The knower paradox concerns a sentence such as, "This sentence is unknown." Now liar sentences can be paired with honest sentences, e.g. "This sentence is true." So suppose there was a sentence that goes, "This sentence is known."

Curry's paradox can involve conditional sentences whose antecedents resemble honest sentences. Is it possible to fuse an honest-knower sentence with the informal structure of a Curry sentence to yield an additional enigmatic sentence? What I mean is, can a paradox be derived from, "If this sentence is known, then φ," where φ is an arbitrary sentence/logical explosion?

Or consider liar cycles such as

A. Sentence B is true.

B. Sentence A is false.

What happens if we switch out "is true/false" for "is known/unknown"? Do we end up with another species of knower paradox?

I also wonder about Yablo's paradox, here. Generally, can we expand upon the knower paradox by combining it with other paradoxes?

  • To engage the mechanism of Curry's paradox you need something weaker than "true", like "meaningful", not stronger like "known". The secret is that x="if x is meaningful, then φ" together with the conditional entails "x is true" by the convention T. Since "x is true" implies "x is meaningful" (presumably) you then get φ by modus ponens. But "x is true" implying "x is known" is not very plausible except to radical anti-realists.
    – Conifold
    Apr 6, 2021 at 5:21
  • 1
    On the relation between the Yablo paradox and the knower paradox, there is a nice paper by Roy Sorensen, 'Yablo's paradox and kindred infinite liars'.
    – sequitur
    Apr 6, 2021 at 22:23

1 Answer 1


Fix your favorite sentence φ and let x be the sentence "If x is known, then φ." In order for x to be false, its hypothesis "x is known" would have to be true, which means a fortiori that x would have to be true. This constitutes a proof of x, and so we know x - from which we get φ.

This is exactly the same "shape" as the argument around the original knower paradox, so if we make the same assumptions we also get a paradox in this case. However, for what it's worth I've never found the relevant assumptions convincing at all, so I don't find the knower paradox (or this Curry-knower paradox) particularly compelling.

An interesting sub-question is whether we could get a more "parsimonious" paradox this way - that is, a paradox which doesn't require us to assume as much about knowability. I don't think that happens, though, since we need a reason why "true-but-unknown" should be problematic. (Keep in mind that in the setting of Godel's theorem "true-but-unprovable" is completely unproblematic, so this is a serious option.) To get any mileage out of this we need a rather strong connection between "is known" and "is true," and this is exactly what (PK)/(KF) give us with the classical knower paradox.

  • So we basically just end up with the same paradoxes (however compelling or not) that we started out with, just conjoined? Apr 6, 2021 at 0:24
  • 1
    @KristianBerry As far as I can tell, yes. To be fair it's hard to rigorously prove a result like that. That said, the Godel point I raise tries to do this: unless we make assumptions on knowability which distinguish it from provability, we're guaranteed (under mild assumptions) to never hit a contradiction since arithmetic itself doesn't break. This shows that in order to get any kind of knowability paradox we need strong rules connecting knowability and truth ... at which point a "truth-to-knowledge shift" doesn't seem particularly exciting anymore. (But that's all heuristic of course.) Apr 6, 2021 at 0:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.