Before considering whether there is a case for the possibility of a causal argument, let us first consider what Thomas's view of the matter is in principle. There are at least two important texts in which Thomas explicitly raises the question of whether the transcendence of God - which entails God's being spoken of 'analogically' - rules out the possibility of inference being valid to God from creatures, and in both his answer is in an unambiguous negative: such inference is not thereby ruled out. The first of these we have already considered:21 on the one hand, Thomas, we saw, maintains that the Christian, who believes in the one true God, and the idolater, who worships some creaturely object as if it were God, contradict each other, which they could not do unless there were something in common between the ways in which they think of God. For unless the idolater was affirming of the idol that it is 'God' in some sense related to that in which the Christian denies that it is God, it could not be the case that the affirmation and the denial were contradictories. Consequently, the Christian's 'God' and the idolater's 'God' cannot be equivocal terms.
20 It goes without saying that no argument for the existence of God could depend upon our knowledge in advance that such analogies do hold, for that would be simply to beg the question.
On the other hand, when the idolater says that this idol is God and the true believer denies it, the word 'God' cannot be used univocally in both cases, for the one is saying of the Creator of all things visible and invisible that it is God, the other that a creaturely idol is God. And the word 'God' cannot be predicated univocally of God and of the creature. Hence, when the pagan and the Christian disagree whether an idol is God, the name 'God' is used, Thomas says, analogically (analogice dicitur).22
It follows then that, for Thomas, there can be formal contradiction between two analogically related propositions. And it follows from that that there can be no objection to there being a formally valid inference between premises and a conclusion analogically related to them across the 'gap' between creatures and God. Why? For the reason which Scotus gives: for if, on his account, an inference is valid only on condition that the terms related to each other by it are such that 'to affirm and to deny [them] of the same subject amounts to a contradiction', then, on Thomas's account, that condition is met by terms which are related to each other analogically. Hence an inference will not, for Thomas, be invalidated by the fact that it connects terms logically related to each other by analogy if, as in the case in question of 'God', to affirm and deny of a bronze statue that it is God amounts to a contradiction. As far as Thomas is concerned, all that is required for the validity of such inferences is that there should be no equivocation between premises and conclusion. That premises and conclusion are related analogically can therefore place no obstacle in the way of the inference between them being logically valid.
If this argument may seem to relate with comparative indirectness to the issue of inference to an analogical conclusion, a second text, found in the Summa contra Gentiles, could not meet the point more squarely. There Thomas considers 'the opinion of those who say that God's existence cannot be demonstrated but can be held by faith alone', and in the course of doing so entertains Milbank's Aristotelian objection to his own view that God's existence is demonstrable: 'if the principles of demonstration have their origin in knowledge of sense, as is shown in the Posterior Analytics, what wholly exceeds every sense and sensible thing seems to be indemonstrable. But the existence of God is such. Therefore it is indemonstrable.'23
But Thomas rejects this counter-argument. If it were valid, he comments, it would prove too much. For on that account - 'if there were no substance knowable beyond sensible substance' - then nothing beyond natural science would be knowable, which even Aristotle denies. He adds that it can be no further objection to the validity of such proof that we
22 ST 1aq13a5 corp.
23 Aquinas, Summa contra Gentiles 1.12.
cannot know the 'essence' of God, and so cannot construct any non-equivocal sequence of premises entailing God's existence, since in proofs ofthe kind in question it is the divine effects24 which function as premises, not the divine nature.25 For we cannot construct an argument for God's existence out of premises definitive of the divine nature, as Anselm famously supposed, that nature being unknown to us - we are in possession of no definition of God in the first place.26 But if the arguments for the existence of God are constructed from premises descriptive of the effects of God in creation, and not from any definition of God, then of course the conclusion of such an argument will have to contain terms not univocally related to those of the premises; it could not be an argument for the existence of God if that were not so, but only for 'just another, creaturely, being'. Hence, the only tests of such an argument's validity could be those of logic; you could not rule out the argument's validity on the grounds alone that the conclusion contained terms not univocally related to the terms of the premises. That, in any case, is pure Scotism.
The same point needs to be made to those who would rule out such a possibility on the rather similar grounds of formal logic. It is often said that from premises employing sense-bound intra-mundane notions of cause (with whatever consistent univocity of sense) you could not in principle conclude to a non-sense-bound extra-mundane cause. For to be non-sense-bound and extra-mundane - and so God - the conclusion would have to contain terms so transcending in meaning that of the terms of the premises as to render the inference invalid. And there seems to be a general principle at stake here: surely, it will be said, the conclusion of a valid inference must be in some way 'contained' in the premises if the conclusion is to be validly 'extracted' from them. But how could God be in any way 'contained' in premises derived from creatures, derived as cause from effect, without God's thereby being conceived of as a cause within, and not of, creation? And how, if not 'contained' in the premises, could an inference from creation to God be justified? So it may well be said.
Of course, since Kant, nearly every philosopher, and as many theologians, have taken this objection to proofs of the existence of God to
24 This is an ellipsis. Kerr is, of course, right (see pp. 196-7 above) that you cannot prove the existence of God from what you know are the divine effects, because that is simply to beg the question. To prove the existence of God is to prove that creatures are 'effects' of a divine creating causality. Nor is Kerr entirely right when he comments that, in saying that the 'five ways' argue from 'effects' to 'cause', Thomas is evidently making creation ex nihilo a presupposition of their validity (Kerr, After Aquinas, p. 59). As I argue at pp. 239-42 below, for Thomas, the five ways do need to be taken in conjunction with the account of creatio ex nihilo in that what shows God to exist is just what shows the world to be created ex nihilo, and so that the world is a divine 'effect'.
be unanswerable. But perhaps one of the reasons it has been taken to be unanswerable is the very great degree of unclarity with which this so-called principle of deductive logic is promoted. The only completely transparent sense in which a conclusion can be said to be 'contained' in the premises of an inference is, once again, that of the petitio principii. Of course, 'if all the apostles are Jews and if Peter is an apostle, then Peter is a Jew' is a case of an argument in which the consequent is 'contained' in the antecedents. But since you would have to know that Peter is an apostle in order to know that the antecedent is true - that all the apostles are Jews - it is hardly the case that the consequent is thus derived from the antecedents. If anything, the major premise is (partially) derived from the conclusion already known. Otherwise than in a tautological case of this kind, there seems to be no very clear way of settling the question of how what is 'contained' in the premises of an inference is to be determined so as to rule out 'something else' appearing in the conclusion, otherwise than to say: a conclusion is 'contained' in a set of premises if and only if it follows from them by means of a logically valid inference employing non-equivocal terms. It might seem as if this is just to turn the tables on the opponent by begging the question. But it is hard to see why one may not do so, at least until some other sense is provided of the expression 'contained in the premises' which can be given a coherent meaning.
That being so, all we need is a logically valid proof of the existence of God meeting the following conditions: first, that no equivocation occur in the premises; secondly, that the conclusion contain terms which are not univocally the same as those contained in the premises, for otherwise the argument could not be said to conclude to God; nor alternatively may terms in the conclusion be equivocally related to the premises, for then the inference could not be logically valid. This, again, appears to be Thomas's view of the matter. For in further response to the 'Aristotelian' objection he simply says that its 'falsity is also shown by the effort of the philosophers who have tried to prove that God exists'27 - if a proof proves, then you will have to abandon any such a priori presuppositions as would entail its impossibility.
Such an argument, as I have said, would not be an argument by analogy from creatures to God, for, God's existence not being presupposed, no such analogy could, short of circularity, be presupposed in the premises, but only entailed in the conclusion. It would, therefore, be an argument to analogy, demonstrating a two-part conclusion. First, in demonstrating the existence of God it would demonstrate that God cannot be named by names univocally predicated of him and of creatures; and second, by
the fact of the argument's validity, it would follow that names of God lay in a degree of continuity with our names for creatures which ruled out their being equivocally predicated - for the validity of the argument would itselfrule out equivocity. Our saying that such names are predicated 'analogically' would therefore get its sense from this double conditioning: we know that we are justified in predicating existence of God from the success of the argument; we should know that the proposition 'God exists' has some meaning from what showed it to be true. But just that same argument's success would also demonstrate that, as predicated of God, we do not otherwise have any grip on what 'exists' in that case means. In short, such an argument would demonstrate simultaneously the need for, and the inseparably mutual logics of, both affirmative and negative theologies. It would thereby demonstrate the possibility and necessity of analogical predication of God, as it would also provide a sense for the expression 'the analogical predication of terms of God'. A term is predicated analogically of creatures and of God when we know from creatures that it must be true of God too, but also know that how it is true of God must be beyond our comprehension.
Was this article helpful?