Charles Hartshorne "Metaphysics and the Modality of Existential Judgments" (1961)

First published in The Relevance of Whitehead, Ivor Leclerc, ed. London: George Allen & Unwin, 1961, pp. 107-121.

Metaphysics and the Modality of Existential Judgments

Charles Hartshorne

What is metaphysics? One may use the word variously. I use it to refer to the study which seeks to clarify our conceptions of the absolute, necessary, strictly universal, infinite, or perfect. These terms are not synonymous; but they all have this in common, that they transcend empirical evidence, in any normal meaning of “empirical.” No experiment can show that there is, or is not, anything absolute, necessary, strictly universal, infinite, or perfect. That we find no exceptions to a law establishes not the slightest rational presumption that none ever will be found by any cognitive being. Some of the Greeks took the stars to be absolute in relation to all lesser things, that is, wholly unaltered by changing relations to these things; but this they could not possibly have had cogent empirical grounds for doing. Nor, I should suppose, could cosmology as an empirical science ever prove, or give any reason for holding, that the cosmos is infinite in space. Or again, we may have evidence that a certain thing is a “perfect” specimen, in some loose sense. However, “perfect” in the strict metaphysical sense means not merely, as measured by this or that concept or standard, but (in some way), by any legitimate concept or standard. It means good, not simply for this or that purpose, but – in whatever way, if any, this can be so –for every possible purpose. No such thing could be empirically detected. The mystic may perhaps experience the Perfect reality, but he cannot “observe” it, in the sense of an empirical science.

The denial of metaphysics consists simply in the assertion that everything so far as knowable (or even conceivable?) is relative, contingent, finite, imperfect, and less than strictly universal.

Metaphysics is distinctive both in subject-matter and in method. The absolute, infinite, necessary, or perfect must be very extraordinary indeed; and this manner in which such a thing could be known to exist, or not to exist, or even to be unknowable or nonsensical, may be expected to reflect this uniqueness. To proclaim a flat monism of intellectual method applied to both this problem and ordinary ones is to try to settle the issue by fiat and definition.

There are two seldom-noticed ambiguities in the foregoing account of metaphysics. The confusions which have clustered about this subject are in no small part due to these ambiguities.

(a) It is one thing to speak of a concrete particular thing as necessary and another to say that there is an idea, property, or kind of thing which must necessarily be exemplified in some suitable concrete particular thing or other, no matter what so long as it exemplifies the idea or property.

In extensional languages, the concept of “individual,” or lowest-level object, must be exemplified; but this does not mean that any particular such object is taken to exist necessarily. We have rather a class which is not allowed to be empty, though the members of it which exist might have been replaced by others. The class of “contingent things” is taken to be necessarily non-empty. There is, I maintain, no contradiction in this.

The same sort of distinction applies to the characterization “absolute.” The validity of the concept of “relative thing” (thing dependent upon certain relations, which might not have obtained) need not itself be relative, but may in a strict sense be absolute. Thus perhaps the truth, “There are relative things,” obtains, not relatively to this or that, but in any possible case and relative to nothing, or (the same) to anything you please. “Nothing is absolute except relativity” might, taken in this sense, be close to the last word concerning both the absolute and the relative.

In some such way the long strife between “absolutists,” or believers in “the absolute,” and “relativists,” or believers in the proposition, “There are no absolutes,” might be resolved. Again, with respect to finite and infinite the same ambiguity appears. Are we to think of some actual infinite thing, or is infinity rather a function of the finite as such, and in this manner perhaps: the totality of conceivable finite realities is precisely and absolutely infinite? What else could it be? Finitude is the cutting off of possible finite things. To be just this length is not to be just that length; every limit is alternative to various other limits; and thus the totality of possible limits can only be rigorously unlimited.

(b) There is another distinction we need to make. It is one thing to say necessary, infinite, or absolute, and another to say Perfect. “Perfect” is a value-term, and we have learned in recent discussions how unsafe it may be to confuse such with neutral terms. If by “necessary” we mean, “in all respects non-contingent,” this is not synonymous with “Perfect,” unless every form of contingency is to be viewed as a disvalue. This is a further question, needing consideration on its merits. Again, “infinite” may be taken as equivalent to “Perfect” only if every form of limit is regarded as bad, a disvalue. But what about the conviction of the Greeks, and perhaps of every artist, that the good is something like the beautiful, and that the merely unlimited is the formless, hence the unbeautiful? Beauty is harmony in contrast; but in sheer infinity there are no definite contrasts and no definite harmonies; there is only the totality of all possible contrasts and harmonies, which is the same as sheer confusion, if taken as actualized. To create beauty is to select, to renounce some possibilities in favour of others. God creating a world must, it seems, do this very thing, and hence he must also select among possible roles for himself as creator and lover of his world: shall he create and love as actual this world or that instead? If then God is Perfect as Creator, he is not in every respect infinite; for if he were he would be the mere totality of his possible roles as creator, but not any actual creator at all. And if you say that he would then be no less good than he is, you are saying that the world adds no value whatever to reality, and that from the definitive or divine point of view it is exactly the same, in terms of value, whatever we do or fail to do. Perhaps (I hope not!) this is good Hinduism, but that it should ever have passed for Christianity is one of the curiosities of human history. A God who “loves” the world cannot be merely indifferent to its existence, and this is true, no matter how “analogical” the term “love” may be here. There is no analogy whatever, on the assumption. To love x is at least to prefer x to nothing at all.

My conclusion is the “Perfect,” in the religiously relevant sense, cannot be identified with “in all respects infinite, absolute, or necessary.” Rather we should say, somewhat as we did under (a), that the Perfection of God must be the absoluteness of his relativity. His essential or generic manner of being relative, in contrast to our way of being relative, is itself non-relative, independent of all things; and this, not his mere absoluteness, is his Perfection. (To see a contradiction here is to overlook the distinction of logical levels involved.) Again, God’s infinity must be something like the infinity of his possible finite roles, not of his actual role. All other individuals are finite, not only in their actual, but also in their possible roles. God could relate himself to any conceivable world; whereas in some worlds you and I would be nothing. However, not even God can relate himself to all conceivable worlds as actualities. For not all are actualities.

Let us now apply these considerations to the main topic of this essay, the modality of existential judgments.

It is a commonplace in logic that ordinary existential statements are contingent. Thus that there is something with the character of being an elephant is true merely in fact, not by any logical necessity. This contingency of ordinary existential statements is by many writers generalized to cover all existential statements. Yet there are good reasons for questioning this generalization. In some logical systems, at least, it is permissible to infer (Ex) fx from (x) fx. In other words, a predicate applying to all objects must apply to some object; or, again, the universe of discourse cannot be empty. The extensional interpretation of language seems to make this requirement inevitable. But it follows that the contingency of ordinary existential statements arises not from their existential reference but from some further characteristic. Moreover, the assertion that thought is contingent if it refers to existence prompts the retort, to what else could it refer? Language? But this is nothing except as a factor in what exists. “Possibilities,” “essence” – ditto. It is existing things or really occurring processes, which make certain other things or processes possible or conceivable. It is instances which alone give universals any foothold in reality – if nothing else, an instance of the universal, “someone thinking of the universal.” It has been said that even mathematics is “the physics of everything,” a study of the forms which the cosmic process might conceivably take.

If contingency does not derive from existential reference alone, from what does it derive? Let us consider what “contingent” means. It means that what is asserted selects among the conceivable possibilities, affirming some and excluding others. To say that some individuals are elephants excludes a world consisting entirely of individuals other than elephants. And such a world is easily conceivable (and doubtless once did exist). If, however, we say, “Some individuals are self-identical,” what do we exclude? We exclude only the case of there being no self-identical individuals, which is the same as no individuals at all. (I use “individuals” here in a sense neutral to the distinction between particular things and particular events.) In what sense is this verbally possible case genuinely conceivable? It is by definition incapable of being known (not only by man, but by any being); for it could not be true that, although there were no individuals, there was an individual knowing this to be so (nor even, that it had “once” been so). Thus whereas “no elephants” might describe a positive and knowable world-state, “no individuals” appears not to describe such a state. And whereas the existence of an elephant excludes from some portion of space-time various otherwise possible kinds of objects, the existence of an individual of wholly unspecified kind does not exclude anything from existing.

It does not exclude universals; for, as Aristotle argued, the universal can exist in individuals, and even those who deny this will contend that the universal exists in its own way or realm, in no fashion interfered with by the reality of individuals. Thus the total emptiness of the call of individual is (i) unknowable and (ii) purely negative, without any positive bearings of any kind.

I suggest that these two traits, unknowability and lack of positive significance, come to the same thing and that what they indicate is lack of consistent meaning. To know is to know something positive, and the knowledge itself must be something positive. Hence any statement which excludes only a mere negative excludes no genuine possibility, and since contingency means selective exclusion of existential possibilities, affirming some and denying others, a statement which excludes only negations cannot be contingent. I propose then, as criterion of non-contingency, the absence of any positive meaning for the denial of a statement, or – the same thing – the failure of the statement to exclude any positive state of affairs. There are individuals excludes no such state; hence it cannot be contingent. Yet it is not meaningless, since in fact there are individuals. Thus the statement is significant, true, and non-contingent. But this is the definition of necessity, in the absolute sense.

There are, then, three modal forms of existential statement: those which contradict every positive existential assertion; those which contradict some positive existential assertions but agree with others; those which contradict no such assertions. In the first case, any conceivable situation would make the statement false, hence it is impossible (there is a contradiction or absurdity somewhere); in the second, some conceivable situations would make the statement false and others would not, hence it is empirical or contingent; in the third case, nothing could make the statement false, hence it is necessary or metaphysically valid.[1]

That the necessary excludes no positive existential assertion is entailed by the common logical dictum that the necessary gives no information about the state of the world. To exclude something from the world is certainly to give such information. But I am proposing that we may convert the rule, “What is necessary is without positive exclusiveness,” so that it runs, “What is without positive exclusiveness is necessary.” If the rule is held not to be convertible, I ask, why not? To me this denial is “counter intuitive.”

We have now to note that although a statement which denies no contingent possibility also affirms no contingent possibility, yet it does not follow that it affirms nothing at all, and “says nothing about the world.” The necessary is indeed compatible with any affirmation you please, but not with any negation you please. Rather it is exclusively positive. It affirms that about the world which would be real no matter what possibilities were actualized, and which therefore cannot be denied except by impossible formulae. If “information” includes reference to the factor which all possible positive states of existence have in common, then necessary statements are informative. And when we refer to such a common aspect of all possibilities, are we talking merely about language? Or has it really been proved that the contrast between what is and what could be is a mere linguistic artifact? If not, then for all that has been shown, the necessary – or that in which all possibilities agree – is also extra-linguistic.

We now take up the most important issue connected with the modal question, the issue of the Ontological Argument. Can a being which is conceived as the source of all contingent being also be conceived as a mere contingent being? If the idea of such a source makes sense at all, it describes the ground of all real possibilities; and the unreality of this ground cannot then be one of the possibilities. The idea of “creator of all things, himself not possibly a creature,” is contradicted if it is taken as a mere possibility which might, or might not, be actualized. For while to be a creature is to be an actualized possibility, to be the universal or ultimate creator is not. There can be no such thing as the possibility of the ultimate creator; there can only be his necessary reality, prior to all contingent alternatives. For it is the Alternativeness itself, the indeterminate-determinable potentiality for possible worlds. So far, at least, the Ontological argument seems to be in the right. Moreover, our test of non-exclusiveness supports the view that the existence of the creator cannot be contingent. A being absolutely unlimited in creative and cognitive capacity cannot, by existing, conflict with the existence of any possible thing; for a state of affairs which the existence of deity would exclude must be a state which deity could not bring about or know. Both the creative power and the cognitive capacity of the divine would thus be conceived as less than perfect! So we infer that whatever the idea of God may mean or fail to mean, it cannot mean a being whose mere existence would be exclusive in any respect. It follows from this, together with our theory of contingency as exclusiveness, that the question of the divine existence does not concern a contingent fact, positive or negative.[2] Deity is logically impossible or necessarily real. Either the divine by existing must exclude everything, or it could exclude nothing; but in the latter case, nothing could make its existence false, while in the former, its existence is a logical impossibility.

It may be argued that the existence of God should at least limit the possibilities of evil in the world, and thus exclude something. Nor would this something be mere negative; since evil is, in some instances, a positive opportunity for heroism, courage, and other good things. And it may be asked if the existence of a deity ideally powerful and good must not exclude this opportunity, by guaranteeing perfect harmony and happiness in all the world. This, however, presupposes that the perfect Will, in creating, would or could choose the details of the world, leaving nothing for created wills to determine. And this, in turn, means that the created will would differ from the supremely creative will, not as imperfect or limited power of choice from perfect or unlimited, but as total helplessness from complete competence, or as zero from infinity. But thereby the very meaning of “choice” or “will” as used in the two cases is destroyed. (Several recent writers on the problem of evil overlook this point.[3]) We can attribute the perfect form of causally-undetermined power of choosing contingent details to deity only because we too have a real, however, imperfect, form of this power (for which a clearer name than “freedom” is “creativity”). That the divine is the ideal form of creativity cannot mean that ours is, or could be, a mere sham. We may have little, or poor, capacity to select among genuinely open alternatives, and God may have much, or in some sense infinite, or all possible, capacity to select among alternatives, but we cannot have none and he all, if our concepts, as used in making the comparison, are to have significance. Accordingly, the supreme Creativity cannot be conceived as an all-determining or monopolistic force, and in so far, its existence cannot guarantee complete harmony. Not that God is “weak,” but that the ideal “strength” is not a monopoly of strength. “All-powerful” properly means, “having power extending to all things,” and able to do for them everything compatible with the element of self-determination without which they would have no individual reality. It follows that the existence of perfection does not exclude the existence of evil, since a multiplicity of (partly) self-determining individuals implies innumerable hazards of conflict and discord. Nor is this argument refuted by pointing out that some evils do not result from human freedom. The argument, fully thought out, is entirely general: the creatures as such must be free.

Nevertheless there is a serious difficultly with the Ontological Argument. For we have seen that the concrete or particular is contingent, while only the universal or (in one sense at least) abstract, e.g., “something exists,” is necessary. And yet it seems that God is thought of as no mere general abstraction, but as a particular and concrete reality, an individual. Indeed, the entire concreteness of the world must be summed up in the divine knowledge. This knowledge must be precisely as particular and exclusive as the world. For if God knows that elephants exist, then this knowledge, being infallible, logically excludes the state of the world free from elephants which otherwise conceivably might have obtained in this geological age. Accordingly, God as including his concrete knowledge is exclusive, and hence, by our principle, contingent.

At this point theologians and philosophical theists, apparently ninety-nine per cent of them, insisted upon what I can only regard as a grave error. They held that not only the existence of deity but also his entire reality were necessary. They refused to make any effective distinction between divine “essence” and divine actuality. But then the Ontological Argument would imply that something concrete, particular, and exclusive was yet necessary, and therefore non-exclusive, and this defies the logic of contingency and necessity altogether. So it was a sound intuition which kept most thinkers, theistic and otherwise, from accepting the Argument. But it is not the Argument which is to be rejected, it is the standard view of God. There is indeed no basis for the dogma that “existence,” in an undetermined sense, cannot be a predicate; but there is a basis for the denial that concrete, particular, exclusive existence can be so. Moreover, as we have seen, divine knowledge that certain contingent things exist is exclusive of alternative possibilities, and hence can only be contingent, like the things. If proposition p is contingent, then so is the knowledge that p is true.; for since p might have been false, there might not have been the knowledge (and could not have been, had p been false). What the possible contradicts cannot be necessary, and “the non-existence of this world” contradicts “God knows that this world exists.”

If we admit, what seems obvious, that any knowledge of the particular, exclusive, and contingent is itself particular, exclusive, and contingent, then it need not and cannot be the full actuality of an all-knowing God which is declared necessary by the Ontological Argument; it can be only his abstract essence, and his bare existence as defined by this essence. God must exist (if deity be logically possible, an assumption I have not tried to justify in this essay), and very likely he must know some world or other perfectly; but he need not have existed knowing this world, for this world need not have existed to be known. The particular actual concrete “state” of the divine knowledge is no more required by the necessity of the divine existence than the particulars of the world are required by the necessity that there be some particulars or other. Not the concrete things, but concreteness as such, is necessary; and not the divine concrete states, but divine concreteness, as such, is what the Argument demonstrates to be non-contingent. God must (assuming his logical possibility) be concrete somehow, but not exactly as he in fact is.

It is an old saying that God is not a member of a class. There is only one God, and he is not a merely possible or contingent God, but rather the one necessary God. We can interpret this as follows: the divine essence is embodied in concrete states, but cannot fail to achieve embodiment in some such states or other; moreover, any such state will be connected with all others so as to form a single personal life, what Whitehead calls a “personally-ordered society,” or a sequence of states interconnected by ideally complete memory and continuity of purpose.

Some readers will say that I do not understand the elementary principle that the divine is not temporal, or subject to the distinction between individual and state, or past and future, or potential and actual. During the past forty years I have encountered this view perhaps a thousand times; however, I have also learned that another, less popular, conception has for centuries had its defenders. After reflecting for many years upon the question, I still see the more usual view as a philosophical and religious blunder.

The divine self-identity or individuality is necessary because it alone is, in its essence (not in its full actuality), free from all particularity or exclusiveness. There is no conceivable world whose existence God could not tolerate and know; and for this very reason there is no world which he must have rather than any other. The divine individual alone has strictly universal functions, and hence it alone, in abstraction from its concrete actual embodiments, is wholly unparticular, non-exclusive, or necessary. It cannot be contingent.

Is God then not a “particular” individual? No, certainly not; he is the universal individual. What do I mean here by “individual”? I mean the unity of a sequence of concrete states of consciousness each connected with the others in the most truly ideal way by omniscient memory and steadfastness of purpose. This is plainly analogous to “individual” in the every day sense, except that this individual, being universal in his role, is unique and without competitor. Being non-localized, he occupies no place from which he excludes other beings, as each of us does at every moment. There is no function exercised by God which any other being could take over in his stead. He is the sole non-competitive, non-exclusive, conscious agent – in his necessary essence quite a general as being itself, but in his contingent actuality containing all the exclusive particularity and concreteness of the real.

Thus the Ontological Argument can be freed of its basic paradox, that it appears to derive the concrete from an abstract definition. On the contrary, the individuality whose existence the argument shows to be necessary is no less abstract than the definition of perfection which furnishes the starting-point of the argument. God merely qua necessary is infinitely less than God in his fullness of actuality. This fullness we cannot know by any abstract argument; and we cannot know it at all, except vaguely, fragmentarily, and uncertainly, and even then, only by reacting to all our experience and knowledge in a fashion for which words can never be adequate.

In this respect, the Zen masters, other mystics, and Wittgenstein are right. There is that which, though it can perhaps in some faint and fragmentary way be “shown,” cannot be said. But this something is not any “essence,” even the necessarily somehow-existing essence of God – the most independent, abstract or general, hence the most sharply definable, of all individual characters – but is rather the unimaginable states by which the worlds, as they come into being, are omnisciently grasped and immortalized. These states constitute not merely that, but how, God exists. No such state can ever be known – save to another such state. It can be felt, “prehended”; and indeed we cannot avoid prehending the concrete deity, but with an overwhelming preponderance of “negative prehensions,” such that they bury in hopeless obscurity all but an infinitesimal portion of the wealth prehended. The God demonstrated by a correct ontological proof, or any other proof, is unspeakably less than the actual God. However, this very truth that there necessarily is an actual God who is unimaginably more than our proofs describe, this too is proved, if the proving process succeeds at all.

Thus Anselm is vindicated in his contention that we can conceive God as incomparably greater than anything we can (explicitly) conceive. We can understand that we cannot under the content of omniscience and that necessarily the existing divine necessity (the common denominator of possibilities) fails to compare in richness of contrast and beauty with the ever-growing integrated totality of contingently actualized possibilities which is the divine consciousness.

According to G. E. M. Anscombe, the ontological argument is a mere fallacy since it amounts to this: (x)Gx → EEx [my notation].

“For all x, if x is God, the x has eternal existence. We can . . . still ask, ‘But is there such a being?’.”[4]

This is not the argument, even as in Anselm, much less as we can put it now. Instead of the somewhat evasive “eternal existence” Anselm has “such that thou canst not be conceived not to exist.” Moreover, Anselm argued not from God, existent or not, but from the idea of God taken as admittedly an existing idea. The question then concerned the status of this idea, as exemplified or unexemplified. He held that the only exemplification which would not contradict the meaning of the idea must be necessary exemplification. Further, that which cannot be contingently exemplified cannot be contingently non-exemplified either. Hence, assuming that there is a genuine idea (that divinity is logically possible), it follows that divinity exists necessarily, since the other three relations to existence have been excluded: the three being contingent exemplification and non-exemplification, and necessary non-exemplification (impossibility).

One should also remember that it is a respectable, even though not universal, doctrine in modal logic to hold that

~(Ex)fx & ~N~(Ex)fx (The function is contingently unexemplified) also to say,

(Ex)fx & ~N(Ex)fx (The function is contingently exemplified) for other functions what makes sense is only:

N(Ex)fx V N~(Ex)fx (The function is either necessarily exemplified or necessarily not exemplified – i.e. impossible)

If all language is taken as extensional, as requiring some realm of objects to refer to, then some function must be such that it necessarily denotes objects, say the function of “being an object.” But note: “infallible knower of all objects, whatever they may be,” is just as general, as noncommittal among objects, as the mere idea of “object” itself. True, it adds the idea of “knowledge,” but this, too, is a universal correlate of objects, unless one wants to try to make sense out of “absolutely unknowable objects.” As for “infallible,” this simply prevents “knows” from being subject to qualification or diminution. We know save so far as we do not know but are ignorant; God, however, knows.

What Anselm intuitively grasped was the complete abstractness or neutrality of the idea of perfection; had he become clearly conscious of the nature of his discovery he must have seen that no such neutral entity could be the total actuality of the supreme existent; he must have altered his idea of deity as necessary through and through, or in all properties.

It is important not to forget that the values of x in existential formulae need not be individuals; they may be states of individuals, “actual entities.” With reference to deity this distinction not only does not disappear, it becomes uniquely important. For here alone the individuality has a different modality from the actual states; elsewhere both are contingent, here, only the states. This is, I submit, the only way – if there is a way – to remove logical absurdities from the idea of God, and to clarify at last the so long misconceived logic of the Ontological Argument.

In the foregoing I have sought to show, not that the traditional Ontological Argument is conclusive, but that the nature of the issues which it raises has been incompletely understood, both by its defenders and by its critics. In a prospective publication, The Logic of Perfection, it will be argued that of twenty criticisms to which the Argument has been or can be subjected some are irrelevant and the remainder are at best inconclusive, when applied to the reformulation of the reasoning, and of its conclusion, which I propose.[5] The problem for the future is this: what is the relation between our criterion for analytic, non-analytic, and analytically false, relative to a specified language system, as given by certain authors, e.g. Richard M. Martin.[6] I have some vague guesses about this, but no more. Upon one point, however, I am tolerably clear: the answer is not furnished ready-made in the pages of Kant, Hume, or the many who do little more than repeat their dicta on this matter, with the ad hoc incorporation of a trick or two from modern logic. We face here large systematic syntactical and semantic issues such as cannot be disposed of by minor revisions of the old shortcuts and dogmas, whether speculative or skeptical, theistic or non-theistic. No more in metaphysics than in physics or biology can we identify our problems with those of the seventeenth or eighteenth century. And the concepts of modality and logical validity will have to be clarified well beyond what has yet been achieved before we can say that the possibility of a cogent Ontological Argument has been provide or disproved.

However, I do venture to regard it as already manifest that the argument not only could not support, it could only tend to refute, the way of defining deity which Anselm, Descartes, and even Kant had in mind. Only a genuinely different type of theism offers any hope of finding a positive use (or perhaps even of evading a negative or anti-theistic use)[7] for the brilliant suggestion of Anselm. Such a new theism also, I believe, offers the only hope of achieving clarity concerning the various remaining theistic argument – and counter-arguments. It is not the classical theologians, but certain heretical ones, Socinus, Fechner, Whitehead, and a number of others, who point the way to the avoidance of the confusion between abstract and concrete (“the fallacy of misplaced concreteness”) which is the pervasive error of much traditional thought and is found among persons of very diverse religious and anti-religious persuasions. The Great (or “perennial”) Tradition in metaphysics or speculative philosophy is one thing, and metaphysics as such (and its possibility or impossibility) is quite another; very much as Classical Physics is one thing, and physics is another. Is there anything very odd or irrational about this distinction? Is not the refusal to make it the attitude which deserves to be branded as odd, nay, in principle obscurantist, a “blocking of the path of inquiry”?

[1] Karl Popper’s lucid classification of existential statements as impossible, empirical, or metaphysical – in the Logic of Scientific Discovery, New York, 1959, p. 633 – according as they contradict all, some, or no positive observation statements (“basic statements”) is equivalent to ours, except that for his purpose there was no need, and for ours there is need, to subdivide the empirically non-testable statements into those whose “unfalsifiability” is relative to human perception, and those which are unfalsifiable absolutely, or relative to any form of positive perception or knowledge, whether human, subhuman, or superhuman. The latter statements may be termed absolutely transempirical or metaphysical. They are sheer necessities. Since I read some parts of Popper’s great book years ago in the earlier (German) edition, it seems just possible that in the recesses of memory I owe my idea of this matter to him. I wish also to salute Popper for his wisdom in having refused, from the outset, to identify the meaningless and human untestable. Such identification is one form of the attempted deification of human nature, which is not the measure of all things. This mistake was original only in ancient Greece.

[2] According to M. Lazerowitz, a necessary proposition is one such that a proposition falsifying it has no descriptive use. Very well, I maintain that “world without God,” or “state of affairs such that there is no God” has no such use. It does not distinguish one conceivable world from another. That it cannot do so is deducible from the definition of God. (See M. Lazerowitz, The Structure of Metaphysics, London, 1955, pp. 265-71. Also Alice ambrose, in “Proof and the Theorem Proved,” Mind, 68 [1959], p. 439.)

[3] E.g., A. Flew, in New Essays in Philosophical Theology, London, 1955, pp. 145-69, esp. 153. For a detailed discussion of the issue, see my “Freedom requires indeterminism and universal causality,” Journal of Philosophy, 55 [1955], pp. 793-800. (The 8th line of this essay should be deleted, and the 9th should begin with “insight,” not with “sight.”)

[4] See her Introduction to Wittgenstein’s Tractatus, London, 1959, p. 15.

[5] In a remarkable number of cases these rebuttals of the standard objections are dupli-cated (I assume independently) by Norman Malcolm, in his admirably reasoned article, “Anselm’s Ontological Arguments,” Philosophic-al Review, 69 (1960), pp. 41-62. My discussion was completed before I knew anything of Malcolm’s; and some of his main points are to be found in my writings of many years ago. I am delighted, however, to have his support. Perhaps now our profession will begin to realize that we cannot rely upon Kant to do our thinking for us in this matter. But Malcolm does not refute the objection that the argument, when used by traditional theists, implies the derivability not only of “existence” but of concrete actuality from an abstract definition. All Malcolm does is to charge the objection with a confusion between “concrete” and “contin-gent.” This will not do; for the burden of p roof is on the man who holds that there is any alter-native to “concrete and contingent” except “more or less abstract,” and in the extreme case of the completely abstract, “necessary.” What can the necessary be if not what all possible concrete states of affairs have in common (a necessary proposition is entailed by any and every proposition), hence a residuum of omission, a deficient mode of realty, if taken merely by itself?

[6] See R. M. Martin, The Notion of Analytic Truth, University of Pennsylvania Press, 1959.

[7] See J . N. Findlay, “Can the existence of God be dis-proved?” Mind LVII (1948).

Back to Hartshorne page