Purtill on Fatalism and Truth

Dr. William Lane Craig

William Lane Craig is Research Professor of Philosophy at Talbot School of Theology in La Mirada, California. He lives in Atlanta, Georgia, with his wife Jan and their two teenage children Charity and John. At the age of sixteen as a junior in high school, he first heard the message of the Christian gospel and yielded his life to Christ. Dr. Craig pursued his undergraduate studies at Wheaton College (B.A. 1971) and graduate studies at Trinity Evangelical Divinity School (M.A. 1974; M.A. 1975), the University of Birmingham (England) (Ph.D. 1977), and the University of Munich (Germany) (D.Theol. 1984). From 1980-86 he taught Philosophy of Religion at Trinity, during which time he and Jan started their family. In 1987 they moved to Brussels, Belgium, where Dr. Craig pursued research at the University of Louvain until 1994.

Richard Purtill's recent contribution to the fatalism debate does not, I think, succeed in the author's intent of proving that the omnitemporality of truth implies fatalism, nor that the past is unchangeable in a non- trivial sense, nor that the consequences of his argument are not detrimental to logic and theology.

Source: "Purtill on Fatalism and Truth." Faith and Philosophy 7 (1990): 229-234.

Richard Purtill's recent contribution to the fatalism debate does not, I think, succeed in the author's intent of proving that the omnitemporality of truth implies fatalism, nor that the past is unchangeable in a non-trivial sense, nor that the consequences of his argument are not detrimental to logic and theology.

His argument gets off to a bad start by misdefining several key concepts. First, by "fatalism" he means the doctrine "that there is nothing we can do now which will make any statement about the future either true or false."{1} But this is not at all what fatalism holds. Purtill's definition leaves open the possibility that I can do something in the future to make a statement about the future now true (or false), e.g., I can do something tomorrow that will make it true (or false) today that "I shall travel to Brussels tomorrow." Normally, what we can do now affects the truth of present-tense statements, not future-tense statements or tenseless statements about the future.{2} if we omit the word "now" from the definition, it still fails to capture the essence of fatalism, for the fatalist does not deny that what we do renders statements true (or false). Rather fatalism denies that we can do anything other than what we shall do, i.e., we cannot act in such a way that a bivalent statement about the future would have a different truth value than the one it has.

Second, Purtill defines the omnitemporality of truth as the doctrine "that any statement which is true at any time is true at all times previous to and all times subsequent to that time."{3} But since Purtill thinks statements are tensed, this definition is wholly incorrect, since future-tense statements become false once the relevant events occur and remain false forever after.{4} Only tenseless propositions outfitted with appropriate dates are omnitemporally true. Though he must deny this doctrine as well, Purtill's real complaint concerns antecedent truth, i.e., the bivalence of future contingent singular statements.

Third, he defines the "unchangeability of the past" as the doctrine "that there is nothing which we can do now which will make any statement about the past either true or false, that is, the past is beyond our control."{5} Again, this definition is multiply flawed. For those who hold to this doctrine are quite willing to allow that we have power to render past-tense statements about soft facts true or false (remember J. T. Saunders's "Caesar died 2009 years prior to my writing this article"?). Moreover, those who hold to the unchangeability of the past can freely admit that we do make past-tense statements true (or false), but still deny that the past is within our control, in the same way that the fatalist can hold that we do make future-tense statements true (or false) but denies that the future is within our control. Finally, many or most non-fatalists agree that the past is not within our control if that means the power to bring about the past or to make past-tense statements true (or false). What the non-fatalist holds and Purtill wants to deny is the doctrine that we can act in such a way that, were we to act in that way, the past would have been different, i.e., different past-tense statements would have been true (or false).

With these confusions cleared up a bit, let us look at Purtill's argument for fatalism. He argues that if future-tense statements are bivalent, then for any such statement p we can form the past-tense statement "It was the case that p." Being in the past tense, this statement is now "unchangeable by me, beyond my control"{6} , in the sense defined above. Since this statement is beyond my control and it entails p, it follows that p is beyond my control, i.e., fatalism is true.

Notice that this argument depends on the assumption that "unchangeability" or, as I prefer, temporal necessity is closed under entailment. Purtill asserts that ". . .it seems as clear as anything in logic can be that the logical consequences of what I cannot change are things I cannot change. . ."{7} But this closure is far from obvious and was denied by Molina, whose view is defended by his gifted translator A. Freddoso.{8} In fact, if all past-tense state- ments are temporally necessary, as Purtill alleges, then nothing could be more obvious than that temporal necessity, like the concept "within one's power,"{9} is not closed under entailment. For if it were, fatalism would follow, and fatalism is simply incoherent, positing as it does a constraint on causally contingent events which is altogether mysterious.

Purtill would no doubt respond as he did at an APA Pacific Division symposium on this subject that "One man's modus ponens is another man's modus tolens," i.e., one may reject in this case either the Principle of Bivalence or the closure of temporal necessity, so that we are simply left with a conflict of intuitions here. Such a stand-off would, however, leave the orthodox theist's position intact; so Purtill attempts to break the deadlock by challenging the Molinist to provide a counter-example to the principle that temporal necessity is closed under entailment. Fine; in another place I have provided examples--drawn from independent discussions of the Special Theory of Relativity, backwards causation, time travel, precognition, and Newcomb's Paradox--to a similar challenge from John Fischer of past events which are as "hard" or fixed as God's past beliefs and which entail or imply future events which are nonetheless within our power.{10} if one holds with Purtill that such events are temporally necessary, then it seems obvious that such necessity is not closed under entailment. In any case, the instances of divine foreknowledge or bivalent future-tense statements are not unique.

Moreover, the Molinist can strengthen his case by arguing that it is plausible that future contingent statements are bivalent. This I have also done in another place.{11} By so doing, he renders plausible the thesis that temporal necessity is not closed under entailment.

The orthodox theist need not embrace the Molinist alternative, however. If he prefers, he can take the Ockhamist position instead (or as well), viz., that the relevant past-tense statements are not temporally necessary.{12} Purtill essays to refute this rejoinder, but his reasoning is vitiated by a fundamental misconception: that "There are two kinds of facts about the past: hard facts which cannot be changed and soft facts which can be changed."{13} But soft facts cannot in fact be changed; given that they are facts, they are as unalterable as hard facts. But they differ from the latter in that they are counterfactually dependent upon future contingents, such that were the future contingent event not to occur, the event expressed by the soft fact would not have occurred. Hence, it is the case, pace Purtill, that the set of future-tense statements true at any past time t is neither growing nor changing. Therefore, Purtill s argument for fatalism fails.{14}

This settles the issue; but the Ockhamist and Molinist might seek to strengthen their case against the would-be fatalist by pointing out the counter-intuitive consequences of denying the Principle of Bivalence.{15} Purtill tries to avoid these consequences by arguing for an infinite multi-valued logic for future contingent statements, the values being interpreted as probability functions. But a fundamental difficulty with this alternative is that it does not seem to make sense to speak of degrees of truth for a proposition. Probability functions are much more plausibly construed as epistemic in nature. A proposition is either true or else it is not-true, and its probability of being one or the other concerns our cognitive relation to it. Purtill would substitute for this simple structure an extravagant complexity without intuitive warrant.

Finally, Purtill seeks to mitigate the theological consequences of the denial of Bivalence, viz., the denial of divine foreknowledge. These are, indeed, serious, both biblically and theologically. Biblically, God's knowledge of future contingents is clearly taught in both Testaments, and numerous examples of prophecy of future contingents may be found.{16} Especially significant Christologically is that such foreknowledge is ascribed to Jesus Christ.

Purtill makes the amazing assertion that "every Biblical passage...about God's knowledge of the future is quite consistent with the view that some statements about the future are neither true nor false and thus are not known by God," and he challenges anyone who thinks otherwise to produce such a biblical passage.{17} But the question is not whether every relevant passage of the Bible is consistent with some statements about the future being unknown to God; the question is whether all the relevant passages of Scripture are so consistent. In any case, it is not difficult to cite passages that assume God's knowledge of important classes of future contingents, for example:

Even before a word is on my tongue, lo, O Lord, thou knowest it altogether. (Psalm 139.4)
Peter, an apostle of Jesus Christ, to the chosen ones. . .according to the foreknowledge of God the Father. . .(I Peter 1.1-2)
Though in a technical sense these statements are compatible with God's not knowing some (other) future contingents, Purtill can hardly admit these, for if God knows our very thoughts before we think them and the identity of the saved before the creation of the world, then any vestige of ignorance left in God about the future will hardly be of much importance to us!

As for the issue of prophecy, Purtill attempts to account for prophetic statements as predictions of either causally determined events or events which God has determined to bring about Himself. But this will hardly do, for Scriptural prophecy is presented as being the revelation of future events which are not present in their causes, and while many prophecies could be construed as statements about God's intentions, the Scriptures contain numerous examples of prophecy concerning events not brought about by God, especially sinful human acts.{18}

Concerning the imagined charge that his view is disrespectful and blasphemous, denigrating the power of God, Purtill answers that ". . .it is not really respectful of God to attribute to Him impossible powers."{19} Granted; but what disrespectful of God is to say that something is impossible for Him when He has revealed it to be the case. If certain Christian philosophers do not find the preferred solutions to the problem of theological fatalism convincing, why not simply admit with the Psalmist

Such knowledge is too wonderful for me; it is high, I cannot attain it. (Psalm 139.6)
rather than jettison a doctrine which is taught with reasonable clarity through- out the Scriptures? Would this not be the better part of intellectual humility?

As for the theological consequences of his position, Purtill does not really discuss these, but raises instead difficulties with the timelessness of God. After Duns Scotus, however, most scholastics rejected the Boethian solution to theological fatalism,{20} so it is not necessary to pursue Purtill's objections to that doctrine. Rather the truly serious theological consequence of Purtill's position is that it renders the doctrine of divine providence and sovereignty virtually unintelligible. For without divine middle knowledge (which entails divine foreknowledge) it seems inexplicable how God could sovereignly direct a world of free creatures toward His previsioned ends without violating their freedom. By contrast, the Ockhamist, or better, Molinist, view of God wins all the advantages Purtill desires in terms of God's dynamic interaction with His creatures, yet without sacrificing either divine foreknowledge or human freedom.


{1} Richard L. Purtill, "Fatalism and the Omnitemporality of Truth," Faith and Philosophy 5 (1988):185.

{2} There are exceptions: e.g., by committing suicide now, I render the proposition "I will have lunch tomorrow" false.

{3} Purtill, "Fatalism," p. 185.

{4} Thus, he errs in stating that his future-tense statement F1 will at some future time be true or its denial be true. On the contrary it is already either true or false.

{5} Purtill, "Fatalism," p. 185.

{6} Ibid, p. 186.

{7} Ibid.

{8} Luis de Molina, On Divine Foreknowledge: Part IV of the "Concordia," trans. and ed. with an Introduction by Alfred J. Freddoso (Ithaca, New York,: Cornell University Press, 1988), pp. 55-62.

{9} See Joshua Hoffmann and Gary Rosenkrantz, "On Divine Foreknowledge and Human Freedom," Philosophical Studies 37 (1980): 289-96.

{10} See William Lane Craig, "Nice Soft Facts: Fischer on Foreknowledge," Religious Studies 25 (1989): 235- 46.

{11} William Lane Craig, Divine Foreknowledge and Human Freedom (Leiden: E. J. Brill, 1990) chap. 4. Viz. (i) the same facts which serve to make past- and present-tense propositions true also serve to make the relevant future-tense propositions true; (ii) if future-tense propositions are neither true nor false, then neither are past-tense propositions; (iii) the tenseless versions of future contingent singular propositions would seem to be always true or false.

{12} See the fine study by Alfred J. Freddoso, "Accidental Necessity and Logical Determinism," Journal of Philosophy 80 (1983): 257-78.

{13} Purtill, "Fatalism," p. 186.

{14} The element of truth in Purtill's argument is that, contrary to the impression given by some recent contributors to the debate, theological fatalism represents, in fact, no advance over logical fatalism. For the temporal necessity ascribed to God's past belief can be more simply ascribed to the past state of affairs constituted by some future-tense proposition's being true (or false). In the standard formulations of the argument for theological fatalism, the premiss concerning God's belief can be replaced with a premiss concerning the antecedent truth of some proposition. For example, A. N. Prior's version actually omits all reference to God, depending merely upon his

6. If it was the case n time units ago that p, then necessarily it was the case n time units ago that p.
And despite his protestations, it seems to me that Pike's version is also so reducible. For one could replace his (31) with
31*. If Jones does A at t2, then it was true at t1 that "Jones does A at t2"
or with
31**. If Jones does A at t2, then it was true at t1 that "Jones will do A at t2."
Any successful attempt to remove the temporal necessity of such states of affairs will inevitably render God's past beliefs temporally contingent as well, as a moment's reflection on Freddoso, "Accidental Necessity," shows.

{15} As I have done in Craig, Foreknowledge and Freedom, chap. 4.

{16} I have done an exegetical study in The Only Wise God(Grand Rapids, Mich.: Baker, 1987), pt. 1.

{17} Purtill, "Fatalism," p. 189.

{18} Purtill seems willing to go so far as to say Judas's denial was inevitable and, hence, predictable because it lay on every future branch. But this is fantastic; on a libertarian view, there must be a branch in which Judas does not deny Christ, otherwise he sins necessarily, not freely. Purtill might back off to the view that on every future branch feasible for God to actualize, Judas sins; but then there is absolutely no way for God to know this apart from middle knowledge, which Purtill must deny. Mere probability is not sufficient for prophecy of future contingents.

{19} Purtill, "Fatalism," p. 190.

{20} See my The Problem of Divine Foreknowledge and Future Contingents from Aristotle to Suarez, Studies in Intellectual History 7 (Leiden: E. J. Brill, 1988), chaps. 5-8.

Copyright (C) William Lane Craig. All Rights Reserved.