Let's assert the existence of an omniscient being for the sake of this question. (Questioning possibility of existence is e.g. done here.)

By its very nature, the being knows it is omniscient. It also knows that it knows etc. But is there any way for it to prove its nature and therefore proving that it is not just delusional? There is no reason to prove it for us as external observers, as we know its nature by assumption. There is no reason to prove it for itself, as it already knows by its nature, and I assume "all knowing" come with "not doubting the knowledge". But given an internal observer, e.g. a observer that does not know by assumption nor by omniscience that the being is omniscient, how - if even possible - could the being prove to the internal observer that it is omniscient and not delusional.

My thoughts: I think a proof in strict logical terms is impossible. No matter how many events the being predicts correctly, the internal observer can never conclude from the finite amount of correct predictions that all (infinitely) possible predictions would be true. The internal observer may be convinced, sure, but he cannot know. Also there is no way to tell for the internal observer if the omniscient being is lying to it (a possibility I would rule out in spirit of the question, but that can nevertheless not be perceived by the internal observer), and occasionally just guessing.

In conclusion, no matter what the omniscient being does, not only can the internal observer not know if the omniscient being is omniscient or delusional in believing to be omniscient, he can also not say if the being is just randomly answering questions which all turn out to be right (the latter one may constitute a question on Philo.SE in its own right).

I should mention I have no clue about philosophy whatsoever, besides talks among friends and getting some bits here and there from Wikipedia and this site.

  • What a coincidence. I asked myself a question "How an omniscient being can really know it is omniscient?" a few days ago. But I forgot to ask it here.
    – rus9384
    May 10, 2018 at 19:16
  • Tomorrow's sports scores or stock prices would do fine.
    – user4894
    May 10, 2018 at 19:32
  • Possibly, it can't because we don't have the possibility to know everything (yes, that depends on definition of everything).
    – rus9384
    May 10, 2018 at 22:24
  • 1
    A proof in strict logical terms of anything whatsoever is impossible, not that the sun will rise tomorrow, not that 1+1=2. One can always suspect that (s)he is scatter brained, delusional, deceived by a demon, etc. In other words, omniscience is in no worse situation than any other claim by this standard, therefore so much the worse for the standard, no knowledge is judged by it. And in practical terms systematic success at unlikely and inexplicable predictions will be convincing enough, and one will know just as we know that the sun will rise tomorrow.
    – Conifold
    May 12, 2018 at 0:07

7 Answers 7


1 The internal observer, X, could know that Y was omniscient if - which is a logical possibility - X was herself omniscient. There is nothing to rule out the existence of two omniscient beings - as in contrast there is to rule out the existence of two omnipotent beings.

2 If X could prove the existence, on whatever grounds, of a perfect being, Y, then it would follow that Y was omniscient, and X could know that Y was omniscient, since a being lacking omniscience would not be perfect.

  • I presume omniscience being part of a perfect being is not just your opinion, but comes from somewhere else. Could you point in the direction?
    – SK19
    May 10, 2018 at 16:09
  • What do you think rules out a duo of omnipotent beings that does not by extension rule out one?
    – Veedrac
    May 10, 2018 at 16:23
  • 1
    One omniscience does not rule out another : two beings could both be omnisicent. But two beings could not both be omnipotent since each would limit the other's power. Best - GT
    – Geoffrey Thomas
    May 10, 2018 at 16:33
  • 1
    @Veedrac If you're interested in a fuller treatment on why there can be at most one omnipotent being, take a look at the Argument from Omnipotence in this SEP article. May 10, 2018 at 22:39
  • 1
    @GeoffreyThomas That's exactly the kind of argument that generalizes to a single god. The cliché "can god create a rock he cannot lift?", for example. Either your idea of omnipotence is too broad to be workable or you must necessarily accept certain limits.
    – Veedrac
    May 10, 2018 at 23:29

There's a few tongue in cheek answers. The first is to point out that the question is moot: if an omniscient being was omniscient, it would know whether there was a way to prove their omniscience to a particular observer or not. And if they did know it was possible, they'd know how to do it.

The other is to approach this as a radical skeptic and ask whether you can ever know anything. Is it plausible that the omniscient being cannot prove to you that they are omniscient, because nothing can be proven to you up to the standard of proof you are demanding.

Beyond that, the question immediately points at the observer. Can the observer be convinced by anything. Is there anything at all that could convince the observer? A scientists might be convinced by a preponderance of evidence. A radically religious person might be convinced by a sign from divinity. Every individual person has their own personal set of approaches through which they may be "convinced by."

This kind of issue occurs quite often in empiricism, which is the study of what we can known from our senses. If I see a glass of water on a table, how can I know that there's actually a glass of water there. Maybe I'm just having a stroke and hallucinating. This line of reasoning is well documented, and you can see opinions on both sides of the issue.

Personally, I would resolve the situation by introducing the concept of "effectively omniscient," which is an entity which has provided sufficient evidence to the observer that the observer finds it reasonable to treat them as if they were omniscient (the general term for it is empirically adequate). Doing that lowers the bar on the proof to something which is accomplishable via empirical means without causing fundamental logical issues to arise. For example, it opens the door to using zero knowledge proofs which can reduce the chance of the omniscient being simply being "lucky" to any arbitrary amount or point out that the observer's understanding of the universe is flawed enough to make the zero knowledge proof ineffective. If their understanding of the universe is flawed enough, one may ponder whether they actually have the right concept of what "omniscient" means anyways.

  • 1
    It did not occur to me that the omniscient being would know of a proof for it's nature or lack thereof, thanks!
    – SK19
    May 10, 2018 at 19:21
  • Why is the first answer “tongue in cheek”? May 11, 2018 at 1:26
  • @JimGarrison I didn't really feel it answered the question, but rather skirted around the question. I found it to be an amusing alternative, but did want to explore the question further.
    – Cort Ammon
    May 11, 2018 at 1:34
  • Omniscience and omnipotence are inherently paradoxical, I find it amusing how much effort goes into justifying them. May 11, 2018 at 19:07

An solely omniscient being would have a harder time proving their omniscience than an all powerful being who might conjure the dead or imbue speech on, say, a squirrel. That said, if I had to prove omniscience I’d go the statistical route. For any event E, there is a finite problem space of events that can occur (eg, coin toss). For any sequence and length of coin toss(es), an omniscient being can predict them with certainty. To prove my omniscience beyond extraordinary luck, I’d set a criterion of outlandish criterion (1000/1000 correct coin toss calls).

This objective test would confound the delusional, though they may object (you weighted the coins!)

  • Do you have a reference to anyone who takes a similar view to this? It would give the reader a place to go for further information. May 11, 2018 at 12:06

Could one prove anything about oneself from within?

Observational bias means because one is the source of both the question and the answer, there is no way of determining interference, because the source is the same for both the question and answer, so by definition interference exists.

Omniscient meaning all knowing, implies there is nothing that can be known that is not known. But other than the assumption, if there is information or knowledge one does not know, you do not know that you do not know, because you do not know.

So the only way to discover one does not know something is to create another who can discover something one does not know, at which point the memory of not knowing something has been discovered so omniscience is not true.

Another problem with the idea of omniscience, is some knowledge is found through derivation and work. If anything takes work to perform to understand the conclusion, before the work is performed the outcome is not known. But with omniscience, the outcome is known beforehand. But then why do the work? Or rather as work is bound by time, nothing cannot be solved. So seeing the outcomes within time, does this mean the outcomes and viewing them deny omniscience because before the observation, the outcome was not known, and after, it was added to the bank of knowledge, admitting there was a gap to knowledge up to that point. This is why I wonder if we can ever grasp this idea, or just see shadows and challenges. And maybe nothing is actually omniscient, just nothing is not knowable. And relative to us, this is omniscience.

Another experiment. If one says, here is a question? And you spend time thinking about it and the answer comes back, did you know the answer before answering the question, or did you derive the answer by working through the question? And if you derived the answer, did it exist before you did this?

Another problem. Infinite knowledge requires infinite storage. Now we know numbers like pi do not repeat into infinity. Now an omniscient being would know all the numbers in this non-repeating group, so must be infinite in knowledge and storage capacity, at least in our universes frame of reference. So is this limitation to know everything in a finite way, the answer can be found, or derived, but is not presently existing, is this omniscience? As far as I know this within the christian biblical setting, this is omniscience.


No matter how many events the being predicts correctly, the internal observer can never conclude from the finite amount of correct predictions that all (infinitely) possible predictions would be true.

You are among friends; you have run into Hume's Problem. In short, Hume argues that simple enumeration of correct predictions offers no information about the next predicted event. After 250 years, the question remains the subject of dispute.


Howson, Colin. 2000. Hume’s problem: induction and the justification of belief. Clarendon Press, Oxford.

"Hume's Problem" < The Problem of Induction < SEP https://plato.stanford.edu/entries/induction-problem/


There can be no such thing as omniscience.any explanation and therefore any item of knowledge procedes from what is being explained (explicandum) to its attempted explanation (explanans) - which means there will always be loose ends - some of which may end up as axioms.Omniscience is a contradiction in terms, not even God can bend or break the rules of logic.


You are assuming that omniscience is a belief in what is true and asking how we can know whether our belief is justified. But true knowledge is 'knowledge by identity' for which knowledge is identical with its object. Such knowledge (and only such knowledge) has no possibility of error, as Aristotle somewhere notes. Omniscience can only be self-knowledge and this is reflected in the ancient Oracles' advice to knowledge seekers.

A complication is that where omniscience is spoken of as a possibility and human potential as in the Perennial tradition, Indian religion, Buddhism, Taoism and so forth, it is also said that at the limit the distinction between knower and known ceases to function such that omniscience would not be knowledge in our usual sense but more like Awareness or Being.

As the Upanishads ask, who is there to understand the understander or know the knower? For omniscience the knower has to become the known and the understander has to become the understood. The idea of knowledge as 'justified true belief' would be inadequate for omniscient knowledge.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .