Charles Hartshorne "A Natural Theology for Our Time: Chapter Three"

A Natural Theology

for Our Time

The Open Court Library of Philosophy

Eugene Freeman, editor, 1967

Charles Hartshorne

Chapter Three: Why There Cannot Be Empirical Proofs

The collapse of natural or rational theology in the eighteenth century had a number of causes, some of which have been touched on. Perhaps the most important of all was an inappropriate definition of deity, inappropriate not only because it failed to express the meaning of worship, the intuitive ground of the idea of God, but also because it involved the philosophically baseless supposition that “greatest possible actuality” makes sense and because it resulted in antinomies from which two millennia failed to find an issue, so long as the definition was retained, but which disappear when it is given up.

Another cause of the “collapse” was the supposed distinction between the “a priori” ontological argument and the “a posteriori” cosmological and design arguments. This is an unclear and essentially erroneous distinction. On the one hand, all those who accepted the ontological argument held that among the implications of the reasoning was the view that in thinking God we in a fashion also experience him, so that the argument, though a priori, is also in a sense experiential. On the other hand, the cosmological and design arguments are not really “empirical” in the sharp sense which goes beyond “experiential” as just used. This sharp sense is the one that Popper first clearly defined: that is empirical in the distinctive sense which some conceivable experience would falsify. It is not enough that experience can illustrate or confirm a proposition; if it is to be usefully called empirical, experience must conceivably be able to disconfirm it. But what advocate of “empirical arguments” for theism (Tennant? Brightman?) has told us how experience might conceivably show that God did not exist? Did Thomas, advocate of the “a posteriori” method, do this? I think not. I hold that this test is decisive. On the one hand, God, being ubiquitous, can be experienced, but then all sorts of things, including the relations of numbers in finite arithmetic, can be experienced, no matter how a priori in the logical sense. But on the other hand, God, being the sole necessarily existing individual, could not possibly be disconfirmed by a contingent fact; and so in the useful or distinctive sense his existence is not “empirical.”

Let us see how this nonempirical status results from our treatment of individuals as subject to the universal rule that they all interact with at least some other individuals. To say, “with some” is to prompt the query: with how many, and which ones? The answer cannot be given a priori: one must point in the perceived world to the particular individuals. But “all the individuals which there are, without possible exception,” is definite a priori. True, we do not know a priori which individuals actually fill the role of interacting with God, but we do not need to know this in order to have defined just the one divine individual. For he alone could in any world have unrestricted scope of interaction. No further principle of individuation is needed or possible. Whereas the role of local interaction simply as such is highly indefinite and specifies no one individual, the role of universal interaction is unique.

The only other unique case specifiable a priori is that of zero interaction. According to this, God may act, but cannot be acted upon. Classical theism chose this zero case as definitive of deity. The choice was doubly blind, in that the universal case was ignored, rather than carefully scrutinized on its merits, and in that no attention was paid to the apparent impossibility of making sense out of an acting agent which cannot interact, or of an individual which cannot relate itself to other individuals. There was even a third oversight, the fact that religious language is full of implications of interaction between God and his creatures. And since the universal form of interaction is just as truly unique a priori as the supposed zero form, nothing is lost by choosing it instead, while much is gained.

If the individuality of God could be specified empirically only, the existence of God could be no more than an empirical or contingent fact. God would then be a mere creature, something which might never have existed, rather than the uncreated creator, presupposition of existence and nonexistence, itself without presupposition. Our ancestors realized this necessity that God’s essential uniqueness be purely conceptual, hence a priori, not factual. But they failed to analyze and generalize sufficiently the problems involved. To say, as they did, that God is conceptually unique because he alone does not react or interact is only a negative description. Negation is parasitic on affirmation. In God himself, at least, his uniqueness must be positive. But the sole possible positive correlative to “absolute,” “independent,” or “uncaused” is “relative,” “dependent” or “caused”; and if God is in no way effect, then his relation to the world must be as mere or uncaused cause. And since universal scope of relativity, or of being effect, is just as a priori as zero scope, but has the advantage over the latter of enabling us to preserve the well-founded rule that to be an individual power is to interact, we have excellent ground for taking universal interaction to be the positive feature in the divine mode of action.

Another consequence of this choice is that the divine “absoluteness” becomes genuinely explanatory of relative existence, instead of being merely opposed to it. The divine interaction is strictly or “absolutely” universal. However, the concrete content of this absolute universality must be relative, varying with the actual world. God interacts with whatever individuals there are, but what individuals are there? This is the contingent, relative, or empirical aspect, yet it is integral to the concrete realization of the absolute aspect. One cannot interact except with existing individuals. But one could, it seems, be without interaction, even were there no other individuals. Thus the negative theology makes creation an irrelevant excrescence. The divine absoluteness, neoclassically conceived, is a relative absoluteness, it is universality, on the one hand specified a priori and obtaining by necessity, but on the other hand always existing concretely with respect to some actual contingent world. All the individuals there are God interacts with, but not all individuals there might be; for one cannot interact with the nonexistent.

Here we see what was wrong with the classical, and Kantian, idea of “most real being.” This was a definition, not with respect to actuality, but with respect to possibilities only. All the value or reality there could be, that God was said to have, no matter what the world might be. But therewith the contrast between actual and possible is hopelessly compromised, and the meaning of both terms is put in doubt. Worse, if possible, our struggle to realize possible values in the world becomes a mere absurdity or impertinence. Act as we may, perfection, all possible value, just is actual if classical theism is correct. It is odd that a theory should demonstrate the absolute irrelevance of the theorizing agent and his achievements.

Instead of ascribing to the Unsurpassable the actualization of all possible perfection, we should ascribe to him the actual possession only of all in fact actualized values. The entire actual world is his to enjoy in all-embracing vision. We should further ascribe to him the potential possession of every possible value. Were such and such a possible value actual for anyone, it would a fortiori be actual for God, who would enjoy unsurpassable knowledge of it. For no mode of possession of value is more absolute than full awareness. Once more we find that unrestricted scope, relative to the actual world, but specified a priori and obtaining by primordial necessity, gives us what we need. The divine actuality is logically coextensive with all actuality and, in this sense, is actuality itself; the divine potentiality is coextensive with all possibility and is possibility itself. Any actual thing God enjoys actually; any possible thing would be his actual possession were it actual for anyone. From this “modal coincidence” it follows that though God can increase in value, he can be surpassed by no other than himself. For any increase anywhere is a fortiori increase in him. He grows, but his mode of growth, as Fechner sagaciously said long ago, is incomparably superior to all other modes. Or, as he put it “the perfection of God is his ideal mode of perfectibility.” If he surpasses himself, it is in an unsurpassable manner.

One reason why this solution of the problem of perfection was so generally and so long overlooked was that it is a rather natural verbal confusion to suppose that the unsurpassable cognitive power must survey all possible as well as all actual values, and since to know a value is, as I have just stated, to possess it, it may seem that the actualization of a possible value gives God’s knowledge no quality it would otherwise lack. Thus if a merely possible valuable quality Q is known to God, he possesses Q no less than if it were actualized. This is the paradox inherent in the verbal expression, “possible world.” If possibilities have, item for item, all the qualities of the corresponding actualities, then actualization is meaningless and indeed adds no value. But this should teach us that possibilities are not to be viewed as qualitatively identical with actualities, apart from some quality-free factor of actualization. Actualization must somehow be qualitative enhancement, or the concept is vacuous.

The solution to the riddle is to see that merely possible qualities are lacking in individual definiteness. There are no possible individuals, but only possible kinds of individuals, possibilities for further individua-tion. As Peirce insisted, possibility is in principle general rather than particular, determinable rather than determinate. It is in some degree indefinite, and so, since value lies in harmonized contrasts and the more definite a thing is the richer the contrasts it can involve, it follows that possible worlds, really worldly possibilities or incompletely definite sorts of worlds, are less rich as objects of knowledge than actual worlds. Thus God does not possess actually all possible value simply by knowing all possibilities.

Since actualization is determination, and “all determination is [partial] negation,” as Spinoza said, actuality as such is finite, and only mere possibility can be strictly infinite. The “absolutely infinite” cannot be determinate; so far from being the fulfillment of all possible richness or plenitude of reality, it is the fulfillment of no possible value. It is the ultimate determinable, abstracting from any and every actual determination, by itself wholly indefinite, and thus empty of contrast and devoid of beauty. Actualization is the acceptance of limitation; it requires choice among incompatible values, this and therefore not that, or that and therefore not this. (I wonder what great philosopher before Whitehead clearly stated this, and system-atically drew the consequences?) God in creating worlds faces such a choice, also. Even for him, to do all possible things is to do nothing. But worship does not require sheer infinity—a formless emptiness at best—taken as actual. It requires that God’s potentiality, what he could be, must be as wide as the absolute infinity of logically possible values. That is, no value can be possible in itself, yet impossible for God.

To worship the infinity of God’s power, what he is capable of being, as though it were simply the same as his actuality is idolatry, so far as I can see. And it implies the sheer irrelevance of all creaturely choices. God is not on one side only of categorial contrasts; he is not merely infinite or merely finite, merely absolute or merely relative, merely cause or merely effect, merely agent or merely patient, merely actual or merely potential, but in all cases both, each in suitable respects or aspects of his living reality, and in such a manner as to make him unsurpassable by another. He is even both joy and sorrow, both happiness and sympathetic participation in our griefs. He is not indeed both goodness and wickedness, but only because the latter is a privation, disregard of the interests of others, and this, like “ignorance of what is”, conflicts with “unsurpassable by another.” But loving participation in grief, like receptivity to influence in general, is no privation, but a positive power, extremely limited in us, unlimited in God.

Neoclassical theism can say and mean, “God is love.” The unsurpassable could not be without love—for even we ourselves would surpass a loveless being. To love is to rejoice with the joys and sorrow with the sorrows of others. Thus it is to be influenced by those who are loved. An entity may be influenced by all, by some, or by no other entities—these are the three cases specifiable a priori. Examples fitting the third or zero case are not hard to find, for abstractions, just so far as they are abstract, are immune to influence. What can one do to change the number two? Also, past events are henceforth immune to influence. Deceased individuals are indeed “impassible.” But living individuals . . . ? It seems that contemporary individuality is recognizable only through interaction. So there are basically just two possibilities, an individual interacting “with some,” or “with all.”

Let us sum up our discussion so far. If for ordinary existence there cannot be conceptual proofs but only empirical or factual evidence, this is because ordinary individuals cannot be defined by concepts alone; if for divine existence, on the contrary, there cannot be empirical evidence but can be conceptual evidence, this is because divine individuality is defined purely conceptually. Not only does the divine existence not conflict with any truly necessary conceptual rule, it is even the sole individual existence which expresses such a rule. The other individuals are just facts, permitted but not called for by any rational reason.

Kant said that “concepts without percepts are empty,” true enough; but the bare existence or individuality of God is indeed empty, the mere universal divine outline of existence without concrete or particular content. Kant’s error was in supposing that the existence of an individual is its entire actuality. Not so; that John Smith, born and baptized forty years ago, still exists leaves unspecified all the concrete details of his actual history since birth. Analogously, that God, never born at all, still exists, always has existed, and always must exist leaves unspecified an unimaginable fullness of particular actuality. Mere “existing deity,” without further information, is indeed in a sense an empty concept. However—it is a concept formally incapable of lacking actualization. It can be actualized in an infinity of alternative ways, but unactualized it cannot be. So the empty conceptual knowledge that God exists does tell us that his individuality is actualized somehow. How it is actualized is for science, revelation, personal experience, some form of empirical knowledge, to tell us, so far as we can know it at all. This of course is not very far. Our knowledge of the concrete divine reality is negligibly small.

There is mystery enough about God, not because his eternal “essence” is inaccessible, or because metaphysics or natural theology is moonshine or impossible, but because particular actuality, even divine actuality, is not metaphysical but empirical. Perception, human or divine, is the only avenue to particular actuality. But in one unique case, it is not the avenue to existence or individual identity. Rather, conceptual, or if you prefer, spiritual, insight is this avenue, for the reason that this individuality is specifiable only a priori or nonsensuously. Whereas empirically individuated entities must exist, if at all, empirically, the entity individuated a priori cannot exist empirically.

The rules for individuality as such permit but do not require anyone particular individual with localized scope of interaction; but they could not even permit, unless they also required, the individual with universal scope. This scope presupposes only the meaning of the rules for its uniqueness, whereas “not universal” is infinitely far from unique. To say that the rules for individuality are neutral to the divine existence is to talk confusedly, since it takes two parties to make a harmony, and the idea of God, or of worship, does not allow the rules to be neutral. They must either be hostile to worship a priori, in which case worship is a kind of nonsense, or else favorable a priori, in which case both atheism and positivism, along with empirical theism, are in some fashion nonsensical. Either everything (in some degree) serves God—everything actual actually, everything possible potentially-or nothing serves or could serve him, and the idea is empty of coherent meaning. Either God is lord over possibility as well as actuality, or he is bare nothing, mere conceptual confusion. All other kinds of individuals are thought of as controlling some actualities or some possibilities; accordingly, they mayor may not exist. This will not do for God, who is all or nothing, inherent in any fact, affirmed in any affirmation, any possible hypothesis, or in none.

Still otherwise, “God” is not simply another word in our language but, if anything rational, a name for the principle back of every word in any possible language. He is not merely another topic to think about, but the all-pervasive medium of knowledge and things known, to recognize whom is a way of thinking about no matter what. The question remains, is it a significant, coherent way of thinking about no matter what? That alone is the question.

If the central issue of natural theology were a scientific one, it would be conceivable that observational or contingent facts might justify a negative decision. But any god with whom facts could conflict is an idol, a fetish, correlative to idolatry, not to genuine worship. Divinity being all actuality in one individual actuality, and all possibility in one individual potentiality or capacity for actuality, a possible divine nonexistence, must, like any other possible reality, be a possible state of God and content of some possible divine knowledge. But a being could not, logically could not, know itself as never having existed. Nor could a divine being know itself, as we know ourselves, as having come into existence after previously not existing. Thus to treat the existential question of natural theology as scientific or factual is to change the subject from the question of the reality of a being whose interaction is strictly universal to that of one whose scope of interaction is in principle limited. But if not all possible values belong even potentially to God, what could prevent a conceivable rival from surpassing him by actualizing these values? And can a being open to rivalry be worshipped in the strict or exalted sense?

To ask, as the empirical theist or atheist does, “Is the world such that it must have been, or could have been, divinely created?” implies two kinds of possible worlds, the one kind requiring (or at least permitting?) a Creator, and the other not. What would distinguish the two kinds? Is it the proportion of good to evil? But at what point in the continuum between more and less evil would a possible world abruptly become compatible with being divinely created? Or is it a question of greater and greater probability of such createdness? Does any of this make sense? I am not joking, for I seriously believe that the empiricist program is at this point nonsensical.

We are told by an English writer that it is a question of whether there be any “utterly senseless” or “unredeemed” evil? What would such a thing be like? I declare in all earnestness I have no idea. Any evil has some value from some perspective, for even to know it exists is to make it contributory to a good, knowledge itself being a good. But any evil is also in some degree a misfortune, and in my opinion the theological “problem of evil” is quite misconceived if it is seen as that of justifying particular evils. Evils are to be avoided where possible; where not, to be mitigated or utilized for good in whatever way possible—but never, for heaven’s sake never, to be metaphysically justified. The voice from the whirlwind in the Book of Job says not one word about evil being good in disguise or not really evil after all, and theologians might perhaps learn from this illustrious example. The justification of evil is not that it is really good or partly good or necessary to good, but that the creaturely freedom from which evils spring, with probability in particular cases and inevitability in the general case, is also an essential aspect of all goods, so that the price of a guaranteed absence of evil would be the equally guaranteed absence of good. Thus not even the nastiest or most conceivably unhelpful evil could have anything to do with the nonexistence of God. Risk of evil and opportunity for good are two aspects of just one thing, multiple freedom; and that one thing is also the ground of all meaning and all existence. This is the sole, but sufficient, reason for evil as such and in general, while as for particular evils, by definition they have no ultimate reason. They are nonrational.

It is argued that a divinely-created world must absolutely lack evil, be devoid of suffering and frustration, as well as of wicked intentions. But could “good” mean anything in a world in which any contrasting term would be totally excluded by omnipotent power? And in such a world how could the creatures, who would have no genuine options, even know what was meant by divine freedom to choose this world out of the totality of possibilities? And if they could not know it, then the envisaged perfect world would be one in which at least the good of creaturely understanding of the creator was quite impossible. Or, if there is not to be even divine choice among possibilities, then God would not be responsible in an ethical sense for good or evil; indeed, he would not be subject to ethical criteria, and hence evils would cease to be relevant as disproofs of his perfection, which is in effect being conceived as having nothing to do with ethics, one way or another.

For creationist or neoclassical metaphysics deity must be the supreme or unsurpassable form of creative freedom. But “supreme” or “unsurpassable” form cannot be the only possible form. In creationist metaphysics, all concrete reality is in principle creative. But then what happens is never, as it stands, simply attributable to “the” creator, but only to deity and the creatures together. Reality is always in part self-created, causa sui, creativity being, in this philosophy, the supreme transcendental. All creatures have creativity above zero, all are creators.

If the evils of the world are not an empirical disproof of deity, what more likely negative evidence could there be? I see no plausible candidate. The “degradation of energy”? Only if the principle were known to be the last word from a cosmic and everlasting perspective, and this, I take it, utterly exceeds any possible observational insight. Only omniscience itself could know this. And omniscience could also alter it, or could know it as its own free decision, and as compatible with its own unsurpassable goodness.

There is one last stand for the empiricist to make: the existence of any world at all is what proves God. So, then, were there nothing worldly, there would also be nothing superworldly or divine? What then would there be—just bare nothing, not anything? And what does this mean? I note the words, but for me they do not add up to a definite and coherent meaning. “There would not be anything”—what do “there” and “be” stand for in this context? There are lions because, somewhere, it is true that “there is a lion.” But “there” gets its meaning from things, not from sheer emptiness. “There” means, in the neighborhood of, or in a certain direction and at a certain distance from, the speaker or the hearer. But with “there is not anything,” even “there is” seems to lose all definiteness. In addition, if I am right that God’s unsurpassability implies that his potentiality is coincident with possibility itself, even the “possibility of nothing” must express something which God could be or have. But universal nonexistence, including his own, he logically could not be or have. It follows that the possibility of “nothing worldly,” if indeed it be a possibility, cannot imply the possibility of divine nonexistence, but only of God existing in solitude. In that case, it is silly to argue, “Deity exists because there is a creature.” For either God is incapable of sheer idleness, of not creating, in which case it is no contingent fact that there is something creaturely but an a priori necessity; or, if he is capable of sheer idleness, then he can and would exist even were there not anything worldly. Take it either way, his existence depends upon no empirical fact.

Of course, our being able to know the divine existence depends upon the empirical fact of our existence. But to say, “we exist, therefore we know God exists” is ambiguous. If it means that only the existent can know, it is correct. If, however, it means that our evidence that God exists is that we exist, it is illogical. The difference between our existing and our not existing can have nothing to do with a difference between God’s existing and his not existing. And this remains true even if we generalize from ourselves to worldly existence in general or as such. For, once more, if God could not exist without some world or other, this would only imply that with his unsurpassable creativity he infallibly provides himself with a world; and then it is no mere fact that there is one. And if (which I question) he could exist without a world, it is illogical to say, “Because there is a world God must exist.” In no case can a sheer necessity obtain because one contingent alternative rather than another is realized. The precritical Kant saw and said this clearly, but in the heat of his critical arguments he partly lost sight of its importance.

The search for an empirical meaning for the existential question about God has failed. If this is what Tillich wants to say by the rather violent method of refusing to admit the word existence in relation to God, “more power to him.” Hume and Kant actually understated the case against empirical proofs; the entire attempt was ludicrously inept. It is a conceptual question, a question of self-understanding, clarity, and consistency.

I know by experience how mightily men will struggle against the foregoing argument. They will say, yes if there is a God, then all possibility is his potentiality, but only if. But (a) there is an “if” only where there is a contrary possibility; and (b) the possibility of a Godless world is the possibility of something which God could not (logically could not) know, were it actual. And thus his power to know is being conceived as subject to limitation, and then I can conceive a superior individual with unlimited power to know, and what you have been calling God is therewith exhibited as only an idol. Suppose God to exist (if this is inconceivable, then indeed our issue is not one of fact but of meaning), yet to exist so that his nonexistence would have been possible, or remains a conceivable alternative. Then God must, so to speak, say to himself, “There is something which conceivably might have been, but which I could not conceivably have known.” Is this the conception of God, or of an idol? It seems to me that it is the latter. We need not and should not define God as “all possible perfections”; for there are incompossible value possibilities, and so the phrase is nonsense. But we should define him as the all-inclusive yet individual actuality, and the all-inclusive yet individual potentiali-ty. And then all—simply all—questions of fact take the form, “Which among divine potential-ities are actualized?” None can take the form, “Are any divine potentialities actualized?” For that implies the meaningless notion of every possibility being unactualized. There are incompatibilities of nonrealization as well as of realization; some possibility must be actualized, and this, in terms of our definition of deity, means that God must exist in some actual state or other.

Is the foregoing a proof of the reality of God? No, it is a disproof of both empirical atheism and empirical theism, of the age-old attempt to treat the theistic question as one of fact. There remains, not yet refuted, a priori theism and positivism (the view that “deity” as defined lacks consistent signifi-cance, other than emotional or pictorial). To combat this last one must argue, not from facts but from concepts (other than that of God). We must say that it is no mere fact which the atheist or positivist is by implication denying, but our basic conceptions, those we use even in talking about fairyland. The antitheist, we may urge, is denying even the possibility of a fact.

In rejecting empirical arguments am I rejecting the Biblical pronouncement, “the heavens declare the glory of God”? No, not exactly that either. The heavens do declare the divine glory, but for the believer. Also, if the latter understands his belief, he sees that any conceivable heavens would declare it also. The actual heavens show only the particular form the divine creation has taken, a different heaven would show a different form; but to suppose that some odd or vicious world might exist which God could not have made is confusion. It splits possibility into two, that portion which is included in divine potentiality and that which is not. This is to limit the divine scope, exactly the aspect of the Unsurpassable (except by self) which in consistency is subject to no limitation. Since any conceivable world, or possible nonworld, by the definition of God, must be possible content of divine experience, no argument should take the form, “Because this possibility has been actualized, God exists, or because that possibility has been actualized, God does not exist.” All such talk is about a fetish, not about God.

What the unbeliever who does not find the divine glory in things must rather do is to persuade himself that the definition of God is logically (not factually) suspect—confused, contradictory, or hopelessly vague. And he can find plausible grounds for these charges. It is precisely the coherence and clarity of the idea of deity which “proofs” must try to display. Grubbing among facts is neither here nor there. Self-understanding is the issue: someone is confused, either the theist, or the nontheist. Which is it? This is the real question.

Note that we have here something like the old doctrine: only faith can relate us to God. Only the sense, trust, or insight that we really mean something by our worship, and are not talking nonsense, or chasing a formless pseudo-conception can give the theistic answer. Observed facts and even laws of nature are neutral to the topic. Laws are, by the definition of “God,” something like divine—that is, unsurpassably influential—decrees, free creations which the universe is inevitably inspired to adopt. Given such and such laws, the theist conceives God making the corresponding decrees; given other laws, he simply conceives God making another set of decrees. Either way he thinks of God only as existent.

But suppose there were no laws? Who knows what this supposition means? Certainly we should not be there if it were true. Would anything thinkable, even “there” be there? Here too, we have an issue of meaning. Could there be a simply lawless reality? Or simply no reality? Such questions are pseudo-factual, or in a broad sense logical, and so is the theistic question. “Nothing” could not be “there” (where?) and neither could the nonexistence of God. But perhaps God could be there (i.e., everywhere). And if he “could” be, then he is. For no possibility can be put back of divine creativity, which is possibility. Or else nonsense.

But, you may say, perhaps it is nonsense? Just this the theistic “proofs” must seek to discredit. It is their only task, except the related one of clarifying and explicating the non-nonsensical and coherent content of the Worshipful or Unsurpassable as such.

Forward to

Chapter Four: Theism, Science, and Religion

Back to

Chapter Two: The Theistic Proofs

Posted April 5, 2007

Back to Hartshorne page