Craig on Thomism and divine simplicity, Part II

Let’s continue our critical look at William Lane Craig’s treatment of Thomism and the doctrine of divine simplicity in Volume IIa of his Systematic Philosophical Theology.  In the first part of this series, I showed that in several ways, Craig badly misunderstands the views of Aquinas and other Thomists.  In this second part, I’ll turn to Craig’s treatment of the more general metaphysical disputes underlying the debate over divine simplicity, and to specific criticisms he raises against the doctrine and the arguments of its defenders.

Metaphysical hand-waving

Analytic philosophers working in metaphysics commonly draw a distinction between “constituent ontologies” and “relational ontologies,” and the contemporary literature on divine simplicity often makes reference to this distinction.  The basic idea is that a constituent ontology holds that an object’s properties are constituents or parts of the object itself, whereas a relational ontology denies this.  Consider some particular dog Fido and his four-leggedness.  For a constituent ontology, Fido’s four-leggedness is a constituent of Fido himself.  But for a relational ontology, it is something Fido has by virtue of his relation to something else – for example, by virtue of his instantiating the universal four-leggedness, considered as an abstract object.  In contemporary discussions, Aristotelianism is commonly taken to be committed to a constituent ontology insofar as it takes a particular thing’s form to be a constituent of the thing itself.  Platonism, by contrast, is said to be committed to a relational ontology insofar as it takes Forms to be, not constituents of particular things themselves, but rather entities existing in a third realm outside of time and space, to which particular things are related by participation. 

It gets more complicated than this, but that much will suffice for present purposes.  Now, the doctrine of divine simplicity as understood by thinkers like Aquinas is often said in contemporary discussions to presuppose a constituent ontology rather than a relational ontology.  For Aquinas, to say that God is powerful and good is not to say that he bears a relation of instantiation or participation to the universals power and goodness considered as abstract objects.  Rather, God’s power and goodness are, to a first approximation, better thought of as constituents of God himself.  But only to a first approximation, because ordinarily, when we consider a thing’s constituents, we think of them as distinct parts of the thing.  For example, the power and goodness of a man are distinct from each other and from the man himself.  But according to the doctrine of divine simplicity, God’s power is identical to his goodness, and indeed both his power and goodness are identical to God himself. 

Now, among the reasons Craig gives for rejecting the Thomistic account of divine simplicity is that he rejects the broader metaphysics in terms of which it is articulated and defended, including constituent ontology.  One problem with his criticisms, though, is that they rest in part on misunderstandings of Aquinas and other Thomists, as we saw in the first post in this series.  Another problem is that the criticisms tend to have a “drive by” quality insofar as they engage with Thomistic ideas in a relatively superficial way, and rely too much on rhetorical shortcuts such as appeals to the conventional wisdom in contemporary philosophy and modern philosophy more generally.  And these two failings of Craig’s approach work in tandem to give his discussion an unearned confidence in the strength of its criticisms of Thomism.

For example, Craig emphasizes that “obviously, with Thomas’ explication of divine simplicity, we find ourselves situated within a metaphysical framework which is so different from that of modernity as to constitute a barrier to understanding” (p. 105).  While he acknowledges that this Scholastic philosophical background cannot be entirely ignored, he warns against paying it too much attention, “lest Christian systematic theology be held prisoner to a metaphysical framework now rejected and regarded as obsolete” (ibid.).  Rather, interpreting and evaluating Aquinas’s teaching on divine simplicity “must also involve a careful translation of the doctrine into a contemporary framework, where it may be assessed for its coherence and plausibility” (ibid., emphasis added). 

But as Craig’s own description of the results makes evident, this method of interpretation only ends up making Aquinas’s position sound more mysterious, not less.  Craig approvingly quotes Peter van Inwagen’s complaint that he “do[es] not understand the words and phrases that are the typical items of the core vocabulary of any given constituent ontology… these are all mysteries to me” (p. 149).  Craig laments that “many Thomists freely deploy the jargon of school metaphysics, but often, I fear, without meaning” (p. 164).  Sounding like Rudolf Carnap or A. J. Ayer, Craig dismisses the claims of one defender of divine simplicity as “flowery nonsense,” and suggests that another’s words merely “evoke an emotional response” but without having “assertoric content” (p. 165).  In their replies to critics, defenders of the traditional doctrine of divine simplicity seem to Craig to be driven to “increasingly implausible and even desperate responses” and are “finally driven to appeal to mystery” (p. 157).

It never seems to occur to Craig that the problem might be precisely that his method of interpretation amounts to a kind of eisegesis that makes the views of Aquinas and other Thomists seem bizarre and unintelligible precisely because he is reading alien contemporary assumptions into them.  For example, as we saw in the previous post, Craig alleges that for Aquinas, God is a “pure, featureless subsistent” (p. 171).  As I showed, that is an absurd characterization given what Aquinas actually says about the divine nature.  But it is a misunderstanding one might naturally fall into if one insists on looking for some familiar contemporary metaphysical view to map Aquinas’s position onto, and can’t find anything closer to it than the theory of bare particulars.  Craig complains that while “properties” are universals, Thomists speak as if they were particulars, yielding claims that seem incoherent (pp. 166-70).   And he sympathizes with Alvin Plantinga’s allegation that Aquinas’s position entails the bizarre conclusion that God is a “property.”  All of this sounds plausible if you use “properties” and related terms the way contemporary analytic philosophers tend reflexively to do, but not at all plausible if you use them the way Scholastics do.  And so on.  (I respond to Plantinga’s critique of divine simplicity at pp. 191-95 of my book Five Proofs of the Existence of God.)

In short, Craig’s approach is like taking a puzzle apart, forcing the pieces back together again in a configuration very different from the original, and then complaining that the resulting image looks like a mess.  Or it is like insisting on wearing rose-colored glasses when evaluating how well a painter has carried out your instructions to paint a house white, and then accusing him of having painted it red instead.

Then there is the question of the metaphysics Craig would recommend in place of Thomism.  It isn’t Platonism or any other brand of relational ontology.  Craig rejects those metaphysical systems too, arguing that if God instantiated universals understood as abstract objects, then he would be metaphysically less fundamental than they are (pp. 150-51), a result he rightly rejects.  And he acknowledges that considerations like this are precisely why many proponents of divine simplicity adopt a constituent ontology.  But Craig insists that there is a third option:

[W]e may reject both in favor of an anti-realist alternative, according to which property talk is a useful and perhaps indispensable façon de parler, meaningful within a certain linguistic framework but without ontological commitment.  We make-believe that there are such things as properties in talking, for example, about God’s perfections.  (p. 151)

This anti-realism is a kind of nominalism, though Craig prefers to avoid that term.  He has long been committed to it, defending it in, for example, his book God Over All: Divine Aseity and the Challenge of Platonism.  But there are two serious problems with it.  First, as I argued in my review of that book, there are insuperable metaphysical difficulties with anti-realism.  If properties are merely convenient fictions within a system of make-believe, then why does that system work as well as it does? 

If we say that it works because God causes the world to behave as if properties really existed, then we face a dilemma.  For why did God choose to make the things that populate the world behave the way they actually do behave, rather than some other way?  For example, why did he decide to make a world where fire burns things, acorns grow into oak trees, lions hunt gazelles, and so on?  Suppose the answer is that no alternative was possible – that any world with fire in it was of necessity one where fire would burn things, any world with acorns in it was one where acorns would grow into oaks, and so on.  Then it looks like there are objective facts about the properties things have and how those properties are related, in which case we’re back to a kind of realism about properties.  But the only alternative to this view seems to be a kind of voluntarism according to which there is no rhyme or reason to how God has chosen to create the world, and he could instead have arbitrarily made it operate in any number of crazy ways – fire freezing things, acorns growing into Volkswagens, and so on.  But Craig himself rejects voluntarism, and there is good reason to do so, because it violates the principle of sufficient reason (something to which Craig is committed).

The second problem is that Craig’s anti-realism leads him to characterize the divine nature in a way that sounds very close to what he accuses advocates of divine simplicity of saying, making pointless the whole exercise of developing an alternative metaphysics so as to avoid divine simplicity!  As we saw in the previous post, Craig accuses Aquinas of holding that God has no properties, and objects that this makes of God a bare or featureless subsistent.  But now Craig himself is suddenly holding that we only “make-believe that there are such things as properties in talking, for example, about God’s perfections” (p. 151, emphasis added).  Indeed, it turns out that “on an anti-realist perspective on properties, God is, indeed, not composed of metaphysical parts and is in that sense simple” (p. 152)!  Nor, contrary to what Craig supposes, does it suffice to distinguish Craig’s view from the one he attributes to Aquinas to say that there is a fundamentum in re or foundation in reality for our attribution of properties to God even if they are convenient fictions (p. 172).  For Aquinas too holds that there is a foundation in reality for our attribution of power, goodness, and the like to God.

One clear difference between Craig and Aquinas is that for Craig, while God is without properties and thus in a sense simple, “so is everything else,” so that there is nothing special in his view about the case of God (p. 152).  But it is hard to see how this helps Craig.  His complaint is that Aquinas’s metaphysics leads to bizarre conclusions about God.  How is it an improvement to adopt instead a metaphysics that leads to similar bizarre conclusions not only about God, but about everything else too?

At one point, Craig warns that “Christian systematic theologians would be very imprudent to marry Christian systematic theology to a particular metaphysical framework” (p. 165).  But that is precisely what he does himself.  Far from letting a thousand flowers bloom, Craig tries to burn any Christian bridges to Thomism or Platonism, accusing the former of agnosticism (p. 160) and the latter of implicit atheism (p. 151).  He denies that his own insistence on anti-realism amounts to marrying Christian theology to a particular metaphysical framework, on the grounds that “many anti-realisms are on offer” (p. 166).  (Craig’s idea of metaphysical diversity reminds one of the scene in The Blues Brothers where a bartender is asked what kind of music they play at the bar, and she replies “both kinds… country and western!”) 

Worse, Craig’s motivation is in part to distance Christian theology from ideas that “modernity” regards as “obsolete” (p. 105), and the result is a position that puts him radically at odds with the Fathers of the Church and with Anselm, Aquinas, and the mainstream tradition of Christian systematic theology in general, as represented by Catholic, Orthodox, and Protestant thinkers alike.  Needless to say, this is a very strange approach to recommend to conservative Christian philosophers and theologians.

Simplicity and aseity

Craig’s unwillingness seriously to engage with Thomistic metaphysics greatly weakness his treatment of the arguments for the traditional doctrine of divine simplicity.  In Question 3 of the Prima Pars of the Summa Theologiae, Aquinas devotes eight articles to arguing for the doctrine.  Craig briefly summarizes some of the arguments in the course of describing Aquinas’s position, but says that “our interest at this point is not so much in the many arguments Aquinas gives in support as in the doctrine of divine simplicity that he enunciates” (p. 136).  In other words, he treats Aquinas’s arguments as of interest primarily as sources for finding out what Aquinas thought the doctrine says, rather than as reasons to believe it. 

When he comes to address arguments for the doctrine, he claims that “today there seems to be a consensus that the traditional arguments… are extraordinarily weak” (p. 143).  He offers no support for this claim, and it isn’t true.  For all I know, there may be such a consensus among critics of the doctrine – though that consensus is not worth much given that some of the most prominent of the critics, like Ryan Mullins and Craig himself, rest their case on what are, demonstrably, egregious misunderstandings of the arguments.  But there is no consensus beyond that, and there are many prominent contemporary thinkers who defend the traditional doctrine on traditional grounds, such as Barry Miller, Brian Davies, and David Bentley Hart. 

Craig does at least in a general way address the thesis that simplicity follows from divine aseity, which is central to the Thomistic approach to the issue.  Aseity is self-existence, which involves existing of necessity rather than contingently, and thus in a manner that in no way depends on anything else.  The traditional view is that only what is simple or non-composite can possess aseity.  For if God were composed of parts, then there would have to be some principle distinct from those parts that accounted for why they are united in God.  And in that case, it would be that principle (whatever it is), rather than God, which is the ultimate reality.  He would depend on it for his existence, and if it ceased operating he would cease to exist.  Hence he would lack aseity.  In response, Craig writes:

This argument is weak, for it assumes without justification that God is not composed of metaphysically inseparable parts, so that God has the property of indivisibility along with his other perfections.  A circle, for example, might have a certain circumference, a certain diameter, and a certain area, which, though distinct, are metaphysically inseparable properties. (p. 145)

What Craig doesn’t see is that the circle example does not help his case, but in fact undermines it.  For while it is true that the circumference, diameter, and area of a circle are metaphysically inseparable properties, circles are nevertheless still contingent things which depend on something outside them to cause them to exist.  Positing metaphysically inseparable parts in God thus hardly suffices to show that he could be necessary and independent, and thus possess aseity, even if he were composite rather than simple.

What Aquinas would say, of course, is that while having a certain circumference, diameter, and area all follow with necessity from the essence of a circle, that essence cannot by itself account for the actual reality of any particular circle.  Existence must, as it were, be added to the essence of a circle in order for circles to be actual beings and not merely potential beings.  And that is going to be true of the essence of absolutely any sort of thing – triangles and trees, parrots and people – except something whose very essence just is existence.  Only that sort of thing does not stand in need of having existence added to its essence.  And precisely for that reason, only that sort of thing can have aseity.  But to be the sort of thing whose essence and existence are identical is the core of being simple or non-composite in the sense described by the traditional doctrine of divine simplicity.

Craig would no doubt reject this line of thought, but his circle example and postulation of metaphysically inseparable parts in God in no way rebut the argument.  On the contrary, they completely miss the point.  So too does what he goes on to say next:

It has also been argued that if God were composite, then there must exist a cause of the composition of his parts, in contradiction to the doctrine of divine aseity… To be at all plausible, the argument must assume that God’s co-eternal parts are more fundamental than God.  But there is no reason for such an assumption.  God and his parts could be on an ontological par and metaphysically inseparable.  Moreover, medievals themselves were familiar with wholes that are explanatorily or metaphysically prior to any parts one cares to specify in them, such as in cases of an object’s potential infinite divisibility.  If the whole is metaphysically prior to its parts, then possessing such parts does not clearly require a cause of their conjunction.  (pp. 145-46)

Like the appeal to the circle example, this is simply a blatant non sequitur.  Yes, a whole and its parts can be on an ontological par, and a whole can even be prior to its parts.  So what?  How does that show that such wholes could possess aseity?  That it shows no such thing is obvious as soon as one considers the usual examples, including the one Craig gives.  A physical object such as a stone is a whole that is prior to the potentially infinite parts one could divide it into, but which aren’t actual until a division occurs.  A living thing such as a bird is prior to its parts insofar as organs like eyes and hearts are what they are only relative to the organism of which they are organs.  But stones and birds are nevertheless contingent and depend on other things for their existence, and thus lack aseity.

In response to the objection that if God has parts, then his existence depends on these parts in a way inconsistent with aseity, Craig appeals to Thomas Morris’s retort that even if God depends on his properties, his properties also depend on him, and in a more fundamental way (p. 149).  To be sure, Craig acknowledges that this would lead to “bootstrapping difficulties” if we took this to entail that God creates his own nature.  We would have a vicious explanatory circle in which God’s existence depends on that of his parts, and the existence of the parts depends in turn on him.  But Craig seems to think there is no problem as long as we instead only draw from Morris’s point the lesson that if God is prior to his parts, then he does not depend on those parts in a manner that threatens aseity.

Once again, though, this is a non sequitur.  A bird is prior to its organs in the obvious sense that what it is to be an organ is determined by the organ’s relation to the whole.  All the same, there is an obvious sense in which the existence of the whole bird depends on the existence of the organs – after all, if they were to die, the whole bird would die as well.  So, again, to show that a whole is in some sense prior to its parts does not suffice to show that it does not depend on those parts in a way incompatible with aseity.  And the example also once again illustrates the deeper point that even a whole that is prior to its parts may have an essence distinct from its existence (as birds do), in which case it will require a cause and thus lack aseity.  Craig’s remarks about parts and wholes simply don’t do anything to get around this problem, certainly not in a way that doesn’t beg the question against the Thomist.

The limits of thought

A crucial defect of criticisms of the Thomistic account of divine simplicity, including Craig’s, is a failure seriously to engage with the thesis that assertions about the divine nature must be understood in an analogical sense rather than either a univocal or equivocal sense.  The tendency is to conduct the discussion as if terms like “essence,” “attribute,” “goodness, “power,” etc. have exactly the same sense when applied to God as when applied to created things.  In particular, the critic raises objections to the doctrine of divine simplicity that could have force only if we assumed that they have this univocal sense.  The critic then confidently declares that these objections show the doctrine to be incoherent or unintelligible or self-contradictory.  Finally, the critic acknowledges in passing that Thomists would object that the crucial terms are not to be understood univocally, but analogically.  The critic then quickly dismisses the notion of analogy on the basis of some stock objection (a reference to Scotus in Craig’s case), and moves on to something else.  The issue of analogical language is treated with shocking superficiality and brevity – as if it were an afterthought or a tangential matter, when in fact it is central to the whole debate.  The failure to grapple with analogy renders the entire critique question-begging at best, if not entirely beside the point.

For example, there is a lot of heavy going in Craig’s discussion about whether Aquinas thinks God has an essence.  Craig treats this as if it were a terribly mysterious exegetical question.  He finally decides, correctly, that Aquinas does think God has an essence, one which is identical to his very being.  But Craig is baffled at how this can be so given that Aquinas denies that God falls under any genus.  For if God does have such an essence, “then God’s act of being must fall under a genus and species like every essence” (p. 157).  Of course, this would be true only if God has an essence in exactly the same way that we do.  But of course, that is precisely what Aquinas denies.  The question of what Aquinas thinks about the matter is mysterious only if one insists on reading him as if he were using “essence” univocally, which of course he is not.

Similarly, Craig is baffled by a remark by Stump to the effect that “God does not have an essence; instead, he is identical with his essence” (quoted in footnote 144 at p. 142).  Doesn’t this seem obviously incoherent or self-contradictory?  No, at least not if one makes a minimal effort to read the statement in light of other Thomistic commitments.  The key word in Stump’s remark is “have.”  A stone or a dog or a human being can be said to “have” an essence in the sense of instantiating some essence distinct from the particular thing that has it.  But God does not “have” an essence in that sense, because he is not one particular instance among others of some shared essence.  Rather, we have in the case of God something that is identical to its essence.  And precisely for that reason, while we can attribute an essence to God, we have to understand that as the claim that there is in him something analogous to what we call an essence in created things, even if it is not exactly the same thing as what we call an essence in those things (since in a created thing, a distinction can be drawn between the essence and the thing which instantiates it).  Whether or not one agrees with this account, there is nothing obviously incoherent or self-contradictory about it.

In his book-length defense of divine simplicity A Most Unlikely God, Barry Miller notes that analogical predication can be understood in terms of the notion of a “limit case.”  As he says:

[A] limit case… is not a member of the series of which it is the limit case.  However, although not belonging to the series, it is not totally unrelated to the members, since it is they and the way in which they are ordered that point to this limit case: although different in kind from the series’ members, it is that towards which those members do point. (p. 8)

For example, consider a sequence of ever decreasing speeds – 4 mph, 3 mph, 2 mph, 1 mph.  0 mph is the limit case of this sequence.  It is not itself a speed, because to move 0 mph is not to move at all.  But it is nevertheless that to which the series points.  Or consider a series of ever shorter lines.  A point would be the limit case of this series, because while it is not itself a line, it is that to which the series points.  Or consider a series of regular polygons with an ever increasing number of sides.  A circle would be the limit case of this series.  It is not itself part of the series, because adding sides will never get you a circle, but only polygons that are ever closer approximations to a circle.  Still, the series points to the circle, as it were.

A limit case differs in kind and not degree from the members of the series that point to it.  Again, 0 mph is not a speed, a point is not a line, and a circle is not a polygon.  At the same time, the members of the series do bear a relation of similarity to the limit case to which they point.  For example, the series of polygons described bears a relation of similarity to a circle, even though a circle is not a polygon.

This gives us a way to understand what it means to say that there is, in God, something analogous to what we call an essence in created things, something analogous to what we call goodness in created things, something analogous to what we call power in created things, and so on.  What we’re attributing to God is not exactly the same thing as what we call essence, goodness, power, etc. in created things.  It is not a matter of him having the same attribute but to a higher degree, just as to be a circle is not the same as being a polygon but with more sides.  At the same time, what we attribute to God is not entirely unrelated to what we attribute to created things, just as a circle is not entirely unrelated to the sequence of polygons.  That is why analogical predications are not univocal, but also not equivocal.

Craig pays scant attention to such considerations, dismissing the notion of analogical predication in a few glib sentences.  But to purport to refute the Thomistic account of divine simplicity in such a manner is like pretending to refute Craig’s kalām cosmological argument while devoting at most a few sentences to his arguments against a beginningless universe.  It simply fails to get to the heart of the issue.

Craig also too casually dismisses Stump’s comparison of analogical discourse about God to the way quantum mechanics describes the nature of matter (pp. 162-63).  As I argued in my recent lecture “The Epistemology of Microphysics,” there are in fact important similarities between the two cases.  In particular, I note that there are similarities between philosopher of science N. R. Hanson’s account of the nature of contemporary physics’ inferences about the nature of unobservable particles, and the Thomistic account of the triplex via by which we draw inferences about the nature of God.  I note that whereas God is pure actuality, the closer we get to the bottom level of the hierarchy of being, namely prime matter, the closer we get to pure potentiality.  And the reason our linguistic and conceptual resources come under strain when we try to describe either God or fundamental particles is that in both cases, the human intellect has to stretch well beyond the middle-range entities which are its natural objects of understanding, and which are compounds of actuality and potentiality. 

I won’t repeat the whole analysis here.  Readers interested in it are directed to the lecture linked to above.  Suffice it for present purposes to note the superficiality of Craig’s criticisms of Stump.  He says, first, that the analogy to quantum mechanics doesn’t help, because properly understood, quantum mechanics does not make any contradictory claims, whereas the Thomistic doctrine of divine simplicity does.  But while he is correct to say that an analogy to quantum mechanics would not save a view that entails contradictions, he is wrong to claim that the Thomistic account entails any.  As we have seen in this article and the previous one, any contradictions Craig thinks he sees in the Thomistic position are a product of his misunderstandings, not of anything Aquinas or other Thomists have said.

Second, Craig says that while we have overwhelming empirical support for quantum mechanics, we have no comparable support for divine simplicity.  Hence, while the evidence might force us into accepting quantum mechanics despite its making some strange claims, there is no similarly overwhelming evidence that forces us into accepting divine simplicity regardless of the strangeness of its claims.

But this just begs the question against the Thomist, and against advocates of the traditional doctrine of divine simplicity more generally.  Consider arguments for the existence of God like the Aristotelian proof, the Neo-Platonic proof, and the Thomistic proof, which I defend in Five Proofs of the Existence of God.  Like other advocates of these arguments, I take them to be metaphysical demonstrations.  They claim to show, not merely that there might be or probably is a divine cause of things, but that there must of necessity be such a cause.  Moreover, each claims that this cause must be of a certain nature – pure actuality in the case of the Aristotelian proof, absolutely non-composite in the case of the Neo-Platonic proof, and something whose existence and essence are identical in the case of the Thomistic proof.  And each of these claims entails, in turn, the traditional doctrine of divine simplicity.

Hence, if the arguments work, we are in a situation analogous to the one that the evidence for quantum mechanics puts us in.  In particular, we are led ineluctably to the conclusion that the doctrine of divine simplicity must be true, despite entailing claims that we might find strange.  Naturally, Craig will disagree with this.  The point, though, is that divine simplicity goes to the very core of natural theology as it is understood in traditions such as Aristotelianism, Neo-Platonism, and Thomism.  It is not some optional extra that we may or may not decide to add on after the project of establishing God’s existence is completed.

Of course, many atheists would be happy to dismiss these theistic arguments and divine simplicity alike, since they are no more impressed by the former than the latter.  But Craig can hardly do so, given that he is among the preeminent contemporary defenders of natural theology and apologetics.  To be sure, the arguments in question are not the kind he has traditionally been prone to defend.  But it would be rash for an apologist to dismiss them out of hand.  And in any event, as I say, to do so in order to undermine the support they give the traditional doctrine of divine simplicity would be to beg the question against Thomists and other defenders of the doctrine.

Craig says more about the doctrine, and I may comment on some of it in a future post.  The aim of this series, though, has been to rebut what he says about the Thomistic approach to the doctrine, specifically.