Compositional Vagueness in Material Objects

I recently attended a conference in Nottingham entitled 'New Directions in Metaphysics', hosted by the University of Nottingham. The venerable Dean Zimmerman was there, giving a talk on property dualism. His talk didn't excite me too much, but when he brought up the well-known problem of the vagueness of composition, it got me thinking again on these questions.

To refresh your memory, the problem of the vagueness of composition arises if we consider a reasonable definition of a material object - say, a human being. Suppose that a human being is defined as being that object that is composed of all of these material elements - say elements a1-an. This seems fairly straightforward, until we consider that it is very hard to, as it were, draw a line around those elements that compose us and those that do not. In fact, when we get a very detailed, close-up view of these elements, it becomes more and more obvious that there are many different elements that are just as entitled to be a part of an object as the ones that we at first chose to countenance. That is, there are many different sets of elements that are up to the task. Or consider another analogous problem: take all of the elements that one might say composes a particular human being (a1-an) and ask yourself whether or not it would be acceptable to define this human being by excluding merely one of these elements - say a3. Repeat and repeat the thought experiment until you reach absurdity.

This seems like a fairly important problem for defining material objects - for once the thought experiment is carried out, there seems to be an extremely large group of acceptable candidates, none of which seem to have any special claim to being the set that defines a particular material object.

An interesting response to this problem is the one given by Peter vanInwagen, who claims that the only material objects that exist are those that are biological - and they are individuated by being a particular group of elements all participating in the biological functioning of an object. Any of those fringe elements that made-up other candidates won't be acceptable then, as they are not aiding in the biological functioning of the object in question - they are, as it were, just idly floating about. Of course, a well known problem for this kind of view is that it entails that things like 'tables' and 'chairs' don't exist. It's a bullet that not many have particularly wanted to bite.

For my part, I think vanInwagen's response is on the right track in a certain respect, in that it individuates material objects using a certain dynamic criterion. It seems to me that the reason that the 'many candidate' problems all get started by assuming that the individuation of a material object must be a merely static - in the case of Lewisian examples, a particular set of elements. There is, of course, something to be said for wishing one's identity criterion remain stable - for instance, you don't want it to be the case that the identity criterion for a material object is constantly changing full stop. But the set-theoretic element identifications are inadequate, I think, because they fail to have any dynamic qualities at all.

I suppose the reason for this is mostly historical . Couple the fact that analytic philosophical methods have always wanted identity to be a stable, unchanging primitive fact with the further fact that the concepts of mathematics have always been considered the most unchanging facts and there's good reason to follow this kind of idea through. But it's hard for me to take any of this seriously in light of the current state of metaphysics, physics and the developments in quantum mechanics. As far as I can tell, all the evidence in quantum mechanical physics points towards defining physical objects as dynamical - the wavefunction is certainly, if anything at all, a dynamic representation of reality. If the wavefunction does represent something physical, it represents something that is not merely 'categorical', and certainly doesn't represent it as simply a set of particular elements; think here only of the fact of wave-particle duality! And metaphysics too, riding the crest of scientific discovery, is edging more and more towards an ontology of dispositions and propensities. So why should we be persuaded by thought experiments involving 'candidates' that are unchanging, static sets of elements?

There's every reason to believe that a better criterion of material object individuation can be given by something that is unchanging and yet has a dynamical aspect - perhaps dispositionally, via (Lorentz) invariant mass; cf. Leibniz, who once defined substance as 'a principle of action and passion'. Under this general criterion, we may define a material object by picking out its fundamental dispositional properties. This kind of analysis certainly has no concern for which elements 'belong' to to an object and which do not and makes the rival candidates of Lewis' thought experiments seem much less interesting.

It might still worry one that this kind of individuating technique leaves open the question of which elements of a set truly belong to a material object and which do not. But I'm not so sure that this is much of anything to truly be worried about. If a material object is defined by (and perhaps, is essentially) a set of fundamental dispositions, asking which material elements belong to it and which do not is not asking anything very interesting about it - unless, of course, one thinks that the contingencies of the moment-to-moment career of the object in question is of any real importance; it certainly isn't important for scientific inquiry at any rate.

A Posteriori Necessities & Natural Kinds

A new article by Nigel Leary ('Natural Kinds: Thick Essentialism or Promiscuous Realism?') suggests to me that Kripkean kind identities may or may not be true a posterori necessary statements. I don't think Leary's argument is incorrect per se, but I do think that perhaps he makes an incorrect equivocation.

The point that Leary makes is that Kripkean identity statements like "Water is H2O" are necessary statements only if we think that 'the essence of a chemical element is its chemical structure'. In short, these identity statements are only necessary insofar as we grant some form of essentialism. But then, the idea that such statements are necessary collapses, since we must assume a form of necessity (namely, essentialism) in order to have necessary statements. "Water is H2O" is a necessary statement iff an object's chemical structure is essential to it.

(Or perhaps the necessity fails because 'Twin Earth' thought experiments aren't wholly out of line, in the sense that they might really show something that is possible in the case of 'Water' being composed of 'XYZ'. But this is more questionable, so I won't get into it here.)

I suppose the pertinent question is: is an object's chemical structure necessary to it? I'm not so sure that it is. And to this extent, I agree with Leary that the Kripkean strategy for deriving a posteriori necessary truths might be, to a large extent, a circular, and so ultimately unsucessful strategy.

But I will say that I think that broadly construed Kripkean a posteriori necessary identity statements are mostly true, at least the ones that concern natural kind identities. This is because, as far as I'm concenred, natural kinds are correctly defined dispositionally. 'What it is to be' a natural kind is just to have a certain dispositional nature. If this is correct - and I do think it is - it is easy to see why I can allow that statements like "Water is H2O" might not be necessary, though they certainly concern natural kinds.

However, Leary is equivocating a bit. To say that a natural kind has a chemical structure essentially could mean that it has a categorical property as a matter of necessity. But it could also mean that it has a certain dispositional nature of necessity - namely, the dispositional nature proper to the pertient mixture of hydrogen and oxygen. If having this particular mixture brings along (or better yet, just is) a certain dispositional nature, I should think that the statement "Water is H2O" expresses a necessary truth.

Dispositional InEssentialism's Implausibility

Given dispositional realism, there are two main positions taken in the debate on whether or not dispositions are essential to their bearers*:

Dispositional Essentialism: Dispositions are the 'what it is to be' for the objects that have them and the existence of these objects is conditional upon their having these dispositions. An object having different dispositions than the ones it does is therefore impossible. Thus, different laws of nature, different objects. (Bird, Ellis & Lierse, Marcus)

Dispositional InEssentialism: Dispositions are contingent aspects of the objects that have them and the existence of these objects is not conditional upon their having these dispositions. An object having different dispositions than the ones it does is possible. Thus, different laws of nature do not imply different objects. (Armstrong, Lewis)

Now, importantly, the debate about dispositional essentialism is a debate about so-called fundamental entities; those 'natural' objects of scientific enquiry - quarks, leptons and electrons. The exemplar question is something like 'Is it possible for an electron to attract another electron'? - ie. could the dispositional nature of an electron have been different?

For the dispositional inessentialist, an electron possibly attracting another electron is accounted for by the fact that the laws of nature are contingent - the 'Contingency Thesis'. And since different laws of nature holding amounts to there being different ways things must act, it is possible that things have different constraints on the way they interact with one another and hence, different dispositional natures.

The problem with this view, I think, have been laid-out fairly successfully by Bird in his Nature's Metaphysics. For one, laws of nature understood as either regularities or nomic-necessitation generalizations both seem untenable; I won't go in to the intricacies here - suffice it to say that regularities supervene, not impose themselves on objects and nomic-necessitiation is seemingly ad hoc logic-chopping. And since an account of dispositional inessentialism entails that the laws of nature are extra-objectual and no such account of laws is plausible, neither is dispositional inessentialism.

Secondly though, and more importantly, is the charge that if the laws of nature are understood as extra-objectual (as I claimed that dispositional inessentialism presupposes), then one must be a quidditivist about properties. That is, one must hold that the dispositional nature of any property is not essential to it. Why is the dispositional inessentialist committed to property quidditivism? Firstly, she is an extra-objectual law theorist - so the laws of nature must determine dispositions, not vice-versa. Secondly, she must accept the following proposition:

(ep) An electron could be a positron.

Now, the dispositional inessentialist will protest: I would not dare say that an electron could be a positron! Rather than accepting (ep), I am making the weaker claim that an electron might have had the disposition to attract other electrons - ie. it could have been positively charged. I am not then, as you are claiming, attempting to make the claim that one kind of entity might have been another fundamentally different kind of entity - surely this is absurd.

But what does the acceptance of (ep) entail? Consider that an electron and positron share all of their structural properties and differ only in the value of their charge. The claim made by (ep) is then not simply that an electron could have had a positive charge. This is because, ex hypothesi, a positron just is an object having all of the structural properties an electron has and being positively charged. So the dispositional inessentialist is committed to (ep) afterall.

You might wonder what accepting (ep) has to do with property quidditivism, since the former concerns entities, the latter properties. But I suggest that the claim that the property 'being negatively charged' might have had the role of attracting negative charges is precisely the mistake that the dispositional inessentialist makes. Why? Because 'repelling negative charges' simply is what it is to be a negative charge. Similarly, 'being negatively charged' simply is what it is to be an electron. So, being a dispositional inessentialist entails the acceptance of (ep). And since (ep) is absurd, so is dispositional inessentialism; and, for that matter, property quidditivism.



*There is another, somewhat middle position that Mumford takes, which I call 'soft dispositional essentialism', according to which objects have their dispositions essentially, but their dispositions are relative to a world. Under soft dispositional essentialism, it is possible for an object to have different dispositions, but not different world-indexed dispositions. This account preserves the Contingency Thesis and dispositional essentialism, but is ultimately unacceptable. In short, it can't be claimed that an object such as an electron has different world-indexed dispositions since, as I suggest, what it is to be an electron is to have a particular dispositional nature. Therefore, any object with a different dispositional nature is conceptually barred from being an electron and so no electron could have different dispositions than the ones it does.