Then there are informal truths: e.g. “the Earth is round”, “the sky is blue”, “Gala apples are red”. You can nitpick them (the Earth isn’t a perfect sphere, the sky is only blue during the day in areas without high pollution, Gala apples may be pinkish or have yellow blotches, or exceptional discoloration), endlessly or until they become formal (possibly by becoming self-referential). But in practice, these are also true (like formal truths; although it’s important to know the difference because…)

The problem is, there’s no line between an informal truth and uncertainty/opinion that isn’t true. Like you know ##FF0000 is red and ##00FF00 is not, but there’s no exact color that separates “red” and “not red” (it depends on person, mood, surroundings…) Consequently, unlike formal truths, informal truths have false implications (“fuzzy logic”). An informal truth can be phrased in a “misleading way”, priming the reader for a false implication (a formal truth can be phrased in a convoluted or unintuitive way, but interpreted formally, never leads to a false implication).

The vast majority of discussion is not formal. Even the smartest people constantly fall for false implications. And this isn’t completely solvable, because we fundamentally can’t formally define everything (too much detail): we tried with GOFAI, it failed and its successor, neural networks, informally defines things like us (by forming a lossy model of the world, then generalizing it).