Mereology and constituent ontology

I’ve just realized that one can motivate belief in bare particulars as follows:

1. Constituent ontology of attribution: A thing has a quality if and only if that quality is a part of it.

2. Universalism: Every plurality has a fusion.

3. Weak supplementation: If x is a proper part of y, then y has a part that does not overlap x.

4. Anti-bundleism: A substance (or at least a non-divine substance) is not the fusion of its qualities.

For, let S be a substance. If S has no qualities, it’s a bare particular, and the argument is done.

So, suppose S has qualities. By universalism, let Q be the fusion of the qualities that are parts of S. This is a part of S by uncontroversial mereology. By anti-bundleism, Q is a proper part of S. By weak supplementation, S has a part P that does not overlap Q. That part has no qualities as a part of it, since if it had any quality as a part of it, it would overlap Q. Hence, P is a bare particular. (And if we want a beefier bare particular, just form the fusion of all such Ps.)

It follows that every substance has a bare particular as a part.

[Bibliographic notes: Sider thinks that something like this argument means that the debate between constituent metaphysicians overlap bare particulars is merely verbal. Not all bare particularists find themselves motivated in this way (e.g., Smith denies 1).]

To me, universalism is the most clearly false claim. And someone who accepts constituent ontology of attribution can’t accept universalism: by universalism, there is fusion of Mt. Everest and my wedding ring, and given constituent ontology, the montaineity that is a part of Everest and the goldenness of my ring will both be qualities of EverestRing, so that EverestRing will be a golden mountain, which is absurd.

But universalism is not, I think, crucial to the argument. We use universalism only once in the argument, to generate the fusion of the qualities of S. But it seems plausible that even if universalism in general is false, there can be a substance S such that there is a fusion Q of its qualities. For instance, imagine a substance that has only one quality, or a substance that has a quality Q₁ such that all its other qualities are parts of Q₁. Applying the rest of the argument to that substance shows that it has a bare particular as a part of it. And if some substances have bare particular parts, plausibly so do all substances (or at least all non-divine substances, say).

If this is right, then we have an argument that:

5. You shouldn’t accept all of: constituent ontology, weak supplementation, anti-bundleism and anti-bare-particularism.

I am an anti-bundleist and an anti-bare-particularist, but constituent ontology seems to have some plausibility to me. So I want to deny weak supplementation. And indeed I think it is plausible to say that the case of a substance that has only one quality is a pretty good counterexample to weak supplementation: that one quality lacks even a weak supplement.

From bundles to bare particulars and back again

Here's a compelling narrative. Start with the bundle theory of substance: substances are nothing but bundles of properties. Then observe that this suffers from serious problems. If a substance is nothing but a bundle of properties, it is unclear how a substance could have had other properties than it does. Further, intuitively it should be possible to have two indiscernible substances--ones with all the same properties. This motivates a move to bare particular theory. According to bundle theory, substances were constituted by one kind of thing: properties. Bare particular theory makes substances be constituted by both properties and a special entity, the bare particular. Introducing the bare particular solves the modal problem, since we can say that the identity of substances is grounded in the identity of the bare particulars, so you can have a substance in one world with different properties than the very same substance in another world, as long as the same bare particular is found in both. Further, there is no difficulty with indiscernibles, as long as you have two bare particulars.

Note that this narrative isn't quite the standard narrative about bare particulars. The standard narrative introduces bare particulars to solve the problem of predication, by making the bare particular be the subject of predication. That standard narrative, however, falls prey to a problem that Andrew Bailey points out: we don't want to say that the bare particular has the ordinary properties of the host substance (for then we get reduplication), but if it does not, then it's not the subject of predication.

So let's stick to my from-bundles-to-bare-particulars narrative. But at this point there is a really interesting move possible, one that was pointed out in my undergraduate metaphysics class by a brilliant freshman, Rose Brugger. According to bare particular theory, substances are constituted by two kinds of things: properties and a bare particular. But Brugger suggested that we take the bare particular to just be an individuating property. Namely, a haecceity.

The result is a really interesting theory. It is a kind of bundle theory. However, first, the motivations for bare particular theory continue to be satisfied: we can ground identity between substances in identity of the haecceity. Second, we solve the puzzle of the mysterious "bareness" of the bare particular: the haecceity isn't some weird propertyless individual, but just a property, albeit a special one. Third, the resulting theory is more parsimonious, because it posits one fewer fundamental category: all it needs are substances and their constituent properties, without a separate category of bare particulars.

The resulting theory is superior to both standard bundle theory and standard bare particular theory, being only slightly more complex than standard bundle theory but solving a number of problems.

Nonsense and "that..." clauses

Suppose the theory of bare particularism is wrong. It is plausible that if it's wrong, it's not that its central claims are false, but rather its central claims are nonsense. It is not so much that "There are bare particulars" is false, as that it fails to express anything. Maybe you're not convinced by this particular example, but if so there are probably some others that you'll find convincing. I suspect that many theories in ontology are such that either they're true or they're nonsense, and they aren't all true. Platonism and trope theory are like that, for instance. I'll use bare particularism as my stand-in for such a case.

Yet we have no hesitation in saying things like:

1. Sally believes that she is partly constituted by a bare particular,

when Sally is a bare particularist. This should trouble us. First of all, analytic orthodoxy holds that in "x believes that s" sentences of this sort (but not in "Sally believes that scientist"!), the "that s" clause refers to the proposition that the sentence "s" expresses.

We could set this orthodoxy aside, and instead of parsing "x believes that s" as predicating a relation of belief between x and the proposition that s, we could take "x believes that" to be a sentential operator. This leads to problems with quantification ("Sally believes some of the things she was just told"), but perhaps those can be solved in some way. But even if we set the orthodoxy aside, we have another problem with (1). We have a sentence that contains a component, outside of quotation marks, that is nonsense, viz., the phrase "bare particular".

The simplest solution to the problem is just to take (1) to be elliptical for some metalinguistic claim like

2. Sally believes that the sentence "Sally is partly constituted by a bare particular" is true.

Or at least, perhaps, that's the charitable way to take (1). Suppose we do that. Then we have the following oddity. Suppose you and I disagree about whether Sally is a bare particularist. You happen to be a bare particularist yourself, but you doubt that Sally is one. So I say (1) while you say:

3. It is not the case that Sally believes that she is partly constituted by a bare particular.

Suppose my use of (1) is elliptical for (2). But your use of (3) is surely not elliptical for the negation of (2), since you have no qualms about bare particularism, and you have no reason to make a metalinguistic claim instead of simply attributing a propositional belief to Sally. So your use of (3) is literal, while my use of (1) is elliptical. But then our claims are not directly contradictory. Maybe that isn't a big deal. And maybe your claim, despite your best intentions to the contrary, is in fact metalinguistic, because the reference magnet in the vicinity of your statement is the proposition expressed by (2).

Another problem with reading (1) as (2) is that it is odd to attribute to Sally beliefs about bits of language. What if Sally thinks, for some good or bad reason, that there are no sentences? Again, maybe there is a reference magnet solution.

A hint of a different solution is provided by this post. That post suggests that there is something more fundamental in the mind than beliefs. There are "doxins", which place constraints on what beliefs are to be attributed to one. It may well be that when Sally accepts bare particularism, she isn't believing any proposition like "that she is constituted by a bare particular", but rather she has the doxin expressible by "The credence of the proposition expressed by 'I am constituted by a bare particular' shall be high." If in fact there were such a proposition, this doxin would allow her to be credited with belief in it. There not being any such proposition, we can't credit her with belief. Rather, we credit her with a doxin that carries a false presupposition, viz., that there is a proposition expressed by "I am constituted by a bare particular". The false presupposition, however, isn't a belief. So she can have that doxin while yet not believing in sentences and the like. There would need to be a lot of work done to defend this.

The issue comes up not just for belief. For instance, one might have a desire that "involves a bare particular". Then one would bring in orektins, from the same post.