What is a logical fallacy?

A "fallacy" is a mistake, and a "logical" fallacy is a mistake in reasoning. There are, of course, other types of mistake than mistakes in reasoning. For instance, factual mistakes are sometimes referred to as "fallacies". However, the Fallacy Files is specifically concerned, not with factual errors, but with logical ones.

In logic, the term "fallacy" is used in two related, but distinct ways. For example:

1. "Argumentum ad Hominem is a fallacy."
2. "Your argument is a fallacy."

In 1, what is called a "fallacy" is a type of argument, so that a "fallacy" in this sense is a type of mistaken reasoning. In 2, it is a specific argument that is said to be a "fallacy", so that in this sense a "fallacy" is an argument which uses bad reasoning.

Clearly, these two senses are related: in 2, the argument may be called a "fallacy" because it is an instance of Argumentum ad Hominem, or some other type of fallacy. In order to keep these two senses distinct, I restrict the term "fallacy" to the first sense. For me, a fallacy is always a kind of argument.

For the second sense, I will say that a specific argument "commits" a fallacy, or is "fallacious". So, in my terminology, 2 above commits a category mistake, for there is no way that your specific argument could be a fallacy. I would say, instead:

3. "Your argument commits a fallacy. It's fallacious."

However, not just any type of mistake in reasoning counts as a logical fallacy. To be a fallacy, a type of reasoning must be potentially deceptive, it must be likely to fool at least some of the people some of the time. Moreover, in order for a fallacy to be worth identifying and naming, it must be a common type of logical error.

History

Aristotle was both the first formal logician—codifying the rules of correct reasoning—and the first informal logician—cataloging types of incorrect reasoning, namely, fallacies. He was both the first to name types of logical error, and the first to group them into categories. The result is his book On Sophistical Refutations.

However, Aristotle's teacher, Plato, deserves credit for being the first philosopher to collect examples of bad reasoning, which is an important preliminary piece of field work before naming and cataloging. Plato's "Euthydemus" preserves a collection of fallacious arguments in dialogue form, putting the perhaps exaggerated examples into the mouths of two sophists. For this reason, fallacious arguments are sometimes called "sophisms" and bad reasoning "sophistry". Aristotle refers to a few of these examples as instances of his named fallacies.

In the centuries since Plato and Aristotle, many great philosophers and logicians have contributed to fallacy studies, among them John Locke, John Stuart Mill, Jeremy Bentham, and Arthur Schopenhauer.

Why study fallacies?

Why study how to reason incorrectly; why not just study how to reason correctly? There are two reasons:

1. Even if you could count on reasoning correctly 100% of the time, you cannot count on others doing so. In logical self-defense, you need to be able to spot poor reasoning, and—more importantly—to understand it. To be able to correct others' mistakes, or to refute them convincingly, you need to understand why they are wrong.
2. Studying formal logic and the rules of correct reasoning is like having a road map that shows how to get from point A to point B. However, even the best navigators sometimes get lost, and it helps if the roads that go nowhere are clearly labeled "DEAD END", "WRONG WAY", or "DO NOT ENTER".

That is what fallacy studies is all about: marking the wrong turns that reasoners are likely to take. Thus, studying fallacies is no substitute for studying the positive principles of good reasoning—learning to navigate through logical space, so to speak. You would not set out on a trip without a road map, hoping to rely upon the "DEAD END" signs to get to your destination. Similarly, the Fallacy Files are no replacement for the study of formal and informal logic, only a supplement.