Does Omniscience Imply Foreknowledge? Craig on Hartshorne

by Donald Wayne Viney

Donald Wayne Viney is Associate Professor of Philosophy in the Social Science Department at Pittsburg State University, Pittsburg, KS 66762. He is the author of Charles Hartshorne and the Existence of God (SUNY Press, 1985).

The following article appeared in Process Studies, pp. 30-37, Vol. 18, Number 1, Spring, 1989. Process Studies is published quarterly by the Center for Process Studies, 1325 N. College Ave., Claremont, CA 91711. Used by permission. This material was prepared for Religion Online by Ted and Winnie Brock.


William Lane Craig’s views of Hartshorne, regarding God’s foreknowledge because of God’s omniscience, are misunderstood and unjust.

One of the ideas for which Charles Hartshorne is known is that God is omniscient but does not have absolute foreknowledge. In several recent works, William Lane Craig argues that there is nothing to commend Hartshorne’s view, that there are positive arguments against it, and that absurd consequences follow from it (PT 103; PS 16:201; cf. OWG 60). Unfortunately Craig misrepresents Hartshorne’s position. As a consequence, he underestimates the strength of Hartshorne’s views of omniscience. The purpose of this paper is two-fold: (1) to set the record straight on Hartshorne’s view of omniscience and (2) to reexamine Craig’s arguments to see what force they have against the position Hartshorne actually endorses. I will argue that there is something to commend Hartshorne’s view, that the positive arguments against it are unsuccessful, and that, as far as Craig has shown, no absurd consequences follow from it.

Hartshorne on Omniscience and Foreknowledge

Hartshorne defines "omniscience" as "knowledge of all things, perfect knowledge" (ER 546). He characterizes divine foreknowledge as God’s view of all events that, from our standpoint, are future (ER 284). It is noteworthy that, as Hartshorne explains it, the concept of foreknowledge does not specify whether the knowledge of future events is temporally prior to those events. The Boethian God who surveys all of time in a single eternal vision has foreknowledge in Hartshorne’s sense even though it is incorrect to say that such a God knows future events in advance (CP 116). Equally, a God whose existence is temporal and who is prescient of all future events has foreknowledge in Hartshorne’s sense. The idea that future events are known -- not the knowing of them as future -- is the essential meaning that Hartshorne assigns to the concept of divine foreknowledge. Craig may be correct that, in their critiques of omniscience, process theists often fail to include the views of those who affirm God’s temporal knowledge of all past, present, and future events (PS 16:198; PT 97). However, this is not a failing in Hartshorne’s writings.

Hartshorne argues that divine foreknowledge does not follow from omniscience unless it can be shown that divine foreknowledge is possible. But divine foreknowledge is not possible unless future events exist, as fully determinate, to be known. Hartshorne denies that future events exist in this sense (JP 60:604). More precisely:

The future is irreducibly potential rather than actual, and this means in some degree, however slight, indeterminate rather than determinate. Becoming is the passage from incomplete definiteness to definiteness. It is creation (MTG 30).

If perfect knowledge is knowledge of the world as it actually exists, then "omniscience is only possible as itself temporal -- as knowing new facts when there are new facts to know, but always knowing all the facts there are at the time" (ER 284). This is Hartshorne’s central argument concerning divine knowledge and is found throughout his writings (HJ 37:251; CSPM 135; AW 13; WVR 15; OOTM 38-39; cf. HPPT). The reasoning is deceptively simple: perfect knowledge knows things as they are. The past is determinate and the future is partly indeterminate. Therefore, perfect knowledge knows the past as determinate and the future as partly indeterminate.

A characteristic feature of Hartshorne’s central argument is that it concerns God’s knowledge of events and not God’s knowledge of truths about events. At one point, Hartshorne alludes to the distinction, made famous by Bertrand Russell, between knowledge of things and knowledge of truths. Knowledge of things sometimes involves an immediate awareness of an object that Russell calls knowledge by acquaintance. Knowledge of truths does not require a direct awareness but applies solely to beliefs, convictions, or judgments about what is the case (PP Chapter V). According to Hartshorne, God’s knowledge is best conceived as a knowledge by acquaintance (ER 547). This is not to say that God cannot have knowledge of truths. However, Hartshorne argues that if God’s knowledge is to be conceived as the highest form of knowledge, it must be conceived as a knowledge by acquaintance. Hartshorne follows Whitehead in viewing divine knowledge as a form of participation in the feelings of the creatures. Without participation, Hartshorne says, omniscience "would be only an abstract and inadequate knowledge of the creatures" (CSPM 263; cf. HPPT).

Since Hartshorne’s central argument concerns God’s knowledge of things, it is a mistake to construe it as concerning God’s knowledge of truths. Craig makes this mistake. According to Craig, Hartshorne’s reason for denying God’s foreknowledge is that "future contingent statements cannot be true. If they cannot be true, then obviously they cannot be known to be true. If they cannot be known to be true, then God cannot know future contingents" (PS 16:199; PT 99). Craig does not say where he found this argument but it is not the argument that Hartshorne presents. Hartshorne’s argument is not that future events cannot be known because statements about them cannot be true; rather, as we have seen, Hartshorne argues that future events cannot be known because they are not fully determinate.

Once one is clear about Hartshorne’s argument, it is apparent that it is much stronger than Craig allows. Brian Haymes notes that knowledge by acquaintance requires the existence of the object known (CKG 28). It makes no sense to say that one knows by acquaintance something that does not exist. Thus, if future events do not exist, then not even God could know them. Or, more perspicuously, in Hartshorne’ s preferred terminology, if future events do not exist as determinate, then God cannot know them as so existing. Craig says that he shares Hartshorne’s view that "the future is not on an ontological par with the present, that is to say, future events do not in any sense exist" (PS 16:199).1 Unless he denies that one cannot know things that do not exist, he should accept Hartshorne’s view that God does not have foreknowledge.

God’s Knowledge of Truths

Craig’s interpretation of Hartshorne’s views may result from his adopting the idea that perfect knowledge is to be construed as a kind of cosmic storehouse of information, past, present, and future. Many accounts of omniscience treat omniscience in this way and discuss only God’s knowledge of truths (cf. PAG 3). While Hartshorne understands omniscience as fundamentally a knowledge by acquaintance. he does not deny that God has knowledge of truths. However, he believes that truth itself has a temporal status.

Truth is some sort of correspondence and the temporal status of the truth is the same as the correspondence. There can be no timeless relation to something whose mode of being is temporal, for relation to X includes X, and if X comes into being, so does the relation (MIGB 74:54-55).

If truths come into being then God would know them as coming into being. Or again, if God knows all and only tine statements, then as new truths emerge in the creative advance, God comes to know them.

If it is true that new facts come into being then the only question is how this truth is to be represented in language. Aristotle apparently believed that propositions about future contingents have an indeterminate truth value (CDI 140). In 1939, Hartshorne defended the Aristotelian view. He argued that if the future is indeterminate as to detail "detailed propositions about it must, to correspond with it, have indeterminate truth values" (APA 27). This view requires that, "as Aristotle seems to have said, the Law of Excluded Middle is valid only of past and present, but not of the future" (APA 28).2 With the publication of Man’s Vision of God (1941), Hartshorne abandoned his earlier position, suggesting that it would be unwise to deny the law of excluded middle for future tensed propositions (MVG 140). In a recent work, Hartshorne reaffirms this conviction:

Aristotle’s way of expressing the indeterminacy of the future seems to consist in suspending the law of excluded middle as applicable to certain propositions . . . I believe that the law of excluded middle as to truth values is best accepted even for propositions about the future. (IO 45).

Hartshorne came to see the Aristotelian solution to the problem of future tensed propositions as unsatisfactory.

If the law of excluded middle is to be retained for propositions concerning future contingents, then how is the indeterminacy of the future to be expressed in language? Hartshorne finds the clue to solving this problem in the unfolding causal history of the universe that makes certain events inevitable and others merely probable. Rather than speaking of indeterminate truth values, however, Hartshorne speaks of indefiniteness as a predicate:

We need no third truth value, but we do need a third type of predicate for future moments of process besides ‘definitely P’ and ‘definitely not P’, namely ‘indefinite with respect to P’. (CSPM 135)

To conceive of indefiniteness as a predicate for future moments of process allows one to put statements about the future into three categories, corresponding to the formally exhaustive triad of all/none/some. Either (a) all causal possibilities include P (definitely P), (b) no causal possibilities include P (definitely not P), or (c) P is included in some but not all causal possibilities (indefinite with respect to P). As Hartshorne notes, this view does not require the sacrifice of the law of excluded middle. "The truth of one of the three (a, b, or c) is the falsity of the other two" (CSPM 135).

Craig claims that Hartshorne’s view of future tensed propositions sacrifices what he calls the traditional law of excluded middle. According to this law, for any statement and its contradictory, one must be true and the other false. According to Craig, "Hartshorne’s position with regard to future contingent statements is that any such statement and its contradictory are both false" (PS 16:198; PT 96). As evidence for his interpretation, Craig notes that, for Hartshorne, statements of the forms ‘x will occur’ and ‘x will not occur’ are both false where x is a future contingent. Craig does not mention Hartshorne’s repeated insistence that, technically, such statements are contraries, not contradictories (JP 61:476; MIGB 74:49; CSPM 16). For Hartshorne, the strict meaning of ‘x will occur’ is ‘all causal possibilities include the occurrence of x’; the strict meaning of ‘x will not occur’ is ‘no causal possibilities include the occurrence of x.’ The meaning Hartshorne assigns the statements makes them contraries, not contradictions. It follows that Hartshorne’s view does not sacrifice the traditional law of excluded middle.3

Although Hartshorne’s view does not violate the traditional law of excluded middle, his view requires that the principle of bivalence be rejected. According to this principle, the probability of something’s being P is 0 if and only if the probability of its being P is not 1 (FP 5:188). For Hartshorne, propositions concerning an indeterminate future have a probability greater than 0 and less than I. On the other hand, for classical theists, like Craig, the probability calculus must be a concession to human ignorance of what the future will hold. For classical theists, there could be no probabilities between 0 and 1 ,from the divine standpoint, except insofar as God understands that, from a creaturely standpoint, the future is uncertain. According to this view, whatever is uncertain for the creatures is certain for God. If God’s standpoint represents the final truth of things, and if classical theism is true, then the principle of bivalence is true. Hartshorne denies the classical theistic view of divine knowledge, and with it, the principle of bivalence.

Hartshorne is aware that the strict meanings that he proposes for ‘x will occur’ and ‘x will not occur’ differ from the ways such expressions are sometimes understood (MIGB 74:52). According to Paul Horwich,

. . . most of us do believe that some definite course of future events will occur whether determined or not. We naïvely make a distinction between what will happen and what must happen. And such a distinction is recognized in practical affairs when we give credit (e.g. by paying off bets) for correct predictions (AT 32).

One may question whether "most of us" really believe that a definite course of future events will occur -- as Horwich knows, Aristotle did not believe this. Nor, apparently, did Charles Dickens when he had Scrooge ask the Ghost of Christmas Yet to Come, "Are these the shadows of the things that Will be, or are they shadows of the things that May be, only?" (CC 128). Nevertheless, it is true that Hartshorne’s technical meanings for future tensed propositions are not necessarily presupposed by the average person. If one says, prior to the 1980 Kentucky Derby, "Genuine Risk will win the race," one does not thereby imply that all causal possibilities include Genuine Risk’s winning the race. But neither, it seems, does one imply that it was true, prior to the race, that Genuine Risk would win. Horwich thinks that the paying of bets indicates otherwise. However, bets are paid only after the race, only after it is definitely the case that Genuine Risk has won. Thus, the truth or falsity relevant to a wager is the truth or falsity of a present or a past tensed statement, not a future tensed statement. The appeal to ordinary language and gambling behavior does not settle the issue of the truth value of future contingent statements.

If there is a brief answer to why one should adopt Hartshorne’s view of future tensed propositions it is that it does not commit one to the principle of bivalence. Richard Purtill notes that, if the principle of bivalence is true, then "probability theory simply collapses into standard propositional logic" (FP 5:188). This consequence seems too high a price to pay to preserve certain intuitions about the omnitemporality of truth.

Craig’s Arguments

Craig offers three arguments against what he considers to be presupposed by Hartshorne’s views of truth and the future. It will be instructive to see what, if any, force they have against the ideas that Hartshorne actually holds.

Before looking at Craig’s arguments, let us briefly examine what he takes to be presupposed by Hartshorne’s views. Craig claims that Hartshorne’s views about truth and the future presuppose "a crude view of truth as correspondence." According to this crude view, "in order for a statement to be true, the reality described by the statement must actually exist at the time at which the statement is true." This view is crude because "only present tense statements require that the events they describe exist contemporaneously with the truth of the statement" (PS 16:199; PT 99).

It is not clear why Craig believes that Hartshorne’s ideas about truth and the future presuppose the crude view of truth as correspondence. We have seen that Hartshorne holds to a kind of correspondence theory of truth, but it is not the view that Craig describes. To repeat, Hartshorne believes that the past is determinate and the future partly indeterminate. This asymmetry of past and future is mirrored linguistically in the triad of predicates, definitely/definitely not/indefinite. Now, none of this presupposes that an event must exist contemporaneously with the truth of the statement. The theory, however, does require that an event must have the ontological status of determinateness in order to have the predicate of definiteness. With this point of clarification, let us turn to Craig’s arguments. For convenience the arguments are labeled Al, A2, and A3.

AI: According to Craig, the truth of a present tense statement entails the prior truth of the future tense version of the same statement. Consider the two statements:

(1) It is raining (uttered on April 13).

(2) It will rain on April 13 (uttered prior to April 13).

Craig says that these statements have the same truth value (at the time of their utterance) because they make the same claim about the facts, viz, rain on April 13. Craig notes the absurdity of supposing, with Hartshorne, that it is possible for (1) to be me and (2) to be false (PS 16:200).

We have seen that Hartshorne construes statements like (2) in a technical sense, as meaning that all causal possibilities include the occurrence of rain on April 13. If rain on April 13 is truly a contingent event, then it is not the case that all causal possibilities include the occurrence of rain on April 13, and in Hartshorne’s technical sense, (2) would be false. This does not mean that (1) cannot be true. All that is entailed by the truth of (1) is that some causal possibilities include the occurrence of rain on April 13. Hartshorne notes that a paradox, similar to the one that concerns Craig, results when a false or inexact scientific law is ‘verified.’ Following Karl Popper’s lead, Hartshorne suggests that the decisive test of a prediction is falsification, not verification (MIGB 74:47). Thus, (2) may be false but ‘verified’ or corroborated by the truth of (1); but the falsity of (1) conclusively falsifies (2). In any event, the apparent absurdity of combining the truth of (1) with the falsity of (2) results from ignoring what Hartshorne means by statements like (2).

Undoubtedly, Craig would not accept Hartshorne’s technical meaning for future tensed propositions. Thus Craig may persist in finding a problem with combining the truth of (1) with the falsity of (2). But, of what philosophical significance is this fact? We have already noted that neither the truth nor the falsity of Hartshorne’s technical meanings are necessarily presupposed in ordinary linguistic usage. Indeed, ordinary language can fuel the intuitions of Hartshorne as well as Craig. As Hartshorne notes, "People commonly hesitate, in this and many other matters, between two or more meanings . . .(MIGB 74:52). Consider, for example, two sentences, uttered prior to April 13:

(2) It will rain on April 13.

(3) It may not rain on April 13.

If (2) is true, then (3) is false. Yet, Craig apparently believes that it is possible for both statements to be true -- for (3) expresses the contingency of rain on April 13 and (2), according to Craig, is neutral as between the contingency and necessity of rain on April 13. To rid himself of the paradox, Craig is free to marshal his own technical meanings for "will" and "may not." From that point, the discussion will involve the relative merits of the technical vocabularies of the competing positions in illuminating and/or solving the relevant philosophical issues. (We have already noted a rather weighty argument in favor of Hartshorne’s view, viz, that it does not commit one to the principle of bivalence.) Given the vagueness of intuitions and the inexactness of ordinary language, it seems advisable to raise the debate to the level of technical vocabularies.

A2: If future contingent statements cannot be true because the events to which they refer do not yet exist, then past contingent statements cannot be true because the events to which they refer no longer exist. But past contingent statements can be true, as even Hartshorne admits. Therefore, future contingent statements can be true.

This argument betrays Craig’s misunderstanding of Hartshorne’s view of truth and the future. For Hartshorne, past contingent statements are true because the reality to which they refer is determinate. The future, however, is partly indeterminate. Thus, the possibility of assigning truth to past contingent statements carries no necessary implication for assigning truth to future contingent statements.

A3: Craig claims that a tensed proposition can be put into a tenseless form. For instance:

(4) It rained on April 13

can be rewritten in tenseless form as:

(5) On April 13, it rains.

Although Craig acknowledges that (4) and (5) are not synonymous, he says that they have the same truth value. But the "tenseless versions are always true or false if they are ever true or false" (PS 16:200). Therefore, future contingent statements can be true.

This argument begs the fundamental question against Hartshorne. Craig offers no reason to suppose that (5) does not become true. He simply asserts that the tenseless version of a statement is "always true or false." According to Hartshorne, "the notion that dates can be assigned from eternity is one of the fairy tales -- or controversial assumptions -- which haunt this subject" (MIGB 74:51).


If one held to the crude view of truth as correspondence, or if one denied the law of excluded middle as applied to future contingent propositions, then Craig’s arguments would pose a serious challenge to one’s views. Hartshorne holds to neither of these views; and this accounts for the fact that Craig’s arguments will not be very convincing to Hartshorneans. Once Hartshorne’s views get a fair hearing, it is clear that they are more subtle and less easily refuted than Craig allows.



APA -- Charles Hartshorne. "Are All Propositions About the Future Either True or False?" Program of the American Philosophical Association: Western Division, April 20-22, 1939, 26-32.

AT -- Paul Horwich. Asymmetries in Time. Cambridge, MA: MIT Press, 1987.

AW -- Charles Hartshorne. Aquinas to Whitehead: Seven Centuries of Metaphysics of Religion. Milwaukee: Marquette University Press, 1979.

CC -- Charles Dickens. A Christmas Carol. New York: James H. Heineman, 1967.

CDI -- Aristotle. Categories and De Interpretatione. Trans. J. L. Ackrill. Oxford: Clarendon, 1963.

CKG -- Brian Haymes. The Concept of the Knowledge of God. New York: St. Martin’s, 1988.

CP -- Boethius. The Consolation of Philosophy. Trans. Richard Green. Indianapolis, IN: Bobbs-Merrill, 1962.

CSPM -- Charles Hartshorne. Creative Synthesis and Philosophic Method. La Salle, IL: Open Court, 1970.

ER -- Charles Hartshorne. "Foreknowledge, Divine" and "Omniscience." An Encyclopedia of Religion. Ed. Vergilius Ferm. Secaucus, NJ: Popular Books, 1945.

FP 5 -- Richard L. Purtill. "Fatalism and the Omnitemporality of Truth." Faith and Philosophy 5:2 (April 1988): 185-192.

HPPT -- Donald Wayne Viney. "God Only Knows? Hartshorne and the Mechanics of Omniscience." Hartshorne, Process Philosophy and Theology. Ed. Robert Kane and Stephen Phillips. Albany, NY: SONY Press, 1989.

HJ 37 -- Charles Hartshorne. "The Reality of the Past, The Unreality of the Future." Hibbert Journal 37:2 (January 1939): 246-257.

IO -- Charles Hartshorne. Insights and Oversights of Great Thinkers An Evaluation of Western Philosophy. Albany, New York: SONY, 1983.

JP 60 -- Charles Hartshorne. "Real Possibility." The Journal of Philosophy 60:21 (Oct. 10, 1963): 593-605.

JP 61 -- Charles Hartshorne. "Deliberation and Excluded Middle." The Journal of Philosophy 61:6 (Sept. 3, 1964): 476-477.

MIGB 74 -- Charles Hartshorne. "The Meaning of ‘Is Going to Be."’ Mind 74:293 (January 1965): 46-58.

MTG -- Charles Hartshorne. "Grounds for Believing in God’s Existence." Meaning, Truth, and God. Ed. Leroy Rouner. Notre Dame & London: University of Notre Dame Press, 1982.

MVG -- Charles Hartshorne. Man’s Vision of God and the Logic of Theism. Chicago: Willet, Clark, 1941.

OOTM -- Charles Hartshorne. Omnipotence and Other Theological Mistakes. Albany, New York: SONY Press, 1984.

OWG -- William Lane Craig. The Only Wise God: The Compatibility of Divine Foreknowledge and Human Freedom. Grand Rapids, MI: Baker Book House, 1987.

PAG -- Jonathan L. Kvanvig. The Possibility of an All-Knowing God. New York: St. Martin’s, 1986.

PP -- Bertrand Russell. The Problems of Philosophy. London: Oxford University Press, 1973.

PS 16 -- William Lane Craig. "Process Theology’s Denial of Divine Foreknowledge." Process Studies 16:3 (Fall 1987): 19S-202.

PT -- William Lane Craig. "Divine Foreknowledge and Future Contingency." Process Theology. Ed. Ronald H. Nash. Grand Rapids, MI: Baker Book House, 1987.

WVR -- Charles Hartshorne and Creighton Peden. Whitehead’s View of Reality. New York: Pilgrim, 1981.



1Although Craig believes that future and present are not on the same ontological footing, he apparently believes that past and future are on the same ontological footing. His second argument against Hartshorne’s view of truth (A2) indicates that he believes that the past and the future are, alike, nonexistent.

2Jules Lequier apparently advocated the Aristotelian view: ". . . between contingent past things and contingent things to come there is this difference: of two contradictory affirmations concerning contingent past things, one is true, the other false; but of two contradictory affirmations concerning contingent things to come, neither the one nor the other is true, both are false" (Oeuvres complètes, ed. Jean Grenier. Neuchatel, Suisse: Editions de la Baconnière. 1952, p. 194).

3 Nor is Hartshorne’s position on the law of excluded middle inconsistent as Craig claims (PS 16:20l; PT 114). The alleged inconsistency arises because, according to Craig, Hartshorne holds that (1) contradictory statements about future contingents are both false, and (2) it is not the case that contradictory statements about future contingents are both false. The weakness in Craig’s allegation, as I have shown, is that Hartshorne does not endorse (I).

In addition to the charge of inconsistency, Craig finds fault with Hartshorne’s claim that the contradictory of "x will occur" is "either x will not occur or, at least, it may not occur" (MVG 101). According to Craig, "The contradictory of any statement ‘p.’ is ‘not-p,’ and not, as Hartshorne would have it, ‘not-p or possibly not-p’" (PS 16:201; PT 114). George Shields is surety correct in interpreting Hartshorne to be speaking loosely, "meaning by ‘negative’ a ‘denial’ of a statement, which can be either a contrary or contradictory" (George W. Shields, "Fate and Logic: Cahn on Hartshorne Revisited," The Southern Journal of Philosophy, 1988, vol. XXVI, n. 3, p. 376). Craig’s eisegesis notwithstanding, Hartshorne’s position on the law of excluded middle is neither inconsistent nor in violation of elementary rules on how to form the negation of a statement.