1. Uncontroversially, ordinary material particulars such as cats and cups have parts, material parts. Equally uncontroversial is that they have properties and stand in relations. That things have properties and stand in relations is a plain Moorean fact beyond the reach of reasonable controversy. After all, my cat is black and he is sleeping next to my blue coffee cup. So far we are at the pre-philosophical level, the level of data. We start philosophizing when we ask what properties are and what it is for a thing to have a property. So the philosophical question is not whether there are properties -- of course there are! -- but what they are. Neither is it a philosophical question whether things have properties -- of course they do! The question concerns how this having is to be understood.
What we want to understand are the nature of properties and the nature of property-possession. Qua ontologist, I don't care what properties there are; I care what properties are. And qua ontologist, I don't care what properties are instantiated; I care what instantiation is.
2. For example, is the blueness of my cup a repeatable entity, a universal, or an unrepeatable entity, a particular (e.g.,a trope)? That is one of several questions one can ask about properties. A second is whether the cup has the property by standing in an external relation to it -- the relation of exemplification -- or by containing it as an ontological or metaphysical part or constituent. Can property-possession be understood quasi-mereologically, as analogous to a part-whole relation? Or is it more like the relation of a thing to a predicate that is true of it? The predicate 'blue' is true of my cup. But no one would get it into his head to think of the word 'blue' as a part of the cup -- in any sense of 'part.' 'Blue' is a word and no concrete material extralinguistic thing has as a word as a part. The relation between 'blue' and the cup to which it applies is external: each term of the relation can exist without the other. Indeed my cup could be blue even if there were no English language and no such word as 'blue.' But if x is an ordinary part or an ontological constituent of y, then y cannot exist without x. So one might analogize properties to predicates and maintain that properties are external to the things that have them and are related to them by exemplification.
3. At a first approximation, the issue that divides constituent ontologists (C-ontologists) and those that N. Wolterstorff rather infelicitously calls 'relational ontologists' (R-ontologists) is whether or not ordinary particulars have ontological or metaphysical parts. C-ontologists maintain that ordinary
particulars have such parts in addition to their commonsense parts, and that among these ontological parts are (some of) the properties of the ordinary particular. R-ontologists deny that ordinary particulars have ontological parts, and consequently deny that ordinary particulars have any of their properties by having them as parts.
4. Let us now examine E. J. Lowe's explanation of the distinction. After reminding us that C-ontologists ascribe to ordinary particulars ontological structure in addition to ordinary mereological structure, he writes: ". . . what is crucial for an ontology to qualify as 'constituent' is that it should maintain that objects have an ontological structure involving 'constituents' which belong to ontological categories other than the category of object itself." ("Essence and Ontology" in Novak et al. eds., Metaphysics: Aristotelian, Scholastic, Analytic, Ontos Verlag 2012, pp. 102-103.) Lowe's characterization of the distinction goes beyond mine in that Lowe requires that the constituents of an object belong to categories other than that of object. An object for Lowe is an Aristotlelian primary (individual) substance. For me it suffices for an ontology to be 'constituent' that it allow that some entities have ontological constituents.
Lowe cites hylomorphism as an example of a constituent ontology. On both Lowe's and my understanding of 'constituent ontology,' hylomorphism is a clear example of a C-ontology. On hylomorphism individual substances are combinations of form and matter where neither the form nor the matter are substances in their own right. But is it true to say or imply, as Lowe does, that forms and matters are members of categories? This strikes me as a strange thing to say or imply. Consider just the forms of individual substances. I would not say that they are members of a category of entity alongside the other categories, but that, on hylomorphism, they are 'principles' (as the Thomists say) invoked in the analysis of individual substances. Form and matter are ontological constituents of an Aristotelian primary substance. But that is not to say that these constituents belong to categories other than that of primary (individual) substance. It is true that the form of a substance is not itself a substance. It does not follow, however, that the form of a substance belongs to an ontological category other than that of substance.
So that is my first quibble with Lowe's explanation. Here is my second. It seems that Lowe's explanation rules out one-category constituent ontologies. Keith Campbell advertises his ontology as 'one-category.' (Abstract Particulars, Basil Blackwell, 1990)) The one category is that of tropes. Everything is either a trope or a construction from tropes. Campbell's is therefore a one-category constituent ontology. Lowe's explanation, however, implies that there must be at least two categories of entity, the category object (individual substance) and one or more categories of entity whose members serve as constituents of objects.
A third problem with Lowe's explanation is that it seems to rule our Bergmann-type C-ontologies that posit bare or thin particulars. Lowe's explanation seems to suggest that the constituents of a particular cannot include any particulars. If a bare particular is a particular, then an ordinary particular has a particular as a constituent in violation of Lowe's explanation. (It is a very interesting question whether a bare particular is a particular. I am tempted to argue that 'bare' functions as an alienans adjective so that a bare particular is not a particular but rather the ontological factor of particularity in an ordinary particular. But this is a separate topic that I will get to in a separate post.)
5. I now want to discuss whether Lowe's four-category ontology succeeds in being neither a C-ontology nor an R-ontology, as he claims.
First of all the question whether it is a C-ontology. Lowe's categorial scheme is approximately as depicted in this diagram:
Lowe speaks of Kinds (substantial universals) being instantiated by Objects (substantial particulars), and of Attributes (non-substantial universals) being instantiated by Modes (non-substantial particulars). Not shown in the above Ontological Square is a diagonal relation of Exemplification running from Attributes (non-substantial universals) to Objects (substantial particulars). Consider, for example, the horse Dobbin. It is an individual substance that instantiates the natural kind horse. Dobbin also has various accidental properties, or Attributes, whiteness, for example. Dobbin exemplifies the universal whiteness. The whiteness of Dobbin, however, is unique to him. It is not a universal, but a particular, albeit a non-substantial particular. It is a Mode (trope) that instantiates the Attribute whiteness. Dobbin is characterized by this Mode, just as the Kind horse is characterized by the Attribute whiteness. On Lowe's scheme there are three distinct relations: Characterization, Instantiatiation, and Exemplification. They relate the members of four distinct fundamental ontological categories: Kinds, Objects, Attributes, and Modes.
Are modes constituents of the objects they characterize? Is Dobbin's whiteness a constituent of Dobbin? If it is, then Lowe's ontology counts as a C-ontology. Lowe plausibly argues that modes are not constituents of objects. I take the argument to be as follows. Modes are identity-dependent on the objects they characterize. Thus Dobbin's whiteness would not be what it is apart from Dobbin and could not exist apart from Dobbin. It follows that the mode in question cannot be an ontological 'building block' out of which Dobbin, together with other items, is constructed. An object is ontologically prior to its modes, which fact entails that modes cannot be constituents of objects.
So far, so good. But what about modes themselves? Do they have constituents? Or are they simple? If modes have constituents, then Lowe's is a C-ontology after all. Dobbin's whiteness could be taken to be Dobbins-exemplifying-the universal whiteness, or it could be taken to be a simple item lacking internal structure, a simple instance of whiteness. If it is a simple item, just an instance of whiteness, then it cannot have any necessary connection to Dobbin or to any object. Why then would it be necessarily identity- and existence-dependent on Dobbin? Why would it be so dependent on any object? There would be nothing about it to ground such a necessary connection. And if it were a simple, then it could very well be a constituent of an object. Lowe's argument against the constituency of the whiteness mode requires that the mode have a necessary connection to Dobbin, that it be the whiteness of Dobbin and of him alone. The mode cannot have that necessary connection unless it is a complex.
If, on the other hand, Dobbin's whiteness is a complex item, then it has as constituents, Dobbin, exemplification, and the universal whiteness, in which case Lowe's ontolology is a C-ontology. For if an ontology has even one category of entity the members of which have ontological constituents, then that ontology is a C-ontology.
My argument can also be put as follows. On Lowe's scheme, modes make up a fundamental category. As fundamental, modes are not derivative from other categories. So it cannot be that a mode is a complex formed by an object's exemplifying an attribute, e.g., Dobbin's exemplifying the non-substantial universal, whiteness. But if modes are simple, why should modes be identity-dependent on objects? It is clear that the whiteness of Dobbin cannot be an ontological part of Dobbin if the whiteness is necessarily tied to Dobbin to be what it is. For then it presupposes the logically antecedent existence of Dobbin. But the only way the whiteness can be necessarily tied to Dobbin is if it is a complex -- which is inconsistent with modes' being a fundamental and irreducible category.