Has any philosopher proposed the following solution to Hempel's raven paradox? The solution is that even if two statements are logically equivalent, an observation may provide evidence for one but not the other.

To put formally, we have ∀x(R(x) → B(x)) ⊣⊢ Q and strong evidential support for the proposition Q, whatever it may be, the only relevant constraint on Q being logically equivalent to ∀x(R(x) → B(x)).

Has anyone else proposed this solution? I would like to see some references.

  • 1
    This question calls for more detail. Dec 10, 2021 at 0:01
  • I don't see how it could work. If two propositions are logically equivalent and I know that they are logically equivalent, then it seems reasonable to suppose that I cannot rationally have a greater credence in one than in the other. So if evidence is understood to be that which rationally updates my degree of credence, any evidence must also apply to the two propositions equally.
    – Bumble
    Dec 10, 2021 at 4:16
  • Wikipedia's article has a long section on rejections of Hempel's logical equivalence condition (and Bayesian view of confirmation by evidence that underwrites it), with references. Could you make this more specific?
    – Conifold
    Dec 10, 2021 at 4:23
  • Any reference to try to address Hume's problem of induction may be helpful for this paradox's solutions if there's some solution... Dec 10, 2021 at 22:28

1 Answer 1


Yes, I believe Hempel’s solution can be improved with the reformulation of Nicod’s Criterion (NC) even though Hempel’s final resolution was unsatisfactory and lacked quantifiability provided by the Bayesian analysis rejecting NC or restricting its scope (Quine, 1969). Most of the solutions with the non-Hempelian view resorted to quantifying the hypothesis or redefining what constitutes an instance of confirmation (Koshy, 2017). The Bayesian solution explicitly makes a case for the irrelevance of contrapositive instances for confirmation—conflict on the assumption to establish Bayesian epistemology, which is not general enough (Clarke, 2010). The solution fails to address the core of the paradox itself (Koshy, 2017). If one approaches the same initial hypothesis, All non-black things are non-ravens (Laetz, 2010). Hempel’s inference could be more satisfied with Charvakas Epistemology and Dinnaga Rules of Inference (Tuske, 1988).

You must log in to answer this question.

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