Against Restricted Composition
Are you looking at a screen right now, or just some pixels arranged screen-wise? I am here to give you the answer!
We usually take ourselves to have quite good grasps of what ordinary objects there are. Here is a computer, there is a table and so on. But to what extent are our everyday judgements about these things true? More specifically, under what conditions does a plurality of objects compose a further object?
This post is really a preliminary to a future post. It is meant to lay some groundwork for a little argument I have been thinking about. For this reason it will mostly be stuff that has been said many times before, and if you have read the SEP article on ordinary objects, there won’t be much new stuff here, but I thought I would make my own case. And of course stay tuned for the next post:
First off, I want to make a distinction about what we are actually talking about. It can be easy to confuse the ontological question we are actually interested in with semantic questions. To make this clear, let’s take the inspiring example of a banana. We can ask the question “do bananas exist?” This may seem like a question that is relevant to our discussion, but I don’t think is. When we ask this question we are basically asking whether there is some state of affairs that satisfies the definition we have in mind for the word “banana” (I argue that there is no answer to the question). But defining a word is not really progress in ontology.
Imagine now that you are holding a banana (whatever that means) in your hand. Now ask the question “does that (pointing at the banana) compose an object?” This is what we are interested in - whether there are actually composite objects that exist beyond our minds and language.
The positions
There are generally three main types of views in mereology:
Compositional Nihilism - There are no composite objects.
Unrestricted Composition - For any two or more objects, they compose a further object.
Restricted Composition - There are certain cases where two or more objects compose a further object, and certain cases where they don’t.
I will refer to these as nihilism, universalism, and particularism respectively. Out of these, particularism is the commonsense position: Cars and turkeys exist, but turcars don’t (the composite between the front half of a turkey and the back half of a car). My position is, however, that particularism is false, and that it is somewhat of a tossup between nihilism and universalism.
No Successful Principle
So my first point is simply that no one seems to have been able to come up with a good principle for when objects compose further objects. We might think things compose an object when they touch each other. This would however mean that my underwear composes an object with you (assuming we are both touching the ground right now, and that I am wearing underwear). At the same time, it would mean that our solar system is not an object. Furthermore, it is not even clear what it means for things to touch, when you get down to the microphysical level.
But I think the real problem is that it is not even really clear what it would mean for a particularistic principle of composition to be successful. Should it just conform to our commonsense ideas of when things compose an object? It is not obvious why this should be a reliable guide to reality. Suppose you Showed a caveman or an alien a computer like this:
Most modern humans would think that there is a single object there: A computer. It is not clear that someone who isn’t used to using computers would think that this would constitute a single object, rather than four separate objects. But suppose you convinced them that it really is a single object. You then show them this:
Naturally they would assume that this was also a single object. Silly them! This is clearly two shoes which happen to be tied together. What explains this difference? I think that there is a simple explanation: In our everyday life it is often useful to think of computers as single objects, but it is rarely useful to think of shoes tied together as single objects. But why should we think that what we find useful in 21st century human society constitutes the objectively right principle of composition? We shouldn’t. So it looks as though even if we were to find a neat principle, which fit all our usual intuitions about composition without counterexamples, we shouldn’t think that it was the right one. But what reason is there then to look for such a principle?
I of course also think that there is no consistent principle which neatly fits our intuitions about composition, and I will give a couple arguments to that effect.
Vagueness
Imagine that you are screwing two wooden planks (James and Rick) together. At the start they will, according to most versions of particularism, not compose an object, but afterwards they will (call it Jack). This raises the question: At what point does Jack come into existence? It is not clear that there is an answer to this. Suppose you say that it is when a screw (Jill) is screwed in. This still leaves a lot of room. How far in must it be in before it starts to compose Jack? Suppose you pick a point. Why do they compose Jack then, and not when Jill was 0.000000001 nm further back?
This is of course a sorites argument:
James and Rick do not compose Jack when Jill is only screwed 1mm in.
If James and Rick do not compose Jack when Jill is only screwed 1mm in, then they do not compose Jack when Jill is screwed 1.0000001mm in.
If James and Rick do not compose Jack when Jill is only screwed 1.0000001mm in, then they do not compose Jack when Jill is screwed 1.0000002mm in.
…
Therefore James and Rick do not compose Jack when Jill is screwed all the way in.
This can also be run in reverse to show that James and Rick always compose Jack. So it looks like there is an inconsistency in particularism: The view says that James and Rick don’t compose Jack at one point, and at another point they do. But the argument shows how this cannot be the case.
There are of course possible responses. One is to simply deny one of the premises by saying that there is some exact cutoff point. This is, however, very implausible, since it would mean that some seemingly arbitrary difference of 0.0000001mm can make the difference for when Jack arises. This is of course not impossible, but it seems to make the rules for composition somewhat arbitrary and brute - there is not really going to be a principled difference between two adjacent cases with a difference of 0.0000001mm, and so it seems like there is just going to be no actual reason why James and Rick compose Jack in the one case and not in the other, other than that just is the way it is.
Another tempting option is one that is quite common in semantic versions of sorites paradoxes: Just saying that there are indeterminate cases of whether the concept in question applies. So consider the common example of baldness. It looks as though a case of someone with 50 (or 100 or whatever) hairs on their head is just a case where there is no answer as to whether they are in fact bald. But there are still cases where someone is definitely bald or definitely not bald. So there is just a range of cases where it is indeterminate whether someone is bald (I personally don’t think this response works in semantic cases either).
The problem with this in the case of composition is that it would mean that it is often indeterminate how many objects exist. When we have James and Rick there are two objects, but when we have James, Rick, and Jack, there are three objects. Thus, on this response, there will sometimes be an indeterminate number of objects in the world. But it is surely absurd that there is not an answer to how many objects exist in the world. This is especially bad for particularism, since one of the big motivations for the view is that it accords with our everyday experience. So if the view ends up entailing an absurdity like this, it looks as though we lose the thing that made the view appealing in the first place.
It is not like this vagueness is a rare phenomenon either. Take a saw and begin to saw any object in half, or begin to chip away a small piece from the object. Near the end of the process it will be unclear whether it is still one object or not.
It is of course clear how nihilists and universalists will respond to this argument: Nihilists will accept the conclusion and universalists will deny premise 1.
The Problem of the Many
Return to the example of a banana (Ben) in your hand. On particularism, we want there to only be a single banana in your hand. But consider now Ben-minus. This is Ben except for a single cellulose molecule (CM). If you were to remove CM from Ben, you would still say that there is a banana in your hand. So it looks as though Ben-minus is also a banana. But this raises a problem: We want to say that there is a single banana in your hand, but we have just shown that there are at least two bananas in your hand (namely Ben and Ben-minus). How do we resolve this?
It is clear how both nihilism and universalism resolve it: Nihilism just says that there are particles arranged banana-wise in your hand, but no actual banana. Universalism says that there is an incredibly high amount of bananas in your hand, and so denies that there is just one.
A particularist might say that a banana is simply the largest group of particles in the object, and so Ben-minus only becomes a banana once CM is removed. Before that it is just a collection of particles which is part of a banana. But now suppose that someone takes an extremely small pair of tweezers and slowly begins to pull CM off of Ben. As we saw earlier, there will not be a good answer as to when Ben ceases to be the banana in your hand and it becomes Ben-minus instead. But when it is unclear which is the banana in the hand, then it is surely true that if Ben is a banana then Ben-minus is a banana. And thus there will be either two or zero bananas in your hand. Both of these answers are of course incompatible with answer that the particularist wants to give: That there is exactly one banana in your hand.
How Crazy Are These Views Really?
When you are first presented with nihilism and universalism, they can sound quite wild or crazy. But I think the more you consider them, the less crazy they really appear. The longer you think about it, the less obvious it becomes that certain arrangements of matter should have some sort of ontological priority over other sorts of matter (at least to my mind). It is also not very surprising that humans would want to prioritize some arrangements over others, even if there is no metaphysical basis for it. There are clearly some arrangements of matter (like computers and bananas) that are more useful to our survival and goals than others (like turcars and banana-minuses), so even on these revisionary ontologies, it is not very surprising that we intuitively think that some objects compose further objects and some don’t. And so I don’t really think these views makes a mess of our intuitions.
I also think we can easily make sense of our talk of composite objects on these views. When I say that there is a single banana in my hand, I don’t mean so in the strictest sense possible. But in everyday life, you don’t really use words in the strictest sense possible (unless you are an obnoxious asshole, or making a joke). So when I have finished drinking a glass of wine, and I say “the glass is empty”, you can tell from the context that I don’t care whether there are still a few drops of wine and some air left in the glass. Likewise, when you ask me “can you give me a banana?”, I know that we are not currently doing metaphysics, and so I don’t say “dId YoU mEaN pArTiClEs ArRaNgEd bAnaNa-WiSe?”, even though that might technically be the right way to say it.
The Toss-up
So now I think we are left with two live possibilities: Universalism and nihilism. Which should we prefer? There is not really, I think, much of a difference between universalism and nihilism in what they result in, since composition is a quite “pointless” property on universalism, because it just applies to everything, much like it is “pointless” on nihilism, since it applies to nothing. So it is hard to give a very principled reason for choosing one over the other.
There is one major difference. On nihilism you can’t have gunky worlds (worlds where there are no smallest parts - everything is a composite object), since there would be no particles to be arranged x-wise. On the other hand you couldn’t have junky worlds on universalism (worlds where every object is a proper part of another object), since there could be no composite object consisting of all objects (that would only be an improper part of itself and not a proper part of anything else). It is easier to imagine a gunky world than a junky world, I think, so this might be more of a strike against nihilism than against universalism.
Universalism does have the problem of inflating our ontology a lot more than nihilism. But the way it does so is very simple, and so not a very big cost to the theory.
I personally feel more inclined towards nihilism than universalism, although I am not sure how much this is just due to a gut feeling rather than a good reason. Although I will say that it does feel unnecessary to postulate the concept of composition, if it has literally no explanatory power with regards to anything, and so unless there are cases where composition provides an explanation for something, we should favor nihilism.
All of this might seem a bit pointless and appear more like a philosophers-circlejerk than something really substantive to some people (I personally do think it is a fun and interesting topic). But I promise that things will get a bit more interesting in the next post.