Anselm's Discovery: A Re-examination of the Ontological Proof for God's Existence Argument
by Charles Hartshorne
Part Two
A Critical Survey of Responses to Anselm's Proof
- Anticipations of the Proof
- A Strange Story
- Guanlio
- The Scholastics: St. Thomas
- Descartes, Gassendi, and Hobbes
- Spinoza
- Ralph Cudworth
- Leibniz
- Hume
- Kant
- Hegel
- Ludwig Feuerbach
- Robert Flint
- W. E. Hocking, Josiah Royce, and George Santayana
- R. G. Collingwood
- Hans Reichenbach
- J. N. Findlay
- Robert S. Hartman
- Jan Berg
- Jerome Shaffer
- Heinrich Scholz and Frederic Fitch
- Conclusions
1. Anticipations of the Proof
In a remarkable article, Prescott Johnson (see Bibliography) seems to succeed in showing that Plato’s dialectic (in The Republic) as means to knowledge of the Good amounts to an ontological argument for the necessary existence of the Good. The lesser ideas are incapable of expressing the principle of order among themselves; in conceiving this order we are conceiving a supreme idea which therefore cannot be lacking in content for our thought. This is—says Johnson—the a posteriori element in Plato’s reasoning. In effect, it is his refutation of positivism. The supreme reality is not inconceivable. (I would here depart somewhat from Johnson by remarking that the conceivability of something is a necessary, not a merely contingent or factual, truth and that it cannot, properly speaking, be known a posteriori. But I shall not attempt to relate this consideration to Plato’s procedure.) The supreme reality is not to be treated as a mere hypothesis. Knowledge of it “requires no assumption,” and “makes no use of images, relying on ideas only.” In short this knowledge is strictly a priori. And no merely contingent existence could be thus known.
[140]
(Nor, I add, would it make sense to say of a contingent reality that it was “superior to existence,” as Plato says of the Good. Only a reality existing in an underived and necessary fashion could be to all things as the sun is to life on the earth.)
Johnson correctly defends Plato’s procedure against the Kantian criticism that the merely possible and the existent cannot differ qualitatively. In regard to the supreme conception, the merely possible is indistinguishable from the impossible. We have not to compare two states of the Good, one as nonexistent and the other as existent; for this is to treat the “beginning” or “principle” of all meaning, value, and reality as a mere possibility which might or might not be actualized. But if it can be so treated it is not the absolute principle at all, and the knowledge of it must be precisely the hypothetical knowledge which Plato contrasts with the highest knowledge.
Johnson does not discuss the relation of the Good to God. This is a difficule topic in Platonic scholarship. Perhaps one is fairly safe in saying that there are grave difficulties in either denying or asserting the identity of the Good with God. Moreover, when in the Timaeus deity is plainly under discussion, we again have ambiguity, for there we seem to confront two Gods, the ‘eternal God and the God that was to be’, or the Demiurge and the World Soul. I wish merely to suggest that Plato is wavering between classical and neoclassical theism, or between the view that deity is pure absoluteness or necessary existence and the view that deity is indeed absoluteness necessarily actualized somehow, but with the particular concrete how or actuality contingent and relative. What in Plato was an unresolved ambiguity, a wise restraint
[141]
in the claim to settle basic issues, in his followers tended to become a premature and unwise resolution of the ambiguity. This unwise resolution, which oversimplifies the religious idea and gives it a fatal one-sidedness, has exacted severe penalties all through the history of thought. One of the penalties has been the failure to clarify the Anselmian problem in a permanently satisfactory manner.
It was a rather close anticipation of Anselm when Aristotle declared, “To be possible and to exist do not differ in eternal things.”1 But Aristotle came even closer:
For what is ‘of necessity’ coincides with what is ‘always’ since that which ‘must be’ cannot possibly ‘not-be’. Hence a thing is eternal if its *being’ is necessary: and if it is eternal, its being is necessary.2
No ecternal thing exists potentially. The reason is this. Every potency is at one and the same time a potency of the opposite; for, while that which is not capable of being present in a subject cannot be present, everything that is capable of being may possibly not be actual . And that which may possxbly not be is perishable, either in the full sense or in the precise sense in which it is said that it possibly may not be. . . Nothing, then, which is in the full sense imperishable is in the full sense [‘in respect of substance’] potentially existent (though there is nothing to prevent its being so in some respect, e.g., potentially of a certain quality or in a certain place); all imperishable things, then, exist actually.3
Some things owe their necessity to something other than themselves; others do not, but are themselves the source of necessity in other things. Therefore the necessary in the primary and strict sense is the
1 Physics, 111, 4. 203, 30. I owe this reference and the literal transladon to my colleague A. P. Brogan. Oxford translation, “In the case of eternal things, what may be must be.”
2 De Generations, 11, 12, 338, 14.
3 Mataphysics, IX, 8. 1050". 8-18. Trans. W. D. Ross.
[142]
simple; for this does not admit of more states than one, so that it cannot even be in one state and also in another; for . . . it would already be in more than one [if it were in all respects necessary]. If, then, there are any things that are eternal and immovable, nothing compuisory or against their nature attaches to them.4
Nothing is by accident perishable. For what is accidental is capable of not being present, but perishableness is one of the attributes that belong of necessity to the things to which they belong; or else one and the same thing may be perishable and imperishable, if perishableness is capable of not belonging to it. Perishableness then must either be the essence or be present in the essence of each perishable thing. The same account holds for imperishableness also; for both are attributes which are present of necessity.5
Certainly Anselm did reason partly as follows: To exist eternally is better than to exist with a temporal beginning or ending; hence God cannot be conceived in the latter fashion; but only that which could not not exist is intrinsically, or for an intelligible reason, without beginning or ending, secure in its eternity. As a Socinian theologian later put it, “that is eternal which cannot not exist.” There is no other criterion for eternal existence, since one cannot wait forever to observe that a thing always goes on existing. Not even God Himself could know it in that way.
Actually Aristotle is superior to Anselm in some respects in this matter (save only that he did not turn his insight into a proof). For one thing, he makes a clear, and virtually neoclassical, distinction between eternity or necessity of mere existence (‘'in respect to substance’), and eternity or necessity with respect to all properties whatever. The former, he
4 Op. cit, V, 5, 1015b, 9-16.
5 Ibid., X, 10, 1059a, 0-7.
[143]
says, does not entail the latter. Precisely, and the neglect or underestimation of the former was the great error of classical theism, an error into which Aristotle himself fell, as is indicated in the next to last of the above quotations. The blunder was natural enough. There are: (a) things contingent or ephemeral both substantially, or as the individuals which they are, and also in the qualities not essential to their individual identity; (b) things (a thing?) existing eternally, necessarily, as the individuals which they are, but not eternal or necessary in all their properties or states; (c) as the extreme or pure case of necessity, things or a thing which could neither fail to exist nor ever in any way be other than it is. The superiority of (b) to (a) seemed clear to Aristotle (as it does to me); he apparently inferred that, by the same principle, (c) must be superior to (b). However, it is not the same principle at all. That it is better to be both contingent and necessary than to be contingent alone does not entail that it is best of all to be necessary alone. It is better for a saw to be both sharp and not sharp, the one in the blade, the other in the handle, than to be not sharp all over; but it is by no means best of all that it be sharp all over and have no handle. To lack necessity even of existence and be therefore wholly contingent is indeed a defect; but this is entirely compatible with its being also a defect to lack all contingent qualities and be wholly necessary. For, if the necessary as such is, as has been shown in this book, extremely abstract, a mere universal common factor, then to be purely necessary is to be purely abstract, totally lacking in concreteness, that is, richness of definite detail and variety.
True, Aristotle seems also to have reasoned: since ‘actuality is prior to potency’, the supremacy and ultimate priority
[144]
must belong to a purely actual being devoid of potentiality for further actualization. But we saw in Part One, Secs. 4, 5, that this assumes an irrelevance of quantity to quality and also of incompossible values to the supreme value, which are anything but noncontroversial. I pass over Aristotle’s suggestion that the supreme reality must be immune to influence by others, a typical piece of Neoplatonic scorn (is not Aristotle the first Neoplatonist?) for passivity, responsiveness, sensitivity to what passes in others.
In the above quotations the Stagirite not only makes an important distinction between two ways of conceiving necessary existence, but he also (1) gives reasons for idendfying modal status with temporally limited and temporally unlimited ways of existing, the contingent having being at most for some time, the necessary always, and (2) with lucid, subtle reasoning shows that such temporal-modal status is itself in all cases necessary, so that the contingent and perishable could not have been necessary or imperishable, and the necessary and imperishable could not have been perishable. This is a form of the modal reduction principle; modal status, including that of nonnecessity, is itself always necessary. And here, two thousand years in advance, is the answer to Kant's charge that the ontological argument must assume that existence is, in general, a predicate, whereas in general it is not. Rather, either contingency (perishableness) or its negative, necessity, (imperishableness) is inherent in any predicate whatsoever: modal status is always a deducible predicate. (See the last of the quotations on page 142.) Kant’s cultural lag on this point is two millennia. Surely Aristotle would have known what Anselm was talking about at least better than even the greatest of Gaunilo’s countless disciples.
[145]
Since neglect of the temporal aspect of modality, or the modal aspect of temporality, is a major defect of the European tradition, it is an interesting question, which I hope to pursue elsewhere, why Aristotle failed so signally to communicate his insights at this point.
Of the long and shameful story of the underestimation of Jews by Christians (in our day by communists as well), one of the least shameful but still interesting chapters is the underestimation of Philo. It is easy to say that Wolfson has exaggerated Philo’s importance; it is harder to find in medieval scholasticism a single statement about God—apart from the Incarnation and some points about the Trinity—which cannot be matched in Philo’s own words. Practically the entire religious metaphysics of fourteen centuries (including both best and worst features) is definitely Philonian. That the divine existence is necessary is repeatedly stated, as follows:
“The virtues of God are founded in truth, existing according to his essence: since God alone exists in essence, on account of which fact, he speaks of necessity about himself, saying, ‘I am that I am".6
“. . . the God who exists in essence, and who is duly thought of in respect of his existence . .."7
“. . .God who exists only in essence . ..”8
“. .. He is full of himself, and He is sufficient for himself . . ."9
Here we have the idea of existence as an identity, therefore
6 Works of Philo Judaeus, trans. C. D. Yonge (London, 1890), I, 282.
7 Op. ois.,, IV, 283.
8 Ibid., 11, 28-29,
9 Ibid., p. 243.
[146]
necessary. Recognition of existence in this superior form of sheer self-existence is not indeed new with Philo, for—besides being in Aristotle—it seems to be as old as monotheism itself, since Ikhnaton expressed it nicely, “Thou of thyself art length of life, men live through thee.” (And not of course men alone, as the grand old hymns make clear enough.) Thus the idea that God’s existence could be just another case of existence in general has always been a failure to comprehend theism. It is three millennia out of date.
The unsurpassability or perfection of God is indicated by Philo as follows:
It is impious to conceive that anything can be better than the cause of all things, since there is nothing equal to him, nothing that is even a little inferior to him; but everything which exists in the world is found to be in its whole genus inferior to God.10
. . . the living God . . . is superior to the good, and more simple than the one, and more ancient than the unity . . .11
His nature is entirely perfect, or rather God is himself the perfection, and completion, and boundary for happiness.12
I submit that these passages are at least as close to the Proslogium as anything in Augustine, who is usually cited in this connection. And they are about four centuries earlier! True, Philo apparently did not see that he had in such considerations the basis for a proof for the divine existence, but then neither did anyone else before Anselm. Augustine’s proof from the superiority of reason to all but Truth, and the identity of
10 Ibid., 1, 196f.
11 Ibid., p. 229.
12 Ibid., IV, 1-2.
[147]
God with Truth, is a version of what I call the epistemic or logical proof, not of the ‘ontological’, which must (if the label is to be of any use) be from the idea of God itself as intrinsically connoting necessity. Nor are proofs, such as Aristotle’s, from ‘degrees of perfection’ to perfection itself ontological. To use the term so widely makes virtually any proof ontological. The essential idea is not of kinds of things, one kind implying another, but of ways of existing, and of one way as selfrealizing and self-certifying, and as such alone appropriate to the all-worshipful being, hence either assertible or deniable on grounds of meaning alone.
A third (vague) aaticipation, though not in our Western tradition, is found in the Taoist principle that the supreme reality (the Tao) is like water, completely without exclusive form of its own, but able to assume the form of any vessel. This makes deity a correlate of being as such, rather than one form of being instead of another. Such absolute nonexclusive flexibility or absence of competitiveness is, I have argued (in this book and elsewhere), identical with noncontingency. The principle is also suggested in the analogy employed by the ancient monists of India that space can take the forms of all the objects in space, and hence is not itself limited by any of these forms. The important thing is to see that the perfection of cognitive capacity, infallibility, is even more clearly without exclusive form, since by definition it must be able to express any form whatsoever (in knowing it). Anselm failed to achieve clarity at this point, and his Greek cast of mind made this inevitable. He was typically Western in exalting ‘masculine’ mastery, power, stability, control, being, absoluteness, while depreciating the feminine, yielding, passive, fluid—that is, becoming and relativity.
[148]
Taoism and Buddhism are closer to the truth here. But Jesus was perhaps closer still. God’s sensitivity registers the fall of the sparrow. This occurrence is a modification of his sympathetic awareness. The absolute responsiveness of universal love is purely noncompetitive, hence in its bare existence wholly noncontingent.
All genuine thought about deity must, indeed, be close to the ontological proof. For it is blasphemous to think of God as merely an additional fact, however great, merely one side of a significant alternative, rather than as the soul of factuality itself and the very basis of all alternativeness, the potential registrant of whatever value or importance ether side of any disjunction can have, hence not subject to intelligible denial.
Ikhnaton (in spite of associating deity peculiarly with sun and sunlight) clearly thought of God as the strictly universal principle of meaning and value, measuring by His love all the forms of existence. He was the God of absolutely all creatures, not just of some. It is only a clarification of this to see that possibility itself must be expressive of the divine, and hence that the ‘possibility of there being no divinity’ formulates an absurdity.
The very notion of creator, introduced into philosophy in the Timaeus, implies the principle of modal equivalence just referred to. For if the reality of the Demiurge actualized a possibility capable of being unactualized then He must be Himself as much in need of a creator as anything else. Thus the Platonic, which is the oldest, formal proof for God breaks down if Anselm’s discovery is a mere sophistry.
And yet no one formulated his Proof or anything much like it before he did. Esser, in his monograph on alleged antici-
[149]
pations of the Proof (see Bibliography), considers a number of pre-Anselmian proofs and rightly dismisses them as not proofs from the mere idea of God. Esser ignores Philo and the most relevant passages in Plato and Aristotle; but I incline to agree with him that before Anselm there was no Ontological Proof.
Anselm’s formula for deity is perhaps less novel. Collingwood mentions some precedents in Boéthius, and he should have mentioned Philo and Augustine. (See Sec. 15, p. 250.) Nevertheless, here, too, Anselm remains distinctive. He alone puts sufficient emphasis upon the difference between greatest, or unsurpassed, in fact and not conceivably surpassable.
No one before Anselm gave so neat a formula for the divine excellence. Correctly interpreted, as to be sure he did not interpret it, it remains without a flaw, precisely as he stated it. I hold that this definition, and the deduction of noncontingency therefrom, constitutes the greatest single step forward in constructive metaphysics taken after Philo and prior to Leibniz. It is also the least understood, the most carelessly treated, by scholars,
2. A Strange Story
The history of discussions concerning the ontological argument might have been that of a collective inquiry into the validity of the reasoning of Prosl. II-IV, with reasonable account taken of later passages. This inquiry might also, after a few centuries perhaps, have led to the discovery of the abstract-concrete paradox as inherent, not in the Argument as such, but in classical theism, yet made more apparent by the Argument. All this might conceivably have happened.
[150]
What we find in fact is rather different: a story of prolonged debate largely, often exclusively, over the thinking of the fictitious Anselm of the Gaunilo tradition, that conveniently naive fellow who made his whole point in Prosl. II and added nothing relevant thereafter, in contrast to the historically demonstrable, keen-witted philosopher whose main point ficst appeared in Prosl. IlI, and was considerably amplified and carefully defended in still later discussions. And the last thing anyone saw clearly was the abstractconcrete paradox, the heart of the whole problem; the problem, however, not alone of the Proof, but of theism itself.
Such is the tale—'stranger than fiction’, though in a sense about a fiction—which we shall now tell, partly in the words of some of the chief participants. The story has at least a bappy ending, for it seems to show that the long-stretchedout farce is nearing its probable dénouement, and that the unconscious falsehoods about the magnificent doctor can scarcely retain their innocence, which has been their strength, much longer.
3. Gaunilo
Whatever the first commentator upon a philosopher may say, be it intelligent or otherwise, he will of course have been the first to say it. Moreover, as human nature is, the chances are that he will be praised for having said it (all the more if there is no other commentator for a hundred years). For his view will probably be a natural interpretation, or misinterpretation, of his subject, and for this reason, and also by the power of suggestion, others will, even more probably, say it after him. But he will always remain the ‘discoverer’.
[151]
Such are the reasons for the fame of Gaunilo. Any very much better reasons the specialists in Anselm, so far as I know, have not detected.
The many signs of limited perspicacity in Gaunilo have been painstakingly detailed by Barth. I shall mention only some of the most important.
Was it smart, or was it not a little stupid, to ‘point out’ to Anselm—what the latter had not only admitted but insisted upon as integral to his theory—that all sorts of things can be conceived as nonexistent? Anselm not only accepted the empirical fact that we often think about nonexistent things, he had a theory, which is more than we know Gaunilo had, about why all things other than God are thus thinkable, or what constitutes the ground of their contingency. Should one then offer the bare truth of general contingency as basis for rejecting the Proof? This is common ground, and is quite compatible with viewing God as the great exception. By definition God is exceptional. Yet to this day, as in the pages of Gaunilo and Kant, reminders of the universal contingency of existence, apart from God, are supposed somehow—it is never quite explained how—to give the Anselmian pause! True, they might well give him pause had Anselm stopped his exposition of the Proof at the end of Chapter II. For only in III is the uniqueness of necessary existence made explicit.
Gaunilo is the originator of the logically bizarre idea, also still widespread, that the admission of God's existence as a fact must be the premise for any proof that He also exists necessarily. There is, unfortunately, some suggestion of this way of looking at the matter in the opening of Prosl. III (following, as it does, upon the alleged proof that God exists in Prosl. II): “thou dost exist so truly that thou canst
[152]
not be conceived not to exist.” But later passages, long before the Reply, and above all in the Reply, make manifest the principle (evident in itself to reflection): facts cannot determine what is logically conceivable; hence it is absurd to talk as though an inconceivability of nonexistence could depend upon a thing’s existing in fact. Necessarily existing in fact? Then the thing could not but exist in fact, and there is no possible nonexistence to disprove by factual inquiry. Contingently? Then we have a contradiction.
What it comes to is that the idea of unconditional or strict logical necessity of existence never makes its appearance at all in Gaunilo’s discourse. Only the words, ‘cannot be conceived not to exist’, are there, not what the words mean. (Those who have accepted the monk’s refutation often fail to include either the idea or the words.) Gaunilo also betrays himself by talking as though the nonexistence even of ordinary things, or especially of oneself, may be inconceivable. This merely confirms, from the other end, his blindness to the modal concepts which are Anselm’s preferred intellectual instruments. Of course, if there is no contingency, there cannot in any distinctive sense be necessity either—and vice versa. Over and over Anselm talks in modal terms, while Gaunilo never clearly does so. Where Anselm says, none greater can be conceived, Gaunilo speaks repeatedly about none greater, or greatest, simpliciter. He allows God hypothetical preeminence in the realm of actual existence, but not in that of possibility or conceivability. Accordingly, to ask, Does God exist? must be to ask whether or not some possibility logically prior to God happens to be actualized. But the Unsurpassably great, hence all-inclusive of actuality and possibility, self-existing no matter what else does or does not
[153]
exist, can have nothing logically prior to it. It is that not to think which is to think nothing, and so not to think. This is Anselm’s world of ideas. Gaunilo has not entered it, even for a trial run. Nor has he, Samson-like, pulled its pillars down from inside. He has not been inside, but only in a fragile and readily detachable antechamber.
The strongest objection to Anselm’s procedure, never clearly seen by Gaunilo, concerns the apparent transition from abstract to concrete, from a brief verbal definition to the incomparable richness and beauty of the divine actuality. The pertinent question is whether or not existence and full actuality are in general coincident and equally concrete; and if not, whether they might not in God be exceptionally distinct and far apart (all aspects of God are somehow exceptional ), so that the existence to be established by the Proof would be as abstract as the unprovable divine actuality is concrete. Had Gaunilo the faintest glimpse of this possibility? Have Gaunilo’s admirers glimpsed it either? Yet is it not relevant, if our question is not merely, was Anselm justified in using his Proof to support his teligious beliefs, but rather, could it justifiably be used to support some religious belief or other, and if so what? This is the more general question; intellectual progress takes its greatest leap forward when higher generality is attained. The fate of classical theism may in the long run be a picayune detail compared to the fate of theism as such.
Gaunilo is abundantly justified, and it is his most relevant point, in asking, Do we really have “the idea of that than which none greater can be conceived” (rather than merely the words)? We have seen (Part One, Sec. 4) the ambiguities and possible contradictions which must be looked to in trying to adjudicate this question. But these are logical
[154]
questions and have nothing to do with particular perceptual facts, to the lack of which Gaunilo refers. If God’s very existence were relative to particular perceptual facts, then He would not be God but an idol. The problem is one of meaning: is worship self-consistent, or is it either contradictory or too indefinite an attitude even to contradict itself?
Considering the reception which Gaunilo has had to this very day, one is moved to ask: has ever a commentator upon a philosopher so long and so much misled so many?
4. The Scholastics: St. Thomas
The reception of the Argument in the twelfth and thirteenth centuries was almost as odd as what happened in its inventor’s own lifetime, or in the modern period. (In this section I am heavily indebted—and deeply grateful—to P. A. Daniels. See Bibliography 1.) In the twelfth century the Proof was simply ignored, so far as our records go. Three conclusions have been drawn from this: all accepted the Proof, all rejected it, they were unacquainted with it. Daniels shows that the last is the most reasonable. In the next three centuries things were dramatically different. Fifteen authors refer to the Proof, of whom the following ten accept it: William of Auxerre, Richard Fischacre, Alexander of Hales, Bonaventura, Matthew of Aquasparta, Johannes Peckham, Nicolaus of Cusa, Aegidius of Rome, William of Ware, and Duns Scotus. Of these at least four, Alexander, Bonaventura, Nicolaus, and Scotus seem to have some appreciation of Prosl. IIl and of the true Anselmian Principle; the rest seem to be thinking largely or exclusively of Prosl. II. Albertus Magnus, Peter of Tarentaise, and Henry
[155]
of Ghent take no position on the Proof; of these, only the first scems to have read past Prosl. II. St. Thomas and his disciple Richard of Middleton reject the Proof; Richard cites only Prosl. II, while Thomas refers (in five different writings) sometimes to this and sometimes to the following chapter; however, where he is explicitly rejecting the Proof (in the two Summas) he mentions only Prosl. II; and where he does mention the other chapter he, in my opinion, misconceives the relationship of the two.
We have then fifteen medieval judges, of whom at most five show that they have the Principle clearly and centrally in mind; one or two others exhibit some conception of it, and the rest, little or none. Of the five having the Principle (as Anselm did) clearly and centrally in mind, four accept the Proof, and the fifth takes no stand. Of the other ten, those who seem not to grasp the centrality of Prosl. III, six accept, two reject, and two give no verdict. Thus even where the Proof was taken at its weakest, still six found it convincing and but two rejected it; and where it was taken at its strongest, four out of five accepted and none rejected it. This seems to show the power of the Proof even when incompletely grasped, and its much greater power when fully grasped. It also shows the blighting influence of Gaunilo’s inability to read beyond Chapter II.
Unfortunately, the example of Thomas has in the end outweighed in prestige all the others put together. Bonaventura's cogent rebuttal of Gaunilo’s ‘island’ analogy has been passed over as though it had never happened, while the objections of Thomas have been treasured. Bonaventura may have been somewhat to blame for this. After grasping the true Principle, he attempted to improve upon it and,
[156]
by a series of steps, reduced it to ‘God is God; therefore He exists’, thus, as Gilson says, “Simplifying the dialectic to the vanishing point.”
How well did Thomas understand the Proof? He seems scarcely to have seen at all how the essential step of the reasoning is back from Prosl. Il to Prosl. 1I, rather than, as admittedly Anselm seemed for a moment to think, in the reverse order. The key to the whole proof is the connection between perfection and the unique kind of existence which is essential, necessary, or self-existence. Thomas knows this connection in his own philosophy, but he denies Anselm’s right to make use of it for an ontological inference. His reasons we shall consider presently; meanwhile, the point is that to assume that this must not be the reasoning is to beg the question. And Thomas in effect makes this assumption. He implies that we must first, as in Chapter II, prove that God exists, and then, as in III, infer that His existence is of the necessary type. But since the first step is not cogent, the whole Proof must be invalid. Accordingly, in refuting Prosl. II in the Summa Theologica (see below), he thinks he has disposed of Anselm’s proposal. But this is not very perceptive, is it? Like contingency, necessity must be a property knowable a priori, the difference being that whereas from the modality ‘necessity of existence’ existence is deducible, from the modality ‘contingency’ it is not. How does this depend upon Prosl. II? By implying that it does, this most influential writer put—or seemed to put—the stamp of his approval upon Gaunilo’s worst mistake, his failure to see that unless one can disprove the deducibility of ‘imperfect’ from ‘contingent’, [and hence (modus tollens) of ‘noncontingent’ from ‘petfect’], he has not dealt the Proof any mortal blow.
[157]
An important feature of Thomas’s atdtude, which distinguishes him to advantage from Kant, is that he does not go to the unwise extreme of denying outright that God is immanent in our thought of Him. For, as he says, all creatures tend toward their good, and God is the Good of all goods. So all creatures are directly related to God and may even, if sentient, be said directly to grasp or experience God. Only, thinks Thomas, this is not a clear or cognitive grasp, as (in the case of the lower animals at least) seems obvious. Hence, runs the reasoning, we cannot use this immanence of deity in our experience as basis of a proof. We have to turn to features of created things, such as change or degrees of value. But if we cannot elevate our direct sense of the supreme Good into a premise for a theistic proof, what reason is there for thinking we can do better with other aspects of experience as means toward so exalted an objective? By scorning the nonsensory awareness of deity in favor of mere sense perception, Thomas is preparing the way for the débacle which theistic proof-making met at the hands of Hume and Kant. He is the greatest single preparer of this misfortune, and Anselm is the man who (properly read) could have prevented it.
Thomas goes part way with Anselm in admitting that of course ‘the nonexistence of God’, adequately understood, is contradictory. By means of other proofs we know that God exists in such a way that there is in Him no separability between essence and existence. So we do know this much about the essence of God, that ‘existence pertains to it’ (Spinoza), but we know this not immediately and directly (through faith) as Anselm seems to imply, but in a slightly roundabout way via knowledge of the creatures. Is this really so wise as many have held? Science has been probing ever more deeply into the
[158]
creatures, and in some ways they seem more mysterious than ever. And if it be said that for a theistic argument we need only understand the creatures as such, the bare idea of a creature, then that almost comes around to the Anselmian position, that one goes a priori from an idea to the divine existence. For what does it add to say, ‘the idea of a creature known to exist’? Would a merely possible creature, or the concept of creature, be independent of deity? The creator is necessary if creatures are to be so much as possible! Why should not the self-understanding of faith be at least as reliable and pertinent here as our ability to see into the meaning of ‘change’, for instance? If change does, in principle, involve deity, then to understand change as such (existent or not) is to understand deity to that extent.
There is, however, another way of interpreting Thomas'’s attitude. Perhaps what he obscurely felt was that Anselm had no right to presume that he did understand his faith. And here, I believe, Thomas would have been correct, for no classical theist as such can understand himself, since his identification of God with the absolute, or with ‘pure actuality’ and selfsufficiency, is a confusion. But if this is Thomas's ground for rejection, then he is himself in the same fix as Anselm. For he too was a classical theist, guilty of the confusion in question.
In the light of the foregoing, the Thomistic objection that we do not know the essence of God, and hence cannot make use of its oneness with His existence, is both relevant and not relevant. It is relevant in that Anselm (like Thomas) was committed by implication to a notion of the divine essence (the total absence in it of relativity or becoming) which could not be right. The objection, however, is not relevant, in that Anselm’s definition, when its ambiguities are left open rather
[159]
than resolved classically, is merely the religious truism that it must be impossible for God to have a conceivable superior. It is here faith which defines the religious question; and this question turns out to be self-answering. To tell us to turn away from faith to the mere existence of the world is to forbid us to take as our problem the rational testing of faith. It is to allow unfaith to set the question. But in metaphysics he who sets the question largely determines what answers can be given.
[Refutation of Proslogium II]
It seems that . . . as soon as the significance of the name God is understood, it is at once seen that God exists. For by this name is signified that thing than which nothing greater can be conceived. But that which exists actually and mentally is greater than that which exists only mentally . . . Thetefore the proposition God exists is selfevident.
On the contrary, no one can mentally admit the opposite of what is self-evident. . . . But the opposite of the proposition God s can be mentally admitted: The fool said in his heart, There is no God. . . .
I answer that a thing can be self-evident in either of two ways: on the one hand, self-evident in itself, though not to us; on the other, selfevident in itself and to us. . . . Therefore I say that this proposition,
God exists, of itself is self-evident, for the predicate is the same as the subject, because God is His own existence, as will be hereafter shown. Now because we do not know the essence of God, the proposition is not self-evident to us, but needs to be demonstrated by things that are more known to us. . . . namely, by His effects . . .
Perhaps not everyone who hears this name God understands it to signify something than which nothing greater can be conceived. Yet, granted that everyone understands this name . . . nevertheless, it does not therefore follow that he understands that what the name signifies exists actually, but only that it exists mentally. Nor can it be argued that it actually exists, unless it be admitted that there actually exists
[160]
something than which nothing greater can be thought; and this precisely is not admitted by those who hold that God does not exist.13
That what the fool says in his heart cannot be self-evidently wrong is correct only if ‘self-evident to us’ means, selfevident to any human being able to say the words with a feeling of understanding them. But this is a rather crude view; Anselm’s discussion of the point in the Reply is subtler. A mere feeling of understanding what we say is not a guarantee that we do understand. Logical relations need not be obvious at first glance to everyone. I fail to see that Anselm had anything to learn from Thomas in this respect. Similarly crude is the disjunction, knowing the essence of God (which in classical theism meant knowing all that God is) and knowing nothing at all about His essential nature, not even that it excludes a conceivable superior. No more is needed for the proof!
The last two of the quoted sentences obviously have relevance at most to the nonmodal form of ontological proof. Even in that reference we seem to be given little but a dogmatic denial that the proof is valid.
A writer who admits that the Thomistic distinction between the two ways of being self-evident does not furnish a good premise for refuting Anselm offers a different passage as turning the trick: No difhculty befalls anyone who posits that God does not exist. For that, for any given thing, either in reality or in the understanding, something greater can be conceived, is 2 difficulty only to him who concedes
13 Symma Theologica, Qu. 2, Art. 1, trans. Fathers of the English Dominican Province. Second and revised edition. (London: Oates & Washburn, 1920), Part I, p. 20.
[161]
that there is in reality something than which a greater cannot be conceived. (Summa contra gentiles. L 11.) 14
The significance of this is explained as follows: atheism may be put in a disjunction: either (a) there is nothing than which a greater cannot be conceived, or (b) there is such a thing, but ‘in the understanding only’. Even if we grant to Anselm that (b) is contradictory, still—it is urged—(a) is certainly free from contradiction. But let us see. (a) may be offered (a1) as a contingent, or (a2) as a necessary truth, If contingent, there must be no logical impossibility in the existence of a not conceivably surpassable being. But since, according to (a), there is in fact no such being, its nonexistence is also taken as possible. It would follow that the not impossible existence of an unsurpassable being could only be contingent existence. But, as we have seen, a contingent being could not be unsurpassable. Thus (a1) is contradictory. Atheism is no contingent truth. There remains (a2). A necessity that, given any being, a greater can be conceived implies the logical impossibility of an unsurpassable being. This, however, is the positivistic not the atheistic tenet. Moreover, if a concept is logically impossible, this can be no mere truth of fact. Modal statements themselves, as Aristotle saw, have the mode of necessity, not of contingency. We conclude that atheism (the merely factual denial of God’s existence) is not saved from contradiction by Thomas. Neither the divine existence nor the divine nonexistence could be a mere fact, i.e., a contingent truth. The question is conceptual not observational. Anselm correctly located the theistic
14 G. B. Matthews, “"Aquinas on Saying that God Doesn’t Exist,” The Monist, 47 (1963), pp. 472-477.
[162]
issue in the logical landscape. Did Aquinas? If indeed the tenability (or at least initial plausibility) of positiviim was his objection, this never becomes very clear in his discussion. (And Gaunilo had already made the point quite as definitely.)
In Thomist circles one often encounters some such formulation as the following:
The idea of God, the infinitely perfect being, does include existence, but only ideal, not real, existence. Therefore, it would be a contradiction if I were to think the infinitely perfect being without thinking it as existent, because I would be affirming and denying existence in the same order (Ordnung); however, a contradiction is not present if I attribute ideal existence to the most perfect being, while leaving the question open whether it exists in ontological reality. (Lehmen, Lebrbuch der Philosophie, Freiburg, 1901, BIL, p. 547, quoted in Esser— see Bibliography.)
This way of talking makes me wonder how stupid I perhaps am, for I can make no clear sense at all out of what is said. What is ‘ideal existence’? Merely that something is thought to exist? (For a legitimate distinction between conceptual and real existence in terms of the contrast between the that and the how of actualization, see Part One, Sec. 18.) But then a necessity to think infinite perfection as (ideally) existing is the necessity to think that it is thought to be thought to exist—and so on. And besides, the defect with which Anselm’s second Argument (against which Esser quotes the above passage) shows that Greatness cannot be combined is the conceivability of failing to have real existence. A being whose not really existing is conceivable is inferior to one whose not really existing is inconceivable. Therefore it is precisely real existence which must be taken as
[163]
inseparable from Greatness. What special merit would there be in ideally existing necessarily while really existing contingently? And if one can only think divinity as really existing then atheism is not thinkable, and only a positivist can reject the conclusion of the Argument. Is that what Lehmen and Esser are trying to say? Then let them for pity’s sake say it. For it is painful to be unable to find sense in what must seem sensible to the many who write in this way.
When it is suggested, as by Esser (p. 36), that while we must think God as existent, still we may also think that he perhaps does not exist ‘in the real order’, I derive from such formulations only this: we must think divinity as existent, but we may also think the proposition, ‘divinity may not exist’. Once more my intelligence fails to arrive at a coherent meaning. Is it our old friend, ‘God exists necessarily if he exists at all’? This seems implied by Esser (p. 35). As I have argued in various places, this expression also means nothing clear and consistent. If it only means, God either fails to exist or else exists eternally and without dependence upon any other existent, then I think (a) it misuses ‘necessarily’, and (b) it implies a radically unintelligibile form of contingency, ie., that something is but might not have been, yet no cause enabled it to be or furnished its real possibility. To be able not to exist yet to owe one’s existence to no actual condition is a combination of ideas that gives me for one ‘logical seasickness’. In addition I have given many reasons for denying that the ‘nonexistence’ of something is conceivable unless the something is competitive, partly exclusive, in its essential nature, so that another thing could exist in its place. But there is no ‘place’ of God which another thing could occupy instead of Him. I deny that an argument can
[164]
be refuted by formulations so full of paradoxes as those just considered.
On the whole, Thomism sheds not much light and some darkness on our topic.
5. Descartes, Gassends, and Hobbes
According to Gilson, it is not provable that Descartes had read Anselm. He may have taken the Argument from Thomas, which would explain why he put it in the weaker form of Prosl. II until, under Gassendi’s prodding, he came to his own version of the second and stronger argument. In any case, there can be no doubt that he knew some of the usual objections —this being the least that everyone has known who ever discussed the subject! On the whole, Descartes did not reach Anselm’s level in this matter. He did, however, furnish an interesting reason for taking the idea of God to be logically admissible. If we doubt, and hence realize our cognitive imperfection, it must mean something to talk about a degree of clarity and distinctness which excludes all doubt, that is, the divine clarity, infallible or omniscient awareness. But Descartes weakened his argument here by claiming—or seeming to claim—absolute clarity and distinctness for us human beings in certain cases.
[Echo of Prosl. I]
Being accustomed in all other things to distinguish between existence and essence, I readily believe that existence can also be disjoined from the essence of God, and that God can therefore be conceived as not actually existing. Buc on closer study, it becomes manifest to me that it is no more possible to separate existence from the essence of God than . . . the idea of 2 mountain from that of a valley. ... Nor may it be objected that though it is indeed necessary to grant that God exists, provided the supposition has antecedently been made
[165]
that God possesses all perfections and that existence is itself one of these perfections, the supposition is not . . . itself necessary . . . It is not indeed necessary that I should at any time be dwelling on the idea of God. None the less, as often as I may be concerned to entertain the thought of first and sovereign being, . . . I must necessarily actribute all perfections to Him . . . And as soon as I take notice that existence is a perfection, I am thereby constrained to conclude that this sovereign being truly exists . . .15
[Echo of Prosl. III?]
In the idea or concept of a thing existence is contained, because we are unable to conceive anything except under the form of an existent; that is, possible or contingent existence is contained in the concept of a limited thing, but necessary and perfect existence in the concept of a supremely perfect thing . . . Necessary existence is contained in the nature or concept of God.
Hence it is true to say of God that necessary existence is in Him, or that God exists.16
Descartes’s critic, Pierre Gassendi, in the following ‘objection’, gives a lucid anticipation of Kant’s principle that existence is not a predicate (or, as Gassendi puts it, a ‘perfection’).
[Refutation of Prosl. II]
You (Descartes) place existence among the Divine perfections, without, however, putting it among the perfections of a triangle or of a mountain, though in exactly similar fashion, and in its own way, it may be said to be a perfection of each. But, sooth to say, existence is a perfection neither in God nor in anything else; it is rather that in the absence of which there is no perfection.
15 Descartes’s Philosophical Writings, seiected and translated by N. K. Smith (London: Macmillan & Co., 1952), pp. 243, 244. [The stronger form of Cartesian ontological argument is omitted from this edition.]
16 From Second Replies to Objections, trans. T. V. Smith and Marjorie Grene. In From Descarses to Kant (Chicago: The University of Chicago Press, 1933), pp. 161-62.
[166]
. that which does not exist has neither perfection nor imperfect:lon, md that which exists and has various perfections, does not have its existence as one of the number of its pcrfecuons, but as that by means of which the thing itself equally with its perfections is in existence . . . Hence neither is existence held to exist in a thing in the way that perfections do, nor if the thing lacks existence is it said to be imperfect (or deprived of a perfection), so much as to be nothing. 17
Here we have one more refutation of the ‘First Ontological Argument’ (to speak with Malcolm). Nonexistence may be no defect in a thing (for there is no thing in the case). However, Anselm had argued, contingency in an existing thing is certainly a defect; and where it could not be, neither could contingent nonexistence, but only sheer impossibility or necessity of existence.
Descartes in his Replies brings this out, though not so fully as Anselm had done:
Nay, necessary existence in the case of God is a true property in the strictest sense of the word, because it belongs to Him and forms part of His essence alone. Hence the existence of a triangle cannot be compared with the existence of God, because existence manifestly has a different relation to essence in the case of God and in the case of a triangle. 18
Did Kant know this exchange between Gassendi and Descartes? It certainly would have been relevant. That existence
17 Philosopbical Works of Descartes, trans. E. S. Haldane and G. R. T. Ross (New York: Dover Publiations, 1955), 1, 186. My colleague Douglas Morgan has pointed out to me that Aristotle had explicidy denied that ‘existence’ is a predicate. Thus it is as clear that Aristotle would have rejected Anselm's first argument as it is that he could scarcely have rejected the second. See Metaphysics 1003 b 264.
18 Op, cit., p. 228. [The Everyman edition of 1922 also contmins this stronger or “second” version of the Proof.]
[167]
is not, in ordinary or contingent cases, a (deducible) property is without prejudice to the status of necessary existence, which is different in kind and inapplicable except to God.
The only advantage Descartes has over Anselm is in the phrases, ‘necessary’ and ‘possible [i.., contingent] existence’. The ideas are certainly Anselmian. And, as we saw in Part One, Sec. 23 (1), Anselm’s language has its merits.
Not only did Descartes encounter a lucid anticipation of Kant’s chief criticism (and rebut it), he also dealt with a clear formulation of the favorite twenteth-century (Wittgensteinian) contention that rational necessity derives from language, and asserts nothing about extralinguistic reality. In other words, modality is purely de dictu and not de re!
Consider the following:
Hobbes: “Reason gives us no conclusions about the nature of things, but only about the terms which designate them, whether, indeed, or not there is a convention (arbitrarily made about their meanings) according to which we join names together.”
Descartes: . . in reasoning we unite not names but things signified by names; and I marvel that the opposite can occur to anyone . . . For, if he [Hobbes] admits that words signify anything, why will he not allow our reasoning to refer to this something that is signified, rather than to words?”'19
It is true that the ontological proof was not the topic which Hobbes had introduced here; but it is clear, I think, what Descartes would have said had that been the topic. I submit
19 Op. cit., Vol 11, pp. 64f. For this reference I am indebted to Dr. Bowman Clarke.
[168]
that it is time for historians to tell us the facts in this case, which are that, down to Wittgenstein, the Anselmian problem had been left nearly where it was in Descartes’s time. Real novelty came with Barth, Koyré, Hartshorne, Findlay, and Malcolm. Most of the others have been treading old trails, often with the airs of pathfinders.
If I overstate, please remember that the contrary overstatements are in a thousand textbooks and works of reference!
What Hobbes actually said (in his Fifth and Seventh Objections) about Descartes’s proof for God, without specifying whether it was the ontological or the proof from the axiom that there must be a cause for the content of every idea, was that in fact we have no idea of God. In short his view on the issue was the positivistic one—an intelligent position.
The logical possibility and consistency of ‘infinite perfection’ follows—Descartes argues—from the fact that he has a clear and distinct idea of it. In short, he has an absolute intuition at this point. Leibniz rightly rejected this claim, as Hobbes had done. And indeed, since Descartes holds to the sheer unity of the divine reality, an absolute intuition of the divine nature should endow us with omniscience at one blow. Moreover, even if, with neoclassical theism, one admits real distinctions between elements and aspects of the divine life, one still is not entitled to put absolute trust in our intuitive grasp of even the most abstract aspect of deity. One cannot oscillate between the appeal to logical inference as against intuition and the appeal to intuition as against inference, in this arbitrary way. The possibility of inferring inconsistency must remain open, and so must that of having a genuine though not infallible intuition of consistency. Leibniz is more penetrating on this issue, though as we shall see, not much more.
[169]
Descartes, however, also argues (condescending to his opponents’ lack of intuition) that perfection is the positive idea, and imperfection derivative from it by negation. One cannot, he thinks, impugn the positive case and leave meaning for the negative. True, he must then defend himself against the observation that ‘infinite’ is negative, and so derivative from the positive idea of finitude or limitation. He reacts by denying the trustworthiness of the linguistic indication in this particular case. By the divine infinity we do not intend a numerical or quantitative unlimitedness, but a reality wholly positive, lacking nothing. There is some plausibility in this, but yet the reasoning is unsound. Limitaton implies a negative element; but it is no mere negation. For, since there are mutually incompatible yet positive values, as we shall see more particularly in discussing Leibniz, to reject all limitation (being this and not that, or that and not this) is to lose all concrete definiteness and become indistinguishable from mere indeterminate potentality for positive value. The linguistic indication which Descartes spurns here might have led him to a great discovery, had he taken it seriously. ‘Perfect’ can be given a wholly positive meaning, but this proves it to be irreducible to sheer infinity. These terms are not synonyms. Perfect in the appropriate religious meaning implies the impossibility of being surpassed by another. This is, insofar, negative. But one can equivalently say it implies the necessity that the X defined as perfect should surpass any being other than itself. Or: In any possible state of reality, X surpasses y, whatever y may be. Sheer infinity is not deducible from this definition. For such a deduction one must assume that a wholly infinite being could also be all-surpassing in richness of actual value, and against this is the argument that to be wholly
[170]
infinite is to be wholly indefinite or indeterminate (as Thomas put it, a wholly ‘indeterminate sea of being’); and it is just not self-evident that this can be distinguished from the totality of the logically possible, in entire abstraction from any actualization. One may verbally stipulate that it is wholly actual, but then one has two absolute infinities, pure possibility as such and ‘pure actuality’ as such. (For some such reason David of Dinant and still others argued that God and prime matter were indistinguishable.) If this is not a paradox, what would be? Self-evident consistency cannot be claimed for any such idea.
Descartes’s weakness, as user of the Argument, was his uncritical acceptance of Classical or Neoplatonic theism. In this he was like countless others, Anselm included.
Descartes, like Bonaventura, Spinoza, Malebranche, and Leibniz, indeed probably all the defenders of the Argument, assigned a different origin to the idea of God than to ideas of sensory things. This nonsensory origin was the point of ‘innate ideas’. Locke's criticisms, as Peirce said, implied a meaning for the phrase which ‘nobody’ had intended. (It may also be fair to say that the intended meaning was not as clear as it ought to have been.) It was of course primarily his sensory theory of knowledge which barred the way to Thomas's acceptance of Anselm’s proposal. According to all the Ontologists, the idea of God comes from God, not merely from creatures to us as creatures, but direct from God to us as His creatures. We do not exclusively infer God, we experience Him spiritually and intellectually. Inferences in this field are, in part at least, reductio ad absurdum arguments against the denial that we know or have God all along, and cannot simply not know or have Him. Bonaventura and Malebranche even
[171]
emphasize the direct awareness of God to such an extent that the Argument is no longer an inference at all, but merely the recognition of the divine givenness as self-existent. (On this point, Jalabert is very illuminating—see Bibliography.) The error consists in not seeing the value of the intellectual experiment by which we test the assertion that God is given with this character of necessary existence. Human intuitions are not so clear, or so easy to put into words without danger of confusion, that we can rest content with their mere assertion. We must also investigate logical relations among our other nonsensory ideas. It turns out that to deny the givenness of God is to deny His existence (for, being conceived as universally immanent, He could not be conceived as nevertheless not given), and that this in turn is to deny even His logical possibility. The positivist is willing to make this denial. But this question, too, is further arguable, and there are reductio ad absurdum arguments concerning it.
What, however, does ‘nonsensory’ amount to? One way to put it would be this. To understand arithmetic or logic one needs no special sense organs, or special physical environment perceived through such organs. One may, if one is not God, require some sense organs or other and some external environment capable of supporting one as a thinking animal. But it makes no difference what the organs or the environment are like, provided thinking (presumably with something like language) can develop freely on a sufficiently conscious level. Given this, arithmetic and logic can arise and be understood. So can the idea and knowledge of God! But the ideas of ‘oxygen’ or of ‘vertebrate’, with the meanings that these terms actually have, are in a different class. Certain special perceptions are required. The geometrical shape of a back-
[172]
bone is not enough to fulfill the requirements of ‘vertebrate’, and it is impossible to put all the requirements into an a priori definition. For we mean something historical by the term, as also by ‘oxygen’, something which has actually arisen in our cosmos in our cosmic epoch, something whose nature is permapently more or less hidden from us, and must be referred to by denotation, by pointing, not by pure description. On the contrary, to grasp what number is, or what God (in His purely necessary aspect) is, one needs no special historical reference or special perceptual experience whatsoever, but only the intelligence to grasp the most universal aspects of absolutely any kind of experience or history. If this is what ‘innate’ (or not sensorily produced) means, then I hold that Locke was mistaken: there are such ideas, and very important they are. No one ever held that we are always conscious of them, or even that all men can be infallibly led to the consciousness of them by any course of argument or instruction. Some men resist arithmetic, or have scarcely had need to consider it beyond the uttermost extreme of simple cases, and there are strong forces opposing the careful consideration of God. But it remains true that no special sensory (including emotional) experiences are logically required, any more for the idea of God than for the ideas of arithmetic.
True, one must have some emotion or other, because the idea of value is involved. God is unsurpasssably great, and great here means having or being ‘whatever it is better to have or be than not to have or be’. But the purely general idea of value, or of better, is all that is required, and it is absurd to suppose that anyone would bother to think if nothing seemed better to him than anything else. Indeed the notion of a value-free experiencing or thinking is nonsensical or contradic-
[173]
tory. To have some emotion or other, and some sensory experience or other, is presupposed by any thinking at all (in spite of so-called thinking machines). Even God, in Neoclassical Theism, has something analogous to sensation. An innate idea is not one which could arise with no sensations or feelings, but one which logically could arise no matter what the sensations or feelings, provided they favored conscious thinking, including thinking about thinking, on sufficiently complex levels, and with sufficient freedom from inhibitions.
How far this is what the rationalists meant I shall not further inquire. It is, I suggest, at least as close to what they meant as anything which Locke set up to attack.
6. Spinoza
Spinoza seems to have been the first, though he was not the last, to employ the Proof in support of another doctrine than classical theism; in his case, a more rigorously formulated version of classical pantheism or Stoicism. In one sense this was logical. If God is a superconcrete yet wholly necessary being, then all concreteness must be within Him—otherwise He is but an abstraction from the total reality—and since, on classical assumptions, nothing contingent can be in Him, all things must be necessary. But then the distinctive meaning of ‘necessary’ is lost!
About the Proof itself Spinoza had no misgivings, and he is well beyond the mere notion that existence is better than nonexistence. His proof of Proposition VII rests on the idea (which had been indicated by Anselm, and developed by Scotus and Thomas Bradwardinus) that what cannot be caused by another cannot exist contingently, which is a version
[174]
of the true Anselmian Principle. Also the first auxiliary proof for Proposition XI is a way of stating still another version, apparently not clearly seen by the magnificent doctor, that contingent existence is competitive, the things which exist preventing (‘annulling’) the existence of many otherwise possible things. Here Spinoza seems to have made a genuine advance. However, he could not do justice to this approach, since if all things are necessary, ‘otherwise possible things’ has no clear meaning.
Spinoza blithely assumes as a matter of faith (he says, ‘intuitive knowledge’) the logical possibility of divinity as he defines it, not noticing how paradoxical it is that he has ostensibly deduced the ‘necessity’ of the entire concrete totality of things from an abstract definition. The inconsistency in the notion of a necessary yet concrete reality is left out of the reckoning.
[Echoes of Proslogium III?]
Prop. VIL It pertains to the nature of substance so exist.
Demonst. There is nothing by which substance can be produced (Corol. Prop. 6). It will therefore be the cause of itself, that is to say (Def. 1), its essence necessarily involves existence, or in other words it pertains to its nature to exist.
Prop. XI. God, or substance consisting of infinite atsributes, each one of which expresses eternal and infinite essence, necessarily exists.
Demonst. If this be denied, conceive, if it be possible, that God does pot exist. Then it follows (Ax.7) that His essence does not involve existence. But this (Prop. 7) is absurd. Therefore God necessarily exists.
Another proof. For the existence or nonexistence of everything there must be a reason or cause. . . . and if it does not exist, there must be a reason or cause which hinders its existence. . . . If, therefore,
[175]
there be no reason nor cause which hinders God from existing, or which negates His existence, we must conclude absolutely that He necessarily exists. But if there be such a reason or cause, it must be either in the nature itself of God or must lie outside it, that is to say, in another substance or another nature. For if the reason lay in a substance of the same nature, the existence of God would be by this very fact admitted. But substance possessing another nature could bave nothing in common with God (Prop. 2), and, therefore, could not give Him existence nor negate it. Since, therefore, the reason or cause which could negate the divine existence cannot be outside the divine nature, it will necessarily, supposing that the divine nature does not exist, be in His Nature itself, which would therefore involve a contradiction. But to affirm this of the Being absolutely infinite and consummately perfect is absurd. Therefore, neither in God nor outside God is thete any reason or cause which can negate His existence, and therefore God necessarily exists.
Another proof. Inability to exist is impotence, and, on the other hand, ability to exist is power, as is self-evident. If, therefore, there is nothing which necessarily exists excepting things finite, it follows that things finite are more powerful than the absolutely infinite Being, and this (as is self-evident) is absurd. . . .20
Spinoza’s view is that all things, granted their causes, are necessary; though in ordinary things, the causes are outside their own natures. But then there is no contrasting term to necessity, no true contingency. The totality of things is as necessary as God. To retain the contrast upon which the meaning of ‘necessary’ depends, we need to admit that things other than God are ultimately contingent. And God transcends their contingency only by taking it wholly into Himself so that it becomes the infinite tolerance of the divine life for alternative states. To this divinely flexible alternativeness there can, therefore, be no alternative. The absolute ‘patience’ of
20 Ethic Demonstrated in Geometrical Order, trans. W. H. White (London: Triibner & Co., 1883), Pt I.
[176]
God for the variety of existence constitutes His immunity to nonexistence. .
However, Spinoza is right in saying that a thing can only fail to exist if something ‘prevents’ it from existing; in other words, all facts are partly positive. But he misconstrues this to mean that effects are necessitated by their causes. They are necessarily prevented from existing if the causes are sufficiently unfavorable, but even favorable causes cannot reduce to zero the creativity—involving contingency—which is becoming itself. One must also take temporal aspects into account in considering how one thing may prevent another from existing. Granted that creativity could, at a certain point, have taken another course than it has, the course actually taken henceforth excludes this other possibility. It was, but no longer is, an ‘open possibility’. But no possibility is closed unless the realization of some incompatible possibility has closed it. Only the future is still an open possibility, and an eternal being, which can never be merely future, can never be an open possibility. Whatever made its nonexistence a fact would also make its existence impossible.
God's absolute “power to exist” is His ability to assimilate any and every causal condition, to make it ‘favorable’ to some appropriate responsive state of His own awareness. This is the opposite of being influenced by nothing other than Himself. Nothing can be merely other or alien to God; all have something ‘in common’ with Him. This neoclassical view is about equally far from that of Spinoza and that of Anselm.
7. Ralph Cudworth
Here is one of the few writers who have arrived at a good
[177]
understanding of the essentially modal structure of the Argument. What is still more rare, he has apparently done this without deriving his understanding from Prosl. III and the Reply. His source seems to have been Descartes (or perhaps Henry More, whose ideas were similar), but Cudworth makes the modal considerations more explicit than Descartes did. He sees that something which could not exist contingently also could not contingently fail to exist; so that to deny its existence is to assert its impossibility.
If God, or a perfect Being, in whose essence is contained necessary existence, be possible or in no way impossible to have been, then He is: . .. for if God were possible, and yet He be not, then is He not pecessary but contingent Being, which is contrary to the hypothesis.21
[This necessity of God's existence] must not be taken hypothetically only . . . that if there be anything absolutely perfect, then its existence was and will be necessary; but also absolutely, that though contradictious things cannot possibly be, and though imperfect things may possibly either be or not be; yet a perfect Being cannot but be; or it is impossible that it should not be.22
How melancholy to reflect that so clear and firm a grasp of the subject (buried, alas, in the third volume of a diffuse and rather formless work dealing with many topics, some of little present-day interest) was published a century before Kant's Critique, and that now, nearly three centuries later, so much of the philosophical world is still (as Barth puts it) “stuck in the dialectic of Prosl. II,” that is, in the nonmodal or false version of the Proof.
21 The True Intellectual System (London, 1895), vol. iii, p. 49. [This edition contains long notes by J. L. Mosheim, which seem to exhibit most of the misunderstandings to which an ontological argument is open.]
22 Op. cit, p. 40.
[178]
Of course Cudworth lacks any realization of the abstract-concrete paradox, and he asserts rather than explicates the incompatibility of perfection with contingency. On this second point, Anselm is more helpful.
8. Leibniz
No more than with Descartes, Spinoza, Cudworth, or most of those we shall have to deal with is there evidence that Leibniz knew the contents of the Proslogium (after the by themselves scarcely intelligible first two chapters). My guess is that he did not. He was, however, too much a metaphysician to be wholly victimized by the Gaunilo tradition. Like Scotus, but first among the moderns, he sees the need of establishing the logical possibility of the theistic concept, and he attempts to meet this need, partly by connecting the problem of logical possibility with the principles which he believes underlie logic generally, an intelligent procedure, if it can be carried through successfully. Like Thomas (and all the great theists) he is clear that if we know anything at all about God we know that He could not exist contingently. Unlike Thomas, but like all the Ontologists, he disbelieves in the sensory origin of the most universal conceptions. God is a direct datum of the soul, always given but not always attended to.
One of the best-known passages of Leibniz concerning the Argument betrays the persistent influence of Anselm’s initial blunder (see Part One, Secs. 6, 19, 20).
To exist is something more than not to exist, or rather, existence adds a degree to grandeur and perfection, and as Descartes states it, existence is itself a perfection. Therefore this
[179]
degree of grandeur and perfection . . . which consists in existence, is in this supreme all-great, all-perfect Being. . . . The Scholastics, not excepting even their Doctor Angelicus, have misunderstood this argument. . . . It is not a paralogism, but it is an imperfect demonstration, which assumes something that must be proved . . . that is, it is tacitly assumed that this idea of the all-great or all-perfect being is possible, and implies no contradiction. And it is already something . . . [to have] proved that, assuming that God is possible, He exists, which is the privilege of divinity alone. We have the right to presume the possiblity of every being, and especially that of God, until someone proves the contrary. So that this argument gives a morally demonstrative conclusion, which declares that, according to the present state of our knowledge, we must judge that God exists.23
In the opening sentences Leibniz seems to be repeating the Prosl. II confusion between saying that the fact of not existing would be a defect in deity and saying that even the bare possibility of not existing would be a defect. Of course the fact entails the possibility, hence whatever excludes the latter excludes the former; but this complex relationship, not the direct and simple exclusion of nonexistence, is the point of the Argument. Only necessary existence can enter into the notion of a kind or essence. In another passage Leibniz does somewhat better.
En disant seulement que Dieu est un étre de soi ou primitif, ems & se c'est-d-dire qui existe par son essence, il est aisé de conclure de cette definition qu'un tel étre, 8'il est possible, existe; ou plutét cette conclusion est un corollaire que se tire immédiatement de la définition et n'en différe presque point. Car I'essence de la chose n'étant que ce qui fait
23 New Essays Concerning Human Understanding, trans, A. G. Langley (New York, 1896), pp. 502f. Reprinted, LaSalle, Illinois: Open Court Publishing Co., 1949.
[180]
sa possibilité, il est bien manifeste qu'exister par son essence, C'est exister par so possibilité Et si I'étre de soi émit défini en termes encore plus approchants, en disant que c'est I'taxe qui doit exister, parce qu'il est possible, il est manifeste que tout ce quon pourra dire contre I'existence d'un tel étre serait de nier sa possibilité.24
Here we have something a little like Prosl. III. Essential or self-existence is the unique existence of deity. But Anselm retains the advantage of having made it clear that one does not ‘define’ God as the necessarily existing being (which leads to the objection that one must then show what necessary existence has to do with divinity), but rather, having defined deity as the worshipful, hence Unsurpassable, one then from this definition derives the trait of necessary existence. Thus it is the Unsurpassable which necessarily exists, not merely the necessarily existent which necessarily exists. Surely the former procedure is superior!
That Leibniz is not very sensitive to the difference between mere existence and necessity of existence as a ‘perfection’ is partly due to certain features of his theory of existence which are well elucidated by Jalabert (see Bibliography). Every essence ‘tends toward’ existence, and insofar includes existence in its concept; but ordinary essences are partly in conflict; they compete with one another for existence, and only Sufficient Reason, or the divine affirmation of the Best compossible set of essences, resolves the conflict. The divine essence, however, is not in competition with the others,
24 Untitled essay on Father Lami's proof for the existence of God. Die philosophische Schriften von Gottfried Wilbelm Leibniz, ed. Gerharde (Berlin: 1887), vol 4, pp. 405-406. [In this and the following quotations I have followed the spelling and punctuation given by Jalabert (p. 82— see Bibliography).]
[181]
being quite independent of them. And its tendency to exist is infinite, and hence could not fail to be fulfilled. ‘Tending to’ means, ‘will if nothing prevents’. With the divine essence, and only with it, nothing could prevent. Hence, a priori, God exists.
This is a much subtler theory than Spinoza’s. Indeed it is the supreme effort of a classical metaphysician to employ the Argument. To say that an essence exists if no competing one does is to say that no fact can be merely negative, that the nonexistence of X is the existence of a world incompatible with X. This, I hold, is quite correct. Spinoza, too, takes nonexistence to imply that the thing is ‘prevented’ from existing; but he spoils this by also saying that all possible things exist, hence there are no such things as competing possibilities. Leibniz wants to preserve the contrasts (with the collapse of which the Argument loses all point) between: that which is necessarily existent, that which is possible but not existent, and that which is possible and also (contingently) existent. Any theist who does not recognize these distinctions is cutting off the limb upon which he would perch.
Had Leibniz been willing to qualify Sufficient Reason, so that truly arbitrary decisions, both divine and otherwise, could be admitted, and to follow through by admitting contingent divine decisions as real qualities of deity, that is to say, had he been willing to break with two thousand years of adamantine resistance to the idea of divine accidents (contingent predicates in God), the distinction between competing world possibilities and the noncompetitive essence of deity might have enabled him to solve the Findlay paradox and legitimately employ the Argument. But what monstrous proposals he would have thought these! Nevertheless, the question remains, can he
[182]
meet the conditions of the Argument on his own more traditional foundations?
Does it really have a coherent meaning to say: (a) there is no impossibility in any of the various incompossible essences, taken one by one, and they are therefore severally really possible; (b) no essence could exist without a divine decree favoring it; (c) only one decree is compatible with the divine wisdom and goodness; (d) that God and His wisdom and goodness exist is logically necessary? Russell holds that this adds up to a miserable self-contradiction or quibble; I think it is a magnificent attempt to find a difference where no room has been left for one. Leibniz is clear that a possible world is correlative to a possible divine decree favoring it; but in what sense is a decree ‘possible’ to a being whose character strictly forbids his making it? If the divine character as eternally necessary uniquely designates one world as best, and therefore alone worthy of existence, is not God by His very essence in competition with the remaining possible worlds? But clearly, in such a competition with absolute necessity, a possibility is simply no possibility! Thus I hold with Russell, and in spite of Jalabert’s skillful pleading on Leibniz’s behalf, that the great attempt is a flat failure. It was, however, a truly great actempt. Consider the problem: there are unrealized possibilities, not all could be realized together, yet severally all are genuinely possible, so that there is no ‘metaphysical’ necessity in the world's detailed character; however, there must be adequate reason for that character; how can this be? A happy thought: suppose one possibility is supreme over all others. God would then—and only then—have a reason for its selection. And God, at least, could not act irrationally or without adequate reason!
[183]
The grandeur of Spinoza was that he made no pretences about saving contingency, but sternly stuck to his denial of it. This, however, was moral even more than intellectual grandeur. For the untenability of such a one-sided modal theory is only too manifest. Leibniz saw that one must at least do better than that. But, alas, Leibniz was not the man to pay the immense price of challenging tradition as rudely as Spinoza had done, though in another and more appropriate place, in its denial of contingent divine properties. Nor was he the man to be so out of tune with the confident science of his age as to resist the attraction of the Phantom of Sufficient Reason of which Spinoza also, in his own way, was a victim. Some things may perhaps be contingent, but this must not mean that they are without ultimate reason. Yet is not that what ‘contingency’ does mean, if anything?
An ultimate reason for the entire world, just as it is, must be exactly as definite and complex as the world. It cannot be the mere abstract notion, ‘best of all’. By a stroke of genius, the dilemma seems to yield to solution: as possible this world is just as definite and complex as the same world as actual; it also, in this definite complexity, surpasses every other possible world. Hence, by what appears almost as magic, a definite world in all its details does deduce itself, so to speak, from the mere abstraction ‘better than any other possible world’. The feat that Spinoza could not quite manage appears to have been accomplished. Spinoza had had to admit that his ‘modes’ do not follow directly from God's essence but rather from antecedent modes; only the totality of modes follows from the essence. But how? What connects the concrete particulars with the bare abstraction, ‘absolutely infinite substance’? Only the verbal bridge: ‘absolutely infi-
[184]
nite reality must consist of all possibilities exhaustively actualized, hence there can be no unrealized possibilities’. But how empty or circular this proof is. If it shows anything it is only this: the intelligibility of Spinoza’s definition of deity (suspiciously like that of other Classical Theists) stands or falls with the intelligibility of ‘all things are necessary’, or ‘what is possible also is’. However, what reason does Spinoza have for assuring us that both stand, rather than both fall? Only his alleged intuitive understanding, which is his secret, or at most that of Classical Metaphysicians generally. Thus the doctrine is a rather transparent though unconscious bluff. In Leibniz the bluff is harder to see through. But is it anything beyond a more complicated bluff?
To allow the theory to pass is to agree not to attack the logical admissibility of ‘one among possible worlds is best’. Yet why should it be any more admissible to suppose this than that one number is greatest, or one possible velocity the swiftest—formulae which, as Leibniz rightly points out repeatedly, one may employ with a feeling of understanding, though when examined they turn out to be logical absurdities. He will not allow Descartes to assume the internal coherence of his definition of God, but asks that this be proved. He makes the same demand of Spinoza. But ‘best possible world’, this he will not undertake to prove consistent (if so, where?) but rather he wants us not to notice the problem of consistency at this point. And how does he know that the demand for an ultimate reason for the nonnecessary is anything more than the demand that the supposedly nonnecessary should be shown necessary after all?
Leibniz argues that if we renounce Sufficient Reason we give up all hope of proofs for God, or indeed of rational knowledge.
[185]
However, he here overlooks some important distinctions. On the one hand, if there is genuine contingency there cannot be an ultimate reason specifying which possibles are actualized. Actualization is brute fact, capricious, undeducible. It is not a concealed syllogism of any kind. On the other hand, for possibilities themselves there must be a reason. Possibilities are not capricious or arbitrary, but rational. Deductive reason is concerned with the rationale of possibilities, not of actualities. We must, however, distinguish between ‘pure’ or eternal possibilities, and spatio-temporally localized possibilities: what is possible in a given time and place. The pure possibilities are wholly rational, and there is no caprice in them. If something is eternally possible, then this is a necessary not a contingent truth. No choice, decision, or selection is presupposed. But what is possible here and now depends upon what has previously happened, including the arbitrary decisions previously made. We can always ask, what events, what brute facts, put the world in a situation in which what could happen had to fall within such and such limits? Real spatiotemporally localized possibilities are narrowly limited, as compared to the realm of pure possibility; this difference, these limitations, constitute causality and make science possible. But the entire difference between pure and real possibility is ultimately arbitrary, brute fact and nothing else. Not that a chaos of innumerable arbitrary decisions could miraculously add up to a world-order. This would indeed be unintelligible, and Leibniz would be right to reject this. What saves the world from being such a chaos, however, is not that there is an ultimate reason for the particular limits of real possibility, ie., for the particular causal laws, but rather, first there is indeed an ultimate reason and necessity that there should
[186]
be orderly limits or laws of some kind, and, second, the preeminence of certain arbitrary decisions, those of God (called in the Bible ‘fiats’), means that all other decisions are swayed and kept within appropriate limits by the influence of these supreme decisions. Like traffic laws, ‘natural laws’ have arbitrary features; but it is not arbitrary that there are laws, and that there is a power to make them. For this, the Reason is indeed sufficient. Mere chaos is not among the possible worlds. Nor is any possible world a godless world, which would be an unordered order, and nonsense. (This is the form the ‘argument from design’ ought to take, but the subject transcends the scope of this book.) So I think there is no need to admit that the principle of rationality which connects the world to God must be the ultrarationalistic one implicitly destructive of contingency itself. And certainly the ontological argument depends upon no such ultrarationalism; quite the contrary, it cannot be well defended without abandoning that doctrine.
Before asking whether the supreme axiom of rational understanding, taken to be that of Sufficient Reason, is absolutely true, or must be renounced or weakened, we should ask the prior question, Is Sufficient Reason itself a sufficiently reasonable formulation of the supreme axiom of rational understanding? I hold that it is a quite unreasonable one. By requiring the principle of rational derivability to apply to everything in every aspect, one prevents it from applying distinctively or effectively to anything. This is the usual penalty for one-sided simplifications in metaphysics. (The contrary one-sidedness of Humian empiricism, which is still very much with us, has its penalties also. It just happens that our age is not so sensitive to them.)
We may be grateful to Leibniz for having put his superb in-
[187]
tellect into the effort to do the impossible, to justify the unjustifiable. For this helps us in our search for the truly justifiable. By making clear the best that could be said for ultrarationalism, Leibniz put us in an excellent position to evaluate that way of thinking. The verdict seems clear enough; the doctrine is mistaken.
The sane principle of sufficient reason implies as corollary that pure possibility consists only of possible kinds and degrees of causal order: a causeless chaos is eternally impossible. Hence, when it was said above that the entitre difference between pure and real possibility is arbitrary, this was not in conflict with what was subsequently said about the element of rationality in the form of divinely-imposed laws. For God could not fail to exist, nor could He fail to impose some order. He does not arbitrarily decide that there shall be order, but only what order. This last, however, is, from an ultimate point of view, simply arbitrary. It has to be so, for the notion of an ulterior reason leads to an idle regress, or to the denial of contingency and therewith any intelligible necessity also. ‘From an ultimate point of view’ means this: in a given situation only certain kinds of decision would be appropriate, and for rejecting inappropriate kinds, there is a sufficient reason. But the situation itself grew out of prior decisions (not those of God alone) and to ask, ‘What is the reason for the entire series of decisions?”’ is to ask a nonsensical question. The design argument so put is, I hold, fallacious. No wonder Hume and Kant could make little or nothing out of it. To imagine God before all creation deciding upon the whole once for all is to slide into the eternalistic dream in which nothing really makes sense. Creation is not a one-step process. Merely possible worlds set no soluble problem, even to the
[188]
All-wise; for where no conditions are fixed, there is no definite problem to solve. Only a world already in being gives God anything to decide about. Not that the world was there before He decided anything, but that His prior decisions and their results were there before any given decision of His, or of anyone else’s. The old doctrine, which Leibniz professes, that God's initial creation and His preservation of the world are but a single act, is a confused one, and it assumes the dubious idea of a beginning of the creative process itself. But all beginnings presuppose the process, not the other way. There is then no one act of creation and no one act of preservation; creation consists of definite steps one after another. Just so did the Myth present the matter. Theologians (with some exceptions) thought themselves wiser. But in truth they had their own intellectualized myth, worse not better than the popular one in not a few respects. Leibniz is the last great philosopher who was really confident about the intellectual myth in question. Our mandate is to do better, by looking in another direction.
Leibniz’s procedure was cleatly not to survey the possible worlds and see that one was best. How could he do this? He had a notion of what makes one world better than another, namely that it integrates more diversity. But ‘best’ would then have to imply an unsurpassable diversity unsurpassably well unified. And what can be meant by unsurpassable diversity, whether or not ideally unified? A greatest possible number of kinds of thing? But since Leibniz himself denies a greatest number, number must be left out of it. If quantity cannot be maximized, then it must be totally abstracted from in conceiving the best world. Also, since there are incompossibles, oot all possible diversity could be integrated. It seems safe
[189]
to conclude that the notion of ‘best possible’ was not independently established, but was deduced from another notion whose consistency was held beyond challenge. This was, of course, Sufficient Reason. Unless one possibility is best, actuality cannot have a reason. But this is hardly a proof that one possibility is best. Since the Principle of Sufficient Reason, applied to the contingent, leads to the reduction of Leibniz’s optimism to the paradox of Spinoza’s necessitarianism, Leibniz has no right to assume the Principle and thence deduce the intelligibility of his theory that one possibility is best. If the latter is to be questioned, the former certainly cannot be secure, and both are dubious, if not clearly false.
Our search for explanations rightly rests upon a hope of finding them, but explaining things by deriving them from premises or conditions is not the absolute end or principle of existence, which is rather happiness (inclusively the divine happiness). Deriving happiness from causes or reasons is something else again. Doing so may increase happiness, but is not its whole content. The interest of logicians in derivations, and of scientists in causal predictions and postdictions, is legitimate; but there are other values, and the simple absolutizing of this one interest (or cluster of interests) in an alleged principle having unlimited right of way is at best a huge risk, and at worst a vast folly. It is not a harmless axiom to be put beyond question.
There is another notion that one must agree not to challenge if Leibniz’s bluff is to remain uncalled. One must admit that a possible world is as definite and complex as the corresponding actual one. This, I hold, reduces the distinction between possible and actual to nullity. Value is in definiteness, and definiteness is ‘the soul of actuality’. Were pos-
[190]
sibility equally definite it would be redundant to actualize it. There is no definite set of sets of ‘possible things’, and therefore it cannot be asked which set of possible things is best. Only properties, not particulars or instances, are possibilities (see below, Sec. 21). Moreover, only the most abstract properties are eternally distinct possibilities among which God might eternally choose. And even then, it is not really a question of choice, since all abstractly possible world forms have a claim to be realized, ‘each in its due season’. But the infinity of these possibilities is such that no actual world process could ever exhaust them. In addition it is a dubious assumption indeed that God can simply choose a world order without leaving anything open for the lesser creative powers of the creatures to further determine.
It is time to turn to Leibniz’s admitted problem of establishing the consistency of his idea of deity.
He has two devices. One is to refer to his Universal Characteristic as showing that simple notions or their objects, ‘simple perfections’, cannot conflict. But this Characteristic never quite came into being, and so far as I know it has not been found possible to construct anything like it on the required basis of simples. The idea of ‘simple perfections’ can only refer to empty abstractions like ‘knowing everything in all its aspects’, or ‘loving all things for all they are worth’. These indeed are mutually compatible, but what about knowing the various possible kinds of worlds (not possible worlds, for there are none) as actual? Here one confronts the exclusive or: this and then not that, or that and then not this. Not all possible sorts of worlds can be actual, hence not all can be known as actual.
[191]
The other device is the following:
Si l'éae nécessaire émit impossible, tous les étres contingents le seraient aussi, et ainsi il n'y aurait rien de possible. Car les étres contingents n'ayant point en eux la raison suffisante de leur existence, il faut recourir a I'Etre nécessaire qui est wltima ratio rerum, la dernidre raison des choses.25
Suppose we grant that there must be something existing necessarily, it does not follow that this something is consistently conceivable as Classical Theists (and Leibniz) conceived God. Therefore, the right of theists of this kind to employ the Argument is not established. I hold that they have no such right. And I think the sole advance Leibniz has made here is to have pointed his finger over and over again at the weakest point in the Anselmian reasoning, the failure to establish the logical possibility of the religious idea as defined. But it never occurred to Leibniz that the price of making good this failure might be to start over again with an untraditional way of defining God, or an untraditional way of construing the traditional definition of God as uniquely perfect.
That the Findlay paradox was not apparent to this philosopher is no cause for surprise. The rationalistic dream of Sufficient Reason implies, so far as it has a clear meaning at all, that there is no paradox, that the eternal and necessary is really completely concrete, and that abstractness is only a human illusion. But if abstractness is an illusion, so is concreteness; for, like necessity and contingency, these concepts stand or fall together. There is a paradox, but unqualified rationalism makes it insoluble. (So does unqualified empiricism.)
25 Jetter to Jaquelot, Nov. 20, 1702. Ibid., p. 44.
[192]
To have the right to employ the Argument, one must make sun-clear that one is not playing fast and loose with the contrast between necessary and contingent. The necessity of God is distinctively and uniquely His, which means that nothing else (except the widest class of individuals as such) has any ultimate necessity, whether one calls it metaphysical, logical, or moral necessity. The existence of God, alone among individual existences, is an eternal or a priori truth, which means that other truths of individual existence (and specific kinds of existence) are not eternal and not a priori. But we find Leibniz holding that for God (who sees things as they are, nota bene) all truth is eternal and a priori. Therewith the game is up, the initial standpoint abandoned, and all clarity lost. To favor the distinctively a priori truth of theism is precisely not to favor the a priori truth of things in general. It is to insist upon the general absence of such truth. Spinoza and Leibniz both quite missed the point here. And of course Anselm had preceded them in the error, since he too believed that in God all truth is timelessly contained in the a priori self-vision of the divine essence.
I feel I must emphasize and insist upon this, since even Boyce Gibson, generally sympathetic as he is to neoclassical theism, seems to feel that the Argument belongs with an eternalistic point of view.26 I say, just the contrary, it belongs with the view that only the bare essence and existence of God, taken as an extreme empty abstraction, is ‘timeless’ or a priori. All else is empirical and in some fashion temporal—even for God,
26 Gibson, A. Boyce, “The Two Strands in Natural Theology.” The Moniss, 47 (1963), pp. 335-364. [In this excellent introduction to what I all neoclassical theism the ontological argument is rejected, though without discussion.]
[193]
indeed especially for God, who does not share the illusions of ultrarationalists! God is not concerned to give ultimate ‘reasons’, whether to us or Himself, for His decisions. He makes decisions such that no other could be better, that is, such that the decisions are unsurpassable for the given situation (concerning all situations at once there is no decision, but only an eternal abstract ideal). But that there is just one best possible decision is a groundless, confused notion. Possible entities, including decisions, are not so definite as to be relatable in any such fashion. Only actual decisions are fully definite. All that is required for divine reasonableness is the negative, it would not have been possible to do better. (This does not mean that a better state of the world at that point in its development was impossible; for the creatures’ decisions also enter into the result, and there is no metaphysical necessity that they should always obey the maxim, never do less well than possible. For their goodness or rationality is not to be defined as perfect.)
Only a concretely temporalistic theology can rightfully employ the Argument. The others will always compromise the uniqueness of the divine eternity and necessity, and the genuineness of its contrast with all else. It is only creative, selfenriching process which can unite harmoniously within itself an abstract, necessary eternal aspect with concrete, contingent, ever partly-new actual states. The contrast with the contingent which is essential to necessity must be within the divine life, not merely between it and something else. Everything must be within that life, without prejudice to the vast and genuine contrasts involved.
Here we may revert to the problem, common to Descartes, Spinoza, and Leibniz, of the ‘simplicity’ and wholly ‘positive’
[194]
character of the divine perfection. Descartes argues that the religious idea cannot be self-inconsistent, since it is simple and without negative elements. Spinoza, pointing out that determination is always (partly) negative, seems to think that, nevertheless, although God must include all the complexity of things, somehow this need not mean that He is limited by any negations. Leibniz has his way of trying to turn the same trick. But since God knows all things, the complexity of the content of His knowledge can only be the greatest complexity there is; moreover, since kinds of worlds are possible which are not actual (if ‘possible’, ‘actual’, ‘necessary’—any of our basic categories—have meaning) and since obviously God knows as actual only the world which is actual, and does not know as actual any of the kinds of world which are merely possible, determination and negation must be admitted as constitutive elements in His reality.
Apart from the ultra-optimism of the Best Possible World, what most discredited Leibniz’s system for his contemporaries and successors was surely his denial of interaction between monads. Even this, I submit, is incongruent with the proper understanding of the Argument. For it put Leibniz in the following position. God’s existence is that upon which all monads utterly, and by strict logical necessity, depend in their every aspect; this absolute dependence, however, is not the supreme example of a principle generally operative in the system; it is rather the sole example. A sole example is not a supreme example. How often this is forgotten! We do not exalt God by giving Him a unique category, like creative power, for His very own. For if A is simply incomparable to B, then it is not inferior or superior to B but simply—incomparable. This is the real objection to a ‘deus ex machina’, where
[195]
no general principle of the system can mediate the comparison between God and other things. The unsurpassable power of God should be the supreme form of ‘power’ in the general sense, exhibited elsewhere in inferior degrees or ‘resemblances’. But Leibniz wants to say, not that there is a weak dependence of one monad upon another and a strong dependence upon God but that there is a zero dependence of one monad upon another (in the logical sense of dependence, and that is what is in question), but infinite dependence of all upon God. True enough, the whole issue is subtly befogged by the doctrine of ‘ideal’ influence via the divine choice of the most harmonious and richly varied whole. But this is like the befogging of the contingency issue by the distinction between direct logical necessity and indirect logical necessity via the necessary goodness of God and its logically predesignated result in the best possible world. These are not solutions but subtle and brilliant evasions. Thus it is stll true that one jumps from monads absolutely independent logically among themselves (or, if you want to insist upon the interdependence, via the divine goodness, monads absolutely interdependent to monads absolutely dependent upon God. There is either no difference or no real comparability at all. How is all this related to the Argument? I hold that the independence or necessity (it is the same) of the bare divine existence (not actuality) from the creatures, its sheer neutrality as between one world and another, implies a supreme dependence of the creatures upon the divine existence (for mutual independence here would be purely absurd) and that this in turn implies nonsupreme forms of dependence, which could only obtain between one creature and another. So Leib
[196]
niz's ‘denial of windows’ was really antitheistic (like every other metaphysical mistake!).
It is notable that our principle, the supreme must not be the sole form of a category, is implicitly Leibnizian. In some ways he adhered to it better than any man before him. Thus he treated the divine freedom as the supreme form of something found also in every creature, namely being moved by the apparent good. Determination by the apparent good is his theory of the will, and it applies even to God. (Of course, with God the apparent good is simply the good, but then even with us there is more or less conformity between appearance and reality.) And Leibniz, to his lasting glory, will not have any weak dualistic compromise, according to which some creatures are moved by the apparent good, and others, mere bits of matter, are not. For though resemblance to God can have all degrees, and the total gamut of these degrees is the same as the range of possible diversity among creatures, the zero degree coincides with nonentity, since being a creature at all means having some positive relation to deity. Here I salute Leibniz for being the first great philosopher (some minor figures in the Renaissance had said it, but not with power) to have really grasped a basic metaphysical truth. There is no room for caprice in considering the mere possibility of creatures; if to be a creature means (speaking loosely) to express something of the nature of God (and what else? ), then anything supreme in God’s essence may be as near to zero, to total absence, in a creature as you please, but present in some positive degree it must be.
What spoiled the picture was that Leibniz denied to both God and creatures any genuine creativity. Giving the nod to the completely defined possible world which most of all
[197]
‘requires’ actualization is not creation and is not freedom. God puts the penny in the right slot, and the right thing comes out. Rather, possible worlds are vague directions for further determination, and the process (to which each creature contributes) of determining the antecedently indeterminate but determinable can never end, since each new creature opens up new real possibilities for advance, and in this way the potentialities are inexhaustible. God has the supreme form of creativity, creatures have lesser forms. One cannot ascend to the divine form of a category one simply lacks. Leibniz (he was in good company here) managed never to intuit creativity at all. His appetition or force which made the monad a ‘spiritual automaton’ is the exact denial of creativity. The dream of explaining and justifying particular things by some ultimate reason or necessity, a superstition if ever there was one, condemned Leibniz to this blindness.
The denial of ‘windows’, though unique to Leibniz, is connected with another implicitly antitheistic tendency of Rationalism, and indeed of most modern philosophy, untl recently at least, which is the calamitous notion that an experience can have simply itself as datum. The monads experience —what? Their own ideas and sentiments, that is, their own mental states, plus—God. God is actually given, not merely an idea of Him, but God Himself. And this must be said, for otherwise the whole system falls to pieces. Sorething besides one’s own experiences must be experienced, or solipsism is not only irrefutable, it is unintelligible how it could even be an issue. And if God is that upon which all absolutely depend, if He is constitutive of the very possibility of things, not to experience God would be to experience nothing at all. But if God is our sole datum, in that sense in which a datum is not
[198]
the experience itself but the thing it is the experience ‘of’, then once more we have a ‘supreme’ which is also ‘sole’. There is but one genuine datum, God. But then nothing mundane can mediate the transition from no datum to the supreme datum. Leibniz would doubtless say, one’s experience of oneself is the mundane datum. But this, in the system, is something absolutely different. For Leibniz, a monad is a single subject of predicates and its self-relations are then identities. But the relation to God is no identity. Here there are two subjects. There is still no mediation.
Like other Classical Theists Leibniz must hold that God’s awareness of the creatures is sheer self-awareness. He intuits His own essence, which includes His choice of the best possible world. But either-or: God could have refrained from selecting that world, so that His selection itself is a contingent reality (meaning, its nonbeing was possible), or He could not have refrained, and then we have Spinozism: all is strictly necessary. If the selection is contingent (and otherwise the world is not contingent) then God must know something contingent in Himself, and this cannot be His mere necessary essence. So God must know something other than Himself taken as necessary, and yet something literally in Himself (can an act of selection be outside the acting agent?). It appears that the divine knowledge and will are far indeed from the reach of this type of Rationalism. Had the abstract-concrete paradox inherent in the recognition of the divine necessity been attended to, the need for a very different approach might have become apparent.
And after all, it is not even enough that God should know what world He selects. He must also know the actualization resulting from the selection. For actuality, we are told, is more
[199]
than mere possibility. Can what is added be a mere additional essence of actuality as such? Surely not. This is just the wrong way to take actuality as a predicate. For ‘possible actuality’ will then already include any essence that might be meant. No, God must know the resulting actual world, not just the possible one, including His selection of it. God, the ancient myth says, beheld the world he had made and *saw that it was good’. He did not merely look at it as possible and see that it wowld be good. Classical Theists have to claim a superior wisdom to Scripture at this point. I should have no objection, if I could see the superiority. But in this case I cannot, and I think I have read a reasonable number of explanations of the doctrine.
Yet the great rationalists had their deep insights. They were perfectly sound in their convicion (and the present age is deluded in going to the contrary extreme) that not all knowledge of existence can be empirical, and that, above all, the bare knowledge that God exists must be of a logically different type from genuinely empirical knowledge. They were correct in holding that there is a proper idea of God which implies a certain simplicity and absence of negation or possible contradiction. They only overlooked the little qualification: what the simple, purely positive idea refers to is not God in all His reality, but only—to use their term but not in their meaning—the divine essence, i.e., what God essentially is, an expression which implies, by contrast, what God is inessentially or accidentally. (*Accidentally’ does not at all—and here many have been misled—imply not really or genuinely. The next time a harshly painful accident occurs to the reader, let him ask himself if it makes sense to say, “This accident has not really happened to me, since it is inessential to my
[200]
nature.” Reality is one thing, the essential or necessary, another, and the latter is but an abstract aspect of the former.) The mere essence of God is simple and wholly positive in certain senses. It has no definite parts and contains nothing of the actual complexity of the world, nor does it even know that complexity. God knows the complexity, but in His accidental not His essential aspect. (Not that He might have failed to know what existed, but that what existed to be known and hence the knowing, might have been indefinitely other than it is.) The divine essence, qualified by necessary existence, as God in His full reality is not, is wholly positive in just the sense that it is strictly noncompetitive, its mere being somehow actualized not contradicting the actualization of any positive possibility you please. Nor is God required by His essence to combine incompossible things, for the essence does not say that He creates or knows all possible things, each actualized, but only that He infallibly knows all actual things (and they at least must be compossible) and that He could and would know any set of compossible things, should it be actualized. In this way the rationalists’ groping search for an idea of perfection which could not be contradictory may, it seems to me, come to its goal. For consistency seems built into the idea, so constructed, just as inconsistency seemed (and I believe was) built into it constructed in the classical way.
Is there not also a good case for Leibniz’s suggestion that the burden of proof is primarily upon those who deny the logical possibility of deity? But Leibniz fails to note that the logical paradoxes in the classical idea of God were clearly pointed to as far back as Carneades, and never genuinely disposed of. Arnauld mentioned one to Leibniz, who virtually admitted
[201]
that it was not to be evaded.27 However, our present situation is that we have an alternative form of theism which also, and with better right, can employ the ontological argument. So I incline to the view that the next move is up to the skeptics,
9. Hume
Hume makes a remarkable concession concerning the possible importance of the ontological argument: he grants that its validity would dispose of the argument against theism based on the evils in the world. And of course, no empirical facts can testify against a logical necessity. Indeed, the argument from evil itself rests on the supposed analytic truth that Greatness must result in a2 world without evil. This, in turn, means that Greatness in God implies an absolute absence of independence or initiative of action in the creatures. For, if they have any such independence, evil may be their doing, for all we could know, not God’s. ('They’ here means creatures generally, not just human beings!) And to say that God should, and as Great logically would, grant no freedom in this sense is to say that a being who can and must deny all genuine independence of action to others is better than one who could and would foster suitable degrees of independence in them. So far from finding this analytically true, some of us find it analytically false. Perhaps ‘omnipotence’, in the sense of a monopoly of power, an infinitely stingy denial of real power to others, is even a mere absurdity. In any case, it
27 Leibniz, Discourse on Metaphysics and Correspondence with Arnauld (LaSalle, Illinois. Open Court Publishing Co., 1924), pp. 96-114 (letters vi, viii).
[202]
has not been validly deduced from Greatness. ‘Greatness’ means having whatever properties it is better to have than not to have, as compared to other conceivable individuals. Perhaps monopolizing freedom is not such a property—is not at all a good thing. One of the beauties of Anselm’s formula is that it frees us (far more than he realized!) from automatic commitment to traditional views about God. It may very well not be ‘best’ to be ‘omnipotent’, in the sense which generates the problem of evil in its classic form.
Hume found this problem crucially important, not so much because of his ‘empiricism’ as because of his quite unempirical espousal of determinism. From the presumed, not observed, petfect regularity of nature he inferred the absolute control of all things by the hypothetical regulator. If God does anything, He must do everything, since the laws leave no indeterminacy. The problem in this form springs from a false a priori, incompatible with the theistic a priori.
The following is Hume’s conception of the ontological proof:
Nothing is demonstrable, unless the contrary implies a contradiction. . . . Whatever we conceive as existent we can also conceive as nonexistent. There is no being, therefore, whose nonexistence implies a contradiction.
It is pretended that the Deity is a necessarily existent being; . . . and that if we knew His whole essence or nature, we should perceive it to be as impossible for Him not to exist as for twice two not to be four. But it is evident that this can never happen, while our faculties remain the same as at present. It will still be possible for us, at any time, to conceive the nonexistence of what we formerly conceived to exist; nor can the mind lie under a necessity of supposing any object to remain always in being; .
But further, why may not the material universe be the necessarily existent Being, according to this pretended explication of necessity. . . . ‘'Any particle of matter, it is said, 'may be com-
[208]
ceived to be annihilated; and any form may be comceived to be altered’ . .. But it secems a great partiality not to perceive that the same argument extends equally to the Deity. . . . It must be some unknown, inconceivable qualities which can make His nonexistence appear impossible, or His attributes unalterable: And no reason can be assigned, why these qualities may not belong to matter. As they are altogether unknown and inconceivable, they can never be proved incompatible with it.28
Here is a blunt challenge. We shall meet it. As Anselm knew, the fool can ‘think’ that God is not, but he cannot genuinely conceive it, understanding what he is saying. Just so Hume. He is here not genuinely asking, in a more than verbal sense, what it is to ‘conceive something to exist’, or not to exist. Is it, perhaps, to imagine experiencing its presence or absence? But how does one experience an absence? As Popper says, by experiencing something positive incompatible with its presence. It has been shown above (Part One, Sec. 15) to be a contradiction that anything positive could contradict the existence and ubiquity of deity. Hence there is no way genuinely to conceive the nonexistence of God. Hume is only showing that we can say in words, “God (perhaps) doesn’t exist.” This disposes of the first two paragraphs of the quotation. They are mere question-begging dogmas.
The ‘quality’ which makes deity necessarily existent is Greatness. The notion of a ‘material universe’ which is Greatest has no clear meaning unless it be supposed equivalent to God, endowed with all the attributes which follow from Greatness, for instance, infinite flexibility in relating itself to possible worlds, while yet remaining genetically identical as an
28 Dialogues Concerning Natural Religion, Part IX. Philosophical Works (London, 1827), vol. ii, p. 496.
[204]
individual. If the material universe is conceived in such fashion it is simply being conceived as God, and ‘material’ loses all its sting.
That the properties of matter can be conceived to alter is not what makes matter contingent; for the very necessity of deity (according to the neoclassical resolution of the ‘nonegreater’ ambiguity) consists precisely in unique alterability, not in unique fixity. The trouble with ‘matter’, however, is that its identity through alterations has no positive meaning. Abstract from all particular forms, and ‘matter’ is but a word; abstract from all particular forms, and mind as Great is still the infinitely flexible correlate of all possible forms, yet able to recognize itself as genetically identical in knowing them, regardless of which ones are actual.
Put in another way: unless mind is the sole ultimate determinable, we have two such determinables, only verbally distinguishable: mind and matter. But one of these must be superfluous; which shall we keep? That we have experience, that we love and know, cannot be denied; the notion of necessary being we need, therefore, is that of Greatness as infinite spiritual capacity for knowing and feeling diverse things. Only because Greatness takes account of particular forms do they have any importance in final perspective; whereas, matter as such, as other than mind, could appreciate or register no value. The necessary being is the ultimate determinable without which determinates would determine nothing. And religious experience tells us how to conceive it. Physics as such cannot do so, and never—"so long as our faculties remain the same as at present””—can it do so.
It scarcely needs saying that Hume shows no sign of having read Anselm—after all, who does?
[205]
Hume makes one positive though unintended contribution to the Anselmian problem. His analysis of the self, often regarded as wholly destructive, is a somewhat crude version of the Buddhist-Whiteheadian view which, as we have seen (Part One, Sec. 11), is of assistance in escaping the abstractconcrete paradox in the employment of the argument. And here, Hume’s empiricism is methodologically in the right. For, in contrast to God, man is a merely empirical reality. His ‘soul’ is nothing but a contingent entity, and the attribution to it of transcendent properties, like absolute self-identity throughout its history (or even indestructibility and eternity! ) belongs in neither a sound rationalism nor a sound empiricism. Only Greatness enjoys perfect genetic self-identity, and the attribution of this perfection to man was one of the sources of the notion that God’s identity must be simply nongenetic, must be sheer immutability. Either there is no perfect case, or anything like it, of genetic identity, or else God is the sole unqualifiedly genidentical being. And if He does not necessarily exist, then perfect genidentity is not even logically possible. To the realization of this point in neoclassicism Hume unknowingly made a valuable contribution by his honest adherence to the empirical method where it was entirely appropriate.
We must be careful, however. Man, as having the idea of God, contains a transcendent element. But this element, as transcendent, has nothing peculiarly human about it. Purified of all contingent and anthropomorphic eccentricities, it coincides with something in God's knowledge of Himself. There is here nothing to support the soul-substance which Hume rejects.
Looking over Hume’s procedure as a whole, from the
[206]
Anselmian point of view, we see the following. (1) Anselm showed that Greatness is inconceivable except as necessarily existing; from which it was a corollary that to deny the conceivability of ‘necessarily existent’ is to affirm, ‘God is inconceivable’. Hence the universal contingency of existence, affirmed by Hume as beyond all exception, is the downright denial even of the thinkability of deity. The stark contradiction between such absolute empiricism and theism does not necessarily refute theism; perhaps it rather refutes absolute empiricism! Moreover, the unqualified validity of empiricism cannot itself be an empirical truth. So Hume is simply appealing to his own a priori, against the religious a priori. It is his say so against that of (theistically) religious mankind.
It can easily be shown that Hume makes such a dogmatic antitheistic decision in more than one additional respect. (2) He posits as an a priori truth that “what is distinguishable is separable,” i.e., there are no internal relationships of any kind among existing things. Yet, among the implications of Greatness is this: at least one sort of internal relationship between existing things must obtain, the relation of dependence of all other existents upon Greatness. Relation to God must be intrinsic to things. Therefore, since to deny Greatness is to declare it absolutely impossible, Hume’s doctrine of universal separability cannot be empirical; for no fact can show something to be absolutely impossible. So here again he is appealing, not to experience but (if anything) to a priori self-evidence; he is again attempting to show, in effect, that theism is logically impossible. Finally, when (3) Hume asserts the truth of absolute determinism, he is contradicting another implication of theism, that free creativity (not simply free in the sense of voluntary and without impediment, but
[207]
rather of determining the otherwise indeterminate, or of adding to the antecedent or presupposed definiteness of reality) is the universal principle of actuality. For, if supreme reality consists in supreme creativity (and the abstract-concrete paradox cannot, I hold, otherwise be resolved), then lesser realities must be lesser—but not zero—forms of such creativity. The step from great to Greatest, inherent in the logic of Anselm’s definition, cannot be from nothing to something; it must be from something to the supreme something. Hence all creatures must have some creative power. Unqualified determinism, therefore, is contradictory of Greatness. If so, then, in this third way too, Hume assumes the logical invalidity of theism (not its simple falsity, for there can be no such thing).
Is it then surprising that the outcome of the Dialogues, resting as the whole discussion does on this triply dogmatic mechanistic and pluralistic positivism, should be ‘skeptical’? The result is built into the method. Absolute empiricism, absolute pluralism, absolute determinism, contradict the existential necessity, the unifying function, and the actual freedom, bountifully overflowing into lesser forms of freedom, which are the very meaning of ‘God’. One must choose, and Hume'’s arguments for his choice merely reiterate the choice as already unwittingly made.
On such an argument in a circle so much modern antimetaphysical and skeptical philosophy is founded! The reckoning with Anselm lies in the future: it has not taken place in the past. Has Hume refuted Anselm? Or has Anselm (not as classical theist, but as theist) refuted Hume? This question has not even been discussed. How could it be, when Anselm was virtually unknown? So a little error in scholarship, collectively compounded through centuries, becomes a great error,
[208]
not of mere detail, but of methodological principle. The central philosophical question, that is, the rationally accessible content, if any, of the central religious question, is not an empirical matter, and empiricism, persisted in to the end, merely makes the muddling of this question the guaranteed outcome of all our philosophical efforts.
Will the great Kant cure the muddle, where the brilliantly lucid Hume could not? Or will he complicate and compound it by adding an alleged class of problems which are neither empirical nor a priori, and concern neither the intelligibly necessary nor the intelligibly contingent, but “something we know not what,” the noumenon? Will he return to the Anselmian problem on at least the level of profundity to which Anselm penetrated, or will he deal with it only in the loose pseudo-Anselmian form current in modern rationalism? The reader perhaps foresees the answers.
10. Kant
Apparently Kant knew nothing of the Proslogium. Even the Cartesian argument he probably thought of chiefly in the forms given to it by Leibniz and, above all, Baumgarten. These forms were, at best, no improvement upon the Anselmian original. Moreover, they employed definitions of God which were less rich in possibilities than Anselm’s own, being more hopelessly committed to ‘platonism’. And indeed Kant, after more than seven centuries, was no freer than Anselm had been to investigate the possibility that the Greek way of construing the idea of God as an absolute and immutable maximum of reality or perfection might be a mistranslation of the religious idea, and that a better translation might
[209]
remove some of the difficulties which philosophy had encountered in dealing with this idea or in attempting to find evidence of its truth. Kant is the last really great representative of classical theism, differing essentially from his predecessors not in his theoretical idea of God, but in his restricting of the grounds for faith to the argument from ethics.
One has only to read the relevant passages in Baumgarten's Metaphysics (see the Bibliography) to realize that from him, as from Leibniz or Wolff, Kant could only learn what I have argued is the wrong version of the Anselmian proof. God’s existence is held to be necessary simply because all His properties or qualifications are necessary. He must have or be everything possible, except such things as—since they connote imperfection—He could not possibly have or be. Thus His entire reality is wholly ‘determined’ by the requirements of His eternal essence. This, I submit, is precisely not the principle of the divine necessity. On the contrary, it is necessary that God be capable of contingent qualifications, and that His capacity for such qualifications be unsurpassable. (For all contingent things are contingent items of the divine knowledge.) The divine existence indeed cannot be contingent, but not because existence in general is a property and God can have no contingent properties, rather, because even the capacity for existing contingently, or for contingently failing to exist, is a defect and only the unique kind of existence, eternal, inevitable, self-existence, is compatible with divinity. The modal distinction between kinds of existence, which Prosl. III brought into the literature, is the only proper basis for the Anselmian proof.
Baumgarten simply follows Leibniz in his proof of the consistency of Classical Theism: all perfections or admirable
[210]
qualities are as such purely positive, and hence no contradiction can arise from attributing them all in maximal degree to God. As Kant saw, this is invalid. To take an example which he did not—but I am tempted to say should have—suggested: God can know a universe of which proposition p is true, or he can know one of which p is false; God cannot do both. Yet either is a possible positive qualification of God. (The negative character of ‘p false’ is not significant here. Suppose ‘p’ stands for ‘there is an oak tree just here’. Then there might have been something else, say an ash tree, just here. Call the assertion that this latter possibility is realized ‘q’. Then God might have known that p, or he might have known that q, but he could not have known that p and also have known that q. Yet, knowing that p and knowing that q are both possible positive qualifications of God. Kant’s own example of incompatible positives was that of velocities in diverse directions. We shall presently consider an objection he might have had to our example.)
Kant saw that Baumgarten’s case cannot be made out; he did not see that the opposed case can be made out, and that Neoclassical Theism may well be the residuary legatee of the inquiry if carried through without fear or favor.
The Argument of Prosl. IIl and subsequent passages was not what Kant refuted; he relied upon authors who had no knowledge of this argument. The Anselm of Prosl. III, also the Descartes of the Replies, as well as Spinoza, are refuted by Kant only in the sense that, as he showed, they had failed to establish a consistent meaning for their view of God. They did not fail, however, because their proof must stand or fall with the predicate-status of existence in general, Existential modality (necessity versus contingency) is
[211]
the predicate required. And that this is not a predicate not only Kant, but the entire philosophical world, has yet to show.
In spite of his negative attitude toward theoretical theism, Kant, in the essay Der Esnzig-Mogliche Bewessgrund des Daseins Gottes, makes a contribution to theism which should never again be lost and which is not invalidated by anything he himself later said. This is the proposition that there is no point in adducing the facts of worldly existence in arguing to a necessary Being. Since the necessary must furnish the ground of the very possibility of things, it is irrelevant to inquire which possibilities are actualized, or whether any are. For the point is that without God as ground of possibility (and this ground the necessary cannpot fail to be) not only would no world be actual, but none would even be possible. This being inconceivable, God as at least potential creator is required, whether or not there are creatures. All theists should feel indebted for this clarification. It was near the surface in the neglected work of Anselm. It was Thomas, as much as anyone, it seems to me, who, with his sensory empiricism, covered up the insight so deeply that only a great disturbance like Kantianism could bring us back to it, in spite of Kant’s later ‘critical’ views.
In his early essay Nova Dilucidatio (1755 ), Kant proposed a quasi-ontological proof, which he later—in the essay above referred to—revised, and still later, in the Critigue, quietly abandoned. Throughout this development, Kant adhered essentially to the same idea of God, though his confidence in our ability to justify it theoretically suffered a complete reversal. God, as absolutely perfect, must be free from all negation
[212]
of being, He must have all predicates so far as positive or expressive of reality. Thus in the Critigue we read:
All negative concepts are . . . derivative, and it is the realities which contain the data and, so to speak, the material, of the transcendental content, by which a complete determination of all chings becomes possible.
If, therefore, our reason postulates a transcendental substratum for all determinations . . . such a substratum is nothing but the idea of the sum total of reality (omnisudo realstasis).
. . . the concept of an ems realisssmum is the concept of an individual being, because of all possible opposite predicates, one, namely, that which absolutely belongs to being, is found in its determination. It is therefore a transcendental sdeal which forms the foundations of the complete determination which is necessary for all that exists, and which constitutes at the same time the highest and complete condition of its possibility, to which all thought of objects, with regard to their contents, must be traced back. It is at the same time the only true ideal of which human reason is capable, because it is in this case alone that a concept of a thing, which in itself is general, is completely determined by itself, and recognized as the representation of an individual.29
What a wonderful account—of certain assumptions of classical theology! All that anything can positively be (all that ‘absolutely belongs to being’) God must be actually and without restriction. This apparently (a) guarantees that He deserves to be worshiped as without flaw; (b) makes Him conform by definition or a priori to the law of excluded middle as the criterion of an individual or wholly determinate reality; (c) makes it possible to argue (in the next paragraph) that “the complete determination of everything [else] depends
29 Critique of Pure Reason, trans. Max Milller (New York and London: The Macmillan Company, 1920), pp. 465-66. This and subsequent citations from this work are by permission of the publisher.
[213]
on the limitation of this total of reality, of which some part is ascribed to the thing, while the rest is excluded from it”; hence (d) shows that the content of the idea of God is the presupposition of possibility as such, so that if it were not real, nothing would be so much as possible. And so, Kant at first thought, he can infer that God exists as the necessary being. For the nonbeing of the ground of all possibility cannot itself be a possibility.
Compared to the usual Prosl. II version of Anselm or Descartes, this proof is not unimpressive. Yet, compared to Prosl. 111, it has no decisive superiority, though it is valuable for the way in which it focuses on some aspects of the problem; and it does have certain defects, above all because it argues, not from the Unsurpassability of deity, which in some sense inheres in the very idea of worship (See Part One, Secs. 4-5), but from a nonreligious and philosophically dubious notion of absolutely complete reality or a priori determinateness. Kant never explicitly repudiates just this proof, but clearly he found it wanting in the end.
What was wrong? According to Kant, a number of things. (1) The needs of our thinking are not legislative for reality: we may need the idea of God, but only as regulative, not as constitutive of nature or supernature. (2) We cannot know that the sum of all predicates as positive is really possible. Positive predicates, as Kant argued in his Versuch den Begriff der negativen Grissen in die Welt-Weisheit einzufiibren, may be “really repugnant” to one another. (3) We cannot know that an unlimited being is the only sort that could exist by absolute necessity. True, we could not deduce such necessity from the concept of a limited being; but then Kant denies that we can really do this from that of the un-
[214]
limited being. We cannot validly deduce anything from so problematic a concept as that of unlimited being, and since a limited being cannot even be defined a priori, as an individual, its necessity can certainly not be deduced. However, neither—he argues—can its contingency. Kaant is really saying lictle more, in all this, than that such contrasts as necessarycontingent, limited-unlimited, transcend our experience and therefore our knowledge. Essentially his attack on the Argument is on positivistic grounds (in the agnostic form: for all we know, ‘God’ has no genuinely possible content).
In seeing that negative concepts presuppose positive ones Kant was close to discovering the competitiveness of contingent existence. For just as one cannot know that x is not white without knowing white, and indeed, without knowing that x has some positive character exclusive of white, so one cannot know that such and such is nonexistent except by knowing that there is something else whose existence excludes the such and such. But then, if an idea is such that its being actualized is omnitolerant of all other forms of positive actualization, its object could not significantly be said to exist unless necessarily.
There is an ambiguity in the assumption, characteristic of classical metaphysics, that to be determinate a thing must relate itself positively or negatively to every possible predicate. Sophocles was non-Shakespearean, but does this mean that the entire quality of Shakespeare must have been available for Sophocles to negate? If qualities can be created, as neoclassical views hold, then so can negative relations to those qualities. Sophocles did not need, for his own definiteness, any relation to the emergent qualities which subsequently appeared in Shakespeare. It is Shakespeare who had to define
[215]
himself relative to Sophocles, not vice versa. This is the asymmetry of time, according to creationist metaphysics. Kant seems not to have had any such idea. But he was right, on this ground too, in his suspicion of the reasoning which led to his ‘transcendental ideal’.
Moreover, the insight that positive qualities can conflict implies more than that we cannot know the possibility of God as an absolute maximum of reality. It implies the impossibility of such a maximum. God could not possibly know as actual all possible worlds, for they are not mutually compatible. Yet the knowledge of an actual world as such is surely something positive. Hence some values must be possible for God but not actual. In addition, if all God’s properties were fixed as actual by a priori necessity, He would have no freedom whatever. And freedom too is positive.
Kant would have interposed here: God does not really ‘know’ or ‘act freely’ in any sense which we could understand. Just as Anselm had said that in God there is not really ‘compassion’, but only its appropriate effects in us, so the redoubtable Konigsberger argues (in the Prolegomena, for instance) that God is not really an intelligence—this he says would be anthropomorphism. However, the world is to be viewed as if made by an intelligent cause. What God may be ‘in Himself’ we can have no idea whatever. Thus, like his medieval predecessors, Kant has his doctrine of theological ‘analogy’.
Certainly knowledge in God is, in its concrete quality, very different from anything we can clearly conceive. But it does not follow and, as we argued in Sections 16 and 17 of Part One, it is not tenable that there are no abstract principles or rules of meaning which apply to all thought, and hence to thought about God. Among such rules are
[216]
these: that actuality differs from possibility by its arbitrary determinateness and that possibilities for concrete values are competitive. Also, that what a thing is in itself, independently of relations, is an abstraction from the full concreteness of the reality containing the thing and its relations. (Thus God merely ‘in Himself’ would be less than God relative to us.) Like other classical theists, Kant appears to reject the evidence we have concerning the most universal aspects of ‘value’ and ‘existence’, and to suppose that the Object of Faith must violate or transcend every rule of concept-formation which experience illustrates. This overlooks the truth that the most general or abstract principles cannot be unknowable, for they are somehow embodied in any concrete thing. One has only to abstract from the specific restrictions; the universal is what remains.
Of course, if we can in principle know nothing about possibility, actuality, incompatibility, abstractness, and concreteness, then we can know nothing about what God is or might be. The indecisive result here is built into the method. The trouble, however, is not with the Argument, as such, rather, with certain features of Kant’s subjectivistic philosophy, together with that traditional (and idolatrous?) identification of God with the absolute or the infinite which Anselm, Descartes, and Leibniz sought to support by the Argument.
Surely Kant did well to drop the search for a proof which would be valid when so used. But had he been a freer mind, he might have drastically reconsidered the question: have we properly conceived God, or is there a better way? Instead he dismissed rational (theoretical) evidence, and fell back upon moral and religious faith alone.
Suppose, however, he had reconsidered his problem: to
[217] define an individual a priori, as essential to the possibility of all other things. One way, the only one he saw, is to posit an absolute maximum of actuality. This failed. What is left? Obviously, to posit an absolute maximum not of actuality but of potentiality. God, simply as necessary, can have neither limited nor unlimited actuality (for no actuality, no particular how of concretization, can be necessary) but only unlimited potentiality. Divine potentiality—thus conceived—is as free from arbitrary limitation as you please, indeed it is absolutely free. Moreover, it is more plausible to relate all possible particular determinations to an infinite potentiality than to an infinite actuality (which indeed, as Kant suspected, is an absurdity). The particular is a determinate under a determinable; the ultimate determinable is the divine creativity. A possible creature is a possible state of the creator-as-having-creatures. But the mere existence of the creator in some state or other is wholly without arbitrary, competitive restriction. It is the disjunction of all restrictions, itself unrestricted. If the necessary is defined in this way, there is no Findlay paradox; for the ultimate determinable, as such, has no arbitrary determinations, and its assertion is logically weak to an infinite degree. Like all necessary statements it is factually ‘empty’, cuts off no genuine possibility.
But how, Kant might have asked at this point, could the ultimate determinable, the least determinate thing, define an individual? In the first place, even an ordinary individual need not and should not be defined as wholly determinate, unless all freedom and chance are to be denied. An individual’s life is of course determined retrospectively, but then there is a new individual (as determinate) each moment.
[218]
The Buddhists alone, before Whitechead, seem to have adequately—or almost adequately—realized this.
Taking this schema of individuality seriously (a course precluded to be sure by Kant's phenomenalistic theory of time), we can still define one individual a priori. To do this we must distinguish between its individual essence and its mere accidents. Only the essence needs to be defined; for its existence merely means that the essence is actualized somehow, in some suitable accidents—any you please so long as they embody the essence. And here the essence is unlimited capacity, e.g., cognitive capacity. Only one individual can have such capacity. In this way we escape the difficulty by which Kant was rightly appalled, the difficulty that positive predicates may conflict. God’s uniqueness is not that He exhaustively actualizes possible value (this is impossible), but that there is no consistent set of possible values He could not enjoy.
Unfortunately, no such view was conceivable for Kant, since he thought as a platonist—with the difference that he is full of skepticism, without a glimmering of hope that the view could either be justified or improved upon theoretically. On the contrary, with his usual eagerness to claim finality, he solemnly pronounces the platonic concept to be ‘the only true ideal’ that human reason can attain to. But there is an ideal he dreamt not of, that of the self-surpassable, otherwise unsurpassable Creativity.
Though Kant's ‘critical’ objections to the Argument take no account of the reasoning of Prosl. III, and though some of them are irrelevant to that reasoning, yet others might be used against it. Thus, e.g., Kant argues: if the reality of God can be inferred from our idea, God must be identical with
[219]
the idea. By this reasoning, as no one seems to have noticed, Kant proves rather too much. For, by its principle, if God can be inferred at all, He must be simply identical with that from which He is inferred; but then not only is no nonquestion-begging proof for God possible, more than that, either God must be identical with the creatures, or else they cannot logically require His existence as cause of their own! Accordingly, it is theism itself which Kant has refuted here, if anything at all! He is really saying that God can in no fashion be ‘immanent’ in the creatures. For if He can be, then He can be in our idea, as a creature! And if in it, then knowable from it.
In addition, if our neoclassical analysis is right, the mere necessary existence of God (His being actualized somehow, no matter how) is not the concrete actuality which realizes this existence, but is an extreme abstraction, such as might well be within our conceptual grasp. And in any case, a universally immanent being cannot but be somehow constitutive of everything, including every thought. So of course God is somehow in our thought of Him, and also in the animal which cannot think Him. According to Wolfson, a central point of the ontological argument is precisely that God is immediately given.30 Moreover, it is deducible from the idea of God that this must be so. Hence, if God is not given, He does not exist, and if He does not exist, then He is either logically impossible or logically contingent—which are here the same since, as Anselm discovered, ‘contingent perfection’ is contradictory.
30 Wolfson, H. A., The Philosophy of Spinoza (Cambridge: Harvard University Press, 1948), vol i, pp. 170f.
[220]
In the same paragraph Kant asks, is the judgment of the divine existence analytic or synthetic? And if the former, does it not beg the question? This is in part a merely technical question, What “follows from the rules of our language” depends upon how we set up those rules. What Anselm discovered was that rules allowing one to treat a proposition as contingent merely because it asserts existence are rules which decide a priori against even the logical possibility of God's existence! Is this a legitimate decision to leave to linguistic rules? Only if such rules have decisive advantages other than that of invalidating the ontological argument—and what other advantages do the rules have? If there are none, it is their adoption, and not the argument, which “begs the question.” Our analysis has suggested that the rules in question have, in truth, grave disadvantages, since they make not only deity but also contingency unintelligible and require us to admit merely negative facts, for example, and other monstrosities.
Like Hume, Kant has put his skeptical results into his initial assumptions. He knows all along that God must either be given as a sensorily perceivable object or be inaccessible. And he knows that God cannot be given as such an object, ergo. . . . No elaborate reasoning is required to draw the conclusion. But the definition of God tells us that this disjunction is invalid. God can neither be given in the obvious fashion of a sense-object, since He is in all such objects and therefore is not peculiarly given in any one of them, nor can He be simply inaccessible, for then He would not be the universal ubiquitous ground of things. And this ground, Anselm would say, “it is better to be than not to be.” Hence God can only be conceived as having such a status. In saying that we
[221]
could not in any fashion know or experience God, Kant is saying that God could not exist.
The great but, even so, calamitously overestimated German philosopher holds that if there be a concept which implies the existence of its object, still that concept is “accepted voluntarily only, and always under the condition that I accept the object of it as given.”31 If Kant were here confessedly representing positivism, not atheism or agnosticism, he would be in the right in the first three words of the quotation: the concept need not be accepted as ‘well formed’ or ‘cognitive’. But the atheist does not reject, he uses, the concept, and if it does imply the necessary existence of its object, then the atheist contradicts himself. He is saying, "God or Greatness, whose appropriate definition (and Anselm'’s argument for its appropriateness has not been refuted) implies that He cannot be conceived not to exist, I yet do conceive as (perhaps) not existing.” If this is allowed, what can be forbidden?
Kant thus fails to define the ground of his negations sharply as between atheism and positivism. Anselm proved that only the latter has a case. Here is another example. Kant says: “In introducing into the concept of a thing, which you wish to think in its possibility only, the concept of its existence, . . . you have been guilty of a contradiction . . . you have achieved nothing, but have only committed a tautology.”?? However, it is simply false that Anselm, Descartes, or Spinoza wished to think God ‘in His possibility only’. They wished to think God in whatever way is compatible with the suitable definition of this term. The definition says it cannot be as
31 Kant, op. cis., p. 479.
32 Ibid,, p. 481.
[222]
‘possible only’; so how can it be? Either as necessarily existent, or not at all, ie., existence must be affirmed, or the concept dropped as not genuine and consistent. If this last was Kant's point, why did he not say so more clearly? Am I deluded or was Kant? If it is I, then forty years have gone by, and I have not been able to see that he was right, though I have made many attempts. Will someone, out of the many who with Kant are so much wiser than I, not help me?
But might not what was neither contingent nor logically impossible yet be really impossible? This seems to be the purport of some of Kant's remarks. However, real possibility or impossibility can only mean, on such and such existing conditions. If, then, the conditions required for a thing are themselves logically possible—and if not, neither is the thing itself—yet are really impossible, we have a vicious regress of conditions of the conditions. To avoid this, we must give up the alleged distinction between real and logical possibility, as applied to our problem. The distinction is quite valid with respect to localized items of reality; for what is logically possible somewhere may lack the required conditions just here, or here. But a nonlocalized reality such as deity cannot be treated in this way. It can have no special conditions whatsoever. If it were not really possible everywhere and always, it would be impossible absolutely and in its very meaning. All possible conditions must be compatible with its existence, and then either real and logical possibility coalesce, or else the idea is not possible at all, whether logically or really.
Let us now consider an extended passage.
Being is evidently not a real predicate, or a concept of something that can be added to the concept of a thing. It is merely the admission of a thing, and of certain determinations in it
[223]
Logically, it is merely the copula of a judgment The proposition, God is almighty, contains two concepts, each having its object, namely, God and almightiness. The small word is, is not an additional predicate, but only serves to put the predicate in relation to the subject If, then, I take the subject (God) with all its predicates (including that of almightiness), and say, God is, or there is a God, I do not put a new predicate to the concept of God, but I only put the subject by itself, with all its predicates, in relation to my concept, as its object Both must contain exactly the same kind of thing, and nothing can have been added to the concept, which expresses possibility only, by my thinking its object as simply given, and saying, it is. And thus the real does not contain more than the possible. A hundred real dollars do not contain a penny more than a hundred possible dollars. For as the latter signify the concept, the former the object and its position by itself, it is clear that, in case the former contained more than the latter, my concept would not express that whole object, and would not therefore be its adequate concept. In my financial position no doubt there exists more by one hundred real dollars, than by their concept only (that is their possibility), because in reality the object is not only contained analytically in my concept, but is added to my concept (which is a determination of my state), synthetically; but the conceived hundred dollars are not in the least increased through the existence which is outside my concept.
By whatever and by however many predicates I may think a thing (even in completely determining it), nothing is really added to it, if I add that the thing exists. Otherwise, it would not be the same that exists, but something more than was contained in the concept, and I could not say that the exact object of my concept existed. Nay, even if 1 were to think in a thing all reality, except one, that one missing reality would not be supplied by my saying that so defective a thing exisws, but it would exist with the same defect with which I thought it; or what exists would be different from what I thought If, then, I try to conceive a being, as the highest reality (without any defect), the question still remains, whether it exists or not. For though in
[224]
my concept there may be wanting nothing of the possible real content of a thing in general, something is wanting in its relations to my whole state of knowledge, namely, that the knowledge of that object should be possible a posteriori also. And here we perceive the cause of our difficulty. If we were concerned with an object of our senses, I could not mistake the existence of a thing for the mere concept of it; for by the concept the object is thought only as in harmony with the general conditions of a possible empirical knowledge, while by its existence it is thought as contained in the whole content of experience. Through this connection with the content of the whole experience, the concept of an object is not in the least increased; our thought has only received through it one more possible perception. If, however, we are thinking existence through the pure category alone, we need not wonder that we cannot find any characteristic to distinguish it from mere possibility.
Whatever, therefore, our concept of an object may contain, we must always step outside it, in order to attribute to it existence. With objects of the senses, this takes place through their connection with any one of my perceptions, according to empirical laws; with objects of pure thought, however, there is no means of knowing their existence, because it would have to be known entitely a priorsi, while our consciousness of every kind of existence, whether immediately by perception, or by conclusions which connect something with perception, belongs entirely to the unity of experience, and any existence outside that field, though it cannot be declared to be absolutely impossible, is a presupposition that cannot be justified by anything.
The concept of a supreme Being is, in many respects, a very useful idea, but, being an idea only, it is quite incapable of increasing, by itself alone, our knowledge with regard to what exists. . . . The analytical characteristic of possibility, which consists in the absence of contradiction in mere positions (realities), cannot be denied to it; but the connection of all real properties in one and the same thing is & synthesis the possibility of which we cannot judge a priori because these realities are not given to us as such, and because, even if this were so, no judgment whatever takes place, it being necessary to look for the
[225]
characteristic of the possibility of synthetical knowledge in experience only, to which the object of an idea can never belong. Thus we see that the celebrated Leibniz is far from having achieved what he thought he had, namely, to understand a priori the possibility of so sublime an ideal Being.
Time and labor therefore are lost on the famous ontological (Cartesian) proof of the existence of a Supreme Being from mere concepts.33
In these famous passages Kant seems scarcely aware that from the standpoint of the second Anselmian or Cartesian Proof the question is not whether ordinary or contingent existence could ever be derivable from the mere concept of a kind of thing, but only whether a uniquely excellent kind of existence, necessary existence, can be derived from a unique concept, that of divine perfection or Greatness.
Insofar as the Critical Kant was aware of the relevant question, his view was apparently similar to that of Hume before him: necessary existence, taken as a quality or excellence, is without intelligible content, so that the idea of it could amount to no more than this: the necessarily existent (supposing the phrase means anything) necessarily exists. No other quality, in short, is logically connected with this alleged quality. We still should not know whar necessarily exists.34 Yet Anselm had tried to exhibit some two-way logical bridges between necessary, essential existence, or self-existence, on the one hand, and eternal existence, existence without parts, dependency, or defect, and thus existence as that of the Unsurpassable or Greatest, on the other. And we have seen
33 Ibid., pp. 483-86.
34 Henrich (see Bibliography I) reaches approximately this conclusion in his study of Kant's criticism of the Proof.
[226]
in Part One that still other and stronger bridges can be constructed or found. Neither Kant nor Hume is adequately aware of the classical literature of this subject, much less of the unexplored possibilities which neoclassicism discloses for accomplishing the end more convincingly.
By remarking that existence is not a predicate, did Kant state clearly the abstract-concrete paradox involved in the Anselmian use of the Proof? It seems not. Existence in the necessary case, for all Kant shows, might be a predicate; but it could not, as necessary, be concrete, or identical with God as concretely actual. There must then in God be a real distinction between His necessary existence and His total reality. Where does Kant even touch on such an idea? Yet the logic of the problem leads us straight to it, unless we have made up our minds not to go in that direction.
Furthermore, Kant overstates his case in denying that ‘existence’ is, in any way, a predicate, or that the existing is in any sense more than the merely possible. There must be some sense in which it is more. And when Kant says, if real dollars contained anything [a penny] more than possible dollars, “my concept would not express the whole object, and would not therefore be its adequate concept,” he gives himself away. For by what right does he assume the possibility of such a thing as an ‘adequate concept’ in this sense? Of course, there are only one hundred pennies in a real, as in a possible, dollar, but neither a dollar nor a penny can be fully expressed in a concept. Only sheer intuition could do this. The conceptual description of a kind of thing may at most account for so much of its quality or value as is expressible in merely abstract terms. But the full quality is not thus expressible. The actual in its unity has quality or value, and this is no ‘predi-
[227]
cate’ or bundle of predicates to which reference can, in a particular case, be made, save by pointing to the concrete and speaking of ‘its’ value. To treat this value as a universal, or a ‘possibility’, is merely verbal, as Bergson (also Dewey) has pointed out; for there is no way to separate it from its actuality. We are only setting up a verbal duplicate of the actuality, and assigning it to another date, in speaking of it as an ‘antecedent’ possibility, if we really mean the unitary quality of the actuality. Thus Anselm and Descartes were not wrong in saying that the actual is greater than the unactualized yet possible. The former has a uniqueness, a concrete definiteness, which mere possibility lacks. And in this unique definiteness is all the richness and beauty of the real world, beside which ‘possible worlds’ are pale shadows. If it were not so, possibility might just as well be left unactualized, and God did nothing when He created the real world.
Yet the foregoing sense in which existence is a predicate is irrelevant to the ontological argument. For, manifestly, that in actuality which is 'more’ than what is conceptually or abstractly expressible for that very reason cannot follow from any concept or definition; hence it cannot be used to establish the necessity of God. And yet Prosl. II and the parallel passages in Descartes give the impression of trying to use it in this way.
Kant seems to feel that it is absurd to suppose that we could know an individual without empirically perceiving it. Yet note that “Such that nothing greater is conceivable” refers to no empirical or contingent fact; Greatness is defined with reference to what could be, not with reference to what is. It is a concept wholly a priori. By contrast, ‘dollar’ cannot be defined without some empirical reference. Moreover, whereas
[228]
a particular dollar exists at or for a certain time—and all dating is empirical—'God exists on August 26th, 1963’ says nothing more than that He exists, and that there is such a date as the one mentioned. For God's existence (not His actuality) is regardless of dates.
True, since God is necessarily somehow actualized, our experience (and all things else) must be in relation to this actuality, and so it may well follow that in some sense, not necessarily conscious, we must intuit, experience, or feel it And how does Kant know that we do not do this? A lack of distinct awareness of something is not a distinct awareness that it is not given. It is merely an absence of conscious perception of the thing as in experience, not a conscious perception of its absolute absence from experience. (Such an absolute absence could only be perceived by a mind with unrestricted cognitive power.) Moreover, it follows from the idea of God that either the idea is illogical, or God exists necessarily and ubiquitously, and from the latter it follows that all experience must have Him as datum, however inaccessible the datum may be to easy and clear conscious detection.
To the query, how do we distinguish God from a mere possibility, we reply, simply by the contradictoriness of ‘mere possibility of Greatness’. Nothing can make Greatness possible, or self-consistent, but its own necessary reality as inevitably somehow actualized. We do not ‘add’ existence to the ‘bare possibility of God’, we deny that this latter phrase has a consistent meaning. Nor do we ‘add to our concept’ of God in affirming His existence; for the only proper concept of God is as existent. He can have no other status.
From a recent article (see Bibliography, Engel) I quote the following:
[229]
Kant . . . goes on to conclude . . . that when I, . . . think a Being as the highest reality, without any defect, the question still remains whether it exists or not—that is, whether (as he goes on to explain) it exists as some part of this (whole) world of my experience. Now the obvious answer . . . is that of course God does not exist in that way. . . . But this simply means that God lacks contingent existence . . . this is hardly a defect, for if he had contingent existence he would not be God, or what we ordinarily understand by that term.
The author adds in a footnote: “if the ontological argument fails to prove what it sets out to prove, the reasons for its failure must be other than those given by Kant” This article is at least a bit of evidence that the almost hypnotic spell of Kant's ‘refutation’ has now been broken. For this we need, no doubt, to thank Norman Malcolm more than anyone else.
We have already dealt, early in this section, with the Kantian contention that we do not know that ‘all realities’ could be combined in one reality, the Supreme Being. In any case, this is the positivistic, not the atheistic, objection to Anselm, and it leaves his refutation of atheism intact, so that he still would have proved something and something important, ie., that God could not be an unrealized possibility. Moreover, the Proof need not, and in neoclassical use does not, take God to be the actual union of all possible realities, the ens realissimum, but only the actual union of all actual realities, and the potential union of all possible ones so far as mutually compossible.
If Kant were to insist that only experience could tell us— but yet it does not and cannot tell us—that perfection, thus defined, is possible—or impossible—we should ask him, who then could know the truth here? Or is it unknowable absolutely? Obviously God could not know His own impossibility! We too, it seemns, cannot know it. What meaning, then, has the
[230]
term ‘impossible’ in this case? And if it has none, what is the dispute about? Certainly God could know—if He could do anything—His own possibility and (the same thing) His own (necessary) existence. Also, it cannot be simply impossible even for us to know it, since by definition the divine existence would be the necessary ground of ours, in its every aspect. Hence, we know must somehow imply the existence of God, and so exclude His impossibility. How then could this be wholly unknowable?
The strongest objection of Kant's is much like one of Thomas's. How can we deduce existence—or anything else— from a divine nature of which we do not initially have a clear conception? Anselm’s answer was that his definition does not claim to tell us anything about God beyond the mere impossibility of something superior to Him. Another answer might be that ‘nature’ is here profoundly ambiguous. If it means, the most abstract, necessary, or merely eternal aspect of God’s reality, then to know this is nothing hopelessly difficult, since we have only to abstract sufficiently from more concrete conceptions, in all of which the most abstract nature must be immanent. Concreteness, not abstractness, is the most baffling thing.
(In view of mathematics and formal logic—compared, say, to psychology—this should be commonplace.) If, in contrast, ‘nature’ means the concrete quality of God as God of this world, and of all created worlds in the near and remote past of the cosmic process, then this nature indeed we cannot know; but it is contingent, and not in question in the argument. There is no more tremendous difference than that between these two aspects of deity.
(If it seems to some remarkable, to the point of absurdity, that one being can unite such contrasting features, the appro-
[231]
priate comment is simple: it is not one whit more remarkable than all the other implicadons of Greatness. Omniscience must somehow unite all things whatsoever. We are back at the positivistic objection, which must always have some plausibility: can God's existence be conceived at all? There is no obvious reason why it should be wholly easy, nor yet simply impossible.)
Kant's entire philosophy rests heavily upon the assumption that theism must mean classical theism. It follows that space and time cannot in any way describe the highest form of reality, for deity is classically viewed as strictly immaterial and immutable. Kant is assuming that Greatness excludes every conceivable kind of becoming as inferior; but this assumption is never really argued, it is never shown that it is better to be so complete that change is irrelevant than to be capable of enrichment of content. Nay, Kant himself holds that we cannot know completeness in this sense to be possible. But if not possible, it is also not better; for there is no virtue in sheer impossibility. Hence Kant does not know that divine Greatness implies immutability, he merely assumes it. Had he gone back to Anselm'’s proposal, he might more easily have seen that he was begging a most important question.
Again, take Kant’s view that noumena could be known only by a wholly active mode of intuition or perception. This is classical theism again, for it assumes that to be wholly without receptivity or passivity is a merit. But on what ground? Are men less receptive than atoms, or incomparably more? The latter seems obvious. Then are men inferior to atoms? Kant was here repeating Anselm in his weakest aspect, his mistaken deduction of the negative theology from Greatness. Greatness is indeed defined negatively, but what it negates is not prop-
[232]
erties in God, but only the possibility of better properties in another individual. This allows God to have any positive property which is wholly good. And receptivity, sensitivity to the realities making up the universe, is indeed wholly good. The more of it a being has, the higher it stands. The absolutely insensitive is the absolutely dead, not the supremely alive. The Platonists (perhaps not Plato) are blind to this truth. Suppose Kant had seen this. He might then also have seen that true knowledge requires conforming the subject to its objects, not sheer making of these objects. (The latter would mean that the known was merely one’s own thinking, one’s own thoughts or feelings.) Kant’s whole epistemology could have been very different—and less subjective—had he realized the arbitrary character of his presupposed idea of God. And the study of Anselm, if ‘critical’ enough, could have taught him this.
Kant repeats, for the phenomenal world, Hume’s determinism, and he does not effectively or ontologically overcome even Hume's radical pluralism, in that he makes, not things or events, but only our experience of things or events, an interconnected whole. Thus he, like Hume, is playing against theism with loaded dice. One of the penalities of accepting the Newtonian conception of absolute causal regularity is that the way was thus barred to an objective or ontological view of modality. (Most logicians are still resolute Newtonians even today in their insistence upon the strict independence of denotation and truth from time.) Only if there is a real contrast between the determinate past and the determinable future can we have a basis for the concept of real possibility, of which real necessity is the most general or abstract aspect. Here is the crux of the modal problem. So of course Kant disclaims any
[233]
ontological knowledge of modality. But it follows immediately that he must disclaim any knowledge of God as necessarily existing, hence any knowledge of God, who only so could exist. Kant's dictum that all our conceptions must find their clue in the temporal structure of experience is correct; however, (a) determinism obscures that structure, and (b) Kant’s phenomenalistic restriction upon the scope of time makes the dictum entail a general agnosticism. According to the idea of divine unsurpassability as entailing divine self-surpassing there must even be an eminent or divine kind of time! So here too Kant has decided against theism before even taking up the subject.
Other features of Kantianism come under the same heading. Thus Kant falls behind Hume in his account of the self by admitting a transcendent timeless soul or noumenal absolute petsonal identity other (it seems) than that of deity, and at least morally certifiable. And he makes still other concessions to the classical theory of ‘substance’. Kant also is farther than Hume was from seeing that classical theism is not the only form theism can take. (Hume's Cleanthes makes an energetic, though not successful, effort to formulate what would have been neoclassical theism had he succeeded.)
When our valiant Konigsberger took it upon himself to declare as “labor lost” all future efforts to discover a valid point in the Argument, was he not “barring the path of inquiry”— alas, only too effectively?
Yet Kant did make a contribution to the Anselmian problem. This was in his contention that the supposedly empirical arguments for God's existence conceal an a priori. If it were in principle impossible to ‘argue from idea to existence’ in respect to God, there could be no theistic argument at all. Kant thus supports Anselm'’s intuition that it is wrong to turn over the
[234]
stones of fact to see if perchance God's existence—or His nonexistence—can be found lurking under them. If anything can indicate God, everything must do so, including particularly the bare idea of God itself. What Kant refuted was not the ontological a priori proof of God—he never clearly stated that (in any form equivalent to Prosl. III). What he refuted was the claim to avoid the a priori in the religious sphere. And indeed, to say that we cannot infer God from the logical possibility of His idea, though we can infer Him from cats, mountains, or our own existence, seems downright frivolous. Either the idea of God is a creature, or it is God's self-knowledge (simply that, or as participated in by a creature); there is within theism no third thing it could be. Either way, its reality entails that of God, if anything whatever can do so. For delivering us from the notion that there is a special order of entity, such as the existing world, or our sense perceptions of this world, with which we must start in order to reach God, we can thank chiefly two men, Anselm and Kant. We can start anywhere, and with anything whatever; the question only is, can we understand it well enough to see the reality of God which, unless theism is absurd rather than false, must be there?
11. Hegel
Hegel's defense of the Proof did it little good, first, because he never properly stated it and second, because his system was too unclear to appeal permanently. Anselm had a lucid mind; he generally used words with a nice exactitude. He meant by God the absolute actualization of all that is desirable and good. This complete actualization did not include that of the world, which was, strictly speaking, superfluous. God might
[235]
not even have been ‘supreme’, since He might have been solus. Deity is absolute self-sufficiency. This at least is clear. Spinoza also was clear enough: God is indeed absolute self-sufficiency, absolute actualization, but because His essence includes the world and because the world is all possibility exhaustively actualized. There is a third view: God is not in every sense selfsufficient, for although He exists independently, He depends for His particular actuality, or how He exists, upon what other things exist. Necessary or absolute in His bare essence and existence as divine, or simply as God, He is yet, in His concrete actuality, contingent, relative, and forever incomplete, because forever in process of further enrichment, value possibilities being inexhaustible. This, roughly stated, is neoclassical theism.
What is Hegel’s position? Perhaps he is a neoclassical theist? On the whole, I think he is a man who is and wants to be in a perpetual systematic muddle between classical theism, classical pantheism, and something like neoclassical theism, with a dose of humanistic atheism, or the self-deification of man, thrown in for good measure.
In his astute but cloudy way, Hegel accepts the Proof as integral to his system. He rightly points out that we should expect the infinite to obey partly different laws from the finite. For Hegel, the infinite, or God, is the absolute unity of subject and object, or thought and reality. Hence the notion of God as mere idea is absurd. As usual in Hegel, there is sense in this but sense entangled in ambiguities. Divinity, for one thing, is not simply identical with infinity; as Hegel himself knew, the merely infinite without the finite is an empty abstraction. But just how the concrete finiteness and the abstract infinity are together in the divine reality—has anyone ever been able to learn this from Hegel? Most of us have had to look else-
[236]
where for any clear light on the topic. That ‘every step toward concreteness is contingent’ is either denied by Hegel or at best admitted in most grudging and unclear fashion. I think it needs to be accepted outright, and without cavil. That the ultimate abstraction is somehow concretized may be necessary, but all else must be contingent. Only so can we avoid “deducing” the logically stronger from the weaker.
(It is only fair to say that the work from which the following is quoted was assembled from notes, partly by students.)
As is well known, the first genuinely metaphysical proof of the cxistence of God took the turn, that God as the idea of the being which unites in itself all reality must also possess the reality of existence. . . . ‘For if it is merely an object of thought it is not the highest thing; 'it can therefore be assumed that it exists: this is greater’ than something merely thought. . . . This is quite right; however, the transition is not exhibited, the subjective understanding is not shown to transcend [aufheben] itself.
. . . for the true proof it is requisite that the procedure should not be according to the [abstract] understanding; but that thoughe should, of its own nature, be shown to negate itself, and . . . to determine itself to existence. And conversely, it must be shown of existence that it is its own dialectic to transcend itself and posit itself as the universal, as thought.35
There is in Hegel’s pages on Anselm not a hint of the content of the vital passages which come after Prosl. II, mostly in the Reply. Thus Hegel essentially re-echoed the Gaunilo tradition, adding the peculiarities of his own dialectic of universal
35 Translated from the text of Georg Wilhelm Friedrich Hegel, Samsliche Werke (Stuttgart: Fr. Frohmanns Verlag, 1959), vol. xix, Vorlesungen dber die Geschichte der Philosophie, bd. iii, pp. 165, 167. [Hegel did a better job on the Cartesian ontological argument, giving some idea of the second or stropger version. See op. cis., p. 347.]
[237]
and concrete or particular, thought and reality. Obviously he is dealing in his own way with the abstract-concrete paradox, but is he making a clear advance toward its resolution? Or is he showing a muddled awareness of problems which are left for someone else to subject to lucid analysis?
12. Ludwig Feuerbach
Because it is arguable that Hegel’s greatest influence has been through the Marxists, it is worth noting that his acceptance of the Gaunilo legend was in a sense echoed in that main source of Marxist atheism, Feuerbach's Essence of Christianity. There, after quoting the Latin of Anselm’s definition of divine Greatness, the author tells us that the Proof runs: nonexistence is a defect (Nichtsein ist ein Mangel), therefore. . . .36 This of course is Prosl. II, for the hundredth time posing as the heart of the matter. One need add only two syllables to get an approximation to the proper form of the major premise: Nicheseinkonnen ist ein Mangel, the possibility of nonexistence is a defect. But to be wise enough to make and understand this addition, one might perhaps need to read what Anselm wrote on the subject. And this, it seems, one does not do, no matter to what school of philosophy one belongs.
Notable also is Feuerbach's assumption that the real existence of God, or of anything else, must be ‘particular’ or ‘empirical’. Much of Feuerbach’s brilliant attack upon theism reads like a diffuse exploitation of the abstractconcrete or Findlay paradox, that God both must and must not be an
36 The Essence of Christianity, trans. George Eliot (New York: Harper's, 1957, p. 198 (ch 20). German edition (Leipzig, 1904), p. 300
(ch. 21).
[238]
empty abstraction, devoid yet not devoid of concreteness or particularity. I ind in this writer no notion of the neoclassical solution of the paradox, that it is the actuality, not the existence, of God which must be particular or empirical, and that the existence is merely the being-somehow-particularized, not the how of particularization. Only the latter need be empirical. It is precisely the extreme abstractness of the divine essence which makes necessary its being somehow actualized or particularized. The essence of the human mind, which Feuerbach would substitute for divinity, is by no means so abstract. As Barth points out, man dies, his actions are pervasively tinged with evil, and he exists not as a single universal individual but as each one of us in our distinct individualities.3” These and many more restrictions upon the pure essence of understanding, love, or will are required to transform these concepts, with which divinity is identified by Feuerbach, into that of humanity. Such restrictions cannot be necessary. Thus Feuerbach, like his ‘bourgeois’ teachers, Kant and Hegel, failed to deal clearly and logically with the Anselmian challenge at its center.
13. Robert Flint
This author belongs with Cudworth as having come closer to Anselm than any of the other moderns considered so far in Part Two. Yet even he could not prevent the Gaunilo tradition from warping his presentation somewhat. Thus, he begins by stating the thought of Prosl. II as though it were the entire
37 See the conclusion of Barth's preface to The Essence of Christianity (1957).
[239]
technical argument, and only then begins to show that he really knows better.
It is heartening to note his awareness that Descartes had two forms of ontological proof (the second coming close to Prosl. III), and he may possibly be the first modern thinker to take clear note of this fact. Is it not permissible to salute this tough Scotchman’s insistence upon doing his own reading and thinking, instead of letting a dozen other persons do it for him?
This reasoning [he has summarized Prosl, II] . . . has commended itself completely to few. Yet it may fairly be doubted whether it has been conclusively refuted, and some of the objections most frequently urged against it are cerwainly inadmissible. . There is . . . no force, as Anselm showed, in the objection of Gaunilo, that the existence of God can no more be inferred from the idea of a perfect being, than the existence of a perfect island is to be inferred from the idea of such an island. There neither is nor can be an idea of an island which is greater and better than any other that can ever be conceived. Anselm could safely promise that he would make Gaunilo a present of such an island when he had really imagined it. Only one being—an infinite, independent, necessary being —can be perfect in the sense of being greater and better than every other conceivable being. The objection that the ideal can never logically yield the real—that the transition from thought to fact must be in every instance illegitimate—is merely an assertion that the argument is fallacious. It is an assertion which cannot fairly be made until the argument has been exposed and refuted. The argument is that a certain thought of God is found necessarily to imply His existence. The objection that existence is not a predicate, and that the idea of a God who exists is not more complete and perfect than the idea of a God who does not exist, is, perhaps, not incapable of being satisfactorily repelled. Mere existence is not a predicate, but specifications or determinations of existence are predicable. Now the argument nowhere implies that existence is a predicate; it implies only that reality, necessity, and independence of existence are predicates of existence; and it implies this on the ground that existence in re can be distinguished from existence
[240]
in concepts, necessary from contingent existence, self-existence from derived existence. Specific distinctions must surely admit of being predicated. That the exclusion of existence—which hete means real and necessary existence—from the idea of God does not leave us with an incomplete idea of God, is not a position, I think, which can be maintained . . . the idea becomes either the idea of a nonentity or the idea of an idea, and not the idea of a perfect being at all. Thus, the argument of Anselm is unwarrantably represented as an argument of four terms instead of three. . . .
The second form of the Cartesian argument is, that God cnnot be thought of as a petfect Being unless He be also thought of as a necessarily existent Being; and that, therefore, the thought of God implies the existence of God. . . . It is futile to meet this by saying that existence ought not to be included in any mere conception, for it is not existence but necessary existence which is included in the conception reasoned from, and that God can be thought of otherwise than as necessarily existent requires to be proved, not assumed. To affirm that existence cannot be given or reached through thought, but only through sense and sensuous experience can prove nothing except the narrowness of the philosophy on which such a thesis is based.38
Flint overlooks the abstract-concrete paradox, and the consequent need to transcend classical theism. But it was fifty years after his writing the above words, at least, before more than a minute fraction of the philosophical world had anything like so definite and accurate a grasp of what either Anselm or Descartes had meant. It is probably still a small fraction.
14. W. E. Hocking, Josiah Royce, and George Santayana
Anselm’s Proof has seldom played a positive role in Ameri-
38 Theism (Edinburgh and London, 1877), pp. 278-80, 282-84 [ Since Flint refers to Cudworth, the latter's rare understanding of the Argument was not quite lost in his vast volumes, after all.]
[241]
can philosophy. Two of our philosophers, however, Hocking and Royce, early in this century called an argument of theirs *ontological’, and one of these said, some fifty years ago, that it was in principle the ‘only’ argument for God. The rather pathetic facts, however, are that neither writer has given a tolerably accurate account of the original Argument and that one of them gives an exceedingly inaccurate account.39 (Since the writer of this commentary is deeply indebted to Professor Hocking as a former teacher, he wishes it were otherwise.) Hocking intended to improve upon Anselm. But alas, his Anselm never existed. In one short paragraph, four incorrect notions concerning the Saint’s procedures are expressed, as anyone who compares the account with the relevant Proslogium pages may ascertain for himself. Nor is there any suspicion of the central principle: that, the nonexistence of which is inconceivable, is greater than that the nonexistence of which is conceivable.
This is how Anselm has been treated—even by many of those with the least motive for misrepresenting him.
In a more constructive and characteristically ingenious passage, Hocking briefly presents his own so<alled ontological argument,
The ontological argument is the answer to the question, May the idea of God be ‘merely subjective?” That answer is, In forming the essence ‘merely subjective, you have at the same time formed the essence ‘not merely subjective’ as in contrast thereto; and God as essence belongs to the ‘not merely subjective.’ Whatever artificiality there is in the argument hails entirely from the
39 “The Ontological Argument in Royce and Others,” Contemporary I1dealism in America, ed. C. Barrett (New York: The Macmillan Compaoy, 1932), p. 49.
[242]
artificiality of the question. The natural situation may be stated thus: the essence of God must be real, because it is an essence inseparable from my continuous consciousness or experience of reality. 40
The second sentence by itself looks like the very argument of Prosl. II which Hocking has rejected. However, taking into account the last sentence, and the setting of the quotation in Hocking’s general position, what we have is not an ontological argument at all, but a Berkeleyesque version of what I call the epistemic or idealistic argument. This version runs: we experience nature, this amounts to experiencing God, since nature is only something in or ‘between’ minds, not a possible thing in itself, and since our perception of nature is not our own doing and can only be God ‘creating us’, at least to that extent. Berkeley had said much of this in his own way: nature is ideas, but not essentially our ideas; hence God’s, which He causes us to share (or approximate to), and thus a ‘language’ whereby God speaks to us. Calling such an argument ‘ontological’ seems to be twisting words to little purpose. The original Argument proceeded directly from the logical possibility or conceivability of God to His necessary existence. Nothing about our human situation relative to nature needed to be invoked. The conceivability of God is not the fact that human beings conceive Him; no failure to conceive can nullify a logical possibility. If we understand that God is conceivable, we also understand, for it is the same, that He necessarily exists. For in this unique case, nonnecessity is equivalent to nonpossibility. The point is related to what nature might be only insofar as all metaphysics is one. There may be a valid form of epistemic argument, but it is not the argument Anselm
40 Op. cit., p. 65.
[243]
invented (and Descartes used) in either its weak or its strong form. It needs another name.
In the same essay—based in part on students’ notes of some of Royce’s lectures—Hocking deals with the Roycean so-called ontological argument. It does not appear that Hocking's teacher came any closer to Anselm than Hocking himself. Nor is this surprising. “Our American Plato” (as Peirce called him) based his views mainly on German philosophy which, almost throughout its history, has been unaware of the real Anselm. And for another reason also, Royce was not in a good position to appreciate the Argument. (He did much better with the epistemic argument.) For Royce, though less clearly and candidly, took the view of the Stoics and Spinoza (and Leibniz by implication) that the eternal essence of God implicated the entire detail of existence. Thus everything is as necessary as God. To use the Proof for this view, with its glaring paradoxes, is to misuse it in one way as Anselm had in another. Classical theism and classical pantheism are the two horns of the classical dilemma. But of the two, it is the Spinozistic doctrine which more obviously must in effect “deduce the concrete from the abstract.” And, just as Hume’s absolute denial of internal relatedness makes nonsense out of theism, so does Royce's absolute denial of external relatedness and hence of contingency. This is an example of the power of Anselm’s discovery; it forces us to regard either theism, or else both extreme monism and extreme pluralism, as self-inconsistent.
That Royce, as a classical pantheist naturally would do, distorted the Proof by attenuating to the vanishing point the distinctiveness of necessary existence compared to existence in general is made quite clear in Hocking’s account. Royce wants to generalize the ontological principle that ‘essence entails exist-
[244]
ence’ by making experience entail the entire universe. Otherwise, he thinks, we cannot escape solipsism. Now, that experience cannot be of nothing, and that it must somehow exhibit the experienced, may indeed be a valid metaphysical axiom. But the inference from our experience to the particular world experienced is one from brute fact to brute fact, from particular to particular. This is almost as far as one can get from the original ontological argument. Royce conceals this from himself by identifying ‘essence’ with whatever is experienced or in the mind. However, Anselm starts from an abstract definition, not a concrete experience or anything concrete, and deduces that no matter what else there may be, God as defined must really exist. This is the total independence or self-sufficiency of God’s existence, its absolute neutrality with respect to all other individual existences. This notion—essential to the Proof—is swept away in Royce’s version (as in Spinoza’s).
Yet Royce is groping toward a metaphysical truth (so subtle is this problem, and so inclusive of all the metaphysical problems). The ontological principle may indeed be applied to more than just God. What it cannot do is apply to individuals other than God; rather, it applies to all abstractions or detetminables on the highest level of generality. God in His merely necessary reality is an abstraction, though a perfectly individual abstraction for all that! But God as such and world as such are correlative, and equally abstract; both must be somehow concretized. However, this our actual world—which was Royce’s theme—is the absolute opposite of the uttermost abstraction or determinable, since it is the uttermost determinate. To lump together these two problems without clear notification of what is going on is confusion indeed. We are being entangled in the pantheistic version of the Findlay paradox inherent in all
[245]
classical thought, without the admission of how paradoxical it is. The Perfect individual can be necessary only because its individuality is totally noncommittal as between particular alternatives of actualization. So far from the entire actual universe following from or being presupposed by its existence, nothing follows, and nothing is presupposed, save that some divine and some worldly actualization there must be. (That world as such is equally noncommittal does not make it a rival necessary being. For world as such is simply the content of the divine knowledge so far as more than mere self-knowledge. It designates, not an individual in the primary sense, but the completely indefinite generalized collection of nondivine individuals.)
As Hocking shows, Royce defined his ideas about the Argument partly by reaction to those held by his brilliant colleague George Santayana. In the first chapter of the latter’s Realm of Truth, we find the following characteristically luminous discussion.
The most real of beings, said St Anselm, necessarily exists: for evidently if it did not exist, far from being most real, it would not be real at all Is then reality, we may ask, the same as existence? And can existence have degrees? St Anselm explains . . . a nonexistent essence would woefully lack moral greatness, perfection, or dignity: it would be a contemptible ghost, a miserable nothing. Undoubtedly for a careladen mind seeking salvation—unless it sought salvation from existence—power, which certainly involves existence, must be the first mark of reality and value. . . .
At the other pole of reflection, on the contrary, as among the Indians or the Eleatics, the most real of things might seem to be pute Being, or the realm of essence, excluding change and existence altogether: because in change and existence there is essential privation. ...
I do not mention this paradox in order to laugh at St Anselm
[246]
or at his many disciples . . . their argument was fallacious and even ridiculous, if by ‘necessary existence’ we understand a necessity artaching to events ar to facts, that is, to contingencies. Yet the same argument breathes a fervent intuition and a final judgment of the spirit, if it intends rather to deny final validity to an existential order which, by definition, is arbitrary, treacherous, and self-destructive: a realm of being over which inessential relations are compulsory and essential relations are powerless. . . . we may come to sce how the maximum of reality mighe logically involve infinity, impassiveness, and eternity: all of which are contrary to the limitation, flux, and craving inherent in existence. No essence, not even this essence of existence, has any power to actualize itself in a fact; nor does such actualization bring to any essence an increment in its logical being. ...
The existence of God is therefore not a necessary truth: for if the proposition is necessary, its terms can only be essences; and the word God itself would then designate a definable idea and would not be a proper name indicating an acrual power. If, on the contrary, the word is such a proper name, and God is a psychological moral being energizing in space and time, then His existence can be proved only by the evidence of these natural manifestations, not by dialectical reasoning upon the meanings of terms.41
Our Americanized Spaniard’s brilliance has not saved him from missing much of the import of Anselm’s reasoning.
First, the proof which he attributes to the Saint is but a loose paraphrase of Prosl. II. In spite of his claim that when he began his studies at Harvard the scholastic proofs for God’s existence were “warm in his mind,” the evidence before us strongly suggests that he either had not read, or when he wrote the quoted passage had forgotten, Prosl. III and Reply I, V, IX.
41 The Realm of Truth (New York: Scribner’s, 1938), pp. 7-10.
[247]
Second, it is assumed, not proved, that all essences have the same relation to existence—of course, the very point at issue. Actualization, it is urged, involves arbitrary limitation, negation, or exclusion of other essences from actualization. Essences being thus competitors for actualization, their existence is always contingent. But what Anselm, in effect, proved is that there is one noncompetitive yet individual essence, wholly independent, in its mere existence, of what else does or does not exist. A dogmatic statement that all essence is competitive is worthless against this explicit argument for an exception to the rule.
To say that “no essence, not even this essence of existence, has any power to actualize itself” prompts the rejoinder, does any essence have power even to ‘subsist’ totally out of relation to existence? We confront the old Aristotelian question. Santayana cannot refute Anselm merely by voting, against Aristotle, for the complete “separability of forms.” (See Part One, Sec. 13.)
That no actualization can “bring to any essence an increment in its logical being” is correct; but it may for all that be true that the denial of actualization to so abstract or noncommittal an essence as existence as such, or God the universal existent, is contradictory of its logical being. In spite of Kant, it is illicit to substitute, Can you add existence to an essence and get a greater essence? for the only relevant question, which is, Can you without contradiction subtract existence from, or deny it to, a certain supreme (and noncompetitive) essence and have the essence at all? Such an impossibility of subtraction does not imply the possibility of a prior addition, rather the contrary! Kant's point here was an ignoratio elenchi. No Anselmian who remembers Prosl. 1II wants to “add” existence to the divine nature. He wants to show that there is no room
[248]
for such an addition, since existence is inherently there. The alternative to ‘God existent’ is not ‘God nonexistent and therefore inferior’—it is contradiction or nonsense.
So what we have is: lack of acquaintance with Anselm, inconclusive and in part irrelevant objections and dogmatic platonism. Here is no secure hiding-place from the Anselmian challenge.
Santayana also begs the question, not only by his extreme platonism, but by his extreme Humianism, or pluralism. Each momentary state of experience (‘spirit’) is independent logically of its predecessors. Mind has no power, only matter. But then of course there is no sense to the idea of God!
True enough, the “actual power” influencing our world transcends any essence; but all that the Proof tries to show, neoclassically interpreted, is that some divine actuality, genidentical with any other, is influencing it.
Santayana and Royce are interesting opposites to one another. Whereas Royce held, in effect, that all genuine essences must be actualized, since the world, in every detail, is the best or only ‘really possible’ choice that the Absolute Wisdom could make among apparent or ‘merely ideal’ possibilides, Santayana had said that any actualization, and even no actualization, is as possible, as thinkable, as the reality which in fact exists. Here, as always, a critical Anselmian takes the moderate position: the divine essence, and all equally general or abstract essences, cannot conceivably be unactualized, but the more particular essences may or may not acquire actualization. Again, whereas Royce held that all details are providential, a theist who has given heed to the abstract-concrete paradox will not say this; but neither will he, with Santayana, see the world as bald fact influenced by no universal ideal or directive.
[249]
So penetrating and various are the implications of Anselm’s discovery. The world has evidently felt a deep need to evade these implications; and meeting this need is a job which has been almost incredibly well done. The chief methodological rule for doing it has been simple: do not study or reflect upon Anselm’s text for yourself, rely upon another. Read only with another’s eyes, reflect only with another’s mind. Scholarly thoroughness, self-reliance, one or both, have nearly always been in scanty supply in this context.
There is a third quality insufficiently manifested in the Anselm debate: intellectual curiosity, the desire to explore ideas as a mathematician explores them with a view to learning what could coherently be thought, regardless of whether it has been thought or not, regardless of what one wishes to believe or disbelieve. How could one formulate the idea of a worshipful being? Is something like the Philonian-Thomistic approach the only possibility—apart from Spinozism, Leibnizianism, Hegelianism, Kantian agnosticism, or a Comtean rejection of the very idea of God? (The inexhaustiveness of the disjunction should be clear.) Like Anselm himself, Santayana did not seriously put this question, and neither did Royce, who dealt only with the issue between agnosticism (or pure mysticism) and some approach to an Hegelianized Leibnizianism.
Thanks to the influence of William James, Hocking did put the question just referred to. He is essentially a neoclassical theist (to the great profit of at least one of his former students), but with a post-Kantian and Roycean background somewhat unfavorable to lucid analysis. We need a larger dose of ‘Latin clarity’, such as characterizes the thought of Anselm.
[250]
15. R. G. Collingwood
This author gives us an admirable short account of the historical background of the Proof. One may doubt if the special role of faith is properly estimated in this account. But on the other hand, one could hardly have a clearer realization than the passage expresses of the enormous systematic philosophical issues which are entwined in this one.
An ingenious argument, not included in the quotation below, is given to show that logic and ethics, like metaphysics, must assert existence as involved in essence. Thus—for instance— in talking about propositions we create propositions, and hence their existence cannot be denied. This seems rather different from the necessity that God should exist; but possibly the consideration is relevant. In any case, if all existence is contingent, metaphysics is a will-o’-the-wisp, and so is much of what has been regarded as philosophy.
Of the three phrases quoted below from Boéthius, the last and most significant is in the Consolations of Philosophy, Book III, X. It is indeed close to Anselm: “nothing better than God can be thought of.” Boéthius uses the formula to prove that God must have perfect goodness, since otherwise, inasmuch as “perfect things were before the imperfect,” thete would be something better than He. Still, this is not Anselm. For the procedure requires us to prove independently that perfect goodness exists, whereas all that is needed is that it be conceivable, since then that which is thought of as not admitting even the thought of a greater must be thought of as perfectly good. Similarly it must be thought of as existing necessarily—again assuming only that necessary existence is conceivable and would be superior to contingency. There is another bit of
[251]
evidence that Boéthius did not have any very penetrating grasp of his own formula: in the very sentence in which he in effect defines God as such that nothing better can be thought, he goes on: “who doubteth but that that is good than which nothing is better?” (id quo melius nihil est), thus relapsing from the modal to the merely existential level. Nevertheless, Collingwood is probably right in assigning this passage the honor of most nearly approximating Anselm’s definition. Very likely, however, it was Augustine, if anyone, who led the Magnificent Doctor to his formula. Thus in De Libero Arbitrio (Vi, 14) we find God referred to as quo nullus est superior, or again quo nihil superior est constiteris. Also in De Doctrina Christiana, L I, C VII, we read: ut aliquid quo nihil melius sit atque sublimius illa cogstatio conetur astingere. (For these citations I am indebted to F. H. Ginascol.) None of these quite say that God is such that nothing greater can be thought, but only that nothing greater exists, or can be demonstrated, or that thought about God tries to arrive at a being to which nothing would be (sit) superior. Perhaps the difference is hairsplitting? If I am right that the modal structure of Anselm’s argument is decisive, nothing is more essential to it than the explicit, unmistakable reference to conceivability. And this we find in Boéthius, not quite in Augustine. But whether this passage of Boéthius actually influenced Anselm is, of course, another question.
Plato had long ago laid it down that to be, and to be knowable are the same (Rep. 476 E); and, in greater detail, that a thought cannot be a mere thought, but must be a thought of something, and of something real (ovros, Parm. 132 B). The neo-Platonists had worked out the conception of God in the metaphysical sense of the word—a being of whom we can say est id quod est, a unity of existence and
[252]
essence, a perfect being (pulcherrimum fortissimumque) such that nihil deo melius excogitari queat (the phrases [except the last] are from Boéthius, De Trinitase). Anselm, putting these two thoughts together, the original Platonic principle that when we really think (but when do we really think, if ever?) we must be thinking of a real object, and the neo-Platonic idea of a perfect being (something which we cannot help conceiving in our minds; but does that guarantee it more than a mere idea?), or rather, pondering on the latter thought until he rediscovered the former as latent within it, realized that to think of this perfect being at all was already to think of him, or it, as existing.
. . . Anselm . . . was careful to explain that his argument apphcd not to thought in general, but only o the thought of one unique object . . . the slightest acquaintance with writers like Boéthius and Augustine is enough to show that he was deliberately referring to the absolute of neo-Platonic metaphysics; and in effect his argument amounts to this, that in the special case of metaphysical thinking the distinction between conceiving something and thinking it to exist is a distinction without a difference.
. . . Of all the legacy of medieval thought, no part was more firmly seized upon than the Ontological Proof by those who laid the foundations of modern thought . . . and it remained the foundation-stone of every successive philosophy until Kant, whose attempt to refute it—perhaps the only occasion on which anyone has rejected it who really understood what it meant—was rightly regarded by his successors as a symptom of that false subjectivism and consequent skepticism from which, in spite of heroic efforts, he never wholly freed himself.
. . . the Proof is not to be dismissed as a quibble . . . what it does prove is that essence involves existence, not always, but in one special case, the case of God in the metaphysical sense: the Dews ssve natura of Spinoza, the Good of Plato, the Being of Aristotle: the object of metaphysical thought. But this means the object of philosophical thought in general; for metaphysics, even if it is regarded as only one among the philosophical sciences, is not unique in its objective reference . . . all philosophical thought . . . partakes of the nature of meta
[253]
physics, which is not a separate philosophical science but a special study of the existential aspect of that same subject-matter whose aspect as truth is studied by logic, and its aspect as goodness by ethics.
Reflection on the history of the Ontological Proof thus offers us a view of philosophy as a form of thought in which essence and existence . . . are conceived as inseparable . . . unlike mathematics or empirical science, philosophy stands committed to maintaining that its subjectmartter is no mere hypothesis, but something actually existing.43
Let it be noted well that, even in this exceptionally sympathetic account, there is no distinct echo of the logic of Prosl. III, As a result the formulation, like so many others, is less cogent than Anselm’s own, when that is taken as he wrote it, without truncation or mutilation. After all, the questions are, why does this essence involve existence? And how do we know that it does? Anselm saw more deeply into these questions than most of those who came before or after him.
The assertion that Kant understood the Proof is of doubtful value in the absence of any evidence that Collingwood had in mind the actual course of the Proof in the ten essential Anselmian pages to which we have repeatedly referred.
16. Hans Reichenbach
Here is our last example of a philosopher refuting an Anselm who never existed, an Anselm whose only wisdom was that existence is one of the predicates which a thing may be conceived to have. Anselm was a man who thought this, but he was not a man who had no other thoughts relevant to the Proof. And only that fictitiously unresourceful man is refuted in the following.
43 Philosophical Method (Oxford, 1933), pp. 124-27.
[254]
[Refuting Prosl. 11]
[Anselm’s] demonstration begins with the definition of God as an infinitely perfect Being; since such a being must have all essential properties, it must also have the property of existence. Therefore, so goes the conclusion, God exists. The premise, in fact, is analytic, because every definition is. Since the statement of God's existence is synthetic, the inference represents a trick by which a synthetic conclusion is derived from an analytic premise.
. . . If it is permissible to derive existence from a definition, we could demonstrate the existence of a cat with three tils by defining such an animal as a cat which has three tmils and which exists. Logically speaking, the fallacy consists in a confusion of universals with particulars. From the definition we can only infer the universal statement that if something is a cat with three tails it exists, which is a true statement. . . . Similarly, we can infer from Anselm’s definition only the statement that if something is an infinitely perfect being it exists, but not that there is such a being.48
Whether, or in what sense, ‘God exists’ is synthetic is of course the very question at issue. In Prosl. III it was shown that ‘Greatest Conceivable’ is not conceivable unless as necessarily existent; hence, since a statement which is conceivable but only as true is in the broad sense analytic, either ‘God exists’ is true and analytic, (true ex vs terminorum) or it is not conceivable, is not a genuine proposition. It certainly is not synthetic in any but a narrow and merely technical sense.
The reductio ad absurdum from the definition of an ‘existent three-tailed cat’ is invalid against the above argument; for a definition which includes existence in the a priori specifications of a thing makes sense only if the form of existence involved
48 The Rise of Scientific Philosophy (Berkeley: University of California Press, 1951), p. 39.
[255]
is necessary existence. But then, since ‘necessarily-existing cat’, like ‘perfect island’, is an absurdity, the reductio is therefore invalid (since anything can be deduced from an absurd concept, taken not to be absurd).
Concerning particular and universal statements: ‘God exists’ is not ‘particular’ in the same sense as ‘cats exist’ is so. As we saw in Part One, ‘Greatness exists’ is, in a relevant sense, as nonparticular as ‘some individuals exist’, which Reichenbach in his logic allows to be an assumption of logic itself.
I think we can dismiss this refutation, along with so many others like it, as missing the mark.
17. J. N. Findlay
The most important contribution since Kant to the Anselmian controversy, on its skeptical side, has in my judgment been made by this author.
The proofs [for God's existence] based on the necessities of thought are universally regarded as fallacious; it is not thought possible to build bridges between mere abstractions and concrete existence . . . Religious people have . . . come to acquiesce in the total absence of any cogent proofs of the Being they believe in; they even find it positively satisfying . . . And nonreligious people . . . don’t so much deny the existence of a God, as the existence of good reasons for believing in Him. We shall, however, maintain that there isn’t room, in the case we are examining, for all these attitudes of . . . doubt. For . . . the Divine Existence can only be conceived, in a religiously satisfactory manner, if we also conceive it as something inescapable and necessary, whether for thought or reality. From which it follows that our modern denial of necessity or rational evidence for such an existence amounts to a demonstration that there cannot be a God.
... We ask . . . whether it isn't wholly anomalous to worship anything lsmited in any thinkable manner. For all limited superior-
[256]
ities are tainted with an obvious relativity, and can be dwarfed in thought by still mightier superiorities, in which process of being dwarfed they lose their claim upon our worshipful attitudes And hence we are led on irresistibly to demand that our religious object should have an unsurpassable supremacy along all avenues, that it should tower infinitely above all other objects. . . . We ask also that it shouldn’t stand surrounded by a world of alien objects, which owe it no allegiance, or set limits to its influence. The proper object of religious reverence must in some manner be all-comprehensive: there mustn’t be anything capable of existing, or displaying any virtue without owing all of these absolutely to this single source. . . . But we are led on to a yet more stringent demand: . . . we can’t help feeling that the worthy object of our worship can never be a thing that merely happens to exist, nor one on which all other objects merely happen to depend. The true object of religious reverence must not be one, merely, to which no actual independent realities stand opposed; it must be one to which such opposition is $mconcesvable. God mustn't merely cover the territory of the actual, but also, with equal comprehensiveness, the territory of the possible. And not only must the existence of other things be unthinkable without Him, but His own nonexistence must be wholly unthinkable in any circumstances. And so we are led on . . . to the barely intelligible notion of a Being in whom Essence and Existence lose their separateness.
It would be quite unsatisfactory from the religious standpoint, if an object merely happened to be wise, good, powerful, and so forth, even to a superlative degree, and if other beings had, as a mere matter of fact, derived their excellence from this single source. . . . Wisdom, kindness and other excellences deserve respect wherever they are manifested, but no being can appropriate them as its personal perquisites, even if it does possess them in a superlative degree. And so an adequate object of our worship must possess its various qualities in some necessary manner. These qualities must be intrinsically incapable of belonging to anything except in so far as they belong primarily to the object of our worship. Again we are led on to a queer and barely intelligible Scholastic doctrine, that God isn't merely good, but is in
[257]
some manner indistinguishable from His own (and anything else’s) goodness.
What, however, are the consequences of these requirements . . .? Plainly (for all who share a contemporary outlook), they entail not only that there isn't a God, but that the Divine Existence is either senseless or impossible. . . . Those who believe in necessary truths which aren’t merely tautological, think that such truths merely connect the possible instances of various characteristics with each other; they don’t expect such truths to tell them whether there wsll be instances of any characteristics. This is the outcome of the whole medieval and Kantian criticism of the Ontological Proof. And, on a yet more modern view of the matter, necessity in propositions merely reflects our use of words, the arbitrary conventions of our langusge. On such a view the Divine Existence could only be a necessary matter if we had made up our minds to speak theistically whatever the empirical circumstances might turn out to be. . . . This wouldn't suffice for the fullblooded worshipper. . . . The religious frame of mind seems, in fact, to be in a quandary; it seems invincibly determined both to eat its cake and have it. It desires the Divine Existence both to have that inescapable character which can, on modern views, only be found where truth reflects an arbitrary convention, and also the character of ‘making a real difference’ which is only possible where truth doesn’t have this merely linguistic basis. We may accordingly deny that modern approaches allow us to remain agnostically poised in regard to God; they force us to come down on the atheistic side. . . . Modern views make it selfevidently absurd (if they don’t make it ungrammatical) to speak of such a Being and attribute existence to Him. It was indeed an ill day for Anselm when he hit upon his famous proof. For on that day he not only laid bare something that is of the essence of an adequate religious object, but also something that entails its necessary nonexistence.44
This is the only refutation of Anselm known to me, at least of those published before 1958, which shows an awareness of
44 “Can God's Existence be Disproved?,” Mind, 57 (1948). Reprinted in Findlay's Language, Mind, and Value. [See Bibliography for a relevant quotation from this book.]
[258]
what Anselm’s Proof in its essential steps actually was. And it puts the matter in a different and clearer light than other refutations. Anselm declared that God could not exist contingently, from which by modal axioms it follows that He could not contingently fail to exist. What then? Obviously, either He could and does exist necessarily, or He necessarily fails to exist, i.e, His existence is logically impossible. Anselm, defended recently by Malcolm, rejects impossibility and infers necessary existence; Findlay rejects the logical possibility of necessary existence and infers the impossibility that God should exist at all. Most critics, so far as they have been at all clear about what they were doing, have by implication at least asserted either that God might contingently exist just as He might contingently fail to exist, or that He might contingently fail to exist even though, should He exist, it must be necessarily. The second alternative—here Anselm, Descartes (not quite so explicit on the point), Cudworth, Flint, Koyré, Malcolm, Findlay, and I agree—is modal nonsense. And the first— here again we agree—contradicts the meaning of ‘God’.
Findlay's ‘disproof’ then is: ‘if the idea of God is logically possible, the axiom of the nondeducibility or contingency of existence cannot be universally applicable; yet considering the manifest impossibility of deducing concrete actuality from mere universals, abstract concepts, or definitions, the axiom must be universally applicable. Not even divine actuality can be an exception’. This is a significant and serious paradox, like those of logic without some equivalent of type theory. It is a mistake to brush it aside as a mere sophistry, thereby refusing a chance to learn, as logicians have learned from their paradoxes. It is this brushing aside of serious difficulties which has partly vitiated nearly every contribution to this controversy.
[259]
That the concrete cannot be deduced from the abstract is an unshakable truth. Findlay is right to say so, and Malcolm here seems simply to miss the point. Findlay is also right, however, with Anselm and many other great men, when he says that God is the necessary being—or nothing thinkable. He is wrong only in supposing that a deducible or necessary divine existence must be actual or concrete. And this is the solution of the paradox: God's existence is not itself an actuality and is as abstract as the concepts from which it is deduced. It is their irreducible content. Nor need ot can it ‘make a difference’ (except that our awareness of it makes a difference to us). It is the particular actuality of God which makes objective differences, and this is not what the Proof proves, but only that there is some such actuality, making some appropriate difference or other. Classical theism cannot admit this solution, which implies a real distinction in God between abstract existence and concrete actuality, the former necessary, the latter contingent. We have seen that this is just what neoclassical theism not only admits but gladly asserts.
The strength of Findlay's argument can be seen in this way. To refute the Anselmian contention of the incompatibility of perfection with contingency, one must maintain that God could conceivably exist contingently. Accordingly, skeptics who reject Findlay’s conclusion are obligated to explain how they conceive an existence as perfect and yet contingent, or how they explain away the appearance of imperfection or essential limitation in the notion of accidental existence. When have they seriously attempted this? They usually do not even see the problem. And when they do, they treat it casually indeed. But it is the heart of the matter, not a detail. Findlay sees this, and strikes a blow at the heart. He grants that an implication
[260]
of the idea of God is noncontingency; but since the idea also implies concreteness, actuality, and this cannot be necessary, the idea has no consistently conceivable relation to existence. On the one hand, religious thought and feeling require a unique or absolute security and logical priority for the divine existence; on the other, logic apparently rules out any such thing; ergo it rules out religious belief.
Findlay thus shows us a new aspect of the Anselmian problem which the older refutations failed to make explicit; but he still does not freely explore the terrain. He also knows too well the conclusion he must reach about the whole matter. That the Proof might, while disposing of the old conclusion, point the way to a different but equally positive one, he did not (at the time he wrote the quoted passages) for a moment imagine. At least, however, he explored the terrain in one rather new direction.
Certainly a theist must be prepared to ‘speak theistically’ no matter what the empirical facts. For he sees the meaning even of ‘possible fact’ in theistic terms; how then can it matter, for the mere question of theism as such, which possibilities are actualized? But a theist need not and will not speak about the concrete divine actuality regardless of the facts. ‘God knows I am innocent’ is inappropriate if in fact I am guilty. But be I innocent or guilty, God exists, and knows me as I am, if in fact I exist at all. And if I do not, then He exists knowing that situation as it is. It drives one almost to despair sometimes that such plain distinctions should be thought esoteric or irrelevant.
The most impressive aspect of Findlay’s reasoning is not his reliance on ‘modern’ views—which is at best rather questionbegging. It is rather his point, at the outset, about the ab-
[261]
stract and the concrete, or—as one may also say—the logically weak and the logically strong. From the former there cannot be a deductive bridge to the latter. This is not modern, it is timeless good sense. But the question remains: is ‘somehow actualized’ any less abstract than ‘perhaps unactualized’? On the assumption that there is a conceivable status of ‘unactualized Greatness’ and another conceivable status of ‘actualized Greatness’, then the assertion of the latter must be less abstract, and logically stronger. But this double conceivability in relation to actualization is just what Anselm and Findlay— wiser than he knows—deny. Naturally, one must admit the double conceivability in ordinary contingent cases; the contrast of God with these conceivably nonexistent natures not only harmonizes with their admission but requires it. The admission of ordinary things as conceivably nonexistent (and therefore not divine) is the same point, put from the other end.
Beyond the abstract-concrete paradox, overcome in neoclassical doctrine, what is left of Findlay’s ‘atheistic’ argument (really ‘positivistic’, in the categorical or dogmatic sense, as I have been using terms) ? Does not what was well designed as an attack on traditional theism become a valuable instrument of defence for a less conventional form of religious thought?
18. Robert S. Hartman
Apparently inspired partly or chiefly by Barth’s fine work, an exciting and imaginative essay on Anselm’s argument has recently appeared. It is not wholly clear what its author would do with the Findlay paradox, yet there is some evidence that he may be aware of it. (The promised sequel to the essay should clarify the question.) This article is another of the
[262]
recent studies which seem to show that to read Anselm carefully is to see that the standard criticisms are far from adequate, and that we are by no means finished with the most famous— and most misunderstood—of all philosophical arguments.
Hartman tries to save Prosl. II by arguing that the existent is richer in properties than the merely possible; hence the ‘richest possible’ thing must be conceived as existent. One can still object that ‘richest conceivable’ thing means, that which would, if it existed, be richest; how then can the Argument get under way? For if the thing existed, it would also have the richness accruing from existence; while if it did not, there would be no ‘it’ characterized by the greatest richness, and hence no question of ‘its’ having a character incompatible with that— hence no contradiction. Only inconceivability-of-nonexistence, as a property, can rule out this escape.
Our author’s suggestion is also open to the grave objection that an ‘absolute maximum of richness’ is doubtfully consistent, as we have seen earlier.
Hartman’s contention that the Proof is ‘not analytic’ should be taken in conjunction with his comparison of it with mathematics. The synthetic element is in the application of a pure logical form to an intuitive datum, in this case religious faith. But (a) nothing is taken from faith except the ‘name’ of God which specifies what the question about His existence is to mean. So the Proof is an analysis of the meaning of ‘God’ for faith, and in this sense is analytic. Furthermore, (b) since the existence of God is, according to faith, the very principle of all actual or possible existence, and truth about it is Truth itself, this existence is no factual matter, in the proper sense of fact, nor, therefore, is its afirmation synthetic in the most usual sense. The Proof shows that the sole way to render faith con-
[263]
sistent is to take it as mecessarsly true. This truth is then analytic in the broad sense that its falsity is either nonsense or contradiction.
Hartman's point of view may be in accord with this. But to be wholly clear as to his meanings one must be acquainted with his own original theory of axiological axiomatics, a doctrine whose careful evaluation belongs to the future.
Proslogion 2-4 consists of four pages . . . All the books and essays written about them . . . would fill libraries . . . What is it in these four pages, that makes them so potent a challenge to the best minds of humanicy? . ..
I suggest that we have in these pages the first and so far the last—that is, the only—example in the history of thought of an entirely new philosophical method, which is the exact opposite of what has usually, both before and after Kant, been regarded as this method. It is neither categorical nor analytic, but axiomatic and synthetic. It is the method of mathematics itself, . . . the most effective cognitional method, that of science, applied to the most sublime subject. ...
The reason that the proof has been so puzling and challenging is that Anselm’s argument presupposes a whole new system of thought of which the argument itself is only a part—Ilike a promontory of a vast hinterland shrouded in fog. The passion aroused by the argument is that of exploring the known unknown—the passion of discovery, which is aroused by the expectation of finding unheard-of treasures in a realm which is known to be there but whose nature is unknown.
. . . A featureless existence is the exigentia of an essence: it is a Problem. The most powerful method of problem solving is the scientific, in particular, the axiomatic method, which posits a formula originating a system of thought applicable to the reality in question. The most famous example of this procedure is Newton, whose gravitational formula originated a system applicable to the whole universe. So Anselm, according to his own account, found a formula applicable to divine reality. The result was for him a matter of
[264]
rejoicing, delectatio—the same kind of joy a scientist feels on breaking through to the properties—the definition—of what he knew was there but did not know what it was.
Behind the Anselmian proof there is hidden an axiomatic system of which the proof itself is a small part . . . in such a system Anselm’s axiom—the name of God: ‘that than which nothing greater can be thought'—would appear not as axiom but as . . . theorem of a system. This system—the meta-Anselmian axiomatic—would have its own axiom, from which the Anselmian name of God would be deduced as a theorem; and the whole proof would become an integral part of the system itself....
. . . Barth makes it clear that the proof is not an analytic but a synthetic or axiomatic one. Anselm posits synthetically the name of God, as a formula, and deduces the proof from it . . . modo geometrico. The proof thus is purely formal: the name, the significance, proves itself. Where the name of God has been announced, heard, and understood, God exists in the cognition of the hearer—but . . . a God who only existed in cognition would be in an intolerable contradiction to His own revealed and believed name; He would have the name of God but would not be what the name says . . . The proof is really a reductio ad absurdum.
Vere ess means God is not only in thought but also over agasnss thought But God does not exist for thought as does any object created by Him: . . . Prosl. 3 . . . proves that God stands over against thinking in the unique manner in which the creator stands over against the thinking of the created creature.
The premise for Anselm is a word . . . but a word of God in the context of His revelation . .. From this . . . it is true, one cannot deduce
His existence; but one can deduce the impossibility of His nonexistence. The proof does not satisfy Gaunilo because, positivist that he is, he seems to want a proof of God based on some sense experience; and such a proof has nothing to do with the Anselmian insellectus fides and would be incompatible with Anselm'’s concept of God.
. .. the third chapter of Proslogion adds a higher stage to the masus and msnus of the second chapter. . .. A being whose existence is independent of the dialectic of cognition and object is maius and belongs to
[265]
a higher stage of being than a being which, no matter how truly it exists, how vere est, is subject to this dialectic; that is, whose existence can hypothetically be denied by the same thinking which may also affirm it The former is an absolute being beyond the opposition of the subjective and objective. It not only exists in truth but it exists ar truth; it is the truth of existence itself. . . .45
19. Jan Berg
Among the few careful attempts to apply the techniques of modern logic to Anselm's text is the one we are now to consider.*® The following symbols are used:
*3x (—x—)"’ ‘there is an x such that —x—’
‘1x(—x—)’ ‘the unique x such that—x—"’
‘G’ abbreviates ‘nihil maius cogitari possit’. Thus ‘G(x)’ means that ‘nothing greater than x can be conceived’.
'~' ‘it is false that’ After several attempts to formulate Anselm’s reasoning in Prosl. 11, attempts which are found to assume what is to be proved (the divine existence), the following formula is said to avoid this assumption:
‘~3y(y =1x(G(x)))>(~G) (x(G (x)))’ This formula says: ‘If there is not a y identical with the unique x such that none greater can be conceived, then the unique x than which none greater can be conceived has the property of being not such that none greater can be conceived’. To avoid the contradiction in the consequent one must deny the antecedent in the initial formula, by asserting:
45 "Prolegomens to a Meta-Anselmian Axiomatic,” Review of Metaphysics 14 (1961), 638, 640, 666f.
46"An Examination of the Ontological Proof,” Theoris 27 (1961), 99106.
[266]
‘3y (y=1x (G) )’ ‘There is a y identical with the unique x such that none greater can be conceived (God exists)’.
Though formally valid, the argument throws no light upon the justification of the initial premise. This is not a logical truism, but depends entirely upon an assumed peculiarity of the predicate G from which it is to follow that its failure to be instantiated in a unique individual would be contradictory. Here Berg has nothing to offer unless it be the Prosl. Il notion that the existent is superior to the nonexistent. The difficulties with this principle are notorious and, as we have seen, are not obviated by rejecting, as Berg does, the dictum ‘existence cannot be a predicate’. (See above, Sec. 10.)
Berg does not himself afirm the correctness of the formula, nor (I think) is he convinced of its correctness. This is not surprising since he makes no use of the modal principle that the necessarily existent is superior to the contingently existent. His formalization is in extensional, not modal, logic. And he thinks Malcolm’s (and my) distinction between the simple and the modal form of the Proof ‘mistaken’. But only the modal form gives a clear reason for imputing a contradiction to the hypothesis of noninstantiation.
If the modal argument is nos distinct, then what is one to make of the assertion in Prosl. III that to be necessarily existent is to be superior to what is contingent? If this principle is valid, why can it not be used as premise for a proof? Will Berg argue that the validity of the principle cannot be knowa unless it is known that something does exist necessarily? But, as Anselm implies, one needs to assume only the conceivability of necessary existence and to have an insight into the superiority of the status so conceived. For, granting these two points, it follows that to apply the property G to a contingently existing
[267]
thing would be contradictory; furthermore, to assert that a property is not instantiated is to imply that its noninstantiation is conceivable, and hence that any possible instance of it would be something existing contingently. Will Berg say that there is no reason to regard necessary existence as superior? But anyone who holds that is so remote from Anselm’s point of view that he might do better than try to construe his text.
Nevertheless it is good that the attempt has been made to express Anselm in nonmodal yet formal terms and that a way has been found to construct in these terms a valid Anselmian argument, valid save (no trifling qualification) for the blindness of its initial premise.
20. Jerome Shaffer
One of the many recent attempts to improve upon traditional criticisms of the Argument discards the contention that the Argument is invalid simply because existence is not a predicate. We can, if we wish, make it a predicate in a certain case by definition. But then, it is urged, there still remains the question, does anything correspond to the definition?
Until further arguments are offered, it seems reasonable to hold that there is nothing logically imptoper in so defining the expression, ‘God’, that *God exists’ is a tautology and *God does not exist’ selfcontradictory. In fact it seems to me that the definition I have given expresses a concept of God (i.e., as necessarily existing) which many people actually accept (just as it is a common conception of Satan that he merely happens to exist). I wish . . . to show that this concept of God can give no support to the religious. I shall argue that no matter what its content, this concept of God is still simply a concept. What must be shown, and what cannot be shown just by an analysis of the concepr, is that there actually exists something which answers to the concept. Even if
[268]
we have here the concept of an object which necessarily exists, a further question remains whether any existent meets the specifications of the concept. The difficulty lies in showing that this further question makes sense, for I have admitted that ‘God exists® is a necessary statement, analytically true, and therefore it looks as if there could be no further question. But that is an illusion. It must however be dispelled.
As a first step, I wish to point out that the concept of God is hardly unique in its capacity to generate a tautological existential statement. . . . Suppose we introduce the word, ‘particular’ to mean ‘object which exists’, and the word ‘nonentity’ to mean ‘object which does not
exist’. Then . .. we might say tautologically, *Particulars exist and nonentities do not exist’. . . The following sentences all have tautological
uses: ‘Existences exist’, ‘Fictitious objects do not exist’, ‘Members of extinct species existed once but no longer exist’ . . . these sentences . . . may be used tautologically in those circumstances in which we wish to include as a necessary feature, as a defining element, notions of existence or nonexistence.
. . Take the tautology, ‘Fictitious objects do not exist’. One mighe think that this means the same as *There are no fictitious objects’. But . . . this is incorrect, for although the former is true the latter is false. There are fictitious objects, many of them—aAlice’s looking glass, Jack’s bean stalk . . . to mention only a few. In general, given a tautology of the form, ‘A’s exist’, we cannot deduce from it, *There are A's’, nor from a tautology of the form, *A’s do not exist’ can we deduce 'There are no A's’. And specifically, given the tautology, ‘God exists’, we cannot deduce from it, 'There is a God’. The statement, ‘God necessarily
exists, but there is no God’, is not self-contradictory.
As it stands the situation is most paradoxical. . . .
It is tempting to try to resolve the paradox in accordance with Aristotle’s principle that ‘there are several senses in which e thing may be said to be’. Then to say that fictitious objects do not exist would be to say that fictitious objects lacked, say, spatio-temparal existence, whereas to say that there are fictitious objects would be to say that they had some other kind of existence—hence no contradiction. . . . But appea]s to the systematic ambiguity of ‘exists’ will not work in all cases. . . . For example, it will be tautologically true that
[269]
particulars exist in precisely the sense of ‘exist’, say, temporal existence, that it might be true that there are no particulars.
A more promising line of argument consists in showing that a tautological existential claim is quite different from a non-tautological existential claim. How are we to explain the difference? Suppose we say that a tautological existential assertion consists in attributing to the subject a special property, the property of necessary existence. We could explain this property by saying, ¢ ls Malcolm, that a being which has this property is such that it is senseless to speak of its non-existence or of its coming into existence or going out of existence or of the existence of anything else as a condition of its existence (pp. 44-59). Now this account will not do. First, the attempt to explain the necessity of the statement by postulating a special property commits us to an infinite regress of propetties, for presumably this special property might not be one which a being just happens to have but one which it necessarily has and which it is senseless to speak of its not havmg, and thus by similar reasoning we are led to necessary necessary-existence, etc. And it is most unclear what these properties could be or how we could distinguish them. But secondly, it is not clear what this property of necessary existence, is, if this is any more than a way of saying that the existential proposition is necessary. Am I making anything clearer when I say that squares, which are necessarily four-sided, have the special property of necessary four-sidedness? A defining property is not a special kind of property. So the tautological character of the existential assertions I have been discussing cannot be explained by postulating a special predicate, necessary existence. Their tautological character arises from nothing but the definition we have stipulated for the subject term. . . .
What lies at the heart of the puzzle about the Ontological Argument is the fact that our concepts have two quite different aspects, marked by the familiar philosophical distinction of intension and extension ... In making assertions about the extension of a concept there are typical forms of expression which we use: * . .. exist’, ‘There are no ... ' '... are scarce’, . . . etc. That such expressions are typically used in assertions about the extension (or lack thereof) of particular concepts is what is correctly brought out in the slogan, * ‘exists’ is not a pred-
[270]
icate.” But the typical use is not the only use. Since any statement, with suitable definitions, can be true by virtue of the meanings of the terms, sentences with existential expressions can be used to express tutological statements. . . . ‘Particulars exist’, when asserted tautologically, is used to make a claim about the meaning of the word, ‘particulars’, and therefore cannot be used to make a claim about the extension of the term. Similarly, if someone uses the sentence, ‘God exists’, tautologically, he tells us only that being an existent is a logical requirement for being God. If, on the other hand, someone asserts, ‘God exists’ nontautologically, then he claims that the term, ‘God’, has extension, applies to some existent. . . . The prima facie plausibility of the Argument comes from the use of a sentence intensionally when the typical use of that sentence is extensional. In this way it conceals the illicit move from an intensional to an extensional statement.
. . . Even when we have an existential tautology like ‘Particulars exist’, or ‘God exises’, it still remains an open question whether the concept of particulars or the concept of God has application, applies to any existent. What is settled at one level is not settled at another. It is important to see that we can go on to settle the question at the other level, too, for we can make it a priors true that the concept has application. For example, let the expression, ‘the concept of God’, mean, ‘a concept which has application and applies to a being such that . . .. Then by definition the concept of God has application; the statement, *The concept of God has application’, is now a tautology, given the definition. But nothing is gained by such a maneuver. ... We have framed a concept, namely the concept of the concept of God, and this concept makes certain statements tautologically true. Yet we can still raise the extensional question, Does this concept refer to any existent? At this level the extensional question would be whether there actually is a concept of God such that this concept has extension, and there is such a concept only if actually there is a God . . . Nothing has been settled except the meaning of a certain expression. . ..
A ... toublesome threat to the intension-extension distinction arises when we try to apply the distinction to certain concepts . . . suppose we ask whether the concept of a number has extension . . .
[271]
What makes this case puzzling is that we have no ides what would count as establishing that the concept of a number has extension or that it does not have extension . . . Nothing would count as showing that the concept of numbers had extension over and above its intensional content, and this is to say that numbers are intensional objects.
The same thing must be said for the existence of God. The most that the Ontological Argument establishes is the intensional object, God, even if this intensional object has the attribute of existence as an intensional feature. To establish that the concept of God has extension requires adducing some additional arguments to show that over and above its intensional features, over and above the content of the concept (or the meaning of the word, ‘God’), the concept of God has extension as well. This additional argument will of necessity have to be an a posteriori argument to the effect that some actual existent answers to the concept. We are thus led to the result that the Onrological Argument of itself alone cannot show the existence of God, in the sense in which the concept is shown to have extension. And this is just as the religious wish it to be. They do not conceive of God as something whose being expresses itself entirely in the concepts and propositions of a language game. They conceive of Him as something which has effects on the world and can in some way be experienced. Here is a crucial respect in which His status is meant to be different from that of numbers. The concept of God is a concept which msght have extension. But some further argument is required to show whether it does or not.47
Let us begin with the fifth and fourth sentences from the end of this clever argumentation. As we have reiterated, there is no necessity to identify the full actuality of God with His bare existence and good reason not to make this identification. God can have particular effects on the world without His mere existence being the entire cause of these effects. Existing neces-
47 Existence, Predication, and the Ontological Argument,” Mind, 71 (1962), esp. pp. 318-325.
[272]
sarily, He can also make contingent ‘decrees’ (and He must make some such decrees or other). It is these free decrees which furnish the concrete reality of deity with which religion is trying, in the feeble way open to man, to deal. So we can reject Shaffer’s attempt to show that the God proved by the Argument is merely intentional or abstract and not the God of religion. It is the God of religion, but only in His most abstract, necessary aspect, which is presupposed by all more concrete and contingent aspects. The necessity is that the class of the latter aspects be nonempty.
A concept can have extension only if it has intension. But there are three sorts of concepts, those whose intension is neutral with respect to whether or not the concept has extension, those which forbid such extension (e.g., ‘absolutely isolated part of the universe’), and those which requsre some extension or other, on pain of falling into mere absurdity. Suppose the class of ‘existents’ were empty. The supposition has no clear meaning. No one could possibly know such a ‘state of affairs’, and it would not in any intelligible sense be a state of affairs. The reason is not in any mere stipulation. The word ‘existent’ might have been used for something other than the widest class which must have members, but then either another word would be used for this class, or language would be left, insofar, incomplete. Our author seems simply to assume that there can be no such necessarily nonempty class. But that is part of what is at issue in an ontological argument. And he fails to remind us that many logicians think it a requirement of logic itself that the widest class of entities should be nonempty.
Specifically, the Argument, neoclassically interpreted, shows that the class ‘divine states, genidentical with one another’, is
[273]
necessarily nonempty, and is indeed inseparable from the class of ‘existents’, since it is essential to any existent to be in relation to divine existence. This universality of relationship between divine existence and all else is implicit in any religiously acceptable definition of God. Hence, so is the impossibility of His nonexistence.
The crucial passage in Shaffer’s argumentation is that in which he rejects—while pretending or appearing to accept it— the idea of necessary existence as definitive of deity. To suppose that the divine freedom from existential contingency can only be a property of our statements, arising from stipulations, is to deny the very possibility of God'’s existence. There is then nothing for an a posteriori proof, or search among facts, to furnish. Only an objective immunity to existential contingency is compatible with what religion means by God (and philosophy ought to mean). Our human contingency, which shows us not to be divine, is likewise no mere matter of propositions and languages; it is rather the real potentiality of the world for doing without us, or for making our existence impossible. The very notion of a corresponding anti-God potentiality is contradictory. (There could of course be an anti-demon potentiality! ) It is clear that the author has only a hazy notion of all this, or of why religion, as he mentions, takes Satan as (at most) something that happens to exist, but God as existing necessarily. When he says, “The question remains, does there actually exist something which answers to the concept of the necessarily existent,” he either means, ‘does there happen to exist such a something’, and then he is talking contradictorily; or he means, ‘is there necessarily such a something?’ But in the latter case the question answers itself. Anything which could be necessary is necessary, by the reduction principle
[274]
of modal logic. Of course, Professor Shaffer can reject this principle. But then #has, and not what he says in this article, is his ground for the rejection of the Argument. And the principle is reasonable.
Three objections to the idea of objective necessity are offered. One is an argument from regress, the necessity of the necessary property, etc. However, the necessity of a necessity, by the reduction principle, is simply the necessity itself, and so is the possibility of the necessity. Similarly, ‘it is true that it is true’ only says, ‘it is true’. As for the contingency of the necessity, the denial of which starts the regress, it is nonsense and needs no denial. The regress seems spurious.
One could, I think, as cogently accuse Shaffer of having a regress on his hands. God ‘exists necessarily’ only means, he says, that we define Him as existing. But then this description of God is either optional or necessary. He will reply, I take it, that it can only be necessary upon some condition or other, such as that one takes the religious tradition seriously at this point. But is the tradition necessary or contingent? Could there be a religion without it, or a rational being without religion? On certain conditions, perhaps there could. But these conditions are either ultimate or they have conditions. In the latter case we have a regress. And in any case, the statement, ‘No being exists necessarily’, is not made true or false by any mere definition, nor yet by any mere fact. A fact or contingent truth cannot be precondition of an unconditional necessity. (Any being can be conditionally necessary; for instance, all past events are necessary for the present’s being as it is.) I say the statement is necessarily false, and any language in which the corresponding necessary truth cannot be expressed is defective.
[275]
Clearly, religion has never meant by the divine necessity that we define God to exist. It has meant that to the divine existence there is no truly conceivable alternative, that God owes his existence to nothing, certainly not to a lucky accident, any more than to a rational cause, or to our ways of talking. His existence is taken to be the presuppositon of any meaning, any value, any existence or nonexistence whatsoever. To allow this requirement to be logically permissible, and then ask, but in fact, is there a being fulfilling the requirement? is to contradict the requirement one has accepted; it is to subject God existentially to contingency, either to causeless chance or to a cause beyond Himself. It is to grant that if God exists He exists either by mere chance, or as caused. But so He could not exist!
Shaffer’s second objection is that it is unclear wherein the objective necessity lies. We have shown that this challenge can be met. It lies in the divine perfection, which involves an absolute correlation between possible and possible-for-God, between ‘reality’ and ‘divinely-known-reality’. Omniscience, universal creativity, self-sufficiency of existence, all the divine attributes, are inseparable from the status of God as pervasive of reality, actual and possible, hence not, in His bare existence, the actualization of this or that particular possibility. The necessary is the universal element in the possible. That element is the functioning of God, not any particular divine functioning, but divine functioning as such.
The third objection is an analogy offered as a reductio ad absurdum. Squares are necessarily four-sided, yet it does not help to say that they have the property of necessary four-sidedness. But why is this locution not helpful? One reason is that ‘contingent four-sidedness’ is not a common expression, whereas ‘con-
[276]
tingent existence’ certainly is (in traditional metaphysics). It is therefore helpful to make the point that no such existence can apply to God, though it does apply to all other individuals. Nor is ‘contingent four-sidedness’ an inevitably foolish expression, for ‘houses’ are only contingently four-sided, since they can be five-sided, six-sided, or circular. By the same token, squares cannot have contingent but only necessary four-sidedness. That this is ucterly trivial is because ‘square’ means by definition having four equal sides. But suppose we say instead that foursided figures necessarily have four angles, or four-angled figures have four sides. This is a truth independent of names and definitions. For whether or no we have a name for the figure between triangle and pentagon in the series of polygoans, the possibility of such a thing and hence of such a name is a necessary truth. And also necessary, regardless of names, is the relation between number of angles and number of sides. Just so the relation between ‘divinity’ and ‘necessary existence’ is not created by names or formulae. Worship implies such a relation, and worship is not a ‘language game’. And here the hypothetical element in the necessity that four-angled figures be four-sided (#f anything is the one it is the other) drops out. For necessary existence means, there mus# be something of the specified kind. To insist that every idea be capable of fitting the hypothetical mold is nothing but the positivistic assumption posing as harmless truism. ‘God’ is not an hypothesis which facts may or may not support but a theory of what it means to be a fact— i.e., to have a certain relation to omniscience. If this view is wrong, it is not merely factually but logically wrong. For it purports to explicate ‘fact’ as such, and to do this without conflicting with the ordinary meaning of the word. If it fails, the result is confusion, not just factual error. Assertions of
[277]
existential necessity cannot be factual errors, but only logical ones. Theism is confusion unless it is an assertion of necessity. According to positivism it is confusion whether or not it is such an assertion. But this, Anselm showed, is the issue, and not any question of mere fact.
Shaffer admits that one can put ‘has application’ into the ‘concept of the concept’ of God. But then, he says, the extensional question becomes, is there such a concept of God? And, he holds, we cannot know this without knowing that there is a God. I reply, there is no concept of the religious object, appropriate to the idea of worship taken with any strictness, which does not imply noncontingency. A concept of the concept of ‘God’ which denies this aspect is a misuse of words. So we come back to the disjunction: either there is a coherent concept of the religious object—and then the divine existence is a closed question—or there is no coherent concept of God, and then positivism, rather than atheism or theism, is correct. Our author has not shown that atheism is self-consistent. The insufficiency of the Argument (by itself alone) of which he speaks is only that it fails to refute positivism. It does, however, refute atheism. That is quite enough to make it a very important discovery. Moreover, in refuting atheism it refutes empiricism. The most burning question of philosophy, it shows, is to be settled, not by observation of facts, but by examination of ideas, meanings.
The remark ‘a defining property is not a special kind of property’ is reasonable, but irrelevant. For the necessary existence of God does not consist merely in His being defined to exist, but in His being defined, and as the One worshiped correctly defined, to exist in the unique manner of noncontingency. That this is a special way of existing the author seems at times to admit and then roundly to deny. But in denying it he is
[278]
either asserting positivism, or else failing to understand the religious idea; while in admitting it he is refuting his own argumentation.
The absurdity of the reasoning appears espedially clearly if we consider such cases of ‘defining things to exist’ as the following, ‘existing dodos’, ‘existing dragons’, and the like. Here there is no serious paradox. A dodo or dragon would exist, if at all, contingently, and hence there cannot be any logical impossibility in the nonexistence of dodos or dragons. To use a definition to make it appear otherwise is so flagrant a misuse of the defining process as to need no special further analysis. It is just the question at issue whether the religious idea is or is not the idea of a thing which would or could exist contingently. Anselm showed that it is not. Most of this author’s discussion is a series of red herrings distracting his and his reader’s attention from this central issue. His most relevant remarks are attempts to prove that all existence must, objectively regarded, be contingent. But on that assumption what Anselm showed must have been, as Findlay rightly points out, the impossibility, not the necessity, of the divine reality.
I conclude that the remarkable ingenuity of this author does not save him from missing—or not clearly seeing—the modal structure of the Anselmian discovery.
21. Heinrich Scholz and Frederic Fisch
Our last two examples are distinguished formal logicians.
I sometimes think that Heinrich Scholz was the noblest human being that I have ever known (I met him in 1949), a theologian who turned from theological studies to formal logic because—and this is characteristic of the man—he thought
[279]
that there was no other equally honest and effective way to further the clarification of theological questions. With Whitehead he is for me the most high-minded and inspiring rationalist of our century. With angelic steadiness he stuck to his ‘Platonic ideal’, as he called it, all through the Nazi period in Germany. Had he not been forced to live almost entirely without a stomach (literally) he might have accomplished far more than this disability, not to mention the spiritual and finally material disorder in his country, permitted him to do.
If any man in Germany was equipped to deal objectively and accurately with the logical structure of Anselm’s reasoning, it should have been Scholz. In one of his lectures, written in 1950-51, he does deal with the topic from the standpoint of ‘logistic logic’,48 And what happens? He falls completely into the trap set for him by Anselm’s blunder in Prosl. II, and Gaunilo’s (and most modern writers’) failure to see beyond this blunder. We are expressly told that only Prosl. II is being considered! So we are prepared for all but a few minor details in the result. I shall not reproduce his formulation of the argument, as it simply duplicates Anselm’s first or nonmodal one. With great precision Scholz then shows how, in terms of modern exact logic, existence cannot without begging the question be taken as a predicate of an individual subject; for without existence there is no subject for the predicate to inhere in. Anselm confused himself here by talking about things existing in the mind, even if not in any further sense, and this, as Scholz argues, will not do, for an ‘individual’, x, ‘merely in the mind’, is not strictly individual
48 Mashesis Universalis: Abbandlungen zur Philosopbie als, sirenger Wissenschaft (Benno Schwabe & Co., Basel, Stuttgart, 1961), pp. 62-74.
[280]
at all, but only a property or class seeking instantiation or members. With this I entirely agree. But the modal argument does not attempt to compare two individuals, one existing, the other not. It compares two hypotheses about the existential status of ‘unsurpassable individual’, and finds one contradictory: the hypothesis, ‘there is an unsurpassable individual which exists contingently’, and the hypothesis, ‘there is an unsurpassable individual which exists necessarily’. Neither hypothesis needs to be taken initially as asserted. For the point is that the first hypothesis is necessarily false, impossible, if the second is even capable of being true. To exist contingently is to be inferior to anything existing necessarily, so that if the latter is possible, the descriptive phrase ‘unsurpassable’ is in the former case used contradictorily. And if ‘necessarily existing’ is not possible, then ‘unsurpassable’ has no definite and coherent meaning. For only that which grounds all possibility, including its own, and the nonexistence of which therefore cannot be a possibility, can be strictly unsurpassable. Hence either the necessary existence or the impossibility, but not the mere nonexistence, of unsurpassability is permissibly assertible. Scholz does, as one might expect, refute, on Anselm’s behalf, the relevance of the perfect-island analogy, on the rather obvious ground that no finite thing such as an island can be the greatest conceivable, so that, whether or not there is a logically admissible idea of greatest conceivable being, there can be none of greatest conceivable island. It is hard to forgive the multitudes who have not allowed their preconvinced minds to hit upon this reply to Gaunilo’s suggestion.
Scholz also explains how Bolzmann and Frege had treated existence as a predicate, but a predicate of a property, the predicate of being nonempty or of having an instance. On this basis
[281]
also the Prosl. II argument will not work. (For we do not compare properties and argue that the better properties are the instantiated ones. In that case, in the words of the proverb, “wishes would be horses and beggars might ride.”) Yet here we must be careful. For as we saw in Part One, Sec. 18, it is inherent in almost all theistic traditions, and Scholz as former theologian must have known this, that God does not have but és His goodness or supreme worth. In other words, the property-individual distinction, symbolized, say by ‘Px’, does not in God have the same structure as in other cases. And this is the very point Anselm had discovered, in a special aspect.
Let us look into this. Some properties are instantiated, but contingently so; others are, equally contingently, uninstantiated. All ordinary properties (those of various kinds of surpassable beings) are contingent in this sense, whether or not they be instantiated. It follows that any instance of ordinary predicates exists contingently; for if its whole species or class might not have existed at all, certainly it might not have done so. But there are predicates which could not be contingently instantiated, since they are logically absurd, like ‘round and square in the same respect’. Anselm’s claim is that ‘unsurpassability’ is a third sort of predicate, which is neither contingent nor impossible, but is necessarily instantiated. Its existence alone is possible. What has Scholz against this? He does not even put the question, naturally enough, since he is considering only Prosl. II If ordinary predicates connote contingency, and if this ordinariness, i.e., this surpassability, is the very reason for the contingency, then by the converse of the same reason, unsurpassability connotes noncontingency. Moreover, that the imperfection of nondivine things, their noneternity, compositeness, dependence, lack of self-sufficiency, is the very reason for their
[282]
contingency is plausibly argued in various ways and in various places by Anselm, and by many before him, including Aristotle. But it clearly follows that even the hypothetical negation of imperfection will imply the negation of contingency. Moreover, the second negation cannot be hypothetical only. For modal status as such is necessary.
It is important to note that existence is no proper or intrinsic predicate of ordinary predicates. That two-handedness exists in my hands is not a part of what is meant by ‘two-handed’. For two and hand might have been two and hand, just as they are, though I had been born one-handed, or not born at all. It is really I, as individual, who exist as instantiating two-handedness. G. E. Moore is wholly unrefuted in his contention that, ‘this thing exists, and might not have’ makes entire good sense. However, since there is no a priori or purely conceptual definition of ‘this thing’, or of me as an individual, this kind of existence as a predicate cannot be deduced a priori, but only found empirically. *“This thing exists’, or ‘I exist’, is in a sense tautological; and yet it is contingent, for neither ‘this thing’ nor ‘I’ might have been possible with this meaning. What is a proper predicate of predicates (this too Aristotle had seen) is their modal status. Wholly limited, dependent, composite, generated beings must also be capable of failing or of having failed to exist; but by exactly the same principle, the Unsurpassable must be incapable of failing or having failed to exist. And here there is no real distinction between Unsurpassable and the Unsurpassable individual; Unsurpassability is its own ‘principle of individuation’. We will return to this point presently.
Not a suspicion of the foregoing is in Scholz’s essay. I add that, so far as I know, no philosopher in Germany since Kant
[283]
(except Anselm’s monographer Hasse) has gone back to Anselm and discovered there the modal, or proper ontological, argument, and dealt carefully with its logic. Instead, all have relied upon Gaunilo, Kant, Hegel, or the nonmodal versions of Anselm, Descartes, or Thomas Aquinas, to inform them in the matter. In this case ‘German thoroughness’ quite failed to operate. More's the pity.
Scholz’s oversight is especially remarkable in that he refers to Barth, who was not guilty of it. My explanation is that, since Barth (as Scholz notes) rejected the Argument as a philosophical one, insisting it is purely theological, and since Scholz was interested in the philosophical question, he did not bother to read Barth in detail. But this is a guess.
Scholz believed in truths applicable to all possible worlds. He called these truths metaphysical, and he expected logistic to help in their elucidation. I wish I could have put the following to him.
First, logic can admit the notion of existential necessity, at least in the form, (x) fx—(Ex) fx; properties universally instantiated cannot be uninstantiated, or in other words, logic cannot deal with a simply empty universe. The widest class cannot be empty. The case for this contention, which Scholz himself accepts, seems to me conclusively made by two recent authors, Jonathan Cohen and William Kneale.49
Second, necessary nonemptiness or instantiatedness is all that we need for necessary existence, even in the divine form_ True enough, if ‘instances’ are thought of as ‘individuals’, the ordinary meaning, then there is trouble in taking any such
49 Jonathan Cohen, The Diversity of Meanings (London, 1962), pp. 255264; Wm. and Martha Kneale, The Developmens of Logic (Oxford, 1962), pp. 706f.
[284]
instance as necessary. (Scholz wonders how one could know a priori that just one greatest individual, God, instantiates divinity as defined.) It is indeed essential to the distinction of type between property and instance that the latter be a contingent illustration. But, as we have seen (Scholz probably did not see it, such was his respect for the theological and metaphysical tradition, which at the same time he did not attempt to defend, since he did not know how to do so with modern instruments of analysis), it is logically possible to distinguish between divinity and its contingent instances, and yet to exclude polytheism, even as possibility, and affirm the necessary existence of a unique divine individual. The key is a doctrine held by Scholz himself, the doctrine that the most concrete particular entities are not enduring individuals but momentary events or states.50 The existence of an individual is the actuality of a certain sort of event-sequence. The sequence can be defined without specifying all the particular events, for we identify a person without committing ourselves to all his adventures past and future. Now this schema can be applied to deity, and I hold must be applied if antinomies are to be avoided. The property of divinity, defined in the neoclassical version of Anselm’s formula, cannot be contingently, but only necessarily, nonempty, and while any of its instantiating ‘states’ is contingent, that there are some such states, and also that any two of them are ‘genidentical’, i.e., in personal sequence, with one another, can and must be necessary. Is there anything in modern logic to forbid this?
You may say that logic finds the notion of necessary instantiation valid at most only with respect to properties that
50 Scholz, op. cit., pp. 405-410.
[285]
are universally applicable. True, but divinity is in a definite, though unique, sense strictly universal. Just as any entity is identical with itself, so is any entity, according to the meaning of theism, related to God as its creator and sole adequate knower. Relativity to the divine is as essential to existence as self-identity. To deny this is to deny not simply the existence but the logical possibility of deity. We have argued this so often that we must leave it at that here. The necessary nonemptiness of the class of divine states is related to the necessary nonemptiness of the only seemingly broader class of states in general (concrete particulars as such) in this way, that for any nondivine state there must be a divine state in which the former is known. The correlation creature-creator cannot be broken up, leaving the mere creature, or the mere creator, without making the very idea of either a meaningless or incoherent notion. Thus, if there must be entities, there must be entities divinely created and known; the only alternative being the rejection of the very question of theism as incoherent.
Had Scholz glimpsed these relationships, would he have had to confess, as he did, that while he could defend metaphysica generalis, as the theory of all possible worlds, he could find no rational approach to metaphysica specialis, or the theory of ‘last things’, including presumably some sort of answer to the theistic question?51 A ‘possible world’ as such is already God, in one of His creative potentialities, and it is and can be neither more nor less than this. True, there are some difficult and important questions concerning the relation of such ideas to the language of pure formal logic. But the locus of the questions at least begins to appear when they are seen from the neo-
51 Ibid., pp. 430-432.
[286]
classical standpoint, whereas from the classical there seems to be, as Scholz found, no promising way to connect logical to theological questions. Anselm had already come closer to the heart of the matter than anyone in the Gaunilo tradition could do. Existence is indeed, in a loose sense, an ateribute, and, moreover, in ordinary cases of special or exclusive attributes, it is a contingent ‘attribute of attributes’. But of universal, all tolerant, or nonspecial attributes, existence is always a necessary attribute, and divinity is in a sense the sum of all nonspecial or strictly universal attributes. The divine knowledge must be able to take on or acquire the form of any object whatever. Hence to assert the existence of this knowledge as such excludes nothing, is not a ‘special’ topic at all, in the relevant sense, but a, or the, strictly universal one. That some particular entities or other must exist means that some such entities must be known to God, whose infallibility means the logical impossibility of a divorce between being and being divinely known. There is only one metaphysics, general metaphysics, but there are levels of explicitness in its resules. If God seems to be left out, this can only mean that something is left obscure.
Whether or not Scholz would have been convinced by all this I am fairly sure that he would not have brushed it aside as irrelevant or unimportant. I believe he would have admitted that these matters deserve further inquiry. And he might have seen that, since the theistic view of ‘possible existent’ is simply, ‘something God might create’, it is absurd to treat the divine existence as among the things which might be and also might not be. The supposition already begs the theistic question by assuming that possibility has a meaning independent of deity. The supremacy imputed to deity includes not merely dependence of what actually exists upon God, but dependence of
[287]
everything whatever, whether actuality or possibility, and whether the latter be real or ‘merely logical’. The notion that the least item of meaning can be independent of the divine existence is a proclamation of atheism, or rather, of positivism. It is not a neutral basis of argument. Once more, there is here no room for a merely factual question; the issue is one of meaning.
It is a great pleasure to be able to end this survey with the consideration of so refreshingly original an essay as the one which Professor Frederic B. Fitch has recently published on “The Perfection of Perfection.”52 He rejects the idea of possible entities exhibiting various properties; he also rejects the idea of existence as a property of individuals, remarking that, whereas the traditional ontological argument seems first to treat God as a perfect possible entity, and then to object that if this entity lacked the attribute of actuality it would be imperfect, what he proposes to show is rather that the attribute of perfection is logically nonempty, ‘existence’ being the nonemptiness of some attribute. Fitch then argues, with remarkable independence of the traditional disputes, that the attribute of perfection must itself be taken as supremely perfect, as its own unique instance. But then, since an empty attribute, one which is merely an attribute, is insofar deficient or imperfect, and this attribute is necessarily perfect, it must also be nonempty. When I first read this, I thought it was more curious than convincing. But further reflection has considerably altered the perspective. At worst, it is wonderful to encounter a writer who has done some thinking of his own on this old topic which
52 The Monist, 47, No, 3 (1963), 466-471.
[288]
has become so surrounded with what one is (no doubt wrongly) tempted to call parrotry.
Let us grant that there are no ‘possible instances’ of predicates but only predicates whose instantiation is possible. However, does it follow that existence is really an attribute of (ordinary kinds of) attributes, with nonexistence a corresponding negative attribute of attributes, their being empty? To have an empty stomach is to have a special kind of stomach, in that the emptiness really qualifies the stomach at the time, giving it a different tension and so on. But an ‘empty attribute’ is not, in any parallel sense, qualified by its emptiness. Here I think (with some difidence, however) that we confront a characteristic tendency of a formal logician to mistake a feature of our symbolic machinery for a reality beyond that machinery. It is the universe, not green gianthood, which is empty of green giants. All discourse is about the actual, though the reference to actuality may be more or less involved and disguised. Even ‘person of normal size’ is, in itself, empty of actual people, though the world is full (too full perhaps) of such entities. Of course, taking things purely extensionally, the class of green giants is indeed empty (but is there such a class, in a nontechnical normal sense of ‘is’?) and that of ‘person of normal size’ is nonempty; but what holds for classes, in this respect, does not hold for properties. The class attribute never contains the individual instances, no matter how many the class itself contains.
However, it does not appear that this point affects Fitch’s argument. For what he wants to show is not merely, as he says, that the attribute of ‘perfection is nonempty’, but rather that it is necessarily nonempty. (This seems quite clear from the structure of his reasoning.) And of course an impossibility of
[289]
being uninstantiated is a very different thing from merely being (contingently) instantiated. Against taking contingent nonemptiness of an attribute as an intrinsic attribute of that attribute is just the contingency itself. A quality must not change in being instantiated, if it can also be uninstantiated; for if it did change then, by an adaptation of one of Kant’s arguments, it would not be that very quality but another which was instantiated. However, if the attribute is incapable of being uninstantiated, then, since being instantiated is the only status we have to recognize for it, there can be no problem of how instantiation would alter or enrich the attribute. The point is that the denial of instantiation is here absurd. We still, to be sure, must be able to distinguish the particular instances from the attribute; however, as explained in Sec. 11 of Part One, this does not entail a further distinction between the attribute and its merely being instantiated somehow. Just as the abstract must be somehow embodied in some actuality (Aristotle’s contention against platonism), so that we mean by ‘an abstraction’ something somehow housed in the concrete (if only by being thought about) and yet no particular concrete entity or set of entities is required by the abstract entity, similarly a necessarily instantiated attribute could be clearly nonidentical with its instances, and yet in its very being, as an attribute, instantiated somehow.
Fitch appatently holds that it is not a fallacy (some have called it the ‘homological fallacy’) to take perfection as its own unique instance, but rather, it is obligatory so to take it. Since I have said that Greatness or Unsurpassability (clearer, in my opinion, than Fitch’s seemingly rather circular ‘highest degree of perfection’) is the very individuality of deity, that which makes God God and no one and nothing else, can I
[290]
object to his saying that the divine essence, attribute, or individuality is the unique case of supreme perfection, ie., of itself? If the values of the variable X are taken to be individuals, then the unique X which is divinely perfect is just the existing divine perfection itself. This is a form of the Findlay paradox. It would constitute a reductio ad absurdum, so far as I can see, were it not for the possibility of taking the values of X to be, not individuals, but states. For here we have a class whose members are contingent, yet through them the class property is necessarily exemplified. That is, the class necessarily has extension (and I have argued, with Cohen, the Kneales, and others, that there must be necessarily nonempty classes, such as the class of concrete endities), and yet the class property could exist with quite other instances. So I am perhaps in agreement, insofar, with Fitch’s double contention that perfection is its own unique instance, and therefore not a mere attribute, since it must have whatever else it needs to be perfect. He seems correct also in holding that being a mere attribute is an imperfection. Certainly, if to be perfect is to surpass all (except self) a mere attribute does not do this, for it is an abstract entity, and the concrete transcends in richness any and every abstraction, its assertion being logically stronger. Moreover, a mere attribute cannot surpass itself, and the Findlay paradox, we saw, can be resolved only through the idea of self-surpassing.
Thus there is a sense in which the attribute of perfection is not itself perfect, for it is not the ‘self-surpassing surpasser of all things’. And now we see that it is not quite correct to identify the divine essence with God as an individual. The divine essence is the individuality of God, but not God as an individual. An individual can surpass itself, but not an individ-
[291]
uality. The distinction is between the ‘defining characteristic’, in Whitehead's careful phrase, of an enduring society or ordered sequence of states or unit events, and the society itself. ‘Society’ is really more exact than ‘sequence’ (and Whitehead's language is often more exact than that of those who condescend to him); for the identity of a society is not dependent upon that of its members. A society is not an extensional class of events in a certain order. What identifies it is not its members, but the characteristic shared among them which is taken as definitive of the society. The Wilson Ornithological Society has an identity which we can project far into the future (barring overwhelming catastrophe), but we do not do this by knowing members of it for future generations and centuries. Similarly, a man has, for his acquaintances or himself, an identity in the future though neither they nor he know a single one of the future events which will actualize this identity. They know that these events will share certain characteristics, such as (barring serious disease) ability to recall various events now in the past, exhibiting certain personality traits (including, from the scientific point of view, a certain unknown but reasonably-posited gene structure which will persist until death).
On my view God is the supreme form of ‘personally-ordered’ society. His defining characteristic, the divine perfection, His gene structure, as it were, is precisely His perfection, His necessary surpassing of all, including self. Each of His states will be the uniquely adequate summing up of the cosmic actuality correlated with it and of all past states of the divine society. And it will be the only society whose defining characteristic could not fail to be actualized in ever new (and greater) states.
In what sense is this defining characteristic itself ‘perfect’? It does not surpass all concrete individuals other than God,
[292]
for an abstracion—a mere attribute—cannot intelligibly be said to surpass concrete things. But we may compare the divine attribute with other attributes. In this comparison we can say that the divine attribute is at least equal to any essence in universality, for it is omnipresent, inherent in reality as such; but at the same time it surpasses all other essences in individuality, since it is the only pure essence which distinguishes an individual from all others. In it alone are universality and individuality completely harmonized or at one. It is the sole individual universal, or universal individual, in which neither the individuality nor the universality needs to be qualified or diluted. A man, for instance, is a sort of universal, in that his defining characteristic keeps receiving thousands of new exemplifications each day that he is alive; but on the one hand the individual identity of a man is a very relative and fluctuating affair (as a conscious person he seems not to exist at all in dreamless sleep, as a rational person in delirium or far-gone intoxication, and his ability to remember his past and his purposes and what he has been and intends to be is an affair of more or less, mostly less from some points of view) and on the other hand, the universality of the man is also a very restricted affair. It spans only a vanishingly small stretch of time, and at any one time is appreciably relevant to a vanishingly small part of space. Again, one might say that the cosmos is an individual, with a strictly universal defining characteristic. But here we must drastically dilute the significance of individual identity. The cosmos is not for itself identical through time. The unity of its sequences of events is not of a high order, even as compared to that of a man. Or, if we wish to say that the cosmos has personal order, we must also say (and I do) that the personally-ordered dominant
[293]
society of the cosmos is God Himself. But the cosmos is not a well-integrated individual other than and rivalling God. Apart from the dominant society which is the divine ‘soul’, we
have but the cosmic body, which is an individual only in something like the sense in which a man’s body, abstracting from his
consciousness, is so. Rather the body is a nonpersonal society of cells (themselves societies of societies), and the cosmos a nonpersonal society of societies on all sorts of levels.
So the unique intersection of unqualified universality with unqualified individuality is indeed a merit elevating the divine essence among essences generally. One may also say that the divine essence is the only fully self-explanatory one. ‘Cosmos’ by itself is a riddle; it is an ordered system, but what orders the parts? It either could cease to be, or its cessation is impossible; take it whichever way you please, you will be baffled. But the personal order of the divine life is the one case in which ‘self-ordering’ and eternal self-maintenance make clear sense. Each new divine state harmonizes itself with its predecessor and with the previous state of the cosmos, somewhat as a man harmonizes himself in each new state with his previous experience and bodily state, but with the decisive difference (among others) that the man must hope, and may easily hope in vain, that the internal and external environment will continue to make it possible for his bodily harmony to survive, whereas with God there is no such problem. First, there is simply no external environment. (Plato saw deeply into the meaning of this in the Timaeus, but his thought on this point was too simply profound for most philosophers since his time.) Second, the ‘internal environment’ is here under a radical control not rivalled even by the influence of a fully conscious man over his own nervous system. Why? No mere mystery. It follows from
[294]
the concept of the divine essence that the divine experience sums up with unique adequacy all the value of the entre actual world, and hence each thing can look to it for guidance and inspiration. Analogously, a man’s brain cells are constantly sensitive to his thoughts and feelings, which, compared to cellular experiences, are doubtless sublime and quasi-divine. Incomparably more must all things be sensitive to the unique beauty and richness of the divine experiences. Thus God needs only to continue adapting Himself to the world, with His unique adaptive skill and power. The world is then bound, for that very reason, to adapt to Him. (The leaders to whom men adjust themselves are those who at least create the illusion that they are in harmony with men and things.) The foregoing is a bare outline. But it perhaps suffices to justify Fitch’s contention that the divine essence, perfection, is itself the supreme essence, the Form elevated above all other forms as such.
Must such a form be more than a mere form, muss it be exemplified? One could fill a volume (we have been doing so) with reasons for an affirmative answer. If the form connotes the sole self-intelligible kind or order, and if the order of everything else is intelligibly derivable from its self-ordering, then that is already a reason. Again, the exact intersection of universality and individuality can hardly fail to be a logically distinctive locus in the intellectual world. Necessity of actualization is not obviously any more distinctive. All other individuals are contingent, true, but then equally all are lacking in the universal relevance of their defining characteristics. All others must be known a posteriori—of course they must, since they can only be defined a posteriori. All others have to ‘fight their way into existence’ (Peirce) against competitive possi-
[295]
bilities; but the divine essence, the capacity to harmonize with and adapt to any situation, for just that reason is noncompetitive. It could not be generated or destroyed; an alleged contingency of its existing would have no intelligible objective reference.
I conclude: Fitch is essentially right. The divine essence, compared to others, is the perfect essence; and everything about this perfection is congruent with its actualization being necessary, and incongruent with its being even possibly unactualized.
Like Scholz, our Yale logician regards existence as an attribute of attributes. In discussing this we need to recall once more our distinction between existence and actuality. The inclusive mode of reality is actuality. This is, as Fitch says, not an attribute in the same sense as other things are attributes. Yet it is a sort of attribute, and precisely not an attribute of attributes either. It is rather an attribute of everything except attributes, whereas (ordinary) attributes are attribuces of some only of the things which are not attributes. Everything except an attribute (or a mere abstraction) is an actuality (or class or system of actualities). Being actual is the same as being wholly definite (conforming to the Law of Excluded Middle as to predicates) or as being wholly particular or concrete. What has concreteness is just any and every concrete thing. Exemplified universals or properties are not literally concrete, concreteness is just not their attribute.
‘Existence’ is the fact that, among the actual things there are one or more which exemplify (‘somehow concretize’, to speak metaphorically or loosely) a specified attribute (including individually distinctive attributes, characteristics definitive of a given individual, such as being the son of So-and-so, or even as being ‘this man’). If an attribute of something is whatever
[296]
may be said about it, then one may call existence, or nonemptiness, an attribute of the essence actualized. But here ‘about’ has an extremely vague meaning, and ‘attribute’ a rather Pickwickian one. It is the exemplifying actualities which really ‘have’ the attribute, not it which has them. Actualization is the fact that something has the attribute. To, or in, the attribute itself this having (being had) is (in ordinary contingent cases) simply nothing.
To be necessarily had by some suitable actualities, however, is very different and is indeed an essence of an essence, an attribute of an attribute. Logically-guaranteed actualization is not just actualization. A man assured of posthumous fame could perhaps rest content, even without any actual fame in his own lifetime. And his satisfaction in the guarantee (so far as he believed in it) would not require that he knew any particular persons who would acclaim him or build upon foundadons he had laid. Analogously, it is nothing to the divine essence (though something to God in His full actuality) by what actualities it exists; nevertheless it is everything to the essence that, since its existence constitutes possibility as such, it could pot fail of actualization. Deprive it of this self-grounded necessity and it loses all its intelligibility and cannot perform any of its assigned roles. Here, in no Pickwickian sense, is indeed an attribute of an attribute. Without existential necessity, we may have an idol, we cannot have deity. It follows, nota bene, that sheer empiricism is antitheism built into a methodological principle. It simply begs the central religious question, and that is all it can in consistency do, so far as that question is concerned.
As to ‘maximal imperfection’, to which Fitch refers, I do not see that this has any clear meaning, unless it means the same
[297]
as ‘bare nothing’. And very likely that is truly ‘the most deficient and imperfect entity in the world’. It also is necessarily unexemplified in actuality, because it is the total denial of actuality, whereas the divine petfection is (in a certain manner) the maximal assertion of it. Classical theism makes this assertion absolute in a sense which runs into totally opaque antinomies; neoclassical theism avoids these by requiring only that the divine actuality be supreme and all-inclusive with respect to whatever is actual and that the divine potentiality account for all that is possible but not actual, by being both the ground of its possibility and that which would fully inherit its richness were it to become actual. Thus all possibility of rivalry with another is made logically impossible. And this exclusion of rivalry, as a great man saw almost nine centuries ago, is the very principle which justifies worship.
22. Conclusions
We have in Part Two considered or mentioned the responses of some forty-four philosophers to the type of argument that Anselm invented. At least twenty of these, including a dozen modern writers, appear to have known virtually nothing of the structure of the Proof as presented in Prosl. III-IV and the Reply, ot in Descartes’s Replies! About fourteen, half of them modern, have had at least a partial understanding of such a structure, in some cases (Descartes, Spinoza, Cudworth) probably not derived directly from Anselm; but of the modern thinkers discussed in Part Two, only Malcolm and Hartman pay explicit attention to the remarkable difference between Anselm’s two accounts, in Prosl. II and III, and deal with the
[298]
obviously pressing problem of their reladonship. One is moved to ask, how careless can we be?
About half of the forty-four may be said to have accepted the Proof, or at least to have seen something in it besides a mere sophistry (including Findlay, who sees it as implicitly a forceful disproof).
Of sixteen refuters of the Proof, three medieval and thirteen modern, few in their refutations so much as mention either of the two principles upon which Prosl. III and a number of later passages turn, but about which nothing is said in Prosl. II. And only Findlay shows much understanding of these principles. Yet among the critics here considered are some of the most influential: Thomas, Hume, Kant. Let us add one more, Bertrand Russell, who in his book on Leibniz and in his History of Western Phslosophy makes it plain that he follows Leibniz and others in accepting the Gaunilo legend as containing all that there was to Anselm’s reasoning.
Thus we have seventeen refutations (counting those by Esser and Lehmen) by philosophers, mostly very famous, scattered over nine centuries. With a solitary exception, these refutations take little or no account of what Anselm said in the latter two-thirds of the four-page statement of the Proof, or in most of the sixteen-page rebuttal of Gaunilo. Is there in the whole history of controversy a parallel case of casual reading of a hotly-controverted text?
One writer (Hocking) both defends and attacks the Proof: defends in relation to his own version and attacks in relation to one which he attributes to Anselm, any resemblance of either version to the historical Anselm being negligible. Hocking’s teacher Royce was similarly remote from the Pros/. III reasoning (as was Royce’s teacher Lotze, who nevertheless
[299]
went through the usual motions of refuting an unidentified or misidentified argument).
Have not philosophy departments in universities everywhere some responsibility to teach students the absurdity of these procedures? No one has to publish a refutation of Anselm; but if he does publish one, can it be less than his intellectual and scholarly duty to know and say whether or not he is dealing with the actual writing and reasoning of Anselm (or Descartes —for he too has been carelessly read, his refuters and even defenders often blandly ignoring his replies to some of the very objections still being made in our own time), rather than a feeble caricature, conveniently susceptible to refutation?
Since Malcolm challenged his colleagues to take Prosl. III duly into account, at least a half-dozen replies have already appeared. As we saw in Part One, Sec. 23, these all leave something to be desired. Perhaps the replies do succeed in showing —hardly so clearly as Findlay and others had already shown it —that classical theism cannot legitimately employ the Proof. However, since no other form of theism is considered, except vaguely and incidentally, the question, ‘Is the trouble with the Proof or with the type or types of theism it has been used to support?’ remains unclarified. The scholarly world, after a little delay, is grudgingly catching up to Anselm and Descartes; but alas, the subject now stands at a different point. (Professors, if intellectually ambitious, are busy people. Who can say whose fault it is if we cannot meet our responsibilities? Perhaps it is that of those who want education to be cheaper than cosmetics, power and luxury in automobiles, smoking, or drink.)
And what about defenders of the Argument? Do they give evidence of having read the two brief texts? As we have seen,
[300]
often not, especially in modern times. Leibniz and Hegel are striking examples. It is as though, for some reason, men recoil from Anselm’s touch, almost as much when pretending to accept his Proof as when harshly rejecting it. Could the reason possibly be that there is an intuition that this Proof is a dangerous weapon? It turns against classical theism and classical pantheism, though both have sometimes used it; it turns against unqualified empiricism so obviously that empiricists have never used it. However, if our analysis is right, there is one doctrine it does not turn against, and that is the theism which has learned to free itself from the Greek overidentification of the divine with one side of the contraries, one-many, absolute-relative, necessary-contingent, being-becoming. If this conclusion is at all correct, then Anselm, without knowing it, had in principle transcended all the older forms of theism and skepticism alike and had furnished mankind with an insttument which we are only now in a position to use correctly.
Empiricism, however, has been right all along in this, that the a priori knowledge of God is at most only an understanding of His purely abstract aspect; while all that is concrete in His reality is to be known, so far as it can be known, through observation, scientific or personal. The God of our world now, and through the geologic ages, is revealed to us partly, perhaps, through Scripture, religious tradition or ritual, partly through science, certainly not through any proof. Proofs can only show that there is a divine actuality for these more concrete or experiential means to reveal, thus giving us an infinitely bare yet balanced, seemingly consistent, and intelligible outline which all our life and aspiration can joyously fill with contingent, more particular values, meanings, and surmises.
[301]
This book has presented the following contentions. Contrary to almost universal belief, the essential principles of Anselm’s Proof are first stated not in Prosl. II, but in Prosl. III. They are: we can conceive something as such that its nonexistence 1s inconceivable; not to be conceivable as nonexistent is greater than to be thus conceivable; therefore, the concept of the greatest conceivable must be of something whose nonexistence is inconceivable.
The relevant objections are: (a) it is a mere assumption that ‘greatest conceivable’ and ‘inconceivable as nogexistent’ are themselves (consistently) conceivable; (b) a greatest conceivable must be either a greatest quantity or else a reality wholly lacking in quantity, in either case a dubious conception; (c) an absolutely greatest being must be incapable of receiving value from, or of being intrinsically related to the world, and therefore of being the loving and conscious God implied by religious attitudes; (d) the Proof appears to be an obviously illicit transition from abstract essence or idea to concrete actuality, or from the logically weaker (less definite) to the logically stronger (more defnite); (e) and consequently (Findlay paradox), if the proposed definition of divinity implies such an illicit transition, it must be judged illogical or contradictory.
All of these objections, except perhaps (a), become inapplicable if one adopts the ‘neoclassical’ version of ‘none greater (is conceivable)’, taking it to mean, ‘none greater—except itself’, ‘Exists necessarily’ then means the same as is ‘somehow actualized in any possible state of affairs’. The particular how of actualization, or the particular actuality concretizing the abstract essence formulated in the definition, is contingent. The necessity that the essence be actualized is due to the absolutely infinite
[302]
range of variability in its possible hows of actualization, or (the same thing) to this range’s being strictly coordinate to that of possible states of affairs. Since no such state would exclude God, He exists ‘no matter what'—in other words, absolutely necessarily. This neoclassical interpretation not only dissolves the Findlay paradox—for the affirmation that divinity is somehow concrete implies nothing whatever as to how it, or anything else, is concrete, and hence is a purely abstract statement—but also removes many other, and more familiar, antinomies of religious metaphysics. It implies neither a greatest possible quantity nor a reality without quantity; the God it describes can receive value from, be relative to, and know and love the world. He need not be ‘immutable, yet active’, free yet wholly necessary. Such antinomies are no longer inescapable.
Whether or not new logical difficulties may be expected to arise from the new doctrine, so that objection (a) remains perhaps in force, is the residual question, upon which historical controversies concerning Anselm (or Descartes) throw no direct or obvious light. For in these controversies the possibility of a neoclassical point of view was entirely overlooked, as it was by Anselm himself. But since the old difficulties disappear, the fair initial presumption in favor of there being some intelligible religious truth justifies a new inquiry into the possibility of a tenable theism. In particular, Kant's rejection of the other theistic proofs calls for radical re-examination.
Finally, the possibility of arriving at a correct estimate of the status and content of metaphysics in general depends in substantial degree upon a proper understanding of Anselm’s discovery that the character of being not conceivably surpassed (by another) must be one of the three: nonsensical, contradictory, or necessarily (somehow) actualized. Between a theism
[303]
which can solve the abstract-concrete paradox inherent in *necessary being’, as traditionally conceived, and a positivistic rejection of the logical possibility of any theism at all, a philosopher must, it seems, make his choice. And since every basic categorial question is connected with this one, Anselm has thus defined our task, in some respects, more sharply than anyone before him and almost everyone after him. Does he not therefore deserve at long last to be paid the minimal compliment of supposing that when he took a number of pages to state his case he needed those pages and had not said all he knew in the first two paragraphs? Granted that in this opening passage, about which so much has been said, there appears to be a notable fallacy, we have but one way to find out whether or not this initial fallacy is essential to the Proof as set forth in subsequent passages. That one way is to read—and reflect upon—those additional passages.
How painfully the fragility of human ‘reason’ is illustrated in this history! However, the hopeful side of the story is that the one right way is now open before us.