Skip to content

Divine Omniscience and Human Free Will

February 26, 2011

For those that are interested in something that’s a bit more technical than my usual, below is my essay on divine omniscience and human free will, which I’ve finally gotten the marks back for. I try to evaluate an argument put forward by Gerard Hughes that attempts to reconcile these two things – unsuccessfully in my opinion. If you want to find out more, read on…

In The Nature of God Gerard Hughes provides an argument that attempts to reconcile divine omniscience with the occurrence of undetermined or free acts. In this essay I will analyse his argument and determine how effective it is at reconciling these two seemingly contradictory ideas. In particular, it will be necessary to establish whether or not the argument successfully allows a reconciliation to take place whilst allowing the notion of a wholly simple God to remain intact.

Hughes main argument can be broken into two parts. In the first, he argues to the conclusion that God’s knowledge is necessary. He then shows how, in his view, this is compatible with human free will. The conclusions of this argument will be controversial in the light of divine simplicity.

The fundamental problem when discussing the nature of God’s knowledge is his relation to time. Hughes says that ‘[There is] no good reason to abandon the timelessness of God’ (Hughes: 102). He is taking the Thomist view of God as necessary, simple and outside time. We cannot therefore used tensed language when discussing what God knows, but rather must resort to untensed statements of what God <know>.

God is therefore given an eternalist position; all times are of equal ontological status to God. From later arguments I believe that Hughes is a presentist, in that he gives the present an ontological privilege. By our being time-bound, it is only the present moment that “exists” in any meaningful sense. We may alter the logical properties of statements or states in the past by our actions in this present, but we cannot alter ‘any intrinsic property of the earlier event’ (Hughes: 104). This position is necessary in order to be able to make statements of the sort given by Hughes using the terms “is”, “was” and “will be” (Hughes: 104). These statements can be made by humans but not by God, who is restricted to the ability to <know> statements such as ‘The sentence “I am writing now” would be true if said by Gerry Hughes at 11.30’ (Hughes: 102). As pointed out by Kelly James Clark, this would be a defect in God ‘if omniscience is the doctrine that God knows all truths’ (Clark: 114), however Hughes has anticipated this objection by saying that

It would be less misleading to say that God <know> all the true statements which God could make; or better, that God <know> whatever states of affairs underpin the truth of all true statements, both those which God could truly make, and those which I might truly make. (Hughes: 102)

The second half of this statement is particularly important when we remember that in the Thomist model God’s knowledge is of things, not statements about things. Hughes has thus established a coherent framework for talking about what God can <know>. God has knowledge by being aware of the state of affairs giving rise to this knowledge, via his superior position with regards to the ontology of past, present and future states.

Hughes’ next step is to show that the things that God <know>, he <know> by necessity. God possesses knowledge of the entire “history” of the cosmos. Hughes tells us that this knowledge ‘might be other than it <be>’ (Hughes: 105). This seems to contradict the necessity of God’s knowledge, but a strong argument is provided that actually helps clarify this necessity.

In what sense should the word “might” be taken in the quote above? According to Hughes’ argument, it is to be taken in the same sense as in the sentence “I might not have written this essay”. Clearly it was a possibility, but only one possibility became actual: I did in fact write this essay. “Might” implies some element of contingency, but I no longer have causal access to the event to which I am referring, which is now accidentally necessary.

The parallel in God to my lack of causal access to the past is God’s lack of causal access to more than one actual history of a time-bound cosmos. From God’s point if views, then, it <be> not accidentally possible that there should be another actual history of this cosmos… In a slightly extended use of the term, the entire history of the cosmos is, from God’s point of view, accidentally necessary. (Hughes: 105)

This resolves the use of the word “might” in a description of God, but immediately presents a further problem. Given that God’s knowledge is accidentally necessary, and that my future actions are part of that knowledge, how do I retain any semblance of free will?

…it is not in our power to bring it about that the actual history of the cosmos is other than <be>, though it is in our power to determine what that actual history <be>. (Hughes: 105)

This seemingly contradictory statement relies on the eternalist view of God that Hughes espouses. For God, a state of affairs that for us is future (and contingent) already has an ontological reality. For us to make this state of affairs present we must use the causal powers that we possess, and we can do so in a contingent way, but any contingent act brings about a state of affairs that already <be>. That is, a future state of affairs may be eternally accidentally necessary from the point of view of God but as yet contingent for us.

Here we arrive at the crux of the argument. The existence of a future state of affairs, already existing for God, relies on our contingent actions. But,

Implicit in the classical position that God is in all respects simple, is the view that the source of God’s knowledge cannot be anything other than God himself. (Hughes: 107)

Hughes now turns to the problem of reconciling this view with the idea that we, as causal agents, affect God’s knowledge in some way.

From the start, Hughes rightly rejects God as an explanation for the occurrence of a contingent event. If free will exists we exercise it independently of God. To posit God as some form of cause of undetermined events would undermine any truly free explanation of human action. We have to admit that an undetermined event such as radioactive decay has no causal explanation for taking place at a particular time. It is this move that is going to cause Hughes problems, as any causation we as humans have on a state of affairs seems, by necessity, to transfer to God’s knowledge of that state of affairs, and hence to God himself.

Consider a free action A that I perform at time t2. At time t1, prior to t2, there are no grounds for believing that A would take place. However, Hughes holds the realist position that it would have been true at t1 to say that A would take place. Grounds for belief and the truth of a statement are to be separated. The truth of the statement (at t1) “Action A will be performed at t2” is a logical consequence of the action A. The epistemic status of knowledge of a state of affairs is determined by the event, and so the truth, and God’s knowledge, depends on the action and event that brought about the state of affairs in question.

Hughes uses an analogy with a football match to illustrate this point well; a goal can only be a match winning goal when the final whistle is blown, and the action of blowing the whistle is what causes the truth of a statement that it was a match winning goal, before or after the whistle is blown.

This argument works. The major argument against it would be to say that a truth claim cannot be true or false before the event to which it refers takes place. In our presentist state this view may be argued for, but given that God has knowledge of “existing” states of affairs that are future from our perspective, we have to acknowledge that the realist view is, in a sense, true for God, who is outside of time. Hughes has thus established that, given the existence of free actions (or at least undetermined events), God’s knowledge is, in some sense, dependent on something other than God.

Though obviously it is not the case that God’s eternal knowledge of contingent events is incomplete until those events occur, it is the case that the occurrence of those events is logically and epistemologically prior to God’s knowledge of them.

The first half of this sentence is presented without justification, and this will turn out to be an important problem. However, up until this point Hughes has presented a coherent and logical argument. There are potential problems, but these are overcome. The challenge presented by the argument is that it must be made to fit into a particular model of God.

Hughes quickly acknowledges that his conclusion causes problems with the notion of simplicity, a notion that Hughes holds as being true.

[The conclusion] weakens the sense in which God’s knowledge can be said to be in all respects necessarily as it is; and hence it weakens the sense in which God’s knowledge can be said to be identical with God’s existence. (Hughes: 112)

Even more damningly,

[It is] unclear whether the necessity of God’s existence automatically transfers to God’s knowledge. I think the conclusion must be that it does not.

If we follow Hughes in his acceptance of the conclusions above, the argument is seriously undermined. He is not trying to just reconcile divine omniscience and human free will. He needs to reconcile these in the setting of a wholly simple God. If there is not a transfer of necessity in God from existence to knowledge, then how can these attributes be said to be identical? Hughes does not provide much discussion of how he proposes to reconcile his conclusion with his model of God, but it is of central importance to the acceptance of his argument as a whole that he is able to do so.

At this point a number of responses are possible. The first is to reject the argument as a whole without question, as it appears that it fails to work in the required setting. Secondly we could follow one reviewer in saying that, sadly, “his resolution of the conflict is too brief to evaluate” (Clark: 114).

A third response, following Sadowsky, would be to ask just what it is that Hughes has shown.

Hughes says rightly that there is a sense in which God’s knowing is epistemologically dependent on the object. But does it follow that it is ontologically dependent on the object? Is the Sun ontologically different because it shines now on this object and now on that object? It depends on the object only in the sense that the object must be there for it to shine upon it. Knowing is something that God does and not something that happens to God. (Sadowsky: 110)

God’s knowledge is of states of affairs. If a state of affairs were to be different than it is, the actual history of the cosmos would be other than it <be>. This would mean that God’s knowledge, and hence God’s existence, would be other than it is. God’s knowledge might, as Hughes says, <be> other than it <be>, but it is not, due to the fact that certain states of affairs have and will arise. God’s knowledge exists because the states of affairs to which it refers exist to God, the eternalist; God’s knowledge of A exists if and only if A exists, and hence ontologically depends on A.

So for a contingent action or event, it is more than just the truth or falsity of a statement that is at stake. The truth of a statement automatically transfers to the existence, in some sense, of God’s knowledge, so if we have successfully argued that God’s knowledge is epistemologically dependent on a contingent event we have to concede that it is ontologically dependent as well.

We thus undermine completely the notion of simplicity of God. God’s knowledge is not necessary; as Hughes proposes, there is no transfer of necessity from God’s existence to God’s knowledge. If existence and knowledge were not identical in God this would not raise a fundamental problem, but that would be to surrender simplicity, something that Hughes does not want to do.

It could be argued that God and his creation belong to completely different ontological “groups”, and use this to deny that any transfer of ontological dependence on to God is possible. If such an argument was successful then it might be the case that, despite epistemological dependence, the existence of God’s knowledge doesn’t depend on his creation. However, this approach is not intellectually satisfying. It is to get around the problem by purposely removing God from the realm in which we can have any meaningful discussion about his attribute. Any discussion of a timeless God who is radically different ontologically from his creation would be very limited in the language that it could use.

Despite successfully providing an explanation of how divine omniscience and human free will can both exist, Hughes’ reconciliation fails in the setting of a wholly simple God. If we are willing to surrender this simplicity, we have an effective and cogent argument, but we lose an aspect of God that is a central tenet of many theologies, and central to Hughes’ beliefs.


Hughes, Gerard, The Nature of God (London: Routledge 1995)

Clark, Kelly James, Book Review, The Philosophical Quarterly 1998 Vol. 48 No. 190 p.113-115

Sadowsky, James, Book Review, New Blackfriars 1996 Vol. 77 Issue 901 p.109-110


No comments yet

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: