Find the error in reasoning

The reasoning model (“thinking”) is an excellent “critic” if used correctly. Ask to look for hidden assumptions, logical substitutions, and test the argument with an example. A classic trap is to confuse connection and cause (correlation and causation). Example: in the summer, both ice cream sales and the number of accidents on the water increase simultaneously. But ice cream doesn’t drown anyone—both values ​​are pulled up by a third reason, heat. Therefore, when you see “A and B grow together,” always ask: is A really the reason for B, or are they both drawn to something third? Other pitfalls are to generalize from one case and argue with a substitute thesis. First technique: ask to come up with a counterexample or an alternative explanation for the same facts. It is easy to find - the terminal is not iron-clad. The second technique is not obvious: ask the model to separate “facts” from “conclusions” in the argument. Error almost always lives in the transition between them: the facts are true, but the conclusion does not follow from them. A typical rookie mistake is to argue emotionally (“nonsense, I don’t agree”). It is more correct to name the type of error and show exactly where the logic jumped from observation to incorrect conclusion. This way criticism becomes verifiable and not a matter of taste.