I asked commenter John whether he thought that temporal parts -- assuming that there are temporal parts -- would count as ontological constituents of an ordinary particular such as an avocado. Here is what he said:
. . . I believe that I would say that the temporal parts of an avocado are ontological constituents of it. A thing's temporal parts are much more like a thing's material parts than any other putative constituent of that object, so I would say that if a thing's material parts are ontological constituents of it, then so too are a thing's temporal parts.
But I don't think I would say that this commits perdurantists to constituent ontology in any interesting sense. I have always understood the contrast between constituent and relational ontologies to be primarily a matter of how a thing relates to its properties: does a thing have properties by standing in some external relation to those properties, or instead by having those properties somehow 'immanent' in it? Perhaps this is wrong. But if it's right, then I would say that perdurantists believe that the temporal parts of a thing are among its ontological constituents, but that this does not commit them to any interesting version of constituent ontology.
John's response is a reasonable one, but it does highlight some of the difficulties in clarifying the difference between constituent ontology (C-ontology) and relational ontology (R-ontology).
One of the difficulties is to specify what exactly is meant by 'ontological constituent.' John takes the material parts of a thing to be ontological constituents of it. I don't. Material parts are ordinary mereological parts. For me, ontological constituents are quasi-mereological metaphysical parts to be contrasted with physical (material) parts. Ontological parts are those parts that contribute to an entity's ontological structure. R-ontologists deny that ordinary concrete particulars have any ontological structure. This is not to deny that they have mereological structure. So R-ontologists have no use for ontological parts (constituents). But they have plenty of use for material parts as we all do. 'Ontological' and 'metaphysical' are interchangeable adjectives in this context.
An avocado is an improper physical part of itself. Among its proper physical parts are the skin, the meat, and the pit. Of course, each of these has proper physical parts, and the parts have parts. All of these parts are parts in the strict mereological sense of 'part.' Now consider the dark green (or greenness) of the skin. It is not a physical or material or spatial part of the skin. I can't peel it off the skin or cut it up or eat it. If it is a part at all, it is a metaphysical part of the skin. And the same goes for every other property of the skin: if is is a part at all, it is a metaphysical part. These metaphysical property-parts together perhaps with some other metaphysical parts (bare or thin particulars, various sorts of nexus, Castanedan ontological operators. . .) make up what we can call the ontological structure of an ordinary particular. This quasi-mereological ontological structure is distinct from the strictly mereological structure of the object in question.
Everyone agrees that things like avocados and aardvarks and asteroids have physical parts. But not all agree that they have in addition metaphysical parts. As I see it, the issue that divides C-ontologists from R-ontologists is the question whether concrete particulars have metaphysical parts in addition to their physical parts where the thing's properties are among its metaphysical parts. C-ontologists say yes; R-ontologists, no.
This is a broader understanding of the difference between C- and R-ontology than John's above. For John the difference is between how concrete particulars have properties. For a C-ontologist, a thing has a property by having it as an ontological constituent. For an R-ontologist, a thing has a property, not by having it as a constituent, but by standing in an external relation to it. That is not wrong, but I think it is too narrow.
John seems to be suggesting that the only ontological constituents there are are properties, and that the only items that have such constituents are ordinary concrete particulars. My understanding is broader. I maintain that among ontological constituents there are or could be other items such as bare or thin particulars, various type of nexus, ontological operators, and perhaps others, in addition to properties (whether taken to be universals or taken to be tropes). I am also open to the possibility that entities other than ordinary concrete particulars could have ontological constituents.
Take God. God is presumably a concrete particular, concrete because causally active, particular because not universal; but surely God is not an ordinary concrete particular, especially if 'ordinary' implies being material. Arguably, God is not related to his attributes; if he were his aseity would be compromised. So I say he has his attributes as constituents. If he is identical to them, as on the doctrine of divine simplicity, then a fortiori he has them as constitutuents, improper constituents.
Return to the humble avocado. Our avocado is green, ripe, soft, etc. So it has properties. This simple observation gives rise to three philosophical questions:
Q1. What are properties?
Q2. What is the item that has the properties?
Q3. What is property-possession? (What is it for an item to have properties?)
I will now contrast one R-ontological answer with one C-ontological answer. What follows are very rough sketches.
One R-ontological answer is this. Properties are abstract objects in a realm apart. They are causally inert, atemporal, nonspatial, not sense-perceivable. Not only do properties not enter into causal relations, they do not induce causal powers in the things that have them. They are what is expressed by such open sentences as '____ is green' analogously as propositions are expressed by such closed sentences as 'Ava is green.' If, per impossibile, God were to annihilate all of these abstract objects, nothing would change in our humble avocado. I say per impossibile because the abstract objects in question are necessary beings. My point is that they do no work here below. They are as irrelevant to what is really going on in the avocado as the predicates 'ripe' and 'green' are.
The item that has properties is just the ordinary concrete thing, the avocado in our example, not a propertyless substratum or any other exotic item. The having is a relation or nonrelational tie that connects the concrete thing to the abstract property.
Now for a C-ontological answer. Properties are universals. Whether or not they can exist unexemplified, when they are exemplified, they enter into the ontological structure of ordinary particulars as metaphysical parts thereof. Thus the greenness of the avocado is 'in' it as a metaphysical part. Same holds for the ripeness, the softness, etc. These universals are empirically detectable and induce causal powers. The thing that has these universals is the avocado viewed as a complex, indeed, as a concrete fact. What makes it particular is a further constituent, the thin particular, which is nonrelationally tied to the universals and unifies them into one thick particular.