One unfortunate feature I’ve noticed in arguments between logically well-trained people and the untrained is a tendency for members of the former group to point out logical errors as if they were counterarguments. This is almost totally ineffective either in changing the mind of your opponent or in convincing neutral observers. There are two main reasons for this failure.
1. Pointing out fallacies is not the same thing as urging someone to reconsider their viewpoint.
Fallacies are problematic because they’re errors in the line of reasoning that one uses to arrive at or support a conclusion. In the same way that taking the wrong route to the movie theater is bad because you won’t get there, committing a fallacy is bad because you’ll be led to the wrong conclusions.
But all that isn’t inherent in the word ‘fallacy’: the vast majority of human beings don’t understand the statement “that’s a fallacy” as “you seem to have been misled by this particular logical error – you should reevaluate your thought process and see if you arrive at the same conclusions without it.” Rather, most people will regard it as an enemy attack,with the result that they will either reject the existence of the fallacy or simply ignore it. If, by some chance, they do acknowledge the error, they’ll usually interpret it as “your argument for that conclusion is wrong – you should argue for the same conclusion in a different way.”
If you’re actually trying to convince someone (as opposed to, say, arguing to appease the goddess Eris) by showing them that the chain of logic they base their current belief on is unsound, you have to say so explicitly. Otherwise saying “fallacy” is about as effective as just telling them that they’re wrong.
2. Pointing out the obvious logical errors that fallacies characterize often obscures the deeper errors that generate the fallacies.
Take as an example the No True Scotsman fallacy. In the canonical example, the Scotsman, having seen a report of a crime, claims that no Scotsman would do such a thing. When presented with evidence of just such a Scottish criminal, he qualifies his claim, saying that no true Scotsman would do such a thing.
The obvious response to such a statement is “Ah, but you’re committing the No True Scotsman fallacy! By excluding any Scotsman who would do such a thing from your reference class, you’re making your statement tautologically true!”
While this is a valid argument, it’s not an effective one. The Scotsman, rather than changing his beliefs about the inherent goodness of all Scots, is likely to just look at you sulkily. That’s because all you’ve done is deprive him of evidence for his belief, not make him disbelieve it – wiped out one of his squadrons, so to speak, rather than making him switch sides in the war. If you were actually trying to make him change his mind, you’d have to have a better model of how it works.
No one is legitimately entranced by a fallacy like No True Scotsman – it’s used strictly as rationalization, not as a faulty but appealing reason to create a belief. Therefore the reason for his belief must lie deeper. In this case, you can find it by looking at what counts for him as evidence. To the Scotsman, the crime committed by the Englishman is an indictment of the English national character, not just the action of an individual. Likewise, a similar crime committed by a Scotsman would be evidence against the goodness of the Scottish character. Since he already believes deeply in the goodness of the Scottish character, he has only two choices: acknowledge that he was wrong about a deeply felt belief, or decide that the criminal was not really Scottish.
The error at the deepest level is that the Scotman possesses an unreasoned belief in the superiority of Scottish character, but it would be impractical at best to argue that point. The intermediate and more important error is that he views national character as monolithic – if Scottish character is better than English character, it must be better across all individuals – and therefore counts the actions of one individual as non-negligible evidence against the goodness of Scotland. If you’re trying to convince him that yes, that criminal really can be a Scotsman, the best way to do so would not be to tell him that he’s comitting a fallacy, but to argue directly against the underlying rationale connecting the individual’s crime and his nationalism. If national character is determined by, say, the ratio of good men to bad men in each nation, then bad men can exist in both England and Scotland without impinging on Scotland’s superiority – and suddenly there’s no reason for the fallacy at all. You’ve disproved his belief and changed his mind, without the word ‘fallacy’ once passing your lips.
You have to sell your idea. Some people have an immediate "ugh" reaction to the idea of selling anything in any way, but there are plenty of white-hat means to do so, e.g. 1 2. This is a completely different skill to debating. It's attracting people to your idea rather than pushing it on them. To herd cats, first work out the local value of tuna.