A Sketch of the Difference between Two Ontological Styles
What it is for a thing to have a property? Ostrich nominalism aside, it is a Moorean fact that things have properties, but the nature of the having is a philosophical problem. The ordinary language 'have' does not wear it correct ontological analysis on its sleeve. My cup is blue. Does the cup have the property of being blue by standing in a relation to it -- the relation of exemplification -- or by containing it as an ontological or metaphysical part or constituent? The root issue that divides constituent ontologists (C-ontologists) and those that N. Wolterstorff calls, rather infelicitously, "relation ontologists" (R-ontologists) is whether or not ordinary particulars have ontological or metaphysical parts.
C-ontologists maintain that (i) ordinary particulars have such parts in addition to their commonsense parts; (ii) that among these ontological parts are (some of) the properties of the ordinary particular; and (iii) that the particular has (some of) its properties by having them as proper parts. 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. Of course, R-ontologists do not deny that (most) ordinary particulars have commonsense parts.
Drawing on some graphic images from D. M. Armstrong, we can say that for C-ontologists ordinary particulars are "layer cakes" while for R-ontologists they are "blobs." 'Blob' conveys the idea that ordinary particulars lack ontological structure in addition to such commonsense structure as spatial structure.
The distinction between these two styles of ontology or two approaches to ontology is not entirely clear, but, I think, clear enough. To take an example, is the blueness of my blue coffee cup an abstract object off in a platonic or quasi-platonic realm apart and only related to the cup I drink from by the external asymmetrical relation of exemplification? That, I take it, is van Inwagen's view. I find it hard to swallow. After all, I see (with the eyes of the head, not the eye of the mind) the blueness at the cup, where the cup is. Phenomenologically, I see (some) properties. So some properties are literally visible. No abstract objects (as PvI and others influenced by Quine use 'abstract objects') are literally visible. Ergo, some properties are not abstract objects.
Here is a second argument. Some properties are either wholly or partially located at the places where the things that have the properties are located. No abstract objects are either wholly or partially located at the places where the things that have the properties are located. Therefore, some properties are not abstract objects. So I am inclined to say that the blueness of my cup is in some unmereological and hard-to-explain sense an ontological proper part or constituent of the cup. It is obviously not the whole of the cup since the cup has other properties. Ordinary particulars are not ontologically structureless 'blobs.'
Needless to say, these two quick little arguments do not decide the matter in favor of C-ontology. And the other arguments I could add won't decide the matter either. But taken cumulatively these arguments give one good reason to reject R-ontology.
It is also worth observing that an ontological constituent needn't be a property. Gustav Bergmann's bare particulars and Armstrong's thin particulars are ontological constituents of ordinary or 'thick' particulars but they are not properties of those particulars. The materia signata of the Thomists is a constituent of material particulars, but not a property of such particulars. So, C-ontology is not just a thesis about properties and how they are had by the things that have them.
So much for ontological background. For more on the two ontological approaches, see my article, "Constituent versus Relational Ontology," Studia Neoaristotelica: A Journal of Analytical Scholasticism, vol. 10, no, 1, 2013, pp. 99-115. Now what relevance does this have for the classically theist doctrine of divine simplicity (DDS)? But first: What is DDS?
The Doctrine of Divine Simplicity
To put it as 'simply' as possible, DDS is the thesis that God is without (proper) parts. (If you want to say that God is an improper part of himself, I'll let that slide.) Being without parts, God is without composition of any sort. It is obvious that God is not a region of space, nor does he occupy a region of space. So he cannot have spatial or material parts. If God is eternal, then he cannot have temporal parts. (And if there are no temporal parts, then God cannot have them even if he is everlasting or omnitemporal.) But he also lacks ontological parts. So the divine attributes cannot be different parts of him in the way that my attributes can be different parts of me on a C-ontology. We can put this by saying that in God there is no real distinction between him and his omni-attributes. He is each attribute, which implies that each attribute is every other attribute. Indeed, there is no distinction in God between God and any of his intrinsic properties. (Each omni-attribute is an intrinsic property, but not conversely.) What's more, there can be no distinction in God between essence and existence, form and matter, act and potency. Since God is in no way composite, he is simple.
And why must God be simple? Because he is absolute, and nothing absolute can be depend for its existence or nature on anything distinct from it. An absolute is what it has. It cannot be compounded of anything that is not absolute or dependent on anything that is not absolute. Why must God be absolute? Because anything less would not be God, a worship-worthy being. These answers are quick and catechetical, but I must invoke my blogospheric privilege and move one.
Plantinga's Critique Misses the Mark
Perhaps the best-known attack on the coherence of DDS is that of A. Plantinga in his Does God Have a Nature? The attack fails because Plantinga foists on the DDS an R-ontology that is foreign to the thought of DDS defenders. If properties are abstract objects, and God is a concrete particular, then of course it would be incoherent to maintain that God and omnsicience are one and the same. For if omniscience is a property, and properties are abstract objects, and abstract objects are causally inert, then the identification of God and omniscience would either render God causally inert -- which would contradict his being concrete -- or it would render omniscience causally active -- in contradiction to its being abstract. More simply, if you think of concreta and abstracta as denizens of radically disjoint realms, as R-ontologists do, then it would be something like a Rylean category mistake to maintain that God is identical to his properties.
More simply still, if God is causally active and no property is causally active (or passive for that matter), would it not be supremely stupid to assert that there is no distinction in reality between God and his properties? Could Augustine, Anselm, Aquinas, Avicenna, et al. have been that stupid? I don't think so. Aquinas was as little Quinean in his understanding of abstracta as Quine was Aquinian. Philosophical theologians under the spell of Quine such as Plantinga and van Inwagen are not well situated to understand such tenets of classical theism as DDS.
It is obvious, then, that DDS is incoherent when read in the light of R-ontology. It is also uncharitable in excelsis to read Aquinas et al. in that light because so reading them makes nonsense of what they say.
Does C-ontology Help with Coherence?
One of the entailments of DDS is that God does not exemplify his nature; he just is his nature. We have seen that this makes no coherent sense on (any version of) R-ontology. But it does make coherent sense on (some versions of) C-ontology. For if God is purely actual with no admixture of potency, wholly immaterial, and free of accidents, then what is left for God to be but his nature? To understand this, one must bear in mind that the divine nature is absolutely unique. As such it is not repeatable: it is not a universal. It is therefore unrepeatable, a particular. What is to prevent it from being identical to God and from being causally active?
If you say that God is an instance of a multiply exemplifiable divine nature, they you are simply reverting to R-ontology and failing to take in the point I just made. God cannot be an instance of a kind, else he would depend on that kind to be what he is. God transcends the distinction between instance and kind. And if you persist in thinking that natures are causally inert abstract objects, then you are simply refusing to think in C-ontological terms.
If you say that I beg the question against the denier of DDS when I say that God transcends the instance-kind distinction, then you miss the point. The concern here is not whether DDS is true or whether there are non-question-begging arguments for it; the concern is whether it makes coherent sense as opposed to being quickly dismissable as guilty of a category mistake.
Another objection one might make is that the divine nature is not simple but complex, and that if God is his nature, then God is complex too. For Plantinga, the nature of a thing is a conjunctive property the conjuncts of which are those properties the thing exemplifies in every possible world in which it exists. On this approach, the divine nature is 'cobbled together' or constructed out of God's essential properties. But then the divine nature is logically and ontologically posterior to those properties. Clearly, no defender of DDS will think of natures in the Plantingian way. He will think of the divine nature as logically and ontologically prior to the properties, and of the properties as manifestations of that unitary nature, a nature the radical unity of which cannot be made sense of on Plantinga's approach.
There are other problematic entailments of DDS. One is that in God, nature and existence are one and the same. On an R-ontology, this makes no coherent sense. But it can be made sense on a C-ontological approach. A fit topic for a separate post.