Is Innate Foreknowledge
Possible to a Temporal God?
by Kel Good
A very interesting attempt to avoid the conclusion that God's foreknowledge is inconsistent with creaturely freedom is an essay entitled "Divine Foreknowledge And Future Contingency,"1 by William Lane Craig. Craig's essay begins with an opening statement of his belief that
The biblical view of divine omniscience entails that God knows all past, present, and future states of affairs. In particular, He knows future contingents . . . Any interpretation which involves a denial of God's knowledge of future contingents is unbiblical and subChristian and must be rejected by the church.2
He then proceeds to prove this thesis in three sections. The first presents the process theology denial of foreknowledge (one which virtually coincides with the Moral Government position). The second examines arguments to prove this view and presents critiques of the same. His final section contains a statement in favour of the original thesis. We will focus on the central section where Craig attacks arguments much like the arguments Moral Government advocates hold to, arguments that suggest foreknowledge and free will are incompatible. Once again he divides this section into three subsections, the first dealing with time theory and truth as correspondence. The second looks at the arguments for theological fatalism. His final subsection distinguishes what Craig considers to be the true basis of divine foreknowledge as opposed to what is usually assumed by those who seek to deny the concept.
The A-Theory Of Time And Truth As Correspondence
Craig begins his first subsection by describing two theories of time.
The A-theory sees time to be in constant change, every moment of history
coming into being in the present and passing into the past. The future
does not yet exist but is created in the dynamic present. On this view
moments of time pass through the three stages of time, constantly changing
until concretised in the past. This is the view Moral Government advocates
hold to. The B-theory of time sees every moment of history existing either
before or after other moments, with each moment staying at the same place
along the time line. This view is essentially the timeless view of history
that Moral Government theology denies. Craig identifies himself with the
A-theory and the belief that the future is genuinely non-existent. He
is going to seek to argue that although God is temporally3 located in the present, he still possesses knowledge of all future free
acts. I am going to argue that in doing so, he forfeits the A-theory position
and must ultimately identify with a B-theory of time.
Craig proceeds by telling us that those who deny foreknowledge on the basis of an A-theory view of time do so through a belief that statements concerning future contingent events (free choices) cannot be true. Since they are not true they are not knowable. If this is so they could not be an item of God's knowledge. Those who deny foreknowledge have held two views regarding the truth status of future contingent statements. On the one hand, men like Charles Hartshorne, in an attempt to preserve the law of excluded middle (for any given statement, it is either A or not A), have said that such statements are false. On the other hand, it has been believed that such statements have no truth-value. This second conclusion appears to be correct, and is based on the fact that the future in no sense exists and hence truth claims cannot be made, correctly or incorrectly, concerning future contingent statements. We are not talking about anything existent when we speak of what will be. Of course the laws of logic can be used of non-existent things when we create them in our minds. When using imagination, the statement "Unicorns have one horn, and no more than one" has a truth-value. It is true. This is because we have created and defined in our minds what a unicorn will be. The same statement applied to reality, what exists, has no truth-value. It is neither true nor false since unicorns do not exist. So it is with statements about future contingent statements. They have no truth-value since they are indeterminate.
Craig points out that this reasoning may need to apply to all future statements since nothing future exists. Usually here statements about causally determined future events are considered to have a truth-value, since they exist in their causes. But what if a free agent changes the causes that exist? Then a statement about the future could be true one day and cease to be true the next. Craig suggests this possibility would require the consistent thinker to deny truth-value to any future statement. It is interesting he should use this example as it proves the precise point we are trying to make. A statement affected by a future free choice is inherently undetermined and hence cannot have a truth-value.
In order to avoid this conclusion and preserve the possibility of foreknowledge, Craig must show that those who deny future contingent statements can be true are in error. He attempts this by seeking a faulty assumption in his opponents' understanding of truth as correspondence. What he needs to show is that a statement about a future contingent event does not require the present existence of the event it describes and/or the present existence of the causes of the event it describes in order to be able to be pronounced true. He urges that only present tense statements require the events to exist in order for them to be true. What is required of future tensed statements is that they describe what will be, how things will happen to work out. If they do this then he feels they can be recognised as true statements. How he can make such a comment without recognising the determinism of it is difficult to understand. If it can be said today what will happen tomorrow, then the future is determined. It in no sense is contingent. Contingency deals entirely with possibilities, not facts. Facts cannot be changed; possibilities may or may not come to be fact.
Three Reasons Things Do Not Have To Exist To Be True
We are given three reasons why we ought to question that a correspondence
theory of truth requires the present existence of the event that such
a statement describes. First, we are to note that for any event described
with a present tense statement, a future tense statement could have been
formulated in the past to describe the same event. Since both statements
describe the same event, and since the present tense statement is true,
does this not require that the future tense statement uttered in the past
was also true? Throughout his essay Craig ignores his own professed A-theory
view of time. The reason the present tense statement is true is because
the event it describes exists. Yesterday the event did not exist, and
might never have. The two statements do not describe exactly the same
event. The present tense describes an event that is happening.
The future tense statement uttered in the past described what might
happen. One cannot ignore the time at which a statement is uttered
and claim that the same event is being described.
Craig's second reason for questioning the necessity of the present existence of the event described by a true statement is that if this is the case, statements about the past cannot be true. The past no longer exists, so how can we make true statements about it? The key is simply that. It no longer exists. It has existed. It has come to be and is now remembered. The future has not yet been caused to exist. This is the fundamental difference between the past and present on the one side, and the future on the other. Both past and present have come to exist. The future never has. Craig's second reason is also no reason.
The final reason we are to doubt the above requirement for a statement's truth status is that tenseless versions of future statements can be constructed by using a time indicator, the date for example, as they can for past and present tense statements. In neither of these last two types of statements is the truth-value lost, even though it now describes the event tenselessly. Since a future tense statement can be worded tenselessly to describe the same events that such tenseless versions of past and present tense describe, a future tense statement must also have a truth-value. Craig uses as his example the invasion of Normandy by the allies and shows that this past event can be described tenselessly. I wonder why he did not use a future event where we do not yet know its truth for his example. This "reason" is erroneous for the same reason as his second one. There is a fundamental difference between the future and the other two time tenses. The former does not exist, and never has. The latter either do or have existed.
Logic Or Fact?
Craig's "reasons" fail to be genuine reasons to doubt the requirement of the present existence of the event described, or at least that it exist in its causes, as a minimum requirement for a statement to have truth status. Since Craig admits the future does not exist, it is impossible that statements describing future contingent events can now have a truth-value. Far from it being the case, as he claims, that the real confusion stems from our failing to distinguish truth status from knowledge status, the real confusion stems directly from the belief that a statement about some event which in no sense exists, can be true or false. Our problem is not that we cannot know future contingent events and that we misinterpret this as meaning that statements of such events do not have a truth value. There is nothing there to know. Craig tries to embarrass the conclusion that statements about future contingents have no truth value by saying that if this was true we would not be able to affirm the truth of such statements as "The Vice President will or will not run for the Presidency,"4 and the falsity of such statements as "The President will both live out his second term and not live out his second term"5 since
statements which are disjunctions or conjunctions depend for their truth value on the truth values of their component sentences. If, then, the component sentences are future contingent statements having no truth value, the entire disjunction or conjunction will also lack a truth-value.6
As usual, our eyes are being diverted from the genuine issue. The truth and falsity of the statements given as examples derive not from what they are discussing (future contingent events) but from the fact that the described hypothetical events form a tautology and a contradiction. One of the two halves of these statement pairs must become true in the future, since they describe the total set of possible states of affairs. Neither half of the given statements need be true now. The laws of logic would apply to any statements whether they describe future, present, or past states of affairs. One does not require a truth-value, but merely a definition of the component sentences to determine the truth or falsity of the hypothetical statements. Given the definitions, we can say the statements are true or false, even if they have no actual application in reality. We can make truth statements regarding Unicorns, based on the definition we give to a Unicorn. It is like a horse, with one horn in its forehead. Based on this definition, the statement A Unicorn has four legs is true. It is not true because in fact the definition of a Unicorn has any real application in the real world. The truth of the statement derives from the assumed definition. So it is with Craigs examples contrasting future possibilities. Their truth or falsity derives from the definitions of what each half of the statement pairs would mean if they occurred. Since each possible outcome and its opposite exhaust the possible outcomes, it is true that one or the other must come to be. The first statement is true because one or the other of these statement halves must come to be. The second statement is false because logically, a particular possible outcome and its opposite cannot both come to be. Again the proposed examples form no difficulty for the view that future contingent statements have no present truth-value.
Craig continues his discussion by seeking to deny the conclusions of
what he calls "theological fatalism. Essentially, it is the claim
that if God necessarily knows what will come to pass, and he foreknows
a certain event, then that event is necessary and determined. This is
of course what Moral Government argues. Craig's answer to this claim is
to point out that in order for the conclusion of the argument - that the
event will come to pass - to be necessary, both preceding premises must
also be necessary. He suggests this is not so since although it is necessary
that God foreknow the future, it is not necessary that God foreknow the
particular event which he did foreknow. It is conceivable that God have
foreknown some other event. Thus, the occurrence of the particular event
is not necessary. Something else might have occurred. Although
it is true that God foreknew that Jones would mow his lawn that Saturday,
it is conceivable that he have foreknown Jones to not mow his lawn.
God knowing the one rather than the other is in no sense absolutely necessary.
We are being subjected here to the famous philosophical distinction between logical and factual necessity. Since there is no logical necessity for that particular event happening, the description of the contrary event is not logically contradictory; there is no logical necessity that that particular event take place. Thus, it is not a necessary event. As is often the case with this distinction between logical and factual necessity, this is interesting but irrelevant. We are talking about the real world. We are not claiming a logical necessity. We are trying to see what in fact is the case. Whether or not we could sub in some other event conceptually, the fact remains that if God is correct in foreknowing the event, then in fact it must happen. It is factually necessary. Nothing else in fact can happen. Craig cannot insist that it is not simply the future act's conceptual possibility of being other than it was, but its contingent nature, which makes God's knowledge of what occurred unnecessary. This is to assume what must be proven; viz. that a foreknown act could in any sense be contingent and free.
Craig's essay is full of the acknowledgement that foreknown actions are factually necessary. He says a future contingent statement is true when it describes what will happen. He later gives examples where if we knew what a being precognized and tried to prevent the event, we would fail. If this is not fatalism, what is? He misses the whole point when he claims that fatalism is incoherent because it says that everything happens with complete contingency and yet is somehow constrained because foreknowledge exists. He seems to think fatalists are arguing that men are free but that somehow God's knowledge of their free choices still fates their free choices. He says fatalism and determinism are not the same because determinism claims everything happens through direct coercion, whereas fatalism claims men's actions are not caused, but free. While this distinction may hold in theory, the point of the fatalist is that if future free acts could be foreknown then such freedom is only an illusion and determinism really is true. Consistent fatalists do not maintain freedom is consistent with a determined future. The point is just this. If there is freedom there can be no determinism. But foreknowledge requires a fatalist view and therefore a determinist view of history. Fatalism is the same as determinism because the fatalist is simply claiming that although we cannot see the connection, in some way what appear to us to be free choices are really caused.
Could We Change The Past?
According to Craig, and here I think he is correct, the difference between past and future is not that the former cannot be changed while the latter can. The future does not exist; there is nothing there to change. The true difference is that the past is causally closed while the future is causally open. Craig admits that the idea that one could change the past is illogical since such reverse causation is the claim that an effect precedes its cause. But on an A-theory view of time the future cause would have no existence whatsoever when its effect occurs. Such action would not be possible since it is impossible for nothing to cause something. But we are told that the claim that God foreknows is not a claim of reverse causation. It is not believed that we can act in such a way that the past be changed. Rather, "we have the power to act in such a way that were we to act in that way, God would have believed differently.7 Craig calls this a "contrary-to-fact or subjunctive conditional statement. It does not express the ability to change or undo the past, but the ability to act such that the past would have been different than it is.8
Nor is this an appeal to some metaphysically objectionable backward causation. I do not retroactively cause cognitions to spring into God's mind. Rather what I shall do determines which future tense statements about my actions are true and false, and God has the essential property of knowing only and all true statements. Thus, he foreknows my future acts.9(Italics mine)
Craig is trying to avoid, in fact the whole purpose of his essay is to
avoid, the necessary conclusion that temporal foreknowledge must be caused
by future events in a reverse causation manner. Note how he distinguishes
between our act causing its truth, and God knowing all true statements.
In other words, our act does not cause God's knowledge of it in the
past. That would be reverse causation. Our act only causes the truth
of the statement about our act. But since God just happens to know everything
that is true, he also knows what our future act will be. Our act does
not cause something in the past therefore; it only causes what is now
the case. But if we acted differently now, this of course would alter
what is true about our act and hence God's past belief would not be changed,
but would have been different from what it was. Thus Craig argues
that "though God's foreknowledge is chronologically prior to our
action, our action is logically prior to God's foreknowledge.10 God's knowledge does not cause our act; our act causes the truth of God's
The strength of this argument rests initially on the success with which Craig has argued that a statement can be true when the events it describes are still future and contingent. This attempt we have seen to be a failure. Since the future does not exist and since a future free choice is totally indeterminate, it cannot have a truth-value that God would know. But all this entirely aside, has Craig escaped the accusation of reverse causation? He thinks he has because he considers God's foreknowledge not as something resulting from his perceptions of the future, but from an innate knowledge of truth.
The Basis Of Divine Foreknowledge
In his final subsection Craig distinguishes between the empiricist and
rationalist basis of foreknowledge. He feels that most if not all attempts
to deny foreknowledge arise from the empiricist assumption that God learns
his knowledge of the future. According to the rationalist model, which
Craig espouses, God does not come to know future events, he just knows.
There is no causal connection between future free events and what God
knows. He simply has an innate knowledge of all true statements, including
those whose truths are determined by future free choices. Their truth
is determined by the free acts themselves, but they can still be said
to be true before they occur since all that is required for their truth
is that they describe what will happen. Again, for Craig this is
not the same as saying what must happen. Though factually certain,
it is not logically necessary.
One gets the impression that it would be an exercise in futility to argue this around much further, when we are willing to admit something will certainly occur and not recognise this as determinism and fatalism. But let us take a final stab. We have seen that future contingent statements describe what does not exist and hence cannot have a truth-value. But for now let us suppose that they do. If God knows all true statements of this sort before they occur by simply knowing them then apparently there is no reverse causation producing God's knowledge. There is no such connection. He simply knows. Is there no more problem? Far from it. Even if God has not "received" this knowledge, he still knows these propositions are true at a time prior to when they are caused to be true. What has caused them to be true? My actions. When have my actions occurred? After the time at which they are known to be true. This is reverse causation. Even if I do not cause God's knowledge directly through reverse causation, I do cause the truth of this knowledge through reverse causation. And Gods knowledge is the effect of this truth. For if it is true for God before it happens, and if it would be false if I did not do the act, then my doing the act is the cause of its being true. This cause follows chronologically the effect of its being true, which effects its being known to be true at a time prior to the initial cause in this chain. It is known to be true before it happens, yet if it did not happen it would not be true. It does not solve the problem to say that if I did something different then God would have believed differently. The issue is whatever I do, the truth of it exists before that which causes it to be true exists.
Craig seems to sense the truth of this accusation for when he describes the relation between God's innate knowledge and the truth of a future contingent statement, he seems to distinguish between the agents act causing the truth of such statements and how the agent will act causing this truth. If my act causes its truth then this is reverse causation as we have seen. Craig seems to choose his words very carefully by stating "what I shall do determines which future tense statements about my actions are true or false, . . ."11 In this case what is causing the truth of such a statement is not my act in the future, but the fact that this is what my act will be. But what else is this fact but the truth that my act will be this? Thus the truth of the statement is caused by the truth of the statement. This is absurd. Only the act itself could cause a statement describing the act to be true. But this act exists as a cause after its effect and Craig himself has pronounced this metaphysically impossible. There is no way to escape the accusation of reverse causation on an A-theory view of time if God is said to foreknow.
The Root Problem
But here we must recognise the root problem in all views of foreknowledge of future free choices. It does not matter whether we say God is timeless and hence knows all of time, or that God is a being of duration and simply has a type of "visionary" knowledge of the future, or that he simply knows all true statements about future free acts. Either way everything must be determined; a B-theory of time to use Craig's terminology. Craig has simply substituted a rationalist description, "innate foreknowledge," for a empiricist version, "visionary knowledge." What have not changed are the frozenness of history and the illusory nature of any alleged freedom this produces. If God has innate foreknowledge and knows all true statements about all points of history, then history is true. It is the only history that could be. Whether God is timeless and "sees the future," or is a being of duration and "knows what is true about the future," it could not be otherwise. Hidden beneath all these intricate arguments, Craig in the end is shown to be a B-theorist in practice, whatever his profession to the contrary. So is everyone who espouses foreknowledge of future free choices. If they are there to be known, or if they are now true, the conclusion is the same. History is frozen and freedom is an illusion. Only a God who lives in time and does not know future free choices could be a God where freedom could exist for himself and his creatures. Craig's view of God as temporal 12 is attractive in that it escapes many of the problems regarding God's power and personality, which various philosophers have shown to be associated with a timeless view of God. Craig cannot make this move and at the same time maintain God's absolute foreknowledge of all future contingents. If God is to be free to exercise his power and personality and respond to his creation in a living way; if all his actions and the actions of his creatures are to escape determinism; then innate foreknowledge cannot be possible to a temporal God.
1. Ronald Nash (ed.), Process Theology (Grand Rapids, Baker Book House, 1987), pg. 95+115.
2. Ibid., pg. 95.
3. This belief on Craig's part does not entail to him a denial of the timelessness of God. In another essay "Creatio Ex Nihilo," Ronald Nash (ed.), Process Theology (Grand Rapids, Baker Book House, 1987), pg. 145+173, he argues against the possibility that there could be an endless duration extending into the past. He uses this argument to prove that the universe had to have had a beginning, hence it must have been created ex nihilo. He says that God's eternity cannot be seen as one of infinite duration into the past but that "God exists timelessly without creation, and with the first event time began." (pg. 151+152) This timelessness does not, however, seem to coincide with the traditional view that God is present to all points of history. Craig definitely believes that upon creating the world God became related to time in such a sense that the future is truly future to himself as well. Craig is somehow claiming that God's existence before creation did not have a before or after.
4. Ibid., pg. 103.
5. Ibid., pg. 103.
6. Ibid., pg. 103.
7. Ibid., pg. 106.
8. Ibid., pg. 106.
9. Ibid., pg. 106+107.
10. Ibid., pg. 106.
11. Ibid., pg. 106.
12. See footnote 3.
©1997 Kel Good, used by permission; This publication may be copied freely as long as no alteration is made to the text. For more information write: Kel Good, 918 - 16 Ave. N.W., Suite 496, Calgary, AB, Canada T2M 0K3, or Email to firstname.lastname@example.org.