What Are We? What Now?
Olson (Eric)
Source: What Are We? A Study in Personal Ontology, Chapter 9 (November 2007: Oxford University Press.)
Paper - Abstract

Paper StatisticsNotes Citing this PaperColour-ConventionsDisclaimer


(OO-L): This chapter proposes that animalism1, the temporal-parts view, and nihilism2 are the best accounts of what we are3. It then takes up metaphysical objections to animalism4 hinted at earlier. It is proposed that animalists5 answer them by endorsing a sparse ontology of material objects. It is then argued that we can work out what we are6 by discovering when composition occurs: if composition is universal, we are temporal parts of animals; if there is no composition, we do not exist; and intermediate theories of composition lead almost inevitably to animalism7. Finally, the view that there is no theory of composition – that composition is brute – is claimed to rule out any good account of what we are8.

  1. Some results
  2. Some opinions
  3. Animalism9 and the thinking-parts problem
  4. Animalism10 and the clay-modelling11 puzzle
  5. Theories of composition
  6. Composition and what we are12
  7. Brutal composition
Click here for Note

Full Text

9.1 Some results
  1. That completes our survey of answers to our question. I have tried to discuss the most interesting and important accounts of what we are13. There are almost certainly other views that I ought to have considered, and views I have passed over too quickly. But I hope I have at least made a good start. What have we learned? Let me try to summarize our main conclusions up to now.
  2. We began with animalism14. If there is a human animal15 located where you are, and it thinks just as you do, it is hard to see how you could be anything other than that animal, or how you could ever know that you are. This “thinking-animal problem” is not only an argument for animalism16, but also a challenge for any other account of what we are17: how can we know we are not the animals that think our thoughts? The most common objection to animalism18 is that it conflicts with popular claims about personal identity: that our identity over time has something to do with psychology, that we have certain mental properties essentially, and that facts about how mental states relate to one another determine how many of us there are at any one time: the psychological individuation19 principle. Animalists20 must reject these claims. They are not alone in this, however: the third claim appears to conflict with almost any account of what we are21.
  3. The constitution view22 says that we are material things constituted by organisms. It is an instance of constitutionalism, the general claim that qualitatively different material objects can be made of the same matter at once. Constitutionalism purports to solve a number of hard metaphysical problems, such as the clay-modelling23 puzzle and the paradox of increase. And if constitutionalism in general is true, it is likely that we are non-animals constituted by animals. Constitutionalists will want to avoid the thinking-animal problem by denying that the animals constituting us are just like us mentally; but it will be hard for them to explain why physically identical beings in identical surroundings should differ mentally. I also complained that the view rules out any good account of when constitution occurs or of what determines our boundaries.
  4. The view that we are brains gets its support from the idea that only brains think in the strictest sense: bigger things can “think” only by having a thinking part. That would solve the thinking-animal problem: we could know that we are not animals because we obviously think in the strictest sense and animals don’t. And anyone who says that we are not brains faces the problem of how we could know that we are not our thinking brains. But the idea that only brains think in the strictest sense turned out to rest on the assumption that all the parts of a genuine thinker must be somehow directly involved in its thinking, and that principle collapsed under scrutiny.
  5. The view that we are temporal parts of animals is based on the general ontology of temporal parts. It offers to solve the metaphysical puzzles about lumps and statues24 and the like that constitutionalism purports to solve, but without claiming that physically identical objects can differ in other respects (though this requires a counterpart-theoretic account of modal25 predication). It also offers a nifty solution to several problems of personal identity. On the other hand, it implies that there are a vast number of thinking beings now sitting in your chair, most of which have very different pasts and futures from those we ascribe to you. Worse, it implies that in the strictest sense we ourselves don’t think at all; only our momentary stages do. We could avoid this by saying that we ourselves are stages – the stage view26 – but that is desperately implausible.
  6. According to the bundle view we are made up of mental states and events. This appears to be the only account of what we are27 that is compatible with the psychological individuation28 principle. But a bundle of mental states doesn’t seem to be the sort of thing that could think. Nor does the bundle view suggest any solution to the thinking-animal problem.
  7. The paradox of increase provides what may be the best argument for our being immaterial substances. The easiest solution to the paradox is to deny that anything can change its parts; and the only things we could be that don’t change their parts, yet have anything like human careers, are immaterial substances. This would also solve the thinking-animal problem by ruling out the existence of animals. Immaterialism faces grave problems, however, about the apparent dependence of the mental on the physical.
  8. Then there is nihilism29, the view that we do not exist. It would solve all the problems about our metaphysical nature in one fell swoop. If nihilism30 is to be a serious claim and not a mad denial of the obvious, however, it must allow that statements apparently about people can be somehow right; so it must be possible to paraphrase those that are right into claims that capture their rightness without implying the existence of people. This will not be an easy task. Nihilism31 also appears to have unwelcome practical consequences: it is about as depressing as solipsism, and threatens to deprive us of any possible reason for acting.
  9. These seven views (eight if you count the stage view)32 have been the main focus of our attention. We also discussed more briefly a number of “minor views”. There is the view that we are temporal parts of brains, and the view that we are material things that brains constitute when they are in the right states. The indeterminate-size view says that it is indeterminate whether we are brains, human organisms, or beings of some intermediate size. According to homunculism, nothing thinks all of your thoughts: each sort of mental activity is carried out by a different part of the brain. There is the view that we are unthinking bundles of thoughts, the view that we are bundles of psychological universals33, and the view that we are something like computer programs. We might be “masses of matter”, which either exist only as long as they are in human form (the momentary-mass view), or endure as long as the particles composing them exist, but are human and able to think for only a moment (the persisting-mass view). There is compound dualism, the view that each of us is composed of an immaterial substance and an organism. Hylomorphism says that each of us is somehow a compound of a particular form and a parcel of matter. Finally there is the view that we are simple material things: either particles in the brain (Lilliputian materialism), or things coinciding with an entire human organism (Brobdingnagian atomism). Although there is something to be said for each of these, those that are not simply variants of the principal views are hard to take seriously.
9.2 Some opinions
  1. I will say no more about the minor views. What about the seven principal views? We have seen that each has its virtues, but also faces worrying objections. And it is hard to get philosophers to agree about how these virtues and vices stack up. Some, for instance, are so impressed with the way the temporal-parts view handles metaphysical puzzles that they are happy to accept its unwelcome implications; others find these implications so repugnant that they will accept almost anything in order to avoid them. Disagreements about such fundamental matters are notoriously hard to resolve.
  2. Rather than leave it at that, though, I will let my hair down and say what I am inclined to think. Many of my opinions about the relative merits of the various views have been evident in previous chapters, despite my attempt to be even-handed. But I can say more. This section is sheer autobiography, and I don’t expect it to persuade anyone. My attempts at persuasion are in the earlier chapters.
  3. As I see it, the brain view and the bundle view are right out. I suppose it is just about conceivable that the brain view might be true. It is certainly better than saying that my brain does my thinking for me. I reject it because there is no reason to believe it. The bundle view, on the other hand, is built on the idea that bundles of thoughts think, which I see as a conceptual mistake.
  4. Immaterialism is more promising. There are arguments for it that, though inconclusive, are not grounded in confusion in the way that the attraction of the brain view is grounded in confusion. On the other hand it faces more than its share of problems. It is shrouded in mystery and impenetrable to reason, yet fascinating – a bit like theism.
  5. Many advocates of the constitution view34 seem to see it as purely angelic. They see objections to it to as based on misunderstanding, rather than as worrying but inconclusive. I have made it plain enough that I disagree. Though the constitution view35 is part of a package that would do a lot of useful work, it strikes me as deeply implausible and, above all, unprincipled. There are too many questions about how constitution is supposed to work that have no answers, yet seem to demand answers. It is, like immaterialism, impenetrable to reason. That leaves nihilism36, the temporal-parts view, and animalism37. I judge these to be the finalists in the contest.
  6. As I have said, I am unsure what to make of nihilism38. It would solve a lot of problems, particularly if it is combined with compositional nihilism39. (Composite objects attract metaphysical problems like ripe fruit attracts flies.) But I am suspicious of the idea that it solves all problems about composite objects without incurring any of its own. I am not sure whether it is possible to tell an adequate story of the world in terms of mereological simples. I also suspect that nihilism40 may have grave implications for ethics, practical reason, and the value of our lives. Because it is unclear to me what follows from nihilism41, however, I have been unable to turn these suspicions into conclusive arguments. This is an area that needs more work. Nihilism42 remains, as they say, a live option.
  7. The consequences of the temporal-parts view are better understood, and to my mind they’re not very nice. I don’t like being forced to accept counterpart theory, which seems to me to make important questions about what it is possible for us to do into matters for arbitrary decision. I find it hard to accept that there are millions of beings sitting here and writing these words, most of which diverge radically in the past and future from what I take to be my own trajectory through space-time. Most of all, I don’t like the claim that all things have all their properties without temporal qualification – an essential feature of four-dimensionalism as we know it. This is what implies that persisting things have properties temporarily only in the sense of timelessly having temporal parts located at different times that have those properties timelessly. It seems to me more or less incredible that I don’t think or act in the strictest sense, but something other than me does; and to say that I do think and act in the strictest sense but exist for only a moment, as the stage view43 has it, is hardly better. That goes against my deepest instincts. I suppose I might be persuaded to put my instincts aside, but only by means of an argument for four-dimensionalism more forceful than any so far proposed.
  8. And animalism44? I cannot hide the fact that I have argued for animalism45 in the past. But it was not my intention to promote it here. Although I suppose I still find animalism46 the best answer to the question of what we are47, that preference rests on my aversion to the temporal parts view and nihilism48; and I don’t expect to persuade anyone that those views are false, apart from those who share my aversion to begin with.
  9. There is more to say, though. So far I have discussed only the best-known objections to animalism49, to do with its implications about personal identity (§§2.8-2.9). I argued that they are less troubling than they might appear. And indeed, they might seem rather slight compared with the problems facing the other accounts of what we are50. But I also said that there are more serious objections to animalism51, and promised to return to them later. Now is the time.
9.3 Animalism52 and the thinking-parts problem
  1. Recall the thinking-brain problem. Doesn’t my brain think my thoughts? And if it does, how do I know that I’m not it? Now we dispensed with one version of this problem, according to which a human animal53 can think only in the sense of having a brain that thinks strictly speaking. That would mean that according to animalism54 I don’t think in the strictest sense, which is absurd. I answered this objection by arguing that there is no reason to suppose that only brains think in the strictest sense, while larger things think only in the sense of having thinking brains as parts. But even if brains are not the only true thinkers, it doesn’t follow that they don’t think at all. Perhaps both brains and whole organisms think in the strictest sense. In that case we should expect every part of an organism that includes a brain to think: heads, upper halves, left-hand-complements, and so on. That would be trouble enough for animalism55. It would not imply that we are not animals, but it would make it hard to see how we could ever know that we are animals. If you think you’re an animal, then your head, which thinks just as you do, ought to think, mistakenly but on the same grounds, that it is an animal. So for all you know you might be your head. Why suppose, then, that you are an animal, rather than a head or a brain or some other thinking part of an animal?
  2. Call this the thinking-parts problem. It is structurally analogous to the thinking-animal problem: both consist in the apparent existence of beings other than ourselves that think our thoughts. Animalists56 need to solve the thinking-parts problem. At any rate they are committed to its having a solution compatible with animalism57. Otherwise – if there is no way of knowing that we are animals rather than brains or heads or the like – there will be no reason to accept animalism58. I find the thinking-parts problem considerably more troubling than the familiar objections to animalism59 – its unintuitive consequences in brain-transplant60 cases, for instance. How bad is it?
  3. Well, it is no worse for animalism61 than it is for many other accounts of what we are62. The same problem arises for any view according to which we are animal-sized things: for the constitution view63 and the temporal-parts view, for instance. They too presuppose that we are able to know that we are not brains or heads; and their advocates have no resources for solving this problem that are not available to animalists64 as well. So the thinking-parts problem is no reason to prefer any other view to animalism65. (Or at least none except nihilism66 and immaterialism, which have resources for solving the thinking-parts problem that are unavailable to anyone else: for instance they can adopt mereological nihilism67.)
  4. Even so, the thinking-parts problem threatens to show that animalism68 is no better than its rivals: even if it gives us no reason to prefer any other account of what we are69 to animalism70, it may leave no reason to prefer animalism71 either. The main reason to suppose that we are animals is the apparent fact that human animals72 think our thoughts. If this really is a fact, then it is hard to see how we could have any reason to suppose that we are anything other than animals. In other words, animalism73 has the virtue of avoiding the thinking-animal problem. But if brains and heads also think our thoughts, it is hard to see how we could have any reason to think that we are animals either. The thinking-parts problem threatens to imply that if anything thinks our thoughts and performs our actions, many beings of different sizes do so: organisms, heads, brains, and many other such things.
  5. That would be a mess. Which of these beings should we be? Well, if the question of what we are74 is the question of what sort of beings think our thoughts, the answer would be that many beings of different sorts do. If the question is what sort of things our personal pronouns and proper names denote, the answer may be that it is indeterminate: they refer ambiguously to brains, heads, human organisms, and many beings of intermediate sizes. Either way, it would apparently be indeterminate whether we are animals, brains, or something in between: the indeterminate-size view mentioned in §4.2.
  6. No one is going to find the indeterminate-size view appealing. But maybe it needn’t come to that. Perhaps animalists75 can solve the thinking-parts problem. How? Well, we saw that the thinking-parts problem is structurally analogous to the thinking-animal problem; so the possible solutions to the thinking-parts problem ought to parallel the possible solutions to the thinking animal76 problem. Let us set aside the possibility of solving the thinking-animal problem by accepting animalism77, for the analogous solution to the thinking-parts problem is to accept the indeterminate-size view, and that is what we wanted to avoid. There are three other possibilities78. One is to deny that human animals79 can think: a psychological solution (§2.5). Another is to accept that human animals80 think our thoughts, and argue that we are nonetheless able to know that we are not those animals: an epistemic solution (§§2.6-2.7). The third is to deny the existence of human animals81: a metaphysical solution (§2.4). Whatever merits these proposed solutions may have, it is easy enough to transform them into solutions to the thinking-parts problem. And if anything, the possible solutions to the thinking-parts problem look rather better than the corresponding solutions to the thinking-animal problem.
  7. A psychological solution to the thinking-parts problem would say that brains, heads, and other spatial parts of human organisms cannot think, or at least not in the way that you and I can. We considered some possible reasons for this in §4.2: brains cannot think because they are not organisms, or because thinking is a maximal property and brains are parts of larger thinking beings, or because the concept we use the word ‘thinking’ to express simply does not apply to brains. Those proposals were not very satisfying. Still, the view that our brains think and act in just the way that we do sounds wrong, even if it is hard to say why it is wrong. Maybe I just haven’t been clever enough to see it. So there may be a good psychological solution to the thinking-parts problem. It certainly looks more promising than a psychological solution to the thinking-animal problem: whatever it is that prevents brains and heads from thinking is unlikely to prevent human animals82 from thinking – unless perhaps it is impossible for a material thing of any sort to think.
  8. What about epistemic solutions? Could we somehow know that we are thinking animals83 rather than thinking heads or brains? Someone might suggest that the meaning of the word ‘person’ prevents it from applying to undetached heads, brains, and other proper parts of human animals84, despite their eminent psychological qualifications. Just why this should be – what it is about the word ‘person’ that explains this surprising gap in its extension – will not be easy to say. In any case, the suggestion would go on to say that our personal pronouns and first-person thoughts refer only to people. This would be something we could know a priori, just in virtue of being competent speakers of English. It would follow that our personal pronouns do not refer to brains and the like, and that we could know this. And since we are whatever our personal pronouns refer to, we could conclude that we are not brains or heads or other parts of human animals85.
  9. This is the personal-pronoun revisionism of §2.7, applied to the thinking-parts problem. As a solution to the thinking-animal problem, personal-pronoun revisionism seems rather desperate: it takes some doing to believe that human animals86, despite being psychologically indistinguishable from ourselves, somehow don’t count as “people”. But the idea that undetached brains don’t count as people, even if they really are as clever as we are, sounds a bit more hopeful. At any rate such things are very different from what we thought people were like. (It sounds wrong to say that brains are people, just as it sounds wrong to say that they think and act.) So an epistemic solution to the thinking-parts problem may have more promise than an epistemic solution to the thinking-animal problem.
  10. Then there is the metaphysical solution: there are no brains or heads or upper halves. (Cases in which a human animal87 has had so many parts cut away that it has itself become a brain or what have you, if that is possible, would be an exception: the proposal is that there are no undetached brains, heads, and so on.) Of course, my head isn’t empty: there are presumably particles “arranged cerebrally” there. But those particles don’t compose anything. Although it is no doubt in some sense correct to say in ordinary circumstances – when we are not doing metaphysics – that human beings have brains, what makes it correct is facts about particles or the like that do not entail the existence of brains. This is the sparse ontology once more. And if there are no brains or heads, we needn’t worry about ruling out the possibility that we might be brains or heads.
  11. The claim that there are no such things as undetached heads is not very plausible. Even so, a metaphysical solution to the thinking-parts problem is surely better than a metaphysical solution to the thinking-animal problem. There is something at least a little bit peculiar about undetached brains and heads – not to mention upper halves and left-hand complements. Their boundaries are at least in part arbitrarily drawn. Aristotle denied their existence, or at least denied that they were substances. Unless every matter-filled region of space, no matter how arbitrary and gerrymandered, contains a material object, we can understand why someone might doubt whether there are such things as undetached heads. If there were no human organisms, on the other hand, just about everything we ever believed about the ontology of material things would be wrong. Organisms look like paradigm cases of material things. To deny their existence is to deny the reality of all ordinary objects.
  12. So although the thinking-parts problem is a serious objection to animalism88, it is at least as bad for its main rivals (leaving aside nihilism89, anyway). And a solution is not beyond hope. Moreover, it looks less threatening to animalism90 than the thinking-animal problem looks to its rivals.
9.4 Animalism91 and the clay-modelling92 puzzle
  1. Even if the thinking-parts problem is no worse for animalism93 than it is for any other popular account of what we are94, however, there are other metaphysical problems that threaten to afflict animalism95 in particular. The clay-modelling96 puzzle, for one, looks like a problem for animalism97 but not for its main rivals. A lump of clay98 modelled into the shape of Thatcher seems able to survive being squashed, but not the clay statue99 of her; yet the lump appears to be the statue100. The most popular ways of reconciling these claims appeal to constitution or temporal parts. Constitutionalists accept that the lump but not the statue101 can survive squashing by saying that they are numerically different and have different persistence conditions102, despite coinciding materially: the lump, as they put it, constitutes the statue103 (§3.2). Four-dimensionalists typically say that the statue104 and the lump are numerically different because one is a proper temporal part of the other, and explain why we appear to attribute different persistence conditions105 to the two objects in terms of counterpart theory (§§5.3-5.4).
  2. Now animalists106 could say the same thing. They could say that statue-shaped107 lumps of clay constitute statues108 and accept constitutionalism; or they could say that statues109 are temporal parts of lumps and accept four-dimensionalism. At any rate these views are formally consistent with animalism110. But animalism111 does not sit easily with either view. Animalism112 does not sit well with constitutionalism because constitutionalism suggests that human animals113 coincide materially with beings that would appear to be mentally just like we are, raising the problem of how we could ever know whether we are the animals.
  3. Constitutionalism suggests that something stands to a human animal114 as a lump of clay stands to a clay statue115 – something that would outlive the animal if it were squashed. If clay statues116 have to be constituted by something, shouldn’t human animals117 have to be constituted by something too? Now suppose you are an animal, and that a lump of something constitutes you. The lump would be physically indistinguishable from you for as long as it constitutes you. It would have the same brain and nervous system as you have (at any rate it would be neurologically identical with you), and the same surroundings. It would show the same behavioral evidence of intelligence and conscious awareness as you do. That suggests that for a while at least, the lump would be conscious and intelligent – indeed, it would be mentally just like you. What grounds could you ever have, in that case, for supposing that you are the animal and not the lump?
  4. Or again: if the right sort of lump of clay in the right circumstances constitutes something that is essentially statue-shaped118, we might expect the right sort of human organism in the right circumstances to constitute something that is essentially able to think. Now no human animal119 is essentially able to think. But the animals we call our bodies would seem to be of the right sort and in the right circumstances to constitute essential thinkers. So constitutionalism suggests that each human animal120 coincides with an essential thinker. Again, how could you ever know which thinker you are?
  5. Now it doesn’t strictly follow from constitutionalism that human animals121 coincide with lumps of something or with essential thinkers. Someone might say that there is a lump of clay constituting a clay statue122 of Thatcher but no lump of anything constituting Thatcher herself. And someone might suppose that lumps of clay sometimes constitute essential statues123 but human organisms never constitute essential thinkers. But that would be surprising. More to the point, it would seem arbitrary and unprincipled – a bit like holding that statues124 of women are constituted by lumps but not statues125 of men. We should have no idea why the two cases – the case of statues126 and the case of organisms – are so different; and we should expect there to be a reason why. For all we know, perhaps, it could be true; but it is hard to see how we could ever know that it was.
  6. So combining animalism127 with constitutionalism raises a problem that animalists128 will want to avoid. They will want to deny that animals constitute or are constituted by anything. Of course, constitutionalism appears to face this sort of problem whether or not it is combined with animalism129: if we are essential thinkers, for instance, we shall coincide materially with human animals130, leaving us wondering how we know we are not those animals. So you might think that adopting constitutionalism would leave animalists131 no worse off than constitutionalists are whether or not they accept animalism132. But this is not so. Just as animalists133 will not want to accept constitutionalism, constitutionalists will not want to accept animalism134. If there are essential thinkers coinciding with animals, the widespread conviction that we are essential thinkers will be a reason to reject animalism135. If constitutionalism is true, in other words, there are better candidates for being us than human organisms. For these reasons animalists136 are ill advised to accept constitutionalism. And we saw in §5.4 why animalism137 fits badly with the ontology of temporal parts: if four-dimensionalism is true, the best candidates for being people, and so the best candidates for being ourselves, would seem not to be animals, but proper temporal parts of animals: parts composed of thinking person-stages, which include no unthinking embryonic138 stages. A four-dimensionalist could be an animalist139 without inconsistency, but it is unsurprising that almost no one holds this combination of views.
  7. So animalists140 will not want to solve the clay-modelling141 puzzle by adopting constitutionalism or four-dimensionalism. They will not want to solve the paradox of increase, the amputation puzzle, or the replacement puzzle in that way either. Nor can they simply accept the paradoxical conclusion that it is impossible for anything to gain or lose parts, for that is incompatible with our being animals. What can they say, then?
  8. There are always the way of funny logic and the way of funny persistence conditions142 (§7.4); but they’re not very plausible. There remains the way of sparse ontology. It says, in the case of the paradox of increase, that when an object acquires a new part, there is then nothing composed of the object’s original parts. If you assimilate an atom, there is no such thing, immediately afterwards, as “all of you but that atom”: you are now composed of your old atoms together with the new one, and the atoms that a moment ago composed you now compose nothing. In the case of the amputation paradox, the solution is that you have no part that you could survive being pared down to: you could survive the loss of your left hand, but there is no such thing, now, as “all of you but your left hand”, your left-hand complement. The way of sparse ontology also implies that there is no such thing as your undetached head or brain – supposing, anyway, that you could be pared down to a head or a brain. This has an important implication: solving the amputation paradox by the way of sparse ontology would solve the thinking-parts problem as well. That is because the entities that would generate the thinking-parts problem also figure in the amputation paradox.
  9. Adopting a sparse ontology of material objects can solve the paradox of increase, the amputation paradox, and the thinking-parts problem. We can use the same strategy to solve the clay-modelling143 and replacement puzzles as well: we can deny that there are any such things as clay statues144 or lumps of clay. There may be tiny particles arranged by sculptors in a statuesque145 fashion, but they don’t compose anything.
  10. In my view this is what animalists146 ought to do: they should solve their metaphysical worries by denying the existence of the entities that would generate them. (Some animalists147 have done it: see van Inwagen 1990, Hoffman and Rosenkrantz 1997, Merricks 2001.) Anyone who likes animalism148 but dislikes the sparse ontology will have to solve the problems we have been considering – both the puzzles about the ontology of material objects and the thinking-parts problem – in some other way; and no other way looks very nice.
  11. This might come as something of a shock. No one wants to be told that there are no such things as lumps or statues149, or undetached brains or heads. It goes against what we learned at mother’s knee. This outcome is especially surprising because animalism150 seemed at first to be so sensible. That our being animals conflicts with things we are inclined to say about brain transplants151 may be something we could live with. But now animalists152 are asking us to give up some of the most ordinary beliefs in the world. What appeared to be a pleasant landscape turns out to have a coal pit running through it.
  12. Let me say three things in defense of the sparse ontology. First, it doesn’t imply that we are mistaken to think that we have heads, or that there are many Greek statues153 in the British Museum. At least it doesn’t imply that these beliefs are mistaken in the way that it would be mistaken to think that we have tails, or that there are many Martian statues154 in the British Museum. As we saw in our discussion of nihilism155, the sparse ontology allows that such beliefs can be taken for all ordinary purposes to be true. We can say, or try to say, what is right about these ordinary beliefs in terms of objects the sparse ontology recognizes, such as simple particles and people: even if there are strictly speaking no heads or statues156, some of our particles are arranged in such a way that they would compose undetached heads if they composed anything, and many particles in the British Museum were arranged more or less as they are now by Greek sculptors. Because our mothers were trying to teach us anatomy and not metaphysics, they were perfectly correct to tell us that we have heads, and we were perfectly correct, as far as our childish purposes were concerned, to believe them. So the conflict between the sparse ontology and our ordinary beliefs is less jolting than may appear. (Some sparse ontologists even deny that there is any conflict at all: see van Inwagen 1990a: §10.)
  13. Second, the other ways of solving the puzzles are not clearly any better. They may even be worse. If you don’t like the sparse ontology, consider the alternative. Think about the ontology of temporal parts, constitutionalism, funny persistence conditions157, and funny logic. Now think about the indeterminate-size view, personal-pronoun revisionism, and the other possible responses to the thinking-parts problem. If the sparse ontology is false, one of each of these groups of repugnant views must be true.
  14. Third, the sparse ontology follows from what most of us will take to be the best approach to another important metaphysical problem: when composition takes place. So it is not only animalists158 who have reason to accept the sparse ontology. I will develop this thought in the next section.
9.5 Theories of composition
  1. Suppose that animalists159 really are best advised to solve the thinking-parts problem, the claymodelling puzzle, the amputation puzzle, and the rest by denying the existence of undetached heads and brains, lumps and statues160, and the other troublesome entities. This proposal may appear not only implausible, but unprincipled too. If there are so many fewer composite objects than we thought, why suppose that there are any at all? In particular, why suppose that there are any human animals161? The proposal is this: there are particles arranged in ever so many ways – undetached-human-headwise, brainwise, clay-lumpwise, statuewise162, anthropomorphically, and so on. When particles are arranged anthropomorphically, they compose something. (And not just something: they compose an organism, a thing that persists as a living thing for many years despite constant material turnover. They don’t compose a mere mass of matter.) But no particles arranged in any of those other ways compose anything. In those cases there are only particles. In other words, the troublesome objects are all unreal, but we human animals163 are an exception. That may sound more like wishful thinking than the verdict of unbiased rational inquiry. Why should particles arranged anthropomorphically compose something, but not particles arranged headwise or statuewise164 or lumpwise? What accounts for the difference? No one ought to accept this fairy-tale ontology unless there is some principled reason why it must be so.
  2. This worry has to do with a general question about composition: When does it take place? Under what conditions do smaller things (let us restrict ourselves to concrete material things) make up or compose something bigger? If we have some things, the xs, what is necessary, and what is sufficient, for there to exist something composed of the xs – something that has all the xs as parts, and all the parts of which share a part with one or more of the xs? What must those things be like, and how must they relate to one another and to their surroundings, in order for them to compose something? This is what van Inwagen has called the special composition question165. Call an answer to this question a theory of composition. The worry is whether any principled theory of composition is consistent with both animalism166 and the sparse ontology we have been considering.
  3. We have already discussed the view that there are no circumstances in which smaller material things compose something bigger: there are no composite material things at all, but only simples. In other words, things compose something if and only if there is only one of those things (by definition everything composes itself). This is compositional nihilism167. It would rule out our being composite objects of any sort. Since animals are composite168, it is incompatible with our being animals. In fact it is incompatible with almost any account of what we are169 apart from nihilism170 and immaterialism.
  4. Another theory of composition is that things always compose something, no matter what they are like in themselves or how they are arranged or situated: composition is universal or unrestricted or automatic. This is compositional universalism. It is of course incompatible with the sparse ontology. It is also highly counterintuitive. It implies that there are vastly more material objects than we would ordinarily have thought. Virtually all material things, on this view, have completely arbitrary boundaries: they are mere “ontological junk”. The proportion of material objects that are dogs or bicycles or planets or anything else of any interesting sort would be about the same as the proportion of regions of the earth’s surface exactly occupied by continents.
  5. What’s wrong with this surplus of objects? Well, for one thing it implies that every dog or bicycle – or person, if we are even partly material things – will almost exactly overlap with a vast number of other beings that differ from it only by having or lacking among its parts a few particles. That will probably make it indeterminate which of the many candidates you are – that is, which one we refer to when we say ‘you’ or use your name. This is the problem of the many171 of §5.8. Universalism also gives us many objects that differ from ourselves in more disturbing ways. It implies that there is, for instance, an object now made up of your upper half and my lower half – a being that is presumably psychologically indistinguishable from you. In fact universalism appears to imply that virtually all the beings that now think your thoughts are pieces of ontological junk. For all you know, it seems, you could be one of them.
  6. There are things that universalists can say to assuage these worries. They can say that it doesn’t matter if it is indeterminate which of the many candidates for being you is you because for all practical purposes we can’t tell them apart anyway. So everything we ordinarily want to say about you comes out true on any assumption about which one you are. And they can say that in ordinary contexts we ignore the arbitrary objects. Our names, pronouns, general terms (such as ‘dog’ and ‘man’), and quantifiers (such as ‘some’ and ‘all’) simply pass them by as if they weren’t there. There is endless room for debate about the extent to which stories like these make universalism easier to believe. But it takes a rather tough-minded philosopher to feel comfortable with it. (We will consider another disputable consequence of universalism in the next section.)
  7. In any case, animalists172 who accept the sparse ontology will reject both compositional nihilism173 and universalism, and many others will too. They will say that some objects compose bigger things and others don’t, depending on what those objects are like and how they relate to one another and (perhaps) to their surroundings. There are some composite objects, so to speak, but not just any: a few special matter-filled regions of space contain (exactly contain) material objects, but most don’t. They will plead for an intermediate theory of composition. Sensible though that may sound, however, it raises our worry once more: if some things compose bigger things and others don’t, which do, which don’t, and why? In particular, why suppose that any good theory of composition will be consistent with both the existence of human animals174 and the nonexistence of undetached heads, clay statues175, and the rest of the troublesome lot?
  8. We should expect an intermediate theory to say that things compose something just when they are somehow unified: when they relate to one another, and to nothing else, in some special causal or spatio-temporal way. Things not unified in this way will be those that would compose arbitrary or gerrymandered objects if they composed anything. The particles that we take to make up a live cat look like good candidates for being unified: a cat is the very opposite of an arbitrary object. The particles that are now located within the North Sea are considerably less likely candidates, and those located either within the North Sea or within the further half of the Andromeda Galaxy are less likely still – though even they are a good deal more unified than most.
  9. What sort of unity might be necessary and sufficient for things to compose something? It is surprisingly hard to say. I don’t know of any answer that has much plausibility on the face of it. I certainly don’t know of any that entails both the existence of most of the familiar furniture of the earth – dogs, bicycles, stones, that sort of thing – and the nonexistence of most of what I have called ontological junk. To my knowledge only two intermediate answers to the special composition question have ever seriously been proposed. Both count as “sparse”.
  10. The first is van Inwagen’s view that things compose something if and only if their activities constitute a biological life – a self-organizing event that maintains the internal structure of an organism (van Inwagen 1990; see also §2.2 above). This has the startling implication that there are no non-living composite objects. The only material things are simples – and I take it that the only material simples are elementary particles – and living things, things with lives. What sort of things have lives? Well, organisms do. Whether anything other than a living organism has a life depends on whether an organism can coincide materially with a non-organism (a point I will return to presently). If not, then the proposal is that the only composite objects are biological organisms.
  11. It is plausible enough to say that things compose something if their activities constitute a life. A life provides the sort of unity that leads us to suppose that the particles caught up in it compose something bigger: living things are paradigm cases of composite objects. But why should things compose something only if their activities constitute a life? Why couldn’t there be non-living composite objects? Van Inwagen’s argument is roughly that nothing other than a biological life provides the right sort of unity: taking any other unity relation to be necessary and sufficient for composition – in particular any sort of physical bonding – leads either to repugnant consequences or to grave logical difficulties.
  12. Call the view that things (or at least material things) compose something if and only if their activities constitute a life biological minimalism. This is not the place for a full-scale evaluation of minimalism. Whatever its merits, though, it is a principled theory of composition. And it has a most convenient consequence for the animalist176: it entails the existence of human animals177 and rules out the existence of nearly all the entities that would cause trouble for animalism178. Undetached heads, proper temporal parts of human animals179, statues180, and lumps, for instance, do not have lives. Neither are they simples. Biological minimalism therefore implies that there are no such things. Thus, not only does the way of sparse ontology enable animalists181 to solve the metaphysical problems they face, but it follows from one of the few principled theories of composition on offer.
  13. Now there is one sort of troublesome entity whose existence biological minimalism does not explicity rule out: things coinciding materially with human animals182. Minimalism is formally consistent with constitutionalism. It says that things compose something just when they are caught up in a life, but it doesn’t say that such things compose only one thing, namely an organism. For all it says, particles caught up in a life might compose both a human organism and a human non-organism. That would be trouble again: an awkward surplus of thinkers.
  14. But as we saw in §3.6, no one will want to combine biological minimalism with constitutionalism. Minimalism provides its own solution to the metaphysical puzzles that constitutionalism was invented to deal with: the clay-modelling183 puzzle, the replacement puzzle, the paradox of increase, and so on. If you’ve already accepted minimalism, there is no point in adopting constitutionalism as well. That would be paying twice for the same thing. The two views are also fundamentally opposed in spirit. Constitutionalism is a rich ontology; minimalism is an austere one. No one is going to suppose that almost none of those particles we think of as composing something actually compose anything, but those few that do compose something compose more than one thing. No one will say that the furniture of the earth consists of nothing but simple particles, organisms, and things coinciding with organisms. That would be like recommending a diet of bread, water, and chocolate fudge cake.
  15. So animalists184 can solve all their metaphysical worries at a stroke by adopting biological minimalism185. That may seem a high price to pay, but it does the job. And it may not be such a great sacrifice after all, compared with the other theories of composition.
  16. The second intermediate theory of composition, advocated by Hoffman and Rosenkrantz, is strikingly similar the first. They agree with van Inwagen that there are living organisms. (They disagree with him about what it is for things to compose a living organism – they propose an account in terms of what they call “functional unity” rather than in terms of lives – but this is a fine point.) The important difference is that on their account there are also non-living composite objects, which they call “mereological compounds”. Things compose a mereological compound, they say, just when they are physically bonded in a certain rigid way, the details of which I won’t try to reproduce. So their theory of composition is roughly this:
      Things compose something if and only if they are either functionally united or rigidly bonded186.
    Call it biological disjunctivism.
  17. Disjunctivism187 implies that there are no parts of organisms except tiny particles. That is because the members of any subset of particles that compose an organism are neither functionally united nor rigidly bonded. They are not rigidly bonded because some of them are in a liquid state: it belongs to the nature of an organism that most of its particles are not rigidly bonded. According to disjunctivism, then, such particles compose neither organisms nor mereological compounds. That is, they compose nothing at all. This makes disjunctivism a sparse ontology of material objects, even if it is less sparse than minimalism. It implies that the heads, shoulders, knees, and toes of our nursery-school ontology do not exist. Nor are there any undetached tails, leaves, or flowers. Like minimalism, then, disjunctivism implies the existence of human animals188 and the nonexistence of the parts of human animals189 that would generate the thinking-parts problem. It also solves the paradox of increase and the amputation paradox, or at any rate the versions of those problems that threaten to imply that organisms cannot change their parts, by the way of sparse ontology. Now because disjunctivism is consistent with the existence of non-living composite objects, it does not by itself do away with all the entities that would make trouble for animalism190. It may not solve the clay-modelling191 puzzle, for instance, which threatens to imply that human animals192 coincide materially with non-animals. Hoffman and Rosenkrantz solve metaphysical puzzles not involving organisms by making the further claim that all mereological compounds have their parts essentially. This claim, together with disjunctivism, leads them to say that all nonliving composite objects are what I called momentary masses (§7.3): things that persist only as long as their particles remain stuck together. So they might solve the clay-modelling puzzle by adopting “lumpism” (§3.2), the view that there are statue-shaped193 lumps of clay but no statues194 – but they reject the problematic claim that there are lumps of flesh on the grounds that the parts of such things would not be rigidly bonded. Because all non-living objects are momentary masses, on their view, there are no ordinary non-living things: what appears to be a persisting stick or stone or statue195 that frequently sheds an atom or two is in reality only a series of shortlived masses. None of these further claims strictly follow from disjunctivism. One could combine disjunctivism with constitutionalism. But that would be hardly more appealing than combining constitutionalism with minimalism.
  18. I won’t try to judge whether disjunctivism is more plausible than biological minimalism, less plausible, or about as bad. The important point is that it would solve the metaphysical worries facing animalism196 just as well as minimalism would.

Comment:
9.6 Composition and what we are197
  1. We have seen that animalists198 can solve the most serious worries facing their view by denying the existence of the troublesome entities that generate them. That may not be the only way of solving the problems, but it is the simplest, and, in my view, the best. We have also seen that this solution is not merely ad hoc, but follows from either of two principled theories of composition, minimalism and disjunctivism. So animalism199 leads very naturally to a certain sort of theory of composition.
  2. Now it seems that the reverse also holds: these two theories of composition lead almost inevitably to animalism200. Biological minimalism rules out the existence of most of the things we could be apart from animals. If the only material things are elementary particles and organisms, it would seem perverse to say that you and I are not organisms. Biological disjunctivists will also say that we are animals, for similar reasons. Someone could combine disjunctivism with immaterialism, or the bundle view, or the constitution view201, or even nihilism202 without formal contradiction; but that would have no appeal.
  3. Minimalism and disjunctivism are not the only possible intermediate theories of composition – that is, the only ones apart from universalism and compositional nihilism203. But it is no accident that they are the only ones that anyone has actually endorsed. Any other theory of composition that I can think of strikes me as considerably less plausible than either of these. If that is right, then anyone who rejects universalism and compositional nihilism204 ought to say that we are animals. In other words, animalism205 will be the best account of what we are206 if some but not all material things compose something bigger. There may perhaps be a good intermediate theory that does not lead almost inevitably to animalism207, but rather supports some other view of what we are208 or leaves it entirely open; but none has ever been proposed.
  4. Here, then, is a bold conjecture: there is an intimate connection between the question of what we are209 and the question of when composition takes place, or what material objects there are. Each theory of composition implies an account of what we are210, or at any rate a narrow range of accounts. Compositional nihilism211 leads to nihilism212 about ourselves (or perhaps to immaterialism). Universalism leads to some version of the temporal-parts view. And anyone who accepts an intermediate theory of composition will find it hard to avoid animalism213. Let us explore this thought.
  5. It is easy to see the link between compositional nihilism214 and nihilism215 about ourselves. Compositional nihilism216 entails that the only things we could be are mereological simples, as that is all there is on that view. That is of course compatible with our being simple immaterial substances, or even simple material things. But I doubt whether anyone tough-minded enough to endorse compositional nihilism217 will be drawn to immaterialism, and simple materialism is an exotic bird. In any case, compositional nihilism218 implies that we are simple things if we exist at all.
  6. How does universalism lead to the temporal-parts view? Nearly all universalists accept four-dimensionalism. They think that elementary particles are composed of temporal parts – particle-stages – and that any particle-stages whatever compose something. That makes it natural to suppose that we are composed of particle-stages, which is a version of the temporal parts view. I will explain.
  7. Suppose, for reductio, that universalism is true and you are not composed of particle-stages, but rather of particles. And suppose for a moment that no particles ever compose more than one thing at once: suppose that constitutionalism is false. (I will return to this assumption presently.) Now consider the particles that currently compose you. Call them the Ps. The Ps did not compose you a month ago. Most of them were not even parts of you then, but were widely dispersed over the earth’s surface. But all the Ps existed then219. And according to universalism they composed something then, for they compose something at every time when they exist. Call the thing the Ps composed then M (for ‘mass of matter’).
  8. Where is M now? Does it still exist? It would seem to. If the Ps compose something whenever they exist, no matter how they are arranged, we should expect them always to compose the same thing. If the arrangement of the Ps makes no difference to whether they compose something, how could it make a difference to which thing they compose (van Inwagen 1990: 77)? In other words, M would seem to be what I earlier called a persisting mass (§7.3): something composed of particles that exists, composed of the same particles, whenever those particles exist. In that case M is now located exactly where you are, and is now composed of the very particles that now compose you. Given that those particles cannot compose two things at once, it follows that you are M. But you are not M. You are not a persisting mass. You were not composed of the Ps a month ago. You were never widely dispersed. Our original assumption, that universalism is true and you are composed of particles, must therefore be false.
  9. What if M doesn’t exist now? What if the thing your current particles composed a month ago no longer exists? Well, why doesn’t it still exist? What caused its demise? Presumably it was that its parts, the Ps, got rearranged in some way. What way? The Ps have been in constant motion relative to each other for the last month. At what point during this period did M cease to exist? Which particular rearrangement of the Ps brought about M’s demise? The most plausible answer is that M ceased to exist as soon as the Ps got rearranged at all. It would be hard to believe that M managed to survive all the rearrangements the Ps have undergone during the past month, only to perish the moment those particles came to be arranged in human form. If a thing cannot survive its particles’ coming to be arranged in human form, surely it cannot survive its particles’ being rearranged in any way at all. And if a thing cannot survive any rearrangement of its particles, then an object whose particles are in constant motion can exist for only a moment. But your particles are in constant motion. It follows that you exist for only a moment: you are a momentary mass. But that is absurd. You are no more a momentary mass than you are a persisting mass. Once again, the assumption that you are composed of particles, given universalism, must be false.
  10. So you could not be composed of particles if universalism is true. What could you be composed of, then? The most obvious answer is that you are composed of particle-stages. None of the Ps are parts of you; rather, your parts include certain temporal parts of the Ps. (Though it may not be easy to say which ones. Animalists220 can say it is those caught up in your biological life, but non-animalists will deny221 this.) And if you are composed of temporal parts of particles, then you yourself have temporal parts.
  11. This is the sort of reasoning that leads universalists to the temporal-parts view. It is not a watertight proof, for it is consistent with universalism that we are composed neither of particles nor of particle-stages. We might be composed of mental states. Or we might be composed of nothing at all, other than ourselves: we might be simple. One might even say that M survived all the rearrangements of the past month but perished when the Ps came to be arranged in human form – another version of the way of funny persistence conditions222 (Rea 1998). But these are eccentric views.
  12. The most obvious weakness in this argument, you might think, is the assumption that constitutionalism is false. Why not say that M still exists now and simply coincides materially with you? More generally, perhaps any particles whatever compose a persisting mass (or a momentary mass; it doesn’t matter), but none of us is a mass; rather, each of us is constituted, at any moment, by a mass. In that case M would stand to you roughly as a lump of clay stands to the clay statue223 made from it, except that you come to be constituted by a different mass every fraction of a second, whereas a statue224 coincides with the same lump for a longer period.
  13. Now I have argued against constitutionalism. But I needn’t rely on those arguments here. Even if constitutionalism is completely unobjectionable, there is no point in bringing it in now, to block the inference from universalism to the temporal-parts view. The proposal we are considering is to answer the special composition question by accepting universalism, and to make this compatible with our being composed of particles rather than particle-stages – different particles at different times – by asserting that we coincide materially with masses. But this only raises a new question about composition: under what circumstances do particles compose something other than a mass? Call this the new composition question. (It is a close relative of the question of when constitution occurs of §3.5.) We should expect it to have an answer if the special composition question has an answer.
  14. Constitutionalists will not want to give a “universalist” answer to the new composition question. They will not say that any particles whatever, no matter how they are arranged or what they are like in themselves, compose something other than a mass. It is bad enough to have to say that the particles now composing you composed anything at all a month ago when they were randomly scattered; it is far worse to say that they composed two things then: not only a persisting mass, but also something other than a mass, something that stood to the mass then as a clay statue225 stands to the lump of clay it is made out of. If universalism is a bloated ontology of material objects, this is double bloating. Nor can constitutionalists give a “nihilistic” answer to the new composition question: they cannot say that particles never compose something other than a mass. They will want to give an intermediate answer: they will say that whether particles compose something other than a mass depends on what they are like and how they are arranged and situated. But then what is the point of answering the special composition question with ‘always’? Whatever intermediate answer constitutionalists can give to the new composition question will do just as well as an answer to the special composition question. If we could say how things have to be arranged and situated in order to compose something other than a mass, why not say that that is what it takes for things to compose anything at all? What is gained by adding a capacious ontology of masses to a moderate ontology of ordinary material things? As far as I can see, nothing at all.
  15. So the combination of universalism and constitutionalism is unappealing. Universalists are under considerable pressure to accept the temporal-parts view, and it is unsurprising that nearly all succumb to it.
  16. If this is right, then our conjecture about a link between when composition occurs and what we are226 would seem to be confirmed. Universalism leads by a devious but fairly secure path to the temporal-parts view. Compositional nihilists will are almost certain to accept nihilism227, or perhaps immaterialism. And we have seen that the only other serious theories of composition on offer, minimalism and disjunctivism, lead almost inevitably to animalism228. I concede that none of these inferences is irresistible. A determined metaphysician could devise an account that combined universalism with animalism229, say, or compositional nihilism230 with the constitution view231, or biological minimalism with immaterialism. But I doubt whether there would be any point in such an exercise. So it appears that a theory of composition would tell us what we are232. At any rate what we say about composition will constrain dramatically what we are able to say, or what it would be sensible to say, about our metaphysical nature. The way to find out what we are233 – or one way, at least – is to find out when composition occurs.
9.7 Brutal composition
  1. Earlier I expressed the opinion that the three best accounts of our metaphysical nature are nihilism234, the temporal-parts view, and animalism235. And I have argued that each of these accounts follows very naturally, if not quite inevitably, from one of the available theories of composition: compositional nihilism236 leads to nihilism237 about ourselves, universalism leads to the temporal-parts view, and intermediate theories lead to animalism238. The dependence also runs the other way: nihilism239 about ourselves leads to compositional nihilism240, the temporal-parts view presupposes universalism, and animalism241 is plausible only on an intermediate theory of composition. If that is right, then we can just about answer the question of what we are242 by giving a theory of composition, and we can just about work out when composition occurs on the basis of what we are243.
  2. This picture may be too neat, however, for I haven’t considered the possibility that there is no true theory of composition – that is, that the special composition question has no answer. Or rather that it has no systematic or principled or intellectually satisfying answer: no complete and non-trivial answer that we could know or write down. It may be that certain conditions are necessary for things to compose something: perhaps objects that exert no causal influence over one another cannot compose anything, for instance. It may also be that certain conditions are sufficient for things to compose something: perhaps things caught up in a biological life always compose something. Even so, there may be no complete, finite set of conditions, each of which is necessary and all of which are jointly sufficient for things to compose something.
  3. Markosian (1998) has called this idea brutal composition – the idea being that whether things compose something is a brute fact, not explainable in terms of any general principle. Its attraction is plain enough. It seems to most of us that some things compose bigger objects and others don’t, ruling out the two “extreme” theories of composition, compositional nihilism244 and universalism. Yet neither biological minimalism nor biological disjunctivism sounds right either: most of us are inclined to believe that there are at least some visible parts of organisms. The fact that no plausible answer to the special composition question compatible with these two convictions has ever been proposed might suggest that the question has no answer.
  4. Although brutal composition is a response to the special composition question, it does not actually answer that question. It doesn’t tell us when composition occurs. It tells us, rather, that we cannot say when composition occurs. So the possibility that composition might be brute challenges the idea that we could find out what we are245 by finding out when composition occurs.
  5. Well, how does brutalism relate to the question of what we are246? It might seem to support animalism247. Brutalists will almost certainly accept the existence of human animals248. Human animals249 are paradigm cases of composite objects: anyone who denies that particles arranged anthropomorphically compose human organisms might as well accept compositional nihilism250. And if there are such things as animals, it is hard to deny that we are animals. Animalism251 and brutal composition also share the same air of humble plausibility – especially when compared to the alternatives, which in both cases are rather wild.
  6. They may even seem an ideal match. Animalism252 would be threatened by the existence of such troublesome entities as thinking parts of animals and material things coinciding materially with animals. I suggested ridding ourselves of these entities by accepting a sparse ontology of material things, such as biological minimalism. Could we not dispense with them at a lower cost by going brutalist?
  7. I think the answer is no. As far as I can see, brutalism is no help in defending animalism253 against the objections we have considered in this chapter. It is hard to combine brutalism with animalism254. In fact it is hard to combine it with any attractive account of what we are255. At any rate the considerations that make brutalism attractive make it hard to say what we are256.
  8. Brutalism would help the animalist257 if the troublesome entities were all arbitrary objects: pieces of ontological junk, things that only friends of universalism would believe in. And some are: undetached hand complements, for instance. But not all. Consider undetached heads: they belong to the nursery-school ontology that we learned as children. Undetached brains belong to the anatomy-textbook ontology that we learned later on. The conviction that there have got to be such things, and not merely particles arranged capitally or cerebrally, is just the sort of thing that leads philosophers to reject minimalism and disjunctivism and retreat to brutalism. If there are any composite objects other than organisms, one is tempted to say, there are surely heads. If there are no heads we may as well accept minimalism. So brutalists are likely to accept the existence of undetached heads and brains, even if they deny the existence of arbitrary parts of human organisms. This will prevent them from giving a metaphysical solution to the thinking-parts problem – that is, from saying that we know we are not heads or brains because there are no such things. And I have not seen any other very satisfactory solution to the thinking-parts problem. Certainly brutal composition suggests none.
  9. For the same reason, brutalists cannot solve the amputation paradox or the paradox of increase by the way of sparse ontology. If there are such things as undetached heads, and if you were pared down to a head and kept alive by life-support machinery, how would you then relate to your head? Brutalists could of course turn to funny logic or funny persistence conditions258, but that would diminish their view’s attraction considerably. If they accept the existence of undetached heads, and accept that you could survive being pared down to a head, and reject the ontology of temporal parts, funny logic, and funny persistence conditions259, they will have to turn to constitutionalism: to say that you and your undetached head would come to coincide materially if the rest of you were cut away. The clay-modelling puzzle will also push brutalists towards constitutionalism. Because they want to give an account of what material objects there are that comes close to what we are ordinarily inclined to say, they will want to accept the existence of statues260 and lumps. So they will be unable to solve the clay-modelling puzzle by denying the existence of statues261 and lumps of clay, or by saying that there is a lump there but no statue262. Nor will they want to solve it by turning to temporal parts, for they reject universalism. Funny logic and funny persistence conditions263 aside, once again, constitutionalism looks like the only alternative. The trouble with brutalism is that the very ontological generosity that makes it attractive gives us many of the material things that make it hard to say what we are264.
  10. Brutalists might try to solve this problem by persisting in their brutalism. They might adopt constitutionalism, and say that we are things constituted by animals. They might go on to say that things constituted by human animals265 can think but human animals266 themselves cannot; and neither can a brain or any other proper part of a human animal267. Why not? What is it about animals or undetached brains that prevents them from thinking? Brutalists might reply that nothing does: their inability to think is simply a brute fact, not consisting in or explainable in terms of other facts. They can’t think, and that’s all there is to it. The question of what it takes for something to be able to think has no answer, or at any rate no systematic or principled or satisfying answer. Or at least we shouldn’t assume that it has an answer, and thus our inability to answer it need not trouble us. They might go brutalist in response to other questions that arise on this view as well, such as when constitution occurs and what determines our boundaries. I complained earlier that many questions about how constitution is supposed to work have no answers, yet seem to demand answers. One might simply reject this demand.
  11. This is not the place to discuss the merits of brutalism as a general philosophical strategy. But the brutalism about thinking that I have suggested here has little of the attraction of brutal composition. The brutalist about composition wants to say that there are such things as organisms, undetached heads, lumps of clay, and clay statues268, and no such things as hand-complements and disconnected objects composed of the upper half of one human being and the lower half of another. That sounds good. The only worry is that it might be unprincipled: we can’t think of any reason why there should be heads but no hand-complements. (Someone might say that hand-complements would be arbitrary objects, while heads are not; but then the question is what this arbitrariness comes to.) Brutalism denies that there must be any such reason why. The brutalist about thinking, on the other hand, claims that human animals269 cannot think but material objects physically indistinguishable from them can. That isn’t plausible on the face of it. Quite the opposite. To say further that there is no explanation for this astonishing state of affairs only makes matters worse. Perhaps we could be warranted in believing that there are such things as undetached heads but no such things as undetached hand complements, even if we have no principled reason for it. But we are surely not warranted, without a principled reason, in denying that human animals270 ever think.
  12. So whereas accepting one of the proposed theories of composition would more or less settle the question of what we are271, accepting brutal composition would not. It would tell us next to nothing about what we are272. In fact those versions of brutal composition that look more attractive than minimalism would make the question of what we are273 very hard to answer. It gives us too many sorts of things that we could be. Despite its initial attraction, then, brutal composition is not so appealing when all is said and done.
  13. If brutalism is false, then it seems that we really can find out what we are274 by finding out when composition occurs. Alternatively, we can find out when composition occurs by finding out what we are275. Or we can try to work out both together. I won’t venture to say which procedure is the best one. In any case, the connection between the two questions ought to make progress on both easier.




In-Page Footnotes

Footnote 78: Or maybe four: inspired by Geach (1997: 310), someone might propose that the animal and its brain are the same thinker, even though they are different material things. Animal and brain are not numerically identical, but neither are they numerically distinct. The proposal is that there is no such thing as numerical identity without qualification, but only a lot of “sortal-relative” identity-like relations. This is a variant of the way of funny logic of §7.4. I find it more or less unintelligible. Consider that one thinker that your animal body and your brain are, on this proposal. How big is it? No satisfactory answer to this question appears to be possible. But surely a material object has to have some size.

Footnote 165: “Special” as opposed to a "general" composition question that does not concern us here. The earliest statement of the special composition question that I know of is Hestevold 1981; see also van Inwagen 1987 and 1990a: 21-32.

Footnote 168: I suppose someone might say that animals are mereologically simple. That may have been Aristotle’s view, and Lowe’s Brobdingnagian atomism (§7.8) is close to it. Combining this with compositional nihilism might enable animalists to dispense with the troublesome entities. It’s a pretty wild idea, though.

Footnote 185: Or at least all the metaphysical worries we have considered. I don’t mean to imply that there are no others. There are always more worries. Most notably, perhaps, there are serious worries to do with “metaphysical vagueness” (see Sider 2001: 120-139, 148-150; van Inwagen 1990a: §§18-19). A discussion of these matters is beyond the scope of this book. In any case they are not unique to animalism.

Footnote 186: More precisely, things compose something if and only if they are either functionally united or they and all their parts are rigidly bonded. It seems that two organisms firmly glued together would be rigidly bonded, yet no one but a universalist, and certainly not Hoffman and Rosenkrantz, would want to say that they would compose something. Another caveat: for technical reasons that we needn’t go into, Hoffman and Rosenkrantz do not actually attempt to give a theory of composition, and the view stated here is extrapolated from their discussion of other matters. I believe that it accurately reflects their intent, however.

Footnote 187: This paragraph has been re-worked in the book.

Footnote 194: It is probably more accurate to say that they deny the existence of lumps of wet clay, because their particles would not be rigidly bonded. But the example, and those that figure in other arguments for constitutionalism, can be modified so as to involve only rigidly bonded particles.

Footnote 219: If you have doubts about whether elementary particles really persist, think of atoms.


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 - Oct 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