Anselm's Discovery: A Re-examination of the Ontological Proof for God's Existence Argument
by Charles Hartshorne
Part One
Necessarily Somehow Actualized: Anselm's Proof in New Perspective
- Blunder or Discovery?
- The Overestimation of Gaunilo
- What the Proof Claims to Prove
- The Definition of God: a Dilemma
- Neoclassical Resolution of the Dilemma
- Existence a Predicate?
- The Second or Strong Form of the Proof
- Malcolm and Findlay: a Fresh Start?
- The Necessary is Abstract
- In What Sense the Proof is Inconclusive
- Predicates, Individuals, and States
- The Role of Faith
- Is the Proof Platonic?
- A Theory of Modality
- Contingency and Observability
- The Proof and Logical Rules
- Anselm's Appeal to Rules
- Refutation of Some Refutations
- The Argument of Proslogium III
- Proslogium II, III, and Anselm’s Principle
- Definite Thought Is about Something
- The Proof and Pantheism
- Some Recent Criticisms of the Proof
- The Proof and the Other Theistic Arguments
1. Blunder or Discovery?
Did Anselm, in his ‘Ontological Argument’—about the year 1070—make one of the greatest intellectual discoveries of all time, or did he merely fall into an interesting blunder? Or was there, in this case as in so many others, a combination of discovery and error? We shall see reason to think that there was indeed a discovery, and a great one, but that Anselm was in part mistaken as to its nature. His critics have clarified certain aspects of the problem to which he pointed, but, alas, have also generally obscured the discovery itself. Nor have his best-known defenders understood it much better.
What was this discovery? In crude anticipatory outline, it was the following: Assuming certain ‘meaning postulates’ (to use Carnap’s helpful phrase) concerning the import of ‘God’ and certain related terms, it follows that the existence of God is a logical or analytic cruth. The meaning postulates can be rejected, but the position then taken is not atheism, as commonly understood (or agnosticism either) but positivism (as I shall use this label), the view that the divine existence is logically impossible. The sole alternative to the necessary truth
[4]
of theism is its logical impossibility. No question of contingent or empirical fact is at stake. Empirical atheism (I shall usually call it simply atheism), holding that there is no objection to the idea of God except that factual evidence concerning it happens to be lacking or negative, is genuinely refuted by ‘the Proof’ (as I shall call it for short); but equally refuted is empirical theism, holding that although atheism is logically unobjectionable the factual evidence favors theism. Unless Anselm made (as so many, but not this writer, believe he made) a mere mistake, empirical theism and empirical atheism are alike logical blunders. If belief in the divine existence even makes sense, unbelief does not, and if unbelief makes sense, belief does not. The issue between them is not one of fact or contingent truth but of meaning. One side or the other is confused. Obviously this resule, if correct, is of great importance for philosophy and religion.
The connection between the idea of God and the inconceivability of nonexistence can be brought out in many ways. Anselm hit upon some of these. For instance, to exist with nonexistence as a conceivable alternative is an inferior manner of existing, compared to existing without such alternative, and hence only the latter mode of existing is compatible with the Unsurpassability (such that “none greater can be conceived”) which defines deity. One can read a hundred standard or important philosophical works in which the Proof is discussed and scarcely find even an unclear statement of this point. Yet Anselm himself expresses it plainly more than once and implies it many times. A grosser failure of scholarship will not, I think, be found, considering that it went without effective challenge for centuries.
Anselm assumed as meaning postulate that belief is logically
[5]
possible, that his idea of deity was consistent. Granted this, he did prove the divine existence. However, should it be granted? Over a thousand years before, Carneades had given impressive reasons for regarding the idea of God as selfinconsistent. Anselm did not refute this contention. Insofar, he did not prove the divine existence.
Here we note another failure of scholarship—or of intellectual perspicacity. To argue cogently for a position is to show the falsity of at least some of its plausible theoretical compettors. A complete demonstration must disprove all of these. Because Anselm failed to accomplish this, it is concluded that he accomplished nothing. But he did accomplish something. He disproved not only atheism but empirical theism as well, and thereby reduced the central religious issue to the forced option, positivism or theism, in either case as logically true or analytic. That he did not resolve this final issue detracts no whit from the accomplishment just stated. One task at a time, if we are serious in our pretensions to sober analysis! The complete justification of theism would require a further argument, that against positivism; but the elimination of two out of four logically possible positions is sheer gain nonetheless.
Since the publication of my Logic of Perfection, in which I present several forms of ontological proof, one of the few attempts at rebuttal amounted to this: a premise from which I reason would not be self-evident to philosophers of all persuasions (i.e., positivists would reject it). Is there any piece of philosophical reasoning for an important position which does not depend upon premises similarly open to challenge from some philosophical school? It is an unwitting compliment to Anselm that he is asked to submit to criteria which perhaps no philosopher could afford to accept. An argument
[6]
is more than a mere mistake if it is cogent against some important philosophical position. It is nice to dream of an argument which would be cogent against every such position except one, but we may have to settle for less than this, or give up the attempt to apply reason to basic issues and turn over the field to Zen Buddhism, or some other form of radical intuitionism, fideism, or no less irrational antifideism.
A common method of seeking to trivialize Anselm’s claim is to hold that ‘necessarily existing’ (to use a non-Anselmian phrase) means only, ‘existing necessarily if it exists at all’. ‘Divinity exists’ is perhaps necessary if true; however, it may be false. Obviously, if a proposition p is not true, it is not necessarily true; so there is a sense in which we may sometimes say, ‘necessary if true’. But what is the meaning of the ‘if . .. then’ in such a case? Can it be: If it is contingently true that divinity exists then it is also necessarily true? No, for this is contradictory. But a proposition which could not be contingently true also could not be contingently false; hence the sole sensible meaning of, ‘if true then necessarily true’, must be, if not necessarily false, then necessarily true. And only of divinity, among distinct kinds of individuals, is there reason to admit that the sole alternative to necessary truth for the assertion of existence is necessary falsity. Moreover, the decision between necessary truth and necessary falsity cannot be empirical or factual but must be logical. So on this interpretation Anselm made a momentous discovery, that one existential question is logical or a priori, not empirical. If this is not important philosophically, what would be?
As an example of this importance, consider the classical argument against theism from the observed facts of evil. If
[7]
the issue is purely logical, no such facts can be relevant. As we shall see, David Hume conceded this relationship between the Proof and the said facts, though he rejected the Proof itself.
There is, however, another possible interpretation of, ‘if true then necessary’: ‘If it is contingently true that God exists, then it is also contingently true that He exists necessarily’. Since ‘contingently’ here refers to logical and ‘necessarily’ to ontological modality, there is no overt contradiction. The one is modality de dictu or in language, the other de re. Nevertheless, we still have not got rid of Anselm’s discovery. For (a) the antitheist has at least been forced to commit himself to a radical disconnectedness between the two kinds of modality (and there are, as we shall see, grave objections to this assumption); and also (b) Anselm himself does not employ the terms ‘necessary’ and ‘contingent’, nor is the distinction between de re and de dictu obviously applicable to his actual procedure. He compares beings whose nonexistence is not conceivable to those whose nonexistence is conceivable and holds that the former must be superior to the lacter, from which he infers that ‘not conceivably surpassable’ must connote ‘not conceivably nonexistent’. And since questions of conceivability are logical not factual, Anselm’s discovery, as I have stated it, still stands. Atheism is ruled out and only positivism survives as competitor to theism.
The effort to trivialize the great discovery has failed. It must be rejected outright, or its importance conceded. Can it reasonably be rejected outright?
This depends upon what is meant by the Proof. If the standard version, given in countless works of reference and other writings, is the essential content of the reasoning, then
[8]
I grant that outright rejection is reasonable. But the reasoning sketched above, which I and others find in Prosl. III and later passages, is in no obvious way (I believe in no legitimate way) reducible to the standard version. Therefore, the standard criticisms fail to justify rejection. Generations of philosophers have deceived themselves in this regard. What they disproved was not Anselm’s complete Argument, but the simplification and corruption which this suffered in Descartes (especially in the Meditations). This corruption was at least threefold. First, Prosl. II was taken as the basic text (known to Descartes perhaps only at second or third hand), not the incomparably supetior Prosl. III. Second, Descartes, by introducing the term ‘necessary’, raised by implication the issue between logical and real modality which Anselm had avoided. Third, by substituting terms like ‘perfect’ for the by no means equivalent conception of unsurpassability, Descartes lost one of the best features of Anselm’s terminology. The supposition that ‘perfect’ and ‘not conceivably surpassable’ are equivalent springs from an illicit conversion. To be absolutely complete or maximal, or to exhaust possible value, is certainly inconsistent with being surpassable. But is being unsurpassable necessarily inconsistent with no¢ being maximal or complete? ‘What is complete is unsurpassable’ may not be convertible. Suppose (we shall see good reason for the supposition) that possible value is in principle inexhaustible, so that the idea of exhausting it is logically absurd. It would follow that either Unsurpassability is likewise logically absurd (and positivism correct) or else Unsurpassability is logically independent of completeness. An advertising slogan quaintly suggests how this might be. “When better cars are built Buick will build them.” Insofar as this is
[9]
taken as absolutely valid for all possible circumstances, Buick cars must as such be unsurpassable, though not maximally good or ‘perfect’. Of course absolute validity could not be imagined for a slogan of this sort. Absolutes of this kind cannot qualify ordinary nondivine things. But some such absolute might very well apply to God. It might even be the sensible meaning of ‘unsurpassable’. I shall argue that it is precisely that. Clearly ‘perfect’ is a poor word to suggest this idea.
Anselm himself, it is true, in effect accepted the equivalence of completeness and unsurpassability. It remains a significant fact that his language was wiser than he knew, and that Descartes’s language was not wiser than Descartes knew.
I have said that the usual reasons do not justify the rejection of Anselm’s Proof. Could better reasons be given? To this my answer is, yes—and no. If by unsurpassable is meant complete, exhaustive of possible value, then it can be cogently argued that neither this nor any other proof is capable of establishing the doctrine. Here I join the critics: Anselm failed to furnish a valid reason supporting his own form of theism. For while atheism cannot fairly defend itself against Anselm, neither can Anselm fairly defend himself against the atheist. Each can point to a fatal flaw in the other’s position. Moreover, Descartes and Leibniz were in no better case. Nevertheless, Anselm made a great discovery, and one which can lend valuable support to theism of a different type, not only against atheism, but against Anselm'’s theism itself. This is the point which has been missed by previous commentators, of no matter what philosophical persuasion, the blind spot for which only the passing of thirty-five generations has brought a cure. Anselm’s reasoning, when its logic is sufficiently care-
[10]
fully attended to, illuminates not only the existence but the nature of God. It requires us to reconceive the import of ‘divine’.
‘Necessary’ is sometimes used to mean, upon certain conditdons which might or might not be fulfilled. This is conditional necessity, the necessity of p, given q. But there is also unconditional necessity, necessity of p, given q or not q—or as von Wright puts it, “upon tautological conditions.” With this latter meaning, it is contradictory to assume that the existence of the thing itself (or the existence of anything else) is condition of its necessary existence, taking ‘condition’ to mean a requirement capable of being unrealized. It is, from his text, quite clear that the necessity Anselm is talking about is unconditional necessity, and hence considerations which require the necessity to be conditional are not relevant. We have nothing to do here with a necessity obtaining only if the requirement that the thing exists is satisfied (whatever such ‘necessity’, surely a very odd or tenuous one, could mean).
Any criticism of the argument, to be enlightening, must be more carefully formulated.
Nor is it admissible to settle the question by fiat, that is, by declaring that existence can in no case be inferred a priori. To reply to a man who argues: whereas all other things can be conceived not to exist, divinity, for certain specified reasons, cannot be so conceived, “My dear sir, not only can all other things be conceived not to exist, but—as you simply must admit—even divinity can be so conceived,” is childish—or if one prefers, remarkably obtuse. It is mere stubbornness, the method of tenacity. We shall have to do better if we would deserve to be called philosophers. The specific
[11]
reasons just referred to must be evaluated before we could have a right to assert the universal contingency of existential assertions.
Anselm’s presentation and defence of his argument occupy slightly over twenty pages, three and one-half in the Proslogsum (Chapters II-V) and seventeen in the reply to Gaunilo (Which I shall call ‘the Reply’). Of those who claim to demonstrate that the argument is a mere sophistry, the majority appear to have read the first page or so (Chapter II), or at least a paraphrase of it in some history, but one would be hard put to it in most cases to furnish evidence that they had read more. Some have perhaps read Gaunilo’s reply at first hand, but if they know that and how Anselm replied, namely with care and (in my opinion) lucidity, as well as serene confidence, to the monk’s criticisms, they have kept this knowledge to themselves. Even when the Reply (he called it his ‘apology’) is referred to, it is usually the least essential portions which are cited. As Koyré and Barth say, the philosophical world has for the most part simply adopted Gaunilo’s point of view, including all its oversights and inability to grasp the subtleties of Anselm.
Does the reader not see a difference which is more than rthetorical between (1) ‘that which exists in reality as well as in the mind is greater than that which exists in the mind alone’, and (2) ‘that whose nonexistence cannot be conceived is greater than that whose nonexistence can be conceived’? This is the point of difference between the reasoning of Chapter II and that of Chapter III. In both cases the writing is almost as lucid as writing well can be; yet one hundred philosophical authors, many of them very famous ones, have proceeded as though it had been beyond their
[12]
capacity to notice the distinction, although it occurs practically within one page! If a difference between a modal statement (as to what can be conceived) and a simple categorical statement is not important in philosophy, what could be? Had the authors attempted to show that the reasoning of the Chapters (II, III) reduces by recognized logical principles to one and the same, then we might still disagree with them, but at least we should know that they had read the work they professed to criticize. As it is we do not know this, and I personally do not believe it. Or, if the glance of these authors did fall upon the pages in question, then, as Dewey once said of one of his critics, they must have been “suffering from never having been able to learn to read.”
Only recently, after Professor Malcolm had called the attention of English-speaking philosophers to the great difference between the logic of Prosl. II and that of Prosl. III, or between Anselm’s ‘first’ and ‘second’ ontological arguments, has one begun to meet the contention that the two arguments must stand or fall together. (See below, Sec. 23, and Malcolm in the Bibliography.) But how odd to take nine centuries even to discuss the question! And why is it that specialists in Anselm do not usually take this position? In the effort to show that an obvious textual problem, with which critics should all along have reckoned, is really no problem at all, is there nothing of the ‘method of tenacity’? Or, if you prefer, of what may be termed ‘the method of convenient ignorance’? It is so much handier to have a fictitious (conveniently simplified and drastically weakened) Anselm to refute than the real one.
If the reader is surprised at my severity, let him hear from
[13]
two distinguished scholarly specialists on Anselm (Koyré and Karl Barth):
Gaunilo seems to have understood the corrections which Saint Anselm addressed to him; at least, he did not reply. The moderns have neglected this highly instructive polemic, and this is why we have seen them repeating since Gassendi, since Kant, and down to our own time (1923 ) the same objections, the same efrors, as thase of the monk of Marmoutiers.1
Only fools and their theological and philosophical advocates, the Gaunilos, are capable of supposing that the measure of existence in general is the measure of God's existence, or of either remaining simply entangled in the dialectic of Prosl. II, or else of taking Prosl. Il to be conditioned by Prosl. II. But it is altogether otherwise: the existence of God is the measure of Existence in general, and if either of the Anselmian chapters ultimately and decisively conditions the other, it is Prosl. III which conditions Prosl. II, not vice versa.2
To repeat, anyone has the right to argue against this thesis of the logical priority of Prosl. IIl (and many subsequent passages) over Prosi. II. But to assume the contrary thesis without argument is what no one who claimed to be a scholar has ever had the right to do. Yet Gaunilo did it, and apparently nearly all followed suit. And they admired Gaunilo for having made things so easy for them!
Perhaps the most astonishing thing of all about this fantastic history is that not only did Aquinas, and many others who rejected the Proof, follow Gaunilo in taking Prosl. II as definitive, but so did Leibniz, Hegel, and many others who were sympathetic to it. Rarely does even a friend of the argument clearly appreciate the content of Chapter III; usually he
1 Koyré, (see Bibliography I), p. 225.
2 Barth (see Bibliography I), p. 178. See also pp. 147, 153.
[14]
takes as primary the formula of the previous chapter, and makes no mention of the wide difference between the two forms of reasoning. And Descartes himself seemed to get to his version of the second form only as a sort of afterthought. (Kant, I suspect, did not know this second form in Descartes.) With such minds leading the way, it is perhaps not too hard to understand how an almost rigid fashion became established. The assumption that only Prosl. II matters was ‘settled’ for centuries.
Here is a sample of what has been going on. In a friend’s office is a row of sixteen books, some rather old, some recent, dealing with the philosophy of religion. Eleven of these undertake to explain and evaluate the ontological argument, in all but one case mentioning Anselm. These eleven discussions vary greatly in quality, but one thing they have in common: they pay no attention to the chief points Anselm makes in Prosl. III-IV and repeats often in the Reply. Only the argument of Prosl. II is dealt with. One author, W. R. Sorley,3 terms that argument, ‘the sum and substance’—which recalls an old German work in which it was called ‘the nerve’ —of the Proof. One could hardly imagine more inappropriate phrases. As we shall see in Sections 19 and 20, Prosl. II is but a blundering preamble or unlucky false start in the development of the Proof. Its major premise is not even a first approximation to the nerve or substance of the eventual argument, still less its sum. Until such notions are dropped, the philosophical world will not have begun to assume the task laid upon it so long ago, to achieve a collective rational
3 W. R. Sorley, Moral Values and the ldea of God (London: Cambridge University Press, 1921).
[15]
evaluation of Anselm’s claim. It is worth adding that where Descartes is brought into the discussions above referred to he is treated in a similarly—though less obviously—truncated way, ignoring the second form of the Proof given in the Replies.
How far can the fallacies of ‘straw man’ and ‘irrelevant conclusion’ be carried and still leave a possibility of significant debate? One hundred philosophers (or should I say ten thousand?) collaborate to show that Prosl. II does not, without amplification or supplementation, provide a satisfactory account of a valid demonstration. What of that? Was it intended to do this? If so, why did its author hurry on, in the second sentence of the next chapter, to introduce two additional principles from which he reinfers the conclusion that God (necessarily) exists? And why did he repeat the major premise of Prosl. II in one short chapter only of the Reply Il but the principles of Prosl. III in the longish Reply I, V, IX? And why, whereas Reply II merely twice reiterates in perfunctory fashion the all-too-simple notion of Prosl. II, do the other three chapters of the Reply referred to offer many ingenious variations upon the more complex theme of Prosl. 11I? The intellectual energy which went into the composition of Prosl. II and Reply II cannot remotely compare with that manifested in the other four chapters. And what have the hundred philosophers to say about the principles so energetically and resourcefully explicated in Prosl. III and Reply 1, V, IX? As a rarely-broken rule, not a single word! For these authors, the chapters might as well have been in invisible ink.
On one conceivable condition the procedure might have been, not justifiable, but excusable: if Anselm had said explicitly that the reasoning of Prosl. III required, among its premises, the conclusion of that of Prosl. II. For then the two
[16]
together would prove something only if the earlier chapter proved its point. To be entirely fair, it is possible, on hasty reading, to take the first sentence of Prosl. III in this sense. However, closer examination shows that no such thing is unambiguously asserted and that if it had been Anselm would have been mistaken about the logical relations of his two arguments. It should be part of a critic’s job to check his judgment on this. Also, and in any case, the intricate, vigorous reasoning of Prosl. Il (and the portions of the Reply which continue its theme) would be interesting in its own right. Unlike that of Prosl. II, it turns upon conceptions of modal logic, and deals explicitly with two kinds of existence, contingent and necessary, rather than with existence versus nonexistence, or merely subjective versus objective (as well as subjective) existence.
When a man takes twenty pages to explain an idea (on which immense issues hang), what rule of textual criticism says that the idea must stand or fall by what is said on the first two of these pages? And who could have imagined that seven centuries might pass before the practice of acting as though there were such a rule would even be emphatically called into question? Perhaps, instead of ‘emphatically’, I should have said ‘rudely’. But can one be altogether polite when gentler admonitions are ignored and people persist in acting as though under a spell whose power has scarcely weakened in all this time, or when almost everyone talks confidently about the import of a text the actual reading of which he has obviously left to some other scholar—who has in fact no more accomplished it than he? How is one who sees all this to act? And suppose he has been at least dimly aware of the situation for forty years, and can count
[17]
on far fewer additional years to convey the message to which, until now, few have been willing to attend. Is not such a one driven to become a bit insistent and dramatic, and not always considerate of everyone’s dignity, or of his own reputadon for modesty? But indeed, it sets up no claim to extraordinary brightness or painstakingness to have escaped, by some lucky set of circumstances, from a strange fashion, almost like a collective hypnosis, of being, in respect to one subject, exceedingly careless or dull-witted. Or is it, perhaps, naively trusting in relying upon other people to read the book which all discuss, but one does not oneself trouble to read?
If this and similar charges are repeated in this commentary a number of times, let the reader not forget that the practices inveighed against have been repeated tens of thousands of times, until they have come to seem beyond the reach of criticism or change. They must be reached and they must change.
It is staggering to think of the five hundred or more works of reference and aids to students in who knows how many languages which more or less grossly misstate the history of so central a matter as the leading proposal ever made to establish a logical connection between conceptual and real existence, and also, by the most direct route, to show the rationality of the central religious belief! How long will these works continue to misinform their readers? Think of the vested interests favoring such continuance. To repeat, the situation staggers the imagination.
Perhaps the whole modern rejection of metaphysics—or the study seeking necessary truths about existence—rests upon similarly shaky foundations. It ought to be, and it is, dangetous in intellectual matters to attack theories we disagree
[18]
with only in their weakest form. In that way, we “play the confidence game upon ourselves” (Peirce)—reinforcing our prejudices over and over with never a chance of escaping from them, even when we encounter those who have thought more deeply than we.
2. The Overestimation of Gaunilo
Anselm was a great mind beside whom Gaunilo was not an intellectual giant. Barth will, I think, convince any patient reader that Gaunilo made certain blunders. Indeed, a patient reader of Anselm will also be convinced of this. Yet to many authors Gaunilo was the real hero of the ancient debate. Consider the following:
This argument . . . found an opponent worthy of Anselmus in Gaunilo, & monk of Marmoutiers in Touraine. Gaunilo emphasizes the difference between thought and being, and points out the fact that we may conceive and imagine a being, and yet that being may not exist. We have as much right to conclude from our idea of an enchanted island in the middle of the ocean that such an island actually exists. The critcism is just. 4
In this summary of Gaunilo’s objections, three points are made: the first is compatible with all that Anselm says; the second disagrees with what he says only if it is intended to apply, not to ordinary ideas alone, but to that of God as well, in which case it begs the question; the third is a loose argument by analogy, which will not stand examination. And so the historian, departing from his reasonable schol-
4 Alfred Weber, History of Philosopby, trans. Frank Thilly (New York: Scribner’s, 1896, 1925), pp. 169-70.
[19]
arly role, presumes—with the barest pretence of argument— to set at naught the chief idea of a great man. And this is a not unusual sample of the treatment Anselm’s Proslogium has received for nine centuries. It depends partly upon the reader how much longer this state of affairs is to continue. Has it ever been shown that ‘perfect island’ and ‘perfect being’ are logically equivalent, for the purpose of Anselm’s argument? Suppose, for the moment, ‘perfect’—or greatest conceivable—does, as Anselm thinks, imply the necessity of existing; it is also true, according not only to Anselm but to almost everyone, that ‘island’ implies contingency. So ‘perfect island’ implies the contradiction, ‘something both necessary and not necessary’. Conclusion: no island could conceivably be perfect in that theological sense (if there be such a sense) which entails existence. And who, indeed, except to win an argument, would ever have pretended to know what could make an island the ‘greatest conceivable’? Can it, then, be maintained that the noncommittal term ‘being’ in ‘perfect being’ connotes contingency and imperfecdon as manifestly as does ‘island’ in ‘perfect island’? Hardly, for whereas nearly everyone expects islands to be contingent and in many ways other than perfect, the great majority of theologians have thought that God must be the perfect and necessary being. So, to assume that this is as plainly absurd as ‘perfect and necessary island’ is to assume that theism itself, as usually understood, is an obvious absurdity. And thus Gaunilo’s much admired reductio ad absurdum reduces to a transparent begging of the question!
The brave historian, continuing to condescend to a dead man long considered fair game, concludes:
[20]
Indeed, the ontological argument would be conclusive, only in case the idea of God and the existence of God in the human mind were identical. If our idea of God is God himself, it is evident that this idea is the immediate and incontrovertible proof of the existence of God. But what the theologian sims to prove is not the existence of the GodIdea of Plato and Hegel, but the existence of the personal God. However that may be, we hardly know what to admire most—St. Anselmus'’s broad and profound conception, or the sagacity of his opponent who, in the seclusion of his cell, anticipates the Transcendental Dialectic of Kant.5
Here we have a fourth criticism, borrowed from Kant As we shall see, what must be inseparable from the idea of God is not His full actuality, but only His bare existence, which is a very different thing.
Weber’s admiration for Gaunilo is another echo of a dubious tradition. No one who has ever studied Anselm with care has failed to be impressed with his greatness of intellect and character; but in what sizable group of educated persons will there not be a Gaunilo or two, that is to say, a well-meaning, bright, but rather conventionally-minded fellow, who tends to greet any subtle and unusual idea with a certain commonsense (‘empirical’) skepticism, and who starts ingenious objections, more or less plausible and more or less relevant, in part understanding the view he is attacking, but in greater part distorting, simplifying, or missing the main point. Every few years, at least, a teacher has a student of this kind. If there is much more in Gaunilo than this, I am not the only reader of Anselm who is unable to see it. To compare such a clever, but essentially commonplace mind with a man of genius seems shocking. And what effort can
5 1bid.
[21]
Weber possibly have made to learn from Anselm’s text how the Saint would—nay, did—answer his critic? Clearly he made virtually no such effortt And why? Because Kant, Hume, Thomas—and Gaunilo (writing before he saw the answer) — had told him to expect nothing valuable there. And so it becomes the historian’s job, not to find meanings, but to impute their absence a priori!
Gaunilo’s substantial contributions, errors and irrelevancies aside, are: (1) his espousal of what we may today term the ‘positivistic’ view (largely anticipated by Carneades) that the initial claim with which Anselm sets out, the availability of an at least logically possible ‘idea of God’, is controversial; (2) his refutation (vaguely anticipatory of Gassendi and Kant) of Anselm’s first or preliminary formulation of his proof (in Chapter II), resting as this apparently does on a dubious conception of ‘existence’ in general; (3) his posing of the quite pertinent question whether the form of the argument, when applied to other topics than the divine existence, would not lead to absurd consequences. What Gaunilo did no# do (but is generally credited with having done) was to justify the affirmative answer to this, in itself, quite proper question. His attempted reductio ad absurdum is easily rebutted. And above all (as Barth remarks) Gaunilo nowhere gives evidence of having got into his head the ultimate form of the argument, which is reached in Chapter III, after a page or two of preliminary skirmishing in the previous chapter. In this oversight his disciples are legion. In view of this long record of haste and carelessness, it behooves all of us to proceed cautiously.
[22]
3. What the Proof Claims to Prove
First, what did Anselm claim to have accomplished by his Proof? Was it to have demonstrated the existence of God to anyone, no matter what his assumptions? But all proofs, as Anselm knew, have premises; whence the premises? A common answer is, from faith. But in that case, it secems, the proof must be unavailing, except to those who already believe, and for them it should be superfluous! And iondeed, if the proof merely derives the existence of God from faith that He exists, it is simply question-begging or circular. Was the thrill of discovery expressed in the Preface to the book concerned with so trivial a matter as that from the premise God exists, one can deduce the conclusion, He exists?
Anselm’s discovery was more subtle and complex than any of the above notions. It amounts to this: there are persons who believe in the divine existence, and these, if they understand their faith, are the only omes who do understand i¢; the others, whether they are believers lacking understanding or ‘unbelievers’, are all people who do not clearly know the meaning of ‘belief in God’. They may, if they are ‘positivists’ (to use a modern word), excuse themselves on the ground that they shrewdly suspect no one else knows the meaning either, because indeed ‘God’ or ‘divinity’ has no clear meaning. (However, the understanding believer may think that he knows better.) But if (like the ‘fool’ of the Psalms) they are atheists, that is, persons who admit that they do find a clear meaning in the central religious question, but yet deny the necessity for an affirmative answer, then, Anselm claims to have shown—and some of us
[23]
find much additional evidence that he is right—they deceive themselves. What they mean by ‘God’ cannot be what the self-understanding believer means by the term.
Alas, this is not quite the whole story. Anselm and most believers who have tried to understand their faith have met— like so many human endeavorers—with only qualified success. It can be shown from the Saint’s writings that he is partially inconsistent—or else much less clear than he means to be—in what he says about God. But it remains nonetheless true that confusion in ideas is the key difficulty, not the mere failure to consult appropriate facts, or the mere presence or absence of faith. No facts can answer an ill-defined question; especially if, as in this case, when adequately understood, it is a self-answering question. So far as it is self-answering, to answer it is a pure matter of logical insight, not of faith. Also, as Anselm in various places tries to show, logical insight can do something toward showing the propriety and logical validity of the question.
Since Anselm’s claim is that only lack of understanding of theistic belief makes possible its rejection, one might have supposed that his critics would have attempted to show that Anselm’s own understanding of belief was faulty and would have realized that to do this they must at least understand his understanding, to which a reasonable first step would have been to read what he had written with some care and thoroughness. Is this what happened? Ah, no. From Gaunilo down through a long list of more illustrious names (including Russell, who was evidently misled by Leibniz) the attitude was: on the basis of the first seven lines of Anselm’s statement of his reasoning, and our own assumed equal or superior grasp of the meaning of the religious term ‘God’, let us reach a de-
[24]
finitive conclusion as to his claim concerning this meaning.
Some readers will be saying: for philosophy, the important thing is whether or not the argument is valid or useful for unbelievers; and this it cannot be if one must first be a believer in order to understand it. But (1) to take this way out is to renounce philosophy’s most important function, which is to clarify the religious question. Science and practical common sense almost take care of themselves, but in facing life, death, and the everlasting, the first and last or strictly cosmic things, man is in great danger of fanatical faith, on the one hand, and cynical despair, on the other. He needs to think about these topics as wisely as he can, and to do this he must cooperate with others, whatever their beliefs, in mutual criticism. This free mutual criticism is the central task of philosophy. So Anselm’s challenge is indeed one which ought not to be brushed aside.
Moreover, (2) it is obvious enough that if, as Anselm holds, the central religious question is self-answering, our whole theory of knowledge must be affected by this truth. We can then no longer assume that the only self-answering questions are trivial or merely linguistic. The general issue of the possibility of metaphysics is here involved. Metaphysical questions are those which, when properly put, are self-answering—and yet are not simply logical or mathematical, at least as logic and mathematics are usually understood and delimited. (Whether they ought to be so understood and delimited is a related and important subject for inquiry.) Metaphysical questions are indeed logical questions, questions essentially about ideas; but since an idea about nothing is not an idea, unless the very idea of ‘nothing’ itself, to say that logical questions are ‘merely’ logical, and therefore
[25]
‘not about existence’, is an antimetaphysical dogma, not a self-evident truth. If Anselm is right, it can be shown to be incorrect. All logical questions (and indeed all questions) are about existence, though not all are equally directly and significantly so. Those logical or self-answering questions which most directly and significantly concern existence are at the same time ‘metaphysical’ questions. Anselm discovered one of the most important of these—does divinity exist? What Anselm proved was the contradictoriness of the negative answer. To reject the positive answer therefore amounts to rejecting the question itself. But the believer is likely to remain serenely confident in his realization that the question is inevitable—on some level of consciousness. It can only be repressed, not rejected.
Such was Anselm’s gift to faith—and not only to faith, to philosophy, which cannot evade its responsibility to deal with faith. That the gift has rarely been accepted proves little, so long as it remains equally true that it has rarely been examined in anything much like the form in which it was offered.
Let us proceed with our own examination.
4. The Definition of God: a Dilemma
The next point to be clarified is the way in which the term God is being used. For a premise of the argument is that even the ‘fool’ who denies God knows, or thinks he knows, the meaning of the word. He is to be confuted by this very meaning. But what is the meaning? Anselm replied with great simplicity: to be God is to be such that ‘none greater can be conceived’. And if you ask about the import of ‘greater’, the reply is, x is greater than y insofar as x is, and
[26]
y is not, something ‘which it is better to be than not to be’. Greater thus means superior, more excellent, more worthy of admiration and respect. Why does our Saint choose this definition? I suppose because he takes it for granted that by ‘God’ is meant the universal object of worship, and if God could have a superior, then only the ignorant or superstitious could worship Him—not all creatures, nor any reasonable creatures. They might fear or admire Him, but not rightly love Him in the unstinted way which is worship.
But now, if a whole system of thought—and Anselm intends no less—is to rest upon a single definition, expressive of faith, ought not the definition to be scrutinized with great care to ascertain if possible whether or not it is (a) really expressive of faith, (b) free from ambiguity, and (c) free from contradiction? Anselm does not seem to think of the matter in this way. Worship requires the unqualified exaltation of its object beyond all possible rivalry, and just this and only this (he thinks) is the content of the definition. How could there be any difhiculty? And if there were contradiction, faith must be simply absurd, which itself, he would think, is absurd. Yet there are difficulties: both serious danger of ambiguity or contradiction and equally serious danger of failing to express faith,
If Anselm largely overlooked these dangers, his critics— those who were in such pell-mell hurry to refute him that they could scarcely pause to read even the four pages of Prosl. II-IV in which the argument was originally presented—these fast-moving critics were not the ones to see the dangers clearly either. Rather it was Leibniz, a defender of the argument (in its Cartesian form, however, or perhaps as reported by Thomas from Prosl. 1I) who brought out the likelihood of
[27]
contradiction in ‘greatest conceivable’, and who also made the best start toward clarifying the ambiguity.
Take any conceivable number. A greater can be conceived. How do we know this is not true of ‘beings’? So much for the possibility of contradiction. The conclusion Leibniz drew was that ‘greatest’ must be taken to mean a purely qualitative, not a quantitative, maximum. Or, as he put it, only those properties can be attributed to God which (unlike quantity) admit a maximal case. The others simply do not apply. Hence the ambiguity in ‘Great’ is to be resolved by the sheer exclusion of quantity. It was, after all, commonplace among the followers of Plato, beginning with Aristotle, and going on through Philo, Plotinus, and Augustine, that deity transcends magnitude altogether. Anselm takes this for granted. God is Great in that He is “whatever it is better to be than not to be.” And better than any size, or number of parts, is being immaterial, simple, and immutable .
What is insufficiently noted here is that quantity may, after all, have a value which is not attainable without it. This consideration is simply ignored, not disproved, in Proslogium, Chapter XVIII, last three paragraphs. The Greek habit (so apparent in Plotinus) of glorying in mere unity (as though absence of contrast were not in principle as deadly as confusion!) is perfectly apparent here. The contention that all having of parts must mean corruptibility is of a piece with certain antitheistic arguments, which Anselm has to combat, against the divine necessity. All existing, his critics say, is contingent existing; similarly, all complexity is disruptible complexity. In both cases, Anselm should have said, the true rule is less simple. It depends upon the kind of thing and its appropriate manner of existing, or of having
[28]
parts. If there can be an eminent, necessary manner of existing, why not an eminent manner of having parts? And we shall see that, in one sense, God may have absolutely no parts while in another sense having more of them than any other being. The old sledgehammer methods in metaphysics need improving. There are ways and ways of having parts, as of existing. So far from its being obvious that the absence of parts is a merit, one of the normal procedures in estimating value is to compare degrees of complexity arising from parts. Beauty of all kinds is unity in variety, and the greater the variety, the greater the value of the unity. A musical chord is as unified as a symphony, but its lack of complexity, its poverty of parts, limits its value most severely. Yet, on the other hand, what could be meant by ‘greatest conceivable variety’? It must mean, all possible variety; and if this is to be unified to constitute a ‘beauty than which none greater is conceivable’, we run into trouble. ‘All possible variety’ is no definite variety at all, but confusion, full of mutual incompatibilities. For, as Leibniz put it, “not all possibles are compossible.” So ‘absolute beauty’—the great Platonic vision—is to all appearances a contradiction.
5. Neoclassical Resolution of the Dilemma
Is there any escape from the dilemma that ‘greatest conceivable quantity’ is impossible and greatest conceivable quality devoid of quantity is, for all we can see, likewise impossible? Fortunately there is, and to find it we need not abandon Anselm’s definition of deity. We need only note the following ambiguity: ‘None greater can be conceived’ may mean, ‘no greater individual’ or it may mean, ‘no
[29]
greater thing or entity’. If the latter, then not only can no other individual be conceived superior; the same individual cannot be conceived superior to itself—that is, all increase is, by definition, excluded from deity. This was indeed the old Platonic argument: the perfect, being complete or maximal in its value, could change only for the worse; but the capacity for such change being a defect, the perfect cannot change at all. This argument, and others like it, all characteristically Greek, became almost the real deity of Christian philosophers. Anselm accepts this Greek doctrine. But therewith the guarantee that the definition expresses faith is gone. We are then interpreting Greek philosophy, not faith. To worship God is indeed to exalt Him above all possible rivalry on the part of other individuals; they must not be able conceivably to surpass Him. (It is unnecessary to say, ‘must not be able to equal Him’, for ‘two equal but not conceivably surpassable beings’ leads to contradictions.) On this condition all may look up to God and worship Him. But if God surpasses God, that does not of itself imply that another individual can surpass Him.
If God is surpassable, even though only by Himself, then He can include quantity in His quality, without the quantity being that presumably impossible thing, an unsurpassable quantity. The divine quantity will be surpassable, but only by God Himself. Now we have none of the contradictions we have been worrying about. God need not be that apparent impossibility, a quality wholly independent of quantity, nor that other impossibility, an unsurpassable quantity. Nor need He actualize all possible value. Yet He can still fully deserve worship by surpassing all conceivable rivals to Himself.
[30]
This resolution of the double ambiguity inherent in the definiion of Greatness is what I call ‘neoclassical theism’ (since as a technical doctrine, it is largely a creation of the last four centuries), while Anselm’s (or Augustine’s or Philo’s) essentially Greek way of resolving it fits the label ‘classical theism’. I believe that the issues Anselm raised cannot be clarified except in terms of the contrast between these two kinds of theism.
Which form of theism best expresses faith? The classical form is certainly intensely Greek. At least, its authors were all saturated with Platonism. This alone does not show the theory to be incompatible with faith, but it suggests that it might be. And there are strong reasons for accepting the suggestion. A God unsurpassable, even by Himself, is a pure ‘absolute’, wholly unreceptive or insensitive toward the world. He is anything else than a God of love, if the word has even a glimmer of meaning in this usage. We shall see the unfortunate consequences, for all his doctrine, of Anselm’s Hellenism, his substitution of the philosophical absolute for the God of religion. Here, not in the Proof, was the essential flaw. True enough, Philo and Anselm developed Greek theology beyond Plato, Aristotle, and the Stoics, but only by going further in a direction already set. What was needed was a new direction.
Our Saint’s bold invention in his Argument had two aspects. (1) He would be the first to deduce, from the definition of the God of faith, not simply God’s existence, which would be trivial enough (for of course, faith means by the all-worshipful, all-important being an at least existing being), but to demonstrate rather his necessary and therefore unique mode of existence (Prosl. III), his inconceivability as monexistent. But (2) Anselm intended also
[31]
to deduce all the knowable attributes of God from the same definition. No philosophical theist ever had a more original and brilliant project. (Beside it—oh well, let us not praise one man by disparaging another!) The trouble is that the definition’s ambiguity—and the implicit contradiction of Anselm’s Greek way of resolving the ambiguity—must dog his steps from beginning to end. However, in that age, one had to philosophize in largely Greek fashion. Anselm did what he could. And his very formulae, interpreted by philosophers not under that compulsion, can, with few changes, be made to express a view less Greek, more Christian, and—so far as has yet been shown—more consistent.
The sources of classical theism are well known: in Anselm’s case, primarily Augustine, back of him Plotinus, back of Plotinus, Philo, and back of all, Plato. The negative theology (from which classical theism cannot consistently distinguish itself) is reasonably complete in Philo. With him the pattern was approximately fixed for a long time to come; the submission of religion to the Greek mode of mysticism lasted unbroken for over a millennium. True, Philo put into the tradition religious feelings which were not Greek, and even an idea of divine and human freedom which was rather new. But the logical pattern is still Greek. The exclusion of quantity and becoming from God is decisive. The technical issues are thereby mostly settled; the rest is mainly an emotive difference. Also Augustine had a more nearly Greek theory of the will than Philo, whose great vision of creativity in God and man paled rather than grew brighter in the Christian. A ‘process philosophy’ of universal creativity had to wait a millennium and a half for its opportunity.
[32]
In considering the relation of the Argument to the two forms of theism it is important to realize that these are not simply different or opposed; they have an area of overlap in which they entirely agree. For, as Leibniz said—and the neoclassicists can say no less—those attributes which are capable of maximization must be maximally present in God (otherwise he would not be unsurpassably Great). However, the newer form of theism asserts, and the older denies, that the attributes incapable of maximization must also be in God, provided they are capable of a form which is self-surpassable only, in which form they too describe God. God is, then, absolutely unsurpassable in whatever respects this is possible. Here there is agreement. He is (He is not) surpassable exclusively by self in whatever respects such exclusive selfsurpassability is possible—here is the sole disagreement.
We shall see that the area of ‘overlap’ spoken of suffices to justify Anselm’s claim that necessity of existence inheres in the definition of Greatness; while the admission of nonmaximal but uniquely self-surpassable properties in God removes the ground from under the strongest of the criticisms of the Argument, the objecdon that a mere concept cannot entail concrete existence.
It is also important to note that in the ‘overlap area’ of the two theisms nearly all of Anselm’s descriptions of God hold. He has indeed no parts, accidents, changes, or passions—in His unsurpassable aspects; but for all that He can have them in His self-surpassable ones. Thus we can honor our forefathers’ wisdom, while not being chained forever to the barren abstractions or negations of Greek philosophical mystcism, which they thought must apply to God through and through.
[33]
6. Existence a Predicate?
Our next point concerns the logical status of ‘exist’. Is ‘existing’, in general, part of the description of a thing, one of its ‘predicates’, or is it something quite different from a predicate? If it is a predicate, then hypothetical descriptions are incomplete. “There is a man on this island.” *“What kind of man, an existing, or a nonexisting man?” No, we do not ordinarily think thus of existing things as a special kind. Any consistently-conceived kind of thing might conceivably exist, no matter what else was true of it. Yet, as Anselm first presents his argument, it appears that he takes existing things to be a different, and superior, kind of thing: hence that which is without conceivable superior must exist. This is the form of the argument which so many, for so long, have triumphantly refuted, on the assumption that it was the essential one. The procedure has this excuse, that not only did Anselm, in his first formulation of the argument (Chapter II), seem to reason in the way just specified, but so did Descartes in his first formulation. (He escaped from it only in dealing with objections in his Replies.)
7.The Second or Strong Form of the Proof
In both Anselm and Descartes, however, the argument is given a second form which need not assume that existence is, in general, a ‘real predicate’; moreover, the existence which in the sole case of God is taken as a predicate is not simply existence in general, but a unique and superior form or manner of existing. This superior form is necessary existence, or existence without conceivable alternative of failing to exist. It may also (Prosl. XII, XXII) be termed self-
[34]
existence, or existence through self—or as Philo had put it, ‘according to essence’. In the Saint’s words: “Thou dost exist so truly that thou canst not be conceived not to exist.” Or again, “Thou art through nothing else than thyself.” In other words, ordinary existence is an inferior or comparatively ‘untrue’ form of existing, existence not through self but through another, always with the threat, the conceivable alternative, of not existing or of never having existed. The existence of all save God is precarious, accidental; and thus to exist is to be inferior to what exists essentially, in its existence beyond the reach of chance or circumstance. This is the Anselmian argument par excellence! The evidence for this is that it is reiterated in a variety of ways (see Prosl. IlI, IV; Reply I, V, IX) all incompatible with the customary interpretation that Anselm must appeal to a general principle of existing things being a special kind. The point is rather this: If God cowuld conceivably fail to exist, He must be something which, "even if it existed” would be less than “that than which none greater can be conceived”’; for we can (it is claimed) conceive of something such that & cannos be concetved mnot to exist, and to be thus is better than to be such that the nonexistence of the thing is conceivable; hence that to which no supersor is to be concesvable must be conceived as such that its nonexistence is inconceivable. He who says be concesves this but believes that it does not, or may not, exist contradicts himself; for be says that he conceives as possible what he also says mo ome can concesve as possible.
The gist of this argument—a stroke of genius if ever there was one—is in the second, third, and fourth sentences of Prosl. III, seven lucid lines. Refutations of Anselm which fail—and
[35]
how many do fail—to quote, paraphrase, or take some account of this short passage (or one of its several equivalents which occur later) may be interesting exercises; they are not, properly speaking, refutations of Anselm!
Our author of so long ago finds it evident—as does this writer—that conceivable nonexistence must be ruled out a priori with respect to Supreme Greatness (we shall sometimes simply say, ‘Greatness’). He tries in various ways (customarily ignored, one need hardly add) to communicate this intuition. A Greatest conceivable which existed merely in fact, or so that its nonexistence might conceivably have obtained instead, would not be Greatest, for thus to exist, confronted with the specter of one’s own conceivable nothingness, is the abysmal weakness which infects, for example, our own existence. It implies ultimate dispensability: existence might not have included me; I am thus absolutely derivative, owing my existence to a happy chance, or to the choice of another. Thus to exist is a defect or limitation. The superior manner of existing would be ‘without conceivable alternative’. So we see that contingency is qualitative, a genuine predicate, even though contingent exsstence, as compared to contingent nonexistence, is (in a sense, and for the purposes of this discussion) qualitatively neutral.
Among the supplementary arguments by which the Saint tries to support his insight are the following. Where nonexistence and existence are alike conceivable, a transition from the first to the second must also, he thinks, be conceivable. (Reply 1, IV, VIL) But any existence resulting from such a transition would be incompatible with the status of Greatest. It would imply a beginning in time, dependence upon causes, circumstances, and the like, all of which are deficiencies.
[36]
Also, what exists contingently is assembled from elements which previously existed, perhaps, in another combination or arrangement.
Again, Greatness, unlike all other predicates, does not exist by virtue of Being or Goodness taking on some accidental form, called Greatness. God is not merely a particulacly good thing but the principle of Good. Were He (per impossibile) not to exist, no good thing could either. And as Anselm’s teacher, Augustine, said, God is Truth itself; without Him nothing could be true. Hence, ‘without Him’ cannot express a possible state of affairs.
8. Malcolm and Findlay: a Fresh Start?
It is almost miraculous that for so many centuries no one, apparently, was really clear that Anselm had presented at least two ontological arguments, rather than one, and that Descartes had followed him in this. The present writer was perhaps the first to insist upon this distinction as close to the surface in the writings of the two authors mentioned.6 However, Flint had suggested it with respect to Descartes, and Barth had fully seen it with respect to Anselm. Recently Norman Malcolm (see Bibliography I) has also arrived at it and has presented it so skilfully as to arouse a good deal of interest. (See Sec. 23.)
A few years before Malcolm’s essay, Findlay (see Part Two, Sec. 17, and Bibliography I) set forth the following position. Anselm was indeed right in holding that deity
6 C. Hartshorne, “The Formal Validity and Real Significance of the Ontological Argument,” Philosophscal Review, 53 (1944), 225-45, esp. p. 234 n.
[37]
must be supposed (if supposed at all) to exist necessarily; for a being worthy of worship could not have the defect that its very existence was contingent or had a conceivable alternative. However, said Findlay, so far from proving the divine existence, by pointing to this requirement, Anselm had really disproved it. For modern logical analysis shows that no existence can be necessary. Concrete or actual existence cannot follow from a mere predicate or abstract definition. Hence divine perfection is impossible.
In my opinion, this criticism was more penetrating than all the classical ones. A merely contingent being would not deserve worship, for we should be revering at most a big and wonderful accident; yet, on the other hand, that a mere abstraction like ‘all-worshipful’ could necessitate a concrete actuality is a logical absurdity. I call this the Findlay paradox or dilemma. Both horns of the dilemma seem unacceptable. However, as often, it may be a trilemma: there may be a third horn. Must one choose between taking divinity as a candidate for contingent existence and supposing that ‘necessary existence’ means the necessity of a particular or concrete actuality? Anselm, so far as I can see, overlooks the dilemma and can offer us no escape from it. This was his major error, resulting as we shall see from his Neoplatonism. But his critics seem to have shared his oversight. They have never focused on the real point, the assumption that the existence of an individual must be concrete or particular and can in no case be abstract and universal.
There is another, more commonly-noted difficulty: if divine or necessary existence is a superior kind of existence, what does it have in common with the inferior forms? For if it has nothing, then we merely equivocate when we speak of God’s
[38]
existing. And what can contingent, and absolutely necessary ‘existence’ have in common? Anselm does not tell us in any satisfactory or clear way.
Is it not evident that an existence which is deducible from an abstract definition—such as that used by Anselm—must itself be abstract? An abstract idea is always neutral as to the particular concrete reality in which it is or may be actualized, and this is inherent in the very meaning of ‘abstract’ and ‘concrete’. Either, then, God’s reality is wholly abstract, or He has a particular concrete actualization which is contingent. But if the concrete reality which actualizes divinity is contingent, what can it mean to say that God’s existence is necessary? The answer, which cannot be found in Anselm or any of his best-known critics, is as follows: to exist is always, and this is the universal meaning, to be somehow actualized in a suitable concrete (and contingent) reality; but whereas in ordinary cases of existence not only is the particular concrete reality contingent, but also it is contingent that there is any concrete reality embodying the predicate. In the divine case, however, the predicate is to be thought of as inevitably actualized somehow, that is, in some suitable concrete reality. Thus contingency has two forms: either (1) both that and how the predicate is actualized or concretized are accidental; or (2) only the how is accidental, while the that is necessary. Existence in general and always means, somehow actualized in a contingent concrete form, just what form, or how actualized, never being necessary. Existence in the superior, or divine form, accordingly, means that the abstract essence (worshipfulness or the impossibility of a superior) is somehow actualized in a suitable contingent concrete form; but here only the how of actualization,
[39]
the particular concrete reality, not that there is some suitable actualization, is contingent.
Of course this ‘neoclassical’ solution implies (what the whole scholastic tradition denied) ‘accidents’ or contingent properties, as well as essental ones, in God. Yet it is not hard to see that Anselm really needs this distinction, not only for his argument, but for other purposes. Thus he says (Prosl. VI, VIII) that God must be a compassionate God, for it is better to be compassionate than not compassionate, Lie., cruel or indifferent to the creatures’ sufferings. On the other hand, since there must be nothing contingent, hence no true ‘passion’, in God, He cannot really be changed by what happens to us. Anselm’s solution was that only in His effects upon His creatures is God compassionate, not in His own reality. We receive the benefits which might be expected to flow from His caring about us and being moved to sorrow by our miseries, but really He remains quite unmoved, literally indifferent, or in an identical state of bliss whatever happens to any creature. This is a form of ‘as if’ doctrine; it is as if God loved us in an intelligible sense, but really all that can be said is that benefits flow from Him to us. We may call this the ‘benefit-machine’ view of divine love. The sun produces crops, as though it cared about our hunger and its appeasement; in reality it cares not. So with God. Is this satisfactory?
If the distinction between the necessarily ‘somehow actualized’ and the contingent ‘how’ of actualization is accepted, then the necessarily existing God can be genuinely compassionate; and the two paradoxes of a necessary yet concrete actuality and the merely ‘as if’ compassionate deity are removed at one stroke.
[40]
From the foregoing it follows that most of a millennium was allowed to go by before the logic of Anselm’s proposal was propetly explored. No one was sufficiently willing to admit that his previous views might have been radically mistaken to undertake a free exploration of Anselm’s problem. Everyone wanted, in some simple way, to accept or reject the argument, and then go on philosophizing as before in other respects. But Anselm's discovery was too fundamental to make such trivial adjustments finally tenable. What Anselm had discovered, or almost discovered, was that existence and actuality (or concreteness) are in principle distinct, and that two kinds of individuals may be conceived, those whose existence and actuality, although distinct, are both contingent and those—or that one—whose actuality but not existence is contingent, this second kind being superior to all others. According to this view, any individual, no matter how superior, exists by virtue of contingent concrete states; but whereas with you or me it is always possible that there should be no such states at all, with God, though any such state is contingent, that there is some such state is necessary.
If God has contingent states, could we not conceive a greater than He by supposing that His contingent state or states had been greater? But this would merely have been God Himself, in a greater state. That God can surpass Himself does not open up a possibility that someone not God could surpass (or equal) God; and only this possibility conflicts with worshipfulness. God may rival Himself, but that justifies no conceit in anyone else.
It may be objected that a greater than God, as we have defined Him, could still be conceived, namely, a being so great that it could not be surpassed even by itself. But this
[41]
idea leads, as we have seen, to paradox and insoluble dilemmas, to absurdity. Absurdity cannot define ‘x greater than any conceivable y’, for absurdity does not define anything.
If the foregoing reasoning is sound, Anselm made a very great discovery, though as so often happens, he only partly understood its nature. His critics saw something of what he overlooked, usually, however, at the cost of missing part of what he had discovered. They realized that nothing concrete, or in that sense actual, can be necessary. Malcolm, to be sure, denies this, saying that it is valid only if by ‘concrete’ we simply mean ‘contingent’. But this will not do. For by necessary we must mean abstract. Let us see more definitely why.
9. The Necessary Is Abstract
A necessary proposition is one whose truth is included in that of any other proposition whatever. For, were this not so, it must be possible for the other proposition to be true while the necessary proposition was false. But the hypothesis is that the proposition cannot be false under any circumstances, since what it affirms is necessary. In this sense, then, as C. I. Lewis has pointed out, a necessary proposition is entailed by any proposition.7 This has been termed a paradox. What has ‘it rained here today’ to do with ‘2 and 2 are four’?
7 For an interesting discussion of problems connected with this, see N. R. Hanson, “A Budget of Cross-Type Inferences,” The Journal of Philosophy, 58 (1961), 449-70. [I do not see that the Argument is cross-type; neither premises nor conclusion are contingent. Yet Hanson implies that he has refuted the Argument.] For C. I. Lewis's view, see A Survey of Symbolic Logic (Berkeley, 1918), pp. 336-339; also Symbolic Logic by Lewis and C. H. Langford (New York: Dover Publications, 1932, 1959), pp. 492-514, esp. 511-514.
[42]
More troublesome still, perhaps, is the consideration that many very complex questions of arithmetic are so difficult that prolonged efforts by many mathematicians have failed to furnish answers. If so simple a premise as the one quoted above entails the answer to all questions of necessary truth it is odd that we cannot simply deduce these answers. But then no one will contend that deducible consequences are always easy actually to deduce. Moreover, all deduction involves logical rules or principles, and ‘finitary’ arithmetic, at least, seems to be inherent in these rules. What follows from a premise is what a person with unrestricted logical powers can arrive at by virtue of understanding the premise. Now from a proposition, say p, a sufficiently logical person can derive p V ¢; hence he can also derive p V~p. In similar though usually less obvious ways any necessary truth, however complex, should be attainable. All one has to do is to weaken or neutralize the elements of irrelevant logical strength in the initial proposition and reason logically about the results. Two difficulties must be overcome in the process. We must be able to abstract from irrelevancies, adequately neutralize contingent alternatives, and still keep our attention upon the extremely attenuated remainder; and in some cases we must perform very complex logical operations (as in the more difficult problems in theory of numbers).
The sense of paradox connected with Lewis's principle arises mainly from this, that in an inference to the necessary, one does not really utilize the distinctive meaning of a contingent premise, but only its nondistinctive kernel. It is a mere illustration, where any illustration we are able to think clearly about would do. Note, however, that even in ordinary inferences we usually discard part of the distinctive mean-
[43]
ing of our premises. Thus the full or unweakened conclusion from ‘Socrates is a man, all men are mortal’ would not be, ‘Socrates is mortal’, but rather, ‘All men, including Socrates, are mortal’, or ‘Socrates, like all other men, is mortal’. Inference to the necessary is simply the completion of the normal weakening or attenuating process of discarding notwanted aspects of assumptions. Only rather trivial inferences fail to exhibit such weakening. For instance, in ‘x is equal to y therefore y is equal to x’ the conclusion is indeed as strong as the premises; but this is exceptional. I submit, therefore, that Lewis is essentially right, and that, just as the necessary is what all possible states of reality have in common, so necessary propositions affirm the neutral universally common element of meaning in ordinary propositions, and accordingly, inferring this element is merely the extreme or limiting case of the attenuation of commitment ordinarily involved in drawing conclusions.
We may here suggest by anticipation how the familiar objection to Anselm that necessity is a matter of propositions not of things, or is de dict» not de re, is going to be dealt with in this book. If reality is essentially creative process, with an aspect of futurity or partial indeterminacy (as Socinus, Lequier, Fechner, Bergson, Peirce, Montague, and some other ‘neoclassical’ philosophers have thought), then objective necessity is merely what all real possibilities have in common, their neutral element, which will be actualized ‘no matter what’ course the creative process may take. This neutral element is creativity in its essential or irreducible aspect, which is inseparable from the necessary aspect of deity. That Anselm had an adequate grasp of this theory of modality can
[44]
certainly not be maintained. No Neoplatonist could have had such a grasp. But his Proof leads us to it if we follow out its implications sufficiently resolutely.
The reader may well be thinking that the assertion of God'’s existence cannot be comparable to such necessities as ‘p V~p’, since, in the latter but not the former, only ‘logical constants’ appear, and since what is not a formality of logic must be empirical, only to be justiied a posteriort. If contemporary logic were a complete system, with a clearly demarcated boundary, and if it were proved that the principle of this boundary coincided with that between a priori and a posteriori, then all would be in order. But who would dare assert that such is the case? How far even set theory is purely ‘logical’ is a moot question, but more than that, the a priori theory of rational inference should, I hold, include some elucidation of two things, the ultimate or completely general theory of concrete entities as such (from which all abstract entities in some fashion derive) and the ultimate or completely general theory of knowing, experience, or awareness, including a theory of givenness, of what it is to experience something. How can the theory of reasoning be complete until these two conditions are met? Logic, as Peirce said (we might mention Husserl here), should include a general or nonpsychological Erkenntnistheorse, theory of knowledge as such. But this is impossible without a theory of reality as such, of what it is to be something knowable. When we have such a completed logic, or theory of the a priori, the idea of God will, I argue, be integral to it. For that idea can be defined with no other equipment than the generalized notion of reality as knowable. God is the X who is not conceivably surpassed, in any categorial way, by another; these categorial ways are all equiv-
[45]
alent, so that any one of them suffices to define God. Thus that God’s knowledge cannot be surpassed (by another) is enough to distinguish Him from all else. If A knows that p and B does not know that p, then A surpasses B cognitively in this respect; hence the unsurpassability of 4 by another means that A knows all there is to be known, or that ‘p entails A knows that p’. So we have defined deity in purely generic or a priori terms. For what empirical fact would be needed in order to form the idea of knowing itself? One would only have to know something, anything at all, and know that it was known, to have the idea. But then one has only to quantify universally in order to distinguish God from all else. Only God has a wnmiversal relation of knowing to things. Ordinary individuals differ from one another in that, while (if they are conscious) they know some things and are ignorant of others, it is impossible to define the distinctive line in each case between the knowledge and the ignorance without mentioning particular empirical facts. But where 4/l is known and there is no ignorance, one needs only a universal statement: No matter what the facts may be, God is the unique X who knows them all. Thus, perhaps: 1 x (z) Kxz. This formula has no special empirical content. Yet it defines deity. (Probably, to be adequate, the formula should be in modal, not extensional, logic.)
Again, take the category of causal influence. Without this idea no statement has a complete meaning, for (to give only one of the reasons) an assertion is what might be believed, and belief with no relation to action, causal influencing, is an empty word. But nothing is easier than to state a definitive characteristic of deity in terms of influencing. If A influences C and B does not influence C, then insofar A surpasses B in causal power. Hence the only way to construe causal unsurpassability
[46]
is to state as minimum that, for any x, 4 influences x. It is not hard to show that only God can meet this condition.
Since ‘God exists necessarily’ means (by Lewis’s principle) that the assertion of His existence must be knowable from any fact whatever, or the denial of any fact, it follows that, not only is it definitive of God that He knows all thete is to know, but also that He is knowable by all minds with sufficient powers of reflection to interpret their experiences adequately. He is in this sense wniversal subject and umsversal object of knowing. The same conversion can be effected with causal influence: not only is God the unique A influencing all things, but God is also the unique A influenced by all things. Thus we have already given four ways of distinguishing God in purely generic or a priori terms. And others are possible. Hence I return to the assertion that the divine existence is entailed by any fact or truth whatever, being as a priori as ‘p V~p’. And indeed, the notion of infallibility means that p entails, God knows that p, and that ~p entails, God knows that ~p, so that ‘p V~p’ and ‘God knows that p or God knows that ~p° are equivalent. They are equally nonempirical or ‘formal’, in the most basic sense. To suppose otherwise is to confuse creaturely knowledge, subject for its degree of truth to chance, circumstance, contingency, with the Creator’s knowledge, whose truth is infallibly guaranteed.
We see then that the idea of God, ‘extralogical’ as, by some current criteria, it is, need not for all that be empirical, unless in a Pickwickian or misleading sense. The same result ensues if we adopt Popper’s criterion of ‘empirical’, namely ‘conceivably falsifiable’. I have shown in a previous work that theism, by its very meaning, is rigorously unfalsifiable. (Positivistically inclined philosophers will then argue that
[47]
it can have no existential import. But as an inference this merely begs the question.)
The term ‘logic’, or ‘formal logic’, can of course be used in the narrow sense now frequently given to it, and this may be advisable. But then the word ‘empirical’ should also be defined sharply, and not be used so widely and vaguely as to coincide with ‘extralogical’. For it is at least as important to have a clear meaning for ‘empirical’ as for ‘formal’, and Popper if anyone has explicated the first of these two meanings. But as so explicated it fails to coincide with ‘extralogical’, taking ‘logical’ in the current strict sense. Moreover, until various questions concerning modal and deontological logics (and contrary-to-fact conditionals), at the very least, are much better settled than they are now, it is dogmatism if not obscurantism to identify what can be known a priori with any currently available view of formal analyticity.
I now call attention to a widely—I am tempted to say universally—neglected consequence of the foregoing. That ‘any proposition entails the necessary truths® itself entails that any proposition no matter how abstract entails such truths. Moreover, since the concrete elements all drop out in the inference to the necessary, one might as well or better start with an abstract premise. Thus instead of arguing, ‘this world exists, therefore God, the necessary being, exists’, one may take as premise, ‘something exists’. For if the conclusion is to be strictly necessary (and if not, the necessary being will be but conditionally necessary and not what the argument seeks), the conclusion must follow just as well from the more abstract premise. And so, ‘divinity exists necessarily’, if true, must follow from ‘something exists’—yes, even from ‘either something exists or nothing exists’. (This is the proper
[48]
form of the cosmological argument, which, as we now see, must be valid if the ontological is so.) But, as is logically selfevident, from an abstract proposition alone only abstract propositions can follow. For the more concrete a proposition is, the logically stronger it is, or the more it asserts; and from the logically weaker, the logically stronger cannot follow. Moreover, a proposition can scarcely be weaker than ‘something exists’. Hence ‘divinity exists’, which follows from it, as from any proposition, must be similarly abstract or weak in what it commits us to. In addition, any necessary proposition whatever that is true of God must be on the same level of abstractness. Take, then, the proposition, ‘God knows that you and I exist’. This proposition has concrete reference and so cannot be necessary; nor can it follow from the proposition, ‘God, an omniscient being, exists’. For this might be true though it were also true that you and I did not exist, and obviously what was false would not be known to be true, even—or especially—by God. Thus Anselm, had his mind been free to reflect without fear or favor on the meaning of God’s necessity, should have seen that God’s necessary existence must be very different indeed from His total concrete or factual reality. The divine necessity is #hat such abstract traits or ‘perfections’ as ‘knowing all there is to know’ must be realized in some concrete form, with respect to some concrete world of knowable things, but not necessarily in the form and with respect to the world which actually obtain.
10. In What Sense the Proof Is
Suppose that what we have said so far is correct; would it follow that God has been proved to exist? Not quite. The
[49]
reason was given by Thomas Aquinas and, more clearly, by
Leibniz: we have not shown that our definition of divine perfection is more than verbal, or as Leibniz puts it, nominal. Consider the definition: ‘necessarily-existing round-square’. To deny its existence is contradictory, for we should be saying that the necessarily true is yet false. However, to assert its existence is also contradictory, for we should be saying that what is round is in the same respect not round. The way out of the maze is to reject the proposed definition as an ill-formed expression, incapable of either truth or falsity. Anselm presents us with this question, Is his definition of God capable of describing anything thinkable? And since, as we have seen, the definition is ambiguous, meaning either ‘none greater except itself’, or ‘none greater simpliciter’, our question becomes a double one. That Anselm’s own meaning, the one last mentioned in the previous sentence, is paradoxical we have pointed out. A paradoxical concept cannot furnish the basis of a cogent argument for the truth of that concept. Whether the alternative construction of the definition can evade the paradoxes, without falling into others of its own, is a question which lies outside the present essay. That it does not face the same paradoxes we have seen.
11. Predicates, Individuals, and States
Modern logic has made a point of the distinction between ‘predicates’, which individual cases may ‘instantiate’ or ‘embody’, and the individual cases themselves. The latter ‘exist’ only in a tautological sense. To be an individual is to exist in the only sense in which an individual can exist. A predicate, in contrast, may have a sort of thinkable reality,
[50]
and yet not exist, that is, not be instantiated. We are told, accordingly, that we should not say, ‘God, or the such and such, exists’, for this is empty tautology, but only, ‘something is divine’, or, ‘there is a divine individual’. But here logicians are in danger of confusing two issues. If in the formula ‘(3x)Dx’ (‘for some x, x is divine’) the values of the variable ‘x' are taken to be individuals, in the usual sense of thing or person, then ‘for some x’ is misleading, since only one individual could be divine. In other words, ‘(3x) Dx & (3y) Dy’ strictly implies ‘(x = y)’. Moreover, the individuality of ‘x’ must here be as abstract as the predicate ‘D’. Thus the mere formula ‘(3x)Dx’ is adequate to state the meaning of the divine existence in its concrete aspect only if the values of ‘x’ are not individuals but rather events or states of individuals. The formula then means, ‘there are individual states which are divine’. This still does not say what is meant by ‘God exists’, unless we further understand that any such state x, and any other y, are necessarily ‘genidentical’ with each other. (Genetic identity is the relaton holding between diverse states of the same individual.) This relationship holds because any divine state must be all-knowing and therefore two such states can have different content only on the assumption that when one, say x, exists, the other, y, does not yet exist, and hence even the all-knowing will not know it; but when y does exist, it must fully know x, and for this and other reasons x and y must constitute successive states of one and the same all-knowing individual.
In addition, we must regard ‘(3x)Dx’ as necessarily, not merely factually, true. Not that the particular value of x is necessary, but that there must be some value or other. The class of ‘values of x’, say a,b,c, etc, cannot be empty,
[51]
though the members of this class might have been other than a,b,c, . . Absolute necessity is always the impossibility that a certain class be empty. This is the same as saying that necessity is always abstract. The class of divine states cannot be empty; moreover, every such state must be in a relation of genidentity with every other, and thus all are states of one and the same necessarily-existing individual. That the states might all have been different does not mean that another individual God might have existed but only that the same God might have existed in other concrete forms.
In a philosophy which, like that of the Buddhists or of Whitehead, takes events or states to be the ultimate units of concrete reality and regards enduring individuals as somewhat abstractly conceived sequences of events, with certain relations to one another, the ultimate values of variables are, of course, events. Moreover, in this sort of philosophy it is rather easy to see that the existence of an individual is always more abstract than the actuality of events. The same individual can exist in a variety of events, and these are never wholly determined by the mere individuality of the sequence. We can know this man as such long before we know all of his states, which indeed could not be known undl he had died. Yet we must know something concrete to identify this man. God, however, is unique in that any state with a certain abstract property of divinity will belong to just the one divine personal sequence and no other. This is the only self-individuating yet radically-abstract property. All other individuals are individuated by something specific or relatively concrete. It follows at once that they cannot exist necessarily. By the same token, God can and indeed must so exist.
So again we see what deep implications are involved in
[52]
Anselm'’s supremely great insight, that God and only God exists by necessity of His (abstract) nature. The abstractness as such escaped Anselm, but the terms of his definition make it quite plain. Nothing but extreme abstractions eater into ‘that than which none greater can be conceived’. This abstractness is the reason why the similarly abstract truth that the defined predicate is ‘somehow embodied’ is necessary and also the reason why this necessary existence cannot be the concrete or total reality of God. But little in scholastic philosophy, or in Kant’s, was favorable to a clear grasp of these two aspects of the Anselmian principle. This principle, in its implications, bursts the bonds of scholasticism. No wonder Thomas rejected the Proof.
The principle, to state it once more, is the necessary noncontingency of the divine existence. What could not be contingent is either necessary or impossible. Anselm assumed that even the foolish atheist would not take God to be impossible (or ‘meaningless’). So he inferred necessary existence. Yet Gaunilo rightly urges (it was not a new contention) the reasonableness, from an unbeliever’s point of view, of the positivistic position, as we might today call it. And it is here, as we have seen, that the real issue lies. The Saint’s attempt (Reply, VIII, IX) to show that his idea of God is produced by quite intelligible procedures, and so cannot be absurd, is insufficient to convince very many of us. Leibniz’s efforts in the same direction were also not successful. This is the main unfinished business, I suggest, in this whole matter. Could God exist, is any idea of his nature intelligible? is the great question. Atheism is no longer a valid issue; but positivism certainly is.
[53]
12. The Role of Faith
Koyré, Karl Barth, and others remind us that Anselm was arguing not with infidels but with believers. Barth admits that the Saint at times seems to forget or deny that there are or can be any true infidels, and to talk as though his argument should convince anyone. Insofar as it does not convince, it follows—even, I think, from Barth’s own exposition—that either the argument has been poorly understood, or the lack of conviction means retreat into the positivistic position, the denial that there is, strictly speaking, an idea of God or a cognitive meaning for ‘God’. To dispel doubt on this score an element of faith is perhaps needed.
Before Barth, Koyré had put into full relief the precise sense in which Anselm’s argument appeals to faith, and is inconclusive against unbelievers.8 The atheist, the man who says that we can conceive God but cannot know that He exists, will be “silenced” by the argument, sf he attends carefully to its structure. For—as Anselm discovered, to his lasting glory—to conceive divinity and 4now that we do so is logically equivalent to knowing that divinity or God exists. But, though the atheist as such is silenced, the unbeliever as such need not be. He has only to shift his ground to the positivistic position: "I do not know that I—or anyone—can conceive God (without falling into contradiction or nonsense).” The moment this happens, the theist must either himself lapse into silence or else enlarge his procedure; he must give some reason other than his faith for supposing that divinity is conceivable. Anselm’s other theistic proofs (in the Monologium) can be used for this purpose, but none too cogently. Like his
8 Koyté, op. cit., pp. 210 ff.
[54]
Proslogium proof they suffer from the classical confusion between concrete and abstract, and they have other defects.
But though the Proof does not refute unbelief of the positivistic variety, it does force the unbeliever to take the radical ground of denying coherent meaning to the religious idea. If the believer in God is sure he understands himself, and does not contradict himself or talk nonsense, he is immune to further attack. This clarifies and focuses the entire problem. And at the same time, it disposes of the claim of sheer empiricism (fairly well represented by Gaunilo) to be an adequate method of dealing with ultimate philosophical issues. Carnap and Wittgenstein are right: the basic philosophical questions are not factual, but semantic. It was Anselm who first clearly revealed this truth. For, as Anselm also knew, divinity is the first principle of metaphysics or ontology: if it is not an empirical topic, neither is anything else that is philosophically essential. When Wittgenstein says, “theology is grammar,” he is really agreeing, insofar, with Anselm. The remaining issue is, how correct is the grammar? Here is where an element of intuition, faith, insight, what you will beyond mere formal reasoning, is inescapable. But so is it in mathematics itself (consider Godel’s discovery).
Koyré speaks of the ‘fact’ that divinity is conceivable. But this is no mere fact, but a necessary truth—or else an impossibility or absurdity. Nothing can merely happen to be conceivable. Conceivability is modal, and all modal classifications (on the purely abstract level here in question) are a priori, not factual in the proper sense. What can be said is rather that the truth or validity of the claim to conceive God is controversial, as between believers and unbelievers. After all, there are controversial topics even in mathematics, in the present
[55]
state of that science; but they are not questions of ‘fact’, save in a dubiously extended meaning of that word. They are questons of logic, or of linguistic rules and their propriety. Or, they are questions of what the eternal Mind, merely as eternal, does or does not see to be valid independently of all factual circumstances. They are not questions of what happens to exist, or come into existence. They are not things ‘made’, hence not facts in the etymological, which is also the systematically most clear and useful, sense.
13. Is the Proof Platonic?
Just as the truth that Anselm’s argument ‘presupposes faith’ is made untrue by being taken too simply, as though the Proof had no secular philosophical importance, so the truth that he was a Platonist is made into an error unless construed with caution. It does not mean that if we are not Platonists we can with impunity ignore the Proof. It does not even mean that the strongest version of the Proof is one with a ‘Platonic’ setting, if that term is interpreted in the most usual way.
Classical theists are, indeed, all ‘Platonists’ in a certain sense even if, like Thomas, they are also Aristotelians. They all think that the universal principle of being can be a sort of superconcrete yet eternal reality; an ‘actus purus’, immune to change and becoming, and yet not an empty abstraction, inferior in value to concrete manifestations of, or creations by, the principle. They think that ‘goodness itself’ must be the most good thing, or the absolute measure of beauty must be the supremely beautiful thing. They commit thus a sort of ‘homological’ fallacy. The eternal and necessary principle
[56]
is, they imply, in no way abstract or inferior to concrete contingent actualities. Anselm’s unawareness of the abstract-concrete paradox was merely one form of this ‘Platonic’ attitude. (Whether or not Plato knew better we need not here inquire. It is important that many of his followers did not.)
Against this ‘Platonizing’ procedure, we may argue: if, as seems reasonable, contingency is in the step from universal to particular, or from more to less universal forms, then it is also (for this is the same) in the step from the more abstract to the more nearly concrete. But then the necessary should be looked for in the opposite direction, facing toward the most abstract!
Does this form of anti-Platonism (or moderate Platonism) invalidate the Argument? Only if one assumes the extreme neo-Platonic or Classical form of Theism as its conclusion. But then the argument is invalid anyway (Findlay paradox). Suppose, however, we take what is often termed the Aristotelian view of universals or forms, that they are not ultimately and absolutely separable from concrete instances, what then becomes of the Proof? Answer: it takes on a neoclassical form. Universals must have some embodiment (if in nothing else, in some mind thinking them). It follows that contingency cannot have its ground in the mere contrast between ‘predicates’ and ‘exemplified predicates’. For some predicates must be exemplified, or there would be nothing to talk about, whether universal or particular. The ground of contingency is rather in the distinction between specific and generic predicates, or between more and less determinate ideas. Specific predicates always involve mutual exclusiveness. They are competitive ways of specializing more general notions, alternative 'determinates’ under higher ‘determinables’; but the
[57]
bare ‘somehow specialized, somehow concretized’, when applied to the highest determinables, is not competitive with anything positive whatever, but only with the ultra-Platonic negation, ‘mere form not specialized, not concretized at all’.
It is one thing to say that each step toward particularity, each increase in the logical strength of our assertions, must involve contingency; it is another, often strangely confused with it, to say that there might be no particularizadon at all “The most general universal is somehow particularized’ is a completely general statement, affirming no definite particular whatever. The contingency of each definite step toward particularity only means that, instead of this or that step, other equally definite steps might have been taken; it does not mean that no definite step might have been taken. To affirm this last as possible is to attribute complete self-sufficiency to the abstract or universal, as I believe not even Plato did.
Applying these considerations to the Proof, we see that the ontological argument is valid if, and only if, the individuality of God is conceivable as a pure determinable, which, like all pure determinables, by the Aristotelian principle (implied by the extensional assumptions of modern logic?) must be particularized and concretized somehow. Although divinity is truly individual, incapable of coexisting with another in its class, yet (as we are about to see) its bare existence, its being ‘somehow actualized’, is quite as abstract (in the here relevant sense), quite as nonspecific or noncompetitive, as ‘reality as such’. It has the same absolutely infinite range of variations, the same unrestricted flexibility. (This does not make the actuality of deity a flabby characterless thing; for the actuality is the contingent bow of em-
[58]
bodiment, not the bare necessary truth ‘embodied somehow’. While God, merely by existing, is not required to forbid this or that nondivine form of existence, by free contingent decision He can do so.) Deity is individual but, according to the neoclassical view, precisely the individual whose definitive functions are strictly universal—such as, knowing everything, influencing and being influenced by everything, related actually to all actual things, potentially to all possible things, coextensive thus with modality itself, and so bound to be instanced no matter what more special abstractions are or are not instanced. None of Anselm’s critics, nor yet his most recent defender Malcolm, seems to have considered this notion of modal coextensiveness as intrinsic to perfection.
The equivalence of modal coextensiveness with Unsurpassability (by others) is manifest: any supposedly rival actuality must be included in the actuality of Greatness (otherwise God would be but a constituent of the total actuality), and what any other individual could be is only a fragment of all that the Unsurpassable could and would be if the other individual were actual. Thus the necessary aspect of deity is simply the ultimate determinable as bound to be embodied in some concrete determinate form. To affirm this inevitability of embodiment as conclusion of the Proof is not to exceed the logical strength of the premises, for the conclusion (and, we shall argue, the whole of metaphysics and pure logic) is already implicit in the minor premise, ‘God is conceivable as not conceivably surpassed’, alone. ‘Somehow-actualized Greatness’ is quite as abstract or ‘weak’ as simple ‘Greatness’ —the concreteness being here wholly in the how of actualization. It is different with ‘somehow-actualized black swans’. For here the kind of thing is itself restrictive or competitive,
[59]
an hypothetical or threatened limitation upon the realization of other possibilities; ‘somehow actualized’ simply affirms the realizaton of this threat. But Greatness does not, of itself, threaten any possibility whatsoever. On the contrary, it is to be thought of as the very being of possibility in general, the power which any possibility would express if actualized.
The reader may be wondering how an ultimate determinable could issue in determinations: there is the old query, can a thing give what it lacks? Here we have exactly the mistake of Platonism (in the bad sense), the notion that all beauty of beautiful concrete things must preexist, or eternally exist, in the Principle or Source of beauty. The notion that creation consists in the mere parceling out of an already completed value is exactly what philosophies somewhat lacking in religious vision might be expected to have. It is the denial of any intelligible creativity, divine or creaturely. To be creative is to add positive determinations to reality, to enrich the totality of things by new values. The ultimate determinable is the supreme creativity; it produces the values, it does not simply pass them out, or down, to the creatures from its Own prior possession.
We seem to have shown that in the neoclassical, or ‘moderately Platonic’ use, the Argument can escape the traditional charges of formal fallacy. True, the acceptance or rejection of either the minor premise (that deity is conceivable) or the Anselmian Principle (existential contingency is a defect, because necessary existence is conceivable and better) will, for each individual, involve, directly or indirectly, an element of intuitive judgment. Is there any argument which will not somewhere require such judgment? Not for nothing is ‘counterintuitive’ used by logicians. Unfortunately, on some
[60]
topics the available supply of intuitive comprehension is today not great. Perhaps it never can be great, but it might be increased.
14. A Theory of Modality
That ordinary predicates neither exist necessarily nor necessarily fail to exist is inherent in their meanings. For they describe a conceivable sort of world which excludes other sorts likewise conceivable, and to do this belongs to their very function as predicates of the usual type. Similarly, self-contradictory predicates, by their mere meanings, necessarily do not exist. In these two cases modal status inheres in the predicate itself. What then is incongruous in there being a third form of predication which, by its very meaning, neither (1) excludes existence nor (2) is neutral to it (existing if this possibility, but not if that possibility, is actualized), but rather (3) requires existence (exists no matter what possibility is actualized)? No impartial person can, | think, deny that there is a certain completeness about this view which has an intellectual appeal. Modal status, it says, is always a priori or logical; but, of the three forms of modality, contingency alone makes existence a question of extralogical facts. The others make it an a priori necessity, positive or negative.
It might be thought that there are four forms of modal status: contingent nonexistence, contingent existence, necessary nonexistence, necessary existence. But the distincton between positive and negative contingency is not, like that between positive and negative necessity, an affair of meaning alone. It belongs thus to a different logical level. Con-
[61]
tradiction requires rejection, and the necessary is that whose contradictory is self-contradictory; thus only the distinction between ‘contingently exists’ and ‘contingently does not exist’ is left open by meaning, or ‘the rules of our language’. It is a radically different sort of distinction, therefore. This corresponds to the logic of the notion of possibility: a predicate which describes no possibility cannot exist, one which describes a factor in every possibility must exist, and only a predicate which describes a factor in some possibilities, but not in others, may or may not exist. These are the relations to possibility as such which a predicate can have. Hence there are three and only three modalities. And all of them are a priori. Only the nonmodal distinction between contingently existent and contingently nonexistent is extralogical or merely factual. It is ‘existence’ in this sense only which is ‘not a predicate’, and it is not a predicate precisely because it is this sort of existence. Also the logical structure which explains its extralogical status equally explains the intralogical status of necessary existence.
Why has this straightforward analysis not long ago become commonplace? One reason seems to be this: ‘possibilities’ are taken to include purely negative ones, like ‘the nonexistence of birds’. So, analogously, ‘the nonexistence of deity’. And then even ‘deity exists’ excludes a possibility, is restrictive or competitive. What is overlooked here is that a negative possibility, if genuine, is not a single definite sort of possibility, but rather an infinite system of possibilities which differ positively as well as negatively among themselves. The ‘nonexistence of birds’ would mean that every part of the world of relevant size and condition was occupied by something other than a feathered, warmblooded vertebrate. The varieties of cases here are unfathomable. This consid-
[62]
eration, that negative possibilities are only an aspect of positive ones taken wholesale, gives us a criterion for distinguishing genuine from merely verbal negations. The nonexistence of a predicate, H, either does or does not imply the disjunction of various positive forms of possibility any one of which would exclude the existence of H. If for H we put divine perfection, no such forms can be specified. It is the same with all comparably abstract properties, e.g., ‘something particular’. No positive possibility is excluded by the bare ‘existence of particulars’, their nature being not further specified. Universals are not crowded out of reality by there being particulars; on the contrary it is only in particulars of some sort, for instance, particular minds, that universals can be met at all. And there being particulars of a kind not further specified sets no restrictions upon the truth or untruth of more specifically defined sorts of particulars. This absence of exclusive particularization in ‘particularization’ itself, taken merely as such, is entirely matched by the absolutely infinite tolerance of ‘divine perfection exists somehow, in some particular concrete form’. Any imaginable being besides God is quite free to exist, so far as this statement is concerned. To say otherwise is to say that something could exist which God could not know existed, and this is to make the idea of His cognitive perfection contradictory.
15. Contingency and Observability
If it be thought arbitrary to take positive possibilities as forming the entirety of possibility, then I appeal to Popper’s invaluable if simple lesson, that the useful meaning of ‘contingent’ or ‘empirical’ is, “capable of being contradicted
[63]
by some, but not all, conceivable observation statements, these being always positive—'here is such and such’.” True, Popper, for his purpose of clarifying the role of science, limits falsi- fying observation to the sort that human beings could make, and thus he excludes such a statement as, ‘Here is a uni- verse everywhere otherwise occupied than by a four-legged feathered creature’ (ie., there are no such creatures). A human being could not observe this, and so it could not falsify, ‘There is somewhere a four-legged feathered crea- ture’. But we can conceive a superhuman mind doing this. It is very different with the statement, ‘Here is a universe everywhere occupied otherwise than by a divine being’. For a divine being could occupy no place and perform no func- tion that anything else could occupy or perform. Hence one would have to observe the bare absence of deity, and this no sort of mind could conceivably do. (All observation is of a presence, not of a mere absence.) If, for instance, one observed a disorderly or ‘badly-ordered’ universe, how, with- out being divine, could one know that a certain degree of dis- order or of evil was more than was compatible with divine rule of a universe of free creatures? Or that creatures should or could wholly lack freedom? (And the disorder and evil must be relative only, or no mind could exist to know it.) Thus the Proof is concerned with a statement which is not, even in an extended Popperian sense, empirical or contingent. Anselm’s *Thou canst not be conceived not to exist’ is correct, if to conceive a negative fact is to conceive a positive one incom- patible with a given hypothesis.
How could any conceivable experience exemplify ‘God does not exist’? God's own experience certainly could not exemplify it. Could any other experience? No, for from the
[64]
very definidon of Greatness, as implying existential self- sufficiency, it follows that God must be able to exist no matter what appears in any experience other than His own. Could such an appearance contradict the existence of God, God would owe His existence, should He exist, to the ab- sence of the hypothetical experience containing this appear- ance. This contradicts His independence of existence from all special conditions. So we have rigorously shown that the nonexistence of God has no conceivable experiential meaning, in terms of divine or nondivine experience. Either then ‘conceivability’ has no essential relation whatever to the testimony of possible experience, or the ‘nonexistence of God’ is inconceivable.
Has the ‘existence of God’ an experiential meaning? (‘Ex- periential’ is not the same as ‘empirical’: the latter con- notes, ‘compatible with some, but not 4/, conceivable experi- ences’; the former, ‘confirmed or manifested at least by some, perhaps by all, conceivable experiences’.) Clearly, God could experience His own existence, if He could do or be anything. And also, in principle, nondivine experience could imply, and hence know, God. For, just as it follows from the idea of Greatness that its existence must depend upon nothing else, so it follows also that the existence of other things must de- pend upon Greatness; and, if all things require God's existence for their own, the occurrence of anything whatever—and so of any nondivine experience—implies the divine existence. Ac- cordingly, whether or not in such experiences there is con- sciousness that God exists depends only upon the level of self- understanding of the experiences. The logical basis for the con- sciousness is given in the definition of Greatness. So, while the nonexistence of God cannot be experientally significant or
[65]
genuinely conceivable, his existence very well can be. Hence, the impossibility of significantly denying ‘Greatness exists’ appears not to be nullified by any comparable difficulty in asserting it. And of course many mystics claim to experience God. Since their claim is compatible with the conception of God, and the claim of falsification is not thus compatible, Anselm’s Principle seems to be vindicated. Greatness is conceivable only as existent, by the very criteria which allow us to conceive either the existence or the nonexistence of any island, dollar, devil, you please.
16. The Proof and Logical Rules
The basic pattern of attacks upon the Argument is this: however exceptional God may be, He cannot be an exception to the ultimate rules of language or of meaning. But this is also the pattern of many positivistic attacks, not simply on the Argument, but on the idea of God itself. Thus God could not be infallible, for to perceive or know is to experience things in a perspective which is bound to put some things in clearer light than others: He could not be both perfect and ‘living’, for to live is to have an environment whose features set limits to the actions of living organisms. In short, God cannot be a sheer exception to the rules governing the meaning of the terms applied to Him. Just so, the antiAnselmian is sure that God could not be an exception to the rule governing ‘exist’—that it is always an arbitrary or contingent determination of the disjunction: to exist or not to exist. So sure is he of this that he does not even need to read and rarely does read—or at least remember—even the four pages of Prosl. II-IV in order to know how wrong they must be. And what is the critic so sure of? Unwittingly or wittingly,
[66]
of this: God is logically impossible. For His very definition requires Him to be an exception to the alleged rule. That this discredits the idea is a conclusion capable of plausible defence—if the critics would but defend it and not confuse the issue by pretence of neutrality concerning God’s logical possibility.
God is by definition an infinite exception! Undl this is seen and admitted by all parties, debate concerning theism is essentially in the dark, and about no one knows what.
But the subtlety of the matter lies of course in this: that if God’s existence simply violates rules, there can be no rational approach to it whatever. Indeed, either there is a sense in which ‘God’ fulfills, rather than violates, the normal logical rules, or the positivist is right and the idea is nonsense. The function of the idea of God must be to ‘fulfill, not destroy’, to constitute the meaning or rationale of things, to establish universal sense, not universal nonsense. Either there must be a more truly universal Rule, which illuminates the lesser rules, and explains why they not only admit but demand an Exception, or else the idea of the Exception is against all logic.
Here again is where classical theism fails. It does not exhibit such a Rule, in appropriate relations to the lesser rules. For instance, it does not clearly say why and how existence is generally but not in the supreme case contingent; or why and how life is generally but not in the supreme case dependent upon an external environment. Yet a rule which does this is nonetheless conceivable. We have indicated it in various ways. To exist competitively is indeed to exist contingently; hence to exist noncontingently is to exist noncompetitively. Again, to exist with but limited capacity
[67]
to respond or ‘adjust’ to things is to be able to achieve clarity and coherence only by responding more to some things and less to others in a graded perspective, and this is precisely what it means to be ‘localized’. But localized beings are competitive; where one is another cannot be. By the same token, to exist with infinite capacity to respond is to have no external environment and no limiting perspective or competitive locus. The rules which set the limitations also prescribe the rule for the absence of limitation. Thus, for instance, to be completely ‘adaptable’ is to be able to exist in any state of affairs; hence to exist wholly noncompetitively or all-tolerantly, i.e. (by the very rule which explains ordinary contingency), noncontingently.
In classical theism one cannot proceed in this way. Not infinite or perfect power of adaptation belongs to God, but simply the absence of any need or capacity to adapt or respond to things (‘impassibility’—an Anselmian doctrine); not perfect love or compassion, but only an ‘as if’ simulation of this, so far as effects upon others are concerned. Here the exception to the rule is not, as in neoclassical theism, built into the rules—it is a sheer violation. Rather than positing infinite adaptability in God, one says rather that God has no need to adapt at all and that what world exists simply makes no difference to His actual state. Thus, one by one, the concepts in terms of which something might be said about God are discarded. He is not the Exceptional fulfillment of the concepts, rather He is their exceptional—indeed absolute—unfulfillment. That this is the way to lose the game of talking with meaning about deity appears especially in this, that one rule which is as surely logical as any, the rule that the necessary must be abstract, empty, entailed by all
[68]
concrete or actual details of existence, but entailing none, is set aside in classical theism by a mere fiat. God must be actual, not a mere abstraction, and yet also wholly necessary. For this exception no rule is possible, and by it no lesser rule is illuminated. This is the real flaw in the Argument as used by nearly all who have employed it.
However, this flaw was not accurately seen by any of the critics either! Or, at least, it was not seen (by Findlay, for example) in conjunction with the relevant consideration that perhaps God need not be taken as simply identical with ‘necessary being’; for while His individuality or essence must be thought of as necessary, how the essence is actualized in actual states of experience or consciousness may yet be contingent. This conjunction of ideas—which preserves as without exception the grand rule that the actual or nonabstract is contingent—was overlooked by nearly all parties alike. Why? Because their approach was unwittingly dogmatic. They had a rigid element of belief that was not subject to revision: e.g., that no form of existence could be necessary, or that no aspect of the divine existence could be contingent; or that theism could not be rationally known as true in any form; or that it was known (or should be believed) to be true in some classical form (including classical pantheism or its equivocal derivative, Hegelianism). The parties scarcely open to conviction on one or other of these points divided the field. They still do largely divide it. But the rigidity seems to be giving way. This is the promise of the present situation.
The promise will be fulfilled only if we become aware of the devices to which tenacity resorts when its beliefs are threatened. We shall be told (probably by some nontheists
[69]
as well as theists) that neoclassical theism is ‘anthropomorphic’. As though infinite adaptability to all possible worlds were not separated by a literally infinite gulf from human adaptability to some things in this world! To put man (and any other mere creature) in his proper place there is no need to allow him to usurp and exhaust categories like adaptation, passibility, change; this is just the way to give man undue importance. For then all that is left as God's province is the emptiness of eternal necessity, void of true freedom and all concreteness! The negative theology was not nearly so modest as it appeared to be. It put God behind the beyond, where He could do nothing and be nothing intelligible to us. All the more freedom for us to attend to our affairs, without relating them positively to God!
If Anselm’s formula, ‘God is whatever it is better to be than not to be’, had been strictly conformed to the negative theology, it would have run, ‘God is not what it is worse to be than not to be’. Would this have improved it? I submit: we do not worship God because of the defects which He does not have. We worship Him for His positive and allencompassing love and beauty.
The use of the Argument by proponents of the negative or classical theology has not been a grand success. The proposal to empiricists that they should use the Argument has caused them to be unable to read. Perhaps another possible use is worth looking into, that by neoclassical theists, those who admit knowable and positive—though extremely abstract—divine properties, as well as largely unknowable concrete divine actualities.
[70]
17. Anselm’s Appeal to Rules
Suppose we can without contradiction conceive that there is a unique x which is unsurpassable, and suppose we can also without contradiction conceive the negation of this. Then either supposition is the notion of an asbsolutely inexplicable brute fact. Ordinary facts may in a relative sense be inexplicable, but in a relative sense at least they are always explicable. They have causes which at least partly explain them. Even if determinism is, as I confidently believe, a false doctrine, still, every existing thing of ordinary kinds has come into existence thanks to causes which made the emergence of some such thing when and where it did, if not inevitable, at least more or less probable. But to be unsurpassable a being must exist thanks to no cause whatever, and without ever coming into existence or being capable of ceasing to exist. An uncaused being such that it might have failed to exist, yet incapable of coming into or going out of existence could only be, in the most absolute sense, inexplicable, a wholly and simply irrational contingent fact. No antecedent fact would illuminate it, and by hypothesis no abstract principle of logical necessity would either. It would just be so. Unlike ordinary ideas of chance, such as those of Peirce and others, which always set limits to the inexplicable aspects of things, we would here have an existence through and through pure chance, in the strict logical sense of having no aspect derivable either from necessity (as in pure mathematics) or from antecedent fact or cause. All else, then, would exist as at least partly explicable, but the inexplicability of this existence would be infinite and total. God exists, He might not have; He does not exist, He might
[71]
have—whichever is true, there can (according to the thoroughgoing empiricist) be no aspect whatever of reason in its being true. It just is.
Thus, whereas the usual view is that Anselm is breaking all rules to establish his case, it is from some points of view his critics who break rules, among them the notion that a contingently existing thing should have some explanation for its existence. Everywhere else mere chance, as entire account of a being’s existence, is ruled out; here it is admitted! Anselm commits no such violation of ordinary principles. He rejects the contingency hypothesis as applicable to this case, holding that in the existence of God there is neither cause nor chance, but the impossibility, inconceivability, of an alternative. God is states the only possible truth about the divine existence; hence there can be no question of why, or how it came about, that God does, rather than does not, exist—except in much the sense in which we may ask why five and seven are twelve, rather than some other number. The sole and sufficient reason is that it must be so; there being no possible alternative, the nonrealization of any such alternative calls for no explanation. How different with contingent facts!
In final metaphysical analysis: that acts occur for which there is no complete causal derivation is not ‘irrational’ if the essential function of reason is to explicate and serve cretivity (rather than to foresee its results); deity, however, cannot be conceived as a mere product of creativity, but only as its supreme and indispensable aspect, whose flexibility is coincident with possibility itself, and is thus on both sides of every contingent alternative, hence itself not contingent but necessary.
[72]
Anselm’s critics, we see, have not been the only philosophers who appeal to the opponent to acknowledge rules and abide by them. Anselm himself made such an appeal. Thus it is a rule that contingent things (generally admitted to be such) go into and out of existence, have parts, depend upon causes, and suffer limitations of capacity inherent in the circumstances of their coming to be, or in the particular parts and causes involved. Adhering to these rules, the Greatest thing, which must at least exist always and could not require for its existence any particular set of parts (for a Greatest set has no meaning) or depend upon any cause for existence, since to depend for existence upon something else is a defect— such a Greatest thing cannot be contingent. Here it is the critics who want to waive the rules. They want to regard one concept, lacking in any of the features found in all the things whose logical contingency is noncontroversial, as nevertheless purely contingent in its actualization.
Moreover, it is no mere inductive generalization that all contingent things have the features specified (and still others of the kind); there seems to be an intelligible connection between these features and their contingency. So, approximately, Anselm thought—and wrote. His arguments on this point have been ignored in more than nine-tenths of the refutations and but carelessly criticized in the others. This is all part of the great master lesson in how not to criticize a philosopher.
And yet, inefhicient as the process has been, it has—after quite a while!—produced some results. It is now possible to see a good way around the situation that Anselm brought about, to relate it not only to Anselm’s form of theism but to the chief logically possible forms, to remove the ambiguity
[73]
in ‘none greater’ and compare the two ways of doing this, to see clearly the distinction between atheism and positivism as alternatives to theism, to generalize the problem of contingency with respect to alternative criteria for the division between necessary and contingent propositions and to relate these to various language systems and their rules. We are, in short, in a position to inquire in this area, rather than merely to debate. So, on the whole, we may after all be at least mildly grateful to those whom we have been viewing with such severity. And certainly we should be grateful to Anselm. He did for us not exactly what he hoped to do, but in some ways far more than he could have dreamt of doing.
18. Refutation of Some Refutations
Gilbert Ryle tells us that the Argument rests upon the use of a ‘systematically misleading expression’. In ‘x exists’, existence is only a ‘bogus predicate’ and that of which it is asserted only a ‘bogus subject’”. And “if existence is not a quality it is not the sort of thing that can be entailed by a quality.”9 Another author says that the verb fo exist “takes us right out of the purely conceptual world,” and therefore “there can never be any logical contradiction in denying that God exists.” 10
In these charges I find unwitting instances of the ‘bad
9 G. Ryle, "Systematically Misleading Expressions,” Proceedings of she Aristotelian Society (1931-32); also in Logic and Language, eds. A GN. Flew and A. Macintyre (New York: Philosophical Library, 1951), pp. 15, 17.
10 New Essays in Philosopbical Theology, eds. A. G. N. Flew and A. Macintyre (London: SCM Press, 1955), p. 34.
[74]
grammar’ or bad logic of which Anselm is accused. First, the notion that ‘existing deity’ is formally analogous, say, to ‘existing tiger’ violates the grammar or logic of the character ‘divine’. Multitudes of theologians have for centuries given reasons for denying that the nature of God can consistently be viewed as a universal ‘predicate’ capable of embodiment in this or that ‘individual’. Thus for instance it was said over and over that God “does not have but is his being or his goodness.” In other words, the unique excellence of God implies a logical-type difference from all other individuals, actual or possible. How often (in old-fashioned language) this sort of thing was repeated! Second, the notion that even the divine existence is entirely extraconceptual is the very one that the Argument, fully understood, purports to disprove; and thus the quoted objection is merely the unsupported denial that the reasoning is valid. Likewise, to suppose that it makes sense to speak of a ‘purely conceptual world’, in the application which the phrase must here have, begs the question.
Ryle analogizes ‘God exists’ to ‘Satan exists’. Thereby he misuses either ‘God’ or ‘Satan’.11 For, though ‘Satan’ is indeed “not the proper name of anything,” what the Argument shows is that ‘God’, by its whole meaning, either stands for no logically possible conception at all, or is the name of an existing individual. It is no proof to the contrary that there is a verbal analogy between ‘Satan exists’ and ‘God exists’. The question of the logical propriety of the analogy is merely one of many ways to put the Anselmian issue. If Anselm is correct, the verbal affinity is indeed ‘systematically
11 Ryle, op. cit.,, p. 16.
[75]
misleading’. The phrases, ‘Nothing is both devilish and alone in being devilish’12 and ‘Nothing is both divine and alone in being divine’, are not logically akin, for all their verbal similarity! For, whereas only one conceivable individual could be divine, its individual uniqueness being thus specifiable by a pure concept, there is no correspondingly definite and selfindividuating concept of devilishness. For instance, only one being could know or love all things; but ‘hater of all beings’ has no clear, coherent, and unique meaning. Without some selflove or love of others there is no self at all.
When Professor Wisdom entitled his essay on the divine existence “Gods,” he was exhibiting very bad grammar indeed.13 Only an idol, not God, could be one among a variety of conceivable gods. It is comic—and also sad—to see how easily prejudice, rather than an alleged principle of method, in itself not unreasonable, may tilt the balance in philosophy. Let us by all means talk grammatically or logically about deity. But who that has read much in the history of religious metaphysics will suppose that this is among the easier tasks of ‘analysis’, to be accomplished by selecting logical devices which have been arrived at in dealing with nondivine things, and blandly applying them to deity as merely one more topic of discourse? This is exactly what Anselm and many others have claimed to show it cannot be!
Do any of the authors discussed in this section show the least inkling of the content of Prosl. III? Or of the Reply to Gaunilo? Or even of the corresponding passages in Descartes
12 Ryle, loc. cis.
13 John Wisdom, “Gods,” Proceedings of the Aristotelian Sociesy (1944); also in Pbilosopby and Psycho-Analysis, pp. 149-68.
[76]
(or Bonaventura)? Must I answer? They do not. The point that, whereas ordinary contingent existence is not a predicate, contingency as such and its negative, necessity as such, are predicates is simply omitted from the discussion.
The admirable clarity of much contemporary British philosophy makes it no great trick to show exactly where it misses the Anselmian point. Using Ryle’s excellent phrase, we may admit that it is a ‘category mistake’ to treat ordinary existence as a quality, deducible from a quality. But has he never heard the saying, virtually as old and persistent as theism itself, that God “transcends the categories,” that they do not apply ‘univocally’ to God? What categories, then, are being mistakenly applied? If it is a category mistake to treat ‘existence’ (in the ordinary contingent sense) as a ‘quality’ (again, in the ordinary restrictive or competitive sense), how does it follow that existence in the ostensibly extraordsnary sense appropriate only to God is not a quality, in an equally extraordinary and uniquely appropriate sense? The contrary might very well follow: it might very well be a category mistake not to treat divine (necessary) existence as a divine (nonrestrictive) quality.
The reasonable questions are, can there be such an extraordinary sense of the categories as theism requires and can this sense be reasonably related to the ordinary senses? But these questions concern primarily, not any argument for theism, but its logical possibility. Why has Findlay been almost the only critic of the Proof who has seen this and has put his cards on the table?
He who attacks the argument can do one of two things, but not reasonably both at once: he can grant, at least for the sake of argument, that ‘God’ means what theists say it means,
[77]
that is, ‘an individual’ who yet is not simply an individual, whose ‘nature’ or quality is not simply a quality, and who ‘exists’, but not simply as other things exist (Tillich overstates this by holding that ‘God exists’ is an inadmissible, atheistic expression); or second, he can refuse to grant all this, and can even contend that an idea which thus violates or ‘transcends’ the basic categories of thought cannot have a rational content. But then he must not say that the argument is illogical because it applies categories in violation of their rules. For the categories, in their usual meaning, with their usual rules, have been explicitly set aside. They are not being applied and hence are not being misapplied either! The words may still be used, but with a frankly distinct meaning. To take God to be simply an individual, simply having a nature or quality, simply existing, is certainly a category mistake, if ever there was one! Deity must itself be a sort of category, and the supreme category, and until #s rules have been investigated, there can be no demonstration that any relevant rules have been violated.
From the above the critic may plausibly draw the conclusion: ‘God’ is an illogical term. He cannot, however, prove this from the assumption that it claims, but fails, to be a case under the ordinary categorial rules. For it makes no such claim. Rather the critic must attack the legitimacy of the pretension (inherent in theism as such) that each category has two levels of possible meaning, the ordinary one and the extraordinary one applicable only to God. And to reject this pretension is to reject theism, rather than just the Argument. Except for Findlay, critics have been playing a game without specifying its rules, and the fact that the game has been played for nine centuries does not provide it with rules,
[78]
nor make it any the less in violation of reason. (Rather, the offence against reason is made the worse by this exhibition of tenacity.) If the usual categorial meanings have their rules, the unusual ones may have theirs. (See above, Secs. 16, 17.)
Another illustration. Wisdom says that the question about the divine mind has two components, a metaphysical, “Can we know other minds than our own?” and a factual, “Is there a special sort of other mind, the divine? (analogous to, Is there a special sort of mind, that of plants?).”14 Now this is a category mistake. For ‘divine mind’ cannot be a special sort of mind, a factual, competitive particularization of ‘other mind’ in general. It is the supposedly universal medium, creator, sustainer of all minds whatever and all things whatever, the transfactual source and bearer of fact as such. Is the “creator of all things, visible and invisible,” in other words, the ground of the possibility of whatever is possible, but one special sort of possible thing? The whole point of what Anselm discovered, if anything, was that the mere ‘existence’ of God is entirely metaphysical, not factual. Findlay draws the plausible conclusion that God is therefore no conceivable actuality, but only an empty or absurd abstraction. Classical theism cannot, I hold, fairly rebut this charge. Neoclassical theism, however, distinguishes existence and actuality, and does this in reference not only to God but to all things. What is exceptional about God is that in Him alone is it possible to treat existence as not only different, but different modally, from actuality, i.e., so that the one is necessary, the other contingent. The ‘actualized somehow’ here covers such an abso-
14 Wisdom, op. cs., pp. 151 £.
[79]
lute infinity of variability in the particular possible hows of actualization that all possibility is included, and hence there can be no possibility that the divine existence will be simply unactualized. This may or may not be a defensible position, but if it is defensible then the usual criticisms of the argument are woefully, ludicrously, beside the point.
The appropriate question is, What categories, if any, apply to God? or How, in what extraordinary manner, do categories apply? not, What follows from the assumption that ordinary categories apply in the ordinary manner? For this assumption renders contradictory the theism supposed (for the sake of argument at least) to form a logically possible position. An atheistic attack on the Argument is one thing—there are rules for dealing with it. A positivistic attack is quite another, and the rules cannot be wholly the same. An attack which hides its affiliations is appealing to no appropriate rules at all. No wonder the debate has lasted so long and has had so little to do with what Anselm actually wrote!
(It is to be noted that atheism has two forms, (1) dogmatic and (2) agnostic: (1) God does not exist, (2) for all we know, God does not exist. Both are excluded by Anselm’s discovery. But there remain two forms of positivism: (1) ‘God’ lacks consistent, cognitive meaning; (2) for all we know, ‘God’ lacks such meaning. Both remain open so far as the argument, taken by itself, and at least as used against unbelievers, is concerned. )
A favorite attack on the Proof is the pronouncement, ‘only propositions, not things, can be necessary’. Very well, but still, what about the proposition, ‘Greatness exists’? If that is necessary, then the proof is granted; if not, it cannot be on the ground that only propositions can be necessary. For this
[80]
is a proposition! And moreover, why not define necessary thing as ‘that the existence of which is affirmable in a necessarily true proposition’? But there is still another weakness in the objection. If a necessary proposition is one which is entailed by any and every proposition, then, since this implies that the meaning of the necessary proposition is included in that of every proposition, we may by legitimate analogy define necessary thing as that which is contained in any and every thing, actual or possible. True, a universal constituent of everything can only be something very abstract, but, as we have seen repeatedly, although Anselm could not have conceded this, in a different type of philosophy from his the implication of abstractness can be accepted.
Another familiar but weak objection is that necessity is always relative to premises or conditions and has no meaning when taken as unconditioned, as it must be taken for the Proof. The objection is weak because the proper meaning of ‘unconditioned’ is, on any conditions you please, or on no matter what conditions, ie., on ‘tautological conditions’ (von Wright). It is neutrality as between all possible alternative premises or conditions. God’s existence follows necessarily—from what? From any statement, or the denial of any statement. Just so ‘2 and 2 are 4’ follows, not from nothing, but from any assumption whatever. The statement is true if it is raining and also if it is not raining. Absolute necessity is that which requires no special assumptions or conditions rather than any others. Thus the notion that ‘necessary’ means necessitated by something in particular is groundless. To be necessary is merely to be common to all possibilities, hence neutral as to which possibility may be actualized. Of course some possibility will be actualized in any case, and so,
[81]
therefore, will the common factor in all possibilities. The notion of something special behind the necessary pulling it invincibly into existence, or holding it there, is an inept analysis. There is no constraint upon possibilities in ensuring the necessary, for the latter is the principle of possibility itself, the creative ground of alternativeness in reality by virtue of which alone there is any contingency. It is not an iron control, but an infinite flexibility, which can stretch as far as possibility itself, without losing self-identity.
No matter what world exists, or what other world instead, God exists, unsurpassably knowing that world. He can accept any world into His knowledge. He can know any truth; hence His own nonexistence is not a possible truth.
Among the anti-Anselmian fallacies which usually go unchallenged is the bold transition from Greatest (or Perfect) entity or individual to Greatest or Perfect ‘of its kind’, as though what followed from the first must also follow from the second! (I am quoting from conversation with a well-known logician. ) This nicely misses the point of the Argument, which is that, for the inference to hold, Greatness itself must be the kind. Otherwise, we have specialization or restrictive limitation, which is always competitive with alternative specializations or restrictions, hence contingent. The very absence of ‘kind’ in the usual sense is what distinguishes Greatness. The necessity that Greatness itself must be the kind rules out, not only ‘greatest island’, but also (in spite of Bertrand Russell, in one of his reckless moments) ‘greatest devil’. Anselm generally chose his words carefully: he said, not greatest god, but greatest something (‘that’). (He should perhaps have said greatest individual or being, but this is still not to specialize or restrict.) Divine capacity can take a form corresponding to
[82]
any conceivable restriction, and therefore is itself unrestricted. What else could ‘infinite’ cognitive or creative capacity mean?
Special kinds of individuals have special capacities and special limitations of capacity; Greatness is ommnicapacity, in whatever sense this is conceivable, ability to experience, know, or deal adequately with quite literally anything—in a uniquely strict sense, ‘a heart for any fate’. Obviously the individual’s own nonexistence is not included in the possible fates, since one can deal with or know that only in restricted ways. Thus, once more, the idea of the ‘nonexistence of the omanicapable’ is not consistent.
Perhaps someone will persist in urging: even an individual of a certain kind would, by Anselm’s reasoning, be better if existing necessarily than otherwise, and hence the ‘best conceivable individual of each kind’ must exist. What this shows is simply that ‘best conceivable of a certain kind’ (no matter what kind, if it is different from Greatness as such) is without consistent meaning. A self-contradictory thing is not a better thing than something else, for it is not a thing. There is, in truth, no absolute ideal for any special kind of thing that one individual could unsurpassably realize. What could be meant by ‘human being than which none could conceivably in any respect be greater’? By contrast, God could and as conceived would possess every positive value which anyone possessed, and hence His capacity for value is not conceivably surpassable. He would unfailingly participate in all joy, all knowledge, and all reality; and thus His actual feeling and knowledge would include and surpass that of every other individual, and be conceivably surpassable only by Himself, ie, it would meet the requirements of neoclassical Great-
[83]
ness. But that the requirements should be conceivable, also the least infringement of them inconceivable, and yet the very nonexistence of the being meeting the requirements should be conceivable, this is a jumble of ideas with no intelligible coherence. Nothing could explain or reduce to logical rules such a combination of absolute accident and absolute necessity. Is it not the opponents of Anselm who, at least as much as anyone defending him, lack respect for ‘rules’? Seemingly they do not care how anarchic thought about God may be, so long only as it is not to be taken as more than a dubious proposal with little or no chance of finding rational support. They perhaps do not want thought about God to be intelligible; the less so the better. By this method, one may be certain of never learning how far intelligibility can go in a given subject.
Still another refutation echoed from book to book is that the Argument establishes only conceptual existence, not real existence. But what either Prosl. II or III proves, if anything at all, is that, supposing Greatness to be conceivable, its failure to have existence in the fullest sense is inconceivable. Not merely that when we think Greatness it must be an object of our thought, as ‘fairies’ are objects when we happen to think about fairies; but that we must think Greatness, if at all, as somehow actualized or really existent. The Argument proves this, or nothing! And when some persist: but still, do we need to think Greatness, existent or not? part of the answer (the full answer would be all the theistic proofs put together) is: at least, no one can be an atheist who does not think about God, so at any rate we are rid of trouble from that quarter! As for positivism, or the denial that we can genuinely think divine perfection, this remains a plausible
[84]
contention, which Anselm does not very seriously try to refute. Here he relies chiefly upon his faith. Sdll, not exclusively, and with subsidiary arguments one can make a good case against positivism. Granted that, as Kant held, the other theistic proofs need help from the ontological, the reverse relation may also obtain, and this without vicious circularity. If the ontological proof by itself refutes atheism, another argument might refute positivism, without leaning on the ontological for that particular accomplishment.
It is a merit of Neoclassicism that, while accepting the Proof, it can nevertheless admit an element of truth in the distinction between merely conceptual and real existence. Here, as everywhere, it softens the clash of extremely opposing opinions by a more moderate or balanced view. It holds that for an idea or essence to exist is for it to be actualized somehow; but that the actual how is never capturable in a concept. The how is the particular actuality; only percepts can exhibit it, never mere concepts. And, except in the case of Greatness, neither a kind of individual nor an individual can be subject to the requirement, ‘necessarily actualized somehow’. But this is because ordinary kinds of individuals involve restrictive or competitive characteristics; they are not all-tolerant in their requirements for existence. Only the nonrestrictive essence can be ‘necessarily actualized somehow’, since any positive state of affairs will do (and a merely negative state is a pseudoconception). But the particular actuality which does the actualizing is indeed a ‘real existence’ which no concept can specify. It is restrictive and contingent. Fortunately, to know that the actualizing must in any case be done somebow is all that we need for knowledge that God Exists.
[85]
In the foregoing we have vindicated, but not on his classical ground, Anselm’s proposition: we can conceive that God is greater than we can conceive. Any concrete reality whatever is greater than we can exhaustively conceive. This is S0, in a radically unique sense, with the divine actuality, for it is the adequate integration of all actuality as so far actualized. Thus we need not leave it to the classical theist to stress how little we can comprehend God.
It has been said that Anselm’s proof takes existence to be an attribute, “whereas it is the bearer of all attributes.” ‘This, however, not only does not refute, it can be used to express, the Anselmian principle. For the character or status, ‘bearer of all attributes’, must itself be an attribute, at least if it could describe an individual. And God or Greatness can be conceived only as this very individual. Any realized predicate whatever must be describable as a predicate of deity, in the form: God knowing that S is P. God is thus the universally presupposed subject of all predication, and his own nonexistence is therefore an impossible predication, or if you prefer, an impossible state of affairs. His mere existence is the essential element in all existence (and nonexistence) whatever. Hence his existence is not possibly unreal.
19. The Argument of Proslogium III
Everyone who has attempted to analyze the logic of informal arguments knows that the explicit premises from which a conclusion is alleged to follow are often insufficient to establish it, but that additional premises which may have been intuitively or confusedly intended can sometimes be supplied from which the conclusion does follow. If, then, a
[86]
man of integrity and keen intellect claims to have proved something, ought we not to make some effort to discover all the premises which he was in a position to appeal to, whether or not he stated them in the very paragraph or two in which he first introduced his proof? And if the man wrote several additional pages, further explicating his thought, and later a few more in reply to objections to his argument (an argument which, if valid, must be highly important), is it a great deal to ask that his reasoning be taken to have at its disposal any additional premises introduced and shown to be reasonable in these two sets of pages?
Suppose, however, that the first commentator does not follow these sound principles, but makes everything depend upon the initial statement given in the first page or so. And suppose his report as to the structure of the entire argument is widely accepted, a structure which fails to supply clear and, upon careful examination, credible premises from which the conclusion logically follows, what is likely to happen? Obviously many will treat the reasoning as a mere sophistry. Thus we can understand Gaunilo’s myriad followers. But then there may also be those who greatly respect the man who made the original claim, or who feel (as mathematicians often do) that some such proof should be possible. They will grope about until with luck they may find the missing premises, or something like them, or at least intuitively supply them without quite knowing that they are doing so. Thus perhaps Descartes and many others who accepted the Proof but show no signs of knowing Prosl. III. Then after many centuries comes another and far greater Gaunilo, David Hume, and still another—whose name needs no telling—of extraordinary intellectual, personal, and literary power, who points out once
[87]
more the gap in the argument, as it is presumed to stand, and having—at this stage in his career—no feeling that such a proof should be possible, rather very much the opposite attitude, he drops, and virtually forbids for all time, the effort to find any missing premises; instead, he drives through the ‘gap’ with all possible energy and éclat. After him come the tens of thousands of Gaunilos in nth degree, and how triumphantly they march through the now—as they feel— comfortably widened breach.
Finally, at long last, more than a century later, the secret begins to emerge (a few had known it long before; for instance, Bonaventura) ;: there were additional premises all along, and not unreasonable ones at that. They were actually stated— briefly but with considerable clarity—in the second sentence of Chapter III, and brilliantly elucidated thereafter, best of all in Reply I and V.
This approximately is the story of Anselm’s argument. One thing is left out in the above account: the reason which made some think that an ontological proof should be possible. My view is this: the more fully one understands the great Definition, or any reasonably equivalent explication of the religious idea, the more one is put in possession of all metaphysical axioms. The more powerfully intuitive the mind, the shorter the step to the ‘missing premises’ in question, indeed to any metaphysical premise whatever.15 The premise, ‘God is conceivable as not conceivably surpassed (by others)’, is implicitly the whole of metaphysics. This is what Anselm proved, when all the implications of his reasoning are thought
15 C. Harwshorne, The Logic of Perfection (LaSalle, Illinois: Open Court Publishing Co., 1962), chapters 3-4.
[88]
out. So of course some minds would in effect bridge the gap in the argument of Proslogium II, a gap which is greater the more purely verbal one’s understanding of the Definition happens to be. And, equally of course, others would not bridge it. Anselm bridged the gap easily enough; he knew that there were not two conceivable states, existence and nonexistence, to be considered for Greatness, but only the one conceivable state. But if Anselm bridged the gap, he also sensed it; and that is why, after having, virtually in the next breath, supplied a new premise (two premises really: that we can conceive of necessary existence, and that this is superior to nonnecessary existence) which bridges it explicitly, he gives perhaps nine subsequent formulations of his Argument, only two of which are of the Prosl. II form, all the others being of a form either identical with, or closely related to, that of Prosl. III. Does this account not make sense out of the textual and historical facts? Has the reader one which makes better sense? If so, I hope that I shall learn what it is.
The principle of Prosl. Il is, as all know: to exist is better than not to exist. Let us call this the ‘false’ Anselmian Principle. Of course it is indeed somehow ‘better to exist than not to exist’ (unless Schopenhauer is metaphysically right); but to define the ‘best possible’ being, and therewith assert its existence, is apparently to make optimism true by definition, not a very impressive procedure. “If God is God, He, being best, must at least exist.” But if God does not exist, then ‘He’ is neither God nor anything else, for there is no ‘He’.
The true Anselmian Principle, which so few know, that of Prosl. III, is, To exist without concesvable alternative of not existing is better than to exist with such alternative;
[89]
hence Greatness is incapable of the latter. Since contingent nonexistence (or ‘merely mental existence’) is here simply excluded, the comparison of it with objective existence— apparently the very basis of the Palse—is actually ruled out by the True Principle. The comparison is not only irrelevant, it is inconceivable. And one has not, in this version, put ‘existence’, in the usual sense, inside a definition. One has rather described a different kind of existence, a kind which resolves no such disjunction as ‘to exist or not to exist’, but transcends it as here inapplicable.
Will any man of sound mind ask us to admit that the False formula is an adequate version of this True one? With a touch of genius, with some luck, and more than a touch of good will, one may possibly succeed in reconstructing or guessing the True from the False Principle; and accordingly, the latter can perhaps be viewed as a crude, distorted simplification of the former. But this is about the most which can be said for it. Yet practically the entire learned world has taken—and takes—the False Principle for the essential form! With what authorization? (a) Gaunilo & Co. so took it; (b) Prosl. II comes before Prosl. III; (c) the False Principle is much handier to refute; yet (d) in refuting it we also refute the ‘True’, since if ‘existing’ cannot be a predicate neither can ‘necessarily existing’, which includes it. These are the reasons that I can imagine for neglecting Chapter III. They are all poor ones. Only (d) even deserves further consideration, and it is fallacious. The ‘existing’ which can be proved not to be a predicate is ‘contingently existing’, the ordinary mode of existence, and that this cannot be a predicate (in the sense required for an ontological inference) is a simple tautology, from the meaning of ‘contingently’. (It is a more subtle,
[90]
implicit tautology from the meaning of ‘ordinary’ or ‘imperfect’.) But this mode of existence is certainly not included in necessary existence. Hence (d) is a fallacy of ambiguity. (See below, Sec. 23, (a) and (n).)
To have prefaced his great third by his so inferior second chapter was indeed a sad blunder of Anselm’s; but need it deprive us of our intelligence or our scholarly conscience? A man who discovers a basic logical principle involved in the most important of all ideas can be allowed some initial groping, one would think, without all the world groping with him forever, merely saying ‘no’ where he said ‘yes’.
In Ueberweg's History of Philosophy (the most learned, ot at least detailed, of all), the two Principles appear in their Latin texts; so far so good. But THE FALSE PRINCIPLE IS PUT IN LARGE LETTERS, thus delicately showing that the point of neither had been grasped, and that the editor was thinking with other peoples’ minds rather than his own—or Anselm’s. [ am asking scholars to laugh, for it seems hopeless to try to be merely solemn about so ludicrous a collective mistake. “This king is naked”—the king being Gaunilo, enthroned in majesty for an amazingly long reign to which his original title was decidedly doubtful—if not wholly ‘illegitimate’. If the true title, that of the author of Prosl. III and the Reply, were at last recognized, we might begin to understand that most brilliant philosophical product of the Middle Ages, the Proslogium!
In fairness to countless people it should also be said that a combination of bad management and very bad luck built a trap for human nature in the Anselm case. It was bad management of Anselm to make his opening and most dubious step in the exposition of the argument appear to be the
[91]
argument itself, or at least, an element upon which the whole must depend: when in truth, if Prosl. II has any point at all, it is only as a loose analogy to the definitive and independent reasoning of the next chapter. (Doubtless Anselm partly deceived himself, as well as others here; he did repeat the Prosl. II reasoning twice in the Reply, where, however, that of Prosl. III or some related argument occurs seven times! and he never seemed clearly to see its weakness—or its triviality, if it only means that, of course, the worshiper cannot take what he worships to be a mere fancy.) It was very bad luck that the first critic should have been so careless a man that he could virtually ignore two-thirds of the very short text he commented upon and that the prejudices of this critic should have been just the ones which would make him seem an oracle to those who had not yet learned the very lesson Anselm wanted to teach. So they did not check his scholarship and forgot to exert any of their own. The lesson was thus cut off at the root, as it were. After a few more reiterations, scarcely anyone suspected there could be doubt, at least as to the main outline of the argument.
Add to all this that Anselm could not well have broken the hold of Greek thought upon his own thinking sufficiently to solve the abstract-concrete paradox by which the credibility of his reasoning was menaced. Here was more than a ‘gap’, to be filled in perhaps by adding a premise; here—unless one took great care in how he resolved the ambiguity in ‘none greater’—was a contradiction, beyond remedy by any mere addition. The critics were not wrong in thinking there was a grave error or danger of error somewhere, but they were too Greek themselves to guess accurately what it might be. When they were not Greek metaphysicians, they were (like Hume)
[92]
Greek skeptics; whereas Anselm’s discovery called for something different from either form of classical thought. The last one hundred years have brought it about. It is the conscious emergence of a theism which can admit the abstractness or inactuality of the merely necessary or eternal, without giving up the notion of the all-surpassing and necessarily (somehow) actualized God.
Anselm’s argument, in Proslogium III and thereafter, has (if the implicit is made explicit) something like the following structure. One and only one of four views must be correct:
A) The property G, of being such that none greater is conceivable, is itself inconceivable (There is no ‘idea of God’).
B) The property G is conceivable, but only as not existing.
C) The property G is conceivable as existing, and also as not existing.
D) The property G is conceivable, but only as existing.
Since A), B), and C) are absurd, D) must be true. A) may be called ‘positivism’, B) seems to have no name, C) is ‘empiricism’, D) is ‘a priori theism’, (in contrast to A) and B), which are the two forms of ‘a priori antitheism’).
Anselm took A) to be counterintuitive; he also had an argument against it (in the Reply).
B) apparently did not occur to him. It might perhaps be illustrated by Kant’s doctrine of the merely regulative use of the Ideal of Reason. Anselm’s main effort was to refute C) as absurd. He has two lines of reasoning, at least, to which we add a third (still others are possible) :
[93]
1) Since ‘inconceivability of nonexistence’ (necessary existence) is conceivable and must exalt what has it above all else, G must imply it.
2) If inconceivability of nonexistence were not conceivable, neither would G be conceivable (to Anselm it seemed evident that both are conceivable).
Conclusion from 1, 2: C) is absurd, whether inconceivability of nonexistence (necessity of existence) is or is not conceivable, hence in any case;
3) Besides, the nonexistence of G has no empirical meaning; yet, by C), it has no a priori meaning either; hence again C) is absurd.
As to 1). Inconceivability of nonexistence guarantees eternity, indestructibility, self-sufficiency, independence of causes of existing; whereas conceivability of nonexistence (contngency) makes any such guarantee unintelligible; hence such contingency is a defect and contradicts G. Therefore, if the nonexistence of G is conceivable, whatever hypothetically may be supposed to exemplify G must at the same time, and contradictorily, be thought not to exemplify it. This point is made over and over again in the Reply, Chapter I, and Chapter V (4th and 5th paragraphs). What is conceivable as nonexistent would be surpassable ‘even if it existed’; for the conceivability of nonexistence must remain, and could not be altered by the fact of existing. Facts do not determine conceivability, modal status is independent of them. Thus, as Aristotle in effect argued long ago, such status can only be necessary. This is the anticipatory rejection, and legitimate rejection, of the stock anti-Anselmian phrase, ‘exists necessarilyif it exists’. The ‘if’, as such, implies the very conceivability
[94]
which one resolution of the alleged alternative must contradict; accordingly, as Malcolm correctly says, the phrase contradicts itself. True, six times in the paragraphs referred to Anselm uses an equivalent of ‘if it exists’; but always, note well, as a step in a reductio ad absurdum argument against the notion that an sf can here properly apply. The fifth paragraph sums it up with lucidity. It is not Greatness to whose existence the ‘if’ is relevant, but a fake substitute, some inferior sort of being whose nonexistence may indeed be conceivable, but whose existence as petfect would be contradictory. It is Anselm who seems to me to have thought clearly at this point, not his critics.
As to 2). G must at least entail eternity, indestructibility, self-sufficiency, and these, the inconceivability of nonexistence; hence if this inconceivability is not conceivable, neither is G.
Therefore, be the inconceivability of nonexistence conceivable or not, G cannot be conceived as in C).
As to 3). To conceive something as existent is to have some idea of an experienceable state of affairs in which it would exhibit itself as existent; to conceive something as nonexistent is to have some idea of an experienceable state of affairs in which its nonexistence would be exhibited. But no experience could establish even the slightest probability of the nonexistence of God, because, since any state of affairs which would amount to the nonexistence of God would be a possibility which must be prevented from realization for God to exist, his existence would thus be made in conception to depend upon the nonrealization of this possibility, a dependence which contradicts the existential self-sufficiency entailed by G. Thus to admit such a possibility is to admit that G's existence is impossible, contrary to C).
[95]
As to B). In spite of Kant, I do not believe an ideal can have a clear and consistent meaning, yet be (for all we can know) incapable of existing. I believe that there are ambiguities or inconsistencies even in the regulative use of the classical idea of God which cannot be overcome without at the same time removing the grounds for denying a constitutive use also. And in fact neither science nor philosophy has found much regulative use for Kant’s version of the idea of God.
A version of B) is the notion of a ‘limiting concept’ (like ‘perfect lever’), an approachable but unreachable limit of thought or action. But the divine perfection, as Anselm defines it, cannot even be approached. There remains always an infinite gulf.
I submit: D) is the most intelligible of the four ways of classifying the property G.
A common objection is, Yes, if we think God, we must think Him as existing, but must we think Him? Well, if we are like the animals and do not even raise the question of God, then of course we need not answer this question; but if we are philosophers, we will raise the question, and then only one answer is logically tenable, the affirmative one. In addition, even the animals, though they cannot explicitly ‘think’ God, must in some way (this follows from the definition of G) relate themselves to Him, and they cannot relate themselves to G as nonexistent, for this is nonsense. So this objection, which Windelband, for example, seems to plume himself upon, is not so very cogent, after all. It is obvious from reading them that both Anselm and Descartes were in a position to deal with it.
Anselm'’s greatest weakness is his failure to dispose of positivism, A), in satisfactory fashion. And with his Neoplaton-
[96]
ic view of the meaning of G, he could not have been cogent at this point.
Here is another version:
(1) It is possible consistently to conceive of an individual (‘God’) as such that none greater can be conceived.
(2) It is possible to conceive of an individual which could not conceivably have failed to exist.
(3) Such an individual (as in 2) must be greater than one which could conceivably have failed to exist (Anselm’s Principle).
(4) It is conceivable that no God (defined as in 1) exists.
(5) Hence the idea of God (defined as in 1) becomes the contradiction, an individual such that (1) none greater can be conceived, and yet (2, 3, 4) such that a greater can be conceived. But according to (1) the idea is consistent. Hence the four premises cannot all be correct.
(6) To escape contradiction, at least one of the premises must be negated.
Negation of (1) and/or (2) constitutes positivism; negation of (3) conflicts with the intuitive discrepancy between contingent existence and unsurpassable excellence. On this intuition there has been, through the ages, a large measure of agreement among theologians. Thus negation of (3) is implicitly positivistic also. Negation of (4) is a priori theism. It removes the contradiction. Premises (1) and (2) are theistic ‘meaning postulates’; they postulate only conceivability. Premise (3) is a corollary of a metaphysical meaning postulate explicating part of the content of ‘contingent existence’, that it connotes dependence, insecurity, surpassability.
[97]
To insist upon (4) is to render theism unintelligible. Meaning postulates in this unique case cannot be neutral to all individual existence; they make the divine existence either necessary or impossible. (4), though atheistic, implicitly begs the question against the conceivability of God, and thus the distinction between positivism and atheism cannot be maintained, except in this sense: atheism is the naive or unwitting form of positivism, positivism the sophisticated form of atheism. The most crucial phase of the argument is what I call ‘Anselm’s Principle’, the incompatibility of Greatness with conceivability of nonexistence, i.e., with contingency. This may be symbolized as
N(p*—>Np*) or N~ (P* & ~Np*)
The proposition affirming the existence of the unique x which is Unsurpassable strictly implies its own necessity; it cannot be merely contingently true.
This principle is affirmed and justiied in the first two sentences of the great third chapter. The point is not simply, “Thou dost in fact exist, and therefore also necessarily,” but rather “since the only form of existence appropriate to thy definition is necessary existence, therefore [barring an irrational, though in our time fashionable, assumption that necessity de re and de dicts are radically independent, an assumption which makes sense at most where conditional, and not as here unconditional, necessity is in question] the only kind of truth which the central religious belief could have is necessary truth.” From this it follows that necessary falsity is the only way in which it could fail to be true if it has any meaning at all.
[98]
It has been asked whether N(p*—>Np*) means that the necessity of p is deducible from p. It is certainly deducible with suitable meaning postulates, and without some such postulates one has no determinate proposition to deal with. To explicate the meaning of ‘Greatness exists’ is to rule out the notion that its nonexistence is conceivable. A proposition whose denial is not consistently conceivable is a necessary proposition, and no proposition is consistently conceivable if what it asserts is inconceivable.
Another, more modern, way, less close to Anselm, of putting the point is this: the statement, Deity exists, implies that we are all God’s creatures, and even possible worlds are merely His possible creatures; hence the statement, Deity does not exist, implies that we—and possible worlds—neither are nor could be anything of the kind. These, therefore, are not two factual hypotheses—they are two languages, two systems of metaphysics, two competitors for the task of elucidating what we can and cannot mean by any of our basic conceptions. They concern, not ‘what are the facts?’, but ‘what is it to be a fact?’. They concern what Heidegger calls Being, the ‘ontological’, not the merely ‘ontic’. The category of fact, not any particular application of the category, is the issue. Prosl. II is a poor text for this reasoning. I even think it difficult to imagine a much worse one. A greater blunder by so great a man might be hard to find.
It is worth noting, as Koyré remarks, that we have no knowledge whatever of Gaunilo’s judgment after he received the Reply. (How many textbooks and histories fail even to tell us that there was a reply!) For all we know, Gaunilo changed his mind concerning the force of his objections.
[99]
20. Proslogium II, III, and Anselm’s Principle
Of all the men of genius who have tried to divine Anselm’s essential point from Prosl. II alone, not one, it seems, has ever arrived at the true formula, the genuine Anselmian Principle, that what is conceivable as nonexistent is inferior to what is not thus conceivable. The derivation might run as follows. If one accepts Prosi. Il as valid, i.e., admits that what exists is greater than what does not, and that therefore absolute Greatness must exist, one may then go on to argue that an existence which is thus deducible must be necessary existence, since otherwise one would have something both logically necessary and logically contingent, a contradiction. To suppose the validity of Prosl. II is therefore to admit that the existence, lack of which would contradict Greatness, is not simple, but necessary, existence. Hence ‘the existent is better than the nonexistent’ must in this case become, ‘the necessarily existent is better than the nonexistent’. However, Greatness must equally be better than the contingently existent, since contingency of existence is a defect. So what we really have is that the necessarily existent is better than the contingent, whether existent or not, and also, as is obvious, better than the impossible, or necessarily nonexistent, since anything is better than the last. Thus the distinctive point about Greatness is that it is better than anything contingent can possibly be. This is the principle of Prosl. III, the essential principle of the Proof. It is derivable from Prosl. II only on the supposition that the reasoning of Prosl. II is valid; and yet to see this validity one needs the Principle. Until we know that God’s existence is deducible, we do not know, from anything said in Prosl. II, that it must be necessary existence which is in question; and until we know this
[100]
we cannot know that it is deducible. Deducibility and necessity come to the same thing. Hence we must have necessary existence in mind throughout. This, however, is the procedure of Prosl. 111
If the preceding paragraph is correct, the historical procedure of making Prosl. II the definitive statement of the Proof was an excellent way to make reasonably sure that the main point of the two chapters taken together would be missed! To refute Prosl. II, taken by itself, is easy enough, since so taken it seems to misuse the idea of existence; yet this does not refute Prosl. III, which turns not upon ordinary or contingent existence, but upon a contrasting modality of existence. And no parade of ordinary things, as existing nondeducibly or contingently, has any direct relevance, save by contrast, to the question of the extraordinary thing of which it is known that it could not exist merely contingently, because we can see, directly and indirectly (and in many ways), that this is a defective and limiting mode of existing, incompatible with Greatness.
Was there any justification or excuse for the implied claim of nearly all the Argument’s opponents that by refuting Prosl. II, they had disposed of Prosl. III? They might defend themselves as follows: “You seem to say above that Prosl. II is valid if and only if Prosl. III is so; but then their validity is equivalent, and to disprove one is to disprove the other.” My answer: I am not speaking in both cases of formal validity but, with respect to Prosl. II, of cogency relative to the meanings which key terms are likely to be given. If existence means general existence, it cannot as a universal rule be taken as a conceptual predicate. One must discern in ‘divinity’ the requirement for a formally, i.e., modally, unique
[101]
type of existence in order to grasp the unique deducibility of existence in this case. It is not that God is conceived as too good to be nonexistent, but that He is conceived as too good to be conceivably nonexistent, to be capable of failing to exist. With all other individual things this impossibility of failure to exist does not come into question at all; only with God. It does indeed follow that if one adequately understands the meaning of Greatness, one understands that God is conceived, if at all, as too good not to exist, but only because He is conceived as too good to be even possibly nonexistent, and what could not be so is not so. However, all this is contained in Prosl. II only through an aspect of the meaning of Unsurpassable Greatness which is not in that chapter explicated for the purpose in hand. Prosl. II, therefore, could have put the critics in touch with the points they had to refute only if they were already aware of all that Prosl. III makes explicit (even then, not fully explicit, since the Reply carries the explication further). But if the critics had seen all that, they would also have seen the irrelevance of their contention about (ordinary or simple) existence not being a deducible predicate. Of course it is not; Anselm in effect not only admits but insists upon this.
As someone has remarked, the reasoning of Prosl. II suffers from the ambiguity in ‘X is thinking of something Unsurpassable’. Does this mean, he is thinking of something which is Unsurpassable—in which case, is not existence presupposed rather than proved? Does it mean, he is thinking of something which would, were it to exist, be Unsurpassable? In that case, that the something does not exist seems no contradiction—at least, not for the mere reason that the something lacks a certain merit attaching to existence in its
[102]
merely ordinary meaning. For why not take the Unsurpassability as hypothetical, a characteristic which something might have were there to be such a something? If there is no such something, then indeed nothing ¢ Unsurpassable, but still we have, it may be claimed, the consistent idea of what it would amount to if something were of this kind. Prosl. III, however, shows that this hypothetical way of dealing with Greatness is illogical, since even the bare conceivability of not existing is a defect, and one which could not be removed by existence, for what is conceivable remains so, no matter what exists. Thus the modal argument of Prosl. III protects the apparently nonmodal reasoning of Prosl. II, but only by showing that it is implicitly modal.
Even an admirer of Anselm ought, I think, to admit that Prosl. II was, at best, a highly misleading first presentation of his case. Contingency is a weakness in a much clearer sense than is mere nonexistence. Unless a thing exists, there are no properties of the thing; and while we may say that at least the idea is there, and that the nonexistence of its object is a weakness of the idea, the trouble is that we seem then to be talking about the properties of the idea rather than of its object. We seem to be committing the ‘homological fallacy’, making the universal an instance of itself. (However, see Part Two, Sec. 21.) But taking contingency as the defect, we can then compare in thought a necessarily-existing and a contingently-existing thing, and decide which must be greater; then, seeing that the former is greater, since (among many reasons) its existence involves no dependence upon anything but its own essential nature, we deduce that if the Supreme is conceived as existent at all, it must be conceived as necessarily existent, ie., as not conceivably nonexistent. But if
[103]
we try to add, ‘yet perhaps it does not exist’ we are simply saying, “perhaps the inconceivable is true.” And then indeed we are talking nonsense, or at least we are using ‘inconceivable’ in a loose sense, whereas the strict sense is what the reasoning requires. For the merit which the Supreme must be conceived to have is strict impossibility of nonexistence, not some fake substitute. Anything with less than this strict impossibility will be inferior to anything with it, and so its denial is incompatible with Greatness.
Thus the reasoning of Prosl. Il is incomparably more powerful—though less simple—than that of Prosl. II, and, therefore, the ‘locus classicus’ for the Argument (to speak with one of the many reference works whose miserable treatment of this topic I have had the misfortune to read) is precisely not the earlier but the later chapter.
Detached from its context, Prosl. Il seems at best no more than an inconsequential truism. The God of religion cannot by the worshiper as such be taken as nonexistent: for to worship something and yet recognize it as a mere fancy is not possible. But this by itself is trivial enough, for perhaps one need not and if intellectually honest will not worship at all! It is quite another thing to discover that not simply the supposition of the actual nonexistence of ‘God’ contradicts the religious use of the term, but even the supposition of the conceivability of this nonexistence does so. For then, not only can the person with faith not conceive the unreality of the object of faith, no one can do so, with or without faith. What a world of difference! (And think of taking nine centuries to arrive at what Anselm had said plainly enough.) The worshiper can know that unbelievers cannot be discussing the idea of God at all. They may be using the word,
[104]
but they cannot have the idea. For the idea forbids the value false and admits only the value necessarily true. One may deny that there is any such idea, may—with Gaunilo— say that there is perhaps but a word (or an emotive meaning only), but the moment one goes beyond that, and tries to separate the idea, admitted as such, from the truth of the idea, one is in Anselm'’s grip, so long as logical rules are obeyed and one understands his definition.
Moreover, since the existence of God is the existence of a universal power, infinitely important to all that is or can be, the discovery that this existence is either necessary or inconceivable means that an entire system of metaphysics is the one or the other. Any corollary of the existence of God is also necessarily true, or else necessarily false (or nonsensical). But these corollaries are the whole of metaphysics, the basic axioms of philosophy. This, too, the magnificent doctor saw, or partly saw.
The notion that we can find God, or the absence of God, by picking among the stones of particular facts, is on the face of it foolish; but Anselm found the precise logical reason: God is not conceivably a fact at all (so far as the bare necessary truth that he exists is concerned). The facts are by definition neutral to this—as to any—necessary truth. (Only the actuality of God is factual. But it we could not expect to prove or disprove since it is incomparably richer than all we can distinctly perceive or imagine.) The evidence for God's existence is then wholly intellectual and spiritual, not sensory. Anselm proved that it must be so, and that, in consequence, the whole of metaphysics must derive its logical structure from this necessity. Has there ever been a greater discovery?
[105]
If there be a way of rescuing Prosl. II from the charge of grave ineptness, if not sheer fallacy, I have not in almost a lifetime of reflection been able to find it. Here I agree with Malcolm. But the modal argument of Prosl. III seems to stand entirely on its own feet, and to need no help from the previous chapter. Here also one may agree with Malcolm. There remains, however, the Findlay paradox, and the necessity of resolving the ambiguity between ‘none greater’ and ‘nothing greater’.
The best that one can do with the Prosl. II version (taken by itself) seems to be as follows: to think of Greatness is to think of something as at least as superior to any unactualized possibility as the corresponding actualized one is. However, this is but a rough analogy, since the sense in which a contingent actuality can be superior to its unactualized possibility is not at all the same as the sense in which the necessary existence of deity is superior to any unrealized possibility, whether that of deity (which would be contradictory) or of anything else. Rather, deity surpasses all other things by the very sense in which its logical ‘possibility’ is to be taken—namely, as equivalent to its necessity. And the reason, not quite seen by Anselm, is that this individual existence is expressible purely conceptually or abstractly. Only God could ‘adequately know all things’, for example. In this abstract idea, the uniqueness of the divine existence is already contained.
But, on that basis, we cannot well deny that either (Findlay paradox) the divine reality is itself abstract (logically empty) or else (solution overlooked by Pindlay) it is not exhausted by its mere individual existence, but has a further definiteness which is genuinely additional to its logical possibility and is therefore contingent. And so it appears once more that
[106]
Anselm’s main mistake is in his idea of God itself, not in his proof.
Moreover, Anselm cannot escape the difficulties of his Neoplatonic idea so easily as Barth seems to imply when he argues that Greatness is not a description of deity but a mere rule for our way of thinking about deity, a mere negation that anything could conceivably be greater. For the positive nature of God which escapes us not only cannot be known by us to be necessary, it cannot even be necessary. The only rules that make sense about ‘necessary’ imply that its referent must be abstract, so that even the ‘description’ which God Himself could give of His ‘essence’ must be logically weak. Hence the admission of divine accidents is obligatory.
There is another consequence of these rules: if God, merely as such, merely as omniscient and the like, can be necessary because—though only because—of His extreme abstractness, then so can ‘world as such’, as merely ‘whatever, other than Himself, God knows’. For this is equally abstract. Thus the ‘contingency of the world’ would become that of this world, not of there being some world or other.
The question of the ontological proof is simply this: what way of thinking about contingency should we adopt, and what are the consequences of this adoption? Our decision will implicitly determine an entire metaphysics. Neither Anselm’s way, nor that of most of his critics, has much to commend it, so far as this writer can see. We need a fresh start.
21. Definite Thought Is about Something
If a man is thinking about fairies and upon being asked what he has in mind replies, “a mere idea,” he may be further
[107]
asked, of what is it the idea? The man could, however, reply: “I am thinking of the world process as productive of all sorts of things, and of the legends of fairies as views about what this process could be imagined to turn up.” The Unsurpassable, however, could not conceivably be turned up, and therefore there can be no factual failure to turn it up, either. There is nothing at all to think about, under the name of God, unless just God Himself. It is a unique case, in which to think (consistently and more than verbally) about a certain kind of individual, and to be in real relation to the one possible such individual, as really existent, is a distinction without a difference. All attempts to find an intelligible difference fail. Few indeed even try to tell us what the intelligible difference might be; they content themselves with repeating the bare words, “God may not exist.” In all other cases, we can explain the sort of thing we mean by a negation of existence; here we cannot.
We can, however, explain, in innumerable ways, what we mean by ‘God exists’. It means, for one thing, that we are fully known, and cannot in the everlasting future ever be unknown, hence that our previous lives are registered where neither moth nor rust doth corrupt and where thieves do not break through nor steal. To say that this is not so, however, has no intelligible meaning, since we cannot conceivably experience the noneverlastingness of the factual content of existence. We have to live and think as though the past were indestructibly real, for otherwise ‘fact’ would have no definite meaning. ‘God’ merely makes this necessary idea more intelligible, that is all. And so with other ideas which we must have: they are all summed up, focused, united, made sense out of, by the idea of God. What is ‘added’ is only the
[108]
more complete intelligibility. God is not (except in the how of His actualization) a fact among facts, but any fact as differing from nothing only through a love which sways and registers all occurrences.
It is to be noted that the only Faith that Anselm’s Proof justifies is faith in God as greatest conceivable being. People’s judgment as to what exalts God over all others may differ at many points. Our task is to distinguish much more carefully than the Saint did between the truly self-answering religious questions and all others. His glory is to have discovered that there are those of the former type, and that they are all corollaries of two questions: What does ‘Unsurpassable’ imply? Are the implications consistent and definite enough to be capable of truth?
22. The Proof and Pantheism
Is Anselm’s argument, as Weber alleges, ‘pantheistic’? This much-abused word has a wide spectrum of meanings. If it connotes the Stoic-Spinozistic view that all things follow necessarily from the eternal divine nature, then the relations to the Argument are as follows. If the necessary reality we prove is taken to be God in His concrete actuality, for instance, God as knowing that you and I exist, then so are we (and all things) necessary. For if we could have failed to exist, then God could have failed to know our existence, since one cannot know as existing what does not exist. As Husik has shown, this is why Spinoza, following Crescas, was a Spinozist. The only escape from such Spinozism, while using the Argument, is via the frank recognition of the Findlay paradox, and the acceptance of its resolution in neoclassicism. The
[109]
Argument, by deducing existence from an abstract definition, proves that the necessary in God must be but an abstraction from His total reality.
Of course we shall still have another kind of ‘pantheism’, if the word is used very broadly. God, as knowing all things, includes them somehow as contents of His knowledge. In addition, if things were simply ‘outside’ God, there would be a greater reality than God, God and the world. (Deny that this is in some way greater and you merely deny that the world has any value whatever. And on that basis you could not know what ‘value’ means.) Pantheism in this sense is simply theism aware of its implications. But obviously the very point of ‘God exists necessarily’ is that other things exist contingently; therefore, when Spinoza denies contingency altogether he is not improving upon Anselm, but is merely cutting off the branch of meaning on which the Anselmian idea must perch. However, the abstract-concrete paradox shows that classicism (the denial of contingent states to God) is a logical blunder equally whether it takes the pantheistic or the theistic form. Apart from that blunder, there is indeed no real issue between theism and pantheism. We all exist in the divine being, as St Paul said. What more could a pantheist want, if he renounces necessitarianism (and determinism)? We are in the free, partly-contingent divine life with our own contingency and freedom. (That this is possible means that the divine life does not consist in mere ‘power’, mere control, but has also a passive aspect, as all life indeed must have. Lequier, Fechner, Varisco, W. P. Montague, E. S. Brightman, and Whitehead are among those who have defended this view.)
[110]
23. Some Recent Criticisms of the Proof
The validity of our position may be tested by considering some recent criticisms of the Anselmian argument, particularly as the argument is presented by Malcolm.16
(a) The ‘second argument’ (from necessary existence as a predicate) is held (Allen, Abelson, Penelhum) to imply the principle of the ‘first argument’, that existence is a predicate, which, as Malcolm himself says, is invalid. Yet that existence in the one case of Perfection is a genuine predicate is quite compatible, as we have seen, with its not being so elsewhere. The critics are indeed right in holding that existence must always have a contingent aspect, for it implies a step from the abstract to a particular concrete case and thus a passage from one logical type to another—a passage, moreover, in the direction of greater definiteness, since the concrete instance must have further qualities not specified in the predicate that is being considered. That such a step in its particularity should be entirely necessary would be sheer contradiction. But the ontological Proof, when employed in a neoclassical system, need not take this step to be necessary. Consider any predicate, H: if it is actualized in a certain way, or by a certain concrete instance, it is always possible to conceive it as actualized in another way, by another instance. If H means human, or ‘rational animal living on the earth’, there might have been such animals different in each case from any which have actually existed. There might also have been no human animals at all; but this is a further
16 See The Phslosophical Review, 70 (1961), for articles by R. E. Allen (pp. 55-66), R. Abelson (pp. 67-84), T. Penelhum (pp. 85-92), P. Henle (pp. 93-101), and G. B. Matthews (pp. 102-3).
[111]
point which, if we abstract from the sort of predicate being considered, and let H stand for predicates in general, is not deducible from the mere contingency of particular instances.
Thus we have:
(1) H is actualized somehow, it might have been actualized otherwise; also it might have been unactualized.
(2) H is actualized somehow, it might have been actualized otherwise; but it is not true that it might have been unactualized.
What contradiction or absurdity is there in (2)? There are propositions whose denial is inadmissible; how do we know that none of these is existential? It cannot be merely because of the type difference between predicate and instance; for we have just shown that this difference is compatible with necessary existential statements, provided that they limit themselves to saying that their predicates are actualized somehow, without specifying how, or in what particular instances.
That divinity is a predicate with no more than one possible concrete instance, so that the necessity of the divine existence refers to a unique concrete actuality, is indeed logical nonsense. No actuality can be necessary. Findlay is here correct, the conclusion aimed at in the classical proof was in reality disproved. But suppose that ‘divine actuality’ means a kind, not a unique instance, of actuality. Would it follow that God is not individual? Yes, if individual means a single determinate actuality. And this is precisely what is at issue between those who hold a substance theory of personality and those who hold the Buddhist-Whiteheadian or event theory. According to the latter, an existing person is a sequence of actualities, several per second presumably (in the human
[112]
case ), not one of which, unless the first one, is necessary to the individual’s being himself. (Leibniz held the contrary opinion, but his view has its penalties.) At each moment, I have the capacity to be actualized the next moment in more than one way. And since in the case of an uncreated being there could be no first actuality, we may drop the qualification, ‘unless the first one’, and say that God could have been Himself even though every actuality in or by which He has existed had been otherwise. (There are, of course, innumerable theologians who deny the applicability of the concept of personality in this sense to God, but there are others who admit it, and a refutation of the Proof can be general and inclusive only if they too are taken into account. This has never been done, to my knowledge.)
(b) Here is the place to consider Abelson’s contention that there is not a single idea of ‘God’, but a loosely-connected family of ideas. This is plausible, but it is also fair to say that only two of these ideas are relevant to the Proof: the paradoxical one which takes the necessary existence of divine perfection to mean a necessary yet unique actuality, and the other which takes it to mean the necessary realization of an individual divine nature in some suitable actualities or other, which actualities being a contingent matter. Either (1) Divine perfection is necessarily realized in the very actuality, or in the very manner, in which it is realized, or (2) it is necessarily realized somehow, in some actuality or actualities able to realize it.
Cortesponding to these two views are two concepts of perfection: (1) a being than which none could conceivably be greater, not even the same being in another actual state; (2) a being than which none could be greater except the
[113]
same being in another actual state. Only (2) escapes the absurdity of a necessary particular actuality, yet it too furnishes a premise for the Proof. For an individuality necessarily actualized somehow is superior to one which merely happens to be actualized. The critics deny this, but on poor grounds. One way to put the Argument is as follows: in religious terms, ‘God is perfect’ means that He is the appropriate object of total devotion, so that all concern, even all interest, has Him for its object. But the interest in the ‘possibility’ that He might not have existed at all—does that have God as object? Any interest in a situation alternative to God would violate the Great Commandment of total devotion. So it can only be contradictory to recognize a being as perfect in the religious sense, yet as existing contingently.
(c) One critic (Allen) grants that within religious language it is scarcely permissible to speak of the nonexistence of deity. And he suggests that one might argue that, outside such language, the question lacks meaning and cannot even be discussed. But then he doubts if this proves anything, since we have to have rules for dealing with languages in general, or with language as such. With this last point I agree. And it is possible, as we have seen (Sec. 16) to set up rules for the ontological argument.
Those who say that ‘existence’, taken on the one hand as a predicate, and on the other as not a predicate, can have no true analogy or common meaning should also, by parity of reasoning, deny that ‘truth’, as analytic and as synthetic, can have common meaning. The parallel is very close. Truth in the one case is a relation of correspondence between language and extralinguistic fact, in the other an intrinsic property of a portion of language itself. For this reason some
[114]
philosophers do indeed reject ‘necessary truth’, and offer ‘validity’ instead. Certainly there is an important difference between the two sorts of correctness of statements. But nevertheless ‘correctness’ is common to them, and similarly ‘actualized somehow’ is common to necessary and contingent existence. Both forms of existence imply ‘capable of being actualized otherwise’, but only in the contingent form is ‘capable of being unactualized’ also implied.
What logical difficulty is there in this? Factual truth and validity have this, among other things, in common, that both valid and true propositions are always mutually compatible (both among themselves and with each other). And indeed, valid propositions are virtually contained in every true proposition (also in every false one), and are a part of what is affirmed in them. Similarly, the divine individuality or existence is in some way contained in every individual, afirmed when it is affirmed, but not denied when it is denied. The critics of the Proof must, it seems, reject this analogy altogether. And so far, they do not show any awareness that this is what they are doing. The fault lies chiefly with theologians, or with theists who have failed to furnish a reasonable analysis of ‘necessity’ which connects it with logical rules exhibiting the relation between necessity in language systems and in reality.
Analytic truth in an extensional or interpreted language is also a sort of correspondence, for it corresponds to every possible object. It is the universally tolerant form of correspondence, just as analytic falsity is the universally intolerant form. The divine existence alone is compatible with any other existence you please, including that of any genuinely conceivable devil.
[115]
(d) The plausibility of the Proof is said to depend upon confusing necessity de re and necessity de dictu, necessity in things and in discourse (Allen). Thus, to exist, God might need to have the property, existing necessarily; but still, the assertion that anything has this property remains contingent. Yet it is arguable that the relation between the two meanings of necessary is closer than is here seen. What can be meant by ‘necessary, if it exists’? The intelligible meaning for objective or extralinguistic necessity is, ‘realized no matter what possibility is actualized’. Thus a “necessary thing” is one required by every possible thing, and this can be so only if it is somehow constitutive of everything. If it be thus universally constitutive, then any language which permits us to deny the statement asserting it allows us by implication to deny simultaneously every positive assertion. And this can only be a defective language. Where necessity is meant, as the Proof intends, unconditionally, there the de re and de dictu forms of necessity should be taken as coextensive. See p. 122(1).
(e) Someone (I do not now find who) has contended, against Malcolm, that the meaning of existence must remain the same, namely contingent existence, existence as not a predicate, no matter what sort of evidence is relevant to establishing it. Thus ‘round-squares do not factually exist’ is necessarily true, even though it is known by analysis, not factual observation. But this overlooks the point that contradictions not only cannot be factually true, they also cannot be necessarily true. They are necessarily neither factual nor necessary. Thus the meaning of ‘The contradictory cannot exist’ takes ‘exist’ in a modally neutral sense, not in the sense of ‘contingently exist’. This modally neutral sense
[116]
is precisely that in which the second ‘exists’ in ‘perfection necessarily exists, therefore it exists’, takes ‘exists’. And the meaning of the whole sentence is, perfecuon is necessarily actualized, therefore it is in truth actualized, ie., present in some concrete actuality or other; or equivalently: necessarily (and therefore truly) some actuality is perfect (possesses adequate knowledge of the totality of the actual, and views it with ideally appropriate appreciation, etc.). ‘Some actuality’ has here a neutral meaning, implying an instance of concreteness as such or in general, which instance is, in addition, an example of the property, ‘belonging to the life of a perfect individual, than which none could be greater (except itself), and which could not be unactualized’. Where is there any logical confusion here?
(f) Is existence ever a ‘real’ predicate? Mr. Allen’s rendering of this Kantian expression is, a term which predicates of a thing more than that it is a (concrete) thing. ‘Actuality’ is surely no such predicate, for it can only say that something actual is actual. But ‘actualizes perfection, which in any possible state of affairs is actualized somehow’, is a real predicate, for only divine actualities can possess it. And it is inherent in the meaning of divine perfection that there must be some such actualities, correlative to whatever nondivine things exist.
(g) A necessary god, it is held, must be extremely abstract, and so not the God of religion (Abelson). But, since the abstraction ‘necessarily actualized somehow’, means, necessarily concretized somehow, and since any concrete state embodying perfection must be worthy of worship, the God proved to exist is indeed the God of religion, but the God of the religion of any and every being capable of worship. One’s
[117]
own religion and one’s own God involve personal elements extrinsic to every general proof or argument. But it does follow from the sense of the Proof that there must be some such personal elements in every case.
(h) Henle contends in an ingenious way that there is (except on certain Neoplatonic assumptions) no contradiction in combining necessary existence with perfection ‘in some respect only’ (supposing that there is none in combining it with perfection in all respects). It is clear from what has been said above that this is incorrect. Only that can be necessary which is infinitely tolerant, in 4/l essential respects, of alternative possibilities, absolutely free from competitiveness in its requirements for existence. Imperfection (even in some respect only) is always competitive, this degree instead of that, or my imperfect aims prevailing instead of someone else’s. God's essential role alone collides with no other. Coincidence with modality itself, flexibility as unbounded as all positive logical possibility, this is the prerequisite if existence is to avoid selecting among possibilities. The assumption here is hardly the ‘Neoplatonism’ to which Henle refers; for it is not the classical idea of perfection at all.
I am surprised that so thoughtful a writer as Henle could think it significant that Anselm did not bother to explain why Gaunilo’s perfect island analogy is irrelevant. What Anselm saw was that God exists necessarily, not because He is perfect in His kind, but because His perfection is His ‘kind’, and because, being individually unique, divine perfection is not in the usual sense a kind at all. On the contrary, ‘island’ is such a kind, and because it is, ‘island uniquely perfect as such’ lacks definite and consistent meaning. Individuals of specific kind, such as islands, have to be identified by rela-
[118]
tively specific descriptions, and finally by pointing to some individual taken as given; hence they are all alike wholly contingent. God escapes such sheer contingency because His identifying characteristic is as universally relevant, and in this sense as abstract, as any idea whatever.
The only sense in which the comparison of one individual with another can point the direction toward God’s necessity is this: some individuals have greater flexibility, greater capacity for preserving significant self-identity through wide changes than others. Hence their identifying traits can be less specific. But to transcend contingency, the traits must be absolutely nonspecific,c and only God has such traits. As for an island, it is scarcely an individual at all; for its identity through change is chiefly a fact for the observer, rather than for the island as a whole, or for the molecules and the like which constitute it. Only by stressing the combination of intrinsic self-identity with flexibility, as a mark of the superiority of (say) high-grade human beings to low-grade, or to any island (unless it be thought of as conscious), can we get any very direct light on the meaning of necessity as inherent in perfection. If the island is conceived as conscious, and this consciousness as individually perfect, unsurpassable, then it will be equivalent to God, save with the addendum, ‘composed of earth or rock, or something of the sort, and surrounded by liquid’. And the combination of the two ideas is not obviously consistent. How can an argument be refuted by showing that a vaguely parallel argument from a doubtfully consistent, or even rather obviously self-contradictory, premise would arrive at a doubtfully tenable conclusion? No mode of reasoning could be safe, under such conditions! Only if the credendals of ‘divine perfection’ to con-
[119]
stitute a coherent conception are at least far better than those of ‘perfect island’ has the argument any cogency. But its credentials are incomparably better. For one thing, we know of innumerable believers in God, and no well-documented instance of a single individual who ever believed in such an island, or gave evidence of seriously thinking it could be believed in. Is this because the islands we know have defects or limitations? Rather it is because simply being an island is already a vast limitation, so that to get rid of defects we must in thought get rid of the character of being an island, or anything much like it.
(i) In spite of Anselm and Malcolm, an eternal reality— it is said—may yet be contingent (Penelhum, Plantinga). We are even told that “for all we know” electrons may always have existed. But physics denies persistent self-identity to electrons! Besides, since there is no conceivable positive empirical test for the capacity to exist without having come into being, nor yet for existing everlastingly, only logical analysis could show this, and according to my understanding, only necessity, the impossibility of not existing, could give support to such notions.
It is worth dwelling for a moment on this matter. Becoming is the pervasive principle of reality; anything which does not become must have some very unusual features indeed, and so must anything which endures forever. The controversy over evolutionism arose partly from the failure on the part of many to see that the burden of proof is upon the assertion of ultimate constancy, not inconstancy. What keeps species stable is at least as much a problem as what makes them change. Natural selection is no less powerful as a conservative and stabilizing force than as a destructive or creative
[120]
one. But in the end, nothing is simply and always stable, except that which could not possibly be otherwise. To survive change (as an individual or species) requires a positive capacity; to transcend all change (by preceding and surviving each and every change) requires absolutely infinite capacity to resist or assimilate change. And absolutely infinite capacity cannot just happen to exist, for it must be the principle of possibility itself; and since ‘principle of its own possible nonexistence’ is nonsense, there can be no such possibility. True, these points may seem less than self-evident. But are their denials any more truly self-evident? I am only urging that the Ontological point of view is no mere sophistry, but an important perspective upon the whole of metaphysics.
One difference between philosophers is in what they expect to understand and how much they want to understand it. When told that an ungenerated and everlasting being might yet exist (or fail to exist) contingently, some of us ask, what understanding could one have of such contingency? The ordinary sources of intelligibility for ‘contingent’ are totally lacking in this alleged case. On a few (to some of us very reasonable) assumptions about becoming or creativity, viewed as the perpetual resolution of the partial indeterminacies constituting the future as such, one can see why all things which come to be must do so contingently. Even the contingency of the laws of nature, as Boutroux showed long ago, can be illuminated in this way. But this mode of rendering contingency intelligible fails utterly with an ungenerated and immortal being. What other mode of explicating contingency have we? So far as I can see, none; we have as sole resource the complete inhibition of all curiosity in the matter. Something might exist always and
[121]
forever which might also forever fail to exist—period. Ask no further questions. No prior indeterminacy would have been resolved, no choice would have been made, no dice thrown, no random process of nature eventuated thus and thus—no! Merely, it is, perhaps, so, and yet it might always have been otherwise. Is there so much instruction in bare words? They remain merely words, and therefore of no real use, untii we can connect the supposed ideas with other ideas in some intelligible way. I suggest, therefore, that if Henle had thought with more acute and persistent curiosity about contingency he would have been able to carry further his generous concession that there may be a good meaning in ‘necessary existence’. It is that meaning which forms one end of the triple disjunction by which contingent is to be understood, the disjunction: totally exclusive, partially exclusive, totally nonexclusive, of (posidve) possibilities; or the disjunction: completely intolerant, partially tolerant, completely tolerant, of all specific possibilities. Complete tolerance is the same as absolute adaptability, unlimited power to preserve self-identity and integrity in response to any world whatever. And this is again only a way of looking at cognitive infallibility. Only knowledge can take a form correlative to anything and everything (the strictly unknowable being the same as nonentity). But only states of perfect knowledge can fully actualize this principle.
For the reasons sketched, should we not side with Anselm in thinking that only the necessary and perfect can be eternal?
(j) One critic appeared to think that a consequence of this—that nothing eternal can be created—is absurd. But why? True, ‘the world’, in its most general possible meaning,
[122]
can be eternal; but in this meaning it is not a particular or individual creature, but merely the bare eternal truth that there are creatures, none first or last in time. And this is an eternal truth only if it be necessary, that is, if God must create without a temporal beginning or end of His creation, or if He could not produce a first or last creature. (This contradicts His perfection of power only if some other being could conceivably produce such a creature. But Aristotle may have been right in thinking that an absolute first or last is here unmeaning. )
(k) It is contended that independence of existence is not uniquely good (Abelson). For bare existence substitute ‘actuality’, or how the predicate exists, and Abelson’s point is an understatement. The divine actuality must be the most sensitive to influence, and in that sense the most dependent, of all. But for its being somehow actualized, no matter how, divinity must be conceived as wholly independent, and ¢hss independence is entirely a merit. Maximal independence and maximal dependence, the one for abstract, and the other for concrete, aspects of individual existence, these are essential to the worshipful perfection of deity.
(l) A shrewd point is made by Matthews when he notes that Anselm puts God above anything we can conceive, so that if, as Malcolm does, one identifies ‘conceivable’ and ‘logically possible’, then God must be logically impossible. The answer is that God is above the humanly conceivable, but not above all conception, including His own. In ‘nothing greater is conceivable’, however, there is no reference to human conceiving; even God Himself cannot conceive another greater than Himself. Only this second, unrestricted use of ‘conceivable’ is equivalent to logical possibility.
[123]
There is, nevertheless, a danger of injustice to Anselm in substituting for his ‘not conceivable as not existing’ the Cartesian ‘necessarily existent’. For then the query seems to arise, are we here dealing with necessity only de re or also de dictu? But if the nonexistence of deity is strictly inconceivable, it cannot be legitimate to assert it, and atheism is ruled out. The strictly inconceivable cannot be said to be even possibly true.
(m) A sixth recent critic of the Ontological Proof assumes that by perfection we have to mean an actuality uniquely defined by its ‘completeness’.?17 Of course it is easily shown that this leads to difficulties. (See Part Two, Sec. 8.) But we do not need to identify perfection with completeness in this sense. Perfection is modal completeness, not actual completeness. No individual actuality can be absolutely complete (exhaustive of possibility), though it can be reladvely so, ie., inclusive of all that is in fact actual, and this by necessity, or in any possible state of affairs. Individual potentiality, however, can be absolutely complete, ie., inclusive of all that is logically possible, qua possible. God's modal completeness means that He could have, not that He does have, any predicate or compossible set of predicates whatever. In this is no impossibility. (Of course, God does not have the anger of the angry man as the man has it, but He has all its actual value, by intuiting with absolute adequacy the man's experience, and that of those at whom the man is angry. He thus has all, in the way of value, which the man has, and incomparably more. And so with other
17 Leslie Armour, “The Ontological Argument and the Concepts of Completeness and Selection,” The Review of Metaphysics, 15 (1961), 280-91.
[124]
actualized predicates.) Our critic has merely resolved, without noticing it, the ambiguity in ‘none could be greater’ in favor of the interpretation, ‘even itself in another state’, rather than in favor of, ‘except itself in another state’. The latter interpretation avoids the difhiculties to which he points, and actually strengthens and clarifies the connection with necessity. Thus he proves nothing against our version of the Proof.
(n) Still another opponent of Anselm (See Bibliography, Zabeeh) reminds us that if the premises of the Ontological Argument are tautological rather than factual, that is, if they are “true in all possible worlds, then . . . they are vacuously true, since they then tell us nothing particular about our world.” He fails to note that by ‘the existence of God’ is not properly intended a particular feature of ‘our’ or ‘the real world’, but only the following purely general status of any world whatever: if that world exists God created it, and if it does not but could exist, God could create it. Naturally, then, ‘God exists’ is true in any possible world; this is one of the many ways of putting Anselm’s point. The divine existence cannot stand or fall with that of a certain world, a certain contingent thing, for this would mean that the creator was a mere creature! God's mere existence is not, in the proper sense, ‘particular’ at all, though His actuality is indeed so.
Our author is right in this: the essential question is whether the idea of God, or that of necessary, uncreated, yet individual, existence, makes sense. Anselm frankly assumed that it did. Against the positivistic contention to the contrary a supplementary argument is needed, and here
[125]
Anselm was in trouble. But with a different philosophy the supplementary argument can perhaps be provided.
Professor Zabeeh’s willingness to beg the question he purports to discuss is rather breathtaking. Thus, leaning on Hume: “it is a brute fact that every object or impression which exists, could cease to exist.” We have Hume’s (and Professor Zabeeh's) word for it. However, I have read Hume, and I find no evidence for the assertion, just the assertion itself (nor does Kant help here). Again, leaning on Ryle: “since things do happen to wear out, it is rational to expect them to wither away . . .” And so the weakness of the creatures is ohne weiteres attributable also to the creator! By what rule of logic? It cannot be God who is here referred to!
It is said, rightly enough, that the mere use of a word cannot establish a necessity. However, as Findlay (and others) have shown, it is the meaning not just of ‘God’ but of the attitude of worship, that it refers to a not conceivably surpassable being. Since no such being could be a contingently existing thing, i.e., a mere creature, either worship is selfcontradictory (rather than merely mistaken), or its object exists necessarily.
Still another attempt to throw new light upon the ancient controversy is Nicholas Rescher’'s argument (see Bibliography) that the real point of the Proof is this: to understand blue one must have experienced blue and insofar must know that there is (or has been) such a thing; similarly, to understand ‘God’ one must have had religious experience, but then one will also know the reality of the object of the experience. Rescher's proposal implies that deity, like redness, is a special sort of thing, to know which requires a special sort of experience. As though the universally im-
[126]
manent were only in certain corners of reality or experience, rather than everywhere! Also the author suggests that the definition from which the old proof started was insignificant from the standpoint of religion. But the definition grew out of religious experience as such, or as worship. God is defined as worshipful because beyond all possibility of rivalry. Yet this definition (when the false interpretation of ‘none other’ is corrected) has an intelligible content which presupposes no special sort of feeling or sensation. Willingness to consider the definition on its merits, to genuinely inquire into its logical implications, does require that one should no¢ be in the emotional state in which many philosophers find themselves. But the analogy with color sense is about as misleading as it well could be. No willingness can help the colorblind man to experience color; but sufficient willingness may enable anyone endowed with intellectual powers to see some point in the modal form of the ontological proof. To appreciate the full value of the Proof is a religious experience, but to see that it offers a legitimate, and not merely silly, line of inquiry it is only requisite that one think with honesty and scholarly care about the matter, going behind the textbooks to sources in writers who have themselves really meditated upon the problem, and not merely copied others, and noting how these writers have dealt with the objections which have become standard. Professor Rescher does not show any sign of realizing that, precisely for religion, God is a principle and not a mere fact, the contrast between the creator and the creatures being indeed a supreme abstraction for all religious persons. The more abstract and universal a factor of reality, the less must one have this experience rather than that to be able to think the factor. And
[127]
the bare existence of deity is strictly universal, hence any experience will do, if it is sufficiently reflective and unconfused concerning its own content. What Anselm proved and meant to prove is that, assuming that the man of faith understands himself, the Fool deceives himself; not as the color-blind man by being simply unable to encounter the thing, but by being unwilling, or emotionally unable to try, to think a certain concept. God in His bare existence is not a mere special quality, but a Universal of universals, the Form of forms. He is infinitely more than this, true enough, but He is not less than this. And this suffices for His necessary existence. Even Thomas admits that all creatures in some sense know their Good, and that this is God. But man reflects upon his good; he cannot be simply unable to know it. The problem is one of self-understanding on the deepest level. The color-blind man can understand himself perfectly on that level.
One final attempt to discredit Anselm’s argument as reformulated by Malcolm runs as follows.18 Malcolm gives the proof in the following steps: (1) If God does not exist His existence is impossible; (2) if He does exist His existence is necessary; (3) His existence is either impossible or necessary; (4) His existence is not impossible; (5) hence it is necessary. Huggett objects: since Malcolm derives (1) from (a) ‘If God does not exist, He cannot come into existence’; and (2) from (b) ‘If He does exist He cannot have come into existence, nor can He cease to exist’; it follows that (3) has no more force than the disjunction: either (p) ‘God is nonexistent and cannot come into existence’, or (q) ‘He exists and
18W/. J. Huggett, “The Nonexistence of Ontological Arguments,” Phslosophscal Review, 71 (1962), 377-79.
[128]
cannot have come into existence or go out of existence’. But then one cannot refute (p) by arguing that God'’s existence is at least conceivable, for (p) says only that coming into existence, not that existence, is inconceivable of God. However, the argument has here been oversimplified. Huggett is assuming that what does not exist and could not come into existence nevertheless might always exist, need not be logically incapable of existing always. But this implies an ultimately total divorce between modal and temporal status which an Anselmian ought to reject. What does ‘might’ mean in nontemporal terms? The intelligible reference is to possible happenings, whose antecedent conditions—at least if we go back far enough—do not uniquely determine their specific outcomes. A critic of Anselm needs to show that contingent existence is intelligible in some terms other than those of coming to exist and ceasing to exist. I am not aware that this can be done. A concept whose representation in reality is contingent is one which can be conceived as realized and also as unrealized; but if both these conditions are conceivable, how can a transition from one to the other be inconceivable? You may answer, because the thing is defined as eternal Yes, but this amounts to conceiving two alternative states of ‘eternity’, and saying that it is logically mere chance, arbitrary fact, which one is realized. But is it not in time that accidents, contingencies, can occur? The theory which the anti-Anselmian is implicitly adopting is that of happenings, accidents, in eternity. The difficulty is not merely etymological. We have no experience of contingency in nontemporal form,; the notion is at best wholly controversial. Save to refute Anselm, it performs no function that I at least can see. Hence it
[129]
is question-begging. Known contingency is confined to things capable of a genesis.19
The very meaning of ‘eternal’ is dubious apart from necessity. Without necessity, nothing could establish the least probability of eternity; one would have to wait, so to speak, forever to see if the thing did always go on existing. Nay more, since the future is potential rather than actual, even God Himself could not know that He would never cease to exist, except by knowing this to be impossible; and how He could at the same time know that His nonexistence was (is?) eternally conceivable is utterly beyond even the vaguest imagining.
Someone may wonder if the foregoing commits us to the view that either the world has had a beginning or it exists necessarily. It does so commit us, but we must not overlook the ambiguity of ‘the world’. Either it means the universe in something like its present state, or at least with the same or similar laws, or it means some universe or other (created and nondivine reality) with no matter what laws. The former is reasonably viewed as contingent and also as having had a genesis; the latter I suspect is necessary and has neither come into being nor can cease to be. It is not, however, a rival to God, for it is not save in the vaguest sense an individual, but only the exceedingly abstract class, nondivine individuals in general. The rivalry which the definition of God excludes is with individuals. Besides, the universe is always wholly embraced and surpassed by the divine knowledge. Its eternity only means that God can never lack a universe to know. Anselm has other ways than the one just considered of
19 C, Harcshorne, "Real Possibility,” The Journal of Philosophy, 60, No. 21 (1963), 593-605.
[130]
discriminating contingent from noncontingent conceptions. Thus if a concept requires a certain composition of parts, how could this be necessary, why not some other parts, in some other arrangement, instead of the specified composition? What Anselm was groping for but did not entirely attain was a theory of arbitrary, exclusive, or competitive conceptions as distinguished from those wholly without arbitrary, exclusive, or competitive features. Becoming as essentially involving real additions to the definiteness of reality must be arbitrary or competitive (why not some other addition?). So with any particular arrangements of parts. So also with any particular degree of imperfection (and all imperfection is a matter of degree). Imperfect things compete for existence, they get in each other’s way, to some extent, and their precise degree of merit is always accidental. Their partial intolerance for other forms of existence is one of the implications of their lack of perfection. But absolute perfection (unsurpassability by another) is no arbitrary degree of value, but the measure of all degrees; and its absolute capacity for dealing with others enables it to be the wholly tolerant, noncompetitive accompaniment of all competitive existence (including God Himself in His contingent or competitive aspects).
Huggett's little essay is one more in the long line of melancholy examples of how eager people are to refute Anselm’s reasoning, rather than to learn from it.
None of the above-mentioned critics seems to worry about the paradox of a God to whom existence can at best, as they seem to suspect, be imputed by sheer faith, but whose allegedly possible nonexistence must even more definitely remain a matter of sheer unfaith, or negative faith. Obviously God could not know His own nonexistence, and if no one else could
[131]
either, then it is simply unknowable. It has been proved above (Sec. 15) that an imperfect mind could know the existence, but not the donexistence, of a perfect mind.
Nor do any of the critics seem to have a clear notion of why existence is in general contingent, unless it is merely because predicates are abstract or indeterminate and existence is concrete and definite. But it is only actuality which is fully concrete, only the particular how or state of concretization; the mere that is always at least relatively abstract, and in the supreme case before us, it is infinitely abstract, and indeed as abstract as genuine or positive possibility itself. Hence the type distinction is here entirely concentrated in the gap between the that and the how. Like so many dichotomies, that of ‘essence and existence’ is simple-minded, and obscures a real distinction. ‘Existence’ is merely a relation of exemplification which actuality (any suitable actuality) has to essence. Thus I exist if my identifying personality traits (gene structure, the property, first-born son of— and—, ot what you will) are somehow embodied in actual events, no matter which. Upon the legitimacy of applying an analogous, but even more radical, distinction to the idea of God depends the possibility of an ontological proof. Most theists have not been in a position to make this application. The central difficulty with the proof, the partly unconscious ground of opposition to it, and probably even of the positivistic denial of the concept of God itself, is in the oversimplified form of this concept presented by classical or neoplatonic theism.
Philosophers should, at long last, give due heed to the manifest difference between existence, the mere abstract truth that an abstraction is somehow concretely embodied, and the actuality, the how, of the embodiment. The ignoring
[132]
of this duality in nearly all discussions of the ontological problem is a marvelous instance of how even centuries of prolonged controversy, involving almost an entire learned world, can still leave a point unnoticed by anyone. The possibility of such collective blindness helps to make intellectual life exciting. There is always a chance of seeing clearly for the first time what implicitly many have been looking for.
The failure to distinguish between existence and actuality seems also responsible for the notion that one might collapse Anselm’s second argument into his first, and refute Prosl. 111 by refuting Prosl. II. The contention that the second argument presupposes the first has been strongly put in Germany by Henrich (see Bibliography), who, in his detailed and, in some respects, admirably careful analysis of the fortunes of the Argument in modern philosophy, stresses the difficulty of explicating ‘necessarily exists’ unless it means that (simple) existence is contained in the essence, and hence derivable a priori. But then the first argument should be sound, since simple existence is being treated as an ateribute. What is missed is the distinction between being actualized somehow and being actualized precisely thus and thus. The latter can never be an attribute, but must be a concrete instance of the attribute, hence contingent; the former can in principle, or in one privileged form, be an attribute, the attribute ‘exemplified in any possible case’. This is noncompetitive exemplification. Competitiveness is clearly an attribute (hypothetical dollars are hypothetical competitors, actual dollars actual competitors), and therefore noncompetitiveness, if there can be such a thing, is also an attribute. But, since it is competitiveness which gives mean-
[133]
ing to nonexistence, the noncompetitive cannot not exist; it must be actualized somehow.
Henrich wrestles valiantly with the apparent abolition in the Argument of the categorial difference between a predicate and its realization, in other words the abstractconctete paradox, and reaches the conclusion that the problem has neither been solved nor definitively shown to be insoluble. He is to be congratulated, for he is right on both counts, so far as the philosophers he discusses are concerned. He also lays down three conditions which a philosophy must meet to have the right to employ the Argument. It must have an idea of divine perfection; it must be aware of the radical difference in general between essence and existence; and it must be able to admit a spiritual reality (an ‘eidos’) quite independent of any act of thinking it. That a philosophy exists today which meets these conditions is a fact apparently unknown to this German author, even though, long before Whitehead and certain other non-Germans, some largely forgotten German thinkers (e.g., Fechner) could have met them.
It is strange that the author’s otherwise magnificent erudition has not saved him from the traditional fault of limiting Anselm’s reasoning to that of Prosl. II. For he imputes only the first argument, from simple existence, to Anselm, while the second, from necessary existence as a perfection, he treats as a post-Anselmian achievement! Thus the Gaunilo legend for the thousandth time poses as a piece of scholarship. Does perfection not entail infinite existential tolerance or adaptability (the infallible can know anything whatever), hence absolute noncompetitiveness, hence the exclusion of nonexistence as meaningless or contradictory? If so, the
[134]
Argument is valid. Yet it takes no step in the direction of describing how, or in what actual concrete form, perfection is realized. The categorial or logical gap between essence and existence, which Henrich stresses, in the divine case becomes an absolutely infinite gap between abstract essence and /or existence, on one side, and concrete actuality on the other.
The trust in mere dichotomies was effectively criticized by Peirce seven or eight decades ago. But the lesson has yet to be learned. The ‘cultural lag’ in philosophy is great indeed. (Thus Henrich in 1960 does not know Prosl. III.) ‘Existence’ can have various positions short of full concreteness or actuality, but in one supreme case it must be as abstract as essence, and hence a priori. The comprehensive or universal contrast of idea or essence is with actuality, not with existence.
Against these authors I maintain: Anselm discovered, and really discovered, the modal uniqueness of the idea of God. What he overlooked, and nearly all his critics equally fail to see, is that, since actuality cannot be necessary, there must be a real duality in God, as in no other being, between necessary existence and contingent actuality.
24. The Proof and the Other Theistic Arguments
There is a final consideration. Philosophers all know that according to Kant the other supposed proofs for God fail unless they can resort to the principle of the ontological proof (that perfection and necessary existence are inseparable). Kant is believed to have refuted the principle, hence to have disposed of all the proofs in one blow. But did he refute the principle? True, he also accused the proofs of other weaknesses.
[135]
But in view of the disreputable history of the criticisms of Anselm, can we trust the criticisms of the other proofs? Admittedly, there is no one locus classicus to be compared to the neglected and scorned Prosl. III and Reply to Gaunilo (though the 10th Book of the Laws has perhaps some title to this honor, and it, too, is underestimated). Still, similar prejudices are involved, such as the belief that theism must mean either classical theism or classical pantheism, or the empiricist notion that a proof for God must be a posteriori. From the empiricist assumption it was deduced—essentially a priori and without any empirical nonsense of reading texts— not only that Anselm’s proof was invalid, but also that the only hope of a proof other than the ontological would be in an inference from some sort of factual observation. In reality, the ontological is not the uniquely a priori proof; it is but one such proof. All the proofs, properly stated, proceed from ideas; but not all from the idea of God itself. And all show that we must either admit some basic idea to be absurd, or take it to be necessarily true, and admit also that this truth entails the necessary existence of the Greatest being. Unbelief is confusion or else belief is confusion. There is no third possibility.
Belief here of course means only belief in God. No special doctrine of any church or group is involved, unless it, or its denial, is deducible from the conceivability of Greatness. Such matters as Anselm’s special beliefs about the Church, the Trinity, and the Incarnation are at best subsidiary to the main religious belief, belief in a God who is unsurpassably all that it is better to be than not to be, just so far as such unsurpassability is logically possible, and who is surpassable exclusively by Himself, so far as this mode of
[136]
exclusive surpassability is possible. Beyond that, He is what He is, and there is little we can know about it except that He knows, and that what He knows is all truth. In what sense does He know? In just the sense in which it is better to know than not to know: infallibly, because this is possible and best, yet not completely and once for all with respect to actualities, because this is contradictory, possibilities for actualization being inexhaustible and therefore never exhausted, even by deity.
To the Anselmian theory that Christ must suffer so that man's offense against God could be pardoned without injury to the divine honor or justice we may object that an author who insists that the divine nature is for us ineffable and full of paradoxes ought not lightly to assume that the values which motivate God in His treatment of man will agree so precisely with the transcendentalized legalistic notions of a certain human society. Medieval notions of ‘honor’ and ‘justice’ may be absurdly inadequate or irrelevant to the cosmic and everlasting. God (in His necessary aspect) is ‘whatever it is better to be than not to be’: perhaps it is not better to be so concerned with legalities and prestige as the deity of Cur Deus Homo is represented as being.
Thus a weakness in Anselm’s thinking is in the ‘human-all-too-human’ valuations on the basis of which he applies the above-quoted formula. The formula, however, transcends these applications and may, if properly construed, have independent validity.