# There Is No Problem Of Induction You’ll often hear that “Popper refuted induction” or “Popper solved the problem of induction” or that “Popper showed induction is false” or “Popper showed induction is a myth” or a hundred other variants of this claim. While all sorta-true, phrasing like this is liable to conflate concepts like _logical entailment_, _reasoning_, and _justifiable assumptions_ into a big muddy mess, as the emphasized sentences in the above quotation highlight. Forget that all important adjective “logical”, and you’re just a few steps away from thinking Popper disproved reasoning itself. To avoid confusion, I suggest a better way to think about it is: Popper realized there _was no problem of induction_, and that the conjectural-deductive approach to knowledge was just fine: > Guess a general law $G$, deduce the logical consequences $E$, and check to see if they are falsified by experiment, argument, or proof. In symbols, Popper realized that while it is impossible for truth to flow from specific to general, $(...e_{i−1} ∧ e_{i} ∧ e_{i+1}...) ⊬ G$      what _is_ possible is for _falsity_ to flow from specific to general: $(...e_{i−1} ∧ ¬e_i ∧ e_{i+1}...)⊢¬G$ and that this is the basis for all science. --- Date: 20250430 Links to: [Logic](Logic.md) Tags: References: * [The Problem of Induction and Machine Learning \| Vaden Masrani](https://vmasrani.github.io/blog/2021/problem-of-induction/)