What Are We? Temporal Parts
Olson (Eric)
Source: What Are We? A Study in Personal Ontology, Chapter 5 (November 2007: Oxford University Press.)
Paper - Abstract

Paper SummaryNotes Citing this PaperText Colour-Conventions


(OO-L): This chapter examines David Lewis's view that we are temporal parts of animals. It examines three arguments for the view that persisting things have temporal parts–four-dimensionalism. One is that it solves the problem of temporary intrinsics. The second is that it solves metaphysical problems about the persistence of material objects without the mystery of constitutionalism–though these solutions require a counterpart-theoretic account of modality. The third is that it solves problems of personal identity–involving fission, for instance–in an attractive way. It is then argued that according to four-dimensionalism, the bearers of such properties as thinking and acting are momentary stages, forcing us to choose between saying that we don't strictly think and saying that we don't persist.

  1. Four-dimensional hunks of matter
  2. Temporary intrinsics
  3. Lumps and statues
  4. The problem of modal incompatibility
  5. Puzzles of personal identity
  6. Thinking animals1 and other worries
  7. Thinking stages
  8. The stage view2
Full Text

5.1 Four-dimensional hunks of matter
  1. Having considered the view that we are spatial parts of human animals3, let us turn now to the idea that we are temporal parts of animals. The idea that animals have temporal parts is not easy to understand. The notion of a temporal part is inherently confusing. Even professional philosophers frequently get it wrong. For numerous reasons it is foreign to our ordinary ways of thinking. Learning to convert ordinary thoughts about persisting things into the ontology of temporal parts is a bit like learning a foreign language – only harder, because the dictionaries and grammar books are incomplete. I will try to explain what it means to say that we are temporal parts of animals, and to explore some of its consequences. Despite my efforts to do this in non-technical terms, however, I fear that some of what I am going to say will sound like Greek4.
  2. You and I are extended in space. (We appear to be, anyway, though a few philosophers disagree. Suppose we are.) We seem to be extended in space by having different parts – spatial parts – located in different places. For instance, you are now both above and below the seat of your chair: partly above it and partly below. That is because you have a part, or several parts, that are above the seat of your chair – wholly above it – and another part, or several parts, that are entirely below it. You fill up space by having parts spread out across it. Whether you have a different part for every bit of space you partly fill may be contentious (it is what van Inwagen (1981) has called the doctrine of arbitrary undetached parts); but it seems clear that you have a great many spatial parts.
  3. We also persist through time. (We appear to, anyway. Suppose we do.) Some philosophers think that we relate to time in much the same way as we relate to space. You have a location in time, just as you have a location in space. And just as you are extended in space, you are also extended in time: most of us extend for a bit less than two meters in space and for a bit less than a century in time. You are extended in time by having temporal parts spread out across time. Just as you are in different places by having different spatial parts located entirely in those places, the idea goes, you are at different times by having different temporal parts entirely located at those times. If you exist both in 2000 and in 2010, that is because you are partly in 2000 and partly in 2010. In other words, you have a temporal part located in 2000 – located wholly during that year – and another temporal part entirely located in 2010. Whether you have a different temporal part for every bit of time you partly fill may be contentious; in any case you have a great many temporal parts. Moreover, every part of you shares a part with one or more of your temporal parts: you are composed of temporal parts.
  4. We all know what spatial parts are. They are just ordinary parts of spatial objects. If we know what ‘part’ means, we know that hands (if there are such things as hands) are parts of human beings. But what are temporal parts? A temporal part of a thing is supposed to be a part of it that incorporates “all of that thing” for as long as the part exists. A temporal part of you is spatially just as big as you are, and located just where you are, while it exists. More precisely, a temporal part of you is a part of you that overlaps all of your parts that exist when that part exists (where overlapping means sharing a part). If something is a temporal part of you, then any part of you that doesn’t overlap that thing will exist only at times when that thing doesn’t exist. So:
      x is a temporal part of y =df x is a part of y, and every part of y that does not overlap x exists only at times when x does not exist.
    Your nose may be a part of you, but it is not a temporal part, for it doesn’t overlap all of your parts that exist when it does. You have parts that exist when your nose does without overlapping it: your feet, for instance. Your nose doesn’t incorporate all of you while it exists: it’s too small to be a temporal part of you. But suppose there is such a thing as your first half. It would be just like you, apart from lasting half as long. It would walk and talk and study philosophy. It would take up all of you for as long as it existed. Any part of you that didn’t overlap it – your final hour, say, or a certain wisp of grey hair – would exist only at times when your first half didn’t exist. Suppose, if you can, that this thing would be a part of you. Then it would be a temporal part of you5.
  5. We ordinarily think of events, such as games – particular matches that occur at particular times and places – as being made up of briefer events or phases, and these seem to be temporal parts of those longer events. A baseball game takes up time by having different parts located at different times. It starts at six and is still going on at ten past seven insofar as one part of it – the first inning, say – starts at six, and another part of it – the third inning – is going on at ten past seven. And each inning takes up “all of the game” while it is going on: any part of the game that doesn’t overlap its third inning exists, or occurs, only when the third inning is not going on.
  6. It also seems that any temporary property we attribute to the game at a particular time is a property that some temporal part of the game occurring at that time has without temporal qualification: the game is dull between six and twenty past because the first inning is dull – not dull at that time, but just plain dull – and exciting at ten past seven because the third inning is exciting. We will return to this important point in the next section.
  7. Perhaps there is a long-running event consisting of everything you ever do and all that ever happens within your boundaries: your career or history. Its temporal parts or phases include your birth, your first day at school, your adolescence, and your reading of this book. The view I want to consider is that you have temporal parts just as your history does. Most of those who take this view say that there is a temporal part of you coinciding with every portion of your history: a part for every time, instantaneous or extended, continuous or discrete, when you exist. So you have an infant-sized temporal part that extends from your birth until your first birthday, a part that extends from the midpoint of your history until your demise, a part that exists throughout June and August 2010 and at no other times, and many more. In fact you and your career may be so similar, on this view, as to be same thing: you might be your history (Quine 1960: 171).
  8. This view is often put by saying that we are "four-dimensional objects", because it says that we are extended in one dimension of time as well as the three dimensions of space (if that is how many spatial dimensions there are). Others put it by saying that we are "space-time worms", because space-time diagrams are typically drawn in a way that makes a person's spatial extent small compared with her temporal extent: six feet looks small on most diagrams compared with 80 years. These descriptions can be misleading. The view that we are space-time worms does not imply that units of time can be converted into units of space – so many seconds to the meter – so that we are literally longer in time than we are in space. We’re not really worm-shaped. And the claim that we are “four-dimensional” by being extended in time is not strictly the same as the view that we are composed of temporal parts, for it may be that we are extended in time and yet lack temporal parts (Mellor 1998: 86). More generally, the name ‘four-dimensionalism’ has been given to a wide variety of views that have nothing to do with temporal parts. It might be better to call the view that all persisting objects are composed of many temporal parts simply ‘the ontology of temporal parts’. However, I will sometimes follow custom and call it four-dimensionalism. I will call the view that we in particular are composed of many temporal parts the temporal-parts view.
  9. Supposing that we exist and are located in time, there are two broad alternatives to the temporal-parts view. One is that we don’t persist at all: what we think of as a single persisting person is really just one momentary person followed immediately by a similar but numerically different momentary person, followed by another, and so on. If we follow the jargon and call momentary temporal parts of things stages, the view is that we are stages rather than worms. We will come to that view in §5.8. The other alternative is that we persist through time but have no temporal parts. (Or at least no proper temporal parts – no temporal parts other than ourselves. The definition of ‘temporal part’ appears to make everything located in time a temporal part of itself.) We don’t fill up time by having temporal parts spread out across it. There is no such thing as your first half.
  10. By itself, the temporal-parts view says little about what we are6: only that we are persisting concrete objects composed of temporal parts. It leaves entirely open what sort of four-dimensional objects we are. For instance it is compatible with our being bundles of mental states or even immaterial substances. But most advocates of the temporal-parts view say that we are material objects – "four-dimensional hunks of matter", as Heller (1990) puts it. We might be animals, or brains; but the more common view is that we are proper temporal parts of animals (see §5.6). Most of what I will say about the temporal-parts view applies equally to all of these variants.
  11. Those who think that we have temporal parts usually accept four-dimensionalism generally: they think that all persisting things are composed of temporal parts. Most say that nothing could persist without having temporal parts: for a thing to exist at different times is just for it to have different temporal parts located at those times. A number of important arguments support the ontology of temporal parts (usefully surveyed in Sider 2001). Two of them are of particular interest for the question of what we are7, and are the subjects of the next two sections.
5.2 Temporary intrinsics
  1. The most common argument for four-dimensionalism is that it offers the best solution to something called the problem of temporary intrinsics or problem of change (Lewis 1986: 202- 204, Sider 2001: 92-98). Intrinsic properties are roughly those a thing has by virtue of the way it is in itself: those it would share with any duplicate of itself. (This is not a satisfactory definition. Intrinsicness is notoriously hard to define. But it will do for present purposes.) Extrinsic properties – those that are not intrinsic – are those a thing has by virtue of the way it relates to things outside itself. Shape is an intrinsic property; being an uncle and being north of the equator are extrinsic. The argument is highly abstract, but it is essential to understanding the ontology of temporal parts (Olson 2006c).
  2. I am now sitting down: I have a bent shape. An hour ago I was standing: I was straight, and not bent. So I am bent, and yet not bent: somehow I both have and lack the intrinsic property of being bent. The same goes whenever something has any intrinsic property temporarily. For that matter, it goes for temporary extrinsic properties: I am now an uncle; five years ago I wasn't; so somehow I both am and am not an uncle. Now it is plain enough that nothing can literally be both bent and not bent. Logic ensures that nothing can both have and lack the same property, or have both that property and its complement. Yet I clearly relate in some way to both the property of being bent and the property of being not bent. What way is it? The problem of temporary intrinsics is to answer this question.
  3. The answer might seem obvious. I am bent now, and not bent an hour ago. For a thing to be bent is for it to be bent at some particular time. If someone is bent, we can always ask when he is bent; and there will always be an answer, even if it is that he always was and always will be bent. A persisting thing has its properties at particular times, and if something has incompatible properties such as bentness and straightness it must have them each at a different time. That, as Aristotle said, is what time is for. the obviousness of this reply might even lead us to wonder whether there is really any problem of temporary intrinsics to be solved. Temporary properties seem problematic, the idea would go, only if we infer from my being bent at some particular time that I am just plain bent, without any temporal qualification at all, and if we likewise infer from my not being bent at some time that I am just plain not bent. But why should that follow? No one ever thought that deleting temporal qualifications was a valid mode of inference, like deleting double negations. And if we don’t commit that fallacy, where is the problem of change?
  4. But four-dimensionalists are not satisfied with this reply. Even if it is right that to be bent is to be bent at a particular time, they say, we need to explain what it is for something to be bent at a time, as opposed to being just plain bent. How does the temporal qualification – my being bent now – block the inference to my being bent simpliciter? If my relation to the properties of bentness and straightness is not simply having, how do I relate to them?
  5. The earlier thought that things don’t just have temporary properties like bentness, but have them at times might suggest that the temporal qualification is built into the having relation. So having or instantiating is a three-place relation involving an object, a property, and a time (or a four-place relation involving two objects, a two-place relation, and a time, or the like). Some things may have some properties without temporal qualification – maybe the number 7 is odd simpliciter – but that is another story.
  6. Four-dimensionalists object that this makes intrinsic properties extrinsic. The proposal is that being bent involves a relation not only to a property, but also to a time. For a thing to be bent, it says, is for it to relate in a certain way to something outside it – a time. Whether a thing is bent or straight depends on how it relates to times, just as whether someone is an uncle depends on how he relates to other people. It follows, the complaint goes, that nothing could be bent in itself: “In itself, considered apart from its relations to other things,” Lewis complains, a thing would have “no shape at all” (1986: 204). Shape would therefore not be an intrinsic property. More generally, no property could be both temporary and intrinsic, for a property could be had temporarily only if that having were relative to a time in a way that would make the property extrinsic. But surely we know that shape is both temporary and intrinsic.
  7. Whatever the merits of this reasoning may be, it has an important consequence: things must have their intrinsic properties timelessly, and not relative to times. Whatever has bentness must be just plain bent, without temporal qualification; otherwise its being bent would involve a relation to a time, making bentness extrinsic. That, the idea goes, is part of the nature of intrinsicness.
  8. Four-dimensionalists solve the problem of temporary intrinsics in a way that respects this consequence. They say that things have different properties at different times by having different temporal parts, located at those times, that have the properties without temporal qualification. I am bent now insofar as the temporal part of me located at this instant – or at any rate some temporal part of me that exists now and only when I am bent – is just plain bent. That part is not temporarily bent, or bent now, or even bent at every time when it exists. It is bent simpliciter, timelessly bent, just as 7 is timelessly odd. Its being bent involves no relation to a time whatever. That makes bentness an intrinsic property, as we always knew it was. Likewise, I am straight, and not bent, at some other time insofar as a temporal part of me located at that other time is timelessly straight. Moreover, I have those parts without temporal qualification. I don’t have a straight temporal part at one time and lack one at another, for having a straight temporal part and not having one are no more compatible than being bent and being straight, and just as intrinsic. (My straight temporal part and my bent temporal part are located at different times; but they are not parts of me at different times.)
  9. This means that I don’t simply have either the property of being bent or the property of being straight. How then do I relate to those properties? Am I bent, or am I straight? The answer is neither: I simply have a bent temporal part and a straight temporal part. (Compare the spatial analog: Am I hand-shaped, or am I foot-shaped? Neither: I simply have a handshaped part and a foot-shaped part.) My relation to the property of bentness is that of having a part that has that property simpliciter. My relation to the property of straightness is that of having another part that has that property simpliciter.
  10. So according to four-dimensionalism, when we say that a thing has different intrinsic properties at different times we are speaking loosely. The strict truth of the matter is that the thing has different temporal parts, located at those times, that have those properties without temporal qualification. In the strictest sense I don’t have the properties of bentness or straightness at all. My properties are not being bent or being straight, but rather having a bent part and having a straight part. If I had bentness and straightness (or being bent and not being bent), I should have to have them relative to times, which four-dimensionalists say would make them objectionably extrinsic. Nothing can strictly have both bentness and straightness, since they are incompatible; and nothing can have either property relative to a time, since they are intrinsic; thus, whatever has bentness must be bent and not straight simpliciter. So the things that are bent or straight are brief temporal parts of me, rather than I myself. I can have the intrinsic properties having a bent part and having a straight part, though, for unlike bentness and straightness they are compatible. I can have them without temporal qualification.
  11. The argument for four-dimensionalism, then, is this: There are intrinsic properties, such as bentness and straightness, that each of us in some sense has. Because these properties are intrinsic, we cannot have them relative to times, but must relate to them in some timeless way. But we cannot simply have them timelessly, for they are incompatible: nothing can be both bent and not bent. Therefore I am bent now only in the sense of having (timelessly) a temporal part located now that is bent simpliciter, and I was straight an hour ago insofar as I have another temporal part located then that is straight simpliciter. It follows that I have at least two temporal parts. Moreover, the fact that my intrinsic properties are continuously changing – the way my atoms are arranged, for instance, is changing continuously, owing to their constant motion – implies that I have a different temporal part for every moment when I exist. And what goes for me goes for concrete objects generally.
  12. Ironically, this argument has the very consequence it accuses the opposing view of having, namely that no property could be both temporary and intrinsic. If, per impossibile, something did first have and then lack an intrinsic property such as bentness, then according to the argument that thing would both have and lack bentness, for bentness, being intrinsic, can only be had or lacked without temporal qualification. According to four-dimensionalism, no intrinsic properties are temporary. Our momentary temporal parts have properties such as bentness, and we have properties such as having bent parts; but nothing has those properties temporarily. (This presumably goes for extrinsic properties as well: I am an uncle now insofar as I have a temporal part located now that is an uncle without temporal qualification. No one will suppose that I bear a timeless, two-place instantiation relation to intrinsic properties and a three-place instantiation relation involving a time to extrinsic properties.)
  13. For this reason and others, four-dimensionalists disagree about the persuasive force of the argument from temporary intrinsics. That is, they disagree about how much better their own solution to the problem of change is than the alternative view that things have temporary properties relative to times8. Even so, they nearly all agree that their solution is true: that persisting, changing objects don’t strictly have temporary intrinsic properties, but instead have short-lived temporal parts that have intrinsic properties timelessly. This seems to belong to the very idea of a temporal part. We will see the importance of this in §5.7.
5.3 Lumps and statues
  1. Four-dimensionalism has important theoretical virtues. One of them is its capacity to solve problems about the identities of material objects in a different way from constitutionalism (see Sider 2001: Ch. 5 for a detailed discussion).
  2. Recall the clay-modelling puzzle (§3.2). A shapeless lump of clay is modelled first into a statue and then into a cube. We want to say that the same lump of clay is first shapeless, then statue-shaped, then cubical, but that the statue is never shapeless or cubical. If this is so, then the statue and the lump must be two different things. Constitutionalists say that the statue and the lump are numerically different even though they coincide materially, making the statue physically indistinguishable from the lump at every time when the statue exists. But we expect physically indistinguishable material things to be indistinguishable in all their intrinsic properties. So their view makes it a mystery how the two objects could differ in their intrinsic properties (§3.4). For instance, what could give them different persistence conditions?
  3. Four-dimensionalists say that the statue is a temporal part of the lump: the largest such part that is statue-shaped. (Suppose the lump is made into a statue only once.) We can illustrate this with a space-time diagram, where the vertical axis represents space and the horizontal axis time: Figure 5.1.
  4. The statue is made at t1and squashed at t2. The whole “worm” is the lump, and the shaded part in the middle is the statue. The lump is composed of earlier temporal parts that have a nondescript shape, later temporal parts that are cubical, and intermediate temporal parts shaped like Thatcher (as it may be); it is these last that compose the statue. So the lump and the statue don't coincide materially, in the sense of there being things that compose both of them: the lump has temporal parts – timelessly has them – that are not parts of the statue, and which do not even overlap any parts of the statue. That is what makes the two objects qualitatively different. We can explain the qualitative differences between the statue and the lump in terms of the differences in their parts: the lump is different from the statue because it, but not the statue, has temporal parts that are not statue-shaped. The lump and the statue are no more alike than a baseball game and its third inning are alike.
  5. In the replacement puzzle (§3.3) we burn a clay statue's arm to ashes and replace it with a new arm made of different clay (where this is the only change of parts in the statue’s career). We want to say that the statue persists through this change, while the lump of clay is destroyed and replaced by a new lump. Constitutionalists say that the statue coincides materially first with one lump and then with another. Four-dimensionalists say that the statue is (timelessly) composed of temporal parts of two different lumps.
  6. In Figure 5.2 The first lump, L1, occupies the forward-slashed space-time region; the second lump, L2, occupies the back-slashed region. If we suppose again that the statue is made at t1and squashed at t2, it occupies the horizontal region located between those times. The earlier part of the statue, on this view, is a temporal part of L1; the later part of the statue is a temporal part of L2. The statue is not identical with either lump; but neither is it composed of the same parts as either lump. Neither lump coincides materially with the statue. Again, the qualitative difference between the statue and the two lumps is explained by their having qualitatively different temporal parts. (The four-dimensionalist’s solution to the amputation puzzle is similar; we will come to it in §7.4.)
5.4 The problem of modal incompatibility
  1. The four-dimensionalists’s story about lumps and statues may sound agreeable, particularly in comparison with constitutionalism. So it may seem to support the claim that persisting objects, including ourselves, are composed of temporal parts. There is, however, something about this story that we might find deeply puzzling.
  2. Suppose we ask why the statue ceases to exist when we squash it, but the lump doesn’t. According to the story there is no physical difference between the statue and the lump when we squash them, for they share all their temporal parts while they both exist. How, then, can objects that are physically identical at a time, and treated in the same way, behave so differently at that time? Isn’t this the very objection we made to constitutionalism? Four-dimensionalists say that we are to explain the qualitative difference between statue and lump in terms of their having qualitatively different temporal parts. That seems to imply that the statue perishes and the lump carries on in the clay-modelling story because the lump has certain temporal parts located after the squashing and the lump hasn’t. But that is no explanation at all. It is like saying that Descartes ceased to exist in 1650 because his career doesn’t extend beyond that year.
  3. If squashing a clay statue really does, necessarily, destroy the statue without destroying the lump it is made of, we expect the reason to be that lumps have something statues lack, namely a power or capacity to survive squashing. Otherwise it would be a mystery why clay statues always perish when squashed, yet lumps of clay never do. If lumps and statues didn’t differ in this way, we should expect them to behave more or less alike when squashed, and we should be unable to say confidently of any statue that if it were squashed it would perish. But such a difference in capacities appears to be incompatible with the ontology of temporal parts.
  4. We can make the problem more vivid by imagining a case where a lump and a statue always coincide. Imagine a lump of clay that comes into being shaped like Thatcher and retains that shape throughout its career without losing any parts. In that case the lump and the statue will share all their parts, temporal and otherwise (at any rate there will be things that compose both the statue and the lump). They will coincide materially. Four-dimensionalists say that in this case the lump is the statue; otherwise they would face the troubles of constitutionalism. Yet the lore of lumps and statues tells us that any lump of clay could survive being squashed, even if it never is. Otherwise their differing behavior under pressure, so to speak, would be a mystery. If this is right, then four-dimensionalism is committed to these three claims:
    … 1. The lump (in our story) could survive squashing.
    … 2. The statue could not survive squashing.
    … 3. The lump = the statue.
    And these claims look inconsistent. If the lump is the statue, how could it – that one thing – be able to survive squashing, yet also unable to survive it?
  5. The same problem arises in the case of people and animals. Most philosophers, including four-dimensionalists, want to say that if your cerebrum9 were removed from your head and kept “alive” and functioning in a vat, while the rest of you were destroyed, you would survive as a detached cerebrum10. (This seems to follow from the conviction that you would go along with your cerebrum11 if that organ were transplanted12.) But no human animal13 could survive if its cerebrum14 were preserved and the rest of it destroyed. That seems to give human people and human animals15 different modal properties: different persistence conditions.
  6. This may not look like a problem if your cerebrum16 actually is removed and kept alive. In that case you are not an animal; you are, rather (according to the temporal-parts view) composed of the pre-operation temporal parts of an animal and the post-operation temporal parts of a cerebrum17. But of course that isn’t going to happen. Your career will coincide exactly with the career of an animal, so that every temporal part of that animal is a part of you and vice versa. You and the animal coincide materially. In that case four-dimensionalists will say that you are that animal. How, then, could it be the case that you, but not the animal, could survive as a detached cerebrum18? Friends of the temporal-parts view will want to say things like these:
    … 4. You could survive as a detached cerebrum19.
    … 5. No animal could survive as a detached cerebrum20.
    … 6. You are an animal.
    And these claims are no more consistent than the others.
  7. Now many four-dimensionalists say that we are not animals, even though there are no cerebrum transplants21. Suppose that having certain mental properties is necessary for being a person; and call a more-or-less momentary temporal part of an organism that has those properties a person-stage22. Then it may be natural to suppose that a person must be something made up entirely of person-stages. (We will come to the question of which person-stages would go to make up a person in a moment.) But because every human animal23 starts out as an embryo24 with no mental properties at all, no human animal25 is composed entirely of person-stages. Alternatively, you might think that the stages of a person must all be in some way psychologically continuous with one another; but the embryonic26 stages of a human animal27 are not psychologically continuous with anything. Either way, it follows that human people, and therefore ourselves, are not animals, but at best proper temporal parts of animals: parts that include no unthinking embryonic28 stages.
  8. But it makes no difference here whether we are animals or temporal parts of animals, for if the persistence conditions of animals have nothing to do with psychology, neither do those of their temporal parts. So the view that we are temporal parts of animals appears to imply this:
    … 4. You could survive as a detached cerebrum29.
    … 7. No temporal part of an animal could survive as a detached cerebrum30.
    … 8. You are a temporal part of an animal.
    And these again look inconsistent.
  9. The general problem these cases illustrate is that things the temporal-parts view says are identical appear to have incompatible modal or dispositional properties – properties that appear to figure in the explanation of why those things cease to exist when they do. Because you are a person, you could survive as a detached cerebrum31; but because you are an animal or a temporal part of an animal, you couldn't. Call this the problem of modal incompatibility.
  10. Some four-dimensionalists take a tough line here: Quine, for instance, denies that there are any modal properties at all (1960: 199), and Heller says that all material objects have their spatio-temporal boundaries essentially (1980: 53). Either view would make at least one proposition of each triad false, or at any rate not true.
  11. But most four-dimensionalists want to accept the modal convictions that feature in the first two claims of each triad. They try to solve the problem by saying that the appearance of inconsistency is an illusion. You could survive as a detached cerebrum32, they say, and the animal or animal part that you are could not survive it; yet these claims are consistent. The reason is that the modal predicate 'could survive as a detached cerebrum33' expresses a different property in 5 and 7 from the one it expresses in 4. When we say that you could survive as a detached cerebrum34, we mean that you qua person could survive it. When we say that no animal could survive as a detached cerebrum35, we mean that no animal qua animal could survive it. So it would be more perspicuous to write 4 and 5 like this:
    … 4*. You, qua person, could survive as a detached cerebrum36.
    … 5*. No animal, qua animal, could survive as a detached cerebrum37.
    This does not express a modal difference between you and your animal body. Both you and that animal have the property of possibly surviving as a detached cerebrum38 qua person, and you both lack the property of possibly surviving as a detached cerebrum39 qua animal. 4* and 5* are therefore consistent with the claim that you are an animal. The same holds, mutatis mutandis, for the other cases.
  12. This means that that there are no absolute, unqualified modal properties of the form being possibly F or being necessarily G, but only qualified or kind-relative modal properties of the form being possibly F qua person and being necessarily G qua animal. If there were such a property as possibly surviving as a detached cerebrum40 – not possibly surviving it qua this or qua that, but just plain possibly surviving it – then we should want to say that you have that property, and that no animal or temporal part of an animal has it, thus raising the problem of modal incompatibility all over again.
  13. So what are these qualified modal properties? What do the qualifications ‘qua person’ and ‘qua animal’ mean? Many views are possible, but the most familiar is counterpart theory (Lewis 1968, 1971, 1986: 248-263, Gibbard 1975). The ordinary view of modal properties is that for a thing to be possibly F is for it to be F at some possible world. Counterpart theory says that for a thing to be possibly F is not for it to be F at a world, but for a “counterpart” of it at a world to be F at that world41. A counterpart of you at a world is a thing at that world that is at least as much like you as anything else at that world, and enough like you that we are willing to treat it as a modal stand-in for you. (The thing most like you at a certain world might be a chimpanzee. If we are not willing to consider how things would be if you were a chimpanzee – if our response to your question, "What if I were a chimpanzee?" were always, "You couldn't be a chimpanzee" – then there would be no counterpart of you in that world.)
  14. So far this is no help. If you have a counterpart at a world that survives as a detached cerebrum42, and no animal (or temporal part of an animal) has such a counterpart, then you cannot be an animal (or a temporal part of an animal). But things can be similar in different ways: something might resemble you in origin and physical composition, say, but not in psychology and behavior. And because the counterpart relation is based on similarity, whether a thing is a counterpart of you may depend on what respects of similarity are relevant in the context. So there appear to be as many different potential counterpart relations as there are respects of similarity. If we say that you are possibly F – that is, that at some world there is a counterpart of you that is F – we don't say anything determinate until we specify a particular counterpart relation – that is, a particular respect of similarity. And it needn't be the same one each time: a modal predicate can express different counterpart relations in different contexts.
  15. So when we say that you could survive as a detached cerebrum43, we might mean that you have a personal counterpart at some world – someone who resembles you in "personal" respects – who survives as a detached cerebrum44 at that world. Now whatever else a personal counterpart of something might be, it ought to be a person. Suppose a person is by definition a being composed of person-stages, each of which is psychologically continuous, in some appropriate way that is open to debate, with every other and with nothing else: a maximal psychological continuer for short45. Then any being that fails to go along with its transplanted cerebrum46 is not a maximal psychological continuer, and therefore not a person, and so not a personal counterpart of you or of anyone else. Thus, every personal counterpart of you whose cerebrum47 is removed and kept alive will go along with that organ. Presumably there are, at some possible worlds, personal counterparts of you whose cerebrums48 are removed and who go along with those organs. (At any rate this is so on the ontology of temporal parts.) And the personal pronoun in the sentence ‘You could survive as a detached cerebrum’49 makes the personal counterpart relation the relevant one. That makes it true to say that you could survive as a detached cerebrum50.
  16. When we say that no animal could survive as a detached cerebrum51, on the other hand, we might mean that no animal has an animal counterpart at any world – something that resembles it in "animal" respects – that survives as a detached cerebrum52 at that world. An animal counterpart of something must presumably be an animal itself. An animal, according to four-dimensionalism, is a thing composed of animal stages related to one another and to nothing else in a certain biological way that I won’t attempt to specify. And because a detached-cerebrum53 stage is not an animal stage, nothing made up partly of detached-cerebrum54 stages is an animal. Thus, no animal, and no animal counterpart of anything, survives as a detached cerebrum55 at any possible world. So nothing has an animal counterpart that survives as a detached cerebrum56. The word 'animal' in the original sentence makes the animal counterpart relation the relevant one. That is why it is true to say that no animal could survive as a detached cerebrum57.
  17. So the claim that you could survive as a detached cerebrum58 but no animal (or temporal part of an animal) could survive it, understood counterpart-theoretically, is compatible with your being an animal (or a temporal part of an animal). Similar remarks apply in other cases: the claim that any lump could survive crushing but no statue could, understood counterpart-theoretically, is compatible with the claim that some statues are lumps.
  18. Whether counterpart theory is the right account of modal properties is disputed, and I will not enter into this dispute here. The point is that friends of the temporal-parts view must either reject the modal claims about people and animals that we have discussed, or else understand them counterpart-theoretically (or in some other way according to which modal predicates express different properties depending on their context). Some philosophers find this a sufficient reason to reject the ontology of temporal parts (van Inwagen 1990b).
  19. Even if we accept counterpart theory, however – or for that matter if we follow Quine and deny that there are any modal properties at all – the question remains: why do statues but not lumps cease to exist when squashed? Four-dimensionalists deny that there is any modal or dispositional difference between a statue and its coincident lump: the lump, they say, hasn’t got a certain capacity to survive that the statue lacks. So things’ modal properties play no role in explaining why they cease to exist when they do. What does it explain it, then? What accounts for the striking difference in the actual behavior of lumps and statues when we squash them?
  20. Four-dimensionalists reply that there is no explanation: the question is based on a misunderstanding of the nature of persisting objects. For every period of time when a lump of clay exists, they say, there is a temporal part of the lump located exactly then. It follows from this that there is a temporal part of our imaginary lump that extends from the time when it is made Thatcher-shaped to the time when it gets squashed. That is the thing we call the statue of Thatcher. Why does that object cease to exist when it gets squashed? What stops it from carrying on for a bit longer and simply changing its shape? The four-dimensionalist’s answer is that nothing stops it. That is simply where its temporal boundary lies. To ask why it comes to an end when it does is like asking why a given period of time comes to an end when it does. Why does the 20th century come to an end at the end of 1999? What stops it from going on for a bit longer? The question is based on a misunderstanding of the nature of periods of time. The only answer we can give is that we wouldn’t call a period of time that extended beyond 1999 the 20th century. But that is an answer to a different question: a question not about why a particular thing ceases to exist when it does, but about why we talk the way we do. It is the same with the lump and the statue. There are temporal parts of the lump that extend beyond the time when we squash it. But we don’t call those things statues. And that is all the answer we can give to the question of why the statue doesn’t continue to exist after we squash it. (That does not mean, however, that nothing could have lasted longer than it did. Descartes could have lived longer, for there are personal counterparts of him that do live longer.)
  21. This is one of the aspects of the temporal-parts ontology that is most foreign to our ordinary thinking. Foreign or not, however, four-dimensionalism is a powerful theory of the nature of persisting objects.
5.5 Puzzles of personal identity
  1. We have seen that the ontology of temporal parts offers an important solution to metaphysical puzzles about the persistence of material things, including ourselves. And we saw that this solution comes at the cost of requiring modal predicates to be inconstant: it implies that in a sense you could survive as a detached cerebrum59, but in another perfectly good sense you couldn’t possibly exist in that state. The view that we are composed of temporal parts – the temporal-parts view – has further important implications about our identity over time.
  2. Recall the “brain-state transfer” machine of §1.7, which records the psychological information encoded in your brain (thereby erasing that organ) and copies it onto another brain (thereby erasing that organ’s previous contents). Some philosophers believe that this process would literally move you from one human organism to another. But we may wonder what sort of thing could move from one animal to another in this way. What sort of thing could be sent as a message by telegraph? Not a material thing, surely.
  3. On the ontology of temporal parts, however, a material thing can move, in a sense, from one human animal60 to another via brain-state transfer. Perhaps no material thing would move from one place to another in the sense of passing through all the points in between; but there would be a material thing that was first in the one place and later in the other. There would be a material thing made up of the temporal parts of the “donor” animal located before the adventure and those of the “recipient” animal located afterwards. That thing would be rational and intelligent in the same way as you are – supposing, at least, that you are material – for it would share your rational, intelligent stages. So the machine would bring it about that a rational, intelligent being is first in one place, associated with one human organism, and later in another place, associated with another organism. Those who believe that we can move from one animal to another via brain-state transfer will say that this being would be you. Thus, the temporal parts view is compatible with the claim that you could swap animal bodies via brain-state transfer.
  4. But the temporal-parts view is also compatible with the claim that you couldn’t swap bodies via brain-state transfer. The animal that ends up with your mental contents in the story does not inherit them from you in anything like the usual way. If you did swap bodies, your mental capacities would not be continuously physically realized. This suggests that you would have a gap in your career: you would cease to exist while the machine does its work, then come into being once more when the work is finished. (Suppose the machine that records the psychological information from the first brain then prints it on sheets of paper; much later the information is typed by hand into a second machine, which then reconfigures the second brain to match the first one. In what form could you exist after the first brain is erased but before the second brain is reconfigured? What sort of thing could you be then?) And you might deny that any material thing could exist intermittently. The temporal-parts view can accommodate this conviction.
  5. The ontology of temporal parts provides three main candidates for being you in this case. First there is the one that stays behind: the human animal61 whose mental contents are erased when the machine does its work and then lives on as a sort of human vegetable. Second, there is the one that ceases to exist: the temporal part of the first animal that extends from its beginning (or from the point where it first acquires mental capacities of the right sort) until its mental contents are erased. Finally, there is the one that gets transferred: the thing composed of the pre-transfer stages of the first animal and the post-transfer stages of the second (or at least those stages of them that have the right mental properties). On the usual ontology of temporal parts, these three beings all exist. The question of what happens to you in the brain-state-transfer cases is the question of which of them is you.
  6. Friends of the temporal-parts view say that this depends on which of them counts as a person, for you are the only person in the story. Now all three candidates are rational, intelligent, and self-conscious, at least for a time, and the ordinary view is that being rational, intelligent, and self-conscious suffices for being a person. (Don’t we call a rational, intelligent, self-conscious being a someone and not a mere something?) But friends of the temporal-parts view deny this. Otherwise there would be far too many people in the world: not only would all three candidates in the brain-state-transfer story count as people, but so would most of your temporal parts. (Remember that on the ontology of temporal parts a being is intelligent at a time by virtue of having a temporal part located then that is intelligent without temporal qualification.) A person, they say, must consist of stages that are not only rational, intelligent, and self-conscious, but which also relate to one another in the right way.
  7. There is room for disagreement about what this way is: whether it is non-branching, continuously physically realized psychological continuity62, non-branching psychological continuity63 of any sort, some wholly non-psychological relation, or what have you. This is the old debate about personal identity over time – the persistence question of §1.6 – transposed into the ontology of temporal parts. Those who say that personal identity over time consists in non-branching, continuously physically realized psychological continuity64 will call the candidate that ceases to exist when the machine does its work the person. Those who say that any sort of psychological continuity65 suffices will call the one that gets transferred the person, while those who think it is something wholly non-psychological will probably say that the candidate that stays behind is the person.
  8. Who is right? They might all be right. According to four-dimensionalism all three candidates exist, and all three satisfy some reasonable definition of the word ‘person’. Those who appear to disagree about which candidate is the person may only be using the word ‘person’ in different senses. More generally, those who appear to disagree about what it takes for a person to persist through time may all be right. They may not disagree at all. They may only be talking about different beings – about “people” in different senses of the word. There are people in the sense of beings composed of person-stages related to one another, and to nothing else, by non-branching, continuously physically realized psychological continuity66 – “conservative people”, we might call them – people in the sense of beings composed of person-stages related to one another by psychological continuity67 of any sort – “liberal people” – and so on. Different accounts of personal identity over time are accounts of the identity over time of different sorts of people. Friends of the conservative psychological-continuity view have given the right account of the persistence of conservative people; those who prefer the liberal psychological-continuity view have given the right account of the persistence of liberal people. For that matter, those who say that our identity is animal identity have given the right account of the persistence of “animal people”. Any proposed account of what it takes for us to persist through time will presumably be the right account of the persistence of some sort of beings that are plausible candidates for being people. Or almost any account. The view that you persist if and only if your immaterial soul continues to exist will not be the right account of the persistence of anything unless there are immaterial souls, and four-dimensionalism in no way implies that there are.
  9. This would vindicate the common view that the facts about personal identity over time are in some sense up to us to decide: they are up to us to decide insofar as it is up to us which sort of “people” to talk about (Olson 1997b, Sider 2001b). No other view of what we are68 (with the possible exception of the bundle view, some versions of which incorporate the ontology of temporal parts) has this consequence. Whether this speaks in favor of the temporal-parts view or against it will be a matter for debate.
  10. Here is one more point about our identity over time. The temporal-parts view is the only account of what we are69 that is consistent with the attractive idea that some sort of psychological continuity70 suffices for us to persist. It seems possible for a person to be psychologically continuous, by anyone’s lights, with two future people: let each of your cerebral hemispheres be transplanted71 into a different head. It is easy to believe that you could survive if one hemisphere were destroyed and the other were transplanted72, so that the operation produced only one being psychologically continuous with you. But it seems impossible for both offshoots to be you, for there are two of them and only one of you, and two things cannot be numerically identical with one thing. (Suppose one of the offshoots had a beard and the other didn’t. If both were you, you would both have a beard and not have a beard at the same time, which most of us regard as a contradiction.) And the claim that just one of the two offshoots would be you would make it a mystery which one it was, and why. This leads many philosophers to say that you could survive a single transplant73 but not a double transplant74: if two beings each got half your brain, that would be the end of you. It follows that psychological continuity75 is not sufficient for you to persist: you may perish even though a later person is psychologically continuous with you. Only "non-branching" psychological continuity76 suffices. But it is hard to believe that you could fail to survive the transplant77 operation merely because it produced two beings psychologically continuous with you rather than one.
  11. Friends of the temporal-parts view are able to say that you can survive the double transplant78 (Lewis 1976). You survive, so to speak, as both offshoots. That is because in fission cases there are, in effect, two of you all along, who share their pre-operative temporal parts but not their post-operative parts. One of them starts when you do and then goes along with your left cerebral hemisphere in the operation; the other has the same beginning, but goes along with your right hemisphere. But they do not coincide materially; they merely overlap, like two railway lines that diverge after having a section of track in common. Thus, as long as someone is psychologically continuous with you in the right way, you survive. It doesn't matter if more than one is.
  12. This has the surprising consequence that if there is fission in your future, there are two people sitting there and thinking your thoughts even now – though they look for all the world, even to themselves, like one. In order to know whether they are two or one – or more generally, to know how many people there are at any one time – we need to know what the future holds. (This is not a case of backwards causation79, however. The fission operation doesn’t cause it to be the case that there are two people there earlier. It simply makes it the case, in a non-causal way, that the pre-fission stages are parts of two people rather than one.)
  13. The story the temporal-parts view tells about fission is surprising in other ways too. You might want to know which person you are. Suppose the one who gets your left hemisphere – "Lefty" – will be in pain after the operation, while the one who gets your right hemisphere, "Righty", will be comfortable. Will you be in pain? The temporal-parts view suggests that when we say 'you', we refer to that person whose stage we are addressing. If that stage is a part of more than one person, as the pre-operative stages in the fission story are, we refer ambiguously to both. So if there is fission in your future, we refer at once to two different people when we say 'you' (supposing we refer at all, anyway). Our sentence 'You will be in pain' says two things at once, one true and one false. It is like saying that the planet between the earth and the sun is cloudy. One of you will be in pain, the other not. But we cannot say that either of those people is you. To say that you are Lefty is at most half right, since only one of you, as it were, is Lefty; the same goes for the claim that you are Righty. Nor can we say that either person is not you. So we can know what happens to all of the people in the story, and yet not know what will happen to you. The question we so urgently want answered – will you be in pain? – is unanswerable, based as it is on the false presupposition that the pronoun ‘you’ refers to only one being, or at any rate that all of the beings it refers to share the same fate80.
  14. Friends of the temporal-parts view will point out that we say similar things about analogous spatial cases, such as railway lines that share their tracks at one place and diverge at another. How many lines there are here depends on what is the case elsewhere. And simple questions such as whether this line goes to Tulsa may have no straightforward answer because the expression ‘this line’ refers ambiguously to two lines, one of which goes to Tulsa and one of which doesn’t. The logic of the case, we might say, is impeccable. But that may not make the story any easier to believe. It may only be a reason to deny that fission is like spatial branching – that is, to deny that we have temporal parts.
  15. And although the proposed view implies that psychological continuity81 is sufficient for our identity over time in the sense that you persist if someone in the future is psychologically continuous with you, it implies that psychological continuity82 is not sufficient for our identity over time in another sense: not every future being who is mentally continuous with you is you (Parfit 1976; see also Lewis 1983). Lefty is not psychologically continuous, before the operation, with Righty as he is afterwards. You can be psychologically continuous with someone other than yourself.

Comment:
5.6 Thinking animals83 and other worries
  1. We saw that the constitution view84 faces a problem about those human animals85 that think our thoughts and perform our actions but which, according to those views, are not ourselves. A similar problem arises for the temporal-parts view. Most four-dimensionalists say that we are not animals, but maximal psychological continuers: beings composed of person-stages, each of which is in some way psychologically continuous with each of the others and with no other stages (§5.4). Since all human animals86 have embryonic87 stages that are not in any way psychologically continuous with anything (and some have senile stages not psychologically continuous with anything), this means that we are at best temporal parts of animals. Yet four-dimensionalists cannot deny that the animals we are parts of think. They say that for a persisting thing to think at a time is for it to have a temporal part located at that time that thinks; and since your thinking stages are all temporal parts of a human animal88, the animal thinks just as you do. So according to the temporal-parts view you share your thoughts with an animal numerically different from yourself. And you ought to wonder, it seems, how you could ever know that you are not the animal.
  2. Four-dimensionalists could avoid the consequence that our bodies are thinking animals89 other than ourselves by saying that we are animals. That would mean giving up the idea that any sort of psychological continuity90 is necessary for us to persist. (They could still say that it is sufficient for us to persist by appealing to counterpart theory.) But it wouldn’t help, for on their view your current stage is a temporal part of a thinking maximal psychological continuer as well as a temporal part of a thinking animal91. So you would still face the problem of how you could know whether you are the animal or the maximal psychological continuer. That is why most animalists92 deny the existence of maximal psychological continuers, and thus reject four-dimensionalism generally.
  3. In fact four-dimensionalism implies that there are all sorts of beings now thinking your thoughts. Your current stage is a temporal part not only of a maximal psychological continuer and an animal, but also of the first or second half of that animal. And it is a part of a vast number of gerrymandered and badly behaved objects: the thing made up of the temporal parts of your animal body located before midnight tomorrow and the temporal parts of Kilimanjaro located thereafter, for instance, and the thing made up of the temporal parts of the moon located before last Friday, the temporal parts of my left ear located after tomorrow, and the temporal parts of your animal body located in between. All of these objects, by virtue of sharing your current thinking stage, now think in the same sense as you do. So there are far more beings sitting in your chair and thinking about philosophy than we thought93.
  4. How could you ever know which of these beings you are? It is bad enough not knowing whether you are the animal or the maximal psychological continuer. Now it seems you ought to wonder whether you might be the first or second half of the animal, or one of the countless gerrymandered objects that share your current stage. What grounds could you possibly have for accepting any of these alternatives? If you really had no idea which thinker you are, you would have no idea what your future holds. The fact that there will be a human person tomorrow who is psychologically continuous with you as you are today would be no reason at all for you to believe that you will be a human person tomorrow. As far as your evidence in the matter would go, you could just as well be an enormous volcano tomorrow.
  5. Faced with this absurd prospect, friends of the temporal-parts view appear to have little other choice than to embrace personal-pronoun revisionism, the linguistic hypothesis discussed in §2.7. The idea is that your current temporal part is a part of only one person. There is room for disagreement about just what counts as a person, but any reasonable definition of 'person' will imply that there is just one being of that sort now thinking your thoughts: just one maximal psychological continuer, for instance.
  6. Or at least this is nearly true. The usual temporal-parts ontology implies that for every person (in any ordinary sense of the term), there is another being just like it but a nanosecond longer or shorter – a being entirely indiscernible from the first by any practical means, and just as good a candidate for being a person. So in reality there is not one person there, but a large class of more or less indiscernible and mostly overlapping people. This is a version of what Unger (1980) calls the problem of the many. It means that it is probably indeterminate which of those beings you are. This is of course rather implausible. Four-dimensionalists point out that we should “count those people as one” for ordinary purposes: when all the people in a certain situation overlap and differ from one another only trivially, those of us who are not engaged in metaphysics describe this by saying that there is just one person there (Lewis 1976, 1993). And perhaps it shouldn’t worry us too much if we cannot know which of these beings you are, or if all ordinary ways of referring to you are ways of referring to all of them ambiguously, since we can’t tell them apart anyway. For that matter, the problem of the many is not obviously unique to the temporal-parts view: anyone who thinks that we are material things will have to say something about those beings, if there are such, that are just like ourselves but larger or smaller by a single particle.
  7. But the important point here has nothing to do with the problem of the many. It has to do not with the surplus of good candidates for being you, but with the surplus of bad candidates: the human animal94 (if you are a maximal psychological continuer), the animal’s first or second half, and all the arbitrary and gerrymandered objects that share your current stage but diverge wildly at other times. Pronoun revisionism says, first, that these beings are not people. Second, personal pronouns, first-person thoughts, and other “personal” referring expressions denote only people. Third: as a competent speaker of English, you know that your first-person thoughts and utterances must refer to a person if they refer at all. So if you know what it is to be a person, you can work out (subject to the caveat about “the many”) which of the beings that share your current stage you are.
  8. This proposal has the repugnant consequence that the vast majority of rational, intelligent speakers are not people. And they cannot refer to themselves in the first person – not, anyway, in the language they actually speak. Their language also prevents them from having first-person thoughts about themselves. That is a strange sort of disability. If the ontology of temporal parts is true, however, and if most of what we say in ordinary life about ourselves and others is right – if it really is true to say that you will be a human person tomorrow and that you were never a satellite of the earth – then it seems that some such linguistic hypothesis must be correct. We may not like it much; but then the alternatives might not be very nice either.
5.7 Thinking stages
  1. I come now to what is perhaps the most serious objection to the temporal-parts view. Recall the problem of temporary intrinsics (§5.2): how can a persisting thing have incompatible properties, such as sitting and standing? Four-dimensionalists say that for a thing to sit at a time is for it to have a temporal part located at that time that has the property of sitting without temporal qualification. I sit now only insofar as my current stage sits. Assuming that I persist, I don’t strictly have the property of sitting at all, but rather the property of having a sitting part. And what goes for sitting goes for temporary properties generally95. Given that persisting things change continuously, it seems that the things that strictly have ordinary properties must literally be momentary: only a momentary being could have a shape without temporal qualification, and any material thing with ordinary properties has a shape.
  2. This means that although certain momentary temporal parts of me have such familiar properties as sitting, sleeping, writing, weighing 150 pounds, and being conscious, I myself have none of those properties. I don’t have the property of sitting, or of sleeping, or of writing. Strictly speaking I have no shape or weight. I have no mental properties, or at any rate none such as being conscious or thinking about Vienna. I have no temporary properties of any sort. If I did, I should have them relative to times, and four-dimensionalists agree that things do not have properties relative to times. I am not even perceived, except perhaps by Berkeley’s God. My stages may be perceived, but if I had the property of being perceived I should also have the property of not being perceived, for being perceived is a temporary property. This is all rather troubling. The temporal-parts view implies that the familiar objects that think and act and are seen and heard are our stages, not ourselves. We ourselves are unobservable theoretical entities.
  3. Four-dimensionalists will reply that we still have these familiar properties in a certain sense. We often attribute to an object the properties of its parts: we say that someone is sunburned when in the strictest sense only a part of her nose is sunburned, or that the day was wet when really only most of the afternoon was. And these descriptions are perfectly correct. So even though I don’t bear a property such as sitting in the strict sense in which my current stage bears it, I nonetheless relate to that property in a way that makes it true to say that I am sitting. That is, I have that property in what Butler (1975) might have called “a loose and popular sense”. The temporal-parts view does not deny that I sit or think in any sense. It is compatible with the way we ascribe ordinary properties to ourselves in ordinary situations.
  4. Of course, more needs to be said, for it isn’t generally right to attribute to an object the properties of its parts. My left foot is a part of me. It weighs about three pounds and is entirely unconscious. Yet it would be absurd to describe this fact by saying that I am foot-shaped, weigh three pounds, and am unconscious. Why is it right to say that I sit by virtue of having a sitting temporal part but wrong to say that I am foot-shaped by virtue of having a foot-shaped spatial part?
  5. Four-dimensionalists will presumably concede that I have the properties of being foot-shaped and being unconscious in the same way as I have the properties of sitting and thinking: my relation to all those properties is the same. They will explain why we say that I am sitting but not that I am foot-shaped by appealing to our interests and expectations. Those interests and expectations make it useful, in ordinary circumstances, to describe me as sitting or thinking when my current temporal part has the property of sitting or thinking, whereas it would not be useful – in fact it would be positively misleading – to describe me as foot-shaped or unconscious, even though I have parts with those properties too. There may be possible circumstances in which it would be right to attribute to me the shape or the mental properties of my left foot, but as things are it isn’t. Just why this should be so is a nice question, but I don’t doubt that it has an answer.
  6. This is all fine and good, but it misses the point. Let us grant that nothing the temporal parts view says conflicts with the way we ascribe ordinary properties to ourselves in ordinary situations. Suppose it is true to say, when we are not doing metaphysics, that we sit and think and are conscious, even if the real bearers of those properties are not ourselves but brief temporal parts of us. Still, don’t we have a deep conviction that we are among the real bearers of such properties as thinking and being conscious? Doesn’t it seem evident not only that you and I think and are conscious in some sense or other, but that we think and are conscious in the strictest possible sense? Of all the things we know about ourselves, isn’t this the most certain? Surely we cannot suppose that we think only in virtue of the fact that some other thing thinks for us. As Chisholm said, if there are two beings thinking these thoughts, one thinking them on its own and the other having its thinking done for it by the first, then I am the first thing and not the second.
  7. Someone is bound to reply that this “thinking-stage problem” is no different in principle from the thinking-brain problem. Nearly everyone believes that we are animals, or at any rate material things the size of animals (things constituted by animals, perhaps). And it follows from this belief (the reply goes) that we think only in the sense of having spatial parts – brains – that think in the strictest sense. So if there is a problem here, it has nothing to do with the temporal-parts view in particular. It is one that nearly everyone shares. Moreover, no one worries about the thinking-brain problem. So why worry about the thinking-stage problem?
  8. Now I think we ought to worry about the thinking-brain problem. I worried about it at some length in the previous chapter. If it really were true that only our brains think in the strictest sense, then in my view we ought to conclude that we are brains.
  9. In fact the thinking-stage problem is more serious than the thinking-brain problem. There are three possible ways of solving the thinking-brain problem. (I don’t regard saying that we “think” only in the sense of having thinking brains as parts to be a solution.) One is to deny that there are such things as brains. An analogous solution to the thinking-stage problem would say that there are no such things as person-stages, which is incompatible with the temporal-parts view. The second and most popular is to deny that only our brains think in the strictest sense. An analogous solution to the thinking-stage problem would be to deny that our stages think in the strictest sense. That too is incompatible with the temporal-parts view. So two of the three possible solutions to the thinking-brain problem are unavailable in the case of the thinking-stage problem. The third is to say that we are brains. Four-dimensionalists could say, analogously, that we are stages. That would solve the problem, though at a considerable cost: that we are momentary stages is no easier to believe than that we are brains. It is, however, an important variant of the temporal-parts view, and deserves a section of its own.
5.8 The stage view96
  1. According to the temporal-parts view it is stages that think and act strictly speaking; yet we are not stages. That is hard to believe. It is also hard to believe that the things that bark and wag their tails and chase postmen are not dogs, but mere parts of dogs, dog-stages. One way to avoid this is to reject the ontology of temporal parts altogether. But another is to say that people and dogs are stages. Call this the stage view97.
  2. If there is any reason to suppose that we are momentary things that do not strictly persist, this is it. There are powerful arguments in support of the ontology of temporal parts, and anyone who holds that view must either say that we are stages or deny that we think or talk or have any other temporary property in the strictest sense. The stage view98 has other virtues too. It avoids the thinking-animal problem by implying that our animal bodies don’t think in the strictest sense. The only true thinker of my current thoughts, it says, is my current stage, which is me. It also gives us the right number of objects existing at any one time (setting aside the problem of the many): it implies that there is just one person sitting here now, even if there is fission in my future, just as we thought – whereas on the temporal-parts view there are at least two (Sider 2001: 188-190).
  3. But although the stage view99 has important advantages over the temporal-parts view, it is rather hard to believe. Most obviously, it implies that we don’t persist through time. This means that you are not the person who began reading this sentence. That was someone else – someone very similar to you, of course, and strongly causally connected to you, but a numerically different being all the same. We are all far younger than we thought. We were never children. We have no past and no future. For that matter, we can’t move, or change. What appears to be a persisting, changing person (or dog or what have you) is in reality only a series of static momentary beings.
  4. Defenders of the stage view100 – and it has its defenders (Sider 1996, 2001: 188-208; Hawley 2001) – are surprisingly easy about this. They respond, as one might have come to expect by now, by insisting that their view is compatible with everything we ordinarily say and believe and care about. Even if the stage view101 is true, they say, certain past and future stages relate to me in ways that lead us, for reasons to do with our interests, to call them by my name. Sider calls such stages “temporal counterparts” of me (personal temporal counterparts, to be more precise). He says that the ordinary belief that I was once a boy is true because I – the current momentary bearer of my name – have as a personal temporal counterpart an earlier boy-stage, which is compatible with the stage view102.
  5. This means that when we say that I was once a boy, we are not asserting or implying that anything persists through time. We are not asserting the numerical identity of any earlier thing with any later thing. When we say such things we are only speaking as if something persisted because that is a convenient way to talk. (You can imagine how inconvenient it would be if we had to call every new momentary object by a different name.) We are doing what Hume called “feigning a continu’d being” (1978: 208). Saying that I was once a boy is rather like saying that the Prime Minister is a man today but was a woman twenty years ago. When we say this we don’t ordinarily mean to imply that anyone has changed sex. We’re not saying that something that is a man today and something that was a woman twenty years ago are one and the same. Rather, we are saying that a man who exists today and a woman who existed twenty years ago relate in some other way: that the woman then held and the man now holds the office of Prime Minister. If we are using the language of identity over time here, we are using it loosely.
  6. When a currently existing stage relates to earlier stages in a way that leads us to speak of them as if they were one – that is, when they are temporal counterparts – we might describe this by saying that the current stage has persisted in the same sort of loose sense. Although the stage view103 denies that people and dogs really persist – that they persist “in the strict philosophical sense of the word”, as Butler would say – it is compatible with their “persisting” in the loose and popular sense in which the Prime Minister has persisted for some two centuries. Stage theorists claim that this “loose and popular persistence” is the only sort of persistence that most of us think or care about. Our ordinary, non-metaphysical thought and talk about identity over time is concerned only with the having of earlier and later temporal counterparts. It is facts about temporal counterparts, not facts about ourselves at other times, that make our ordinary statements and beliefs about our persistence true or false. Our practical attitudes are likewise based not on strict identity, but on the personal temporal-counterpart relation: we hold people responsible for the actions of their earlier temporal counterparts, and each of us has a special, selfish concern for the well-being of her later temporal counterparts (we encountered a view like this in §2.8). So according to the stage view104 we do persist, in the only sense that matters. Whether things persist in the strict philosophical sense, stage theorists say, is of interest only to metaphysicians. Thus, only a metaphysician will object to the stage view105 on the grounds that it rules out our persisting through time. To anyone else this is a mere technical detail.
  7. This story is not easy to believe. The stage view106 certainly appears to conflict with things we ordinarily believe and say. When we say, in ordinary contexts, that I was once a boy, we seem to be asserting that I myself once had the property of being a boy, and not that I relate in a certain way to some other being that had that property. Even if the story were true, though, it would not entirely silence the complaint. The fact that most of us find the stage view107 more or less incredible shows that it contradicts something that most of us believe. I, for one, believe that I persist in the strict philosophical sense, and not merely that certain past beings are temporal counterparts of me, even if the existence of such beings makes it correct for ordinary, non-metaphysical purposes to describe me loosely as having existed in the past. I don’t believe that this conviction of mine is eccentric, or that it came about only as a result of my philosophical training. It may be that it is a metaphysical conviction, and that our ordinary, non-philosophical beliefs do not imply it. But it is no less widely or deeply held for all that.
  8. In any case, the stage view108 has further troubling consequences. For one, it is incompatible with any plausible view of what it takes for us to persist through time – that is, of personal identity over time. It conflicts, for instance, with the view that our identity over time consists in some sort of psychological continuity109. In fact it implies that no sort of continuity at all, psychological or otherwise, suffices for us to persist. Of course, it may still suffice for us to persist in a loose and popular sense: the fact that some future person inherits her mental and physical properties from you in some appropriate way might make her a personal temporal counterpart of you. But it does not suffice for her to be you.
  9. Someone might find this consequence not merely implausible, but incoherent. No momentary being, you might think, could count as a person: anyone who believes, or seems to believe, that people never persist through time has just not got the concept of a person, just as someone who believed, or seemed to believe, that cats are fuzzy toys would not have the concept of a cat. But if it belongs to the concept of a person that people ordinarily persist in the strict philosophical sense, then people cannot be stages. If it also belongs to the concept of a person that people have the property of thinking, then the ontology of temporal parts in general is inconsistent with the existence of people.
  10. If the stage view110 is compatible with the existence of people, however, it gives us far too many people. You might have thought that this book has only one author. (That's what it says on the cover.) Not so, according to the stage view111: even if only one philosopher was at work on it at any one time, a vast number of momentary philosophers successively took over the job – though none stayed at it long enough to write even a single word.
  11. Finally, the stage view112 implies that none of the people who exist are the ones we know and love. Consider Socrates. What stage is he? According to the stage view113, plenty of stages successively bore the name 'Socrates'. But nothing could make it the case that just one of those stages, rather than another, was Socrates. We couldn't discover that Socrates – the teacher of Plato and the wisest man in Athens – existed for only an instant during the evening of the 6th of August, 417 BC. No stage could be Socrates. If Socrates can only be a stage, then there can be no such thing as Socrates. The same goes for the rest of us. What stage could I be? You might say that I am the current bearer of my name: the current "Olson-stage". (Set aside the inconvenient fact that we cannot refer uniquely to any one momentary stage.) But why should I be that stage? Choosing it would be just as arbitrary as choosing a moment during an August evening as the instant when Socrates existed.
  12. There are many clever things that stage theorists can say in response to these complaints. They can say, for instance, that because all the authors of this book are personal temporal counterparts of one another, there is a perfectly good sense in which only one person wrote it. And they can say that because there are many successive bearers, of the right sort, of the name ‘Socrates’, it is true to say that Socrates existed, even if none of those things is Socrates. That is, they can reply to these complaints in much the same way as they reply to the objection that the stage theory114 violates our conviction that we persist through time. And the objections can be sharpened, as before, in a way that makes them immune to those replies, if perhaps less forceful.
  13. Whatever the outcome of this debate may be, four-dimensionalists must choose between the temporal-parts view, according to which we persist but don’t strictly think or act, and the stage view115, according to which we think and act but don’t strictly persist. It would be nice if we could avoid this dilemma.




In-Page Footnotes

Footnote 4: The most comprehensive exposition of the ontology of temporal parts is Sider 2001. Heller 1990: Ch. 1 and Hawley 2001 are also recommended.

Footnote 5: I say more about what a temporal part is in Olson 2006c.

Footnote 8: There is more than one alternative to the four-dimensionalist’s account of change, but the one I have sketched here seems the most plausible. Haslanger 2003 is a useful guide to this messy debate.

Footnote 22: This assumes that a more-or-less momentary object can have the sorts of mental properties characteristic of people: self-consciousness, for instance. And that may be doubtful. Four-dimensionalists generally say that a momentary thing can have mental properties if it has the right causal antecedents (Sider 2001: 197f.). What something does for a moment might count as thinking in part because of what other beings do at earlier (and perhaps later) times, in something like the way that an event counts as the turning point in a war because of what happens earlier and later. Given what we said earlier about temporary intrinsics, four-dimensionalism appears to require this. It implies that thinking is an extrinsic property. The subject of momentary thinkers will come up again at the end of §7.4.

Footnote 41: For the sake of simplicity I am here indulging in the fiction that there are things that exist only in other possible worlds: things that don’t actually exist but might have existed. Although Lewis believes that this fiction is true, counterpart theory can be formulated in a way that is free of this implication (Stalnaker 1986).

Footnote 45: This has the interesting implication that personhood is an extrinsic property: whether a thing made up of appropriately connected person-stages is a person depends on whether it connects in that way to any other person-stages. If personhood were intrinsic, the temporal-parts view would have the absurd consequence that every connected temporal part of a person is a person.

Footnote 80: Blackburn (1997: 181) makes a similar complaint. The semantic theory developed in Perry 1972 is intended to avoid this consequence.

Footnote 93: At any rate all four-dimensionalists that I know of believe in these objects. They accept universal composition for momentary stages: for any momentary temporal parts of any objects whatever, there is an object that those things compose. One might be able to accept an ontology of temporal parts without holding this: perhaps one could say that only some momentary stages compose something, while others don’t. But that would prevent one from saying many of the things four-dimensionalists want to say: see for instance §5.4 and Sider 2001: 120-139.

Footnote 95: The property of being 43 years old might seem to be an exception: I don’t have it now by having a temporal part located now that is 43 years old. Most four-dimensionalists will deny that there is any such property. Otherwise things would have to have it relative to times, and four-dimensionalists generally hold that all things have all their properties without temporal qualification. They will say that the sentence ‘Olson is 43 years old’, uttered at a time t, expresses a true proposition if and only if Olson is born 43 years before t. So I don’t temporarily have the property being 43 years old; I timelessly have the property being born 43 years before t.


Text Colour Conventions (see disclaimer)

  1. Blue: Text by me; © Theo Todman, 2018
  2. Mauve: Text by correspondent(s) or other author(s); © the author(s)



© Theo Todman, June 2007 - July 2018. Please address any comments on this page to theo@theotodman.com. File output:
Website Maintenance Dashboard
Return to Top of this Page Return to Theo Todman's Philosophy Page Return to Theo Todman's Home Page