That's entirely wrong. The whole point of the article is that there are two distinct types of information present in all problems of this type: information abou

That's entirely wrong. The whole point of the article is that there are two distinct types of information present in all problems of this type: information about the problem (which the author calls certification, and which is assumed to be true), and then information presented in the problem, which - in almost every problem I've ever seen of this type - has statements that aren't true in it. > "If you break the assumption that you can trust all > logical statements in the problem" That is never the assumption - you are normally told explicitly that you can trust some and not others.