It is possible to contest this construction on Thomas's strategy in the 'five ways' on a number of counts, however, two in particular deserving comment. The first is the importation into Thomas's thinking, at just the wrong point, of the 'Augustinian' principle that you cannot know relative degrees of imperfection in creatures without first knowing, in some way, whether by 'intuition', or as the object of some 'experience', supreme perfection to exist. Thomas certainly maintains that there could be the degrees of goodness which we perceive only if there were an absolute good in which created goods participated; indeed, for Thomas, 'a participating x' and 'a created x' are extensionally, and possibly also intentionally, equivalent. But whether or not it means the same to say that a thing 'participates in another' as to say that it 'is created', what is certainly true is that anything which participates in another is created. Moreover, what shows that the goods we perceive are participating, and so created, goods, is whatever shows the supreme good to be, and to be their Creator. Hence, so far as concerns what is at stake here, Milbank is right: there are but two possible ways of reading Thomas's strategy. The first is that there is a logically valid proof, which starts from the degrees of created goodness which we perceive and concludes to the existence of the supreme good which is their creating cause - in which case, general standards of validity of proof preclude there being, presupposed to our perception of these
5 Fergus Kerr, After Aquinas: Visions of Thomism, Oxford: Blackwell, 2002, p. 59. But see note 24 below.
6 In fact any such presupposition would have to be no more than implicit, since there is absolutely no indication that Thomas admits to making it in ST 1a q2 a3.
created goods and required as a premise, some prior knowledge of the supreme good's existence. The second is that presupposed to the five ways there is some such prior 'intuition' or 'experience' of the supreme good, such as would be required to perceive the created goods from which the 'arguments' proceed as participating and as created - in which case an argument from motion could not be, and probably could not have been intended to be, a logically valid proof. And of course Milbank settles for the latter to the exclusion of the former.
But why settle for that second option? In particular, why settle for it in view of the fact that lying before us in the text of Summa Theologiae 1a q2 a3 corp. is just the evidence we need that Thomas chose the first option and that he thought the existence of a supreme good can be proved - the so-called 'fourth way'? For here Thomas does not say: we know that there are degrees of goodness in things only because we know a priori that there is a supreme good, there being no proof of anything in that, and Thomas is explicit about his argument-strategy here: it is meant as proof, for this, he says, is the fourth of the ways in which Deum esse probari potest.1 Besides, there is no evidence whatever that Thomas thinks the general proposition to be true - as Anselm, Bonaventure, Descartes, and above all Augustine certainly did - that we can perceive relative degrees of a quality only if we have prior knowledge of what would count as the maximal degree of it. In fact Thomas says the inverse of what Milbank claims for him: he argues that we know there must be a supreme good because there are degrees of goodness in things. As he puts it, we 'meet with'8 greater and lesser degrees of goodness and truth in things anyway, and the argument then goes on to show how it follows that there must be some maximal degree of such qualities, not that we could not know of such degrees of goodness and truth unless we already had some 'glimpse' of them in their maximal degree. And whatever one thinks of the validity of such an inference, or of the further, insufficiently explained, inference that whatever accounts for our capacity to judge degrees of goodness must be the cause of them, the argument is clearly presented as an inference, moreover to a cause, 'which we call God'. Hence, even if, as Milbank believes, the force of the argument from motion (the 'first way') depends upon the presupposition that there is a supreme good which is the end of all motions, the proposition that there is a supreme good is itselfcapable of
7 Thomas's choice of words rules out Milbank's suggestion that the five ways are intended as demonstrationes in some weaker sense than 'strict proof: potest probari is as strong as you can get in point of apodeicticity.
8 'Invenitur enim in rebus aliquid magis et minus bonum, et verum, et nobile; et sic de aliis huiusmodi.' ST 1a q2 a3 corp.
being demonstrated (potest probari) by an independent proof (the 'fourth way').
Which brings us to the second, and more critical, misreading. It seems crucial to Milbank's case - though of course it is not the whole of it -that Thomas could not consistently have intended the five ways to be logically valid proofs of the existence of God, for it is Milbank's view that logically valid proofs of the existence of God are impossible, short of engagement in a form of 'Scotist onto-theology' - which Thomas, by anticipation, rejected. Here, we are brought back in turn to that persistent tendency within Milbank's work to play Thomas off against Scotus, a tendency which does so on the one hand to Scotus' disadvantage, as we have already had cause to observe, and on the other hand on terms and rules of contest which, remarkably, owe far more to Scotus than they do to Thomas, as we shall now see.
It is, in the first place, Scotus, not Thomas, who anywhere says that demonstration of the existence of God depends logically upon 'being' being predicated univocally of God and of creatures, such that analogical predication would rule out scientific demonstration:
The active intellect and the sense image cannot give rise to a concept that, with respect to the object revealed in the sense image, is not univocal but rather, in accordance with an analogical relationship, is altogether other and higher than the object. It follows that such an 'other', analogous, concept will never arise in the intellect in our present state. Also it would thus follow that one could not naturally have any concept of God - which is false.9
Thomas nowhere says any such thing - in fact, as we shall see, he explicitly anticipates Scotus' objection, and rejects it. Moreover, it is Milbank, not Thomas, who repeats this Scotist nostrum as a crucial step in his account of why Thomas cannot be offering proofs from the natural light of reason: 'one can point out', he says, that in the realm of metaphysics even the relative certainty profered by reason is very weak. For scientific demonstration proper depends, for Aquinas after Aristotle, on a univocity of terms answering to a univocity between causes and effects. For Aquinas, this contention disallowed a transgeneric 'science' in the strictest sense . . . Aquinas . . . by identifying God with non-generic esse, and by specifically excluding God from genus and from substance in the sense either of distinct
9 'Sed conceptus qui non esset univocus obiecto relucenti in phantasmate, sed omnino alius, prior, ad quem ille habeat analogiam, non potest fieri virtute intellectus agentis et phantasmatis; ergo talis conceptus alius, analogus qui ponitur, naturaliter in intellectus viatoris numquam erit, - et ita non poterit haberi naturaliter aliquis conceptus de Deo, quod est falsum.' Duns Scotus, Ordinatio 1 d3 n. 36; Frank and Wolter, Duns Scotus, pp. 112-13.
essence or self-standing individual . . . also ensures that there can be only an analogical or not strictly scientific approach to the divine. Hence . . . his 'demonstrations' of God's existence can only be meant to offer weakly probable modes of argument and very attenuated 'showings'. ('Intensities', pp. 454-5)
Twice in this passage Milbank stakes a claim on Thomas's behalf for the impossibility of a strictly scientific demonstration which depends on inference between terms predicated analogically. But to conclude from this that no scientific demonstration of the existence of God is possible -on the grounds, of course, that any sense to a term predicated of God can be only 'analogically' related to the sense that term has of creatures -is to misunderstand equally what is required of valid inference, of the structure of the arguments for the existence of God in Aquinas, and of the logic of analogy. We shall see later in this chapter that, on the score of univocity, what is required of any valid inference is no more than that any term occurring more than once in the premises of a valid inference (the so-called 'middle term') is used in the same sense (univocally) on every occasion of its occurrence. It is not required that there be a univocity of terms in premises and conclusion. It is true that, as Milbank says, Aristotle goes further and maintains that no conclusion may follow in a valid demonstration whose terms are not predicated according to a sense univocally the same as those same terms occurring in the premises. In the Posterior Analytics Aristotle makes this clear:
It is impossible to prove a fact by transition from another genus, e.g. a geometrical fact by arithmetic . . . every proof has its own subject-genus. Therefore the genus must be either the same, or the same in some respect, if proof is to be transferable; otherwise it is impossible; for the extremes and the middle term must be drawn from the same genus, since if they are not connected per se, they are accidental to one another . . . nor can one science prove the propositions of another, unless the subjects of the one fall under those of the other, as is the case with optics and geometry, or with harmonics and arithmetic . . .10
and no doubt Milbank takes comfort from the exception which Aristotle makes to this general rule of syllogistic inference. For, on his account, a philosophical proof of the existence of God would have validity only as 'falling under' the principles of revealed theology in the same way as the subject-matter of optics 'falls under' that of geometry. Milbank's way of putting it is that the arguments for the existence of God gain what little power of proof they possess from their equivalent 'subalternation' to an already given, revealed, and so faith-based premise - the given experience of divine perfection. But, according to Milbank, otherwise than within that theological subalternation the five ways have no probative power of
10 Aristotle, Posterior Analytics 1.7, 75a38-b18.
any kind. That, however, as I have argued, makes no sense at all of what 'proof is. For that subalternation guarantees nothing even 'weakly' or 'probabilistically' to Thomas's arguments by way of proof. Since, on the contrary, that 'subalternation' reduces them to straightforwardly invalid arguments, committing the fallacy of petitio principii, we are once again brought back to a choice of readings of Thomas's 'five ways'. Either Thomas means what he says - and he says the 'five ways' are proofs -in which case he at least cannot suppose them to beg the question; or else Milbank is right, and Thomas's conclusion ('God exists') follows only from a conjunction of premises which include an explicitly theological presupposition - in which case neither he nor Milbank have any business describing them as 'proofs' of any sort, 'weak' or 'strong'.
Be that as it may, Milbank's objection in principle to strict proof of God's existence appears to rest on the supposition that if transgeneric demonstration is invalid, then an inference which purported to transgress the boundary between any created genus and God, who is beyond every genus, must by at least the same token be invalid.11 But this is a significant non sequitur, the full and disastrous theological consequences of which I shall examine later in this chapter. At this point let us confine ourselves to saying, first, that it is a non sequitur. To suppose without more ado that because an inference is invalid by the fallacy of equivocation if it crosses from one genus to another it must be at least as invalid if it crosses from generic being to God, who is beyond every genus, is to suppose, without more ado, that the gap to be crossed between one genus and another and the gap to be crossed between generic being and God are logically the same kinds of gap, only - one supposes - 'bigger' in the latter case. And, secondly, it is a significant non sequitur because that, once again, is exactly the supposition which Scotus makes. For though Scotus, like Milbank and Thomas, denies that God belongs to any genus, and though Scotus, unlike Milbank and Thomas, so construes the 'gap' between God and creatures as to be logically of the same kind as that between one genus and another, Milbank, like Scotus and unlike Thomas, holds that inference could cross the gap between creatures and God only if that gap fell univocally within a common genus. It is beside the point at this stage that, unlike Scotus, Milbank thinks it to follow that inference cannot cross it, since it is impossible that terms could be predicable univocally of God and of creatures. For Milbank's assumption about inference and univocity is Scotist, not Thomist. If this must have the consequence that Scotus is simply confused about how God and creatures differ, it equally means that Milbank has, conversely, conflated what holds of the logic of inference
11 Milbank and Pickstock, Truth in Aquinas, p. 28.
between genera with what holds of the logic of inference between generic being and God. And this, in turn, entails a Scotist confusion on Milbank's part concerning the logic of analogical predication.
Was this article helpful?