In the previous post, I discussed modal arguments for substance dualism. Here I defend two more arguments. Both have to do with mereology, the study of the relation between parts and wholes. Mereology deals with the nature of composite objects, objects which are composed of parts. The first argument is, as far as I know, more or less original to me, though it relies heavily on terrain mapped by others. The second comes mostly from the philosopher Peter Unger. (Like Chalmers, Unger is an atheist, so again, you don’t need to worry about a hidden agenda.)

The First Argument

Some mereological questions have already come up in these posts, such as whether a certain composite object (my brain) can survive being divided in the way described in the transplant scenario, or whether a different composite object (my body) can survive having its parts replaced incredibly quickly. So far I have talked, for the sake of argument, as though there are substantive, objective answers to questions like this. But you might think this is a mistake. Maybe you think: “Look, we know what happened in the transplant case: you took some matter and put it over here, and some other matter and put it over here. And that’s all there is to know. Talking as if there’s some further question about whether ‘my brain’ survived, understood as some object over and above this matter, is a confusion, a type of reification: it involves taking certain ordinary ways of talking too literally.”

If you think this, good. I’ve got an argument for you.

Our ordinary view of mereology has a few components. One is that, when we have a number of distinct substances, they sometimes compose an additional, composite object, and sometimes don’t. So, for instance, people think that the matter which is currently in my body did not compose anything five hundred years ago, when it was scattered all around the earth, but that it currently does compose something over and above the individual bits of matter (namely, my body). They think there’s some matter in front of me which composes a new object, my desk, but they don’t think that my body and the desk together compose an object (even when they are touching), or that the desk and the paperweight on it do, or that the left half of my brain and the moon do. In a similar vein, they think that, when a composite object does exist, it can survive some changes to its parts and not others. For instance, they think that a building can survive having a window replaced, but not being blown up and rebuilt with entirely new materials. And they think my body could survive the loss of a single cell, but not that it could survive having its material return to the scattered-around-the-earth state it was in five hundred years ago.

A final component of the commonsense view of mereology is that these facts about composition are objective and mind-independent. That is, people don’t think the parts of my desk compose an object, while my brain and my desk don’t, because we decided that that’s true. Rather, they think it’s just how things are. They think the same of the claim that some things are furry (my cats’ tails) and some things are both photovoltaic and on Mars (the solar panels on a Mars rover) but that nothing has one part which is furry and another part which is both photovoltaic and on Mars (a composite object comprised of both my cats’ tails and the solar panels on a Mars rover).

However, it’s extremely hard to spell out exactly what the principles governing composition are supposed to be on the commonsense view. For instance, you might think it has something to do with the parts of a thing being attached, but attachment is neither necessary nor sufficient for commonsense composition. On the one hand, intuitively, gluing my forehead to the desk does not mean that I and the desk now compose an additional whole. On the other hand, a large pyramid made of stones does intuitively seem like a composite object, even if the stones are not attached with mortar and are instead just resting on one another. Similarly, someone might think a hill is a composite object, even though the bits of dirt in it are not attached–it’s really just a pile of dirt. But then, they probably wouldn’t think that about a pile of clothes. Further, the commonsense significance of attachment may differ depending on the context. If I glue a book to a desk, can whether they compose something be affected by my putting them in an exhibition and calling them a single work of art?

In fact, reflecting on commonsense mereology leads many philosophers to think that some part of it has to be false. Our ordinary judgments about mereology seem too complicated and arbitrary, and too closely related to which ways of categorizing things we happen to find useful, to be credibly understood as expressions of mind-independent truths. Various famous puzzles help bring this out. One is the Ship of Theseus. Theseus’ ship is getting old, and each day he replaces one of the planks. Typically people don’t think that changing a single plank would destroy the ship. But by the end of the year, all of the original parts have been changed out. Is it still the same ship? If not, at what point did it cease being the same ship? If so, suppose I was following Theseus and gathering up the planks, and using them to build a new ship of my own. At the end of the year, I have a ship identical to the original, including being made out of the same parts. Why isn’t that a better candidate for being the original ship? Or are they somehow both supposed to be the original ship–even though Theseus and I are sailing them in different directions? One starts to worry that any answer will look arbitrary.

Another famous problem concerns the statue and the clay. Here’s a clay statue. How many composite objects exist in the space it takes up? People ordinarily think the answer is “one”–the statue exists there. But other ordinary judgments seem to suggest that there are actually two composite objects there, the statue and the lump of clay which composes it. Consider: if we smush it flat, the statue is destroyed, but the lump of clay still exists. But one and the same thing cannot exist while also ceasing to exist. So they must be separate objects, with the clay constituting the statue, right?  But this view itself faces serious problems. For instance, it seems problematic to think that two things can have exactly the same parts while being distinct objects, and it seems hard to explain why the statue and clay have different modal properties if they share all the same parts. Further, not only does it seem odd to think that the statue and the clay are two overlapping objects rather than just one, it also seems hard to avoid an infinite profusion of overlapping objects. What about the group of fundamental particles making up the lump of clay, which would still exist even if we tore apart the molecules in the clay–do they compose their own, third object? Or consider: we are presumably tempted to say that the statue comes to exist when the sculptor molds it into that shape, but that nothing new comes to exist when the shape of a lump of clay is just changed slightly in a way with no significance to us. But this is just because we care about statues. Suppose members of some other species cared just as much about, say, whether the lump is more than three inches tall: is there also a separate “taller-than-three-inches-lump” which is destroyed after the statue but before the clay is pounded flat–and so on for every other height in between?

So, my own view is that commonsense mereology–understood as a view which attempts to incorporate our ordinary judgments about which composites do and don’t exist while also understanding these judgments as expressing objective, mind-independent truths–has to be false. There are three main alternatives to it, more or less. One is mereological nihilism. Where moral nihilists claim that nothing is really right or wrong, mereological nihilists claim that there aren’t really any composite objects–composition never occurs, and none of the puzzles above even get off the ground. This means that the only objects which really exist are simple ones not composed of any parts. In the case of physical objects, these “physical simples” might be whatever the simplest particles are, or maybe point-sized bits in a field, or whatever. On (my favored version of) this view, putatively composite objects are really just “simples arranged desk-wise” (or however else-wise); there is no additional thing, the desk, which exists as something over and above the parts. Speaking as if there are really desks is fine for ordinary purposes; it is just a loose way of speaking, and accurate by the standards appropriate to ordinary language. (To use a common example, in the same way, I think I say something true if I say that the sun has moved below the horizon, even though it was really the earth, not the sun, that moved.) It’s just important that we not let this way of talking confuse us when we start doing philosophy and trying to speak strictly.

A personal story

The first time I ever thought seriously about mereology was when I went to a talk about the Ship of Theseus as a sophomore. (Incidentally, this was also the first time I ever heard the name “Peter van Inwagen.”) By the end of the talk it seemed obvious to me that mereological nihilism was true, and it has basically seemed that way to me ever since. It is a very simple view, and giving up the view that there are really, literally composite objects never seemed to me like a big deal. Further, it immediately solves the puzzles about composition mentioned above. Neither the statue nor the clay really exist. There are just simples which can be put into different arrangements. No need to worry about overlapping objects and all that mess. Similarly, Theseus’ ship doesn’t really exist; the story is just about rearranging physical simples, and that’s all there is to it. These puzzles are just pseudo-problems, consequences of our taking ordinary ways of talking too literally.

However, there is at least one very prominent objection to nihilism. Surely I exist–really and literally. But if I am some sort of physical object, like my body or brain, I am composite. So, assuming physicalism, there must be at least one composite object, and nihilism is false. One option (favored by van Inwagen) is modifying nihilism to produce organicism, the view that the only composite objects are living organisms. So tables and chairs don’t really exist, but you and I, as human animals, do. This has never seemed plausible to me: I think it’s ad hoc, and I worry that it leads to problematic forms of vagueness while allowing the same puzzles we were initially worried about to re-arise in a different form. So another nihilist option–and I think the more popular one–is just to deny that I (or you) really exist (see sec. 7 here). But that doesn’t seem right!

You can see where this is going. The dualist can say that I am not a composite object, but am instead a simple, immaterial soul. We thus reconcile a plausible view of mereology with the obvious fact that I exist. (Briefly, another fun connection between dualism and mereology: many of the main arguments against substance dualism are all equally good arguments for nihilism. Accordingly, people who reject dualism on one of these grounds have also reason to be nihilists. But it turns out that being a nihilist should lead one to be a substance dualist after all.)

A second view is what we might call mereological subjectivism. This rejects the mind-independence part of commonsense mereology in favor of the view that the principles governing composition somehow depend upon what we think, or decide, or are disposed to say, or whatever. In this way, it is analogous to ethical subjectivism. This view also has its attractions. It allows us to uphold commonsense views about when composition does and doesn’t occur. Further, on this view, the fact that the principles of composition seem so closely tied to what we happen to care about no longer seems like an objection; of course they are, since we are responsible for the principles having the content they do. Further still, at least the worry about a profusion of objects in the statue and the clay case is solved: there aren’t “taller than three inches” lumps and the like because we don’t care about them. And finally, even some of the elements of arbitrariness or vagueness which arise in the course of mereological puzzles may not be so bad on this view: since the principles of composition are things we just made up, maybe it isn’t surprising if they break down in unusual cases, or have silly or arbitrary elements.

However, if we are composite objects, mereological subjectivism also causes problems. One issue is circularity. If we are composite objects, and composite objects depend for their existence on us, then how did the first one of us come into being? (One answer, suggested by my old colleague Andrew Brenner in his paper “How to be a Mereological Anti-Realist,” is that God decided that our parts would compose a whole. Of course that option won’t be attractive to atheists.) Another issue is just that it doesn’t seem true that my existence, or yours, depends upon human activity in this way. (Consider a new, more humane method of execution: instead of lethal injection, we simply wipe murderers from existence by decreeing that their parts no longer compose a whole.) ^[1] On the other hand, subjectivism avoids these problems if we are not composite objects, but are instead simple, immaterial souls.

The final major view on offer is mereological universalism. It asserts that composition always occurs–for any two things, there is a new whole which they compose. It therefore rejects commonsense mereology, but by departing in the opposite direction from the nihilist. On this view, there really are objects that are partly furry and partly photovoltaic and on Mars, and they exist in just the same sense as other composite objects. Mereological universalism has certain advantages. For instance, it allows us to say that the objects we ordinarily think exist–my desk, etc.–really do exist, but while avoiding charges of arbitrariness or anthropocentrism (composition always occurs, but we only take note of it when the composite object is somehow significant to us). Using universalism to answer the puzzles about the Ship of Theseus and the statue and the clay takes some additional doing. I won’t go into details, but the usual move is to conjoin it with belief in so-called temporal parts, and even then some additional work is needed to handle objects which share all their temporal parts, but don’t do so essentially. These moves are controversial, but I’ll grant them here  The most pertinent fact about universalism is that it easily allows for us to exist as composite objects.

However, mereological universalism also has some problems. It fits less well with commonsense mereology than does subjectivism, ^[2] and not really any better than does nihilism–it countenances the composites we usually think exist, but also all kinds of composites we usually think don’t exist. At the same time, it’s less simple than nihilism: it requires an additional, mind-independent relation, composition, along with all kinds of attendant objects. (I will also note, again, that nihilism has seemed obviously true to me every time I’ve ever thought about it. Admittedly, this isn’t really an argument unless you trust my judgment.)

So: I’m prepared to say that all three of the views we’ve discussed here avoid at least the major problems faced by commonsense mereology. However, universalism is the only view really able to accomodate the claim that we are composite objects, and universalism faces other disadvantages as compared to nihilism and subjectivism. So I think this fact provides support for the view that we are not composite objects, but are instead simple, immaterial souls. Further, we are about to see that mereological universalism strengthens the second mereological argument.

The Second Argument

Here’s a second mereological argument. Its general structure was developed by Peter Unger, though the exact spin I will place on the argument here also owes a debt to Hedda Hassel Mørch. Consider that some properties are intrinsic while others are extrinsic. Something’s intrinsic properties have to do, roughly, with how it is in itself, while its extrinsic properties have to do, roughly, with how it relates to other things. So, for instance, my computer’s mass is an intrinsic property; it has to do with how the computer is, in and of itself. On the hand, “being located in the same house as two cats” is an extrinsic property of the computer. When the cats leave the house, there is some sense in which this affects the computer–it loses the property “being located in the same house as two cats.” But there is another, fairly clear sense in which this change is not a change to the computer, taken in and of itself. Of course, this is not to say that the cats can’t causally affect the which intrinsic properties the computer possesses. They could knock it over and break it, for instance. Indeed, their mere presence results in a very slight gravitational effect on the computer. But the point is that the cats affect the intrinsic properties by the computer by causing changes in how it is in itself; their being in the house is not itself part of how the computer is in itself.

Notice that whether something is conscious seems like an intrinsic property: as Mørch notes, “Intuitively, one cannot change whether someone is conscious merely by changing the properties of other things,” unless those changes in other things somehow cause a relevant change in the person themselves. Notice also that, intuitively, there is only one conscious subject associated with my body; there is not, so to speak, a colony of billions and billions of subject sitting in this chair, thinking my thoughts, having my experiences. However, these two judgments–about the intrinsicality of consciousness and the existence of just a single subject here–may pose a serious threat to materialism. Here’s why.

Suppose I am some physical object–my brain, perhaps. In that case, presumably, the reason why my brain is a conscious subject will have something to do with some macro-level feature of how its parts are arranged–with its information-processing functions, or whatever. But notice that whatever this macro-level property is, it would not lose this property if some tiny part of it–a single neuron, say, or a single molecule–were suddenly teleported away while everything else was left as is. There is no single particle which is such that its removal would prevent the existence of a conscious subject (just as there is no single pixel on your screen which is such that its removal would prevent this from being a legible sentence).

But now consider Brain-, which consists of all the matter in my brain except for a single particle. Why isn’t Brain- also conscious, constituting a subject separate from me? We agreed that my brain minus that particle would have whatever macro-level physical property allegedly amounts to being a conscious subject, so it seems that Brain- does have it. And if consciousness is an intrinsic property, then it seems like whether Brain- is conscious should not depend on anything outside itself–such as whether the additional particle is actually present. So it now looks like Brain- should in fact be conscious, alongside my brain. And so should all sorts of other physical objects which are just like Brain-, except with a different particle missing, or a few particles missing, or whatever. So it looks like materialism, together with the intrinsicality of consciousness, commits us to saying that there are in fact an unfathomable number of subjects present in the body of any conscious being.

One possible response to this is that Brain- is not really a distinct object, so that there is nothing to experience these conscious properties. There is just my brain; while its parts form a new composite object, the matter in Brain- does not. But notice that this response is clearly ruled out by mereological universalism, which holds that absolutely any set of objects are such that there is a further object composed of them. And we saw in the first half of this post that mereological universalism is (I claimed) the best mereological theory which doesn’t lead more-or-less immediately to substance dualism (at least on the assumption that we exist). So if we avoid the arguments of the previous section by adopting mereological universalism, we can’t subsequently avail ourselves of this objection by claiming that some particles do not form composites.

On the other hand, consider substance dualism. It is of course compatible with the thought that there are billions and billions of subjects associated with my body–I suppose there could, in principle, be any number of souls causally related to it. But it also allows us to reconcile the more intuitive view is that there is only one subject with the view that consciousness is an intrinsic property. For on substance dualism, neither my brain, nor Brain-, nor any other physical objects are themselves conscious. Rather, consciousness is an intrinsic property of my soul. These physical objects stand in certain causal relations to my soul and affect consciousness by causing changes to how my soul is in and of itself. It is no violation of intrinsicality to say that my soul stands in causal relations with a large cloud of many different, slightly overlapping brains (if we want to say that). It might be that my soul, receiving input from vastly many slightly different physical objects, nonetheless constitutes a single subject of experience, just as (to borrow an example from Unger) a television set receiving signals from vastly many slightly different transmitters might nonetheless (if set up in the right way) display a single, clear image. Or as Mørch puts the point, the problem here “could also be solved by adopting the dualist view that consciousness is causally produced by purely physical systems… This would preserve INTRINSICALITY because nothing prevents intrinsic properties from being caused by, as opposed to identical to, extrinsic properties.” (She in fact rejects the dualist option due to worries about mental causation. I will address some of those in my next post.)

Footnotes