The moves that make an argument look stronger than it is: what each one is, how it sounds in a debate, and when the same move is fair.
A fallacy is a flaw in reasoning, not a false claim. The conclusion may even be true; the argument just doesn't get there. Most fallacies borrow the shape of a good argument (citing an expert, generalising from cases, drawing a consequence) and skip the step that would make it hold. That's why they persuade.
Name a fallacy when it carries weight in the argument, and say exactly where: quote the sentence, say which step fails, and what would fix it. Then answer the point anyway if you can. A named fallacy is the start of a reply, not the end of one. When scoring a debate, cite it against the side that leaned on it.
Every fallacy here has a fair twin: citing a real expert in their field, a survey wide enough to generalise, a consequence that truly follows. Check the "not a fallacy when" note before calling one. Rhetoric, humour and strong words aren't fallacies by themselves, and showing an argument is fallacious doesn't make its conclusion false; assuming it does is a fallacy of its own.
Attacking the person instead of the argument.
"Of course he'd say that, he never went to a proper university", and the evidence goes unanswered.
Not a fallacy when the person is the question: a narrator's reliability in hadith criticism, or a witness with a stake in what they testify.
Misrepresenting the other side, then refuting that.
One says "this hadith has a weak chain"; the other answers "so you reject the whole Sunnah!"
A fair summary isn't a straw man, nor is drawing a consequence that really follows from what they said. The test: would they accept it as their view?
Cutting a quote so it says something else.
Reading "woe to those who pray" and stopping there, as if the verse condemned prayer.
A short quote is fine as long as it keeps the meaning; check it against the full text.
Changing the subject to avoid the question.
Asked for evidence that a report is authentic, he talks about the other side's mistakes in an earlier debate.
A digression that comes back and serves the point isn't one, nor is objecting that a question falls outside the agreed topic.
Assuming the conclusion in the premises.
"This book is true because it says it is true."
Being consistent isn't circular, nor is citing a text to someone who accepts it. The fallacy is using the conclusion against someone who disputes it.
Asking the other side to disprove a claim.
"Prove to me this event didn't happen, or it stands."
Whoever claims must prove, and that includes someone who claims to know a negative. Debaters may also agree in the conditions on who proves first.
Raising the bar once the evidence is given.
He asked for one source; given it, he wanted an original manuscript; given that, a contemporary witness.
Asking about the evidence just given isn't moving the goalposts. The fallacy is changing the condition after it was met.
Presenting two options when there are more.
"Either you accept everything in this book, or you reject religion altogether."
An either/or is sound when the options really are exhaustive: a whole number is even or odd.
Concluding about all from a few cases.
"I met three followers of this school who couldn't cite their evidence, so they're all ignorant."
A broad survey or a representative sample is good evidence, and a single counterexample is enough to refute a universal claim.
Relying on who said it instead of the evidence.
"A famous physicist said so, so the question of law is settled."
Citing experts in their own field, or an authority the other side accepts, is fair. The fallacy is an authority outside their field, or a name in place of evidence that was asked for.
Treating a claim as true because many hold it.
"A billion people believe this; they can't all be wrong."
When the question is what people do, such as custom or what a word means, numbers are evidence. Consensus, for those who hold it binding, has its own conditions; it isn't a head count.
Claiming one step must lead to a chain of outcomes.
"If we allow criticism of this report, all hadith will go, then the Quran, then nothing is left."
When each link is shown to follow from the one before, it's a valid argument, as in blocking the means (sadd al-dhara'i) for those who apply it, with its conditions.
A question that assumes something unproven.
"When will you admit your sources were altered?"
A question built on something both sides granted, or that was shown earlier, isn't loaded. The answer to a loaded one is to reject its premise, not to say yes or no.
Answering a criticism by accusing the critic.
Asked about a contradiction in his school, he says "yours has more".
Showing the critic applies a double standard, rejecting here what they accept elsewhere, is a fair point (an ilzam). It still doesn't answer the objection itself.
Rattling off many claims too fast to answer.
In a five-minute turn, he lists twenty objections without developing any, then counts every one left unanswered as conceded.
Several independent arguments are fine when each is developed enough to be answered. That's why some debate conditions allow one point per turn.
