Categories of Logical Fallacies

Core Idea

Fallacies are reasoning errors—arguments appearing plausible that fail to support conclusions. Major categories: fallacies of relevance (premises don't address the conclusion), fallacies of weak induction (premises insufficiently support conclusion), and fallacies of ambiguity (unclear language masks failed reasoning).

Explainer

You already know from arguments-premises-and-conclusions that a good argument has two requirements: the premises must be relevant to the conclusion, and they must provide adequate support for it. The taxonomy of fallacies maps directly onto these requirements. When an argument fails because its premises are irrelevant — they don't address the conclusion at all — it commits a fallacy of relevance. When the premises are related to the conclusion but don't provide enough support for it — the inference is too weak — it commits a fallacy of weak induction. A third category, fallacies of ambiguity, captures failures where unclear or shifting language creates the illusion of valid reasoning.

Fallacies of relevance are often emotionally compelling even though they're logically beside the point. The *ad hominem* attacks the person making the argument rather than the argument itself — the speaker's character or motives are irrelevant to whether their premises are true. The *appeal to inappropriate authority* cites an impressive-sounding source that lacks expertise on the specific question at hand. The *straw man* attacks a distorted, weakened version of the opponent's position — it looks like refutation, but the actual position has not been engaged. The diagnostic question for relevance fallacies is always: do the premises actually address the conclusion, or are they changing the subject to something easier to attack?

Fallacies of weak induction have premises that are genuinely relevant to the conclusion but don't do enough evidential work. *Hasty generalization* draws a broad conclusion from too small or unrepresentative a sample — the premises do connect to the conclusion, but the evidence is insufficient. *False cause* (post hoc, ergo propter hoc) infers causation from temporal succession alone — correlation is relevant evidence of causation, but correlation alone is far from sufficient. *Appeal to ignorance* treats the absence of disconfirming evidence as positive evidence — the connection to the conclusion is real but the support is too thin. Recognizing weak induction requires shifting from the question "does this connect?" to "how much support does this actually provide?"

Fallacies of ambiguity exploit multiple meanings or grammatical vagueness in the premises. *Equivocation* uses the same word in two different senses across the argument, making an invalid inference appear valid — for example, using "law" to mean both "physical law" and "legal statute" within a single argument. *Amphiboly* arises from structural ambiguity in a sentence that makes the premise's meaning indeterminate. These are particularly insidious because the argument may be formally valid if the meaning stays fixed — the failure is that the meaning quietly shifts. Spotting ambiguity fallacies requires asking: if I pin down a single, consistent meaning for each key term, does the argument still go through?