differently. Those who assume points of time and thus are temporal absolutists have to analyse a longer duration in terms of a greater number of time points or rather a greater interval of time and a date as a certain interval of time. According to the absolutist view things which are located at a greater number of time points, at a longer interval of time are persistent things. Those ontologists who reject time points advocate temporal relations instead and are therefore called the temporal relationists. If they are consistent they assay persistence also relationally, e.g. by the relation 'lasts longer than'. They take persistent things to be characterised by being longer than some standard not too ephemeral thing. Presentists have still another view of duration although they are absolutists. They do not think of duration not as an interval of time points but as a movement through the time points of an interval. Some ontologists deny that there are any persistent things. They claim that there are only things of a minimal duration and what seem to be persistent things is in reality the circumstance that the things of minimal duration form a temporal series. The continuity which sometimes presents itself to us in perception is grounded by them customarily on a close causal

connection between the members of the respective temporal series. However, they have to assume that a series is not an entity, i.e., not an existent. If they do that they cannot attribute to it duration since duration can be, like any other property or relation, attributed only to existents and thus only the members of the series and not to the series as a whole. Under this view one cannot say though that duration underlies the phenomenon of persistence. Moreover, all things have in this view the same duration, namely a minimal duration which may even be a zero duration defining for a time point in the strict sense. If a series is taken to be an entity and if it is thought to be the persistent thing a duration should be attributable to it. Yet, it seems difficult to ascribe a duration to the complex (the series) just because it is a complex. Obviously, temporal series with gaps (corresponding to temporally interrupted objects) have no duration. Thus it seems that at least not all temporal series have a duration. However, the temporal series which correspond to persistent objects must not have gaps. If it had gaps it would be out of the question. It would lack persistence. Therefore, the difficulty of attributing a duration is not practical one. The interval between the time points of the first and the last member of the series is an obvious and perfect solution. The principal difficulty of it is that durations can hardly be attributes of simple as well as of complex entities and of the members of the series as well as of the series as whole. Primarily, duration is attributable to the members of the series. If we add up the scale values of durations the result means only that the members have certain durations. We cannot take the result as an attribute of the series as a whole. Moreover, the members of the series are mostly thought to be point-like and thus to have no duration. Zero durations can, of course, not added up to a positive duration of the series. To summarise: the ontological grounding of duration depends on the general view of time (whether it is absolutist or relational) and on the view of the things which have durations; one can distinguish three ontological assays of the persistent object: the strong series assay with series as entities, the weak series assay with series as non-entities and the substance solution with a simple persistent thing. However, one cannot assess an assay in isolation. One has to take into account the whole ontology into which it belongs, not only its view of time. Therefore, I will sketch my ontology before offering my solution to the problem of persistence The world consists of things, facts, and categorial forms. Things and facts have categorial forms. Its categorial form makes a thing either an individual or a universal of a certain type. Facts are made by their form either exemplification facts or negative facts or conjunctive facts or

186

disjunctive fact or conditional facts or general-quantified facts or at-leastone-quantified fact. Exemplification facts ground the possession of properties and the holding of relations. Thus an individual and a one-place universal of the first order form an exemplification fact. So do two individuals and a two-place universal of the first order or a universal of the first order and a one-place universal of the second order or two universals of the first order and a two-place universal of the second order etc. Exemplification facts consist thus of individuals and universals and their togetherness in the fact has a certain form which is an additional entity but not an additional constituent. The facts of the other kinds (subcategories) consist of facts. Negative facts and quantified facts have one fact as a constituent and the form of negation or the form of one of the two kinds of quantification, The remaining facts consist of more than one fact and have the forms of conjunction, disjunction or conditional. Thus facts have a large scope in the world. But there are also things in it, either individuals or universals. Now, the problem of persistence seems to concern mainly individuals. Individuals persist. Some individuals persist, some don't. Universals and facts are not that kind of entities which are able to persist because they are not in time. Only individuals are in time. That has to be explained immediately, lest a customary misunderstanding arises from the mereological metaphor of "being in". This metaphor suggests the idea that time is a container. The suggestion is wrong. Time like space is merely a group of relations (universals with more than one place). It is much too narrow and wrong to take the world to be what is "in time and space". Time and space are nothing but certain relations. They belong to the content of the world and are not the form of the world. To support this claim one can point out that universals and facts are perceived although they do not stand in temporal and spatial relations. Naturally, that is a claim partly based on phenomenological data but also on ontological theory. Persistence, as was stated in the beginning, is a matter of duration and duration is based in the ontological assay offered here on the two relational universals 'lasts as long as' and 'lasts longer than'. What about the customary definition of persistence given by D. Lewis: "An individual persists iff it lasts longer than some standard individual with a short duration. something persists iff, somehow or other, it exists at various times."1 Though Lewis pretends to give a universally acceptable definition of persistence, he seems to think in absolutist terms, i.e., in terms of the 1

Lewis 1986, 203.

187

view of time according to which there are time points (temporal absoluta). This is the opposite of the view here advocated according to which there only temporal relations. However, an absolutist would usually rather represent the duration of an ordinary object by an interval, not by several time points. Apparently, Lewis thinks of a duration in terms of a running through a series of point-like presents. That is, as was mentioned already, the presentist view which Lewis explicitly rejects. Presentism is to a large extent the view with which we all grow up and it remains the view of most people. Lewis could not free himself from it. There is a implicit and hidden presentism in Lewis. Philosophers moulded by mathematical logic take for granted that time points can easily be introduced into the relational view. They are mistaken. What corresponds in the relational view to a time point would be the overlapping with the individual in a very short time signal (a fact). Now, there is no such instantaneous time signal corresponding to the moments of the absolutist. Moreover, in the relational ontology time points would be relational properties, universals and there are strong arguments against the existence of such properties. Hence, Lewis' Definition of persistence is not applicable in a relational ontology of time. That is true also of his distinction between and his definitions of perdurance and endurance which go thus: “Something perdures iff its persists by having different temporal parts, or stages, at different times, though no part of it is wholly present at more than one time; whereas it endures iff it persists by being wholly present at more than one time".2 Lewis talks with respect to persistence of “existing at a time” and with respect to perdurance and endurance of “being present at a time”. Both seem to refer to a localisation at time points. In the definitions of the two views of persistence the presentism again comes through in that a longer durations is not analysed as a localisation at an interval of time points which is how the temporal absolutist would customarily analyse duration but as localisation at more than one time point. Lewis distinction between the two views also depends on it. According to the endurance view a persistent thing is localised as a whole at several time points while according to the perdurance view it has temporal parts which are located at one time point only. It is not clear whether in the perdurance view the thing as a whole is taken to be located at time points. It sounds as if it were not. This would mean, strictly speaking, that the thing as a whole has no duration. What else could it mean to say that the whole thing is present at times by its temporal parts. Now, according to the standard view only intervals between time points are 2

Lewis 1986, 203.

188

durations but time points are not. Time points are zero durations. That entails that temporal parts don't have durations either. If a time would be a duration it would include more than one time point. We see that Lewis' distinction is unsatisfactory even if restricted to the absolutist view of time. As an alternative to Lewis' division between the view that persisting objects endure and the view that they perdure I propose the division between the view that the persisting thing is nevertheless simple and the view that it is a temporal complex, more precisely a temporal series. The divisions have the presumably the same extensions. But the latter division has the advantage of not being restricted to absolutist views of time, of being free of presentist thinking and of not mixing temporal location and existence. The division between the view that ordinary objects are simple and the view that they are temporal complexes has to be further subdivided. There is the view that ordinary objects are simple but in spite of that have inherents, i.e. other individuals inhere in them (it is the view I take) and the traditional view which does not countenance such inherents. Call the latter the substance view and the former the multifarious substance view3. The complex view has to be subdivided in accordance with the opposition between those which hold that a series is an entity which exists in addition to the members and those which hold that there exist only the members. Call the latter the weak serial view of persistence4 and the former the strong serial view5. One could say that according to the weak serial view the changing ordinary thing does not persist in that there is no entity in addition to the members of the series which is the same as itself. However, the advocates of this view usually offer an explanation of the phenomenon of persistence, namely that the members of the series are causally closely connected and that they are more closely connected causally to each other than to entities not in the series. To which views of the persistent things are Lewis' definitions of “endure” and “perdure” related? “Endure” to the substance view, in a presentist version though, as was mentioned already. “Perdure” presumably to the strong serial view since the definition says that the perduring thing “persists by having different temporal parts, or stages, at different times.” However, there is a certain inconsistency in Lewis' It is developed in a different terminology in Tegtmeier 1992, §§8-10. The term “temporal part” used there is substituted here by “inheres” because the former led to misunderstandings of my view as a serial view and a kind of four-dimensionalism. 4 It is advocated in a sophisticated ontological version by Bergmann 2003, 355ff. 5 It is developed in a sophisticated ontological way by Grossmann 1983, §§36-39. 3

189

explanations since he connects existence with time points which is the presentist view and which excludes temporal series from the realm of existents, since its duration exceed by far that of a time point. Hence, Lewis understanding of existence seems to lead to the weak serial view if it is applied to temporal series.6 The substance view with inherents seems to me the best view. With simple things the unity of the persisting thing is not in doubt nor is its having a duration. Since it does not consist of its inherents its duration is independent of them. The main difficulty of an ontology with simple things known since Parmenides7 arises in the analysis of change. Change is the substitution of a property by another which is incompatible with it and implies its negation. When the light of a lamp changes from red to green that implies that the light afterwards is not red. Let us symbolise the light of the lamp by l and the two incompatible colours red and green by R and G, then the change means in an ontology with negative and conjunctive facts that the conjunctive fact R(l) & G(l) obtains which (together with the general fact of the incompatibility of red and green) logically implies that the fact R(l) & not R(l) obtains. However, the law of non-contradiction entails that such a fact does not obtain. Hence, there must be something wrong with the ontological analysis. The absolutist with respect to time can draw on his time points here. He takes the possession of properties to involve time points. Customarily, he would use the phrase: l is G at t1 and l is R at t2. But what does that mean ontologically? Take R and G to be universals, I as a particular. Time points may be either universals or particulars. In any case, the predicative connection (if it is assumed) has to connect not only particulars and the universals grounding their properties but also always time points. If one wants to render the possession of a property at a time point the only way which suggests itself is to represent the particular and the time point by an ordered pair. This would turn however each predicate into a relational predicate. There would be no nonrelational predicates left. Moreover, it would amount to a transition from the substance view to the serial view, the members of the series being the combinations between a particular and a time point. The one thing would be replaced by a combination of it and a time points similar to the members of temporal series of the serial views. For the temporal relationist time points are not available. He rejects them. Mainly, because we never perceive time points. What we perceive is always a simultaneity to some reference process such as the motion of a 6 7

s. Tegtmeier 2002. Contemporary analytic philosophers credit D. Lewis with its discovery.

190

clock or of the earth or of other heavenly bodies or to some time signal. We saw that relational substitutes of time points are not available either. Hence the substance ontologist who is temporal relationist has to look for another solution of the contradiction of change. My solution is based on the one hand on the relation of inherence which holds between some individuals and on the other hand on the claim that there are more individuals and substances than is usually supposed. Where we usually realise only one individual there is a multitude of individuals related by the relation of inherence. There is not only the lamp with the changing colours but also the lamp with the red light and the lamp with the green light. The first lamp is neither red nor green. It has only the property of being a lamp. It has only what have traditionally been called the substantial or essential properties. When we perceive the lamp with the red light we usually do not mean that it is always red. We are well aware that it is red only for some time. Hence, it suggests itself that what we perceive is the following quantified fact: for some individual x, which stands to the substance lamp before us in the relation of inherence, x is red. The inherents are also lamps. I call individuals which have inherents (i.e. in which other individuals inhere, which stand as first terms in the relation of inherence to other individuals) and have the substantial properties only substances. According to my view, all individuals, be they substances or not, have a duration, have a duration greater than zero and the duration of the substance is not derived from the durations of their inherents though there are laws connecting them. Hence the problem of persistence as that of duration arises also for inherents (the problem of change does not). Inherents can differ with respect to duration in contrast to the serial view according to which all members have minimal durations. The durations of inherents are connected with change. There is some inherent of the substance, e.g., of the lamp the duration of which corresponds to how long it was green in one case and another corresponding to how long it was red in one case. Having a duration means for the temporal relationist, who assays not only temporal location but also duration relationally, that there is another particular which either lasts longer or last the same time or it last longer than that other particular. If there are facts with the relation ‘lasts longer’ this implies that individuals have durations of different lengths. It follows also that simultaneity (which is one of the relations in the time group besides 'being earlier than' and 'overlapping temporally') is not a point-to-point affair (in contrast to the serial views). Rather the relation 'simultaneous with', which holds between individuals, requires that the respective individuals have the same duration. There is besides the

191

relations 'earlier than' and 'simultaneous with' also the relation 'overlaps temporally' which implies the holding of the relation 'last longer than'. If the individual x temporally overlaps the individual y than x last longer than y and x has a longer duration than y. What prevents the contradiction is that for each kind of universal each individual exemplifies only one universal. Thus the lamp with the red light is another individual than the lamp with the green light. Both stand in the relation of inherence to yet another individual which has neither red nor green light but which is a lamp. This is the subject of the change of colour. Thus an individual changes by individuals which inhere in it. Only substances as defined above change. Individuals which inhere in a substance are overlapped temporally by it. Having a longer duration does not imply for an individual that other individuals inhere in it, though it makes it more probable. What obviously implies the inherence of other individuals is change. At any rate the analysis of change by substances with inherents does not lead into contradictions. Substances are individuals with inherents and hence changing individuals. Individuals which don't have inherents do not change. Thus inherents which do not have inherents do not change. Are there inherents which have inherents, i.e., substances which inhere in other substances? That would entail that there is substantial change, i.e. substantial properties are substituted by other substantial properties, e.g., when a tadpole is turned into a frog. I follow Avicenna in holding that there is not only substantial change in the sense of a transition between substantial properties but also that the changing entities is a substance which he continues to call like Aristotle “prime matter”. 8 I depart from Avicenna in taking the subject of substantial change to be a fully individuated entity, namely an individual. Being individuated means being diverse (i.e. numerically different) from all other entities (existents). In the ontology with individuals, universals, facts and forms advocated here each entity is individuated. The substances in which substances inhere cannot have the substantial properties which are interchanged in the substantial change lest the contradiction of change occurs. The substance which survives the substantial change has to have other substantial properties than the inhering substances. In the case of the substantial change from tadpole to frog the substantial property of the surviving substance could be called “being a fish-reptile”. And this property must be involved in the substantial properties of the tadpole as well as in that of the frog. s. Avicenna: Das Buch der Genesung der Seele (transl. M. Horten). Frankfurt 1960, 2. Abhandlung 4. Kapitel. 8

192

Closely connected with the phenomenon of persistence is that of unity through time and change. With the substance view the solution is ready: the simplicity of the substance which is thought to be unaffected by the inherence of the accidents. According to the multifarious substance view the unity is based on the holding of the relation of inherence to one and the same substance and on the simplicity of that substance. That several individuals inhere in a substance does not imply that it is complex, that it consists of those subjects. Rather it means only that there holds a certain relation between the individuals. Only facts are complex. They consist of things and facts. Like the classical substances the substances advocated here are simple. They are connected to their inherents by virtue of certain relational facts with the relation of inherence. Hence when we meet some ordinary object again which has changed in the meantime it is literally the same object, according to the multifarious substance view, though we mostly meet also another inherent of it. The unity of the changing object seems to be better grounded by the strong serial view since it takes the changing thing actually to consist of those things which, according to the multifarious substance view, merely stand to it in the relation of inherence. The ontological advantage of the latter view is that the substance can have inherents and still be simple. It seems that with the strong serial view it is more clearly the persisting thing, taken as a temporal series, which changes. However, even with the multifarious substance view it is clear that the inherents as such are not the subjects of change. The change incorporated in the temporal series of inherents with their opposing attributes has to be attributed to the substance in which they inhere. There is no alternative. Change involves a chain, a temporal series of individuals inhering in one and the same individual. The latter is the one which changes. However, it changes by virtue of the inherents. As to properties which are also involved in any change they cannot belong to the changing individual since they are always incompatible. If a thing changes from round to oval it has to be either not round nor not oval since being round implies not being oval and having both is therefore contradictory. The conclusion drawn in the ontological analysis of change offered by the multifarious substance view is that the round individual is not the same individual as the oval one though they inhere in the same individual. They form a temporal series which is based on a relational fact with the two inhering individuals and the relation 'earlier than' as constituents and which means that there is a transition from the round to the oval individual. This phenomenon is well-known from the numerous changes we constantly perceive and the relation 'earlier than' is meant to be the ontological ground of it. Thus according to the ontological

193

assay proposed here there is a transition between the inherents of the changing individual but no transition at the changing individual itself. The weak serial view implies the non-existence of the persisting thing, as was mentioned already; nevertheless it grounds in a way the unity of the temporal series of momentary objects by the close causal connections between them and by lawful connections based on substantial properties. Under the strong serial view the unity of the changing thing is as close as that of the series in the respective ontology. Since a series is in any case a complex the unity of the changing thing is not as close according to the strong serial view as according the the multifarious substance view which makes that thing a simple. The unity is close still in the traditional substance view. This does not assay change in terms of a series adjoined to the simple changing thing but in no way impairs the unity of the changing thing. It views change like we customarily do as a substitution of properties. The price it pays is presentism. That is a high price since presentism run into several grave difficulties. Moreover, it is a highly implausible view which amounts to reducing the world to a present in the attempt to avoid contradiction. Basically, the contradiction of change is eliminated by expelling the initial state before the change simply from the world.9 It has been argued by McTaggart that a series cannot be a change because the positions of its members are fixed. He thinks that one has to mobilise the series by a moving present. McTaggart's argument made a permanent impression until today. Nevertheless, it arises from a misunderstanding of the temporal series generated by Russell's relation 'earlier than' and of the task of ontological analysis. I emphasised that 'earlier than' is the relation of transition. It grounds the temporal aspect of change. McTaggart seems to take the task of ontology to be the simulation rather than the analysis of phenomena. Moreover, he can make his point only as long as he does not complete his analysis, namely as long as he does not analyse the moving of the present. This moving has in a Russellian frame to be analysed with the relation 'earlier than' corresponding to succession of the presentness of successive members of Russell's original earlier-series.10 Customarily, persistence and existence are closely associated, not to say equated, with respect to concrete objects. The duration of a substance is understood as the span of its existence. That is an implication of the view that becoming is coming into existence and passing away is a 9

cf. Tegtmeier 1994 cf. Tegtmeier 2007

10

194

transition into non-existence. However, this view leads also into contradictions and the contradictions of becoming cannot be resolved in the same way as the contradictions of change just because there are no inherents (i.e. no inherent individuals) before the beginning of a substance. Hence existence has to be separated from becoming and persistence.11 Persistence is sometimes explicated as identity through time and change. Now, if “identity” is taken as “numerical sameness” (sometimes called “strict identity) the explication is in accordance with the substance and the strong serial view but not with the weak serial view. For the advocate of the weak serial view persistence is not numerical sameness. However, it is doubtful whether the strong serial view as D. Lewis advocates it (namely as perdurance) entails that persistence is sameness. Due to an implicit presentism, Lewis assumes a successive localisation in time which furnishes a temporal location for the members (mostly called “temporal parts”) but not for the series as a whole and he associates existence and temporal location. That seems to exclude a temporal continuity of the whole object. Insofar Lewis continues the empiricist tradition aiming at dissolving the Aristotelian substance and failing to acknowledge the complexes which serve as substitutes. In a way I would vindicate that position because I think that persistent temporal series don't work because of the entanglement of temporal and spatial complexity. REFERENCES Bergmann, G. 2003 Collected Works II, Frankfurt/Lancaster: Ontos. Grossmann, R. 1983 The Categorial Structure of the World, Bloomington: University of Indiana Press. Lewis, D. 1986 On the Plurality of Worlds, Oxford and Cambridge (Mass.): Blackwell. Tegtmeier, E. 1992 Grundzüge einer kategorialen Ontologie, Freiburg and München: Alber. Tegtmeier, E. 1994 Der Hyperdynamismus in der Ontologie der Zeit. Eine Auseinandersetzung mit Brentano und McTaggart. Logos 1. - 1999 “Parmenides' Problem of Becoming and its Solution”, History of Philosophy and Philosophical Analysis, 2. − 2002 „Warum Lewis' Unterscheidung zwischen Mitdauern (perdure) und Währen (endure) verfehlt ist“, in W. Löffler (Hrsg.), Substanz und Identität, Paderborn: mentis, 83-95. − 2007 Three Flawed Distinctions in the Philosophy of Time. Metaphysica 8.

11

s. Tegtmeier 1999.

195

The Editor

CHRISTIAN KANZIAN Institute of Philosophy, Theological Faculty, University of Innsbruck, Austria [email protected]

List of Authors

STEVE BARKER (co-author with Dowe) Department of Philosophy, University of Nottingham, Nottingham, UK [email protected] ARKADIUSZ CHRUDZIMSKI Department of Philosophy, University of Salzburg, Austria Department of Philosophy, University of Szczecin, Poland [email protected] PHIL DOWE (co-author with Barker) Philosophy, HPRC, University of Queensland, Brisbane, Australia [email protected] PIERRE GRENON (co-author with Smith) Institute for Formal Ontology and Medical Information Science, University of Saarland / Department of Philosophy, University of Geneva, Germany / Switzerland [email protected] LUDGER JANSEN Department of Philosophy, University of Rostock, Germany [email protected] KRISTIE MILLER Philosophy, HPRC, University of Queensland, Brisbane, Australia [email protected]

JONATHAN LOWE Department of Philosophy, Durham University, UK [email protected] UWE MEIXNER Department of Philosophy, University of the Saarland, Germany [email protected] EDMUND RUNGGALDIER Institute of Philosophy, Theological Faculty, University of Innsbruck, Austria [email protected] JOHANNA SEIBT Department of Philosophy, University of Aarhus, Denmark [email protected] BARRY SMITH (co-author with Grenon) Institute for Formal Ontology and Medical Information Science, University of Saarland / Department of Philosophy, University at Buffalo, Germany / USA [email protected] PETER SIMONS Department of Philosophy, University of Leeds, UK [email protected] ERWIN TEGTMEIER Department of Philosophy, University of Mannheim, Germany [email protected]

198

PhilosophischeAnalyse PhilosophicalAnalysis 1 Herbert Hochberg Russell, Moore and Wittgenstein The Revival of Realism

8 Rafael Hüntelmann Existenz und Modalität Eine Studie zur Analytischen Modalontologie

2

9 Andreas Bächli / Klaus Petrus Monism

ISBN 3-937202-00-5 334 pp., Hardcover € 94,00

Heinrich Ganthaler Das Recht auf Leben in der Medizin Eine moralphilosophische Untersuchung

ISBN 3-937202-01-3 167 pp., Hardcover € 58,00

3 Ludger Jansen Tun und Können Ein systematischer Kommentar zu Aristoteles’ Theorie der Vermögen im neunten Buch der „Metaphysik“ ISBN 3-937202-02-1 302 pp., Hardcover € 70,00

4 Manuel Bremer Der Sinn des Lebens Ein Beitrag zur Analytischen Religionsphilosophie ISBN 3-937202-03-X 134 pp., Hardcover € 58,00

5 Georg Peter Analytische Ästhetik Eine Untersuchung zu Nelson Goodman und zur literarischen Parodie ISBN 3-937202-04-8, 332 pp. Hardcover € 94,00

6 Wolfram Hinzen / Hans Rott Belief and Meaning Essays at the Interface ISBN 3-937202-05-6 250 pp., Hardcover € 58,00

7 Hans Günther Ruß Empirisches Wissen und Moralkonstruktion Eine Untersuchung zur Möglichkeit von Brückenprinzipien in der Natur- und Bioethik ISBN 3-937202-06-4 208 pp., Hardcover € 58,00

ISBN 3-937202-07-2 189 pp., Hardcover € 58,00

ISBN 3-937202-19-6 340 pp., Hardcover € 70,00

10 Maria Elisabeth Reicher Referenz, Quantifikation und ontologische Festlegung ISBN 3-937202-39-0 ca. 300 pp., Hardcover € 89,00

11 Herbert Hochberg / Kevin Mulligan Relations and Predicates ISBN 3-937202-51-X 250 pp., Hardcover € 74,00

12 L. Nathan Oaklander C. D. Broad's Ontology of Mind ISBN 3-937202-97-8 105 pp., Hardcover € 39,00

13 Uwe Meixner The Theory of Ontic Modalities ISBN 3-938793-11-2 374 pages, Hardcover,€ 79,00

14 Donald W. Mertz Realist Instance Ontology and its Logic ISBN 3-938793-33-3 252 pp., Hardcover, EUR 79,00

15 N. Psarros / K. Schulte-Ostermann (Eds.) Facets of Sociality ISBN 3-938793-39-2 370 pp., Hardcover, EUR 98,00

16 Markus Schrenk The Metaphysics of Ceteris Paribus Laws ISBN 13: 978-3-938793-42-8 192pp, Hardcover, EUR 79,00

EditedBy • HerbertHochberg • RafaelHüntelmann ChristianKanzian • RichardSchantz • ErwinTegtmeier

PhilosophischeAnalyse PhilosophicalAnalysis 17 Nicholas Rescher Interpreting Philosophy The Elements of Philosophical Hermeneutics ISBN 978-3-938793-44-2 190pp., Hardcover € 89,00

18 Jean-Maurice Monnoyer(Ed.) Metaphysics and Truthmakers ISBN 978-3-938793-32-9 337 pp., Hardcover € 98,00

19 Fred Wilson Acquaintance, Ontology, and Knowledge Collected Essays in Ontology ISBN 978-3-938793-58-9 XX, 726., Hardcover, EUR 159,00

20 Laird Addis, Greg Jesson, and Erwin Tegtmeier (Eds.) Ontology and Analysis Essays and Recollections about Gustav Bergmann ISBN 978-3-938793-69-5 312 pp., Hardcover, EUR 98,00

EditedBy • HerbertHochberg • RafaelHüntelmann ChristianKanzian • RichardSchantz • ErwinTegtmeier