Omniscience and why God lacks it

In order to show that God is not omniscient, Lovering first argues that being omniscient requires something beyond having maximal propositional knowledge, namely, that one always has maximal experiential knowledge. Lovering says that by ‘experiential knowledge’ he means:

This definition of experiential knowledge is not quite suited for Lovering's purposes, since it ultimately only requires of experiential knowledge that its object be non-propositional, and yet people are non-propositional, as are many objects and properties that Lovering would not want to count as experientially knowable. Perhaps, rather than defining ‘experiential knowledge’, it suffices just to state that X is experientially knowable iff there is something it is like to X.

Maximal experiential knowledge is knowledge of everything experientially knowable. Lovering offers two reasons for including maximal experiential knowledge in the ‘concept of omniscience’. First, unless there is reason to exclude them, ‘it seems the concept of omniscience should include all varieties of knowledge’.Footnote ³ And Lovering says no such reason has been offered. If he is right, then, since experiential knowledge is a kind of knowledge, an omniscient being should have it maximally. Lovering's second reason is this. If God's omniscience is limited to maximal propositional knowledge, then there is a greater conceivable being than God, namely, one otherwise like God but who has maximal experiential knowledge. But tradition has it that God is the greatest conceivable being. So omniscience must include maximal experiential knowledge.

Lovering goes on to argue that God cannot always have both maximal experiential knowledge and maximal propositional knowledge, on the grounds that God's having maximal knowledge of one kind at a time is incompatible with his having maximal knowledge of the other at that time. Why is it incompatible? Because, according to Lovering, if God has maximal experiential knowledge then he knows what it's like not to know some truth, in which case he lacks maximal propositional knowledge.Footnote ⁴ And if God has maximal propositional knowledge then he doesn't know what it's like not to know a truth (since he has always known all truths), in which case he lacks maximal experiential knowledge. Either way, God is not omniscient since he fails to have a required maximum of some kind of knowledge.

God knows what he needs to know

Lovering's main thesis targets God's omniscience, but there is nothing particularly problematic about his omniscience. If Lovering's main argument succeeds, it is because the concept of omniscience is incoherent, having incompatible requirements, and so no-one could be omniscient. One may then wonder whether Lovering's account of omniscience is correct. I will challenge Lovering's arguments by making three main points. The first is that his account of omniscience is insufficiently motivated. In particular, there is no reason to suppose omniscience requires maximal experiential knowledge. Recall that Lovering claims that, unless there is reason not to, we should hold that the concept of omniscience includes (the maximum of) all varieties of knowledge. An initial response grants his claim, but notes that there is reason not to include all varieties in the concept of omniscience: doing so straightforwardly makes the concept incoherent. The concept at least appears coherent prior to the inclusion, and that apparent coherence is worth preserving.

Another response to Lovering's claim challenges it. Why, in the absence of a reason not to, should we include in the concept of omniscience (the maximum of) all so-called varieties of knowledge? It is not obvious that all examples of certain varieties concern omniscience at all. Some ‘knowledge-how’, for instance, may simply concern omnipotence. Here it helps to consider some examples that motivate anti-intellectualism about knowledge-how, the view that knowledge-how and propositional knowledge are distinct. If someone doesn't know how to think rationally, or how to ride a bicycle, or how to catch a fly ball, or how to speak a language, that could simply be due to a lack of skill.Footnote ⁵ If God per impossibile could not do any of these things, that may ultimately just be a strike against his power, not against his knowledge, since God may just lack a relevant ability or disposition. At any rate, it seems that not all knowledge-how helps make one great – indeed, some knowledge-how is better not to have – and so it is implausible that omniscience requires it. Consider, for instance, knowing how to act more morally than one acts. Is knowing how to do this necessary for omniscience or for greatness? Morally imperfect beings can know how to do this, but God doesn't, and this seems to be to God's credit! God does not know how to do any better than he does, and this is no strike against his omniscience or greatness. It's plausible that neither greatness nor omniscience requires the maximum of all so-called varieties of knowledge, and so there is no reason so far to suppose that omniscience requires maximal experiential knowledge.

Lovering's second reason for supposing that omniscience requires maximal experiential knowledge is that a being who has this knowledge is conceivably greater than an otherwise similar being who lacks it, and yet God is supposed to be the greatest conceivable being. However, this at best only shows that maximal experiential knowledge is great-making, not that it should be a requirement on omniscience. The fact that the greatest conceivable being has a certain property is no reason to suppose that that property is required for omniscience. Lovering has at most shown that God must have maximal experiential knowledge, not that the concept of omniscience includes it. Still, if Lovering's ultimate goal is to challenge traditional theism, he could just revise his thesis, arguing that two properties required for maximal greatness (maximal experiential knowledge and maximal propositional knowledge) are incompatible, and so there cannot be a maximally great being, and so either God cannot be maximally great or (if part of what it is to be God is to be maximally great) there can be no God. A key question, then, is: is maximal experiential knowledge great-making?

It seems not. In fact, Lovering's problem case of experiential knowledge, that of knowing what it's like not to know a truth, supports this. It seems that, other things being equal, a subject who doesn't know what it's like not to know a truth is greater than one who does. Perhaps it only seems otherwise because additional experiential knowledge is in many cases ‘better-making’. Nonetheless it's best not to know what it's like not to know a truth. However, Lovering does go on to argue that knowing what it's like not to know a truth is great-making.Footnote ⁶ He does so on the grounds that knowing what this is like allows for one to have proper sympathy for a person ignorant of a truth, since sympathy involves the partial sharing of another's psychological state.

But without being ignorant one can share the relevant part of an ignorant person's psychological state, and so without being ignorant one can have proper sympathy for an ignorant person. To have this sympathy, it is enough to have experiential knowledge of something sufficiently similar to or related to ignorance. On many theistic views, God knows what it's like to miss out on something, or at least to be in a relationship with a being that is missing out on something. On these theistic views humankind, or at least many people, do not have a proper unity with God. Since God knows what it's like to be estranged from people, God knows what it's like to lack something that it would be better (for creatures, at least) to have. This experiential knowledge enables God to have proper sympathy for the ignorant, who lack something it would be better (for them, at least) to have. What's more, if sympathy just requires a relevant sharing of a psychological state, this requirement says nothing about how that psychological state must arise. There is no clear reason why one couldn't simply share this psychological state without sharing the experience that brings it about – one could enter into this state because of drugs, imaginative endeavours, a miracle, or what have you. And there is no clear reason why God, being omnipotent, couldn't simply will himself into this psychological state. So, God can have proper sympathy without knowing what it's like not to know a truth. There is still no reason to suppose that that bit of experiential knowledge is or is required for something that is great-making. So there is no reason to suppose that omniscience, or something else about God, requires him to know what it's like not to know a truth.Footnote ⁷ Even if Lovering goes on to show that God lacks maximal experiential knowledge or that God doesn't know what it's like not to know a truth, God's omniscience and perfection are not impugned.

My second point about Lovering's arguments can grant that omniscience, or some other feature of God, requires that God has maximal experiential knowledge or at least requires that he knows what it's like not to know a truth. The point is that, if there is something it is like not to know a truth, there is nothing qualitatively distinct about this state. That is, one can be in a state other than ignorance and still have the ‘what it's like aspect’ of ignorance. But before I argue this, why suppose in the first place that there is something it is like not to know a truth? Lovering's answer is that when he becomes aware of the fact that he doesn't know the truth-value of There is intelligent life on another planet, a ‘qualitative experiential shift occurs, one perhaps best described as from a harmony of sorts to a disharmony of sorts’, and he claims that similar cases abound, though he thinks that precisely what it's like varies from proposition to proposition.Footnote ⁸ Notice that Lovering identifies a qualitative experiential shift at the point of his becoming aware of his ignorance. This merely suggests either of two things: (1) that there is something it is like to know a truth consciously (in this case, the truth is that one doesn't know a particular truth), or (2) that there is something it is like to be consciously ignorant of a particular proposition. If (1), then so far it seems that God can know what it is like not to know a truth, since knowing what that is like is no different from knowing what it's like to know a truth. If (2), then so far there's no reason to think there is something it is like to be ignorant (simpliciter) of a proposition. It's not as if everyone who has never formed any belief about a certain proposition p, and who therefore doesn't know p and who therefore is ignorant about p, shares a distinct qualitative state! For there to be something it's like to be in a mental state, the state must at least be occurrent.Footnote ⁹ And if Lovering's example is any guide, there may be nothing it's like not to know a particular proposition, only something it is like to attend to one's lack of this knowledge.

Now, perhaps if God doesn't know what it is like to attend to this, he isn't omniscient. In order see whether God knows what this is like, it will help to get clear on the following: just what is this like? It's doubtful that there is some qualitative experience had by all and only people who are consciously ignorant of some proposition p – some people might just form a certain mental image, others might just feel distressed, and others might just feel relieved. Would it help to suppose that what it's like is person-relative? No. It's doubtful that there's even something it's distinctively like for S consciously to know that p. It's even more doubtful that S's consciously not knowing has any distinctive character. Nothing uniquely, positively describes what being in this state is like. You cannot search for a positive feeling of ignorance in order to find out whether you are consciously ignorant, in the way that you can, say, search for a positive feeling of itchiness in order to find out whether you have an itch. On some occasions, what it's like for one to be consciously ignorant of some p could feel qualitatively identical to, say, being consciously less than certain about p; and on other occasions one could have a qualitatively different experience when consciously ignorant of p.Footnote ¹⁰ There is nothing distinctive about what it's like not to know a particular proposition or about what it's like to be consciously ignorant of that proposition.

Perhaps Lovering would agree with this, and simply means that there is something it's like for one to be consciously ignorant of p, but that there is nothing distinctive about one's not knowing p. Perhaps he would grant that what it's like for one to be consciously ignorant of p is in some cases qualitatively identical to other experiences, and so one could know what it's like for one to be consciously ignorant of p by having some other experiential knowledge.

This brings me to my third point. Lovering makes the crucial ‘assumption’ that, if S knows what it's like not to know a truth, then S at some time t does not know a truth.Footnote ¹¹ This assumption allows Lovering to infer that if God knows what it's like not to know a truth, then God lacks maximal propositional knowledge, and so God is not omniscient. Lovering defends this assumption against some objections, but ultimately does not argue for it, perhaps because it is at least at first glance plausible. However (and this is my third point), the crucial assumption is false. As I have made clear, there is nothing distinctive about what it's like consciously not to know a truth (hereafter, for simplicity I drop the ‘consciously’). Now it is important to note that what it's like to know a truth is qualitatively indistinguishable from believing a proposition with high confidence (or perhaps with good enough evidence – nothing here hangs on this distinction). One can at t merely seem to know p, discover at t+1 that one doesn't in fact know p, and then at t+2 come to know p. In at least some cases, merely seeming to know is phenomenally or qualitatively indistinguishable from knowing – the experiences at t and t+2 in some case are identical. To be Gettiered with respect to p (that is, to have a mere justified true belief that p) qualitatively feels no different than to know p. If it felt any different, it would be much easier to solve the Gettier problem, that is, the unsolved problem of identifying what, besides a justified true belief, constitutes knowledge. And in many cases (especially in the sorts of cases that help motivate external-world scepticism), having a justified false belief that p feels qualitatively no different from knowing that p. For example, Smith is like us and knows he has hands, but Smith's mentally identical twin (Smith*) is massively deceived and has no hands, and so he doesn't know he has hands. But Smith and Smith* are, by stipulation, mentally identical. There can be no qualitative difference in their experiences, and so no difference between the experiences of knowing and of being justifiedly ignorant in this case.

Now, if in at least some cases what it's like to know a proposition is qualitatively indistinguishable from what it's like merely to believe a proposition with high confidence, then it is doubtful that what it's like not to know some truth differs from what it's like not to believe some proposition with high confidence. And God certainly knows what the latter is like. In particular, God doesn't believe with high confidence any falsehoods, and he presumably knows what that's like, and that's all there is to knowing what it's like not to know a truth. And so, God, or any other omniscient being, can know what it's like not to know a truth and yet always know all truths. Lovering's assumption is false, and so his conclusion that God is not omniscient is unsupported.

Lovering could in reply grant this, but then revise his crucial assumption so that it concerns a different bit of experiential knowledge: if S knows what it's like to be uncertain as to whether some proposition is true, then S at some time t does not know a truth. It would seem that Lovering could then infer that, if God has maximal experiential knowledge (or knowledge of what it's like to be uncertain about whether a proposition is true) then he lacks maximal propositional knowledge, and so God is not omniscient.

Unfortunately, even this revised assumption is doubtful. To see this, we should first try to get clear on what it is to be uncertain whether some proposition is true. Perhaps this uncertainty is just a simultaneous lack of confidence in p and in not-p. But God knows what it's like not to be confident about p, for any p that is false. This may be enough for God to know what uncertainty is like, or enough for him to know via imagination what uncertainty is like, since God is familiar with the kinds of states that are constitutive of uncertainty. Perhaps uncertainty is a state of appreciating an approximate counterbalance of evidence for and against p. Or, perhaps uncertainty is a state of having failed (after trying) to appreciate a preponderance of evidence for or against p. But God can focus on a subset of his evidence and see that the subset is counterbalanced for and against p, and then God can imagine what it's like to have no more than this subset of evidence concerning p, and thereby know what it's like to be uncertain about whether p is true. And, using only this subset of evidence, God can try and fail to appreciate a preponderance of evidence for or against p, and then God can imagine what it's like to have no more than this subset of evidence concerning p, and thereby know what uncertainty is like. We are able to engage in these sorts of imaginative endeavours and to know things as a result, so it's hard to see why God could not. In the absence of a motivated account of uncertainty that differs from the above ones, it seems that God can know what it's like to be uncertain while still always knowing all truths. The attractive revision of Lovering's crucial assumption does not help his case.

I will offer four further doubts about Lovering's original assumption and the revised assumption (though for simplicity I will just discuss the original). Lovering argues that it's not clear whether God can know what it's like not to know a truth by directly perceiving an ignorant person's mental states.Footnote ¹² After all, perhaps all God knows by this direct perception is what it's like for that person not to know a truth, and so God still fails to know what it's like not to know a truth. But it is clear enough that God can by this direct perception know what it's like not to know a truth. For, if God cannot, then an implausible scepticism about other minds seems to follow. If God can't know what a mental state is like by directly perceiving that state when another person is in it, how could ordinary people like us know anything about the mental states of others? We can't directly perceive these states. And if direct perception of these states is not enough to know what they are like, then the indirect perception available to us is not enough either. So, each person could know what her own mental states are like, but couldn't know what any mental state of any other person is like. Jones could know what it's like (for Jones) to be in pain, but she couldn't know what it's like for any other subject to be in pain. This is an unattractive and implausible outcome, one adequately avoided by granting that God can by direct perception of an ignorant person's mind know what it's like not to know a truth. And if God can do this, he can do it while knowing all truths, and so Lovering's crucial assumption is false.

A second doubt about Lovering's assumption is this. If one has propositional knowledge of the relevant facts, then, according to several popular and influential views in the philosophy of mind, when one comes to know what it's like not to know a truth, either (a) one would learn nothing new (see Churchland (Reference Churchland1989) ), or (b) one would at most gain a new recognitional and imaginative ability (see Lewis (Reference Lewis and Lewis1983) ), or (c) one would only come to know an already known fact in a new way (see Tye (Reference Tye1995) ), or (d) one would merely come to have acquaintance knowledge with a new property (see Conee (Reference Conee1994); acquaintance knowledge is knowledge of a thing, not of a fact). If any of these views is correct, then Lovering's assumption is false. One could know what it's like not to know a truth and still have maximal propositional knowledge. On (a), having maximal propositional knowledge is sufficient for knowing what it's like not to know a truth. And on (b), (c), and (d), knowing what it's like not to know a truth is compatible with maximal propositional knowledge. So Lovering's assumption requires independent reason to doubt all of these influential views in the philosophy of mind – and the disjunction of these views is endorsed by the majority of philosophers of mind – but Lovering offers none.

The final two doubts about Lovering's assumption are available only on specific or on less traditional theistic views. On Open Theism, God does not know and is not certain about some propositions concerning the future free actions of creatures, but still knows all knowable truths and is therefore omniscient. And according to Christianity, God is three persons, one of whom was incarnate and may have been ‘emptied’ of certain knowledge. The incarnate person could have experienced uncertainty and not known some truths, and so God could (as one person) know what these things are like while (as another person) knowing all truths. Any reason in favour of Open Theism or Christianity therefore counts against Lovering's assumption.