Regular Arguments [CF Article]

Formal logic, by the way, means logic where the argument must be true due to its form, without needing to understand its content. E.g. if you know “If X is true, then Y is true.” and you know “X is true” then you can deduce “Y is true”. That argument doesn’t depend on what X and Y are, beyond them being propositions (statements that can be true or false). This is not the kind of argument people actually make when, e.g., debating the merits of the current U.S. president. Some people claim that’s bad and that regular people are arguing wrong, but I think the philosophers are wrong and should learn to understand regular arguments.

A particular informal argument might be wrong, but you can’t refute it just by pointing out that it isn’t deductive.

Agreed.

(bad) Summary: The vast majority of arguments that people make in normal situations are not deductive, and could not be translated into deductive arguments. Instead, they are “regular arguments,” which are incomplete / informal arguments which are intended to convey some select information to the listener to help him get to the answer on his own. E.g. a regular argument might contain an explanation and some refutations of some initial criticisms the listener might have.


I definitely agree that truly deductive arguments are rare and hard to use. A good argument to bring up against the people who think that all arguments could be turned into deduction would be to talk about mathematics:

Even in mathematics, which I’d think should be the pinnacle of deductive arguments, the arguments which actually convince people are usually regular arguments. Making these loose mathematical explanations into actual proofs is often a completely different step from understanding them, and often involves different techniques and different modes of thinking. It is usually only undertaken after the prover understands why what he’s proving is true. The final published proofs are deductive, but even still, most of them haven’t been translated into individual steps that could be checked by formal proof assistant software (doing that translation is extremely time-consuming and difficult even for simple stuff, and I have heard that it is next-to-impossibly hard for some subfields).